Merge with /pub/scm/linux/kernel/git/torvalds/linux-2.6.git

Signed-off-by: Steve French <sfrench@us.ibm.com>

476 files changed, 14200 insertions, 10528 deletions

diff --git a/fs/9p/9p.h b/fs/9p/9p.h
index 0cd374d94717..94e2f92ab2e8 100644
--- a/fs/9p/9p.h
+++ b/fs/9p/9p.h
@@ -8,9 +8,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -235,6 +234,7 @@ struct Tcreate {
 	struct v9fs_str name;
 	u32 perm;
 	u8 mode;
+	struct v9fs_str extension;
 };
 
 struct Rcreate {
@@ -348,8 +348,6 @@ int v9fs_t_attach(struct v9fs_session_info *v9ses, char *uname, char *aname,
 
 int v9fs_t_clunk(struct v9fs_session_info *v9ses, u32 fid);
 
-int v9fs_t_flush(struct v9fs_session_info *v9ses, u16 oldtag);
-
 int v9fs_t_stat(struct v9fs_session_info *v9ses, u32 fid,
 		struct v9fs_fcall **rcall);
 
@@ -366,7 +364,7 @@ int v9fs_t_remove(struct v9fs_session_info *v9ses, u32 fid,
 		  struct v9fs_fcall **rcall);
 
 int v9fs_t_create(struct v9fs_session_info *v9ses, u32 fid, char *name,
-		  u32 perm, u8 mode, struct v9fs_fcall **rcall);
+	u32 perm, u8 mode, char *extension, struct v9fs_fcall **rcall);
 
 int v9fs_t_read(struct v9fs_session_info *v9ses, u32 fid,
 		u64 offset, u32 count, struct v9fs_fcall **rcall);
@@ -374,3 +372,4 @@ int v9fs_t_read(struct v9fs_session_info *v9ses, u32 fid,
 int v9fs_t_write(struct v9fs_session_info *v9ses, u32 fid, u64 offset,
 		 u32 count, const char __user * data,
 		 struct v9fs_fcall **rcall);
+int v9fs_printfcall(char *, int, struct v9fs_fcall *, int);
diff --git a/fs/9p/Makefile b/fs/9p/Makefile
index 2f4ce43f7b6c..87897f84dfb6 100644
--- a/fs/9p/Makefile
+++ b/fs/9p/Makefile
@@ -1,10 +1,9 @@
-obj-$(CONFIG_9P_FS) := 9p2000.o
+obj-$(CONFIG_9P_FS) := 9p.o
 
-9p2000-objs := \
+9p-objs := \
 	trans_fd.o \
-	trans_sock.o \
 	mux.o \
-	9p.o \
+	fcall.o \
 	conv.o \
 	vfs_super.o \
 	vfs_inode.o \
@@ -14,5 +13,6 @@ obj-$(CONFIG_9P_FS) := 9p2000.o
 	vfs_dentry.o \
 	error.o \
 	v9fs.o \
-	fid.o
+	fid.o \
+	fcprint.o
 
diff --git a/fs/9p/conv.c b/fs/9p/conv.c
index bf1f10067960..a767e05b60bf 100644
--- a/fs/9p/conv.c
+++ b/fs/9p/conv.c
@@ -8,9 +8,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -536,6 +535,7 @@ struct v9fs_fcall *v9fs_create_tversion(u32 msize, char *version)
 	return fc;
 }
 
+#if 0
 struct v9fs_fcall *v9fs_create_tauth(u32 afid, char *uname, char *aname)
 {
 	int size;
@@ -559,6 +559,7 @@ struct v9fs_fcall *v9fs_create_tauth(u32 afid, char *uname, char *aname)
       error:
 	return fc;
 }
+#endif  /*  0  */
 
 struct v9fs_fcall *
 v9fs_create_tattach(u32 fid, u32 afid, char *uname, char *aname)
@@ -664,7 +665,8 @@ struct v9fs_fcall *v9fs_create_topen(u32 fid, u8 mode)
 	return fc;
 }
 
-struct v9fs_fcall *v9fs_create_tcreate(u32 fid, char *name, u32 perm, u8 mode)
+struct v9fs_fcall *v9fs_create_tcreate(u32 fid, char *name, u32 perm, u8 mode,
+	char *extension, int extended)
 {
 	int size;
 	struct v9fs_fcall *fc;
@@ -672,6 +674,9 @@ struct v9fs_fcall *v9fs_create_tcreate(u32 fid, char *name, u32 perm, u8 mode)
 	struct cbuf *bufp = &buffer;
 
 	size = 4 + 2 + strlen(name) + 4 + 1;	/* fid[4] name[s] perm[4] mode[1] */
+	if (extended && extension!=NULL)
+		size += 2 + strlen(extension);	/* extension[s] */
+
 	fc = v9fs_create_common(bufp, size, TCREATE);
 	if (IS_ERR(fc))
 		goto error;
@@ -680,6 +685,8 @@ struct v9fs_fcall *v9fs_create_tcreate(u32 fid, char *name, u32 perm, u8 mode)
 	v9fs_put_str(bufp, name, &fc->params.tcreate.name);
 	v9fs_put_int32(bufp, perm, &fc->params.tcreate.perm);
 	v9fs_put_int8(bufp, mode, &fc->params.tcreate.mode);
+	if (extended)
+		v9fs_put_str(bufp, extension, &fc->params.tcreate.extension);
 
 	if (buf_check_overflow(bufp)) {
 		kfree(fc);
diff --git a/fs/9p/conv.h b/fs/9p/conv.h
index 26a736e4a2e7..dd5b6b1b610f 100644
--- a/fs/9p/conv.h
+++ b/fs/9p/conv.h
@@ -8,9 +8,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -33,14 +32,14 @@ int v9fs_deserialize_fcall(void *buf, u32 buflen, struct v9fs_fcall *rcall,
 void v9fs_set_tag(struct v9fs_fcall *fc, u16 tag);
 
 struct v9fs_fcall *v9fs_create_tversion(u32 msize, char *version);
-struct v9fs_fcall *v9fs_create_tauth(u32 afid, char *uname, char *aname);
 struct v9fs_fcall *v9fs_create_tattach(u32 fid, u32 afid, char *uname,
 	char *aname);
 struct v9fs_fcall *v9fs_create_tflush(u16 oldtag);
 struct v9fs_fcall *v9fs_create_twalk(u32 fid, u32 newfid, u16 nwname,
 	char **wnames);
 struct v9fs_fcall *v9fs_create_topen(u32 fid, u8 mode);
-struct v9fs_fcall *v9fs_create_tcreate(u32 fid, char *name, u32 perm, u8 mode);
+struct v9fs_fcall *v9fs_create_tcreate(u32 fid, char *name, u32 perm, u8 mode,
+	char *extension, int extended);
 struct v9fs_fcall *v9fs_create_tread(u32 fid, u64 offset, u32 count);
 struct v9fs_fcall *v9fs_create_twrite(u32 fid, u64 offset, u32 count,
 	const char __user *data);
diff --git a/fs/9p/debug.h b/fs/9p/debug.h
index fe551032788b..4228c0bb3c32 100644
--- a/fs/9p/debug.h
+++ b/fs/9p/debug.h
@@ -5,9 +5,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -30,6 +29,7 @@
 #define DEBUG_MUX		(1<<5)
 #define DEBUG_TRANS		(1<<6)
 #define DEBUG_SLABS	      	(1<<7)
+#define DEBUG_FCALL		(1<<8)
 
 #define DEBUG_DUMP_PKT		0
 
diff --git a/fs/9p/error.c b/fs/9p/error.c
index e4b6f8f38b6f..981fe8ecd780 100644
--- a/fs/9p/error.c
+++ b/fs/9p/error.c
@@ -11,9 +11,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
diff --git a/fs/9p/error.h b/fs/9p/error.h
index a9794e85fe51..5f3ca522b316 100644
--- a/fs/9p/error.h
+++ b/fs/9p/error.h
@@ -12,9 +12,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
diff --git a/fs/9p/9p.c b/fs/9p/fcall.c
index f86a28d1d6a6..71742ba150c4 100644
--- a/fs/9p/9p.c
+++ b/fs/9p/fcall.c
@@ -1,5 +1,5 @@
 /*
- *  linux/fs/9p/9p.c
+ *  linux/fs/9p/fcall.c
  *
  *  This file contains functions to perform synchronous 9P calls
  *
@@ -8,9 +8,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -148,13 +147,13 @@ v9fs_t_clunk(struct v9fs_session_info *v9ses, u32 fid)
 	return ret;
 }
 
+#if 0
 /**
  * v9fs_v9fs_t_flush - flush a pending transaction
  * @v9ses: 9P2000 session information
- * @tag: tid to release
+ * @tag: tag to release
  *
  */
-
 int v9fs_t_flush(struct v9fs_session_info *v9ses, u16 oldtag)
 {
 	int ret;
@@ -171,6 +170,7 @@ int v9fs_t_flush(struct v9fs_session_info *v9ses, u16 oldtag)
 
 	return ret;
 }
+#endif
 
 /**
  * v9fs_t_stat - read a file's meta-data
@@ -332,8 +332,8 @@ v9fs_t_remove(struct v9fs_session_info *v9ses, u32 fid,
  */
 
 int
-v9fs_t_create(struct v9fs_session_info *v9ses, u32 fid, char *name,
-	      u32 perm, u8 mode, struct v9fs_fcall **rcp)
+v9fs_t_create(struct v9fs_session_info *v9ses, u32 fid, char *name, u32 perm,
+	u8 mode, char *extension, struct v9fs_fcall **rcp)
 {
 	int ret;
 	struct v9fs_fcall *tc;
@@ -341,7 +341,9 @@ v9fs_t_create(struct v9fs_session_info *v9ses, u32 fid, char *name,
 	dprintk(DEBUG_9P, "fid %d name '%s' perm %x mode %d\n",
 		fid, name, perm, mode);
 
-	tc = v9fs_create_tcreate(fid, name, perm, mode);
+	tc = v9fs_create_tcreate(fid, name, perm, mode, extension,
+		v9ses->extended);
+
 	if (!IS_ERR(tc)) {
 		ret = v9fs_mux_rpc(v9ses->mux, tc, rcp);
 		kfree(tc);
diff --git a/fs/9p/fcprint.c b/fs/9p/fcprint.c
new file mode 100644
index 000000000000..583e827baebd
--- /dev/null
+++ b/fs/9p/fcprint.c
@@ -0,0 +1,346 @@
+/*
+ *  linux/fs/9p/fcprint.c
+ *
+ *  Print 9P call.
+ *
+ *  Copyright (C) 2005 by Latchesar Ionkov <lucho@ionkov.net>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to:
+ *  Free Software Foundation
+ *  51 Franklin Street, Fifth Floor
+ *  Boston, MA  02111-1301  USA
+ *
+ */
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/errno.h>
+#include <linux/fs.h>
+#include <linux/idr.h>
+
+#include "debug.h"
+#include "v9fs.h"
+#include "9p.h"
+#include "mux.h"
+
+static int
+v9fs_printqid(char *buf, int buflen, struct v9fs_qid *q)
+{
+	int n;
+	char b[10];
+
+	n = 0;
+	if (q->type & V9FS_QTDIR)
+		b[n++] = 'd';
+	if (q->type & V9FS_QTAPPEND)
+		b[n++] = 'a';
+	if (q->type & V9FS_QTAUTH)
+		b[n++] = 'A';
+	if (q->type & V9FS_QTEXCL)
+		b[n++] = 'l';
+	if (q->type & V9FS_QTTMP)
+		b[n++] = 't';
+	if (q->type & V9FS_QTSYMLINK)
+		b[n++] = 'L';
+	b[n] = '\0';
+
+	return scnprintf(buf, buflen, "(%.16llx %x %s)", (long long int) q->path,
+		q->version, b);
+}
+
+static int
+v9fs_printperm(char *buf, int buflen, int perm)
+{
+	int n;
+	char b[15];
+
+	n = 0;
+	if (perm & V9FS_DMDIR)
+		b[n++] = 'd';
+	if (perm & V9FS_DMAPPEND)
+		b[n++] = 'a';
+	if (perm & V9FS_DMAUTH)
+		b[n++] = 'A';
+	if (perm & V9FS_DMEXCL)
+		b[n++] = 'l';
+	if (perm & V9FS_DMTMP)
+		b[n++] = 't';
+	if (perm & V9FS_DMDEVICE)
+		b[n++] = 'D';
+	if (perm & V9FS_DMSOCKET)
+		b[n++] = 'S';
+	if (perm & V9FS_DMNAMEDPIPE)
+		b[n++] = 'P';
+	if (perm & V9FS_DMSYMLINK)
+		b[n++] = 'L';
+	b[n] = '\0';
+
+	return scnprintf(buf, buflen, "%s%03o", b, perm&077);
+}
+
+static int
+v9fs_printstat(char *buf, int buflen, struct v9fs_stat *st, int extended)
+{
+	int n;
+
+	n = scnprintf(buf, buflen, "'%.*s' '%.*s'", st->name.len,
+		st->name.str, st->uid.len, st->uid.str);
+	if (extended)
+		n += scnprintf(buf+n, buflen-n, "(%d)", st->n_uid);
+
+	n += scnprintf(buf+n, buflen-n, " '%.*s'", st->gid.len, st->gid.str);
+	if (extended)
+		n += scnprintf(buf+n, buflen-n, "(%d)", st->n_gid);
+
+	n += scnprintf(buf+n, buflen-n, " '%.*s'", st->muid.len, st->muid.str);
+	if (extended)
+		n += scnprintf(buf+n, buflen-n, "(%d)", st->n_muid);
+
+	n += scnprintf(buf+n, buflen-n, " q ");
+	n += v9fs_printqid(buf+n, buflen-n, &st->qid);
+	n += scnprintf(buf+n, buflen-n, " m ");
+	n += v9fs_printperm(buf+n, buflen-n, st->mode);
+	n += scnprintf(buf+n, buflen-n, " at %d mt %d l %lld",
+		st->atime, st->mtime, (long long int) st->length);
+
+	if (extended)
+		n += scnprintf(buf+n, buflen-n, " ext '%.*s'",
+			st->extension.len, st->extension.str);
+
+	return n;
+}
+
+static int
+v9fs_dumpdata(char *buf, int buflen, u8 *data, int datalen)
+{
+	int i, n;
+
+	i = n = 0;
+	while (i < datalen) {
+		n += scnprintf(buf + n, buflen - n, "%02x", data[i]);
+		if (i%4 == 3)
+			n += scnprintf(buf + n, buflen - n, " ");
+		if (i%32 == 31)
+			n += scnprintf(buf + n, buflen - n, "\n");
+
+		i++;
+	}
+	n += scnprintf(buf + n, buflen - n, "\n");
+
+	return n;
+}
+
+static int
+v9fs_printdata(char *buf, int buflen, u8 *data, int datalen)
+{
+	return v9fs_dumpdata(buf, buflen, data, datalen<16?datalen:16);
+}
+
+int
+v9fs_printfcall(char *buf, int buflen, struct v9fs_fcall *fc, int extended)
+{
+	int i, ret, type, tag;
+
+	if (!fc)
+		return scnprintf(buf, buflen, "<NULL>");
+
+	type = fc->id;
+	tag = fc->tag;
+
+	ret = 0;
+	switch (type) {
+	case TVERSION:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Tversion tag %u msize %u version '%.*s'", tag,
+			fc->params.tversion.msize, fc->params.tversion.version.len,
+			fc->params.tversion.version.str);
+		break;
+
+	case RVERSION:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Rversion tag %u msize %u version '%.*s'", tag,
+			fc->params.rversion.msize, fc->params.rversion.version.len,
+			fc->params.rversion.version.str);
+		break;
+
+	case TAUTH:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Tauth tag %u afid %d uname '%.*s' aname '%.*s'", tag,
+			fc->params.tauth.afid, fc->params.tauth.uname.len,
+			fc->params.tauth.uname.str, fc->params.tauth.aname.len,
+			fc->params.tauth.aname.str);
+		break;
+
+	case RAUTH:
+		ret += scnprintf(buf+ret, buflen-ret, "Rauth tag %u qid ", tag);
+		v9fs_printqid(buf+ret, buflen-ret, &fc->params.rauth.qid);
+		break;
+
+	case TATTACH:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Tattach tag %u fid %d afid %d uname '%.*s' aname '%.*s'",
+			tag, fc->params.tattach.fid, fc->params.tattach.afid,
+			fc->params.tattach.uname.len, fc->params.tattach.uname.str,
+			fc->params.tattach.aname.len, fc->params.tattach.aname.str);
+		break;
+
+	case RATTACH:
+		ret += scnprintf(buf+ret, buflen-ret, "Rattach tag %u qid ", tag);
+		v9fs_printqid(buf+ret, buflen-ret, &fc->params.rattach.qid);
+		break;
+
+	case RERROR:
+		ret += scnprintf(buf+ret, buflen-ret, "Rerror tag %u ename '%.*s'",
+			tag, fc->params.rerror.error.len,
+			fc->params.rerror.error.str);
+		if (extended)
+			ret += scnprintf(buf+ret, buflen-ret, " ecode %d\n",
+				fc->params.rerror.errno);
+		break;
+
+	case TFLUSH:
+		ret += scnprintf(buf+ret, buflen-ret, "Tflush tag %u oldtag %u",
+			tag, fc->params.tflush.oldtag);
+		break;
+
+	case RFLUSH:
+		ret += scnprintf(buf+ret, buflen-ret, "Rflush tag %u", tag);
+		break;
+
+	case TWALK:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Twalk tag %u fid %d newfid %d nwname %d", tag,
+			fc->params.twalk.fid, fc->params.twalk.newfid,
+			fc->params.twalk.nwname);
+		for(i = 0; i < fc->params.twalk.nwname; i++)
+			ret += scnprintf(buf+ret, buflen-ret," '%.*s'",
+				fc->params.twalk.wnames[i].len,
+				fc->params.twalk.wnames[i].str);
+		break;
+
+	case RWALK:
+		ret += scnprintf(buf+ret, buflen-ret, "Rwalk tag %u nwqid %d",
+			tag, fc->params.rwalk.nwqid);
+		for(i = 0; i < fc->params.rwalk.nwqid; i++)
+			ret += v9fs_printqid(buf+ret, buflen-ret,
+				&fc->params.rwalk.wqids[i]);
+		break;
+
+	case TOPEN:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Topen tag %u fid %d mode %d", tag,
+			fc->params.topen.fid, fc->params.topen.mode);
+		break;
+
+	case ROPEN:
+		ret += scnprintf(buf+ret, buflen-ret, "Ropen tag %u", tag);
+		ret += v9fs_printqid(buf+ret, buflen-ret, &fc->params.ropen.qid);
+		ret += scnprintf(buf+ret, buflen-ret," iounit %d",
+			fc->params.ropen.iounit);
+		break;
+
+	case TCREATE:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Tcreate tag %u fid %d name '%.*s' perm ", tag,
+			fc->params.tcreate.fid, fc->params.tcreate.name.len,
+			fc->params.tcreate.name.str);
+
+		ret += v9fs_printperm(buf+ret, buflen-ret, fc->params.tcreate.perm);
+		ret += scnprintf(buf+ret, buflen-ret, " mode %d",
+			fc->params.tcreate.mode);
+		break;
+
+	case RCREATE:
+		ret += scnprintf(buf+ret, buflen-ret, "Rcreate tag %u", tag);
+		ret += v9fs_printqid(buf+ret, buflen-ret, &fc->params.rcreate.qid);
+		ret += scnprintf(buf+ret, buflen-ret, " iounit %d",
+			fc->params.rcreate.iounit);
+		break;
+
+	case TREAD:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Tread tag %u fid %d offset %lld count %u", tag,
+			fc->params.tread.fid,
+			(long long int) fc->params.tread.offset,
+			fc->params.tread.count);
+		break;
+
+	case RREAD:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Rread tag %u count %u data ", tag,
+			fc->params.rread.count);
+		ret += v9fs_printdata(buf+ret, buflen-ret, fc->params.rread.data,
+			fc->params.rread.count);
+		break;
+
+	case TWRITE:
+		ret += scnprintf(buf+ret, buflen-ret,
+			"Twrite tag %u fid %d offset %lld count %u data ",
+			tag, fc->params.twrite.fid,
+			(long long int) fc->params.twrite.offset,
+			fc->params.twrite.count);
+		ret += v9fs_printdata(buf+ret, buflen-ret, fc->params.twrite.data,
+			fc->params.twrite.count);
+		break;
+
+	case RWRITE:
+		ret += scnprintf(buf+ret, buflen-ret, "Rwrite tag %u count %u",
+			tag, fc->params.rwrite.count);
+		break;
+
+	case TCLUNK:
+		ret += scnprintf(buf+ret, buflen-ret, "Tclunk tag %u fid %d",
+			tag, fc->params.tclunk.fid);
+		break;
+
+	case RCLUNK:
+		ret += scnprintf(buf+ret, buflen-ret, "Rclunk tag %u", tag);
+		break;
+
+	case TREMOVE:
+		ret += scnprintf(buf+ret, buflen-ret, "Tremove tag %u fid %d",
+			tag, fc->params.tremove.fid);
+		break;
+
+	case RREMOVE:
+		ret += scnprintf(buf+ret, buflen-ret, "Rremove tag %u", tag);
+		break;
+
+	case TSTAT:
+		ret += scnprintf(buf+ret, buflen-ret, "Tstat tag %u fid %d",
+			tag, fc->params.tstat.fid);
+		break;
+
+	case RSTAT:
+		ret += scnprintf(buf+ret, buflen-ret, "Rstat tag %u ", tag);
+		ret += v9fs_printstat(buf+ret, buflen-ret, &fc->params.rstat.stat,
+			extended);
+		break;
+
+	case TWSTAT:
+		ret += scnprintf(buf+ret, buflen-ret, "Twstat tag %u fid %d ",
+			tag, fc->params.twstat.fid);
+		ret += v9fs_printstat(buf+ret, buflen-ret, &fc->params.twstat.stat,
+			extended);
+		break;
+
+	case RWSTAT:
+		ret += scnprintf(buf+ret, buflen-ret, "Rwstat tag %u", tag);
+		break;
+
+	default:
+		ret += scnprintf(buf+ret, buflen-ret, "unknown type %d", type);
+		break;
+	}
+
+	return ret;
+}
diff --git a/fs/9p/fid.c b/fs/9p/fid.c
index c4d13bf904d2..b7608af07ce8 100644
--- a/fs/9p/fid.c
+++ b/fs/9p/fid.c
@@ -4,9 +4,8 @@
  *  Copyright (C) 2005, 2006 by Eric Van Hensbergen <ericvh@gmail.com>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
diff --git a/fs/9p/fid.h b/fs/9p/fid.h
index 1fc2dd08d75a..aa974d6875c3 100644
--- a/fs/9p/fid.h
+++ b/fs/9p/fid.h
@@ -4,9 +4,8 @@
  *  Copyright (C) 2005 by Eric Van Hensbergen <ericvh@gmail.com>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
diff --git a/fs/9p/mux.c b/fs/9p/mux.c
index ea1134eb47c8..3e5b124a7212 100644
--- a/fs/9p/mux.c
+++ b/fs/9p/mux.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2004-2005 by Latchesar Ionkov <lucho@ionkov.net>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -31,6 +30,7 @@
 #include <linux/poll.h>
 #include <linux/kthread.h>
 #include <linux/idr.h>
+#include <linux/mutex.h>
 
 #include "debug.h"
 #include "v9fs.h"
@@ -69,7 +69,7 @@ struct v9fs_mux_data {
 	int msize;
 	unsigned char *extended;
 	struct v9fs_transport *trans;
-	struct v9fs_idpool tidpool;
+	struct v9fs_idpool tagpool;
 	int err;
 	wait_queue_head_t equeue;
 	struct list_head req_list;
@@ -110,7 +110,7 @@ static void v9fs_pollwait(struct file *filp, wait_queue_head_t * wait_address,
 static u16 v9fs_mux_get_tag(struct v9fs_mux_data *);
 static void v9fs_mux_put_tag(struct v9fs_mux_data *, u16);
 
-static DECLARE_MUTEX(v9fs_mux_task_lock);
+static DEFINE_MUTEX(v9fs_mux_task_lock);
 static struct workqueue_struct *v9fs_mux_wq;
 
 static int v9fs_mux_num;
@@ -142,7 +142,7 @@ void v9fs_mux_global_exit(void)
  *
  * The current implementation returns sqrt of the number of mounts.
  */
-inline int v9fs_mux_calc_poll_procs(int muxnum)
+static int v9fs_mux_calc_poll_procs(int muxnum)
 {
 	int n;
 
@@ -166,7 +166,7 @@ static int v9fs_mux_poll_start(struct v9fs_mux_data *m)
 
 	dprintk(DEBUG_MUX, "mux %p muxnum %d procnum %d\n", m, v9fs_mux_num,
 		v9fs_mux_poll_task_num);
-	up(&v9fs_mux_task_lock);
+	mutex_lock(&v9fs_mux_task_lock);
 
 	n = v9fs_mux_calc_poll_procs(v9fs_mux_num + 1);
 	if (n > v9fs_mux_poll_task_num) {
@@ -225,7 +225,7 @@ static int v9fs_mux_poll_start(struct v9fs_mux_data *m)
 	}
 
 	v9fs_mux_num++;
-	down(&v9fs_mux_task_lock);
+	mutex_unlock(&v9fs_mux_task_lock);
 
 	return 0;
 }
@@ -235,7 +235,7 @@ static void v9fs_mux_poll_stop(struct v9fs_mux_data *m)
 	int i;
 	struct v9fs_mux_poll_task *vpt;
 
-	up(&v9fs_mux_task_lock);
+	mutex_lock(&v9fs_mux_task_lock);
 	vpt = m->poll_task;
 	list_del(&m->mux_list);
 	for(i = 0; i < ARRAY_SIZE(m->poll_waddr); i++) {
@@ -252,7 +252,7 @@ static void v9fs_mux_poll_stop(struct v9fs_mux_data *m)
 		v9fs_mux_poll_task_num--;
 	}
 	v9fs_mux_num--;
-	down(&v9fs_mux_task_lock);
+	mutex_unlock(&v9fs_mux_task_lock);
 }
 
 /**
@@ -279,8 +279,8 @@ struct v9fs_mux_data *v9fs_mux_init(struct v9fs_transport *trans, int msize,
 	m->msize = msize;
 	m->extended = extended;
 	m->trans = trans;
-	idr_init(&m->tidpool.pool);
-	init_MUTEX(&m->tidpool.lock);
+	idr_init(&m->tagpool.pool);
+	init_MUTEX(&m->tagpool.lock);
 	m->err = 0;
 	init_waitqueue_head(&m->equeue);
 	INIT_LIST_HEAD(&m->req_list);
@@ -383,7 +383,7 @@ v9fs_pollwait(struct file *filp, wait_queue_head_t * wait_address,
 /**
  * v9fs_poll_mux - polls a mux and schedules read or write works if necessary
  */
-static inline void v9fs_poll_mux(struct v9fs_mux_data *m)
+static void v9fs_poll_mux(struct v9fs_mux_data *m)
 {
 	int n;
 
@@ -634,6 +634,14 @@ static void v9fs_read_work(void *a)
 			goto error;
 		}
 
+		if ((v9fs_debug_level&DEBUG_FCALL) == DEBUG_FCALL) {
+			char buf[150];
+
+			v9fs_printfcall(buf, sizeof(buf), m->rcall,
+				*m->extended);
+			printk(KERN_NOTICE ">>> %p %s\n", m, buf);
+		}
+
 		rcall = m->rcall;
 		rbuf = m->rbuf;
 		if (m->rpos > n) {
@@ -739,6 +747,13 @@ static struct v9fs_req *v9fs_send_request(struct v9fs_mux_data *m,
 
 	v9fs_set_tag(tc, n);
 
+	if ((v9fs_debug_level&DEBUG_FCALL) == DEBUG_FCALL) {
+		char buf[150];
+
+		v9fs_printfcall(buf, sizeof(buf), tc, *m->extended);
+		printk(KERN_NOTICE "<<< %p %s\n", m, buf);
+	}
+
 	req->tag = n;
 	req->tcall = tc;
 	req->rcall = NULL;
@@ -761,9 +776,8 @@ static struct v9fs_req *v9fs_send_request(struct v9fs_mux_data *m,
 	return req;
 }
 
-static inline void
-v9fs_mux_flush_cb(void *a, struct v9fs_fcall *tc, struct v9fs_fcall *rc,
-		  int err)
+static void v9fs_mux_flush_cb(void *a, struct v9fs_fcall *tc,
+			      struct v9fs_fcall *rc, int err)
 {
 	v9fs_mux_req_callback cb;
 	int tag;
@@ -901,6 +915,7 @@ v9fs_mux_rpc(struct v9fs_mux_data *m, struct v9fs_fcall *tc,
 	return err;
 }
 
+#if 0
 /**
  * v9fs_mux_rpcnb - sends 9P request without waiting for response.
  * @m: mux data
@@ -924,6 +939,7 @@ int v9fs_mux_rpcnb(struct v9fs_mux_data *m, struct v9fs_fcall *tc,
 	dprintk(DEBUG_MUX, "mux %p tc %p tag %d\n", m, tc, req->tag);
 	return 0;
 }
+#endif  /*  0  */
 
 /**
  * v9fs_mux_cancel - cancel all pending requests with error
@@ -963,7 +979,7 @@ static u16 v9fs_mux_get_tag(struct v9fs_mux_data *m)
 {
 	int tag;
 
-	tag = v9fs_get_idpool(&m->tidpool);
+	tag = v9fs_get_idpool(&m->tagpool);
 	if (tag < 0)
 		return V9FS_NOTAG;
 	else
@@ -972,6 +988,6 @@ static u16 v9fs_mux_get_tag(struct v9fs_mux_data *m)
 
 static void v9fs_mux_put_tag(struct v9fs_mux_data *m, u16 tag)
 {
-	if (tag != V9FS_NOTAG && v9fs_check_idpool(tag, &m->tidpool))
-		v9fs_put_idpool(tag, &m->tidpool);
+	if (tag != V9FS_NOTAG && v9fs_check_idpool(tag, &m->tagpool))
+		v9fs_put_idpool(tag, &m->tagpool);
 }
diff --git a/fs/9p/mux.h b/fs/9p/mux.h
index 9473b84f24b2..e90bfd32ea42 100644
--- a/fs/9p/mux.h
+++ b/fs/9p/mux.h
@@ -7,9 +7,8 @@
  *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -50,8 +49,6 @@ void v9fs_mux_destroy(struct v9fs_mux_data *);
 int v9fs_mux_send(struct v9fs_mux_data *m, struct v9fs_fcall *tc);
 struct v9fs_fcall *v9fs_mux_recv(struct v9fs_mux_data *m);
 int v9fs_mux_rpc(struct v9fs_mux_data *m, struct v9fs_fcall *tc, struct v9fs_fcall **rc);
-int v9fs_mux_rpcnb(struct v9fs_mux_data *m, struct v9fs_fcall *tc,
-	v9fs_mux_req_callback cb, void *a);
 
 void v9fs_mux_flush(struct v9fs_mux_data *m, int sendflush);
 void v9fs_mux_cancel(struct v9fs_mux_data *m, int err);
diff --git a/fs/9p/trans_fd.c b/fs/9p/trans_fd.c
index 5b2ce21b10fa..94e0a7fd9fc2 100644
--- a/fs/9p/trans_fd.c
+++ b/fs/9p/trans_fd.c
@@ -1,15 +1,16 @@
 /*
  * linux/fs/9p/trans_fd.c
  *
- * File Descriptor Transport Layer
+ * Fd transport layer.  Includes deprecated socket layer.
  *
- *  Copyright (C) 2005 by Latchesar Ionkov <lucho@ionkov.net>
- *  Copyright (C) 2005 by Eric Van Hensbergen <ericvh@gmail.com>
+ *  Copyright (C) 2006 by Russ Cox <rsc@swtch.com>
+ *  Copyright (C) 2004-2005 by Latchesar Ionkov <lucho@ionkov.net>
+ *  Copyright (C) 2004-2005 by Eric Van Hensbergen <ericvh@gmail.com>
+ *  Copyright (C) 1997-2002 by Ron Minnich <rminnich@sarnoff.com>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -25,6 +26,7 @@
  */
 
 #include <linux/config.h>
+#include <linux/in.h>
 #include <linux/module.h>
 #include <linux/net.h>
 #include <linux/ipv6.h>
@@ -40,89 +42,119 @@
 #include "v9fs.h"
 #include "transport.h"
 
+#define V9FS_PORT 564
+
 struct v9fs_trans_fd {
-	struct file *in_file;
-	struct file *out_file;
+	struct file *rd;
+	struct file *wr;
 };
 
 /**
- * v9fs_fd_recv - receive from a socket
+ * v9fs_fd_read- read from a fd
  * @v9ses: session information
  * @v: buffer to receive data into
  * @len: size of receive buffer
  *
  */
-
-static int v9fs_fd_recv(struct v9fs_transport *trans, void *v, int len)
+static int v9fs_fd_read(struct v9fs_transport *trans, void *v, int len)
 {
-	struct v9fs_trans_fd *ts = trans ? trans->priv : NULL;
+	int ret;
+	struct v9fs_trans_fd *ts;
 
-	if (!trans || trans->status != Connected || !ts)
-		return -EIO;
+	if (!trans || trans->status == Disconnected || !(ts = trans->priv))
+		return -EREMOTEIO;
 
-	return kernel_read(ts->in_file, ts->in_file->f_pos, v, len);
+	if (!(ts->rd->f_flags & O_NONBLOCK))
+		dprintk(DEBUG_ERROR, "blocking read ...\n");
+
+	ret = kernel_read(ts->rd, ts->rd->f_pos, v, len);
+	if (ret <= 0 && ret != -ERESTARTSYS && ret != -EAGAIN)
+		trans->status = Disconnected;
+	return ret;
 }
 
 /**
- * v9fs_fd_send - send to a socket
+ * v9fs_fd_write - write to a socket
  * @v9ses: session information
  * @v: buffer to send data from
  * @len: size of send buffer
  *
  */
-
-static int v9fs_fd_send(struct v9fs_transport *trans, void *v, int len)
+static int v9fs_fd_write(struct v9fs_transport *trans, void *v, int len)
 {
-	struct v9fs_trans_fd *ts = trans ? trans->priv : NULL;
-	mm_segment_t oldfs = get_fs();
-	int ret = 0;
+	int ret;
+	mm_segment_t oldfs;
+	struct v9fs_trans_fd *ts;
 
-	if (!trans || trans->status != Connected || !ts)
-		return -EIO;
+	if (!trans || trans->status == Disconnected || !(ts = trans->priv))
+		return -EREMOTEIO;
+
+	if (!(ts->wr->f_flags & O_NONBLOCK))
+		dprintk(DEBUG_ERROR, "blocking write ...\n");
 
 	oldfs = get_fs();
 	set_fs(get_ds());
 	/* The cast to a user pointer is valid due to the set_fs() */
-	ret = vfs_write(ts->out_file, (void __user *)v, len, &ts->out_file->f_pos);
+	ret = vfs_write(ts->wr, (void __user *)v, len, &ts->wr->f_pos);
 	set_fs(oldfs);
 
+	if (ret <= 0 && ret != -ERESTARTSYS && ret != -EAGAIN)
+		trans->status = Disconnected;
 	return ret;
 }
 
-/**
- * v9fs_fd_init - initialize file descriptor transport
- * @v9ses: session information
- * @addr: address of server to mount
- * @data: mount options
- *
- */
-
-static int
-v9fs_fd_init(struct v9fs_session_info *v9ses, const char *addr, char *data)
+static unsigned int
+v9fs_fd_poll(struct v9fs_transport *trans, struct poll_table_struct *pt)
 {
-	struct v9fs_trans_fd *ts = NULL;
-	struct v9fs_transport *trans = v9ses->transport;
+	int ret, n;
+	struct v9fs_trans_fd *ts;
+	mm_segment_t oldfs;
 
-	if((v9ses->wfdno == ~0) || (v9ses->rfdno == ~0)) {
-		printk(KERN_ERR "v9fs: Insufficient options for proto=fd\n");
-		return -ENOPROTOOPT;
-	}
+	if (!trans || trans->status != Connected || !(ts = trans->priv))
+		return -EREMOTEIO;
 
-	ts = kmalloc(sizeof(struct v9fs_trans_fd), GFP_KERNEL);
+	if (!ts->rd->f_op || !ts->rd->f_op->poll)
+		return -EIO;
 
-	if (!ts)
-		return -ENOMEM;
+	if (!ts->wr->f_op || !ts->wr->f_op->poll)
+		return -EIO;
 
-	ts->in_file = fget( v9ses->rfdno );
-	ts->out_file = fget( v9ses->wfdno );
+	oldfs = get_fs();
+	set_fs(get_ds());
 
-	if (!ts->in_file || !ts->out_file) {
-		if (ts->in_file)
-			fput(ts->in_file);
+	ret = ts->rd->f_op->poll(ts->rd, pt);
+	if (ret < 0)
+		goto end;
 
-		if (ts->out_file)
-			fput(ts->out_file);
+	if (ts->rd != ts->wr) {
+		n = ts->wr->f_op->poll(ts->wr, pt);
+		if (n < 0) {
+			ret = n;
+			goto end;
+		}
+		ret = (ret & ~POLLOUT) | (n & ~POLLIN);
+	}
 
+      end:
+	set_fs(oldfs);
+	return ret;
+}
+
+static int v9fs_fd_open(struct v9fs_session_info *v9ses, int rfd, int wfd)
+{
+	struct v9fs_transport *trans = v9ses->transport;
+	struct v9fs_trans_fd *ts = kmalloc(sizeof(struct v9fs_trans_fd),
+					   GFP_KERNEL);
+	if (!ts)
+		return -ENOMEM;
+
+	ts->rd = fget(rfd);
+	ts->wr = fget(wfd);
+	if (!ts->rd || !ts->wr) {
+		if (ts->rd)
+			fput(ts->rd);
+		if (ts->wr)
+			fput(ts->wr);
 		kfree(ts);
 		return -EIO;
 	}
@@ -133,84 +165,145 @@ v9fs_fd_init(struct v9fs_session_info *v9ses, const char *addr, char *data)
 	return 0;
 }
 
-
-/**
- * v9fs_fd_close - shutdown file descriptor
- * @trans: private socket structure
- *
- */
-
-static void v9fs_fd_close(struct v9fs_transport *trans)
+static int v9fs_fd_init(struct v9fs_session_info *v9ses, const char *addr,
+			char *data)
 {
-	struct v9fs_trans_fd *ts;
-
-	if (!trans)
-		return;
-
-	ts = xchg(&trans->priv, NULL);
+	if (v9ses->rfdno == ~0 || v9ses->wfdno == ~0) {
+		printk(KERN_ERR "v9fs: Insufficient options for proto=fd\n");
+		return -ENOPROTOOPT;
+	}
 
-	if (!ts)
-		return;
+	return v9fs_fd_open(v9ses, v9ses->rfdno, v9ses->wfdno);
+}
 
-	trans->status = Disconnected;
-	if (ts->in_file)
-		fput(ts->in_file);
+static int v9fs_socket_open(struct v9fs_session_info *v9ses,
+			    struct socket *csocket)
+{
+	int fd, ret;
+
+	csocket->sk->sk_allocation = GFP_NOIO;
+	if ((fd = sock_map_fd(csocket)) < 0) {
+		eprintk(KERN_ERR, "v9fs_socket_open: failed to map fd\n");
+		ret = fd;
+	      release_csocket:
+		sock_release(csocket);
+		return ret;
+	}
 
-	if (ts->out_file)
-		fput(ts->out_file);
+	if ((ret = v9fs_fd_open(v9ses, fd, fd)) < 0) {
+		sockfd_put(csocket);
+		eprintk(KERN_ERR, "v9fs_socket_open: failed to open fd\n");
+		goto release_csocket;
+	}
 
-	kfree(ts);
+	((struct v9fs_trans_fd *)v9ses->transport->priv)->rd->f_flags |=
+	    O_NONBLOCK;
+	return 0;
 }
 
-static unsigned int
-v9fs_fd_poll(struct v9fs_transport *trans, struct poll_table_struct *pt)
+static int v9fs_tcp_init(struct v9fs_session_info *v9ses, const char *addr,
+			 char *data)
 {
-	int ret, n;
-	struct v9fs_trans_fd *ts;
-	mm_segment_t oldfs;
+	int ret;
+	struct socket *csocket = NULL;
+	struct sockaddr_in sin_server;
+
+	sin_server.sin_family = AF_INET;
+	sin_server.sin_addr.s_addr = in_aton(addr);
+	sin_server.sin_port = htons(v9ses->port);
+	sock_create_kern(PF_INET, SOCK_STREAM, IPPROTO_TCP, &csocket);
+
+	if (!csocket) {
+		eprintk(KERN_ERR, "v9fs_trans_tcp: problem creating socket\n");
+		return -1;
+	}
 
-	if (!trans)
-		return -EIO;
+	ret = csocket->ops->connect(csocket,
+				    (struct sockaddr *)&sin_server,
+				    sizeof(struct sockaddr_in), 0);
+	if (ret < 0) {
+		eprintk(KERN_ERR,
+			"v9fs_trans_tcp: problem connecting socket to %s\n",
+			addr);
+		return ret;
+	}
 
-	ts = trans->priv;
-	if (trans->status != Connected || !ts)
-		return -EIO;
+	return v9fs_socket_open(v9ses, csocket);
+}
 
-	oldfs = get_fs();
-	set_fs(get_ds());
+static int
+v9fs_unix_init(struct v9fs_session_info *v9ses, const char *addr, char *data)
+{
+	int ret;
+	struct socket *csocket;
+	struct sockaddr_un sun_server;
+
+	if (strlen(addr) > UNIX_PATH_MAX) {
+		eprintk(KERN_ERR, "v9fs_trans_unix: address too long: %s\n",
+			addr);
+		return -ENAMETOOLONG;
+	}
 
-	if (!ts->in_file->f_op || !ts->in_file->f_op->poll) {
-		ret = -EIO;
-		goto end;
+	sun_server.sun_family = PF_UNIX;
+	strcpy(sun_server.sun_path, addr);
+	sock_create_kern(PF_UNIX, SOCK_STREAM, 0, &csocket);
+	ret = csocket->ops->connect(csocket, (struct sockaddr *)&sun_server,
+			sizeof(struct sockaddr_un) - 1, 0);
+	if (ret < 0) {
+		eprintk(KERN_ERR,
+			"v9fs_trans_unix: problem connecting socket: %s: %d\n",
+			addr, ret);
+		return ret;
 	}
 
-	ret = ts->in_file->f_op->poll(ts->in_file, pt);
+	return v9fs_socket_open(v9ses, csocket);
+}
 
-	if (ts->out_file != ts->in_file) {
-		if (!ts->out_file->f_op || !ts->out_file->f_op->poll) {
-			ret = -EIO;
-			goto end;
-		}
+/**
+ * v9fs_sock_close - shutdown socket
+ * @trans: private socket structure
+ *
+ */
+static void v9fs_fd_close(struct v9fs_transport *trans)
+{
+	struct v9fs_trans_fd *ts;
 
-		n = ts->out_file->f_op->poll(ts->out_file, pt);
+	if (!trans)
+		return;
 
-		ret &= ~POLLOUT;
-		n &= ~POLLIN;
+	ts = xchg(&trans->priv, NULL);
 
-		ret |= n;
-	}
+	if (!ts)
+		return;
 
-end:
-	set_fs(oldfs);
-	return ret;
+	trans->status = Disconnected;
+	if (ts->rd)
+		fput(ts->rd);
+	if (ts->wr)
+		fput(ts->wr);
+	kfree(ts);
 }
 
-
 struct v9fs_transport v9fs_trans_fd = {
 	.init = v9fs_fd_init,
-	.write = v9fs_fd_send,
-	.read = v9fs_fd_recv,
+	.write = v9fs_fd_write,
+	.read = v9fs_fd_read,
 	.close = v9fs_fd_close,
 	.poll = v9fs_fd_poll,
 };
 
+struct v9fs_transport v9fs_trans_tcp = {
+	.init = v9fs_tcp_init,
+	.write = v9fs_fd_write,
+	.read = v9fs_fd_read,
+	.close = v9fs_fd_close,
+	.poll = v9fs_fd_poll,
+};
+
+struct v9fs_transport v9fs_trans_unix = {
+	.init = v9fs_unix_init,
+	.write = v9fs_fd_write,
+	.read = v9fs_fd_read,
+	.close = v9fs_fd_close,
+	.poll = v9fs_fd_poll,
+};
diff --git a/fs/9p/trans_sock.c b/fs/9p/trans_sock.c
deleted file mode 100644
index 44e830697acb..000000000000
--- a/fs/9p/trans_sock.c
+++ /dev/null
@@ -1,334 +0,0 @@
-/*
- * linux/fs/9p/trans_socket.c
- *
- * Socket Transport Layer
- *
- *  Copyright (C) 2004-2005 by Latchesar Ionkov <lucho@ionkov.net>
- *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
- *  Copyright (C) 1997-2002 by Ron Minnich <rminnich@sarnoff.com>
- *  Copyright (C) 1995, 1996 by Olaf Kirch <okir@monad.swb.de>
- *
- *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
- *
- *  This program is distributed in the hope that it will be useful,
- *  but WITHOUT ANY WARRANTY; without even the implied warranty of
- *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- *  GNU General Public License for more details.
- *
- *  You should have received a copy of the GNU General Public License
- *  along with this program; if not, write to:
- *  Free Software Foundation
- *  51 Franklin Street, Fifth Floor
- *  Boston, MA  02111-1301  USA
- *
- */
-
-#include <linux/config.h>
-#include <linux/in.h>
-#include <linux/module.h>
-#include <linux/net.h>
-#include <linux/ipv6.h>
-#include <linux/errno.h>
-#include <linux/kernel.h>
-#include <linux/un.h>
-#include <asm/uaccess.h>
-#include <linux/inet.h>
-#include <linux/idr.h>
-#include <linux/file.h>
-
-#include "debug.h"
-#include "v9fs.h"
-#include "transport.h"
-
-#define V9FS_PORT 564
-
-struct v9fs_trans_sock {
-	struct socket *s;
-	struct file *filp;
-};
-
-/**
- * v9fs_sock_recv - receive from a socket
- * @v9ses: session information
- * @v: buffer to receive data into
- * @len: size of receive buffer
- *
- */
-
-static int v9fs_sock_recv(struct v9fs_transport *trans, void *v, int len)
-{
-	int ret;
-	struct v9fs_trans_sock *ts;
-
-	if (!trans || trans->status == Disconnected) {
-		dprintk(DEBUG_ERROR, "disconnected ...\n");
-		return -EREMOTEIO;
-	}
-
-	ts = trans->priv;
-
-	if (!(ts->filp->f_flags & O_NONBLOCK))
-		dprintk(DEBUG_ERROR, "blocking read ...\n");
-
-	ret = kernel_read(ts->filp, ts->filp->f_pos, v, len);
-	if (ret <= 0) {
-		if (ret != -ERESTARTSYS && ret != -EAGAIN)
-			trans->status = Disconnected;
-	}
-
-	return ret;
-}
-
-/**
- * v9fs_sock_send - send to a socket
- * @v9ses: session information
- * @v: buffer to send data from
- * @len: size of send buffer
- *
- */
-
-static int v9fs_sock_send(struct v9fs_transport *trans, void *v, int len)
-{
-	int ret;
-	mm_segment_t oldfs;
-	struct v9fs_trans_sock *ts;
-
-	if (!trans || trans->status == Disconnected) {
-		dprintk(DEBUG_ERROR, "disconnected ...\n");
-		return -EREMOTEIO;
-	}
-
-	ts = trans->priv;
-	if (!ts) {
-		dprintk(DEBUG_ERROR, "no transport ...\n");
-		return -EREMOTEIO;
-	}
-
-	if (!(ts->filp->f_flags & O_NONBLOCK))
-		dprintk(DEBUG_ERROR, "blocking write ...\n");
-
-	oldfs = get_fs();
-	set_fs(get_ds());
-	ret = vfs_write(ts->filp, (void __user *)v, len, &ts->filp->f_pos);
-	set_fs(oldfs);
-
-	if (ret < 0) {
-		if (ret != -ERESTARTSYS)
-			trans->status = Disconnected;
-	}
-
-	return ret;
-}
-
-static unsigned int v9fs_sock_poll(struct v9fs_transport *trans,
-	struct poll_table_struct *pt) {
-
-	int ret;
-	struct v9fs_trans_sock *ts;
-	mm_segment_t oldfs;
-
-	if (!trans) {
-		dprintk(DEBUG_ERROR, "no transport\n");
-		return -EIO;
-	}
-
-	ts = trans->priv;
-	if (trans->status != Connected || !ts) {
-		dprintk(DEBUG_ERROR, "transport disconnected: %d\n", trans->status);
-		return -EIO;
-	}
-
-	oldfs = get_fs();
-	set_fs(get_ds());
-
-	if (!ts->filp->f_op || !ts->filp->f_op->poll) {
-		dprintk(DEBUG_ERROR, "no poll operation\n");
-		ret = -EIO;
-		goto end;
-	}
-
-	ret = ts->filp->f_op->poll(ts->filp, pt);
-
-end:
-	set_fs(oldfs);
-	return ret;
-}
-
-
-/**
- * v9fs_tcp_init - initialize TCP socket
- * @v9ses: session information
- * @addr: address of server to mount
- * @data: mount options
- *
- */
-
-static int
-v9fs_tcp_init(struct v9fs_session_info *v9ses, const char *addr, char *data)
-{
-	struct socket *csocket = NULL;
-	struct sockaddr_in sin_server;
-	int rc = 0;
-	struct v9fs_trans_sock *ts = NULL;
-	struct v9fs_transport *trans = v9ses->transport;
-	int fd;
-
-	trans->status = Disconnected;
-
-	ts = kmalloc(sizeof(struct v9fs_trans_sock), GFP_KERNEL);
-
-	if (!ts)
-		return -ENOMEM;
-
-	trans->priv = ts;
-	ts->s = NULL;
-	ts->filp = NULL;
-
-	if (!addr)
-		return -EINVAL;
-
-	dprintk(DEBUG_TRANS, "Connecting to %s\n", addr);
-
-	sin_server.sin_family = AF_INET;
-	sin_server.sin_addr.s_addr = in_aton(addr);
-	sin_server.sin_port = htons(v9ses->port);
-	sock_create_kern(PF_INET, SOCK_STREAM, IPPROTO_TCP, &csocket);
-	rc = csocket->ops->connect(csocket,
-				   (struct sockaddr *)&sin_server,
-				   sizeof(struct sockaddr_in), 0);
-	if (rc < 0) {
-		eprintk(KERN_ERR,
-			"v9fs_trans_tcp: problem connecting socket to %s\n",
-			addr);
-		return rc;
-	}
-	csocket->sk->sk_allocation = GFP_NOIO;
-
-	fd = sock_map_fd(csocket);
-	if (fd < 0) {
-		sock_release(csocket);
-		kfree(ts);
-		trans->priv = NULL;
-		return fd;
-	}
-
-	ts->s = csocket;
-	ts->filp = fget(fd);
-	ts->filp->f_flags |= O_NONBLOCK;
-	trans->status = Connected;
-
-	return 0;
-}
-
-/**
- * v9fs_unix_init - initialize UNIX domain socket
- * @v9ses: session information
- * @dev_name: path to named pipe
- * @data: mount options
- *
- */
-
-static int
-v9fs_unix_init(struct v9fs_session_info *v9ses, const char *dev_name,
-	       char *data)
-{
-	int rc, fd;
-	struct socket *csocket;
-	struct sockaddr_un sun_server;
-	struct v9fs_transport *trans;
-	struct v9fs_trans_sock *ts;
-
-	rc = 0;
-	csocket = NULL;
-	trans = v9ses->transport;
-
-	trans->status = Disconnected;
-
-	if (strlen(dev_name) > UNIX_PATH_MAX) {
-		eprintk(KERN_ERR, "v9fs_trans_unix: address too long: %s\n",
-			dev_name);
-		return -ENOMEM;
-	}
-
-	ts = kmalloc(sizeof(struct v9fs_trans_sock), GFP_KERNEL);
-	if (!ts)
-		return -ENOMEM;
-
-	trans->priv = ts;
-	ts->s = NULL;
-	ts->filp = NULL;
-
-	sun_server.sun_family = PF_UNIX;
-	strcpy(sun_server.sun_path, dev_name);
-	sock_create_kern(PF_UNIX, SOCK_STREAM, 0, &csocket);
-	rc = csocket->ops->connect(csocket, (struct sockaddr *)&sun_server,
-		sizeof(struct sockaddr_un) - 1, 0);	/* -1 *is* important */
-	if (rc < 0) {
-		eprintk(KERN_ERR,
-			"v9fs_trans_unix: problem connecting socket: %s: %d\n",
-			dev_name, rc);
-		return rc;
-	}
-	csocket->sk->sk_allocation = GFP_NOIO;
-
-	fd = sock_map_fd(csocket);
-	if (fd < 0) {
-		sock_release(csocket);
-		kfree(ts);
-		trans->priv = NULL;
-		return fd;
-	}
-
-	ts->s = csocket;
-	ts->filp = fget(fd);
-	ts->filp->f_flags |= O_NONBLOCK;
-	trans->status = Connected;
-
-	return 0;
-}
-
-/**
- * v9fs_sock_close - shutdown socket
- * @trans: private socket structure
- *
- */
-
-static void v9fs_sock_close(struct v9fs_transport *trans)
-{
-	struct v9fs_trans_sock *ts;
-
-	if (!trans)
-		return;
-
-	ts = trans->priv;
-
-	if ((ts) && (ts->filp)) {
-		fput(ts->filp);
-		ts->filp = NULL;
-		ts->s = NULL;
-		trans->status = Disconnected;
-	}
-
-	kfree(ts);
-
-	trans->priv = NULL;
-}
-
-struct v9fs_transport v9fs_trans_tcp = {
-	.init = v9fs_tcp_init,
-	.write = v9fs_sock_send,
-	.read = v9fs_sock_recv,
-	.close = v9fs_sock_close,
-	.poll = v9fs_sock_poll,
-};
-
-struct v9fs_transport v9fs_trans_unix = {
-	.init = v9fs_unix_init,
-	.write = v9fs_sock_send,
-	.read = v9fs_sock_recv,
-	.close = v9fs_sock_close,
-	.poll = v9fs_sock_poll,
-};
diff --git a/fs/9p/transport.h b/fs/9p/transport.h
index 91fcdb94b361..b38a4b8a41ce 100644
--- a/fs/9p/transport.h
+++ b/fs/9p/transport.h
@@ -7,9 +7,8 @@
  *  Copyright (C) 2004 by Eric Van Hensbergen <ericvh@gmail.com>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c
index 61352491ba36..d37416eb5791 100644
--- a/fs/9p/v9fs.c
+++ b/fs/9p/v9fs.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -51,7 +50,7 @@ enum {
 	Opt_port, Opt_msize, Opt_uid, Opt_gid, Opt_afid, Opt_debug,
 	Opt_rfdno, Opt_wfdno,
 	/* String options */
-	Opt_name, Opt_remotename,
+	Opt_uname, Opt_remotename,
 	/* Options that take no arguments */
 	Opt_legacy, Opt_nodevmap, Opt_unix, Opt_tcp, Opt_fd,
 	/* Error token */
@@ -67,7 +66,7 @@ static match_table_t tokens = {
 	{Opt_rfdno, "rfdno=%u"},
 	{Opt_wfdno, "wfdno=%u"},
 	{Opt_debug, "debug=%x"},
-	{Opt_name, "name=%s"},
+	{Opt_uname, "uname=%s"},
 	{Opt_remotename, "aname=%s"},
 	{Opt_unix, "proto=unix"},
 	{Opt_tcp, "proto=tcp"},
@@ -116,7 +115,7 @@ static void v9fs_parse_options(char *options, struct v9fs_session_info *v9ses)
 		if (!*p)
 			continue;
 		token = match_token(p, tokens, args);
-		if (token < Opt_name) {
+		if (token < Opt_uname) {
 			if ((ret = match_int(&args[0], &option)) < 0) {
 				dprintk(DEBUG_ERROR,
 					"integer field, but no integer?\n");
@@ -158,7 +157,7 @@ static void v9fs_parse_options(char *options, struct v9fs_session_info *v9ses)
 		case Opt_fd:
 			v9ses->proto = PROTO_FD;
 			break;
-		case Opt_name:
+		case Opt_uname:
 			match_strcpy(v9ses->name, &args[0]);
 			break;
 		case Opt_remotename:
@@ -289,7 +288,7 @@ v9fs_session_init(struct v9fs_session_info *v9ses,
 	/* set global debug level */
 	v9fs_debug_level = v9ses->debug;
 
-	/* id pools that are session-dependent: FIDs and TIDs */
+	/* id pools that are session-dependent: fids and tags */
 	idr_init(&v9ses->fidpool.pool);
 	init_MUTEX(&v9ses->fidpool.lock);
 
diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h
index f337da7a0eec..c134d104cb28 100644
--- a/fs/9p/v9fs.h
+++ b/fs/9p/v9fs.h
@@ -5,9 +5,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -91,6 +90,3 @@ void v9fs_session_cancel(struct v9fs_session_info *v9ses);
 #define V9FS_DEFUSER	"nobody"
 #define V9FS_DEFANAME	""
 
-/* inital pool sizes for fids and tags */
-#define V9FS_START_FIDS 8192
-#define V9FS_START_TIDS 256
diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h
index a759278acaae..f867b8d3e973 100644
--- a/fs/9p/v9fs_vfs.h
+++ b/fs/9p/v9fs_vfs.h
@@ -5,9 +5,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -40,8 +39,8 @@
 
 extern struct file_system_type v9fs_fs_type;
 extern struct address_space_operations v9fs_addr_operations;
-extern struct file_operations v9fs_file_operations;
-extern struct file_operations v9fs_dir_operations;
+extern const struct file_operations v9fs_file_operations;
+extern const struct file_operations v9fs_dir_operations;
 extern struct dentry_operations v9fs_dentry_operations;
 
 struct inode *v9fs_get_inode(struct super_block *sb, int mode);
diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c
index 8100fb5171b7..efda46fb64d9 100644
--- a/fs/9p/vfs_addr.c
+++ b/fs/9p/vfs_addr.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
diff --git a/fs/9p/vfs_dentry.c b/fs/9p/vfs_dentry.c
index 12c9cc926b71..062daa6000ab 100644
--- a/fs/9p/vfs_dentry.c
+++ b/fs/9p/vfs_dentry.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -51,7 +50,7 @@
  *
  */
 
-int v9fs_dentry_delete(struct dentry *dentry)
+static int v9fs_dentry_delete(struct dentry *dentry)
 {
 	dprintk(DEBUG_VFS, " dentry: %s (%p)\n", dentry->d_iname, dentry);
 	return 1;
diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c
index cd5eeb032d64..e32d5971039b 100644
--- a/fs/9p/vfs_dir.c
+++ b/fs/9p/vfs_dir.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -205,7 +204,7 @@ int v9fs_dir_release(struct inode *inode, struct file *filp)
 	return 0;
 }
 
-struct file_operations v9fs_dir_operations = {
+const struct file_operations v9fs_dir_operations = {
 	.read = generic_read_dir,
 	.readdir = v9fs_dir_readdir,
 	.open = v9fs_file_open,
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index de3a129698da..083dcfcd158e 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -69,29 +68,30 @@ int v9fs_file_open(struct inode *inode, struct file *file)
 
 	fid = v9fs_get_idpool(&v9ses->fidpool);
 	if (fid < 0) {
-			eprintk(KERN_WARNING, "newfid fails!\n");
-			return -ENOSPC;
-		}
+		eprintk(KERN_WARNING, "newfid fails!\n");
+		return -ENOSPC;
+	}
 
 	err = v9fs_t_walk(v9ses, vfid->fid, fid, NULL, NULL);
 	if (err < 0) {
-			dprintk(DEBUG_ERROR, "rewalk didn't work\n");
+		dprintk(DEBUG_ERROR, "rewalk didn't work\n");
 		goto put_fid;
 	}
 
-	vfid = kmalloc(sizeof(struct v9fs_fid), GFP_KERNEL);
-	if (vfid == NULL) {
-		dprintk(DEBUG_ERROR, "out of memory\n");
-		goto clunk_fid;
-		}
-
-		/* TODO: do special things for O_EXCL, O_NOFOLLOW, O_SYNC */
-		/* translate open mode appropriately */
+	/* TODO: do special things for O_EXCL, O_NOFOLLOW, O_SYNC */
+	/* translate open mode appropriately */
 	omode = v9fs_uflags2omode(file->f_flags);
 	err = v9fs_t_open(v9ses, fid, omode, &fcall);
 	if (err < 0) {
 		PRINT_FCALL_ERROR("open failed", fcall);
-		goto destroy_vfid;
+		goto clunk_fid;
+	}
+
+	vfid = kmalloc(sizeof(struct v9fs_fid), GFP_KERNEL);
+	if (vfid == NULL) {
+		dprintk(DEBUG_ERROR, "out of memory\n");
+		err = -ENOMEM;
+		goto clunk_fid;
 	}
 
 	file->private_data = vfid;
@@ -106,15 +106,12 @@ int v9fs_file_open(struct inode *inode, struct file *file)
 
 	return 0;
 
-destroy_vfid:
-	v9fs_fid_destroy(vfid);
-
 clunk_fid:
 	v9fs_t_clunk(v9ses, fid);
 
 put_fid:
 	v9fs_put_idpool(fid, &v9ses->fidpool);
-		kfree(fcall);
+	kfree(fcall);
 
 	return err;
 }
@@ -269,7 +266,7 @@ v9fs_file_write(struct file *filp, const char __user * data,
 	return total;
 }
 
-struct file_operations v9fs_file_operations = {
+const struct file_operations v9fs_file_operations = {
 	.llseek = generic_file_llseek,
 	.read = v9fs_file_read,
 	.write = v9fs_file_write,
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 3ad8455f8577..133db366d306 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -7,9 +7,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -255,8 +254,8 @@ struct inode *v9fs_get_inode(struct super_block *sb, int mode)
 }
 
 static int
-v9fs_create(struct v9fs_session_info *v9ses, u32 pfid, char *name,
-	u32 perm, u8 mode, u32 *fidp, struct v9fs_qid *qid, u32 *iounit)
+v9fs_create(struct v9fs_session_info *v9ses, u32 pfid, char *name, u32 perm,
+	u8 mode, char *extension, u32 *fidp, struct v9fs_qid *qid, u32 *iounit)
 {
 	u32 fid;
 	int err;
@@ -271,14 +270,14 @@ v9fs_create(struct v9fs_session_info *v9ses, u32 pfid, char *name,
 	err = v9fs_t_walk(v9ses, pfid, fid, NULL, &fcall);
 	if (err < 0) {
 		PRINT_FCALL_ERROR("clone error", fcall);
-		goto error;
+		goto put_fid;
 	}
 	kfree(fcall);
 
-	err = v9fs_t_create(v9ses, fid, name, perm, mode, &fcall);
+	err = v9fs_t_create(v9ses, fid, name, perm, mode, extension, &fcall);
 	if (err < 0) {
 		PRINT_FCALL_ERROR("create fails", fcall);
-		goto error;
+		goto clunk_fid;
 	}
 
 	if (iounit)
@@ -293,7 +292,11 @@ v9fs_create(struct v9fs_session_info *v9ses, u32 pfid, char *name,
 	kfree(fcall);
 	return 0;
 
-error:
+clunk_fid:
+	v9fs_t_clunk(v9ses, fid);
+	fid = V9FS_NOFID;
+
+put_fid:
 	if (fid >= 0)
 		v9fs_put_idpool(fid, &v9ses->fidpool);
 
@@ -348,7 +351,7 @@ error:
 	return ERR_PTR(err);
 }
 
-struct inode *
+static struct inode *
 v9fs_inode_from_fid(struct v9fs_session_info *v9ses, u32 fid,
 	struct super_block *sb)
 {
@@ -474,7 +477,7 @@ v9fs_vfs_create(struct inode *dir, struct dentry *dentry, int mode,
 		flags = O_RDWR;
 
 	err = v9fs_create(v9ses, dfid->fid, (char *) dentry->d_name.name,
-		perm, v9fs_uflags2omode(flags), &fid, &qid, &iounit);
+		perm, v9fs_uflags2omode(flags), NULL, &fid, &qid, &iounit);
 
 	if (err)
 		goto error;
@@ -550,7 +553,7 @@ static int v9fs_vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	perm = unixmode2p9mode(v9ses, mode | S_IFDIR);
 
 	err = v9fs_create(v9ses, dfid->fid, (char *) dentry->d_name.name,
-		perm, V9FS_OREAD, &fid, NULL, NULL);
+		perm, V9FS_OREAD, NULL, &fid, NULL, NULL);
 
 	if (err) {
 		dprintk(DEBUG_ERROR, "create error %d\n", err);
@@ -614,6 +617,7 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 
 	sb = dir->i_sb;
 	v9ses = v9fs_inode2v9ses(dir);
+	dentry->d_op = &v9fs_dentry_operations;
 	dirfid = v9fs_fid_lookup(dentry->d_parent);
 
 	if (!dirfid) {
@@ -681,8 +685,6 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 		goto FreeFcall;
 
 	fid->qid = fcall->params.rstat.stat.qid;
-
-	dentry->d_op = &v9fs_dentry_operations;
 	v9fs_stat2inode(&fcall->params.rstat.stat, inode, inode->i_sb);
 
 	d_add(dentry, inode);
@@ -1009,11 +1011,13 @@ static int v9fs_readlink(struct dentry *dentry, char *buffer, int buflen)
 
 	/* copy extension buffer into buffer */
 	if (fcall->params.rstat.stat.extension.len < buflen)
-		buflen = fcall->params.rstat.stat.extension.len;
+		buflen = fcall->params.rstat.stat.extension.len + 1;
 
-	memcpy(buffer, fcall->params.rstat.stat.extension.str, buflen - 1);
+	memmove(buffer, fcall->params.rstat.stat.extension.str, buflen - 1);
 	buffer[buflen-1] = 0;
 
+	dprintk(DEBUG_ERROR, "%s -> %.*s (%s)\n", dentry->d_name.name, fcall->params.rstat.stat.extension.len,
+		fcall->params.rstat.stat.extension.str, buffer);
 	retval = buflen;
 
       FreeFcall:
@@ -1073,7 +1077,7 @@ static void *v9fs_vfs_follow_link(struct dentry *dentry, struct nameidata *nd)
 	if (!link)
 		link = ERR_PTR(-ENOMEM);
 	else {
-		len = v9fs_readlink(dentry, link, strlen(link));
+		len = v9fs_readlink(dentry, link, PATH_MAX);
 
 		if (len < 0) {
 			__putname(link);
@@ -1110,10 +1114,7 @@ static int v9fs_vfs_mkspecial(struct inode *dir, struct dentry *dentry,
 	struct v9fs_session_info *v9ses;
 	struct v9fs_fid *dfid, *vfid;
 	struct inode *inode;
-	struct v9fs_fcall *fcall;
-	struct v9fs_wstat wstat;
 
-	fcall = NULL;
 	inode = NULL;
 	vfid = NULL;
 	v9ses = v9fs_inode2v9ses(dir);
@@ -1126,7 +1127,7 @@ static int v9fs_vfs_mkspecial(struct inode *dir, struct dentry *dentry,
 	}
 
 	err = v9fs_create(v9ses, dfid->fid, (char *) dentry->d_name.name,
-		perm, V9FS_OREAD, &fid, NULL, NULL);
+		perm, V9FS_OREAD, (char *) extension, &fid, NULL, NULL);
 
 	if (err)
 		goto error;
@@ -1149,23 +1150,11 @@ static int v9fs_vfs_mkspecial(struct inode *dir, struct dentry *dentry,
 		goto error;
 	}
 
-	/* issue a Twstat */
-	v9fs_blank_wstat(&wstat);
-	wstat.muid = v9ses->name;
-	wstat.extension = (char *) extension;
-	err = v9fs_t_wstat(v9ses, vfid->fid, &wstat, &fcall);
-	if (err < 0) {
-		PRINT_FCALL_ERROR("wstat error", fcall);
-		goto error;
-	}
-
-	kfree(fcall);
 	dentry->d_op = &v9fs_dentry_operations;
 	d_instantiate(dentry, inode);
 	return 0;
 
 error:
-	kfree(fcall);
 	if (vfid)
 		v9fs_fid_destroy(vfid);
 
@@ -1225,7 +1214,7 @@ v9fs_vfs_link(struct dentry *old_dentry, struct inode *dir,
 	}
 
 	name = __getname();
-	sprintf(name, "hardlink(%d)\n", oldfid->fid);
+	sprintf(name, "%d\n", oldfid->fid);
 	retval = v9fs_vfs_mkspecial(dir, dentry, V9FS_DMLINK, name);
 	__putname(name);
 
@@ -1254,6 +1243,8 @@ v9fs_vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
 		return -EINVAL;
 
 	name = __getname();
+	if (!name)
+		return -ENOMEM;
 	/* build extension */
 	if (S_ISBLK(mode))
 		sprintf(name, "b %u %u", MAJOR(rdev), MINOR(rdev));
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index d05318fa684e..b0a0ae509c00 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -8,9 +8,8 @@
  *  Copyright (C) 2002 by Ron Minnich <rminnich@lanl.gov>
  *
  *  This program is free software; you can redistribute it and/or modify
- *  it under the terms of the GNU General Public License as published by
- *  the Free Software Foundation; either version 2 of the License, or
- *  (at your option) any later version.
+ *  it under the terms of the GNU General Public License version 2
+ *  as published by the Free Software Foundation.
  *
  *  This program is distributed in the hope that it will be useful,
  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ -262,7 +261,7 @@ static struct super_operations v9fs_super_ops = {
 };
 
 struct file_system_type v9fs_fs_type = {
-	.name = "9P",
+	.name = "9p",
 	.get_sb = v9fs_get_sb,
 	.kill_sb = v9fs_kill_super,
 	.owner = THIS_MODULE,
diff --git a/fs/Kconfig b/fs/Kconfig
index e9749b0eecd8..e207be68d4ca 100644
--- a/fs/Kconfig
+++ b/fs/Kconfig
@@ -859,18 +859,6 @@ config RAMFS
 	  To compile this as a module, choose M here: the module will be called
 	  ramfs.
 
-config RELAYFS_FS
-	tristate "Relayfs file system support"
-	---help---
-	  Relayfs is a high-speed data relay filesystem designed to provide
-	  an efficient mechanism for tools and facilities to relay large
-	  amounts of data from kernel space to user space.
-
-	  To compile this code as a module, choose M here: the module will be
-	  called relayfs.
-
-	  If unsure, say N.
-
 config CONFIGFS_FS
 	tristate "Userspace-driven configuration filesystem (EXPERIMENTAL)"
 	depends on EXPERIMENTAL
@@ -1567,6 +1555,7 @@ config RPCSEC_GSS_SPKM3
 	select CRYPTO
 	select CRYPTO_MD5
 	select CRYPTO_DES
+	select CRYPTO_CAST5
 	help
 	  Provides for secure RPC calls by means of a gss-api
 	  mechanism based on the SPKM3 public-key mechanism.
diff --git a/fs/Makefile b/fs/Makefile
index 1db711319c80..f3a4f7077175 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -10,7 +10,7 @@ obj-y :=	open.o read_write.o file_table.o buffer.o  bio.o super.o \
 		ioctl.o readdir.o select.o fifo.o locks.o dcache.o inode.o \
 		attr.o bad_inode.o file.o filesystems.o namespace.o aio.o \
 		seq_file.o xattr.o libfs.o fs-writeback.o mpage.o direct-io.o \
-		ioprio.o pnode.o drop_caches.o
+		ioprio.o pnode.o drop_caches.o splice.o
 
 obj-$(CONFIG_INOTIFY)		+= inotify.o
 obj-$(CONFIG_EPOLL)		+= eventpoll.o
@@ -91,7 +91,6 @@ obj-$(CONFIG_AUTOFS4_FS)	+= autofs4/
 obj-$(CONFIG_ADFS_FS)		+= adfs/
 obj-$(CONFIG_FUSE_FS)		+= fuse/
 obj-$(CONFIG_UDF_FS)		+= udf/
-obj-$(CONFIG_RELAYFS_FS)	+= relayfs/
 obj-$(CONFIG_SUN_OPENPROMFS)	+= openpromfs/
 obj-$(CONFIG_JFS_FS)		+= jfs/
 obj-$(CONFIG_XFS_FS)		+= xfs/
diff --git a/fs/adfs/adfs.h b/fs/adfs/adfs.h
index f6cd01352cc8..29217ff36d44 100644
--- a/fs/adfs/adfs.h
+++ b/fs/adfs/adfs.h
@@ -85,7 +85,7 @@ void __adfs_error(struct super_block *sb, const char *function,
 
 /* dir_*.c */
 extern struct inode_operations adfs_dir_inode_operations;
-extern struct file_operations adfs_dir_operations;
+extern const struct file_operations adfs_dir_operations;
 extern struct dentry_operations adfs_dentry_operations;
 extern struct adfs_dir_ops adfs_f_dir_ops;
 extern struct adfs_dir_ops adfs_fplus_dir_ops;
@@ -94,7 +94,7 @@ extern int adfs_dir_update(struct super_block *sb, struct object_info *obj);
 
 /* file.c */
 extern struct inode_operations adfs_file_inode_operations;
-extern struct file_operations adfs_file_operations;
+extern const struct file_operations adfs_file_operations;
 
 static inline __u32 signed_asl(__u32 val, signed int shift)
 {
diff --git a/fs/adfs/dir.c b/fs/adfs/dir.c
index 0b4c3a028076..7b075fc397da 100644
--- a/fs/adfs/dir.c
+++ b/fs/adfs/dir.c
@@ -196,7 +196,7 @@ out:
 	return ret;
 }
 
-struct file_operations adfs_dir_operations = {
+const struct file_operations adfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= adfs_readdir,
 	.fsync		= file_fsync,
diff --git a/fs/adfs/file.c b/fs/adfs/file.c
index afebbfde6968..1014b9f2117b 100644
--- a/fs/adfs/file.c
+++ b/fs/adfs/file.c
@@ -19,17 +19,13 @@
  *
  *  adfs regular file handling primitives           
  */
-#include <linux/errno.h>
 #include <linux/fs.h>
-#include <linux/fcntl.h>
-#include <linux/time.h>
-#include <linux/stat.h>
 #include <linux/buffer_head.h>			/* for file_fsync() */
 #include <linux/adfs_fs.h>
 
 #include "adfs.h"
 
-struct file_operations adfs_file_operations = {
+const struct file_operations adfs_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.mmap		= generic_file_mmap,
diff --git a/fs/adfs/super.c b/fs/adfs/super.c
index 243963228d10..252abda0d200 100644
--- a/fs/adfs/super.c
+++ b/fs/adfs/super.c
@@ -241,7 +241,8 @@ static int init_inodecache(void)
 {
 	adfs_inode_cachep = kmem_cache_create("adfs_inode_cache",
 					     sizeof(struct adfs_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (adfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/affs/affs.h b/fs/affs/affs.h
index 0c6799f2137a..a43a876742b8 100644
--- a/fs/affs/affs.h
+++ b/fs/affs/affs.h
@@ -192,9 +192,9 @@ extern void   affs_dir_truncate(struct inode *);
 extern struct inode_operations	 affs_file_inode_operations;
 extern struct inode_operations	 affs_dir_inode_operations;
 extern struct inode_operations   affs_symlink_inode_operations;
-extern struct file_operations	 affs_file_operations;
-extern struct file_operations	 affs_file_operations_ofs;
-extern struct file_operations	 affs_dir_operations;
+extern const struct file_operations	 affs_file_operations;
+extern const struct file_operations	 affs_file_operations_ofs;
+extern const struct file_operations	 affs_dir_operations;
 extern struct address_space_operations	 affs_symlink_aops;
 extern struct address_space_operations	 affs_aops;
 extern struct address_space_operations	 affs_aops_ofs;
diff --git a/fs/affs/dir.c b/fs/affs/dir.c
index 548efd0ee98c..5d9649fa1814 100644
--- a/fs/affs/dir.c
+++ b/fs/affs/dir.c
@@ -17,7 +17,7 @@
 
 static int affs_readdir(struct file *, void *, filldir_t);
 
-struct file_operations affs_dir_operations = {
+const struct file_operations affs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= affs_readdir,
 	.fsync		= file_fsync,
diff --git a/fs/affs/file.c b/fs/affs/file.c
index f72fb776ecdf..7076262af39b 100644
--- a/fs/affs/file.c
+++ b/fs/affs/file.c
@@ -25,7 +25,7 @@ static struct buffer_head *affs_get_extblock_slow(struct inode *inode, u32 ext);
 static int affs_file_open(struct inode *inode, struct file *filp);
 static int affs_file_release(struct inode *inode, struct file *filp);
 
-struct file_operations affs_file_operations = {
+const struct file_operations affs_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/affs/super.c b/fs/affs/super.c
index aaec015a16e4..4d7e5b19e5cd 100644
--- a/fs/affs/super.c
+++ b/fs/affs/super.c
@@ -98,7 +98,8 @@ static int init_inodecache(void)
 {
 	affs_inode_cachep = kmem_cache_create("affs_inode_cache",
 					     sizeof(struct affs_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (affs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/afs/cmservice.c b/fs/afs/cmservice.c
index 9eef6bf156ab..3d097fddcb7a 100644
--- a/fs/afs/cmservice.c
+++ b/fs/afs/cmservice.c
@@ -94,7 +94,7 @@ static struct rxrpc_service AFSCM_service = {
 	.error_func	= afscm_error,
 	.aemap_func	= afscm_aemap,
 	.ops_begin	= &AFSCM_ops[0],
-	.ops_end	= &AFSCM_ops[sizeof(AFSCM_ops) / sizeof(AFSCM_ops[0])],
+	.ops_end	= &AFSCM_ops[ARRAY_SIZE(AFSCM_ops)],
 };
 
 static DECLARE_COMPLETION(kafscmd_alive);
diff --git a/fs/afs/dir.c b/fs/afs/dir.c
index 5c61c24dab2a..a6dff6a4f204 100644
--- a/fs/afs/dir.c
+++ b/fs/afs/dir.c
@@ -32,7 +32,7 @@ static int afs_d_delete(struct dentry *dentry);
 static int afs_dir_lookup_filldir(void *_cookie, const char *name, int nlen,
 				  loff_t fpos, ino_t ino, unsigned dtype);
 
-struct file_operations afs_dir_file_operations = {
+const struct file_operations afs_dir_file_operations = {
 	.open		= afs_dir_open,
 	.readdir	= afs_dir_readdir,
 };
diff --git a/fs/afs/file.c b/fs/afs/file.c
index 150b19227922..7bb716887e29 100644
--- a/fs/afs/file.c
+++ b/fs/afs/file.c
@@ -28,7 +28,7 @@ static int afs_file_release(struct inode *inode, struct file *file);
 #endif
 
 static int afs_file_readpage(struct file *file, struct page *page);
-static int afs_file_invalidatepage(struct page *page, unsigned long offset);
+static void afs_file_invalidatepage(struct page *page, unsigned long offset);
 static int afs_file_releasepage(struct page *page, gfp_t gfp_flags);
 
 struct inode_operations afs_file_inode_operations = {
@@ -212,7 +212,7 @@ int afs_cache_get_page_cookie(struct page *page,
 /*
  * invalidate part or all of a page
  */
-static int afs_file_invalidatepage(struct page *page, unsigned long offset)
+static void afs_file_invalidatepage(struct page *page, unsigned long offset)
 {
 	int ret = 1;
 
@@ -238,11 +238,11 @@ static int afs_file_invalidatepage(struct page *page, unsigned long offset)
 			if (!PageWriteback(page))
 				ret = page->mapping->a_ops->releasepage(page,
 									0);
+			/* possibly should BUG_ON(!ret); - neilb */
 		}
 	}
 
 	_leave(" = %d", ret);
-	return ret;
 } /* end afs_file_invalidatepage() */
 
 /*****************************************************************************/
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index ab8f87c66319..72febdf9a35a 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -64,7 +64,7 @@ extern struct cachefs_index_def afs_cache_cell_index_def;
  * dir.c
  */
 extern struct inode_operations afs_dir_inode_operations;
-extern struct file_operations afs_dir_file_operations;
+extern const struct file_operations afs_dir_file_operations;
 
 /*
  * file.c
@@ -105,7 +105,7 @@ extern struct cachefs_netfs afs_cache_netfs;
  * mntpt.c
  */
 extern struct inode_operations afs_mntpt_inode_operations;
-extern struct file_operations afs_mntpt_file_operations;
+extern const struct file_operations afs_mntpt_file_operations;
 extern struct afs_timer afs_mntpt_expiry_timer;
 extern struct afs_timer_ops afs_mntpt_expiry_timer_ops;
 extern unsigned long afs_mntpt_expiry_timeout;
diff --git a/fs/afs/mntpt.c b/fs/afs/mntpt.c
index 31ee06590de5..4e6eeb59b83c 100644
--- a/fs/afs/mntpt.c
+++ b/fs/afs/mntpt.c
@@ -32,7 +32,7 @@ static struct dentry *afs_mntpt_lookup(struct inode *dir,
 static int afs_mntpt_open(struct inode *inode, struct file *file);
 static void *afs_mntpt_follow_link(struct dentry *dentry, struct nameidata *nd);
 
-struct file_operations afs_mntpt_file_operations = {
+const struct file_operations afs_mntpt_file_operations = {
 	.open		= afs_mntpt_open,
 };
 
diff --git a/fs/afs/proc.c b/fs/afs/proc.c
index 9c81b8f7eef0..101d21b6c037 100644
--- a/fs/afs/proc.c
+++ b/fs/afs/proc.c
@@ -37,7 +37,7 @@ static struct seq_operations afs_proc_cells_ops = {
 	.show	= afs_proc_cells_show,
 };
 
-static struct file_operations afs_proc_cells_fops = {
+static const struct file_operations afs_proc_cells_fops = {
 	.open		= afs_proc_cells_open,
 	.read		= seq_read,
 	.write		= afs_proc_cells_write,
@@ -53,7 +53,7 @@ static ssize_t afs_proc_rootcell_write(struct file *file,
 				       const char __user *buf,
 				       size_t size, loff_t *_pos);
 
-static struct file_operations afs_proc_rootcell_fops = {
+static const struct file_operations afs_proc_rootcell_fops = {
 	.open		= afs_proc_rootcell_open,
 	.read		= afs_proc_rootcell_read,
 	.write		= afs_proc_rootcell_write,
@@ -77,7 +77,7 @@ static struct seq_operations afs_proc_cell_volumes_ops = {
 	.show	= afs_proc_cell_volumes_show,
 };
 
-static struct file_operations afs_proc_cell_volumes_fops = {
+static const struct file_operations afs_proc_cell_volumes_fops = {
 	.open		= afs_proc_cell_volumes_open,
 	.read		= seq_read,
 	.llseek		= seq_lseek,
@@ -101,7 +101,7 @@ static struct seq_operations afs_proc_cell_vlservers_ops = {
 	.show	= afs_proc_cell_vlservers_show,
 };
 
-static struct file_operations afs_proc_cell_vlservers_fops = {
+static const struct file_operations afs_proc_cell_vlservers_fops = {
 	.open		= afs_proc_cell_vlservers_open,
 	.read		= seq_read,
 	.llseek		= seq_lseek,
@@ -124,7 +124,7 @@ static struct seq_operations afs_proc_cell_servers_ops = {
 	.show	= afs_proc_cell_servers_show,
 };
 
-static struct file_operations afs_proc_cell_servers_fops = {
+static const struct file_operations afs_proc_cell_servers_fops = {
 	.open		= afs_proc_cell_servers_open,
 	.read		= seq_read,
 	.llseek		= seq_lseek,
diff --git a/fs/afs/super.c b/fs/afs/super.c
index d6fa8e5999df..53c56e7231ab 100644
--- a/fs/afs/super.c
+++ b/fs/afs/super.c
@@ -341,7 +341,7 @@ static struct super_block *afs_get_sb(struct file_system_type *fs_type,
 
 	sb->s_flags = flags;
 
-	ret = afs_fill_super(sb, &params, flags & MS_VERBOSE ? 1 : 0);
+	ret = afs_fill_super(sb, &params, flags & MS_SILENT ? 1 : 0);
 	if (ret < 0) {
 		up_write(&sb->s_umount);
 		deactivate_super(sb);
diff --git a/fs/aio.c b/fs/aio.c
index aec2b1916d1b..e41e932ba489 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -122,10 +122,9 @@ static int aio_setup_ring(struct kioctx *ctx)
 	info->nr = 0;
 	info->ring_pages = info->internal_pages;
 	if (nr_pages > AIO_RING_PAGES) {
-		info->ring_pages = kmalloc(sizeof(struct page *) * nr_pages, GFP_KERNEL);
+		info->ring_pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
 		if (!info->ring_pages)
 			return -ENOMEM;
-		memset(info->ring_pages, 0, sizeof(struct page *) * nr_pages);
 	}
 
 	info->mmap_size = nr_pages * PAGE_SIZE;
diff --git a/fs/autofs/autofs_i.h b/fs/autofs/autofs_i.h
index 990c28da5aec..a62327f1bdff 100644
--- a/fs/autofs/autofs_i.h
+++ b/fs/autofs/autofs_i.h
@@ -146,7 +146,7 @@ struct autofs_dir_ent *autofs_expire(struct super_block *,struct autofs_sb_info
 
 extern struct inode_operations autofs_root_inode_operations;
 extern struct inode_operations autofs_symlink_inode_operations;
-extern struct file_operations autofs_root_operations;
+extern const struct file_operations autofs_root_operations;
 
 /* Initializing function */
 
diff --git a/fs/autofs/dirhash.c b/fs/autofs/dirhash.c
index 5ccfcf26310d..3fded389d06b 100644
--- a/fs/autofs/dirhash.c
+++ b/fs/autofs/dirhash.c
@@ -92,7 +92,7 @@ struct autofs_dir_ent *autofs_expire(struct super_block *sb,
 			;
 		dput(dentry);
 
-		if ( may_umount(mnt) == 0 ) {
+		if ( may_umount(mnt) ) {
 			mntput(mnt);
 			DPRINTK(("autofs: signaling expire on %s\n", ent->name));
 			return ent; /* Expirable! */
diff --git a/fs/autofs/root.c b/fs/autofs/root.c
index 870e2cf33016..9cac08d6a873 100644
--- a/fs/autofs/root.c
+++ b/fs/autofs/root.c
@@ -26,7 +26,7 @@ static int autofs_root_rmdir(struct inode *,struct dentry *);
 static int autofs_root_mkdir(struct inode *,struct dentry *,int);
 static int autofs_root_ioctl(struct inode *, struct file *,unsigned int,unsigned long);
 
-struct file_operations autofs_root_operations = {
+const struct file_operations autofs_root_operations = {
 	.read		= generic_read_dir,
 	.readdir	= autofs_root_readdir,
 	.ioctl		= autofs_root_ioctl,
diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h
index 385bed09b0d8..57c4903614e5 100644
--- a/fs/autofs4/autofs_i.h
+++ b/fs/autofs4/autofs_i.h
@@ -3,6 +3,7 @@
  * linux/fs/autofs/autofs_i.h
  *
  *   Copyright 1997-1998 Transmeta Corporation - All Rights Reserved
+ *   Copyright 2005-2006 Ian Kent <raven@themaw.net>
  *
  * This file is part of the Linux kernel and is made available under
  * the terms of the GNU General Public License, version 2, or at your
@@ -13,6 +14,7 @@
 /* Internal header file for autofs */
 
 #include <linux/auto_fs4.h>
+#include <linux/mutex.h>
 #include <linux/list.h>
 
 /* This is the range of ioctl() numbers we claim as ours */
@@ -40,14 +42,6 @@
 
 #define AUTOFS_SUPER_MAGIC 0x0187
 
-/*
- * If the daemon returns a negative response (AUTOFS_IOC_FAIL) then the
- * kernel will keep the negative response cached for up to the time given
- * here, although the time can be shorter if the kernel throws the dcache
- * entry away.  This probably should be settable from user space.
- */
-#define AUTOFS_NEGATIVE_TIMEOUT (60*HZ)	/* 1 minute */
-
 /* Unified info structure.  This is pointed to by both the dentry and
    inode structures.  Each file in the filesystem has an instance of this
    structure.  It holds a reference to the dentry, so dentries are never
@@ -62,6 +56,7 @@ struct autofs_info {
 
 	struct autofs_sb_info *sbi;
 	unsigned long last_used;
+	atomic_t count;
 
 	mode_t	mode;
 	size_t	size;
@@ -82,27 +77,41 @@ struct autofs_wait_queue {
 	int hash;
 	int len;
 	char *name;
+	u32 dev;
+	u64 ino;
+	uid_t uid;
+	gid_t gid;
+	pid_t pid;
+	pid_t tgid;
 	/* This is for status reporting upon return */
 	int status;
-	atomic_t notified;
+	atomic_t notify;
 	atomic_t wait_ctr;
 };
 
 #define AUTOFS_SBI_MAGIC 0x6d4a556d
 
+#define AUTOFS_TYPE_INDIRECT     0x0001
+#define AUTOFS_TYPE_DIRECT       0x0002
+#define AUTOFS_TYPE_OFFSET       0x0004
+
 struct autofs_sb_info {
 	u32 magic;
 	struct dentry *root;
+	int pipefd;
 	struct file *pipe;
 	pid_t oz_pgrp;
 	int catatonic;
 	int version;
 	int sub_version;
+	int min_proto;
+	int max_proto;
 	unsigned long exp_timeout;
+	unsigned int type;
 	int reghost_enabled;
 	int needs_reghost;
 	struct super_block *sb;
-	struct semaphore wq_sem;
+	struct mutex wq_mutex;
 	spinlock_t fs_lock;
 	struct autofs_wait_queue *queues; /* Wait queue pointer */
 };
@@ -165,8 +174,10 @@ int autofs4_expire_multi(struct super_block *, struct vfsmount *,
 extern struct inode_operations autofs4_symlink_inode_operations;
 extern struct inode_operations autofs4_dir_inode_operations;
 extern struct inode_operations autofs4_root_inode_operations;
-extern struct file_operations autofs4_dir_operations;
-extern struct file_operations autofs4_root_operations;
+extern struct inode_operations autofs4_indirect_root_inode_operations;
+extern struct inode_operations autofs4_direct_root_inode_operations;
+extern const struct file_operations autofs4_dir_operations;
+extern const struct file_operations autofs4_root_operations;
 
 /* Initializing function */
 
@@ -175,13 +186,6 @@ struct autofs_info *autofs4_init_ino(struct autofs_info *, struct autofs_sb_info
 
 /* Queue management functions */
 
-enum autofs_notify
-{
-	NFY_NONE,
-	NFY_MOUNT,
-	NFY_EXPIRE
-};
-
 int autofs4_wait(struct autofs_sb_info *,struct dentry *, enum autofs_notify);
 int autofs4_wait_release(struct autofs_sb_info *,autofs_wqt_t,int);
 void autofs4_catatonic_mode(struct autofs_sb_info *);
@@ -199,12 +203,22 @@ static inline int autofs4_follow_mount(struct vfsmount **mnt, struct dentry **de
 	return res;
 }
 
+static inline u32 autofs4_get_dev(struct autofs_sb_info *sbi)
+{
+	return new_encode_dev(sbi->sb->s_dev);
+}
+
+static inline u64 autofs4_get_ino(struct autofs_sb_info *sbi)
+{
+	return sbi->sb->s_root->d_inode->i_ino;
+}
+
 static inline int simple_positive(struct dentry *dentry)
 {
 	return dentry->d_inode && !d_unhashed(dentry);
 }
 
-static inline int simple_empty_nolock(struct dentry *dentry)
+static inline int __simple_empty(struct dentry *dentry)
 {
 	struct dentry *child;
 	int ret = 0;
@@ -216,3 +230,6 @@ static inline int simple_empty_nolock(struct dentry *dentry)
 out:
 	return ret;
 }
+
+void autofs4_dentry_release(struct dentry *);
+
diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c
index dc39589df165..b8ce02607d66 100644
--- a/fs/autofs4/expire.c
+++ b/fs/autofs4/expire.c
@@ -4,7 +4,7 @@
  *
  *  Copyright 1997-1998 Transmeta Corporation -- All Rights Reserved
  *  Copyright 1999-2000 Jeremy Fitzhardinge <jeremy@goop.org>
- *  Copyright 2001-2003 Ian Kent <raven@themaw.net>
+ *  Copyright 2001-2006 Ian Kent <raven@themaw.net>
  *
  * This file is part of the Linux kernel and is made available under
  * the terms of the GNU General Public License, version 2, or at your
@@ -16,7 +16,7 @@
 
 static unsigned long now;
 
-/* Check if a dentry can be expired return 1 if it can else return 0 */
+/* Check if a dentry can be expired */
 static inline int autofs4_can_expire(struct dentry *dentry,
 					unsigned long timeout, int do_now)
 {
@@ -41,14 +41,14 @@ static inline int autofs4_can_expire(struct dentry *dentry,
 		     attempts if expire fails the first time */
 		ino->last_used = now;
 	}
-
 	return 1;
 }
 
-/* Check a mount point for busyness return 1 if not busy, otherwise */
-static int autofs4_check_mount(struct vfsmount *mnt, struct dentry *dentry)
+/* Check a mount point for busyness */
+static int autofs4_mount_busy(struct vfsmount *mnt, struct dentry *dentry)
 {
-	int status = 0;
+	struct dentry *top = dentry;
+	int status = 1;
 
 	DPRINTK("dentry %p %.*s",
 		dentry, (int)dentry->d_name.len, dentry->d_name.name);
@@ -63,9 +63,14 @@ static int autofs4_check_mount(struct vfsmount *mnt, struct dentry *dentry)
 	if (is_autofs4_dentry(dentry))
 		goto done;
 
-	/* The big question */
-	if (may_umount_tree(mnt) == 0)
-		status = 1;
+	/* Update the expiry counter if fs is busy */
+	if (!may_umount_tree(mnt)) {
+		struct autofs_info *ino = autofs4_dentry_ino(top);
+		ino->last_used = jiffies;
+		goto done;
+	}
+
+	status = 0;
 done:
 	DPRINTK("returning = %d", status);
 	mntput(mnt);
@@ -73,78 +78,124 @@ done:
 	return status;
 }
 
+/*
+ * Calculate next entry in top down tree traversal.
+ * From next_mnt in namespace.c - elegant.
+ */
+static struct dentry *next_dentry(struct dentry *p, struct dentry *root)
+{
+	struct list_head *next = p->d_subdirs.next;
+
+	if (next == &p->d_subdirs) {
+		while (1) {
+			if (p == root)
+				return NULL;
+			next = p->d_u.d_child.next;
+			if (next != &p->d_parent->d_subdirs)
+				break;
+			p = p->d_parent;
+		}
+	}
+	return list_entry(next, struct dentry, d_u.d_child);
+}
+
+/*
+ * Check a direct mount point for busyness.
+ * Direct mounts have similar expiry semantics to tree mounts.
+ * The tree is not busy iff no mountpoints are busy and there are no
+ * autofs submounts.
+ */
+static int autofs4_direct_busy(struct vfsmount *mnt,
+				struct dentry *top,
+				unsigned long timeout,
+				int do_now)
+{
+	DPRINTK("top %p %.*s",
+		top, (int) top->d_name.len, top->d_name.name);
+
+	/* If it's busy update the expiry counters */
+	if (!may_umount_tree(mnt)) {
+		struct autofs_info *ino = autofs4_dentry_ino(top);
+		if (ino)
+			ino->last_used = jiffies;
+		return 1;
+	}
+
+	/* Timeout of a direct mount is determined by its top dentry */
+	if (!autofs4_can_expire(top, timeout, do_now))
+		return 1;
+
+	return 0;
+}
+
 /* Check a directory tree of mount points for busyness
  * The tree is not busy iff no mountpoints are busy
- * Return 1 if the tree is busy or 0 otherwise
  */
-static int autofs4_check_tree(struct vfsmount *mnt,
-	       		      struct dentry *top,
-			      unsigned long timeout,
-			      int do_now)
+static int autofs4_tree_busy(struct vfsmount *mnt,
+	       		     struct dentry *top,
+			     unsigned long timeout,
+			     int do_now)
 {
-	struct dentry *this_parent = top;
-	struct list_head *next;
+	struct autofs_info *top_ino = autofs4_dentry_ino(top);
+	struct dentry *p;
 
-	DPRINTK("parent %p %.*s",
+	DPRINTK("top %p %.*s",
 		top, (int)top->d_name.len, top->d_name.name);
 
 	/* Negative dentry - give up */
 	if (!simple_positive(top))
-		return 0;
-
-	/* Timeout of a tree mount is determined by its top dentry */
-	if (!autofs4_can_expire(top, timeout, do_now))
-		return 0;
-
-	/* Is someone visiting anywhere in the tree ? */
-	if (may_umount_tree(mnt))
-		return 0;
+		return 1;
 
 	spin_lock(&dcache_lock);
-repeat:
-	next = this_parent->d_subdirs.next;
-resume:
-	while (next != &this_parent->d_subdirs) {
-		struct dentry *dentry = list_entry(next, struct dentry, d_u.d_child);
-
+	for (p = top; p; p = next_dentry(p, top)) {
 		/* Negative dentry - give up */
-		if (!simple_positive(dentry)) {
-			next = next->next;
+		if (!simple_positive(p))
 			continue;
-		}
 
 		DPRINTK("dentry %p %.*s",
-			dentry, (int)dentry->d_name.len, dentry->d_name.name);
-
-		if (!simple_empty_nolock(dentry)) {
-			this_parent = dentry;
-			goto repeat;
-		}
+			p, (int) p->d_name.len, p->d_name.name);
 
-		dentry = dget(dentry);
+		p = dget(p);
 		spin_unlock(&dcache_lock);
 
-		if (d_mountpoint(dentry)) {
-			/* First busy => tree busy */
-			if (!autofs4_check_mount(mnt, dentry)) {
-				dput(dentry);
-				return 0;
+		/*
+		 * Is someone visiting anywhere in the subtree ?
+		 * If there's no mount we need to check the usage
+		 * count for the autofs dentry.
+		 * If the fs is busy update the expiry counter.
+		 */
+		if (d_mountpoint(p)) {
+			if (autofs4_mount_busy(mnt, p)) {
+				top_ino->last_used = jiffies;
+				dput(p);
+				return 1;
+			}
+		} else {
+			struct autofs_info *ino = autofs4_dentry_ino(p);
+			unsigned int ino_count = atomic_read(&ino->count);
+
+			/* allow for dget above and top is already dgot */
+			if (p == top)
+				ino_count += 2;
+			else
+				ino_count++;
+
+			if (atomic_read(&p->d_count) > ino_count) {
+				top_ino->last_used = jiffies;
+				dput(p);
+				return 1;
 			}
 		}
-
-		dput(dentry);
+		dput(p);
 		spin_lock(&dcache_lock);
-		next = next->next;
-	}
-
-	if (this_parent != top) {
-		next = this_parent->d_u.d_child.next;
-		this_parent = this_parent->d_parent;
-		goto resume;
 	}
 	spin_unlock(&dcache_lock);
 
-	return 1;
+	/* Timeout of a tree mount is ultimately determined by its top dentry */
+	if (!autofs4_can_expire(top, timeout, do_now))
+		return 1;
+
+	return 0;
 }
 
 static struct dentry *autofs4_check_leaves(struct vfsmount *mnt,
@@ -152,58 +203,68 @@ static struct dentry *autofs4_check_leaves(struct vfsmount *mnt,
 					   unsigned long timeout,
 					   int do_now)
 {
-	struct dentry *this_parent = parent;
-	struct list_head *next;
+	struct dentry *p;
 
 	DPRINTK("parent %p %.*s",
 		parent, (int)parent->d_name.len, parent->d_name.name);
 
 	spin_lock(&dcache_lock);
-repeat:
-	next = this_parent->d_subdirs.next;
-resume:
-	while (next != &this_parent->d_subdirs) {
-		struct dentry *dentry = list_entry(next, struct dentry, d_u.d_child);
-
+	for (p = parent; p; p = next_dentry(p, parent)) {
 		/* Negative dentry - give up */
-		if (!simple_positive(dentry)) {
-			next = next->next;
+		if (!simple_positive(p))
 			continue;
-		}
 
 		DPRINTK("dentry %p %.*s",
-			dentry, (int)dentry->d_name.len, dentry->d_name.name);
-
-		if (!list_empty(&dentry->d_subdirs)) {
-			this_parent = dentry;
-			goto repeat;
-		}
+			p, (int) p->d_name.len, p->d_name.name);
 
-		dentry = dget(dentry);
+		p = dget(p);
 		spin_unlock(&dcache_lock);
 
-		if (d_mountpoint(dentry)) {
-			/* Can we expire this guy */
-			if (!autofs4_can_expire(dentry, timeout, do_now))
-				goto cont;
-
+		if (d_mountpoint(p)) {
 			/* Can we umount this guy */
-			if (autofs4_check_mount(mnt, dentry))
-				return dentry;
+			if (autofs4_mount_busy(mnt, p))
+				goto cont;
 
+			/* Can we expire this guy */
+			if (autofs4_can_expire(p, timeout, do_now))
+				return p;
 		}
 cont:
-		dput(dentry);
+		dput(p);
 		spin_lock(&dcache_lock);
-		next = next->next;
 	}
+	spin_unlock(&dcache_lock);
+	return NULL;
+}
+
+/* Check if we can expire a direct mount (possibly a tree) */
+static struct dentry *autofs4_expire_direct(struct super_block *sb,
+					    struct vfsmount *mnt,
+					    struct autofs_sb_info *sbi,
+					    int how)
+{
+	unsigned long timeout;
+	struct dentry *root = dget(sb->s_root);
+	int do_now = how & AUTOFS_EXP_IMMEDIATE;
+
+	if (!sbi->exp_timeout || !root)
+		return NULL;
+
+	now = jiffies;
+	timeout = sbi->exp_timeout;
+
+	/* Lock the tree as we must expire as a whole */
+	spin_lock(&sbi->fs_lock);
+	if (!autofs4_direct_busy(mnt, root, timeout, do_now)) {
+		struct autofs_info *ino = autofs4_dentry_ino(root);
 
-	if (this_parent != parent) {
-		next = this_parent->d_u.d_child.next;
-		this_parent = this_parent->d_parent;
-		goto resume;
+		/* Set this flag early to catch sys_chdir and the like */
+		ino->flags |= AUTOFS_INF_EXPIRING;
+		spin_unlock(&sbi->fs_lock);
+		return root;
 	}
-	spin_unlock(&dcache_lock);
+	spin_unlock(&sbi->fs_lock);
+	dput(root);
 
 	return NULL;
 }
@@ -214,10 +275,10 @@ cont:
  *  - it is unused by any user process
  *  - it has been unused for exp_timeout time
  */
-static struct dentry *autofs4_expire(struct super_block *sb,
-				     struct vfsmount *mnt,
-				     struct autofs_sb_info *sbi,
-				     int how)
+static struct dentry *autofs4_expire_indirect(struct super_block *sb,
+					      struct vfsmount *mnt,
+					      struct autofs_sb_info *sbi,
+					      int how)
 {
 	unsigned long timeout;
 	struct dentry *root = sb->s_root;
@@ -241,7 +302,7 @@ static struct dentry *autofs4_expire(struct super_block *sb,
 		struct dentry *dentry = list_entry(next, struct dentry, d_u.d_child);
 
 		/* Negative dentry - give up */
-		if ( !simple_positive(dentry) ) {
+		if (!simple_positive(dentry)) {
 			next = next->next;
 			continue;
 		}
@@ -249,31 +310,36 @@ static struct dentry *autofs4_expire(struct super_block *sb,
 		dentry = dget(dentry);
 		spin_unlock(&dcache_lock);
 
-		/* Case 1: indirect mount or top level direct mount */
+		/*
+		 * Case 1: (i) indirect mount or top level pseudo direct mount
+		 *	   (autofs-4.1).
+		 *	   (ii) indirect mount with offset mount, check the "/"
+		 *	   offset (autofs-5.0+).
+		 */
 		if (d_mountpoint(dentry)) {
 			DPRINTK("checking mountpoint %p %.*s",
 				dentry, (int)dentry->d_name.len, dentry->d_name.name);
 
-			/* Can we expire this guy */
-			if (!autofs4_can_expire(dentry, timeout, do_now))
+			/* Can we umount this guy */
+			if (autofs4_mount_busy(mnt, dentry))
 				goto next;
 
-			/* Can we umount this guy */
-			if (autofs4_check_mount(mnt, dentry)) {
+			/* Can we expire this guy */
+			if (autofs4_can_expire(dentry, timeout, do_now)) {
 				expired = dentry;
 				break;
 			}
 			goto next;
 		}
 
-		if ( simple_empty(dentry) )
+		if (simple_empty(dentry))
 			goto next;
 
 		/* Case 2: tree mount, expire iff entire tree is not busy */
 		if (!exp_leaves) {
 			/* Lock the tree as we must expire as a whole */
 			spin_lock(&sbi->fs_lock);
-			if (autofs4_check_tree(mnt, dentry, timeout, do_now)) {
+			if (!autofs4_tree_busy(mnt, dentry, timeout, do_now)) {
 				struct autofs_info *inf = autofs4_dentry_ino(dentry);
 
 				/* Set this flag early to catch sys_chdir and the like */
@@ -283,7 +349,10 @@ static struct dentry *autofs4_expire(struct super_block *sb,
 				break;
 			}
 			spin_unlock(&sbi->fs_lock);
-		/* Case 3: direct mount, expire individual leaves */
+		/*
+		 * Case 3: pseudo direct mount, expire individual leaves
+		 *	   (autofs-4.1).
+		 */
 		} else {
 			expired = autofs4_check_leaves(mnt, dentry, timeout, do_now);
 			if (expired) {
@@ -297,7 +366,7 @@ next:
 		next = next->next;
 	}
 
-	if ( expired ) {
+	if (expired) {
 		DPRINTK("returning %p %.*s",
 			expired, (int)expired->d_name.len, expired->d_name.name);
 		spin_lock(&dcache_lock);
@@ -325,7 +394,7 @@ int autofs4_expire_run(struct super_block *sb,
 	pkt.hdr.proto_version = sbi->version;
 	pkt.hdr.type = autofs_ptype_expire;
 
-	if ((dentry = autofs4_expire(sb, mnt, sbi, 0)) == NULL)
+	if ((dentry = autofs4_expire_indirect(sb, mnt, sbi, 0)) == NULL)
 		return -EAGAIN;
 
 	pkt.len = dentry->d_name.len;
@@ -351,17 +420,22 @@ int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt,
 	if (arg && get_user(do_now, arg))
 		return -EFAULT;
 
-	if ((dentry = autofs4_expire(sb, mnt, sbi, do_now)) != NULL) {
-		struct autofs_info *de_info = autofs4_dentry_ino(dentry);
+	if (sbi->type & AUTOFS_TYPE_DIRECT)
+		dentry = autofs4_expire_direct(sb, mnt, sbi, do_now);
+	else
+		dentry = autofs4_expire_indirect(sb, mnt, sbi, do_now);
+
+	if (dentry) {
+		struct autofs_info *ino = autofs4_dentry_ino(dentry);
 
 		/* This is synchronous because it makes the daemon a
                    little easier */
-		de_info->flags |= AUTOFS_INF_EXPIRING;
+		ino->flags |= AUTOFS_INF_EXPIRING;
 		ret = autofs4_wait(sbi, dentry, NFY_EXPIRE);
-		de_info->flags &= ~AUTOFS_INF_EXPIRING;
+		ino->flags &= ~AUTOFS_INF_EXPIRING;
 		dput(dentry);
 	}
-		
+
 	return ret;
 }
 
diff --git a/fs/autofs4/inode.c b/fs/autofs4/inode.c
index 2d3082854a29..fde78b110ddd 100644
--- a/fs/autofs4/inode.c
+++ b/fs/autofs4/inode.c
@@ -3,6 +3,7 @@
  * linux/fs/autofs/inode.c
  *
  *  Copyright 1997-1998 Transmeta Corporation -- All Rights Reserved
+ *  Copyright 2005-2006 Ian Kent <raven@themaw.net>
  *
  * This file is part of the Linux kernel and is made available under
  * the terms of the GNU General Public License, version 2, or at your
@@ -13,6 +14,7 @@
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/file.h>
+#include <linux/seq_file.h>
 #include <linux/pagemap.h>
 #include <linux/parser.h>
 #include <linux/bitops.h>
@@ -46,6 +48,7 @@ struct autofs_info *autofs4_init_ino(struct autofs_info *ino,
 	ino->size = 0;
 
 	ino->last_used = jiffies;
+	atomic_set(&ino->count, 0);
 
 	ino->sbi = sbi;
 
@@ -64,10 +67,19 @@ struct autofs_info *autofs4_init_ino(struct autofs_info *ino,
 
 void autofs4_free_ino(struct autofs_info *ino)
 {
+	struct autofs_info *p_ino;
+
 	if (ino->dentry) {
 		ino->dentry->d_fsdata = NULL;
-		if (ino->dentry->d_inode)
+		if (ino->dentry->d_inode) {
+			struct dentry *parent = ino->dentry->d_parent;
+			if (atomic_dec_and_test(&ino->count)) {
+				p_ino = autofs4_dentry_ino(parent);
+				if (p_ino && parent != ino->dentry)
+					atomic_dec(&p_ino->count);
+			}
 			dput(ino->dentry);
+		}
 		ino->dentry = NULL;
 	}
 	if (ino->free)
@@ -145,20 +157,44 @@ static void autofs4_put_super(struct super_block *sb)
 		autofs4_catatonic_mode(sbi); /* Free wait queues, close pipe */
 
 	/* Clean up and release dangling references */
-	if (sbi)
-		autofs4_force_release(sbi);
+	autofs4_force_release(sbi);
 
 	kfree(sbi);
 
 	DPRINTK("shutting down");
 }
 
+static int autofs4_show_options(struct seq_file *m, struct vfsmount *mnt)
+{
+	struct autofs_sb_info *sbi = autofs4_sbi(mnt->mnt_sb);
+
+	if (!sbi)
+		return 0;
+
+	seq_printf(m, ",fd=%d", sbi->pipefd);
+	seq_printf(m, ",pgrp=%d", sbi->oz_pgrp);
+	seq_printf(m, ",timeout=%lu", sbi->exp_timeout/HZ);
+	seq_printf(m, ",minproto=%d", sbi->min_proto);
+	seq_printf(m, ",maxproto=%d", sbi->max_proto);
+
+	if (sbi->type & AUTOFS_TYPE_OFFSET)
+		seq_printf(m, ",offset");
+	else if (sbi->type & AUTOFS_TYPE_DIRECT)
+		seq_printf(m, ",direct");
+	else
+		seq_printf(m, ",indirect");
+
+	return 0;
+}
+
 static struct super_operations autofs4_sops = {
 	.put_super	= autofs4_put_super,
 	.statfs		= simple_statfs,
+	.show_options	= autofs4_show_options,
 };
 
-enum {Opt_err, Opt_fd, Opt_uid, Opt_gid, Opt_pgrp, Opt_minproto, Opt_maxproto};
+enum {Opt_err, Opt_fd, Opt_uid, Opt_gid, Opt_pgrp, Opt_minproto, Opt_maxproto,
+	Opt_indirect, Opt_direct, Opt_offset};
 
 static match_table_t tokens = {
 	{Opt_fd, "fd=%u"},
@@ -167,11 +203,15 @@ static match_table_t tokens = {
 	{Opt_pgrp, "pgrp=%u"},
 	{Opt_minproto, "minproto=%u"},
 	{Opt_maxproto, "maxproto=%u"},
+	{Opt_indirect, "indirect"},
+	{Opt_direct, "direct"},
+	{Opt_offset, "offset"},
 	{Opt_err, NULL}
 };
 
 static int parse_options(char *options, int *pipefd, uid_t *uid, gid_t *gid,
-			 pid_t *pgrp, int *minproto, int *maxproto)
+			 pid_t *pgrp, unsigned int *type,
+			 int *minproto, int *maxproto)
 {
 	char *p;
 	substring_t args[MAX_OPT_ARGS];
@@ -225,6 +265,15 @@ static int parse_options(char *options, int *pipefd, uid_t *uid, gid_t *gid,
 				return 1;
 			*maxproto = option;
 			break;
+		case Opt_indirect:
+			*type = AUTOFS_TYPE_INDIRECT;
+			break;
+		case Opt_direct:
+			*type = AUTOFS_TYPE_DIRECT;
+			break;
+		case Opt_offset:
+			*type = AUTOFS_TYPE_DIRECT | AUTOFS_TYPE_OFFSET;
+			break;
 		default:
 			return 1;
 		}
@@ -243,6 +292,10 @@ static struct autofs_info *autofs4_mkroot(struct autofs_sb_info *sbi)
 	return ino;
 }
 
+static struct dentry_operations autofs4_sb_dentry_operations = {
+	.d_release      = autofs4_dentry_release,
+};
+
 int autofs4_fill_super(struct super_block *s, void *data, int silent)
 {
 	struct inode * root_inode;
@@ -251,7 +304,6 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	int pipefd;
 	struct autofs_sb_info *sbi;
 	struct autofs_info *ino;
-	int minproto, maxproto;
 
 	sbi = (struct autofs_sb_info *) kmalloc(sizeof(*sbi), GFP_KERNEL);
 	if ( !sbi )
@@ -263,13 +315,17 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	s->s_fs_info = sbi;
 	sbi->magic = AUTOFS_SBI_MAGIC;
 	sbi->root = NULL;
+	sbi->pipefd = -1;
 	sbi->catatonic = 0;
 	sbi->exp_timeout = 0;
 	sbi->oz_pgrp = process_group(current);
 	sbi->sb = s;
 	sbi->version = 0;
 	sbi->sub_version = 0;
-	init_MUTEX(&sbi->wq_sem);
+	sbi->type = 0;
+	sbi->min_proto = 0;
+	sbi->max_proto = 0;
+	mutex_init(&sbi->wq_mutex);
 	spin_lock_init(&sbi->fs_lock);
 	sbi->queues = NULL;
 	s->s_blocksize = 1024;
@@ -285,38 +341,46 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	if (!ino)
 		goto fail_free;
 	root_inode = autofs4_get_inode(s, ino);
-	kfree(ino);
 	if (!root_inode)
-		goto fail_free;
+		goto fail_ino;
 
-	root_inode->i_op = &autofs4_root_inode_operations;
-	root_inode->i_fop = &autofs4_root_operations;
 	root = d_alloc_root(root_inode);
-	pipe = NULL;
-
 	if (!root)
 		goto fail_iput;
+	pipe = NULL;
+
+	root->d_op = &autofs4_sb_dentry_operations;
+	root->d_fsdata = ino;
 
 	/* Can this call block? */
 	if (parse_options(data, &pipefd,
 			  &root_inode->i_uid, &root_inode->i_gid,
-			  &sbi->oz_pgrp,
-			  &minproto, &maxproto)) {
+			  &sbi->oz_pgrp, &sbi->type,
+			  &sbi->min_proto, &sbi->max_proto)) {
 		printk("autofs: called with bogus options\n");
 		goto fail_dput;
 	}
 
+	root_inode->i_fop = &autofs4_root_operations;
+	root_inode->i_op = sbi->type & AUTOFS_TYPE_DIRECT ?
+			&autofs4_direct_root_inode_operations :
+			&autofs4_indirect_root_inode_operations;
+
 	/* Couldn't this be tested earlier? */
-	if (maxproto < AUTOFS_MIN_PROTO_VERSION ||
-	    minproto > AUTOFS_MAX_PROTO_VERSION) {
+	if (sbi->max_proto < AUTOFS_MIN_PROTO_VERSION ||
+	    sbi->min_proto > AUTOFS_MAX_PROTO_VERSION) {
 		printk("autofs: kernel does not match daemon version "
 		       "daemon (%d, %d) kernel (%d, %d)\n",
-			minproto, maxproto,
+			sbi->min_proto, sbi->max_proto,
 			AUTOFS_MIN_PROTO_VERSION, AUTOFS_MAX_PROTO_VERSION);
 		goto fail_dput;
 	}
 
-	sbi->version = maxproto > AUTOFS_MAX_PROTO_VERSION ? AUTOFS_MAX_PROTO_VERSION : maxproto;
+	/* Establish highest kernel protocol version */
+	if (sbi->max_proto > AUTOFS_MAX_PROTO_VERSION)
+		sbi->version = AUTOFS_MAX_PROTO_VERSION;
+	else
+		sbi->version = sbi->max_proto;
 	sbi->sub_version = AUTOFS_PROTO_SUBVERSION;
 
 	DPRINTK("pipe fd = %d, pgrp = %u", pipefd, sbi->oz_pgrp);
@@ -329,6 +393,7 @@ int autofs4_fill_super(struct super_block *s, void *data, int silent)
 	if ( !pipe->f_op || !pipe->f_op->write )
 		goto fail_fput;
 	sbi->pipe = pipe;
+	sbi->pipefd = pipefd;
 
 	/*
 	 * Take a reference to the root dentry so we get a chance to
@@ -356,6 +421,8 @@ fail_dput:
 fail_iput:
 	printk("autofs: get root dentry failed\n");
 	iput(root_inode);
+fail_ino:
+	kfree(ino);
 fail_free:
 	kfree(sbi);
 fail_unlock:
diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c
index 62d8d4acb8bb..84e030c8ddd0 100644
--- a/fs/autofs4/root.c
+++ b/fs/autofs4/root.c
@@ -4,7 +4,7 @@
  *
  *  Copyright 1997-1998 Transmeta Corporation -- All Rights Reserved
  *  Copyright 1999-2000 Jeremy Fitzhardinge <jeremy@goop.org>
- *  Copyright 2001-2003 Ian Kent <raven@themaw.net>
+ *  Copyright 2001-2006 Ian Kent <raven@themaw.net>
  *
  * This file is part of the Linux kernel and is made available under
  * the terms of the GNU General Public License, version 2, or at your
@@ -30,9 +30,9 @@ static int autofs4_dir_close(struct inode *inode, struct file *file);
 static int autofs4_dir_readdir(struct file * filp, void * dirent, filldir_t filldir);
 static int autofs4_root_readdir(struct file * filp, void * dirent, filldir_t filldir);
 static struct dentry *autofs4_lookup(struct inode *,struct dentry *, struct nameidata *);
-static int autofs4_dcache_readdir(struct file *, void *, filldir_t);
+static void *autofs4_follow_link(struct dentry *, struct nameidata *);
 
-struct file_operations autofs4_root_operations = {
+const struct file_operations autofs4_root_operations = {
 	.open		= dcache_dir_open,
 	.release	= dcache_dir_close,
 	.read		= generic_read_dir,
@@ -40,14 +40,14 @@ struct file_operations autofs4_root_operations = {
 	.ioctl		= autofs4_root_ioctl,
 };
 
-struct file_operations autofs4_dir_operations = {
+const struct file_operations autofs4_dir_operations = {
 	.open		= autofs4_dir_open,
 	.release	= autofs4_dir_close,
 	.read		= generic_read_dir,
 	.readdir	= autofs4_dir_readdir,
 };
 
-struct inode_operations autofs4_root_inode_operations = {
+struct inode_operations autofs4_indirect_root_inode_operations = {
 	.lookup		= autofs4_lookup,
 	.unlink		= autofs4_dir_unlink,
 	.symlink	= autofs4_dir_symlink,
@@ -55,6 +55,14 @@ struct inode_operations autofs4_root_inode_operations = {
 	.rmdir		= autofs4_dir_rmdir,
 };
 
+struct inode_operations autofs4_direct_root_inode_operations = {
+	.lookup		= autofs4_lookup,
+	.unlink		= autofs4_dir_unlink,
+	.mkdir		= autofs4_dir_mkdir,
+	.rmdir		= autofs4_dir_rmdir,
+	.follow_link	= autofs4_follow_link,
+};
+
 struct inode_operations autofs4_dir_inode_operations = {
 	.lookup		= autofs4_lookup,
 	.unlink		= autofs4_dir_unlink,
@@ -82,87 +90,7 @@ static int autofs4_root_readdir(struct file *file, void *dirent,
 
 	DPRINTK("needs_reghost = %d", sbi->needs_reghost);
 
-	return autofs4_dcache_readdir(file, dirent, filldir);
-}
-
-/* Update usage from here to top of tree, so that scan of
-   top-level directories will give a useful result */
-static void autofs4_update_usage(struct vfsmount *mnt, struct dentry *dentry)
-{
-	struct dentry *top = dentry->d_sb->s_root;
-
-	spin_lock(&dcache_lock);
-	for(; dentry != top; dentry = dentry->d_parent) {
-		struct autofs_info *ino = autofs4_dentry_ino(dentry);
-
-		if (ino) {
-			touch_atime(mnt, dentry);
-			ino->last_used = jiffies;
-		}
-	}
-	spin_unlock(&dcache_lock);
-}
-
-/*
- * From 2.4 kernel readdir.c
- */
-static int autofs4_dcache_readdir(struct file * filp, void * dirent, filldir_t filldir)
-{
-	int i;
-	struct dentry *dentry = filp->f_dentry;
-
-	i = filp->f_pos;
-	switch (i) {
-		case 0:
-			if (filldir(dirent, ".", 1, i, dentry->d_inode->i_ino, DT_DIR) < 0)
-				break;
-			i++;
-			filp->f_pos++;
-			/* fallthrough */
-		case 1:
-			if (filldir(dirent, "..", 2, i, dentry->d_parent->d_inode->i_ino, DT_DIR) < 0)
-				break;
-			i++;
-			filp->f_pos++;
-			/* fallthrough */
-		default: {
-			struct list_head *list;
-			int j = i-2;
-
-			spin_lock(&dcache_lock);
-			list = dentry->d_subdirs.next;
-
-			for (;;) {
-				if (list == &dentry->d_subdirs) {
-					spin_unlock(&dcache_lock);
-					return 0;
-				}
-				if (!j)
-					break;
-				j--;
-				list = list->next;
-			}
-
-			while(1) {
-				struct dentry *de = list_entry(list,
-						struct dentry, d_u.d_child);
-
-				if (!d_unhashed(de) && de->d_inode) {
-					spin_unlock(&dcache_lock);
-					if (filldir(dirent, de->d_name.name, de->d_name.len, filp->f_pos, de->d_inode->i_ino, DT_UNKNOWN) < 0)
-						break;
-					spin_lock(&dcache_lock);
-				}
-				filp->f_pos++;
-				list = list->next;
-				if (list != &dentry->d_subdirs)
-					continue;
-				spin_unlock(&dcache_lock);
-				break;
-			}
-		}
-	}
-	return 0;
+	return dcache_readdir(file, dirent, filldir);
 }
 
 static int autofs4_dir_open(struct inode *inode, struct file *file)
@@ -170,8 +98,16 @@ static int autofs4_dir_open(struct inode *inode, struct file *file)
 	struct dentry *dentry = file->f_dentry;
 	struct vfsmount *mnt = file->f_vfsmnt;
 	struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
+	struct dentry *cursor;
 	int status;
 
+	status = dcache_dir_open(inode, file);
+	if (status)
+		goto out;
+
+	cursor = file->private_data;
+	cursor->d_fsdata = NULL;
+
 	DPRINTK("file=%p dentry=%p %.*s",
 		file, dentry, dentry->d_name.len, dentry->d_name.name);
 
@@ -180,12 +116,15 @@ static int autofs4_dir_open(struct inode *inode, struct file *file)
 
 	if (autofs4_ispending(dentry)) {
 		DPRINTK("dentry busy");
-		return -EBUSY;
+		dcache_dir_close(inode, file);
+		status = -EBUSY;
+		goto out;
 	}
 
+	status = -ENOENT;
 	if (!d_mountpoint(dentry) && dentry->d_op && dentry->d_op->d_revalidate) {
 		struct nameidata nd;
-		int empty;
+		int empty, ret;
 
 		/* In case there are stale directory dentrys from a failed mount */
 		spin_lock(&dcache_lock);
@@ -195,13 +134,13 @@ static int autofs4_dir_open(struct inode *inode, struct file *file)
 		if (!empty)
 			d_invalidate(dentry);
 
-		nd.dentry = dentry;
-		nd.mnt = mnt;
 		nd.flags = LOOKUP_DIRECTORY;
-		status = (dentry->d_op->d_revalidate)(dentry, &nd);
+		ret = (dentry->d_op->d_revalidate)(dentry, &nd);
 
-		if (!status)
-			return -ENOENT;
+		if (!ret) {
+			dcache_dir_close(inode, file);
+			goto out;
+		}
 	}
 
 	if (d_mountpoint(dentry)) {
@@ -212,25 +151,29 @@ static int autofs4_dir_open(struct inode *inode, struct file *file)
 		if (!autofs4_follow_mount(&fp_mnt, &fp_dentry)) {
 			dput(fp_dentry);
 			mntput(fp_mnt);
-			return -ENOENT;
+			dcache_dir_close(inode, file);
+			goto out;
 		}
 
 		fp = dentry_open(fp_dentry, fp_mnt, file->f_flags);
 		status = PTR_ERR(fp);
 		if (IS_ERR(fp)) {
-			file->private_data = NULL;
-			return status;
+			dcache_dir_close(inode, file);
+			goto out;
 		}
-		file->private_data = fp;
+		cursor->d_fsdata = fp;
 	}
-out:
 	return 0;
+out:
+	return status;
 }
 
 static int autofs4_dir_close(struct inode *inode, struct file *file)
 {
 	struct dentry *dentry = file->f_dentry;
 	struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
+	struct dentry *cursor = file->private_data;
+	int status = 0;
 
 	DPRINTK("file=%p dentry=%p %.*s",
 		file, dentry, dentry->d_name.len, dentry->d_name.name);
@@ -240,26 +183,28 @@ static int autofs4_dir_close(struct inode *inode, struct file *file)
 
 	if (autofs4_ispending(dentry)) {
 		DPRINTK("dentry busy");
-		return -EBUSY;
+		status = -EBUSY;
+		goto out;
 	}
 
 	if (d_mountpoint(dentry)) {
-		struct file *fp = file->private_data;
-
-		if (!fp)
-			return -ENOENT;
-
+		struct file *fp = cursor->d_fsdata;
+		if (!fp) {
+			status = -ENOENT;
+			goto out;
+		}
 		filp_close(fp, current->files);
-		file->private_data = NULL;
 	}
 out:
-	return 0;
+	dcache_dir_close(inode, file);
+	return status;
 }
 
 static int autofs4_dir_readdir(struct file *file, void *dirent, filldir_t filldir)
 {
 	struct dentry *dentry = file->f_dentry;
 	struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
+	struct dentry *cursor = file->private_data;
 	int status;
 
 	DPRINTK("file=%p dentry=%p %.*s",
@@ -274,7 +219,7 @@ static int autofs4_dir_readdir(struct file *file, void *dirent, filldir_t filldi
 	}
 
 	if (d_mountpoint(dentry)) {
-		struct file *fp = file->private_data;
+		struct file *fp = cursor->d_fsdata;
 
 		if (!fp)
 			return -ENOENT;
@@ -289,27 +234,26 @@ static int autofs4_dir_readdir(struct file *file, void *dirent, filldir_t filldi
 		return status;
 	}
 out:
-	return autofs4_dcache_readdir(file, dirent, filldir);
+	return dcache_readdir(file, dirent, filldir);
 }
 
-static int try_to_fill_dentry(struct vfsmount *mnt, struct dentry *dentry, int flags)
+static int try_to_fill_dentry(struct dentry *dentry, int flags)
 {
-	struct super_block *sb = mnt->mnt_sb;
-	struct autofs_sb_info *sbi = autofs4_sbi(sb);
-	struct autofs_info *de_info = autofs4_dentry_ino(dentry);
+	struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
+	struct autofs_info *ino = autofs4_dentry_ino(dentry);
 	int status = 0;
 
 	/* Block on any pending expiry here; invalidate the dentry
            when expiration is done to trigger mount request with a new
            dentry */
-	if (de_info && (de_info->flags & AUTOFS_INF_EXPIRING)) {
+	if (ino && (ino->flags & AUTOFS_INF_EXPIRING)) {
 		DPRINTK("waiting for expire %p name=%.*s",
 			 dentry, dentry->d_name.len, dentry->d_name.name);
 
 		status = autofs4_wait(sbi, dentry, NFY_NONE);
-		
+
 		DPRINTK("expire done status=%d", status);
-		
+
 		/*
 		 * If the directory still exists the mount request must
 		 * continue otherwise it can't be followed at the right
@@ -317,34 +261,36 @@ static int try_to_fill_dentry(struct vfsmount *mnt, struct dentry *dentry, int f
 		 */
 		status = d_invalidate(dentry);
 		if (status != -EBUSY)
-			return 0;
+			return -ENOENT;
 	}
 
 	DPRINTK("dentry=%p %.*s ino=%p",
 		 dentry, dentry->d_name.len, dentry->d_name.name, dentry->d_inode);
 
-	/* Wait for a pending mount, triggering one if there isn't one already */
+	/*
+	 * Wait for a pending mount, triggering one if there
+	 * isn't one already
+	 */
 	if (dentry->d_inode == NULL) {
 		DPRINTK("waiting for mount name=%.*s",
 			 dentry->d_name.len, dentry->d_name.name);
 
 		status = autofs4_wait(sbi, dentry, NFY_MOUNT);
-		 
+
 		DPRINTK("mount done status=%d", status);
 
 		if (status && dentry->d_inode)
-			return 0; /* Try to get the kernel to invalidate this dentry */
-		
+			return status; /* Try to get the kernel to invalidate this dentry */
+
 		/* Turn this into a real negative dentry? */
 		if (status == -ENOENT) {
-			dentry->d_time = jiffies + AUTOFS_NEGATIVE_TIMEOUT;
 			spin_lock(&dentry->d_lock);
 			dentry->d_flags &= ~DCACHE_AUTOFS_PENDING;
 			spin_unlock(&dentry->d_lock);
-			return 1;
+			return status;
 		} else if (status) {
 			/* Return a negative dentry, but leave it "pending" */
-			return 1;
+			return status;
 		}
 	/* Trigger mount for path component or follow link */
 	} else if (flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY) ||
@@ -363,19 +309,87 @@ static int try_to_fill_dentry(struct vfsmount *mnt, struct dentry *dentry, int f
 			spin_lock(&dentry->d_lock);
 			dentry->d_flags &= ~DCACHE_AUTOFS_PENDING;
 			spin_unlock(&dentry->d_lock);
-			return 0;
+			return status;
 		}
 	}
 
-	/* We don't update the usages for the autofs daemon itself, this
-	   is necessary for recursive autofs mounts */
-	if (!autofs4_oz_mode(sbi))
-		autofs4_update_usage(mnt, dentry);
+	/* Initialize expiry counter after successful mount */
+	if (ino)
+		ino->last_used = jiffies;
 
 	spin_lock(&dentry->d_lock);
 	dentry->d_flags &= ~DCACHE_AUTOFS_PENDING;
 	spin_unlock(&dentry->d_lock);
-	return 1;
+	return status;
+}
+
+/* For autofs direct mounts the follow link triggers the mount */
+static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd)
+{
+	struct autofs_sb_info *sbi = autofs4_sbi(dentry->d_sb);
+	int oz_mode = autofs4_oz_mode(sbi);
+	unsigned int lookup_type;
+	int status;
+
+	DPRINTK("dentry=%p %.*s oz_mode=%d nd->flags=%d",
+		dentry, dentry->d_name.len, dentry->d_name.name, oz_mode,
+		nd->flags);
+
+	/* If it's our master or we shouldn't trigger a mount we're done */
+	lookup_type = nd->flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY);
+	if (oz_mode || !lookup_type)
+		goto done;
+
+	/*
+	 * If a request is pending wait for it.
+	 * If it's a mount then it won't be expired till at least
+	 * a liitle later and if it's an expire then we might need
+	 * to mount it again.
+	 */
+	if (autofs4_ispending(dentry)) {
+		DPRINTK("waiting for active request %p name=%.*s",
+			dentry, dentry->d_name.len, dentry->d_name.name);
+
+		status = autofs4_wait(sbi, dentry, NFY_NONE);
+
+		DPRINTK("request done status=%d", status);
+	}
+
+	/*
+	 * If the dentry contains directories then it is an
+	 * autofs multi-mount with no root mount offset. So
+	 * don't try to mount it again.
+	 */
+	spin_lock(&dcache_lock);
+	if (!d_mountpoint(dentry) && list_empty(&dentry->d_subdirs)) {
+		spin_unlock(&dcache_lock);
+
+		status = try_to_fill_dentry(dentry, 0);
+		if (status)
+			goto out_error;
+
+		/*
+		 * The mount succeeded but if there is no root mount
+		 * it must be an autofs multi-mount with no root offset
+		 * so we don't need to follow the mount.
+		 */
+		if (d_mountpoint(dentry)) {
+			if (!autofs4_follow_mount(&nd->mnt, &nd->dentry)) {
+				status = -ENOENT;
+				goto out_error;
+			}
+		}
+
+		goto done;
+	}
+	spin_unlock(&dcache_lock);
+
+done:
+	return NULL;
+
+out_error:
+	path_release(nd);
+	return ERR_PTR(status);
 }
 
 /*
@@ -384,47 +398,43 @@ static int try_to_fill_dentry(struct vfsmount *mnt, struct dentry *dentry, int f
  * yet completely filled in, and revalidate has to delay such
  * lookups..
  */
-static int autofs4_revalidate(struct dentry * dentry, struct nameidata *nd)
+static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
-	struct inode * dir = dentry->d_parent->d_inode;
+	struct inode *dir = dentry->d_parent->d_inode;
 	struct autofs_sb_info *sbi = autofs4_sbi(dir->i_sb);
 	int oz_mode = autofs4_oz_mode(sbi);
 	int flags = nd ? nd->flags : 0;
-	int status = 1;
+	int status = 0;
 
 	/* Pending dentry */
 	if (autofs4_ispending(dentry)) {
 		if (!oz_mode)
-			status = try_to_fill_dentry(nd->mnt, dentry, flags);
-		return status;
+			status = try_to_fill_dentry(dentry, flags);
+		return !status;
 	}
 
 	/* Negative dentry.. invalidate if "old" */
 	if (dentry->d_inode == NULL)
-		return (dentry->d_time - jiffies <= AUTOFS_NEGATIVE_TIMEOUT);
+		return 0;
 
 	/* Check for a non-mountpoint directory with no contents */
 	spin_lock(&dcache_lock);
 	if (S_ISDIR(dentry->d_inode->i_mode) &&
 	    !d_mountpoint(dentry) && 
-	    list_empty(&dentry->d_subdirs)) {
+	    __simple_empty(dentry)) {
 		DPRINTK("dentry=%p %.*s, emptydir",
 			 dentry, dentry->d_name.len, dentry->d_name.name);
 		spin_unlock(&dcache_lock);
 		if (!oz_mode)
-			status = try_to_fill_dentry(nd->mnt, dentry, flags);
-		return status;
+			status = try_to_fill_dentry(dentry, flags);
+		return !status;
 	}
 	spin_unlock(&dcache_lock);
 
-	/* Update the usage list */
-	if (!oz_mode)
-		autofs4_update_usage(nd->mnt, dentry);
-
 	return 1;
 }
 
-static void autofs4_dentry_release(struct dentry *de)
+void autofs4_dentry_release(struct dentry *de)
 {
 	struct autofs_info *inf;
 
@@ -462,12 +472,13 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s
 	DPRINTK("name = %.*s",
 		dentry->d_name.len, dentry->d_name.name);
 
+	/* File name too long to exist */
 	if (dentry->d_name.len > NAME_MAX)
-		return ERR_PTR(-ENAMETOOLONG);/* File name too long to exist */
+		return ERR_PTR(-ENAMETOOLONG);
 
 	sbi = autofs4_sbi(dir->i_sb);
-
 	oz_mode = autofs4_oz_mode(sbi);
+
 	DPRINTK("pid = %u, pgrp = %u, catatonic = %d, oz_mode = %d",
 		 current->pid, process_group(current), sbi->catatonic, oz_mode);
 
@@ -519,7 +530,7 @@ static struct dentry *autofs4_lookup(struct inode *dir, struct dentry *dentry, s
 	 * doesn't do the right thing for all system calls, but it should
 	 * be OK for the operations we permit from an autofs.
 	 */
-	if ( dentry->d_inode && d_unhashed(dentry) )
+	if (dentry->d_inode && d_unhashed(dentry))
 		return ERR_PTR(-ENOENT);
 
 	return NULL;
@@ -531,6 +542,7 @@ static int autofs4_dir_symlink(struct inode *dir,
 {
 	struct autofs_sb_info *sbi = autofs4_sbi(dir->i_sb);
 	struct autofs_info *ino = autofs4_dentry_ino(dentry);
+	struct autofs_info *p_ino;
 	struct inode *inode;
 	char *cp;
 
@@ -564,6 +576,10 @@ static int autofs4_dir_symlink(struct inode *dir,
 
 	dentry->d_fsdata = ino;
 	ino->dentry = dget(dentry);
+	atomic_inc(&ino->count);
+	p_ino = autofs4_dentry_ino(dentry->d_parent);
+	if (p_ino && dentry->d_parent != dentry)
+		atomic_inc(&p_ino->count);
 	ino->inode = inode;
 
 	dir->i_mtime = CURRENT_TIME;
@@ -590,11 +606,17 @@ static int autofs4_dir_unlink(struct inode *dir, struct dentry *dentry)
 {
 	struct autofs_sb_info *sbi = autofs4_sbi(dir->i_sb);
 	struct autofs_info *ino = autofs4_dentry_ino(dentry);
+	struct autofs_info *p_ino;
 	
 	/* This allows root to remove symlinks */
 	if ( !autofs4_oz_mode(sbi) && !capable(CAP_SYS_ADMIN) )
 		return -EACCES;
 
+	if (atomic_dec_and_test(&ino->count)) {
+		p_ino = autofs4_dentry_ino(dentry->d_parent);
+		if (p_ino && dentry->d_parent != dentry)
+			atomic_dec(&p_ino->count);
+	}
 	dput(ino->dentry);
 
 	dentry->d_inode->i_size = 0;
@@ -611,6 +633,7 @@ static int autofs4_dir_rmdir(struct inode *dir, struct dentry *dentry)
 {
 	struct autofs_sb_info *sbi = autofs4_sbi(dir->i_sb);
 	struct autofs_info *ino = autofs4_dentry_ino(dentry);
+	struct autofs_info *p_ino;
 	
 	if (!autofs4_oz_mode(sbi))
 		return -EACCES;
@@ -625,8 +648,12 @@ static int autofs4_dir_rmdir(struct inode *dir, struct dentry *dentry)
 	spin_unlock(&dentry->d_lock);
 	spin_unlock(&dcache_lock);
 
+	if (atomic_dec_and_test(&ino->count)) {
+		p_ino = autofs4_dentry_ino(dentry->d_parent);
+		if (p_ino && dentry->d_parent != dentry)
+			atomic_dec(&p_ino->count);
+	}
 	dput(ino->dentry);
-
 	dentry->d_inode->i_size = 0;
 	dentry->d_inode->i_nlink = 0;
 
@@ -640,6 +667,7 @@ static int autofs4_dir_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 {
 	struct autofs_sb_info *sbi = autofs4_sbi(dir->i_sb);
 	struct autofs_info *ino = autofs4_dentry_ino(dentry);
+	struct autofs_info *p_ino;
 	struct inode *inode;
 
 	if ( !autofs4_oz_mode(sbi) )
@@ -662,6 +690,10 @@ static int autofs4_dir_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 
 	dentry->d_fsdata = ino;
 	ino->dentry = dget(dentry);
+	atomic_inc(&ino->count);
+	p_ino = autofs4_dentry_ino(dentry->d_parent);
+	if (p_ino && dentry->d_parent != dentry)
+		atomic_inc(&p_ino->count);
 	ino->inode = inode;
 	dir->i_nlink++;
 	dir->i_mtime = CURRENT_TIME;
@@ -745,7 +777,7 @@ static inline int autofs4_ask_umount(struct vfsmount *mnt, int __user *p)
 {
 	int status = 0;
 
-	if (may_umount(mnt) == 0)
+	if (may_umount(mnt))
 		status = 1;
 
 	DPRINTK("returning %d", status);
diff --git a/fs/autofs4/waitq.c b/fs/autofs4/waitq.c
index 394ff36ef8f1..142ab6aa2aa1 100644
--- a/fs/autofs4/waitq.c
+++ b/fs/autofs4/waitq.c
@@ -3,7 +3,7 @@
  * linux/fs/autofs/waitq.c
  *
  *  Copyright 1997-1998 Transmeta Corporation -- All Rights Reserved
- *  Copyright 2001-2003 Ian Kent <raven@themaw.net>
+ *  Copyright 2001-2006 Ian Kent <raven@themaw.net>
  *
  * This file is part of the Linux kernel and is made available under
  * the terms of the GNU General Public License, version 2, or at your
@@ -33,7 +33,7 @@ void autofs4_catatonic_mode(struct autofs_sb_info *sbi)
 	sbi->catatonic = 1;
 	wq = sbi->queues;
 	sbi->queues = NULL;	/* Erase all wait queues */
-	while ( wq ) {
+	while (wq) {
 		nwq = wq->next;
 		wq->status = -ENOENT; /* Magic is gone - report failure */
 		kfree(wq->name);
@@ -45,7 +45,6 @@ void autofs4_catatonic_mode(struct autofs_sb_info *sbi)
 		fput(sbi->pipe);	/* Close the pipe */
 		sbi->pipe = NULL;
 	}
-
 	shrink_dcache_sb(sbi->sb);
 }
 
@@ -98,7 +97,10 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi,
 
 	pkt.hdr.proto_version = sbi->version;
 	pkt.hdr.type = type;
-	if (type == autofs_ptype_missing) {
+	switch (type) {
+	/* Kernel protocol v4 missing and expire packets */
+	case autofs_ptype_missing:
+	{
 		struct autofs_packet_missing *mp = &pkt.missing;
 
 		pktsz = sizeof(*mp);
@@ -107,7 +109,10 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi,
 		mp->len = wq->len;
 		memcpy(mp->name, wq->name, wq->len);
 		mp->name[wq->len] = '\0';
-	} else if (type == autofs_ptype_expire_multi) {
+		break;
+	}
+	case autofs_ptype_expire_multi:
+	{
 		struct autofs_packet_expire_multi *ep = &pkt.expire_multi;
 
 		pktsz = sizeof(*ep);
@@ -116,7 +121,34 @@ static void autofs4_notify_daemon(struct autofs_sb_info *sbi,
 		ep->len = wq->len;
 		memcpy(ep->name, wq->name, wq->len);
 		ep->name[wq->len] = '\0';
-	} else {
+		break;
+	}
+	/*
+	 * Kernel protocol v5 packet for handling indirect and direct
+	 * mount missing and expire requests
+	 */
+	case autofs_ptype_missing_indirect:
+	case autofs_ptype_expire_indirect:
+	case autofs_ptype_missing_direct:
+	case autofs_ptype_expire_direct:
+	{
+		struct autofs_v5_packet *packet = &pkt.v5_packet;
+
+		pktsz = sizeof(*packet);
+
+		packet->wait_queue_token = wq->wait_queue_token;
+		packet->len = wq->len;
+		memcpy(packet->name, wq->name, wq->len);
+		packet->name[wq->len] = '\0';
+		packet->dev = wq->dev;
+		packet->ino = wq->ino;
+		packet->uid = wq->uid;
+		packet->gid = wq->gid;
+		packet->pid = wq->pid;
+		packet->tgid = wq->tgid;
+		break;
+	}
+	default:
 		printk("autofs4_notify_daemon: bad type %d!\n", type);
 		return;
 	}
@@ -162,23 +194,31 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 {
 	struct autofs_wait_queue *wq;
 	char *name;
-	int len, status;
+	unsigned int len = 0;
+	unsigned int hash = 0;
+	int status;
 
 	/* In catatonic mode, we don't wait for nobody */
-	if ( sbi->catatonic )
+	if (sbi->catatonic)
 		return -ENOENT;
 	
 	name = kmalloc(NAME_MAX + 1, GFP_KERNEL);
 	if (!name)
 		return -ENOMEM;
 
-	len = autofs4_getpath(sbi, dentry, &name);
-	if (!len) {
-		kfree(name);
-		return -ENOENT;
+	/* If this is a direct mount request create a dummy name */
+	if (IS_ROOT(dentry) && (sbi->type & AUTOFS_TYPE_DIRECT))
+		len = sprintf(name, "%p", dentry);
+	else {
+		len = autofs4_getpath(sbi, dentry, &name);
+		if (!len) {
+			kfree(name);
+			return -ENOENT;
+		}
 	}
+	hash = full_name_hash(name, len);
 
-	if (down_interruptible(&sbi->wq_sem)) {
+	if (mutex_lock_interruptible(&sbi->wq_mutex)) {
 		kfree(name);
 		return -EINTR;
 	}
@@ -190,19 +230,19 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 			break;
 	}
 
-	if ( !wq ) {
+	if (!wq) {
 		/* Can't wait for an expire if there's no mount */
 		if (notify == NFY_NONE && !d_mountpoint(dentry)) {
 			kfree(name);
-			up(&sbi->wq_sem);
+			mutex_unlock(&sbi->wq_mutex);
 			return -ENOENT;
 		}
 
 		/* Create a new wait queue */
 		wq = kmalloc(sizeof(struct autofs_wait_queue),GFP_KERNEL);
-		if ( !wq ) {
+		if (!wq) {
 			kfree(name);
-			up(&sbi->wq_sem);
+			mutex_unlock(&sbi->wq_mutex);
 			return -ENOMEM;
 		}
 
@@ -212,24 +252,47 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 		wq->next = sbi->queues;
 		sbi->queues = wq;
 		init_waitqueue_head(&wq->queue);
-		wq->hash = dentry->d_name.hash;
+		wq->hash = hash;
 		wq->name = name;
 		wq->len = len;
+		wq->dev = autofs4_get_dev(sbi);
+		wq->ino = autofs4_get_ino(sbi);
+		wq->uid = current->uid;
+		wq->gid = current->gid;
+		wq->pid = current->pid;
+		wq->tgid = current->tgid;
 		wq->status = -EINTR; /* Status return if interrupted */
 		atomic_set(&wq->wait_ctr, 2);
-		atomic_set(&wq->notified, 1);
-		up(&sbi->wq_sem);
+		atomic_set(&wq->notify, 1);
+		mutex_unlock(&sbi->wq_mutex);
 	} else {
 		atomic_inc(&wq->wait_ctr);
-		up(&sbi->wq_sem);
+		mutex_unlock(&sbi->wq_mutex);
 		kfree(name);
 		DPRINTK("existing wait id = 0x%08lx, name = %.*s, nfy=%d",
 			(unsigned long) wq->wait_queue_token, wq->len, wq->name, notify);
 	}
 
-	if (notify != NFY_NONE && atomic_dec_and_test(&wq->notified)) {
-		int type = (notify == NFY_MOUNT ?
-			autofs_ptype_missing : autofs_ptype_expire_multi);
+	if (notify != NFY_NONE && atomic_read(&wq->notify)) {
+		int type;
+
+		atomic_dec(&wq->notify);
+
+		if (sbi->version < 5) {
+			if (notify == NFY_MOUNT)
+				type = autofs_ptype_missing;
+			else
+				type = autofs_ptype_expire_multi;
+		} else {
+			if (notify == NFY_MOUNT)
+				type = (sbi->type & AUTOFS_TYPE_DIRECT) ?
+					autofs_ptype_missing_direct :
+					 autofs_ptype_missing_indirect;
+			else
+				type = (sbi->type & AUTOFS_TYPE_DIRECT) ?
+					autofs_ptype_expire_direct :
+					autofs_ptype_expire_indirect;
+		}
 
 		DPRINTK("new wait id = 0x%08lx, name = %.*s, nfy=%d\n",
 			(unsigned long) wq->wait_queue_token, wq->len, wq->name, notify);
@@ -240,14 +303,14 @@ int autofs4_wait(struct autofs_sb_info *sbi, struct dentry *dentry,
 
 	/* wq->name is NULL if and only if the lock is already released */
 
-	if ( sbi->catatonic ) {
+	if (sbi->catatonic) {
 		/* We might have slept, so check again for catatonic mode */
 		wq->status = -ENOENT;
 		kfree(wq->name);
 		wq->name = NULL;
 	}
 
-	if ( wq->name ) {
+	if (wq->name) {
 		/* Block all but "shutdown" signals while waiting */
 		sigset_t oldset;
 		unsigned long irqflags;
@@ -282,19 +345,19 @@ int autofs4_wait_release(struct autofs_sb_info *sbi, autofs_wqt_t wait_queue_tok
 {
 	struct autofs_wait_queue *wq, **wql;
 
-	down(&sbi->wq_sem);
-	for ( wql = &sbi->queues ; (wq = *wql) != 0 ; wql = &wq->next ) {
-		if ( wq->wait_queue_token == wait_queue_token )
+	mutex_lock(&sbi->wq_mutex);
+	for (wql = &sbi->queues ; (wq = *wql) != 0 ; wql = &wq->next) {
+		if (wq->wait_queue_token == wait_queue_token)
 			break;
 	}
 
-	if ( !wq ) {
-		up(&sbi->wq_sem);
+	if (!wq) {
+		mutex_unlock(&sbi->wq_mutex);
 		return -EINVAL;
 	}
 
 	*wql = wq->next;	/* Unlink from chain */
-	up(&sbi->wq_sem);
+	mutex_unlock(&sbi->wq_mutex);
 	kfree(wq->name);
 	wq->name = NULL;	/* Do not wait on this queue */
 
diff --git a/fs/bad_inode.c b/fs/bad_inode.c
index e172180a1d8c..80599ae33966 100644
--- a/fs/bad_inode.c
+++ b/fs/bad_inode.c
@@ -22,7 +22,7 @@ static int return_EIO(void)
 
 #define EIO_ERROR ((void *) (return_EIO))
 
-static struct file_operations bad_file_ops =
+static const struct file_operations bad_file_ops =
 {
 	.llseek		= EIO_ERROR,
 	.aio_read	= EIO_ERROR,
diff --git a/fs/befs/datastream.c b/fs/befs/datastream.c
index 785f6b2d5d10..b7d6b920f65f 100644
--- a/fs/befs/datastream.c
+++ b/fs/befs/datastream.c
@@ -118,7 +118,7 @@ befs_fblock2brun(struct super_block *sb, befs_data_stream * data,
  * befs_read_lsmylink - read long symlink from datastream.
  * @sb: Filesystem superblock 
  * @ds: Datastrem to read from
- * @buf: Buffer in wich to place long symlink data
+ * @buf: Buffer in which to place long symlink data
  * @len: Length of the long symlink in bytes
  *
  * Returns the number of bytes read
diff --git a/fs/befs/linuxvfs.c b/fs/befs/linuxvfs.c
index 2d365cb8eec6..68ebd10f345d 100644
--- a/fs/befs/linuxvfs.c
+++ b/fs/befs/linuxvfs.c
@@ -64,7 +64,7 @@ static const struct super_operations befs_sops = {
 /* slab cache for befs_inode_info objects */
 static kmem_cache_t *befs_inode_cachep;
 
-static struct file_operations befs_dir_operations = {
+static const struct file_operations befs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= befs_readdir,
 };
@@ -427,7 +427,8 @@ befs_init_inodecache(void)
 {
 	befs_inode_cachep = kmem_cache_create("befs_inode_cache",
 					      sizeof (struct befs_inode_info),
-					      0, SLAB_RECLAIM_ACCOUNT,
+					      0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					      init_once, NULL);
 	if (befs_inode_cachep == NULL) {
 		printk(KERN_ERR "befs_init_inodecache: "
@@ -561,7 +562,7 @@ befs_utf2nls(struct super_block *sb, const char *in,
  * @sb: Superblock
  * @src: Input string buffer in NLS format
  * @srclen: Length of input string in bytes
- * @dest: The output string in UTF8 format
+ * @dest: The output string in UTF-8 format
  * @destlen: Length of the output buffer
  * 
  * Converts input string @src, which is in the format of the loaded NLS map,
diff --git a/fs/bfs/bfs.h b/fs/bfs/bfs.h
index 1fbc53f14aba..9d791004b21c 100644
--- a/fs/bfs/bfs.h
+++ b/fs/bfs/bfs.h
@@ -49,11 +49,11 @@ static inline struct bfs_inode_info *BFS_I(struct inode *inode)
 
 /* file.c */
 extern struct inode_operations bfs_file_inops;
-extern struct file_operations bfs_file_operations;
+extern const struct file_operations bfs_file_operations;
 extern struct address_space_operations bfs_aops;
 
 /* dir.c */
 extern struct inode_operations bfs_dir_inops;
-extern struct file_operations bfs_dir_operations;
+extern const struct file_operations bfs_dir_operations;
 
 #endif /* _FS_BFS_BFS_H */
diff --git a/fs/bfs/dir.c b/fs/bfs/dir.c
index 5af928fa0449..26fad9621738 100644
--- a/fs/bfs/dir.c
+++ b/fs/bfs/dir.c
@@ -70,7 +70,7 @@ static int bfs_readdir(struct file * f, void * dirent, filldir_t filldir)
 	return 0;	
 }
 
-struct file_operations bfs_dir_operations = {
+const struct file_operations bfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= bfs_readdir,
 	.fsync		= file_fsync,
diff --git a/fs/bfs/file.c b/fs/bfs/file.c
index 807723b65daf..d83cd74a2e4e 100644
--- a/fs/bfs/file.c
+++ b/fs/bfs/file.c
@@ -17,7 +17,7 @@
 #define dprintf(x...)
 #endif
 
-struct file_operations bfs_file_operations = {
+const struct file_operations bfs_file_operations = {
 	.llseek 	= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/bfs/inode.c b/fs/bfs/inode.c
index 3af6c73c5b5a..55a7a78332f8 100644
--- a/fs/bfs/inode.c
+++ b/fs/bfs/inode.c
@@ -257,7 +257,8 @@ static int init_inodecache(void)
 {
 	bfs_inode_cachep = kmem_cache_create("bfs_inode_cache",
 					     sizeof(struct bfs_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (bfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c
index c2eac2a50bd2..537893a16014 100644
--- a/fs/binfmt_elf.c
+++ b/fs/binfmt_elf.c
@@ -500,17 +500,22 @@ out:
 #define INTERPRETER_AOUT 1
 #define INTERPRETER_ELF 2
 
+#ifndef STACK_RND_MASK
+#define STACK_RND_MASK 0x7ff		/* with 4K pages 8MB of VA */
+#endif
 
 static unsigned long randomize_stack_top(unsigned long stack_top)
 {
 	unsigned int random_variable = 0;
 
-	if (current->flags & PF_RANDOMIZE)
-		random_variable = get_random_int() % (8*1024*1024);
+	if (current->flags & PF_RANDOMIZE) {
+		random_variable = get_random_int() & STACK_RND_MASK;
+		random_variable <<= PAGE_SHIFT;
+	}
 #ifdef CONFIG_STACK_GROWSUP
-	return PAGE_ALIGN(stack_top + random_variable);
+	return PAGE_ALIGN(stack_top) + random_variable;
 #else
-	return PAGE_ALIGN(stack_top - random_variable);
+	return PAGE_ALIGN(stack_top) - random_variable;
 #endif
 }
 
@@ -1334,7 +1339,7 @@ static int fill_psinfo(struct elf_prpsinfo *psinfo, struct task_struct *p,
 
 	i = p->state ? ffz(~p->state) + 1 : 0;
 	psinfo->pr_state = i;
-	psinfo->pr_sname = (i < 0 || i > 5) ? '.' : "RSDTZW"[i];
+	psinfo->pr_sname = (i > 5) ? '.' : "RSDTZW"[i];
 	psinfo->pr_zomb = psinfo->pr_sname == 'Z';
 	psinfo->pr_nice = task_nice(p);
 	psinfo->pr_flag = p->flags;
@@ -1465,12 +1470,11 @@ static int elf_core_dump(long signr, struct pt_regs * regs, struct file * file)
 		read_lock(&tasklist_lock);
 		do_each_thread(g,p)
 			if (current->mm == p->mm && current != p) {
-				tmp = kmalloc(sizeof(*tmp), GFP_ATOMIC);
+				tmp = kzalloc(sizeof(*tmp), GFP_ATOMIC);
 				if (!tmp) {
 					read_unlock(&tasklist_lock);
 					goto cleanup;
 				}
-				memset(tmp, 0, sizeof(*tmp));
 				INIT_LIST_HEAD(&tmp->list);
 				tmp->thread = p;
 				list_add(&tmp->list, &thread_list);
diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c
index 5b3076e8ee90..a2e48c999c24 100644
--- a/fs/binfmt_elf_fdpic.c
+++ b/fs/binfmt_elf_fdpic.c
@@ -572,8 +572,7 @@ static int create_elf_fdpic_tables(struct linux_binprm *bprm,
 	csp -= sizeof(unsigned long);
 	__put_user(bprm->argc, (unsigned long *) csp);
 
-	if (csp != sp)
-		BUG();
+	BUG_ON(csp != sp);
 
 	/* fill in the argv[] array */
 #ifdef CONFIG_MMU
diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c
index 108d56bbd0d0..69f44dcdb0b4 100644
--- a/fs/binfmt_flat.c
+++ b/fs/binfmt_flat.c
@@ -36,6 +36,7 @@
 #include <linux/personality.h>
 #include <linux/init.h>
 #include <linux/flat.h>
+#include <linux/syscalls.h>
 
 #include <asm/byteorder.h>
 #include <asm/system.h>
@@ -426,6 +427,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 	int i, rev, relocs = 0;
 	loff_t fpos;
 	unsigned long start_code, end_code;
+	int ret;
+	int exec_fileno;
 
 	hdr = ((struct flat_hdr *) bprm->buf);		/* exec-header */
 	inode = bprm->file->f_dentry->d_inode;
@@ -450,7 +453,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 		 */
 		if (strncmp(hdr->magic, "#!", 2))
 			printk("BINFMT_FLAT: bad header magic\n");
-		return -ENOEXEC;
+		ret = -ENOEXEC;
+		goto err;
 	}
 
 	if (flags & FLAT_FLAG_KTRACE)
@@ -458,14 +462,16 @@ static int load_flat_file(struct linux_binprm * bprm,
 
 	if (rev != FLAT_VERSION && rev != OLD_FLAT_VERSION) {
 		printk("BINFMT_FLAT: bad flat file version 0x%x (supported 0x%x and 0x%x)\n", rev, FLAT_VERSION, OLD_FLAT_VERSION);
-		return -ENOEXEC;
+		ret = -ENOEXEC;
+		goto err;
 	}
 	
 	/* Don't allow old format executables to use shared libraries */
 	if (rev == OLD_FLAT_VERSION && id != 0) {
 		printk("BINFMT_FLAT: shared libraries are not available before rev 0x%x\n",
 				(int) FLAT_VERSION);
-		return -ENOEXEC;
+		ret = -ENOEXEC;
+		goto err;
 	}
 
 	/*
@@ -478,7 +484,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 #ifndef CONFIG_BINFMT_ZFLAT
 	if (flags & (FLAT_FLAG_GZIP|FLAT_FLAG_GZDATA)) {
 		printk("Support for ZFLAT executables is not enabled.\n");
-		return -ENOEXEC;
+		ret = -ENOEXEC;
+		goto err;
 	}
 #endif
 
@@ -490,14 +497,27 @@ static int load_flat_file(struct linux_binprm * bprm,
 	rlim = current->signal->rlim[RLIMIT_DATA].rlim_cur;
 	if (rlim >= RLIM_INFINITY)
 		rlim = ~0;
-	if (data_len + bss_len > rlim)
-		return -ENOMEM;
+	if (data_len + bss_len > rlim) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	/* check file descriptor */
+	exec_fileno = get_unused_fd();
+	if (exec_fileno < 0) {
+		ret = -EMFILE;
+		goto err;
+	}
+	get_file(bprm->file);
+	fd_install(exec_fileno, bprm->file);
 
 	/* Flush all traces of the currently running executable */
 	if (id == 0) {
 		result = flush_old_exec(bprm);
-		if (result)
-			return result;
+		if (result) {
+			ret = result;
+			goto err_close;
+		}
 
 		/* OK, This is the point of no return */
 		set_personality(PER_LINUX);
@@ -527,7 +547,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 			if (!textpos)
 				textpos = (unsigned long) -ENOMEM;
 			printk("Unable to mmap process text, errno %d\n", (int)-textpos);
-			return(textpos);
+			ret = textpos;
+			goto err_close;
 		}
 
 		down_write(&current->mm->mmap_sem);
@@ -542,7 +563,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 			printk("Unable to allocate RAM for process data, errno %d\n",
 					(int)-datapos);
 			do_munmap(current->mm, textpos, text_len);
-			return realdatastart;
+			ret = realdatastart;
+			goto err_close;
 		}
 		datapos = realdatastart + MAX_SHARED_LIBS * sizeof(unsigned long);
 
@@ -564,7 +586,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 			printk("Unable to read data+bss, errno %d\n", (int)-result);
 			do_munmap(current->mm, textpos, text_len);
 			do_munmap(current->mm, realdatastart, data_len + extra);
-			return result;
+			ret = result;
+			goto err_close;
 		}
 
 		reloc = (unsigned long *) (datapos+(ntohl(hdr->reloc_start)-text_len));
@@ -582,7 +605,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 				textpos = (unsigned long) -ENOMEM;
 			printk("Unable to allocate RAM for process text/data, errno %d\n",
 					(int)-textpos);
-			return(textpos);
+			ret = textpos;
+			goto err_close;
 		}
 
 		realdatastart = textpos + ntohl(hdr->data_start);
@@ -627,7 +651,8 @@ static int load_flat_file(struct linux_binprm * bprm,
 			printk("Unable to read code+data+bss, errno %d\n",(int)-result);
 			do_munmap(current->mm, textpos, text_len + data_len + extra +
 				MAX_SHARED_LIBS * sizeof(unsigned long));
-			return result;
+			ret = result;
+			goto err_close;
 		}
 	}
 
@@ -690,8 +715,10 @@ static int load_flat_file(struct linux_binprm * bprm,
 			unsigned long addr;
 			if (*rp) {
 				addr = calc_reloc(*rp, libinfo, id, 0);
-				if (addr == RELOC_FAILED)
-					return -ENOEXEC;
+				if (addr == RELOC_FAILED) {
+					ret = -ENOEXEC;
+					goto err_close;
+				}
 				*rp = addr;
 			}
 		}
@@ -718,8 +745,10 @@ static int load_flat_file(struct linux_binprm * bprm,
 			relval = ntohl(reloc[i]);
 			addr = flat_get_relocate_addr(relval);
 			rp = (unsigned long *) calc_reloc(addr, libinfo, id, 1);
-			if (rp == (unsigned long *)RELOC_FAILED)
-				return -ENOEXEC;
+			if (rp == (unsigned long *)RELOC_FAILED) {
+				ret = -ENOEXEC;
+				goto err_close;
+			}
 
 			/* Get the pointer's value.  */
 			addr = flat_get_addr_from_rp(rp, relval, flags);
@@ -731,8 +760,10 @@ static int load_flat_file(struct linux_binprm * bprm,
 				if ((flags & FLAT_FLAG_GOTPIC) == 0)
 					addr = ntohl(addr);
 				addr = calc_reloc(addr, libinfo, id, 0);
-				if (addr == RELOC_FAILED)
-					return -ENOEXEC;
+				if (addr == RELOC_FAILED) {
+					ret = -ENOEXEC;
+					goto err_close;
+				}
 
 				/* Write back the relocated pointer.  */
 				flat_put_addr_at_rp(rp, addr, relval);
@@ -752,6 +783,10 @@ static int load_flat_file(struct linux_binprm * bprm,
 			stack_len);
 
 	return 0;
+err_close:
+	sys_close(exec_fileno);
+err:
+	return ret;
 }
 
 
diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c
index 6a7b730c206b..d73d75591a39 100644
--- a/fs/binfmt_misc.c
+++ b/fs/binfmt_misc.c
@@ -600,7 +600,7 @@ static ssize_t bm_entry_write(struct file *file, const char __user *buffer,
 	return count;
 }
 
-static struct file_operations bm_entry_operations = {
+static const struct file_operations bm_entry_operations = {
 	.read		= bm_entry_read,
 	.write		= bm_entry_write,
 };
@@ -668,7 +668,7 @@ out:
 	return count;
 }
 
-static struct file_operations bm_register_operations = {
+static const struct file_operations bm_register_operations = {
 	.write		= bm_register_write,
 };
 
@@ -715,7 +715,7 @@ static ssize_t bm_status_write(struct file * file, const char __user * buffer,
 	return count;
 }
 
-static struct file_operations bm_status_operations = {
+static const struct file_operations bm_status_operations = {
 	.read		= bm_status_read,
 	.write		= bm_status_write,
 };
diff --git a/fs/bio.c b/fs/bio.c
index 1f3bb501c262..eb8fbc53f2cd 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -25,11 +25,12 @@
 #include <linux/module.h>
 #include <linux/mempool.h>
 #include <linux/workqueue.h>
+#include <linux/blktrace_api.h>
 #include <scsi/sg.h>		/* for struct sg_iovec */
 
 #define BIO_POOL_SIZE 256
 
-static kmem_cache_t *bio_slab;
+static kmem_cache_t *bio_slab __read_mostly;
 
 #define BIOVEC_NR_POOLS 6
 
@@ -38,7 +39,7 @@ static kmem_cache_t *bio_slab;
  * basically we just need to survive
  */
 #define BIO_SPLIT_ENTRIES 8	
-mempool_t *bio_split_pool;
+mempool_t *bio_split_pool __read_mostly;
 
 struct biovec_slab {
 	int nr_vecs;
@@ -635,12 +636,10 @@ static struct bio *__bio_map_user_iov(request_queue_t *q,
 		return ERR_PTR(-ENOMEM);
 
 	ret = -ENOMEM;
-	pages = kmalloc(nr_pages * sizeof(struct page *), GFP_KERNEL);
+	pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
 	if (!pages)
 		goto out;
 
-	memset(pages, 0, nr_pages * sizeof(struct page *));
-
 	for (i = 0; i < iov_count; i++) {
 		unsigned long uaddr = (unsigned long)iov[i].iov_base;
 		unsigned long len = iov[i].iov_len;
@@ -1095,6 +1094,9 @@ struct bio_pair *bio_split(struct bio *bi, mempool_t *pool, int first_sectors)
 	if (!bp)
 		return bp;
 
+	blk_add_trace_pdu_int(bdev_get_queue(bi->bi_bdev), BLK_TA_SPLIT, bi,
+				bi->bi_sector + first_sectors);
+
 	BUG_ON(bi->bi_vcnt != 1);
 	BUG_ON(bi->bi_idx != 0);
 	atomic_set(&bp->cnt, 3);
@@ -1123,16 +1125,6 @@ struct bio_pair *bio_split(struct bio *bi, mempool_t *pool, int first_sectors)
 	return bp;
 }
 
-static void *bio_pair_alloc(gfp_t gfp_flags, void *data)
-{
-	return kmalloc(sizeof(struct bio_pair), gfp_flags);
-}
-
-static void bio_pair_free(void *bp, void *data)
-{
-	kfree(bp);
-}
-
 
 /*
  * create memory pools for biovec's in a bio_set.
@@ -1149,8 +1141,7 @@ static int biovec_create_pools(struct bio_set *bs, int pool_entries, int scale)
 		if (i >= scale)
 			pool_entries >>= 1;
 
-		*bvp = mempool_create(pool_entries, mempool_alloc_slab,
-					mempool_free_slab, bp->slab);
+		*bvp = mempool_create_slab_pool(pool_entries, bp->slab);
 		if (!*bvp)
 			return -ENOMEM;
 	}
@@ -1182,15 +1173,12 @@ void bioset_free(struct bio_set *bs)
 
 struct bio_set *bioset_create(int bio_pool_size, int bvec_pool_size, int scale)
 {
-	struct bio_set *bs = kmalloc(sizeof(*bs), GFP_KERNEL);
+	struct bio_set *bs = kzalloc(sizeof(*bs), GFP_KERNEL);
 
 	if (!bs)
 		return NULL;
 
-	memset(bs, 0, sizeof(*bs));
-	bs->bio_pool = mempool_create(bio_pool_size, mempool_alloc_slab,
-			mempool_free_slab, bio_slab);
-
+	bs->bio_pool = mempool_create_slab_pool(bio_pool_size, bio_slab);
 	if (!bs->bio_pool)
 		goto bad;
 
@@ -1243,18 +1231,18 @@ static int __init init_bio(void)
 		scale = 4;
 
 	/*
-	 * scale number of entries
+	 * Limit number of entries reserved -- mempools are only used when
+	 * the system is completely unable to allocate memory, so we only
+	 * need enough to make progress.
 	 */
-	bvec_pool_entries = megabytes * 2;
-	if (bvec_pool_entries > 256)
-		bvec_pool_entries = 256;
+	bvec_pool_entries = 1 + scale;
 
 	fs_bio_set = bioset_create(BIO_POOL_SIZE, bvec_pool_entries, scale);
 	if (!fs_bio_set)
 		panic("bio: can't allocate bios\n");
 
-	bio_split_pool = mempool_create(BIO_SPLIT_ENTRIES,
-				bio_pair_alloc, bio_pair_free, NULL);
+	bio_split_pool = mempool_create_kmalloc_pool(BIO_SPLIT_ENTRIES,
+						     sizeof(struct bio_pair));
 	if (!bio_split_pool)
 		panic("bio: can't create split pool\n");
 
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 6e50346fb1ee..af88c43043d5 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -86,16 +86,12 @@ EXPORT_SYMBOL(set_blocksize);
 
 int sb_set_blocksize(struct super_block *sb, int size)
 {
-	int bits = 9; /* 2^9 = 512 */
-
 	if (set_blocksize(sb->s_bdev, size))
 		return 0;
 	/* If we get here, we know size is power of two
 	 * and it's value is between 512 and PAGE_SIZE */
 	sb->s_blocksize = size;
-	for (size >>= 10; size; size >>= 1)
-		++bits;
-	sb->s_blocksize_bits = bits;
+	sb->s_blocksize_bits = blksize_bits(size);
 	return sb->s_blocksize;
 }
 
@@ -135,9 +131,10 @@ blkdev_get_block(struct inode *inode, sector_t iblock,
 
 static int
 blkdev_get_blocks(struct inode *inode, sector_t iblock,
-		unsigned long max_blocks, struct buffer_head *bh, int create)
+		struct buffer_head *bh, int create)
 {
 	sector_t end_block = max_block(I_BDEV(inode));
+	unsigned long max_blocks = bh->b_size >> inode->i_blkbits;
 
 	if ((iblock + max_blocks) > end_block) {
 		max_blocks = end_block - iblock;
@@ -238,7 +235,7 @@ static int block_fsync(struct file *filp, struct dentry *dentry, int datasync)
  */
 
 static  __cacheline_aligned_in_smp DEFINE_SPINLOCK(bdev_lock);
-static kmem_cache_t * bdev_cachep;
+static kmem_cache_t * bdev_cachep __read_mostly;
 
 static struct inode *bdev_alloc_inode(struct super_block *sb)
 {
@@ -265,10 +262,13 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 	    SLAB_CTOR_CONSTRUCTOR)
 	{
 		memset(bdev, 0, sizeof(*bdev));
-		sema_init(&bdev->bd_sem, 1);
-		sema_init(&bdev->bd_mount_sem, 1);
+		mutex_init(&bdev->bd_mutex);
+		mutex_init(&bdev->bd_mount_mutex);
 		INIT_LIST_HEAD(&bdev->bd_inodes);
 		INIT_LIST_HEAD(&bdev->bd_list);
+#ifdef CONFIG_SYSFS
+		INIT_LIST_HEAD(&bdev->bd_holder_list);
+#endif
 		inode_init_once(&ei->vfs_inode);
 	}
 }
@@ -312,14 +312,15 @@ static struct file_system_type bd_type = {
 	.kill_sb	= kill_anon_super,
 };
 
-static struct vfsmount *bd_mnt;
+static struct vfsmount *bd_mnt __read_mostly;
 struct super_block *blockdev_superblock;
 
 void __init bdev_cache_init(void)
 {
 	int err;
 	bdev_cachep = kmem_cache_create("bdev_cache", sizeof(struct bdev_inode),
-			0, SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|SLAB_PANIC,
+			0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
+				SLAB_MEM_SPREAD|SLAB_PANIC),
 			init_once, NULL);
 	err = register_filesystem(&bd_type);
 	if (err)
@@ -492,6 +493,300 @@ void bd_release(struct block_device *bdev)
 
 EXPORT_SYMBOL(bd_release);
 
+#ifdef CONFIG_SYSFS
+/*
+ * Functions for bd_claim_by_kobject / bd_release_from_kobject
+ *
+ *     If a kobject is passed to bd_claim_by_kobject()
+ *     and the kobject has a parent directory,
+ *     following symlinks are created:
+ *        o from the kobject to the claimed bdev
+ *        o from "holders" directory of the bdev to the parent of the kobject
+ *     bd_release_from_kobject() removes these symlinks.
+ *
+ *     Example:
+ *        If /dev/dm-0 maps to /dev/sda, kobject corresponding to
+ *        /sys/block/dm-0/slaves is passed to bd_claim_by_kobject(), then:
+ *           /sys/block/dm-0/slaves/sda --> /sys/block/sda
+ *           /sys/block/sda/holders/dm-0 --> /sys/block/dm-0
+ */
+
+static struct kobject *bdev_get_kobj(struct block_device *bdev)
+{
+	if (bdev->bd_contains != bdev)
+		return kobject_get(&bdev->bd_part->kobj);
+	else
+		return kobject_get(&bdev->bd_disk->kobj);
+}
+
+static struct kobject *bdev_get_holder(struct block_device *bdev)
+{
+	if (bdev->bd_contains != bdev)
+		return kobject_get(bdev->bd_part->holder_dir);
+	else
+		return kobject_get(bdev->bd_disk->holder_dir);
+}
+
+static void add_symlink(struct kobject *from, struct kobject *to)
+{
+	if (!from || !to)
+		return;
+	sysfs_create_link(from, to, kobject_name(to));
+}
+
+static void del_symlink(struct kobject *from, struct kobject *to)
+{
+	if (!from || !to)
+		return;
+	sysfs_remove_link(from, kobject_name(to));
+}
+
+/*
+ * 'struct bd_holder' contains pointers to kobjects symlinked by
+ * bd_claim_by_kobject.
+ * It's connected to bd_holder_list which is protected by bdev->bd_sem.
+ */
+struct bd_holder {
+	struct list_head list;	/* chain of holders of the bdev */
+	int count;		/* references from the holder */
+	struct kobject *sdir;	/* holder object, e.g. "/block/dm-0/slaves" */
+	struct kobject *hdev;	/* e.g. "/block/dm-0" */
+	struct kobject *hdir;	/* e.g. "/block/sda/holders" */
+	struct kobject *sdev;	/* e.g. "/block/sda" */
+};
+
+/*
+ * Get references of related kobjects at once.
+ * Returns 1 on success. 0 on failure.
+ *
+ * Should call bd_holder_release_dirs() after successful use.
+ */
+static int bd_holder_grab_dirs(struct block_device *bdev,
+			struct bd_holder *bo)
+{
+	if (!bdev || !bo)
+		return 0;
+
+	bo->sdir = kobject_get(bo->sdir);
+	if (!bo->sdir)
+		return 0;
+
+	bo->hdev = kobject_get(bo->sdir->parent);
+	if (!bo->hdev)
+		goto fail_put_sdir;
+
+	bo->sdev = bdev_get_kobj(bdev);
+	if (!bo->sdev)
+		goto fail_put_hdev;
+
+	bo->hdir = bdev_get_holder(bdev);
+	if (!bo->hdir)
+		goto fail_put_sdev;
+
+	return 1;
+
+fail_put_sdev:
+	kobject_put(bo->sdev);
+fail_put_hdev:
+	kobject_put(bo->hdev);
+fail_put_sdir:
+	kobject_put(bo->sdir);
+
+	return 0;
+}
+
+/* Put references of related kobjects at once. */
+static void bd_holder_release_dirs(struct bd_holder *bo)
+{
+	kobject_put(bo->hdir);
+	kobject_put(bo->sdev);
+	kobject_put(bo->hdev);
+	kobject_put(bo->sdir);
+}
+
+static struct bd_holder *alloc_bd_holder(struct kobject *kobj)
+{
+	struct bd_holder *bo;
+
+	bo = kzalloc(sizeof(*bo), GFP_KERNEL);
+	if (!bo)
+		return NULL;
+
+	bo->count = 1;
+	bo->sdir = kobj;
+
+	return bo;
+}
+
+static void free_bd_holder(struct bd_holder *bo)
+{
+	kfree(bo);
+}
+
+/**
+ * add_bd_holder - create sysfs symlinks for bd_claim() relationship
+ *
+ * @bdev:	block device to be bd_claimed
+ * @bo:		preallocated and initialized by alloc_bd_holder()
+ *
+ * If there is no matching entry with @bo in @bdev->bd_holder_list,
+ * add @bo to the list, create symlinks.
+ *
+ * Returns 1 if @bo was added to the list.
+ * Returns 0 if @bo wasn't used by any reason and should be freed.
+ */
+static int add_bd_holder(struct block_device *bdev, struct bd_holder *bo)
+{
+	struct bd_holder *tmp;
+
+	if (!bo)
+		return 0;
+
+	list_for_each_entry(tmp, &bdev->bd_holder_list, list) {
+		if (tmp->sdir == bo->sdir) {
+			tmp->count++;
+			return 0;
+		}
+	}
+
+	if (!bd_holder_grab_dirs(bdev, bo))
+		return 0;
+
+	add_symlink(bo->sdir, bo->sdev);
+	add_symlink(bo->hdir, bo->hdev);
+	list_add_tail(&bo->list, &bdev->bd_holder_list);
+	return 1;
+}
+
+/**
+ * del_bd_holder - delete sysfs symlinks for bd_claim() relationship
+ *
+ * @bdev:	block device to be bd_claimed
+ * @kobj:	holder's kobject
+ *
+ * If there is matching entry with @kobj in @bdev->bd_holder_list
+ * and no other bd_claim() from the same kobject,
+ * remove the struct bd_holder from the list, delete symlinks for it.
+ *
+ * Returns a pointer to the struct bd_holder when it's removed from the list
+ * and ready to be freed.
+ * Returns NULL if matching claim isn't found or there is other bd_claim()
+ * by the same kobject.
+ */
+static struct bd_holder *del_bd_holder(struct block_device *bdev,
+					struct kobject *kobj)
+{
+	struct bd_holder *bo;
+
+	list_for_each_entry(bo, &bdev->bd_holder_list, list) {
+		if (bo->sdir == kobj) {
+			bo->count--;
+			BUG_ON(bo->count < 0);
+			if (!bo->count) {
+				list_del(&bo->list);
+				del_symlink(bo->sdir, bo->sdev);
+				del_symlink(bo->hdir, bo->hdev);
+				bd_holder_release_dirs(bo);
+				return bo;
+			}
+			break;
+		}
+	}
+
+	return NULL;
+}
+
+/**
+ * bd_claim_by_kobject - bd_claim() with additional kobject signature
+ *
+ * @bdev:	block device to be claimed
+ * @holder:	holder's signature
+ * @kobj:	holder's kobject
+ *
+ * Do bd_claim() and if it succeeds, create sysfs symlinks between
+ * the bdev and the holder's kobject.
+ * Use bd_release_from_kobject() when relesing the claimed bdev.
+ *
+ * Returns 0 on success. (same as bd_claim())
+ * Returns errno on failure.
+ */
+static int bd_claim_by_kobject(struct block_device *bdev, void *holder,
+				struct kobject *kobj)
+{
+	int res;
+	struct bd_holder *bo;
+
+	if (!kobj)
+		return -EINVAL;
+
+	bo = alloc_bd_holder(kobj);
+	if (!bo)
+		return -ENOMEM;
+
+	mutex_lock(&bdev->bd_mutex);
+	res = bd_claim(bdev, holder);
+	if (res || !add_bd_holder(bdev, bo))
+		free_bd_holder(bo);
+	mutex_unlock(&bdev->bd_mutex);
+
+	return res;
+}
+
+/**
+ * bd_release_from_kobject - bd_release() with additional kobject signature
+ *
+ * @bdev:	block device to be released
+ * @kobj:	holder's kobject
+ *
+ * Do bd_release() and remove sysfs symlinks created by bd_claim_by_kobject().
+ */
+static void bd_release_from_kobject(struct block_device *bdev,
+					struct kobject *kobj)
+{
+	struct bd_holder *bo;
+
+	if (!kobj)
+		return;
+
+	mutex_lock(&bdev->bd_mutex);
+	bd_release(bdev);
+	if ((bo = del_bd_holder(bdev, kobj)))
+		free_bd_holder(bo);
+	mutex_unlock(&bdev->bd_mutex);
+}
+
+/**
+ * bd_claim_by_disk - wrapper function for bd_claim_by_kobject()
+ *
+ * @bdev:	block device to be claimed
+ * @holder:	holder's signature
+ * @disk:	holder's gendisk
+ *
+ * Call bd_claim_by_kobject() with getting @disk->slave_dir.
+ */
+int bd_claim_by_disk(struct block_device *bdev, void *holder,
+			struct gendisk *disk)
+{
+	return bd_claim_by_kobject(bdev, holder, kobject_get(disk->slave_dir));
+}
+EXPORT_SYMBOL_GPL(bd_claim_by_disk);
+
+/**
+ * bd_release_from_disk - wrapper function for bd_release_from_kobject()
+ *
+ * @bdev:	block device to be claimed
+ * @disk:	holder's gendisk
+ *
+ * Call bd_release_from_kobject() and put @disk->slave_dir.
+ */
+void bd_release_from_disk(struct block_device *bdev, struct gendisk *disk)
+{
+	bd_release_from_kobject(bdev, disk->slave_dir);
+	kobject_put(disk->slave_dir);
+}
+EXPORT_SYMBOL_GPL(bd_release_from_disk);
+#endif
+
 /*
  * Tries to open block device by device number.  Use it ONLY if you
  * really do not have anything better - i.e. when you are behind a
@@ -574,7 +869,7 @@ static int do_open(struct block_device *bdev, struct file *file)
 	}
 	owner = disk->fops->owner;
 
-	down(&bdev->bd_sem);
+	mutex_lock(&bdev->bd_mutex);
 	if (!bdev->bd_openers) {
 		bdev->bd_disk = disk;
 		bdev->bd_contains = bdev;
@@ -605,21 +900,21 @@ static int do_open(struct block_device *bdev, struct file *file)
 			if (ret)
 				goto out_first;
 			bdev->bd_contains = whole;
-			down(&whole->bd_sem);
+			mutex_lock(&whole->bd_mutex);
 			whole->bd_part_count++;
 			p = disk->part[part - 1];
 			bdev->bd_inode->i_data.backing_dev_info =
 			   whole->bd_inode->i_data.backing_dev_info;
 			if (!(disk->flags & GENHD_FL_UP) || !p || !p->nr_sects) {
 				whole->bd_part_count--;
-				up(&whole->bd_sem);
+				mutex_unlock(&whole->bd_mutex);
 				ret = -ENXIO;
 				goto out_first;
 			}
 			kobject_get(&p->kobj);
 			bdev->bd_part = p;
 			bd_set_size(bdev, (loff_t) p->nr_sects << 9);
-			up(&whole->bd_sem);
+			mutex_unlock(&whole->bd_mutex);
 		}
 	} else {
 		put_disk(disk);
@@ -633,13 +928,13 @@ static int do_open(struct block_device *bdev, struct file *file)
 			if (bdev->bd_invalidated)
 				rescan_partitions(bdev->bd_disk, bdev);
 		} else {
-			down(&bdev->bd_contains->bd_sem);
+			mutex_lock(&bdev->bd_contains->bd_mutex);
 			bdev->bd_contains->bd_part_count++;
-			up(&bdev->bd_contains->bd_sem);
+			mutex_unlock(&bdev->bd_contains->bd_mutex);
 		}
 	}
 	bdev->bd_openers++;
-	up(&bdev->bd_sem);
+	mutex_unlock(&bdev->bd_mutex);
 	unlock_kernel();
 	return 0;
 
@@ -652,7 +947,7 @@ out_first:
 	put_disk(disk);
 	module_put(owner);
 out:
-	up(&bdev->bd_sem);
+	mutex_unlock(&bdev->bd_mutex);
 	unlock_kernel();
 	if (ret)
 		bdput(bdev);
@@ -714,7 +1009,7 @@ int blkdev_put(struct block_device *bdev)
 	struct inode *bd_inode = bdev->bd_inode;
 	struct gendisk *disk = bdev->bd_disk;
 
-	down(&bdev->bd_sem);
+	mutex_lock(&bdev->bd_mutex);
 	lock_kernel();
 	if (!--bdev->bd_openers) {
 		sync_blockdev(bdev);
@@ -724,9 +1019,9 @@ int blkdev_put(struct block_device *bdev)
 		if (disk->fops->release)
 			ret = disk->fops->release(bd_inode, NULL);
 	} else {
-		down(&bdev->bd_contains->bd_sem);
+		mutex_lock(&bdev->bd_contains->bd_mutex);
 		bdev->bd_contains->bd_part_count--;
-		up(&bdev->bd_contains->bd_sem);
+		mutex_unlock(&bdev->bd_contains->bd_mutex);
 	}
 	if (!bdev->bd_openers) {
 		struct module *owner = disk->fops->owner;
@@ -746,7 +1041,7 @@ int blkdev_put(struct block_device *bdev)
 		bdev->bd_contains = NULL;
 	}
 	unlock_kernel();
-	up(&bdev->bd_sem);
+	mutex_unlock(&bdev->bd_mutex);
 	bdput(bdev);
 	return ret;
 }
@@ -792,7 +1087,7 @@ struct address_space_operations def_blk_aops = {
 	.direct_IO	= blkdev_direct_IO,
 };
 
-struct file_operations def_blk_fops = {
+const struct file_operations def_blk_fops = {
 	.open		= blkdev_open,
 	.release	= blkdev_close,
 	.llseek		= block_llseek,
diff --git a/fs/buffer.c b/fs/buffer.c
index a9b399402007..23f1f3a68077 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -160,12 +160,7 @@ int sync_blockdev(struct block_device *bdev)
 }
 EXPORT_SYMBOL(sync_blockdev);
 
-/*
- * Write out and wait upon all dirty data associated with this
- * superblock.  Filesystem data as well as the underlying block
- * device.  Takes the superblock lock.
- */
-int fsync_super(struct super_block *sb)
+static void __fsync_super(struct super_block *sb)
 {
 	sync_inodes_sb(sb, 0);
 	DQUOT_SYNC(sb);
@@ -177,7 +172,16 @@ int fsync_super(struct super_block *sb)
 		sb->s_op->sync_fs(sb, 1);
 	sync_blockdev(sb->s_bdev);
 	sync_inodes_sb(sb, 1);
+}
 
+/*
+ * Write out and wait upon all dirty data associated with this
+ * superblock.  Filesystem data as well as the underlying block
+ * device.  Takes the superblock lock.
+ */
+int fsync_super(struct super_block *sb)
+{
+	__fsync_super(sb);
 	return sync_blockdev(sb->s_bdev);
 }
 
@@ -201,7 +205,7 @@ int fsync_bdev(struct block_device *bdev)
  * freeze_bdev  --  lock a filesystem and force it into a consistent state
  * @bdev:	blockdevice to lock
  *
- * This takes the block device bd_mount_sem to make sure no new mounts
+ * This takes the block device bd_mount_mutex to make sure no new mounts
  * happen on bdev until thaw_bdev() is called.
  * If a superblock is found on this device, we take the s_umount semaphore
  * on it to make sure nobody unmounts until the snapshot creation is done.
@@ -210,25 +214,13 @@ struct super_block *freeze_bdev(struct block_device *bdev)
 {
 	struct super_block *sb;
 
-	down(&bdev->bd_mount_sem);
+	mutex_lock(&bdev->bd_mount_mutex);
 	sb = get_super(bdev);
 	if (sb && !(sb->s_flags & MS_RDONLY)) {
 		sb->s_frozen = SB_FREEZE_WRITE;
 		smp_wmb();
 
-		sync_inodes_sb(sb, 0);
-		DQUOT_SYNC(sb);
-
-		lock_super(sb);
-		if (sb->s_dirt && sb->s_op->write_super)
-			sb->s_op->write_super(sb);
-		unlock_super(sb);
-
-		if (sb->s_op->sync_fs)
-			sb->s_op->sync_fs(sb, 1);
-
-		sync_blockdev(sb->s_bdev);
-		sync_inodes_sb(sb, 1);
+		__fsync_super(sb);
 
 		sb->s_frozen = SB_FREEZE_TRANS;
 		smp_wmb();
@@ -264,7 +256,7 @@ void thaw_bdev(struct block_device *bdev, struct super_block *sb)
 		drop_super(sb);
 	}
 
-	up(&bdev->bd_mount_sem);
+	mutex_unlock(&bdev->bd_mount_mutex);
 }
 EXPORT_SYMBOL(thaw_bdev);
 
@@ -327,31 +319,24 @@ int file_fsync(struct file *filp, struct dentry *dentry, int datasync)
 	return ret;
 }
 
-static long do_fsync(unsigned int fd, int datasync)
+long do_fsync(struct file *file, int datasync)
 {
-	struct file * file;
-	struct address_space *mapping;
-	int ret, err;
-
-	ret = -EBADF;
-	file = fget(fd);
-	if (!file)
-		goto out;
+	int ret;
+	int err;
+	struct address_space *mapping = file->f_mapping;
 
-	ret = -EINVAL;
 	if (!file->f_op || !file->f_op->fsync) {
 		/* Why?  We can still call filemap_fdatawrite */
-		goto out_putf;
+		ret = -EINVAL;
+		goto out;
 	}
 
-	mapping = file->f_mapping;
-
 	current->flags |= PF_SYNCWRITE;
 	ret = filemap_fdatawrite(mapping);
 
 	/*
-	 * We need to protect against concurrent writers,
-	 * which could cause livelocks in fsync_buffers_list
+	 * We need to protect against concurrent writers, which could cause
+	 * livelocks in fsync_buffers_list().
 	 */
 	mutex_lock(&mapping->host->i_mutex);
 	err = file->f_op->fsync(file, file->f_dentry, datasync);
@@ -362,21 +347,31 @@ static long do_fsync(unsigned int fd, int datasync)
 	if (!ret)
 		ret = err;
 	current->flags &= ~PF_SYNCWRITE;
-
-out_putf:
-	fput(file);
 out:
 	return ret;
 }
 
+static long __do_fsync(unsigned int fd, int datasync)
+{
+	struct file *file;
+	int ret = -EBADF;
+
+	file = fget(fd);
+	if (file) {
+		ret = do_fsync(file, datasync);
+		fput(file);
+	}
+	return ret;
+}
+
 asmlinkage long sys_fsync(unsigned int fd)
 {
-	return do_fsync(fd, 0);
+	return __do_fsync(fd, 0);
 }
 
 asmlinkage long sys_fdatasync(unsigned int fd)
 {
-	return do_fsync(fd, 1);
+	return __do_fsync(fd, 1);
 }
 
 /*
@@ -431,8 +426,10 @@ __find_get_block_slow(struct block_device *bdev, sector_t block)
 	if (all_mapped) {
 		printk("__find_get_block_slow() failed. "
 			"block=%llu, b_blocknr=%llu\n",
-			(unsigned long long)block, (unsigned long long)bh->b_blocknr);
-		printk("b_state=0x%08lx, b_size=%u\n", bh->b_state, bh->b_size);
+			(unsigned long long)block,
+			(unsigned long long)bh->b_blocknr);
+		printk("b_state=0x%08lx, b_size=%zu\n",
+			bh->b_state, bh->b_size);
 		printk("device blocksize: %d\n", 1 << bd_inode->i_blkbits);
 	}
 out_unlock:
@@ -496,7 +493,7 @@ static void free_more_memory(void)
 	wakeup_pdflush(1024);
 	yield();
 
-	for_each_pgdat(pgdat) {
+	for_each_online_pgdat(pgdat) {
 		zones = pgdat->node_zonelists[gfp_zone(GFP_NOFS)].zones;
 		if (*zones)
 			try_to_free_pages(zones, GFP_NOFS);
@@ -801,8 +798,7 @@ void mark_buffer_dirty_inode(struct buffer_head *bh, struct inode *inode)
 	if (!mapping->assoc_mapping) {
 		mapping->assoc_mapping = buffer_mapping;
 	} else {
-		if (mapping->assoc_mapping != buffer_mapping)
-			BUG();
+		BUG_ON(mapping->assoc_mapping != buffer_mapping);
 	}
 	if (list_empty(&bh->b_assoc_buffers)) {
 		spin_lock(&buffer_mapping->private_lock);
@@ -865,8 +861,8 @@ int __set_page_dirty_buffers(struct page *page)
 		}
 		write_unlock_irq(&mapping->tree_lock);
 		__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
+		return 1;
 	}
-	
 	return 0;
 }
 EXPORT_SYMBOL(__set_page_dirty_buffers);
@@ -1119,8 +1115,7 @@ grow_dev_page(struct block_device *bdev, sector_t block,
 	if (!page)
 		return NULL;
 
-	if (!PageLocked(page))
-		BUG();
+	BUG_ON(!PageLocked(page));
 
 	if (page_has_buffers(page)) {
 		bh = page_buffers(page);
@@ -1527,8 +1522,7 @@ void set_bh_page(struct buffer_head *bh,
 		struct page *page, unsigned long offset)
 {
 	bh->b_page = page;
-	if (offset >= PAGE_SIZE)
-		BUG();
+	BUG_ON(offset >= PAGE_SIZE);
 	if (PageHighMem(page))
 		/*
 		 * This catches illegal uses and preserves the offset:
@@ -1598,11 +1592,10 @@ EXPORT_SYMBOL(try_to_release_page);
  * point.  Because the caller is about to free (and possibly reuse) those
  * blocks on-disk.
  */
-int block_invalidatepage(struct page *page, unsigned long offset)
+void block_invalidatepage(struct page *page, unsigned long offset)
 {
 	struct buffer_head *head, *bh, *next;
 	unsigned int curr_off = 0;
-	int ret = 1;
 
 	BUG_ON(!PageLocked(page));
 	if (!page_has_buffers(page))
@@ -1629,19 +1622,18 @@ int block_invalidatepage(struct page *page, unsigned long offset)
 	 * so real IO is not possible anymore.
 	 */
 	if (offset == 0)
-		ret = try_to_release_page(page, 0);
+		try_to_release_page(page, 0);
 out:
-	return ret;
+	return;
 }
 EXPORT_SYMBOL(block_invalidatepage);
 
-int do_invalidatepage(struct page *page, unsigned long offset)
+void do_invalidatepage(struct page *page, unsigned long offset)
 {
-	int (*invalidatepage)(struct page *, unsigned long);
-	invalidatepage = page->mapping->a_ops->invalidatepage;
-	if (invalidatepage == NULL)
-		invalidatepage = block_invalidatepage;
-	return (*invalidatepage)(page, offset);
+	void (*invalidatepage)(struct page *, unsigned long);
+	invalidatepage = page->mapping->a_ops->invalidatepage ? :
+		block_invalidatepage;
+	(*invalidatepage)(page, offset);
 }
 
 /*
@@ -1743,6 +1735,7 @@ static int __block_write_full_page(struct inode *inode, struct page *page,
 	sector_t block;
 	sector_t last_block;
 	struct buffer_head *bh, *head;
+	const unsigned blocksize = 1 << inode->i_blkbits;
 	int nr_underway = 0;
 
 	BUG_ON(!PageLocked(page));
@@ -1750,7 +1743,7 @@ static int __block_write_full_page(struct inode *inode, struct page *page,
 	last_block = (i_size_read(inode) - 1) >> inode->i_blkbits;
 
 	if (!page_has_buffers(page)) {
-		create_empty_buffers(page, 1 << inode->i_blkbits,
+		create_empty_buffers(page, blocksize,
 					(1 << BH_Dirty)|(1 << BH_Uptodate));
 	}
 
@@ -1785,6 +1778,7 @@ static int __block_write_full_page(struct inode *inode, struct page *page,
 			clear_buffer_dirty(bh);
 			set_buffer_uptodate(bh);
 		} else if (!buffer_mapped(bh) && buffer_dirty(bh)) {
+			WARN_ON(bh->b_size != blocksize);
 			err = get_block(inode, block, bh, 1);
 			if (err)
 				goto recover;
@@ -1938,6 +1932,7 @@ static int __block_prepare_write(struct inode *inode, struct page *page,
 		if (buffer_new(bh))
 			clear_buffer_new(bh);
 		if (!buffer_mapped(bh)) {
+			WARN_ON(bh->b_size != blocksize);
 			err = get_block(inode, block, bh, 1);
 			if (err)
 				break;
@@ -2093,6 +2088,7 @@ int block_read_full_page(struct page *page, get_block_t *get_block)
 
 			fully_mapped = 0;
 			if (iblock < lblock) {
+				WARN_ON(bh->b_size != blocksize);
 				err = get_block(inode, iblock, bh, 0);
 				if (err)
 					SetPageError(page);
@@ -2414,6 +2410,7 @@ int nobh_prepare_write(struct page *page, unsigned from, unsigned to,
 		create = 1;
 		if (block_start >= to)
 			create = 0;
+		map_bh.b_size = blocksize;
 		ret = get_block(inode, block_in_file + block_in_page,
 					&map_bh, create);
 		if (ret)
@@ -2674,6 +2671,7 @@ int block_truncate_page(struct address_space *mapping,
 
 	err = 0;
 	if (!buffer_mapped(bh)) {
+		WARN_ON(bh->b_size != blocksize);
 		err = get_block(inode, iblock, bh, 0);
 		if (err)
 			goto unlock;
@@ -2760,6 +2758,7 @@ sector_t generic_block_bmap(struct address_space *mapping, sector_t block,
 	struct inode *inode = mapping->host;
 	tmp.b_state = 0;
 	tmp.b_blocknr = 0;
+	tmp.b_size = 1 << inode->i_blkbits;
 	get_block(inode, block, &tmp, 0);
 	return tmp.b_blocknr;
 }
@@ -3012,7 +3011,7 @@ out:
 }
 EXPORT_SYMBOL(try_to_free_buffers);
 
-int block_sync_page(struct page *page)
+void block_sync_page(struct page *page)
 {
 	struct address_space *mapping;
 
@@ -3020,7 +3019,6 @@ int block_sync_page(struct page *page)
 	mapping = page_mapping(page);
 	if (mapping)
 		blk_run_backing_dev(mapping->backing_dev_info, page);
-	return 0;
 }
 
 /*
@@ -3051,68 +3049,6 @@ asmlinkage long sys_bdflush(int func, long data)
 }
 
 /*
- * Migration function for pages with buffers. This function can only be used
- * if the underlying filesystem guarantees that no other references to "page"
- * exist.
- */
-#ifdef CONFIG_MIGRATION
-int buffer_migrate_page(struct page *newpage, struct page *page)
-{
-	struct address_space *mapping = page->mapping;
-	struct buffer_head *bh, *head;
-	int rc;
-
-	if (!mapping)
-		return -EAGAIN;
-
-	if (!page_has_buffers(page))
-		return migrate_page(newpage, page);
-
-	head = page_buffers(page);
-
-	rc = migrate_page_remove_references(newpage, page, 3);
-	if (rc)
-		return rc;
-
-	bh = head;
-	do {
-		get_bh(bh);
-		lock_buffer(bh);
-		bh = bh->b_this_page;
-
-	} while (bh != head);
-
-	ClearPagePrivate(page);
-	set_page_private(newpage, page_private(page));
-	set_page_private(page, 0);
-	put_page(page);
-	get_page(newpage);
-
-	bh = head;
-	do {
-		set_bh_page(bh, newpage, bh_offset(bh));
-		bh = bh->b_this_page;
-
-	} while (bh != head);
-
-	SetPagePrivate(newpage);
-
-	migrate_page_copy(newpage, page);
-
-	bh = head;
-	do {
-		unlock_buffer(bh);
- 		put_bh(bh);
-		bh = bh->b_this_page;
-
-	} while (bh != head);
-
-	return 0;
-}
-EXPORT_SYMBOL(buffer_migrate_page);
-#endif
-
-/*
  * Buffer-head allocation
  */
 static kmem_cache_t *bh_cachep;
@@ -3140,7 +3076,7 @@ static void recalc_bh_state(void)
 	if (__get_cpu_var(bh_accounting).ratelimit++ < 4096)
 		return;
 	__get_cpu_var(bh_accounting).ratelimit = 0;
-	for_each_cpu(i)
+	for_each_online_cpu(i)
 		tot += per_cpu(bh_accounting, i).nr;
 	buffer_heads_over_limit = (tot > max_buffer_heads);
 }
@@ -3189,6 +3125,9 @@ static void buffer_exit_cpu(int cpu)
 		brelse(b->bhs[i]);
 		b->bhs[i] = NULL;
 	}
+	get_cpu_var(bh_accounting).nr += per_cpu(bh_accounting, cpu).nr;
+	per_cpu(bh_accounting, cpu).nr = 0;
+	put_cpu_var(bh_accounting);
 }
 
 static int buffer_cpu_notify(struct notifier_block *self,
@@ -3205,8 +3144,11 @@ void __init buffer_init(void)
 	int nrpages;
 
 	bh_cachep = kmem_cache_create("buffer_head",
-			sizeof(struct buffer_head), 0,
-			SLAB_RECLAIM_ACCOUNT|SLAB_PANIC, init_buffer_head, NULL);
+					sizeof(struct buffer_head), 0,
+					(SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
+					SLAB_MEM_SPREAD),
+					init_buffer_head,
+					NULL);
 
 	/*
 	 * Limit the bh occupancy to 10% of ZONE_NORMAL
diff --git a/fs/char_dev.c b/fs/char_dev.c
index 21195c481637..4e1b849f912f 100644
--- a/fs/char_dev.c
+++ b/fs/char_dev.c
@@ -19,6 +19,7 @@
 #include <linux/kobject.h>
 #include <linux/kobj_map.h>
 #include <linux/cdev.h>
+#include <linux/mutex.h>
 
 #ifdef CONFIG_KMOD
 #include <linux/kmod.h>
@@ -28,7 +29,7 @@ static struct kobj_map *cdev_map;
 
 #define MAX_PROBE_HASH 255	/* random */
 
-static DECLARE_MUTEX(chrdevs_lock);
+static DEFINE_MUTEX(chrdevs_lock);
 
 static struct char_device_struct {
 	struct char_device_struct *next;
@@ -88,13 +89,13 @@ out:
 
 void *acquire_chrdev_list(void)
 {
-	down(&chrdevs_lock);
+	mutex_lock(&chrdevs_lock);
 	return get_next_chrdev(NULL);
 }
 
 void release_chrdev_list(void *dev)
 {
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	kfree(dev);
 }
 
@@ -145,13 +146,11 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor,
 	int ret = 0;
 	int i;
 
-	cd = kmalloc(sizeof(struct char_device_struct), GFP_KERNEL);
+	cd = kzalloc(sizeof(struct char_device_struct), GFP_KERNEL);
 	if (cd == NULL)
 		return ERR_PTR(-ENOMEM);
 
-	memset(cd, 0, sizeof(struct char_device_struct));
-
-	down(&chrdevs_lock);
+	mutex_lock(&chrdevs_lock);
 
 	/* temporary */
 	if (major == 0) {
@@ -186,10 +185,10 @@ __register_chrdev_region(unsigned int major, unsigned int baseminor,
 	}
 	cd->next = *cp;
 	*cp = cd;
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	return cd;
 out:
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	kfree(cd);
 	return ERR_PTR(ret);
 }
@@ -200,7 +199,7 @@ __unregister_chrdev_region(unsigned major, unsigned baseminor, int minorct)
 	struct char_device_struct *cd = NULL, **cp;
 	int i = major_to_index(major);
 
-	down(&chrdevs_lock);
+	mutex_lock(&chrdevs_lock);
 	for (cp = &chrdevs[i]; *cp; cp = &(*cp)->next)
 		if ((*cp)->major == major &&
 		    (*cp)->baseminor == baseminor &&
@@ -210,7 +209,7 @@ __unregister_chrdev_region(unsigned major, unsigned baseminor, int minorct)
 		cd = *cp;
 		*cp = cd->next;
 	}
-	up(&chrdevs_lock);
+	mutex_unlock(&chrdevs_lock);
 	return cd;
 }
 
@@ -251,7 +250,7 @@ int alloc_chrdev_region(dev_t *dev, unsigned baseminor, unsigned count,
 }
 
 int register_chrdev(unsigned int major, const char *name,
-		    struct file_operations *fops)
+		    const struct file_operations *fops)
 {
 	struct char_device_struct *cd;
 	struct cdev *cdev;
@@ -407,7 +406,7 @@ static void cdev_purge(struct cdev *cdev)
  * is contain the open that then fills in the correct operations
  * depending on the special file...
  */
-struct file_operations def_chr_fops = {
+const struct file_operations def_chr_fops = {
 	.open = chrdev_open,
 };
 
@@ -465,9 +464,8 @@ static struct kobj_type ktype_cdev_dynamic = {
 
 struct cdev *cdev_alloc(void)
 {
-	struct cdev *p = kmalloc(sizeof(struct cdev), GFP_KERNEL);
+	struct cdev *p = kzalloc(sizeof(struct cdev), GFP_KERNEL);
 	if (p) {
-		memset(p, 0, sizeof(struct cdev));
 		p->kobj.ktype = &ktype_cdev_dynamic;
 		INIT_LIST_HEAD(&p->list);
 		kobject_init(&p->kobj);
@@ -475,7 +473,7 @@ struct cdev *cdev_alloc(void)
 	return p;
 }
 
-void cdev_init(struct cdev *cdev, struct file_operations *fops)
+void cdev_init(struct cdev *cdev, const struct file_operations *fops)
 {
 	memset(cdev, 0, sizeof *cdev);
 	INIT_LIST_HEAD(&cdev->list);
diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES
index 3e0fcc394810..8a2de038882e 100644
--- a/fs/cifs/CHANGES
+++ b/fs/cifs/CHANGES
@@ -178,7 +178,7 @@ improperly zeroed buffer in CIFS Unix extensions set times call.
 Version 1.25
 ------------
 Fix internationalization problem in cifs readdir with filenames that map to 
-longer UTF8 strings than the string on the wire was in Unicode.  Add workaround
+longer UTF-8 strings than the string on the wire was in Unicode.  Add workaround
 for readdir to netapp servers. Fix search rewind (seek into readdir to return 
 non-consecutive entries).  Do not do readdir when server negotiates 
 buffer size to small to fit filename. Add support for reading POSIX ACLs from
diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c
index 8219c02145a7..d4b713e5affb 100644
--- a/fs/cifs/cifsfs.c
+++ b/fs/cifs/cifsfs.c
@@ -476,7 +476,7 @@ cifs_get_sb(struct file_system_type *fs_type,
 
 	sb->s_flags = flags;
 
-	rc = cifs_read_super(sb, data, dev_name, flags & MS_VERBOSE ? 1 : 0);
+	rc = cifs_read_super(sb, data, dev_name, flags & MS_SILENT ? 1 : 0);
 	if (rc) {
 		up_write(&sb->s_umount);
 		deactivate_super(sb);
@@ -580,7 +580,7 @@ struct inode_operations cifs_symlink_inode_ops = {
 #endif 
 };
 
-struct file_operations cifs_file_ops = {
+const struct file_operations cifs_file_ops = {
 	.read = do_sync_read,
 	.write = do_sync_write,
 	.readv = generic_file_readv,
@@ -604,7 +604,7 @@ struct file_operations cifs_file_ops = {
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
 
-struct file_operations cifs_file_direct_ops = {
+const struct file_operations cifs_file_direct_ops = {
 	/* no mmap, no aio, no readv - 
 	   BB reevaluate whether they can be done with directio, no cache */
 	.read = cifs_user_read,
@@ -623,7 +623,7 @@ struct file_operations cifs_file_direct_ops = {
 	.dir_notify = cifs_dir_notify,
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
-struct file_operations cifs_file_nobrl_ops = {
+const struct file_operations cifs_file_nobrl_ops = {
 	.read = do_sync_read,
 	.write = do_sync_write,
 	.readv = generic_file_readv,
@@ -646,7 +646,7 @@ struct file_operations cifs_file_nobrl_ops = {
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
 
-struct file_operations cifs_file_direct_nobrl_ops = {
+const struct file_operations cifs_file_direct_nobrl_ops = {
 	/* no mmap, no aio, no readv - 
 	   BB reevaluate whether they can be done with directio, no cache */
 	.read = cifs_user_read,
@@ -665,7 +665,7 @@ struct file_operations cifs_file_direct_nobrl_ops = {
 #endif /* CONFIG_CIFS_EXPERIMENTAL */
 };
 
-struct file_operations cifs_dir_ops = {
+const struct file_operations cifs_dir_ops = {
 	.readdir = cifs_readdir,
 	.release = cifs_closedir,
 	.read    = generic_read_dir,
@@ -692,7 +692,8 @@ cifs_init_inodecache(void)
 {
 	cifs_inode_cachep = kmem_cache_create("cifs_inode_cache",
 					      sizeof (struct cifsInodeInfo),
-					      0, SLAB_RECLAIM_ACCOUNT,
+					      0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					      cifs_init_once, NULL);
 	if (cifs_inode_cachep == NULL)
 		return -ENOMEM;
@@ -734,10 +735,8 @@ cifs_init_request_bufs(void)
 		cERROR(1,("cifs_min_rcv set to maximum (64)"));
 	}
 
-	cifs_req_poolp = mempool_create(cifs_min_rcv,
-					mempool_alloc_slab,
-					mempool_free_slab,
-					cifs_req_cachep);
+	cifs_req_poolp = mempool_create_slab_pool(cifs_min_rcv,
+						  cifs_req_cachep);
 
 	if(cifs_req_poolp == NULL) {
 		kmem_cache_destroy(cifs_req_cachep);
@@ -767,10 +766,8 @@ cifs_init_request_bufs(void)
 		cFYI(1,("cifs_min_small set to maximum (256)"));
 	}
 
-	cifs_sm_req_poolp = mempool_create(cifs_min_small,
-				mempool_alloc_slab,
-				mempool_free_slab,
-				cifs_sm_req_cachep);
+	cifs_sm_req_poolp = mempool_create_slab_pool(cifs_min_small,
+						     cifs_sm_req_cachep);
 
 	if(cifs_sm_req_poolp == NULL) {
 		mempool_destroy(cifs_req_poolp);
@@ -804,10 +801,8 @@ cifs_init_mids(void)
 	if (cifs_mid_cachep == NULL)
 		return -ENOMEM;
 
-	cifs_mid_poolp = mempool_create(3 /* a reasonable min simultan opers */,
-					mempool_alloc_slab,
-					mempool_free_slab,
-					cifs_mid_cachep);
+	/* 3 is a reasonable minimum number of simultaneous operations */
+	cifs_mid_poolp = mempool_create_slab_pool(3, cifs_mid_cachep);
 	if(cifs_mid_poolp == NULL) {
 		kmem_cache_destroy(cifs_mid_cachep);
 		return -ENOMEM;
diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index b4dcdc2052a0..4e829dc672a6 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -61,10 +61,10 @@ extern struct inode_operations cifs_file_inode_ops;
 extern struct inode_operations cifs_symlink_inode_ops;
 
 /* Functions related to files and directories */
-extern struct file_operations cifs_file_ops;
-extern struct file_operations cifs_file_direct_ops; /* if directio mount */
-extern struct file_operations cifs_file_nobrl_ops;
-extern struct file_operations cifs_file_direct_nobrl_ops; /* if directio mount */
+extern const struct file_operations cifs_file_ops;
+extern const struct file_operations cifs_file_direct_ops; /* if directio mount */
+extern const struct file_operations cifs_file_nobrl_ops;
+extern const struct file_operations cifs_file_direct_nobrl_ops; /* if directio mount */
 extern int cifs_open(struct inode *inode, struct file *file);
 extern int cifs_close(struct inode *inode, struct file *file);
 extern int cifs_closedir(struct inode *inode, struct file *file);
@@ -76,7 +76,7 @@ extern int cifs_lock(struct file *, int, struct file_lock *);
 extern int cifs_fsync(struct file *, struct dentry *, int);
 extern int cifs_flush(struct file *);
 extern int cifs_file_mmap(struct file * , struct vm_area_struct *);
-extern struct file_operations cifs_dir_ops;
+extern const struct file_operations cifs_dir_ops;
 extern int cifs_dir_open(struct inode *inode, struct file *file);
 extern int cifs_readdir(struct file *file, void *direntry, filldir_t filldir);
 extern int cifs_dir_notify(struct file *, unsigned long arg);
diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c
index a5941872d3f7..d705500aa283 100644
--- a/fs/cifs/cifssmb.c
+++ b/fs/cifs/cifssmb.c
@@ -5023,7 +5023,7 @@ SetEARetry:
 	parm_data->list_len = cpu_to_le32(count);
 	parm_data->list[0].EA_flags = 0;
 	/* we checked above that name len is less than 255 */
-	parm_data->list[0].name_len = (__u8)name_len;;
+	parm_data->list[0].name_len = (__u8)name_len;
 	/* EA names are always ASCII */
 	if(ea_name)
 		strncpy(parm_data->list[0].name,ea_name,name_len);
diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c
index ddd11fa15528..1d0ca3eaaca5 100644
--- a/fs/cifs/dir.c
+++ b/fs/cifs/dir.c
@@ -139,9 +139,9 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode,
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -316,9 +316,9 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode,
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 	if(full_path == NULL)
 		rc = -ENOMEM;
 	else if (pTcon->ses->capabilities & CAP_UNIX) {
diff --git a/fs/cifs/fcntl.c b/fs/cifs/fcntl.c
index a7a47bb36bf3..ec4dfe9bf5ef 100644
--- a/fs/cifs/fcntl.c
+++ b/fs/cifs/fcntl.c
@@ -86,9 +86,9 @@ int cifs_dir_notify(struct file * file, unsigned long arg)
 	cifs_sb = CIFS_SB(file->f_dentry->d_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(file->f_dentry);
-	up(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 
 	if(full_path == NULL) {
 		rc = -ENOMEM;
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index 2c093de26225..5c497c529772 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -203,9 +203,9 @@ int cifs_open(struct inode *inode, struct file *file)
 		}
 	}
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(file->f_dentry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -1377,7 +1377,7 @@ int cifs_fsync(struct file *file, struct dentry *dentry, int datasync)
 	return rc;
 }
 
-/* static int cifs_sync_page(struct page *page)
+/* static void cifs_sync_page(struct page *page)
 {
 	struct address_space *mapping;
 	struct inode *inode;
@@ -1391,16 +1391,18 @@ int cifs_fsync(struct file *file, struct dentry *dentry, int datasync)
 		return 0;
 	inode = mapping->host;
 	if (!inode)
-		return 0; */
+		return; */
 
 /*	fill in rpages then 
 	result = cifs_pagein_inode(inode, index, rpages); */ /* BB finish */
 
 /*	cFYI(1, ("rpages is %d for sync page of Index %ld ", rpages, index));
 
+#if 0
 	if (rc < 0)
 		return rc;
 	return 0;
+#endif
 } */
 
 /*
diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c
index b21038b99fc2..e842ce9f6547 100644
--- a/fs/cifs/inode.c
+++ b/fs/cifs/inode.c
@@ -163,9 +163,9 @@ int cifs_get_inode_info_unix(struct inode **pinode,
 
 		if (num_of_bytes < end_of_file)
 			cFYI(1, ("allocation size less than end of file"));
-		cFYI(1,
-		     ("Size %ld and blocks %ld",
-		      (unsigned long) inode->i_size, inode->i_blocks));
+		cFYI(1, ("Size %ld and blocks %llu",
+			(unsigned long) inode->i_size,
+			(unsigned long long)inode->i_blocks));
 		if (S_ISREG(inode->i_mode)) {
 			cFYI(1, ("File inode"));
 			inode->i_op = &cifs_file_inode_ops;
@@ -574,9 +574,9 @@ int cifs_unlink(struct inode *inode, struct dentry *direntry)
 
 	/* Unlink can be called from rename so we can not grab the sem here
 	   since we deadlock otherwise */
-/*	down(&direntry->d_sb->s_vfs_rename_sem);*/
+/*	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);*/
 	full_path = build_path_from_dentry(direntry);
-/*	up(&direntry->d_sb->s_vfs_rename_sem);*/
+/*	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);*/
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -717,9 +717,9 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode)
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -802,9 +802,9 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry)
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -1136,9 +1136,9 @@ int cifs_setattr(struct dentry *direntry, struct iattr *attrs)
 			rc = 0;
 	}
 		
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 	if (full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
diff --git a/fs/cifs/link.c b/fs/cifs/link.c
index ce86ec69fe01..9562f5bba65c 100644
--- a/fs/cifs/link.c
+++ b/fs/cifs/link.c
@@ -48,10 +48,10 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode,
 /* No need to check for cross device links since server will do that
    BB note DFS case in future though (when we may have to check) */
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	fromName = build_path_from_dentry(old_file);
 	toName = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 	if((fromName == NULL) || (toName == NULL)) {
 		rc = -ENOMEM;
 		goto cifs_hl_exit;
@@ -103,9 +103,9 @@ cifs_follow_link(struct dentry *direntry, struct nameidata *nd)
 
 	xid = GetXid();
 
-	down(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&direntry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&direntry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&direntry->d_sb->s_vfs_rename_mutex);
 
 	if (!full_path)
 		goto out_no_free;
@@ -164,9 +164,9 @@ cifs_symlink(struct inode *inode, struct dentry *direntry, const char *symname)
 	cifs_sb = CIFS_SB(inode->i_sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&inode->i_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&inode->i_sb->s_vfs_rename_sem);
+	mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);
 
 	if(full_path == NULL) {
 		FreeXid(xid);
@@ -232,9 +232,9 @@ cifs_readlink(struct dentry *direntry, char __user *pBuffer, int buflen)
 
 /* BB would it be safe against deadlock to grab this sem 
       even though rename itself grabs the sem and calls lookup? */
-/*       down(&inode->i_sb->s_vfs_rename_sem);*/
+/*       mutex_lock(&inode->i_sb->s_vfs_rename_mutex);*/
 	full_path = build_path_from_dentry(direntry);
-/*       up(&inode->i_sb->s_vfs_rename_sem);*/
+/*       mutex_unlock(&inode->i_sb->s_vfs_rename_mutex);*/
 
 	if(full_path == NULL) {
 		FreeXid(xid);
diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c
index 405d4b7ec3ac..2f6e2825571e 100644
--- a/fs/cifs/readdir.c
+++ b/fs/cifs/readdir.c
@@ -197,10 +197,10 @@ static void fill_in_inode(struct inode *tmp_inode,
 
 	if (allocation_size < end_of_file)
 		cFYI(1, ("May be sparse file, allocation less than file size"));
-	cFYI(1,
-	     ("File Size %ld and blocks %ld and blocksize %ld",
-	      (unsigned long)tmp_inode->i_size, tmp_inode->i_blocks,
-	      tmp_inode->i_blksize));
+	cFYI(1, ("File Size %ld and blocks %llu and blocksize %ld",
+		(unsigned long)tmp_inode->i_size,
+		(unsigned long long)tmp_inode->i_blocks,
+		tmp_inode->i_blksize));
 	if (S_ISREG(tmp_inode->i_mode)) {
 		cFYI(1, ("File inode"));
 		tmp_inode->i_op = &cifs_file_inode_ops;
@@ -404,9 +404,9 @@ static int initiate_cifs_search(const int xid, struct file *file)
 	if(pTcon == NULL)
 		return -EINVAL;
 
-	down(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_lock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(file->f_dentry);
-	up(&file->f_dentry->d_sb->s_vfs_rename_sem);
+	mutex_unlock(&file->f_dentry->d_sb->s_vfs_rename_mutex);
 
 	if(full_path == NULL) {
 		return -ENOMEM;
diff --git a/fs/cifs/xattr.c b/fs/cifs/xattr.c
index 777e3363c2a4..3938444d87b2 100644
--- a/fs/cifs/xattr.c
+++ b/fs/cifs/xattr.c
@@ -62,9 +62,9 @@ int cifs_removexattr(struct dentry * direntry, const char * ea_name)
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
                                                                                      
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -116,9 +116,9 @@ int cifs_setxattr(struct dentry * direntry, const char * ea_name,
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -223,9 +223,9 @@ ssize_t cifs_getxattr(struct dentry * direntry, const char * ea_name,
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
@@ -341,9 +341,9 @@ ssize_t cifs_listxattr(struct dentry * direntry, char * data, size_t buf_size)
 	cifs_sb = CIFS_SB(sb);
 	pTcon = cifs_sb->tcon;
 
-	down(&sb->s_vfs_rename_sem);
+	mutex_lock(&sb->s_vfs_rename_mutex);
 	full_path = build_path_from_dentry(direntry);
-	up(&sb->s_vfs_rename_sem);
+	mutex_unlock(&sb->s_vfs_rename_mutex);
 	if(full_path == NULL) {
 		FreeXid(xid);
 		return -ENOMEM;
diff --git a/fs/coda/cache.c b/fs/coda/cache.c
index c607d923350a..5d0527133266 100644
--- a/fs/coda/cache.c
+++ b/fs/coda/cache.c
@@ -51,7 +51,7 @@ void coda_cache_clear_all(struct super_block *sb)
         struct coda_sb_info *sbi;
 
         sbi = coda_sbp(sb);
-        if (!sbi) BUG();
+	BUG_ON(!sbi);
 
 	atomic_inc(&permission_epoch);
 }
diff --git a/fs/coda/cnode.c b/fs/coda/cnode.c
index 23aeef5aa814..4c9fecbfa91f 100644
--- a/fs/coda/cnode.c
+++ b/fs/coda/cnode.c
@@ -120,8 +120,7 @@ void coda_replace_fid(struct inode *inode, struct CodaFid *oldfid,
 	
 	cii = ITOC(inode);
 
-	if (!coda_fideq(&cii->c_fid, oldfid))
-		BUG();
+	BUG_ON(!coda_fideq(&cii->c_fid, oldfid));
 
 	/* replace fid and rehash inode */
 	/* XXX we probably need to hold some lock here! */
diff --git a/fs/coda/coda_int.h b/fs/coda/coda_int.h
new file mode 100644
index 000000000000..9e6338fea514
--- /dev/null
+++ b/fs/coda/coda_int.h
@@ -0,0 +1,13 @@
+#ifndef _CODA_INT_
+#define _CODA_INT_
+
+extern struct file_system_type coda_fs_type;
+
+void coda_destroy_inodecache(void);
+int coda_init_inodecache(void);
+int coda_fsync(struct file *coda_file, struct dentry *coda_dentry,
+	       int datasync);
+
+#endif  /*  _CODA_INT_  */
+
+
diff --git a/fs/coda/dir.c b/fs/coda/dir.c
index 8f1a517f8b4e..71f2ea632e53 100644
--- a/fs/coda/dir.c
+++ b/fs/coda/dir.c
@@ -27,6 +27,8 @@
 #include <linux/coda_cache.h>
 #include <linux/coda_proc.h>
 
+#include "coda_int.h"
+
 /* dir inode-ops */
 static int coda_create(struct inode *dir, struct dentry *new, int mode, struct nameidata *nd);
 static struct dentry *coda_lookup(struct inode *dir, struct dentry *target, struct nameidata *nd);
@@ -50,7 +52,6 @@ static int coda_dentry_delete(struct dentry *);
 /* support routines */
 static int coda_venus_readdir(struct file *filp, filldir_t filldir,
 			      void *dirent, struct dentry *dir);
-int coda_fsync(struct file *, struct dentry *dentry, int datasync);
 
 /* same as fs/bad_inode.c */
 static int coda_return_EIO(void)
@@ -81,7 +82,7 @@ struct inode_operations coda_dir_inode_operations =
 	.setattr	= coda_setattr,
 };
 
-struct file_operations coda_dir_operations = {
+const struct file_operations coda_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= coda_readdir,
diff --git a/fs/coda/file.c b/fs/coda/file.c
index 30b4630bd735..7c2642431fa5 100644
--- a/fs/coda/file.c
+++ b/fs/coda/file.c
@@ -24,6 +24,8 @@
 #include <linux/coda_psdev.h>
 #include <linux/coda_proc.h>
 
+#include "coda_int.h"
+
 /* if CODA_STORE fails with EOPNOTSUPP, venus clearly doesn't support
  * CODA_STORE/CODA_RELEASE and we fall back on using the CODA_CLOSE upcall */
 static int use_coda_close;
@@ -286,7 +288,7 @@ int coda_fsync(struct file *coda_file, struct dentry *coda_dentry, int datasync)
 	return err;
 }
 
-struct file_operations coda_file_operations = {
+const struct file_operations coda_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= coda_file_read,
 	.write		= coda_file_write,
diff --git a/fs/coda/inode.c b/fs/coda/inode.c
index 04a73fb4848f..ada1a81df6bd 100644
--- a/fs/coda/inode.c
+++ b/fs/coda/inode.c
@@ -31,6 +31,8 @@
 #include <linux/coda_fs_i.h>
 #include <linux/coda_cache.h>
 
+#include "coda_int.h"
+
 /* VFS super_block ops */
 static void coda_clear_inode(struct inode *);
 static void coda_put_super(struct super_block *);
@@ -69,7 +71,7 @@ int coda_init_inodecache(void)
 {
 	coda_inode_cachep = kmem_cache_create("coda_inode_cache",
 				sizeof(struct coda_inode_info),
-				0, SLAB_RECLAIM_ACCOUNT,
+				0, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
 				init_once, NULL);
 	if (coda_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/coda/pioctl.c b/fs/coda/pioctl.c
index 127714936c66..214822be87bd 100644
--- a/fs/coda/pioctl.c
+++ b/fs/coda/pioctl.c
@@ -36,7 +36,7 @@ struct inode_operations coda_ioctl_inode_operations =
 	.setattr	= coda_setattr,
 };
 
-struct file_operations coda_ioctl_operations = {
+const struct file_operations coda_ioctl_operations = {
 	.owner		= THIS_MODULE,
 	.ioctl		= coda_pioctl,
 };
diff --git a/fs/coda/psdev.c b/fs/coda/psdev.c
index 6a3df88accfe..6c6771db36da 100644
--- a/fs/coda/psdev.c
+++ b/fs/coda/psdev.c
@@ -48,12 +48,9 @@
 #include <linux/coda_psdev.h>
 #include <linux/coda_proc.h>
 
-#define upc_free(r) kfree(r)
+#include "coda_int.h"
 
-/* 
- * Coda stuff
- */
-extern struct file_system_type coda_fs_type;
+#define upc_free(r) kfree(r)
 
 /* statistics */
 int           coda_hard;         /* allows signals during upcalls */
@@ -345,7 +342,7 @@ static int coda_psdev_release(struct inode * inode, struct file * file)
 }
 
 
-static struct file_operations coda_psdev_fops = {
+static const struct file_operations coda_psdev_fops = {
 	.owner		= THIS_MODULE,
 	.read		= coda_psdev_read,
 	.write		= coda_psdev_write,
@@ -394,8 +391,6 @@ out:
 MODULE_AUTHOR("Peter J. Braam <braam@cs.cmu.edu>");
 MODULE_LICENSE("GPL");
 
-extern int coda_init_inodecache(void);
-extern void coda_destroy_inodecache(void);
 static int __init init_coda(void)
 {
 	int status;
diff --git a/fs/compat.c b/fs/compat.c
index 5333c7d7427f..7f8e26ea427c 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -114,6 +114,7 @@ asmlinkage long compat_sys_newlstat(char __user * filename,
 	return error;
 }
 
+#ifndef __ARCH_WANT_STAT64
 asmlinkage long compat_sys_newfstatat(unsigned int dfd, char __user *filename,
 		struct compat_stat __user *statbuf, int flag)
 {
@@ -134,6 +135,7 @@ asmlinkage long compat_sys_newfstatat(unsigned int dfd, char __user *filename,
 out:
 	return error;
 }
+#endif
 
 asmlinkage long compat_sys_newfstat(unsigned int fd,
 		struct compat_stat __user * statbuf)
@@ -1474,10 +1476,9 @@ int compat_do_execve(char * filename,
 	int i;
 
 	retval = -ENOMEM;
-	bprm = kmalloc(sizeof(*bprm), GFP_KERNEL);
+	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
 	if (!bprm)
 		goto out_ret;
-	memset(bprm, 0, sizeof(*bprm));
 
 	file = open_exec(filename);
 	retval = PTR_ERR(file);
@@ -1638,15 +1639,6 @@ void compat_set_fd_set(unsigned long nr, compat_ulong_t __user *ufdset,
  * This is a virtual copy of sys_select from fs/select.c and probably
  * should be compared to it from time to time
  */
-static void *select_bits_alloc(int size)
-{
-	return kmalloc(6 * size, GFP_KERNEL);
-}
-
-static void select_bits_free(void *bits, int size)
-{
-	kfree(bits);
-}
 
 /*
  * We can actually return ERESTARTSYS instead of EINTR, but I'd
@@ -1685,7 +1677,7 @@ int compat_core_sys_select(int n, compat_ulong_t __user *inp,
 	 */
 	ret = -ENOMEM;
 	size = FDS_BYTES(n);
-	bits = select_bits_alloc(size);
+	bits = kmalloc(6 * size, GFP_KERNEL);
 	if (!bits)
 		goto out_nofds;
 	fds.in      = (unsigned long *)  bits;
@@ -1719,7 +1711,7 @@ int compat_core_sys_select(int n, compat_ulong_t __user *inp,
 	compat_set_fd_set(n, exp, fds.res_ex);
 
 out:
-	select_bits_free(bits, size);
+	kfree(bits);
 out_nofds:
 	return ret;
 }
@@ -2168,9 +2160,12 @@ asmlinkage long compat_sys_nfsservctl(int cmd, struct compat_nfsctl_arg __user *
 
 	default:
 		err = -EINVAL;
-		goto done;
+		break;
 	}
 
+	if (err)
+		goto done;
+
 	oldfs = get_fs();
 	set_fs(KERNEL_DS);
 	/* The __user pointer casts are valid because of the set_fs() */
diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c
index c666769a875d..d2c38875ab29 100644
--- a/fs/compat_ioctl.c
+++ b/fs/compat_ioctl.c
@@ -72,6 +72,7 @@
 #include <linux/i2c-dev.h>
 #include <linux/wireless.h>
 #include <linux/atalk.h>
+#include <linux/blktrace_api.h>
 
 #include <net/sock.h>          /* siocdevprivate_ioctl */
 #include <net/bluetooth/bluetooth.h>
@@ -1521,8 +1522,7 @@ static struct {
 	{ ATM_QUERYLOOP32,   ATM_QUERYLOOP }
 };
 
-#define NR_ATM_IOCTL (sizeof(atm_ioctl_map)/sizeof(atm_ioctl_map[0]))
-
+#define NR_ATM_IOCTL ARRAY_SIZE(atm_ioctl_map)
 
 static int do_atm_iobuf(unsigned int fd, unsigned int cmd, unsigned long arg)
 {
@@ -1823,7 +1823,7 @@ static struct {
 	{ FDWERRORGET32, FDWERRORGET }
 };
 
-#define NR_FD_IOCTL_TRANS (sizeof(fd_ioctl_trans_table)/sizeof(fd_ioctl_trans_table[0]))
+#define NR_FD_IOCTL_TRANS ARRAY_SIZE(fd_ioctl_trans_table)
 
 static int fd_ioctl_trans(unsigned int fd, unsigned int cmd, unsigned long arg)
 {
diff --git a/fs/configfs/configfs_internal.h b/fs/configfs/configfs_internal.h
index f70e46951b37..3f4ff7a242b9 100644
--- a/fs/configfs/configfs_internal.h
+++ b/fs/configfs/configfs_internal.h
@@ -72,9 +72,9 @@ extern void configfs_release_fs(void);
 
 extern struct rw_semaphore configfs_rename_sem;
 extern struct super_block * configfs_sb;
-extern struct file_operations configfs_dir_operations;
-extern struct file_operations configfs_file_operations;
-extern struct file_operations bin_fops;
+extern const struct file_operations configfs_dir_operations;
+extern const struct file_operations configfs_file_operations;
+extern const struct file_operations bin_fops;
 extern struct inode_operations configfs_dir_inode_operations;
 extern struct inode_operations configfs_symlink_inode_operations;
 
diff --git a/fs/configfs/dir.c b/fs/configfs/dir.c
index ca60e3abef45..8ed9b06a9828 100644
--- a/fs/configfs/dir.c
+++ b/fs/configfs/dir.c
@@ -1027,7 +1027,7 @@ static loff_t configfs_dir_lseek(struct file * file, loff_t offset, int origin)
 	return offset;
 }
 
-struct file_operations configfs_dir_operations = {
+const struct file_operations configfs_dir_operations = {
 	.open		= configfs_dir_open,
 	.release	= configfs_dir_close,
 	.llseek		= configfs_dir_lseek,
diff --git a/fs/configfs/file.c b/fs/configfs/file.c
index 3921920d8716..f499803743e0 100644
--- a/fs/configfs/file.c
+++ b/fs/configfs/file.c
@@ -322,7 +322,7 @@ static int configfs_release(struct inode * inode, struct file * filp)
 	return 0;
 }
 
-struct file_operations configfs_file_operations = {
+const struct file_operations configfs_file_operations = {
 	.read		= configfs_read_file,
 	.write		= configfs_write_file,
 	.llseek		= generic_file_llseek,
diff --git a/fs/cramfs/inode.c b/fs/cramfs/inode.c
index 8ad52f5bf255..9efcc3a164e8 100644
--- a/fs/cramfs/inode.c
+++ b/fs/cramfs/inode.c
@@ -22,16 +22,17 @@
 #include <linux/cramfs_fs_sb.h>
 #include <linux/buffer_head.h>
 #include <linux/vfs.h>
+#include <linux/mutex.h>
 #include <asm/semaphore.h>
 
 #include <asm/uaccess.h>
 
 static struct super_operations cramfs_ops;
 static struct inode_operations cramfs_dir_inode_operations;
-static struct file_operations cramfs_directory_operations;
+static const struct file_operations cramfs_directory_operations;
 static struct address_space_operations cramfs_aops;
 
-static DECLARE_MUTEX(read_mutex);
+static DEFINE_MUTEX(read_mutex);
 
 
 /* These two macros may change in future, to provide better st_ino
@@ -250,20 +251,20 @@ static int cramfs_fill_super(struct super_block *sb, void *data, int silent)
 	memset(sbi, 0, sizeof(struct cramfs_sb_info));
 
 	/* Invalidate the read buffers on mount: think disk change.. */
-	down(&read_mutex);
+	mutex_lock(&read_mutex);
 	for (i = 0; i < READ_BUFFERS; i++)
 		buffer_blocknr[i] = -1;
 
 	/* Read the first block and get the superblock from it */
 	memcpy(&super, cramfs_read(sb, 0, sizeof(super)), sizeof(super));
-	up(&read_mutex);
+	mutex_unlock(&read_mutex);
 
 	/* Do sanity checks on the superblock */
 	if (super.magic != CRAMFS_MAGIC) {
 		/* check at 512 byte offset */
-		down(&read_mutex);
+		mutex_lock(&read_mutex);
 		memcpy(&super, cramfs_read(sb, 512, sizeof(super)), sizeof(super));
-		up(&read_mutex);
+		mutex_unlock(&read_mutex);
 		if (super.magic != CRAMFS_MAGIC) {
 			if (!silent)
 				printk(KERN_ERR "cramfs: wrong magic\n");
@@ -366,7 +367,7 @@ static int cramfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		mode_t mode;
 		int namelen, error;
 
-		down(&read_mutex);
+		mutex_lock(&read_mutex);
 		de = cramfs_read(sb, OFFSET(inode) + offset, sizeof(*de)+256);
 		name = (char *)(de+1);
 
@@ -379,7 +380,7 @@ static int cramfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		memcpy(buf, name, namelen);
 		ino = CRAMINO(de);
 		mode = de->mode;
-		up(&read_mutex);
+		mutex_unlock(&read_mutex);
 		nextoffset = offset + sizeof(*de) + namelen;
 		for (;;) {
 			if (!namelen) {
@@ -410,7 +411,7 @@ static struct dentry * cramfs_lookup(struct inode *dir, struct dentry *dentry, s
 	unsigned int offset = 0;
 	int sorted;
 
-	down(&read_mutex);
+	mutex_lock(&read_mutex);
 	sorted = CRAMFS_SB(dir->i_sb)->flags & CRAMFS_FLAG_SORTED_DIRS;
 	while (offset < dir->i_size) {
 		struct cramfs_inode *de;
@@ -433,7 +434,7 @@ static struct dentry * cramfs_lookup(struct inode *dir, struct dentry *dentry, s
 
 		for (;;) {
 			if (!namelen) {
-				up(&read_mutex);
+				mutex_unlock(&read_mutex);
 				return ERR_PTR(-EIO);
 			}
 			if (name[namelen-1])
@@ -447,7 +448,7 @@ static struct dentry * cramfs_lookup(struct inode *dir, struct dentry *dentry, s
 			continue;
 		if (!retval) {
 			struct cramfs_inode entry = *de;
-			up(&read_mutex);
+			mutex_unlock(&read_mutex);
 			d_add(dentry, get_cramfs_inode(dir->i_sb, &entry));
 			return NULL;
 		}
@@ -455,7 +456,7 @@ static struct dentry * cramfs_lookup(struct inode *dir, struct dentry *dentry, s
 		if (sorted)
 			break;
 	}
-	up(&read_mutex);
+	mutex_unlock(&read_mutex);
 	d_add(dentry, NULL);
 	return NULL;
 }
@@ -474,21 +475,21 @@ static int cramfs_readpage(struct file *file, struct page * page)
 		u32 start_offset, compr_len;
 
 		start_offset = OFFSET(inode) + maxblock*4;
-		down(&read_mutex);
+		mutex_lock(&read_mutex);
 		if (page->index)
 			start_offset = *(u32 *) cramfs_read(sb, blkptr_offset-4, 4);
 		compr_len = (*(u32 *) cramfs_read(sb, blkptr_offset, 4) - start_offset);
-		up(&read_mutex);
+		mutex_unlock(&read_mutex);
 		pgdata = kmap(page);
 		if (compr_len == 0)
 			; /* hole */
 		else {
-			down(&read_mutex);
+			mutex_lock(&read_mutex);
 			bytes_filled = cramfs_uncompress_block(pgdata,
 				 PAGE_CACHE_SIZE,
 				 cramfs_read(sb, start_offset, compr_len),
 				 compr_len);
-			up(&read_mutex);
+			mutex_unlock(&read_mutex);
 		}
 	} else
 		pgdata = kmap(page);
@@ -511,7 +512,7 @@ static struct address_space_operations cramfs_aops = {
 /*
  * A directory can only readdir
  */
-static struct file_operations cramfs_directory_operations = {
+static const struct file_operations cramfs_directory_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= cramfs_readdir,
diff --git a/fs/dcache.c b/fs/dcache.c
index 11dc83092d4a..19458d399502 100644
--- a/fs/dcache.c
+++ b/fs/dcache.c
@@ -34,9 +34,8 @@
 #include <linux/swap.h>
 #include <linux/bootmem.h>
 
-/* #define DCACHE_DEBUG 1 */
 
-int sysctl_vfs_cache_pressure = 100;
+int sysctl_vfs_cache_pressure __read_mostly = 100;
 EXPORT_SYMBOL_GPL(sysctl_vfs_cache_pressure);
 
  __cacheline_aligned_in_smp DEFINE_SPINLOCK(dcache_lock);
@@ -44,7 +43,7 @@ static seqlock_t rename_lock __cacheline_aligned_in_smp = SEQLOCK_UNLOCKED;
 
 EXPORT_SYMBOL(dcache_lock);
 
-static kmem_cache_t *dentry_cache; 
+static kmem_cache_t *dentry_cache __read_mostly;
 
 #define DNAME_INLINE_LEN (sizeof(struct dentry)-offsetof(struct dentry,d_iname))
 
@@ -59,9 +58,9 @@ static kmem_cache_t *dentry_cache;
 #define D_HASHBITS     d_hash_shift
 #define D_HASHMASK     d_hash_mask
 
-static unsigned int d_hash_mask;
-static unsigned int d_hash_shift;
-static struct hlist_head *dentry_hashtable;
+static unsigned int d_hash_mask __read_mostly;
+static unsigned int d_hash_shift __read_mostly;
+static struct hlist_head *dentry_hashtable __read_mostly;
 static LIST_HEAD(dentry_unused);
 
 /* Statistics gathering. */
@@ -325,10 +324,13 @@ static struct dentry * __d_find_alias(struct inode *inode, int want_discon)
 
 struct dentry * d_find_alias(struct inode *inode)
 {
-	struct dentry *de;
-	spin_lock(&dcache_lock);
-	de = __d_find_alias(inode, 0);
-	spin_unlock(&dcache_lock);
+	struct dentry *de = NULL;
+
+	if (!list_empty(&inode->i_dentry)) {
+		spin_lock(&dcache_lock);
+		de = __d_find_alias(inode, 0);
+		spin_unlock(&dcache_lock);
+	}
 	return de;
 }
 
@@ -486,6 +488,7 @@ repeat:
 			continue;
 		}
 		prune_one_dentry(dentry);
+		cond_resched_lock(&dcache_lock);
 		goto repeat;
 	}
 	spin_unlock(&dcache_lock);
@@ -599,10 +602,6 @@ resume:
 		 */
 		if (!list_empty(&dentry->d_subdirs)) {
 			this_parent = dentry;
-#ifdef DCACHE_DEBUG
-printk(KERN_DEBUG "select_parent: descending to %s/%s, found=%d\n",
-dentry->d_parent->d_name.name, dentry->d_name.name, found);
-#endif
 			goto repeat;
 		}
 	}
@@ -612,10 +611,6 @@ dentry->d_parent->d_name.name, dentry->d_name.name, found);
 	if (this_parent != parent) {
 		next = this_parent->d_u.d_child.next;
 		this_parent = this_parent->d_parent;
-#ifdef DCACHE_DEBUG
-printk(KERN_DEBUG "select_parent: ascending to %s/%s, found=%d\n",
-this_parent->d_parent->d_name.name, this_parent->d_name.name, found);
-#endif
 		goto resume;
 	}
 out:
@@ -794,11 +789,12 @@ struct dentry *d_alloc_name(struct dentry *parent, const char *name)
  
 void d_instantiate(struct dentry *entry, struct inode * inode)
 {
-	if (!list_empty(&entry->d_alias)) BUG();
+	BUG_ON(!list_empty(&entry->d_alias));
 	spin_lock(&dcache_lock);
 	if (inode)
 		list_add(&entry->d_alias, &inode->i_dentry);
 	entry->d_inode = inode;
+	fsnotify_d_instantiate(entry, inode);
 	spin_unlock(&dcache_lock);
 	security_d_instantiate(entry, inode);
 }
@@ -850,6 +846,7 @@ struct dentry *d_instantiate_unique(struct dentry *entry, struct inode *inode)
 	list_add(&entry->d_alias, &inode->i_dentry);
 do_negative:
 	entry->d_inode = inode;
+	fsnotify_d_instantiate(entry, inode);
 	spin_unlock(&dcache_lock);
 	security_d_instantiate(entry, inode);
 	return NULL;
@@ -980,6 +977,7 @@ struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry)
 		new = __d_find_alias(inode, 1);
 		if (new) {
 			BUG_ON(!(new->d_flags & DCACHE_DISCONNECTED));
+			fsnotify_d_instantiate(new, inode);
 			spin_unlock(&dcache_lock);
 			security_d_instantiate(new, inode);
 			d_rehash(dentry);
@@ -989,6 +987,7 @@ struct dentry *d_splice_alias(struct inode *inode, struct dentry *dentry)
 			/* d_instantiate takes dcache_lock, so we do it by hand */
 			list_add(&dentry->d_alias, &inode->i_dentry);
 			dentry->d_inode = inode;
+			fsnotify_d_instantiate(dentry, inode);
 			spin_unlock(&dcache_lock);
 			security_d_instantiate(dentry, inode);
 			d_rehash(dentry);
@@ -1173,6 +1172,9 @@ void d_delete(struct dentry * dentry)
 	spin_lock(&dentry->d_lock);
 	isdir = S_ISDIR(dentry->d_inode->i_mode);
 	if (atomic_read(&dentry->d_count) == 1) {
+		/* remove this and other inotify debug checks after 2.6.18 */
+		dentry->d_flags &= ~DCACHE_INOTIFY_PARENT_WATCHED;
+
 		dentry_iput(dentry);
 		fsnotify_nameremove(dentry, isdir);
 		return;
@@ -1339,6 +1341,7 @@ already_unhashed:
 
 	list_add(&dentry->d_u.d_child, &dentry->d_parent->d_subdirs);
 	spin_unlock(&target->d_lock);
+	fsnotify_d_move(dentry);
 	spin_unlock(&dentry->d_lock);
 	write_sequnlock(&rename_lock);
 	spin_unlock(&dcache_lock);
@@ -1682,7 +1685,8 @@ static void __init dcache_init(unsigned long mempages)
 	dentry_cache = kmem_cache_create("dentry_cache",
 					 sizeof(struct dentry),
 					 0,
-					 SLAB_RECLAIM_ACCOUNT|SLAB_PANIC,
+					 (SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
+					 SLAB_MEM_SPREAD),
 					 NULL, NULL);
 	
 	set_shrinker(DEFAULT_SEEKS, shrink_dcache_memory);
@@ -1706,10 +1710,10 @@ static void __init dcache_init(unsigned long mempages)
 }
 
 /* SLAB cache for __getname() consumers */
-kmem_cache_t *names_cachep;
+kmem_cache_t *names_cachep __read_mostly;
 
 /* SLAB cache for file structures */
-kmem_cache_t *filp_cachep;
+kmem_cache_t *filp_cachep __read_mostly;
 
 EXPORT_SYMBOL(d_genocide);
 
diff --git a/fs/dcookies.c b/fs/dcookies.c
index f8274a8f83bd..8749339bf4f6 100644
--- a/fs/dcookies.c
+++ b/fs/dcookies.c
@@ -23,6 +23,7 @@
 #include <linux/mm.h>
 #include <linux/errno.h>
 #include <linux/dcookies.h>
+#include <linux/mutex.h>
 #include <asm/uaccess.h>
 
 /* The dcookies are allocated from a kmem_cache and
@@ -36,10 +37,10 @@ struct dcookie_struct {
 };
 
 static LIST_HEAD(dcookie_users);
-static DECLARE_MUTEX(dcookie_sem);
-static kmem_cache_t * dcookie_cache;
-static struct list_head * dcookie_hashtable;
-static size_t hash_size;
+static DEFINE_MUTEX(dcookie_mutex);
+static kmem_cache_t *dcookie_cache __read_mostly;
+static struct list_head *dcookie_hashtable __read_mostly;
+static size_t hash_size __read_mostly;
 
 static inline int is_live(void)
 {
@@ -114,7 +115,7 @@ int get_dcookie(struct dentry * dentry, struct vfsmount * vfsmnt,
 	int err = 0;
 	struct dcookie_struct * dcs;
 
-	down(&dcookie_sem);
+	mutex_lock(&dcookie_mutex);
 
 	if (!is_live()) {
 		err = -EINVAL;
@@ -134,7 +135,7 @@ int get_dcookie(struct dentry * dentry, struct vfsmount * vfsmnt,
 	*cookie = dcookie_value(dcs);
 
 out:
-	up(&dcookie_sem);
+	mutex_unlock(&dcookie_mutex);
 	return err;
 }
 
@@ -157,7 +158,7 @@ asmlinkage long sys_lookup_dcookie(u64 cookie64, char __user * buf, size_t len)
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	down(&dcookie_sem);
+	mutex_lock(&dcookie_mutex);
 
 	if (!is_live()) {
 		err = -EINVAL;
@@ -192,7 +193,7 @@ asmlinkage long sys_lookup_dcookie(u64 cookie64, char __user * buf, size_t len)
 out_free:
 	kfree(kbuf);
 out:
-	up(&dcookie_sem);
+	mutex_unlock(&dcookie_mutex);
 	return err;
 }
 
@@ -290,7 +291,7 @@ struct dcookie_user * dcookie_register(void)
 {
 	struct dcookie_user * user;
 
-	down(&dcookie_sem);
+	mutex_lock(&dcookie_mutex);
 
 	user = kmalloc(sizeof(struct dcookie_user), GFP_KERNEL);
 	if (!user)
@@ -302,7 +303,7 @@ struct dcookie_user * dcookie_register(void)
 	list_add(&user->next, &dcookie_users);
 
 out:
-	up(&dcookie_sem);
+	mutex_unlock(&dcookie_mutex);
 	return user;
 out_free:
 	kfree(user);
@@ -313,7 +314,7 @@ out_free:
 
 void dcookie_unregister(struct dcookie_user * user)
 {
-	down(&dcookie_sem);
+	mutex_lock(&dcookie_mutex);
 
 	list_del(&user->next);
 	kfree(user);
@@ -321,7 +322,7 @@ void dcookie_unregister(struct dcookie_user * user)
 	if (!is_live())
 		dcookie_exit();
 
-	up(&dcookie_sem);
+	mutex_unlock(&dcookie_mutex);
 }
 
 EXPORT_SYMBOL_GPL(dcookie_register);
diff --git a/fs/debugfs/file.c b/fs/debugfs/file.c
index d575452cd9f7..66a505422e5c 100644
--- a/fs/debugfs/file.c
+++ b/fs/debugfs/file.c
@@ -39,7 +39,7 @@ static int default_open(struct inode *inode, struct file *file)
 	return 0;
 }
 
-struct file_operations debugfs_file_operations = {
+const struct file_operations debugfs_file_operations = {
 	.read =		default_read_file,
 	.write =	default_write_file,
 	.open =		default_open,
@@ -213,7 +213,7 @@ static ssize_t write_file_bool(struct file *file, const char __user *user_buf,
 	return count;
 }
 
-static struct file_operations fops_bool = {
+static const struct file_operations fops_bool = {
 	.read =		read_file_bool,
 	.write =	write_file_bool,
 	.open =		default_open,
@@ -251,3 +251,49 @@ struct dentry *debugfs_create_bool(const char *name, mode_t mode,
 }
 EXPORT_SYMBOL_GPL(debugfs_create_bool);
 
+static ssize_t read_file_blob(struct file *file, char __user *user_buf,
+			      size_t count, loff_t *ppos)
+{
+	struct debugfs_blob_wrapper *blob = file->private_data;
+	return simple_read_from_buffer(user_buf, count, ppos, blob->data,
+			blob->size);
+}
+
+static struct file_operations fops_blob = {
+	.read =		read_file_blob,
+	.open =		default_open,
+};
+
+/**
+ * debugfs_create_blob - create a file in the debugfs filesystem that is
+ * used to read and write a binary blob.
+ *
+ * @name: a pointer to a string containing the name of the file to create.
+ * @mode: the permission that the file should have
+ * @parent: a pointer to the parent dentry for this file.  This should be a
+ *          directory dentry if set.  If this paramater is NULL, then the
+ *          file will be created in the root of the debugfs filesystem.
+ * @blob: a pointer to a struct debugfs_blob_wrapper which contains a pointer
+ *        to the blob data and the size of the data.
+ *
+ * This function creates a file in debugfs with the given name that exports
+ * @blob->data as a binary blob. If the @mode variable is so set it can be
+ * read from. Writing is not supported.
+ *
+ * This function will return a pointer to a dentry if it succeeds.  This
+ * pointer must be passed to the debugfs_remove() function when the file is
+ * to be removed (no automatic cleanup happens if your module is unloaded,
+ * you are responsible here.)  If an error occurs, NULL will be returned.
+ *
+ * If debugfs is not enabled in the kernel, the value -ENODEV will be
+ * returned.  It is not wise to check for this value, but rather, check for
+ * NULL or !NULL instead as to eliminate the need for #ifdef in the calling
+ * code.
+ */
+struct dentry *debugfs_create_blob(const char *name, mode_t mode,
+				   struct dentry *parent,
+				   struct debugfs_blob_wrapper *blob)
+{
+	return debugfs_create_file(name, mode, parent, blob, &fops_blob);
+}
+EXPORT_SYMBOL_GPL(debugfs_create_blob);
diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c
index d4f1a2cddd47..85d166cdcae4 100644
--- a/fs/debugfs/inode.c
+++ b/fs/debugfs/inode.c
@@ -191,7 +191,7 @@ static int debugfs_create_by_name(const char *name, mode_t mode,
  */
 struct dentry *debugfs_create_file(const char *name, mode_t mode,
 				   struct dentry *parent, void *data,
-				   struct file_operations *fops)
+				   const struct file_operations *fops)
 {
 	struct dentry *dentry = NULL;
 	int error;
diff --git a/fs/devfs/base.c b/fs/devfs/base.c
index b621521e09d4..52f5059c4f31 100644
--- a/fs/devfs/base.c
+++ b/fs/devfs/base.c
@@ -856,14 +856,14 @@ static int devfsd_close(struct inode *inode, struct file *file);
 #ifdef CONFIG_DEVFS_DEBUG
 static ssize_t stat_read(struct file *file, char __user *buf, size_t len,
 			 loff_t * ppos);
-static struct file_operations stat_fops = {
+static const struct file_operations stat_fops = {
 	.open = nonseekable_open,
 	.read = stat_read,
 };
 #endif
 
 /*  Devfs daemon file operations  */
-static struct file_operations devfsd_fops = {
+static const struct file_operations devfsd_fops = {
 	.open = nonseekable_open,
 	.read = devfsd_read,
 	.ioctl = devfsd_ioctl,
@@ -1842,8 +1842,8 @@ static int try_modload(struct devfs_entry *parent, struct fs_info *fs_info,
 
 static struct inode_operations devfs_iops;
 static struct inode_operations devfs_dir_iops;
-static struct file_operations devfs_fops;
-static struct file_operations devfs_dir_fops;
+static const struct file_operations devfs_fops;
+static const struct file_operations devfs_dir_fops;
 static struct inode_operations devfs_symlink_iops;
 
 static int devfs_notify_change(struct dentry *dentry, struct iattr *iattr)
@@ -2061,11 +2061,11 @@ static int devfs_open(struct inode *inode, struct file *file)
 	return err;
 }				/*  End Function devfs_open  */
 
-static struct file_operations devfs_fops = {
+static const struct file_operations devfs_fops = {
 	.open = devfs_open,
 };
 
-static struct file_operations devfs_dir_fops = {
+static const struct file_operations devfs_dir_fops = {
 	.read = generic_read_dir,
 	.readdir = devfs_readdir,
 };
diff --git a/fs/devpts/inode.c b/fs/devpts/inode.c
index bfb8a230bac9..14c5620b5cab 100644
--- a/fs/devpts/inode.c
+++ b/fs/devpts/inode.c
@@ -18,6 +18,7 @@
 #include <linux/mount.h>
 #include <linux/tty.h>
 #include <linux/devpts_fs.h>
+#include <linux/parser.h>
 
 #define DEVPTS_SUPER_MAGIC 0x1cd1
 
@@ -32,39 +33,60 @@ static struct {
 	umode_t mode;
 } config = {.mode = 0600};
 
+enum {
+	Opt_uid, Opt_gid, Opt_mode,
+	Opt_err
+};
+
+static match_table_t tokens = {
+	{Opt_uid, "uid=%u"},
+	{Opt_gid, "gid=%u"},
+	{Opt_mode, "mode=%o"},
+	{Opt_err, NULL}
+};
+
 static int devpts_remount(struct super_block *sb, int *flags, char *data)
 {
-	int setuid = 0;
-	int setgid = 0;
-	uid_t uid = 0;
-	gid_t gid = 0;
-	umode_t mode = 0600;
-	char *this_char;
-
-	this_char = NULL;
-	while ((this_char = strsep(&data, ",")) != NULL) {
-		int n;
-		char dummy;
-		if (!*this_char)
+	char *p;
+
+	config.setuid  = 0;
+	config.setgid  = 0;
+	config.uid     = 0;
+	config.gid     = 0;
+	config.mode    = 0600;
+
+	while ((p = strsep(&data, ",")) != NULL) {
+		substring_t args[MAX_OPT_ARGS];
+		int token;
+		int option;
+
+		if (!*p)
 			continue;
-		if (sscanf(this_char, "uid=%i%c", &n, &dummy) == 1) {
-			setuid = 1;
-			uid = n;
-		} else if (sscanf(this_char, "gid=%i%c", &n, &dummy) == 1) {
-			setgid = 1;
-			gid = n;
-		} else if (sscanf(this_char, "mode=%o%c", &n, &dummy) == 1)
-			mode = n & ~S_IFMT;
-		else {
-			printk("devpts: called with bogus options\n");
+
+		token = match_token(p, tokens, args);
+		switch (token) {
+		case Opt_uid:
+			if (match_int(&args[0], &option))
+				return -EINVAL;
+			config.uid = option;
+			config.setuid = 1;
+			break;
+		case Opt_gid:
+			if (match_int(&args[0], &option))
+				return -EINVAL;
+			config.gid = option;
+			config.setgid = 1;
+			break;
+		case Opt_mode:
+			if (match_octal(&args[0], &option))
+				return -EINVAL;
+			config.mode = option & ~S_IFMT;
+			break;
+		default:
+			printk(KERN_ERR "devpts: called with bogus options\n");
 			return -EINVAL;
 		}
 	}
-	config.setuid  = setuid;
-	config.setgid  = setgid;
-	config.uid     = uid;
-	config.gid     = gid;
-	config.mode    = mode;
 
 	return 0;
 }
diff --git a/fs/direct-io.c b/fs/direct-io.c
index 27f3e787faca..910a8ed74b5d 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -86,12 +86,12 @@ struct dio {
 	unsigned first_block_in_page;	/* doesn't change, Used only once */
 	int boundary;			/* prev block is at a boundary */
 	int reap_counter;		/* rate limit reaping */
-	get_blocks_t *get_blocks;	/* block mapping function */
+	get_block_t *get_block;		/* block mapping function */
 	dio_iodone_t *end_io;		/* IO completion function */
 	sector_t final_block_in_bio;	/* current final block in bio + 1 */
 	sector_t next_block_for_io;	/* next block to be put under IO,
 					   in dio_blocks units */
-	struct buffer_head map_bh;	/* last get_blocks() result */
+	struct buffer_head map_bh;	/* last get_block() result */
 
 	/*
 	 * Deferred addition of a page to the dio.  These variables are
@@ -129,6 +129,7 @@ struct dio {
 	/* AIO related stuff */
 	struct kiocb *iocb;		/* kiocb */
 	int is_async;			/* is IO async ? */
+	int io_error;			/* IO error in completion path */
 	ssize_t result;                 /* IO result */
 };
 
@@ -210,9 +211,9 @@ static struct page *dio_get_page(struct dio *dio)
 
 /*
  * Called when all DIO BIO I/O has been completed - let the filesystem
- * know, if it registered an interest earlier via get_blocks.  Pass the
+ * know, if it registered an interest earlier via get_block.  Pass the
  * private field of the map buffer_head so that filesystems can use it
- * to hold additional state between get_blocks calls and dio_complete.
+ * to hold additional state between get_block calls and dio_complete.
  */
 static void dio_complete(struct dio *dio, loff_t offset, ssize_t bytes)
 {
@@ -250,6 +251,10 @@ static void finished_one_bio(struct dio *dio)
 			    ((offset + transferred) > dio->i_size))
 				transferred = dio->i_size - offset;
 
+			/* check for error in completion path */
+			if (dio->io_error)
+				transferred = dio->io_error;
+
 			dio_complete(dio, offset, transferred);
 
 			/* Complete AIO later if falling back to buffered i/o */
@@ -406,7 +411,7 @@ static int dio_bio_complete(struct dio *dio, struct bio *bio)
 	int page_no;
 
 	if (!uptodate)
-		dio->result = -EIO;
+		dio->io_error = -EIO;
 
 	if (dio->is_async && dio->rw == READ) {
 		bio_check_pages_dirty(bio);	/* transfers ownership */
@@ -488,7 +493,7 @@ static int dio_bio_reap(struct dio *dio)
  * The fs is allowed to map lots of blocks at once.  If it wants to do that,
  * it uses the passed inode-relative block number as the file offset, as usual.
  *
- * get_blocks() is passed the number of i_blkbits-sized blocks which direct_io
+ * get_block() is passed the number of i_blkbits-sized blocks which direct_io
  * has remaining to do.  The fs should not map more than this number of blocks.
  *
  * If the fs has mapped a lot of blocks, it should populate bh->b_size to
@@ -501,7 +506,7 @@ static int dio_bio_reap(struct dio *dio)
  * In the case of filesystem holes: the fs may return an arbitrarily-large
  * hole by returning an appropriate value in b_size and by clearing
  * buffer_mapped().  However the direct-io code will only process holes one
- * block at a time - it will repeatedly call get_blocks() as it walks the hole.
+ * block at a time - it will repeatedly call get_block() as it walks the hole.
  */
 static int get_more_blocks(struct dio *dio)
 {
@@ -519,8 +524,6 @@ static int get_more_blocks(struct dio *dio)
 	 */
 	ret = dio->page_errors;
 	if (ret == 0) {
-		map_bh->b_state = 0;
-		map_bh->b_size = 0;
 		BUG_ON(dio->block_in_file >= dio->final_block_in_request);
 		fs_startblk = dio->block_in_file >> dio->blkfactor;
 		dio_count = dio->final_block_in_request - dio->block_in_file;
@@ -529,6 +532,9 @@ static int get_more_blocks(struct dio *dio)
 		if (dio_count & blkmask)	
 			fs_count++;
 
+		map_bh->b_state = 0;
+		map_bh->b_size = fs_count << dio->inode->i_blkbits;
+
 		create = dio->rw == WRITE;
 		if (dio->lock_type == DIO_LOCKING) {
 			if (dio->block_in_file < (i_size_read(dio->inode) >>
@@ -537,13 +543,14 @@ static int get_more_blocks(struct dio *dio)
 		} else if (dio->lock_type == DIO_NO_LOCKING) {
 			create = 0;
 		}
+
 		/*
 		 * For writes inside i_size we forbid block creations: only
 		 * overwrites are permitted.  We fall back to buffered writes
 		 * at a higher level for inside-i_size block-instantiating
 		 * writes.
 		 */
-		ret = (*dio->get_blocks)(dio->inode, fs_startblk, fs_count,
+		ret = (*dio->get_block)(dio->inode, fs_startblk,
 						map_bh, create);
 	}
 	return ret;
@@ -778,11 +785,11 @@ static void dio_zero_block(struct dio *dio, int end)
  * happily perform page-sized but 512-byte aligned IOs.  It is important that
  * blockdev IO be able to have fine alignment and large sizes.
  *
- * So what we do is to permit the ->get_blocks function to populate bh.b_size
+ * So what we do is to permit the ->get_block function to populate bh.b_size
  * with the size of IO which is permitted at this offset and this i_blkbits.
  *
  * For best results, the blockdev should be set up with 512-byte i_blkbits and
- * it should set b_size to PAGE_SIZE or more inside get_blocks().  This gives
+ * it should set b_size to PAGE_SIZE or more inside get_block().  This gives
  * fine alignment but still allows this function to work in PAGE_SIZE units.
  */
 static int do_direct_IO(struct dio *dio)
@@ -942,7 +949,7 @@ out:
 static ssize_t
 direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode, 
 	const struct iovec *iov, loff_t offset, unsigned long nr_segs, 
-	unsigned blkbits, get_blocks_t get_blocks, dio_iodone_t end_io,
+	unsigned blkbits, get_block_t get_block, dio_iodone_t end_io,
 	struct dio *dio)
 {
 	unsigned long user_addr; 
@@ -964,13 +971,14 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
 
 	dio->boundary = 0;
 	dio->reap_counter = 0;
-	dio->get_blocks = get_blocks;
+	dio->get_block = get_block;
 	dio->end_io = end_io;
 	dio->map_bh.b_private = NULL;
 	dio->final_block_in_bio = -1;
 	dio->next_block_for_io = -1;
 
 	dio->page_errors = 0;
+	dio->io_error = 0;
 	dio->result = 0;
 	dio->iocb = iocb;
 	dio->i_size = i_size_read(inode);
@@ -1171,7 +1179,7 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
 ssize_t
 __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 	struct block_device *bdev, const struct iovec *iov, loff_t offset, 
-	unsigned long nr_segs, get_blocks_t get_blocks, dio_iodone_t end_io,
+	unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
 	int dio_lock_type)
 {
 	int seg;
@@ -1267,7 +1275,7 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 		(end > i_size_read(inode)));
 
 	retval = direct_io_worker(rw, iocb, inode, iov, offset,
-				nr_segs, blkbits, get_blocks, end_io, dio);
+				nr_segs, blkbits, get_block, end_io, dio);
 
 	if (rw == READ && dio_lock_type == DIO_LOCKING)
 		release_i_mutex = 0;
diff --git a/fs/dnotify.c b/fs/dnotify.c
index f3b540dd5d11..f932591df5a4 100644
--- a/fs/dnotify.c
+++ b/fs/dnotify.c
@@ -21,9 +21,9 @@
 #include <linux/spinlock.h>
 #include <linux/slab.h>
 
-int dir_notify_enable = 1;
+int dir_notify_enable __read_mostly = 1;
 
-static kmem_cache_t *dn_cache;
+static kmem_cache_t *dn_cache __read_mostly;
 
 static void redo_inode_mask(struct inode *inode)
 {
diff --git a/fs/dquot.c b/fs/dquot.c
index 1966c890b48d..6b3886920939 100644
--- a/fs/dquot.c
+++ b/fs/dquot.c
@@ -103,12 +103,12 @@
  * (these locking rules also apply for S_NOQUOTA flag in the inode - note that
  * for altering the flag i_mutex is also needed).  If operation is holding
  * reference to dquot in other way (e.g. quotactl ops) it must be guarded by
- * dqonoff_sem.
+ * dqonoff_mutex.
  * This locking assures that:
  *   a) update/access to dquot pointers in inode is serialized
  *   b) everyone is guarded against invalidate_dquots()
  *
- * Each dquot has its dq_lock semaphore. Locked dquots might not be referenced
+ * Each dquot has its dq_lock mutex. Locked dquots might not be referenced
  * from inodes (dquot_alloc_space() and such don't check the dq_lock).
  * Currently dquot is locked only when it is being read to memory (or space for
  * it is being allocated) on the first dqget() and when it is being released on
@@ -118,9 +118,9 @@
  * spinlock to internal buffers before writing.
  *
  * Lock ordering (including related VFS locks) is the following:
- *   i_mutex > dqonoff_sem > iprune_sem > journal_lock > dqptr_sem >
- *   > dquot->dq_lock > dqio_sem
- * i_mutex on quota files is special (it's below dqio_sem)
+ *   i_mutex > dqonoff_sem > journal_lock > dqptr_sem > dquot->dq_lock >
+ *   dqio_mutex
+ * i_mutex on quota files is special (it's below dqio_mutex)
  */
 
 static DEFINE_SPINLOCK(dq_list_lock);
@@ -281,8 +281,8 @@ static inline void remove_inuse(struct dquot *dquot)
 
 static void wait_on_dquot(struct dquot *dquot)
 {
-	down(&dquot->dq_lock);
-	up(&dquot->dq_lock);
+	mutex_lock(&dquot->dq_lock);
+	mutex_unlock(&dquot->dq_lock);
 }
 
 #define mark_dquot_dirty(dquot) ((dquot)->dq_sb->dq_op->mark_dirty(dquot))
@@ -321,8 +321,8 @@ int dquot_acquire(struct dquot *dquot)
 	int ret = 0, ret2 = 0;
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 
-	down(&dquot->dq_lock);
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dquot->dq_lock);
+	mutex_lock(&dqopt->dqio_mutex);
 	if (!test_bit(DQ_READ_B, &dquot->dq_flags))
 		ret = dqopt->ops[dquot->dq_type]->read_dqblk(dquot);
 	if (ret < 0)
@@ -343,8 +343,8 @@ int dquot_acquire(struct dquot *dquot)
 	}
 	set_bit(DQ_ACTIVE_B, &dquot->dq_flags);
 out_iolock:
-	up(&dqopt->dqio_sem);
-	up(&dquot->dq_lock);
+	mutex_unlock(&dqopt->dqio_mutex);
+	mutex_unlock(&dquot->dq_lock);
 	return ret;
 }
 
@@ -356,7 +356,7 @@ int dquot_commit(struct dquot *dquot)
 	int ret = 0, ret2 = 0;
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	spin_lock(&dq_list_lock);
 	if (!clear_dquot_dirty(dquot)) {
 		spin_unlock(&dq_list_lock);
@@ -373,7 +373,7 @@ int dquot_commit(struct dquot *dquot)
 			ret = ret2;
 	}
 out_sem:
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 	return ret;
 }
 
@@ -385,11 +385,11 @@ int dquot_release(struct dquot *dquot)
 	int ret = 0, ret2 = 0;
 	struct quota_info *dqopt = sb_dqopt(dquot->dq_sb);
 
-	down(&dquot->dq_lock);
+	mutex_lock(&dquot->dq_lock);
 	/* Check whether we are not racing with some other dqget() */
 	if (atomic_read(&dquot->dq_count) > 1)
 		goto out_dqlock;
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	if (dqopt->ops[dquot->dq_type]->release_dqblk) {
 		ret = dqopt->ops[dquot->dq_type]->release_dqblk(dquot);
 		/* Write the info */
@@ -399,31 +399,57 @@ int dquot_release(struct dquot *dquot)
 			ret = ret2;
 	}
 	clear_bit(DQ_ACTIVE_B, &dquot->dq_flags);
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 out_dqlock:
-	up(&dquot->dq_lock);
+	mutex_unlock(&dquot->dq_lock);
 	return ret;
 }
 
 /* Invalidate all dquots on the list. Note that this function is called after
  * quota is disabled and pointers from inodes removed so there cannot be new
- * quota users. Also because we hold dqonoff_sem there can be no quota users
- * for this sb+type at all. */
+ * quota users. There can still be some users of quotas due to inodes being
+ * just deleted or pruned by prune_icache() (those are not attached to any
+ * list). We have to wait for such users.
+ */
 static void invalidate_dquots(struct super_block *sb, int type)
 {
 	struct dquot *dquot, *tmp;
 
+restart:
 	spin_lock(&dq_list_lock);
 	list_for_each_entry_safe(dquot, tmp, &inuse_list, dq_inuse) {
 		if (dquot->dq_sb != sb)
 			continue;
 		if (dquot->dq_type != type)
 			continue;
-#ifdef __DQUOT_PARANOIA
-		if (atomic_read(&dquot->dq_count))
-			BUG();
-#endif
-		/* Quota now has no users and it has been written on last dqput() */
+		/* Wait for dquot users */
+		if (atomic_read(&dquot->dq_count)) {
+			DEFINE_WAIT(wait);
+
+			atomic_inc(&dquot->dq_count);
+			prepare_to_wait(&dquot->dq_wait_unused, &wait,
+					TASK_UNINTERRUPTIBLE);
+			spin_unlock(&dq_list_lock);
+			/* Once dqput() wakes us up, we know it's time to free
+			 * the dquot.
+			 * IMPORTANT: we rely on the fact that there is always
+			 * at most one process waiting for dquot to free.
+			 * Otherwise dq_count would be > 1 and we would never
+			 * wake up.
+			 */
+			if (atomic_read(&dquot->dq_count) > 1)
+				schedule();
+			finish_wait(&dquot->dq_wait_unused, &wait);
+			dqput(dquot);
+			/* At this moment dquot() need not exist (it could be
+			 * reclaimed by prune_dqcache(). Hence we must
+			 * restart. */
+			goto restart;
+		}
+		/*
+		 * Quota now has no users and it has been written on last
+		 * dqput()
+		 */
 		remove_dquot_hash(dquot);
 		remove_free_dquot(dquot);
 		remove_inuse(dquot);
@@ -439,7 +465,7 @@ int vfs_quota_sync(struct super_block *sb, int type)
 	struct quota_info *dqopt = sb_dqopt(sb);
 	int cnt;
 
-	down(&dqopt->dqonoff_sem);
+	mutex_lock(&dqopt->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (type != -1 && cnt != type)
 			continue;
@@ -474,7 +500,7 @@ int vfs_quota_sync(struct super_block *sb, int type)
 	spin_lock(&dq_list_lock);
 	dqstats.syncs++;
 	spin_unlock(&dq_list_lock);
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 
 	return 0;
 }
@@ -515,7 +541,7 @@ static int shrink_dqcache_memory(int nr, gfp_t gfp_mask)
 /*
  * Put reference to dquot
  * NOTE: If you change this function please check whether dqput_blocks() works right...
- * MUST be called with either dqptr_sem or dqonoff_sem held
+ * MUST be called with either dqptr_sem or dqonoff_mutex held
  */
 static void dqput(struct dquot *dquot)
 {
@@ -540,6 +566,10 @@ we_slept:
 	if (atomic_read(&dquot->dq_count) > 1) {
 		/* We have more than one user... nothing to do */
 		atomic_dec(&dquot->dq_count);
+		/* Releasing dquot during quotaoff phase? */
+		if (!sb_has_quota_enabled(dquot->dq_sb, dquot->dq_type) &&
+		    atomic_read(&dquot->dq_count) == 1)
+			wake_up(&dquot->dq_wait_unused);
 		spin_unlock(&dq_list_lock);
 		return;
 	}
@@ -576,11 +606,12 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type)
 		return NODQUOT;
 
 	memset((caddr_t)dquot, 0, sizeof(struct dquot));
-	sema_init(&dquot->dq_lock, 1);
+	mutex_init(&dquot->dq_lock);
 	INIT_LIST_HEAD(&dquot->dq_free);
 	INIT_LIST_HEAD(&dquot->dq_inuse);
 	INIT_HLIST_NODE(&dquot->dq_hash);
 	INIT_LIST_HEAD(&dquot->dq_dirty);
+	init_waitqueue_head(&dquot->dq_wait_unused);
 	dquot->dq_sb = sb;
 	dquot->dq_type = type;
 	atomic_set(&dquot->dq_count, 1);
@@ -590,7 +621,7 @@ static struct dquot *get_empty_dquot(struct super_block *sb, int type)
 
 /*
  * Get reference to dquot
- * MUST be called with either dqptr_sem or dqonoff_sem held
+ * MUST be called with either dqptr_sem or dqonoff_mutex held
  */
 static struct dquot *dqget(struct super_block *sb, unsigned int id, int type)
 {
@@ -656,7 +687,7 @@ static int dqinit_needed(struct inode *inode, int type)
 	return 0;
 }
 
-/* This routine is guarded by dqonoff_sem semaphore */
+/* This routine is guarded by dqonoff_mutex mutex */
 static void add_dquot_ref(struct super_block *sb, int type)
 {
 	struct list_head *p;
@@ -732,13 +763,9 @@ static void drop_dquot_ref(struct super_block *sb, int type)
 {
 	LIST_HEAD(tofree_head);
 
-	/* We need to be guarded against prune_icache to reach all the
-	 * inodes - otherwise some can be on the local list of prune_icache */
-	down(&iprune_sem);
 	down_write(&sb_dqopt(sb)->dqptr_sem);
 	remove_dquot_ref(sb, type, &tofree_head);
 	up_write(&sb_dqopt(sb)->dqptr_sem);
-	up(&iprune_sem);
 	put_dquot_list(&tofree_head);
 }
 
@@ -938,8 +965,8 @@ int dquot_initialize(struct inode *inode, int type)
 	unsigned int id = 0;
 	int cnt, ret = 0;
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return 0;
 	down_write(&sb_dqopt(inode->i_sb)->dqptr_sem);
@@ -1002,8 +1029,8 @@ int dquot_alloc_space(struct inode *inode, qsize_t number, int warn)
 	int cnt, ret = NO_QUOTA;
 	char warntype[MAXQUOTAS];
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode)) {
 out_add:
 		inode_add_bytes(inode, number);
@@ -1051,8 +1078,8 @@ int dquot_alloc_inode(const struct inode *inode, unsigned long number)
 	int cnt, ret = NO_QUOTA;
 	char warntype[MAXQUOTAS];
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return QUOTA_OK;
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
@@ -1095,8 +1122,8 @@ int dquot_free_space(struct inode *inode, qsize_t number)
 {
 	unsigned int cnt;
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode)) {
 out_sub:
 		inode_sub_bytes(inode, number);
@@ -1131,8 +1158,8 @@ int dquot_free_inode(const struct inode *inode, unsigned long number)
 {
 	unsigned int cnt;
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return QUOTA_OK;
 	down_read(&sb_dqopt(inode->i_sb)->dqptr_sem);
@@ -1171,8 +1198,8 @@ int dquot_transfer(struct inode *inode, struct iattr *iattr)
 	    chgid = (iattr->ia_valid & ATTR_GID) && inode->i_gid != iattr->ia_gid;
 	char warntype[MAXQUOTAS];
 
-	/* First test before acquiring semaphore - solves deadlocks when we
-         * re-enter the quota code and are already holding the semaphore */
+	/* First test before acquiring mutex - solves deadlocks when we
+         * re-enter the quota code and are already holding the mutex */
 	if (IS_NOQUOTA(inode))
 		return QUOTA_OK;
 	/* Clear the arrays */
@@ -1266,9 +1293,9 @@ int dquot_commit_info(struct super_block *sb, int type)
 	int ret;
 	struct quota_info *dqopt = sb_dqopt(sb);
 
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	ret = dqopt->ops[type]->write_file_info(sb, type);
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 	return ret;
 }
 
@@ -1324,7 +1351,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 	struct inode *toputinode[MAXQUOTAS];
 
 	/* We need to serialize quota_off() for device */
-	down(&dqopt->dqonoff_sem);
+	mutex_lock(&dqopt->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		toputinode[cnt] = NULL;
 		if (type != -1 && cnt != type)
@@ -1353,7 +1380,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 		dqopt->info[cnt].dqi_bgrace = 0;
 		dqopt->ops[cnt] = NULL;
 	}
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 	/* Sync the superblock so that buffers with quota data are written to
 	 * disk (and so userspace sees correct data afterwards). */
 	if (sb->s_op->sync_fs)
@@ -1366,7 +1393,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 	 * changes done by userspace on the next quotaon() */
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++)
 		if (toputinode[cnt]) {
-			down(&dqopt->dqonoff_sem);
+			mutex_lock(&dqopt->dqonoff_mutex);
 			/* If quota was reenabled in the meantime, we have
 			 * nothing to do */
 			if (!sb_has_quota_enabled(sb, cnt)) {
@@ -1378,7 +1405,7 @@ int vfs_quota_off(struct super_block *sb, int type)
 				mark_inode_dirty(toputinode[cnt]);
 				iput(toputinode[cnt]);
 			}
-			up(&dqopt->dqonoff_sem);
+			mutex_unlock(&dqopt->dqonoff_mutex);
 		}
 	if (sb->s_bdev)
 		invalidate_bdev(sb->s_bdev, 0);
@@ -1419,7 +1446,7 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
 	/* And now flush the block cache so that kernel sees the changes */
 	invalidate_bdev(sb->s_bdev, 0);
 	mutex_lock(&inode->i_mutex);
-	down(&dqopt->dqonoff_sem);
+	mutex_lock(&dqopt->dqonoff_mutex);
 	if (sb_has_quota_enabled(sb, type)) {
 		error = -EBUSY;
 		goto out_lock;
@@ -1444,17 +1471,17 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id)
 	dqopt->ops[type] = fmt->qf_ops;
 	dqopt->info[type].dqi_format = fmt;
 	INIT_LIST_HEAD(&dqopt->info[type].dqi_dirty_list);
-	down(&dqopt->dqio_sem);
+	mutex_lock(&dqopt->dqio_mutex);
 	if ((error = dqopt->ops[type]->read_file_info(sb, type)) < 0) {
-		up(&dqopt->dqio_sem);
+		mutex_unlock(&dqopt->dqio_mutex);
 		goto out_file_init;
 	}
-	up(&dqopt->dqio_sem);
+	mutex_unlock(&dqopt->dqio_mutex);
 	mutex_unlock(&inode->i_mutex);
 	set_enable_flags(dqopt, type);
 
 	add_dquot_ref(sb, type);
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 
 	return 0;
 
@@ -1462,7 +1489,7 @@ out_file_init:
 	dqopt->files[type] = NULL;
 	iput(inode);
 out_lock:
-	up(&dqopt->dqonoff_sem);
+	mutex_unlock(&dqopt->dqonoff_mutex);
 	if (oldflags != -1) {
 		down_write(&dqopt->dqptr_sem);
 		/* Set the flags back (in the case of accidental quotaon()
@@ -1550,14 +1577,14 @@ int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d
 {
 	struct dquot *dquot;
 
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!(dquot = dqget(sb, id, type))) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	do_get_dqblk(dquot, di);
 	dqput(dquot);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1619,14 +1646,14 @@ int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d
 {
 	struct dquot *dquot;
 
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!(dquot = dqget(sb, id, type))) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	do_set_dqblk(dquot, di);
 	dqput(dquot);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1635,9 +1662,9 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
   
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!sb_has_quota_enabled(sb, type)) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	mi = sb_dqopt(sb)->info + type;
@@ -1647,7 +1674,7 @@ int vfs_get_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	ii->dqi_flags = mi->dqi_flags & DQF_MASK;
 	ii->dqi_valid = IIF_ALL;
 	spin_unlock(&dq_data_lock);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1656,9 +1683,9 @@ int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 {
 	struct mem_dqinfo *mi;
 
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	if (!sb_has_quota_enabled(sb, type)) {
-		up(&sb_dqopt(sb)->dqonoff_sem);
+		mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 		return -ESRCH;
 	}
 	mi = sb_dqopt(sb)->info + type;
@@ -1673,7 +1700,7 @@ int vfs_set_dqinfo(struct super_block *sb, int type, struct if_dqinfo *ii)
 	mark_info_dirty(sb, type);
 	/* Force write to disk */
 	sb->dq_op->write_info(sb, type);
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	return 0;
 }
 
@@ -1794,7 +1821,8 @@ static int __init dquot_init(void)
 
 	dquot_cachep = kmem_cache_create("dquot", 
 			sizeof(struct dquot), sizeof(unsigned long) * 4,
-			SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|SLAB_PANIC,
+			(SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
+				SLAB_MEM_SPREAD|SLAB_PANIC),
 			NULL, NULL);
 
 	order = 0;
diff --git a/fs/efs/dir.c b/fs/efs/dir.c
index 777c614ff360..17f5b2d3c16a 100644
--- a/fs/efs/dir.c
+++ b/fs/efs/dir.c
@@ -10,7 +10,7 @@
 
 static int efs_readdir(struct file *, void *, filldir_t);
 
-struct file_operations efs_dir_operations = {
+const struct file_operations efs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= efs_readdir,
 };
diff --git a/fs/efs/super.c b/fs/efs/super.c
index afc4891feb36..dff623e3ddbf 100644
--- a/fs/efs/super.c
+++ b/fs/efs/super.c
@@ -81,7 +81,7 @@ static int init_inodecache(void)
 {
 	efs_inode_cachep = kmem_cache_create("efs_inode_cache",
 				sizeof(struct efs_inode_info),
-				0, SLAB_RECLAIM_ACCOUNT,
+				0, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
 				init_once, NULL);
 	if (efs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 4284cd31eba6..242fe1a66ce5 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -34,6 +34,7 @@
 #include <linux/eventpoll.h>
 #include <linux/mount.h>
 #include <linux/bitops.h>
+#include <linux/mutex.h>
 #include <asm/uaccess.h>
 #include <asm/system.h>
 #include <asm/io.h>
@@ -46,7 +47,7 @@
  * LOCKING:
  * There are three level of locking required by epoll :
  *
- * 1) epsem (semaphore)
+ * 1) epmutex (mutex)
  * 2) ep->sem (rw_semaphore)
  * 3) ep->lock (rw_lock)
  *
@@ -67,9 +68,9 @@
  * if a file has been pushed inside an epoll set and it is then
  * close()d without a previous call toepoll_ctl(EPOLL_CTL_DEL).
  * It is possible to drop the "ep->sem" and to use the global
- * semaphore "epsem" (together with "ep->lock") to have it working,
+ * semaphore "epmutex" (together with "ep->lock") to have it working,
  * but having "ep->sem" will make the interface more scalable.
- * Events that require holding "epsem" are very rare, while for
+ * Events that require holding "epmutex" are very rare, while for
  * normal operations the epoll private "ep->sem" will guarantee
  * a greater scalability.
  */
@@ -274,22 +275,22 @@ static struct super_block *eventpollfs_get_sb(struct file_system_type *fs_type,
 /*
  * This semaphore is used to serialize ep_free() and eventpoll_release_file().
  */
-static struct semaphore epsem;
+static struct mutex epmutex;
 
 /* Safe wake up implementation */
 static struct poll_safewake psw;
 
 /* Slab cache used to allocate "struct epitem" */
-static kmem_cache_t *epi_cache;
+static kmem_cache_t *epi_cache __read_mostly;
 
 /* Slab cache used to allocate "struct eppoll_entry" */
-static kmem_cache_t *pwq_cache;
+static kmem_cache_t *pwq_cache __read_mostly;
 
 /* Virtual fs used to allocate inodes for eventpoll files */
-static struct vfsmount *eventpoll_mnt;
+static struct vfsmount *eventpoll_mnt __read_mostly;
 
 /* File callbacks that implement the eventpoll file behaviour */
-static struct file_operations eventpoll_fops = {
+static const struct file_operations eventpoll_fops = {
 	.release	= ep_eventpoll_close,
 	.poll		= ep_eventpoll_poll
 };
@@ -451,15 +452,6 @@ static void ep_poll_safewake(struct poll_safewake *psw, wait_queue_head_t *wq)
 }
 
 
-/* Used to initialize the epoll bits inside the "struct file" */
-void eventpoll_init_file(struct file *file)
-{
-
-	INIT_LIST_HEAD(&file->f_ep_links);
-	spin_lock_init(&file->f_ep_lock);
-}
-
-
 /*
  * This is called from eventpoll_release() to unlink files from the eventpoll
  * interface. We need to have this facility to cleanup correctly files that are
@@ -477,10 +469,10 @@ void eventpoll_release_file(struct file *file)
 	 * cleanup path, and this means that noone is using this file anymore.
 	 * The only hit might come from ep_free() but by holding the semaphore
 	 * will correctly serialize the operation. We do need to acquire
-	 * "ep->sem" after "epsem" because ep_remove() requires it when called
+	 * "ep->sem" after "epmutex" because ep_remove() requires it when called
 	 * from anywhere but ep_free().
 	 */
-	down(&epsem);
+	mutex_lock(&epmutex);
 
 	while (!list_empty(lsthead)) {
 		epi = list_entry(lsthead->next, struct epitem, fllink);
@@ -492,7 +484,7 @@ void eventpoll_release_file(struct file *file)
 		up_write(&ep->sem);
 	}
 
-	up(&epsem);
+	mutex_unlock(&epmutex);
 }
 
 
@@ -607,7 +599,7 @@ sys_epoll_ctl(int epfd, int op, int fd, struct epoll_event __user *event)
 	switch (op) {
 	case EPOLL_CTL_ADD:
 		if (!epi) {
-			epds.events |= POLLERR | POLLHUP;
+			epds.events |= POLLERR | POLLHUP | POLLRDHUP;
 
 			error = ep_insert(ep, &epds, tfile, fd);
 		} else
@@ -621,7 +613,7 @@ sys_epoll_ctl(int epfd, int op, int fd, struct epoll_event __user *event)
 		break;
 	case EPOLL_CTL_MOD:
 		if (epi) {
-			epds.events |= POLLERR | POLLHUP;
+			epds.events |= POLLERR | POLLHUP | POLLRDHUP;
 			error = ep_modify(ep, epi, &epds);
 		} else
 			error = -ENOENT;
@@ -819,9 +811,9 @@ static void ep_free(struct eventpoll *ep)
 	 * We do not need to hold "ep->sem" here because the epoll file
 	 * is on the way to be removed and no one has references to it
 	 * anymore. The only hit might come from eventpoll_release_file() but
-	 * holding "epsem" is sufficent here.
+	 * holding "epmutex" is sufficent here.
 	 */
-	down(&epsem);
+	mutex_lock(&epmutex);
 
 	/*
 	 * Walks through the whole tree by unregistering poll callbacks.
@@ -843,7 +835,7 @@ static void ep_free(struct eventpoll *ep)
 		ep_remove(ep, epi);
 	}
 
-	up(&epsem);
+	mutex_unlock(&epmutex);
 }
 
 
@@ -1615,7 +1607,7 @@ static int __init eventpoll_init(void)
 {
 	int error;
 
-	init_MUTEX(&epsem);
+	mutex_init(&epmutex);
 
 	/* Initialize the structure used to perform safe poll wait head wake ups */
 	ep_poll_safewake_init(&psw);
diff --git a/fs/exec.c b/fs/exec.c
index 0b515ac53134..950ebd43cdc3 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -127,7 +127,7 @@ asmlinkage long sys_uselib(const char __user * library)
 	struct nameidata nd;
 	int error;
 
-	error = __user_path_lookup_open(library, LOOKUP_FOLLOW, &nd, FMODE_READ);
+	error = __user_path_lookup_open(library, LOOKUP_FOLLOW, &nd, FMODE_READ|FMODE_EXEC);
 	if (error)
 		goto out;
 
@@ -477,7 +477,7 @@ struct file *open_exec(const char *name)
 	int err;
 	struct file *file;
 
-	err = path_lookup_open(AT_FDCWD, name, LOOKUP_FOLLOW, &nd, FMODE_READ);
+	err = path_lookup_open(AT_FDCWD, name, LOOKUP_FOLLOW, &nd, FMODE_READ|FMODE_EXEC);
 	file = ERR_PTR(err);
 
 	if (!err) {
@@ -616,6 +616,15 @@ static int de_thread(struct task_struct *tsk)
 		kmem_cache_free(sighand_cachep, newsighand);
 		return -EAGAIN;
 	}
+
+	/*
+	 * child_reaper ignores SIGKILL, change it now.
+	 * Reparenting needs write_lock on tasklist_lock,
+	 * so it is safe to do it under read_lock.
+	 */
+	if (unlikely(current->group_leader == child_reaper))
+		child_reaper = current;
+
 	zap_other_threads(current);
 	read_unlock(&tasklist_lock);
 
@@ -632,7 +641,7 @@ static int de_thread(struct task_struct *tsk)
 		 * synchronize with any firing (by calling del_timer_sync)
 		 * before we can safely let the old group leader die.
 		 */
-		sig->real_timer.data = current;
+		sig->tsk = current;
 		spin_unlock_irq(lock);
 		if (hrtimer_cancel(&sig->real_timer))
 			hrtimer_restart(&sig->real_timer);
@@ -699,22 +708,30 @@ static int de_thread(struct task_struct *tsk)
 		remove_parent(current);
 		remove_parent(leader);
 
-		switch_exec_pids(leader, current);
+
+		/* Become a process group leader with the old leader's pid.
+		 * Note: The old leader also uses thispid until release_task
+		 *       is called.  Odd but simple and correct.
+		 */
+		detach_pid(current, PIDTYPE_PID);
+		current->pid = leader->pid;
+		attach_pid(current, PIDTYPE_PID,  current->pid);
+		attach_pid(current, PIDTYPE_PGID, current->signal->pgrp);
+		attach_pid(current, PIDTYPE_SID,  current->signal->session);
+		list_add_tail(&current->tasks, &init_task.tasks);
 
 		current->parent = current->real_parent = leader->real_parent;
 		leader->parent = leader->real_parent = child_reaper;
 		current->group_leader = current;
 		leader->group_leader = leader;
 
-		add_parent(current, current->parent);
-		add_parent(leader, leader->parent);
+		add_parent(current);
+		add_parent(leader);
 		if (ptrace) {
 			current->ptrace = ptrace;
 			__ptrace_link(current, parent);
 		}
 
-		list_del(&current->tasks);
-		list_add_tail(&current->tasks, &init_task.tasks);
 		current->exit_signal = SIGCHLD;
 
 		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
@@ -751,7 +768,6 @@ no_thread_group:
 		/*
 		 * Move our state over to newsighand and switch it in.
 		 */
-		spin_lock_init(&newsighand->siglock);
 		atomic_set(&newsighand->count, 1);
 		memcpy(newsighand->action, oldsighand->action,
 		       sizeof(newsighand->action));
@@ -768,7 +784,7 @@ no_thread_group:
 		write_unlock_irq(&tasklist_lock);
 
 		if (atomic_dec_and_test(&oldsighand->count))
-			sighand_free(oldsighand);
+			kmem_cache_free(sighand_cachep, oldsighand);
 	}
 
 	BUG_ON(!thread_group_leader(current));
@@ -1143,10 +1159,9 @@ int do_execve(char * filename,
 	int i;
 
 	retval = -ENOMEM;
-	bprm = kmalloc(sizeof(*bprm), GFP_KERNEL);
+	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
 	if (!bprm)
 		goto out_ret;
-	memset(bprm, 0, sizeof(*bprm));
 
 	file = open_exec(filename);
 	retval = PTR_ERR(file);
diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c
index b3dbd716cd3a..d672aa9f4061 100644
--- a/fs/ext2/dir.c
+++ b/fs/ext2/dir.c
@@ -416,8 +416,7 @@ void ext2_set_link(struct inode *dir, struct ext2_dir_entry_2 *de,
 
 	lock_page(page);
 	err = page->mapping->a_ops->prepare_write(NULL, page, from, to);
-	if (err)
-		BUG();
+	BUG_ON(err);
 	de->inode = cpu_to_le32(inode->i_ino);
 	ext2_set_de_type (de, inode);
 	err = ext2_commit_chunk(page, from, to);
@@ -554,8 +553,7 @@ int ext2_delete_entry (struct ext2_dir_entry_2 * dir, struct page * page )
 		from = (char*)pde - (char*)page_address(page);
 	lock_page(page);
 	err = mapping->a_ops->prepare_write(NULL, page, from, to);
-	if (err)
-		BUG();
+	BUG_ON(err);
 	if (pde)
 		pde->rec_len = cpu_to_le16(to-from);
 	dir->inode = 0;
@@ -660,7 +658,7 @@ not_empty:
 	return 0;
 }
 
-struct file_operations ext2_dir_operations = {
+const struct file_operations ext2_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= ext2_readdir,
diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h
index 00de0a7312a2..9f74a62be555 100644
--- a/fs/ext2/ext2.h
+++ b/fs/ext2/ext2.h
@@ -138,6 +138,9 @@ extern void ext2_set_inode_flags(struct inode *inode);
 extern int ext2_ioctl (struct inode *, struct file *, unsigned int,
 		       unsigned long);
 
+/* namei.c */
+struct dentry *ext2_get_parent(struct dentry *child);
+
 /* super.c */
 extern void ext2_error (struct super_block *, const char *, const char *, ...)
 	__attribute__ ((format (printf, 3, 4)));
@@ -151,12 +154,12 @@ extern void ext2_write_super (struct super_block *);
  */
 
 /* dir.c */
-extern struct file_operations ext2_dir_operations;
+extern const struct file_operations ext2_dir_operations;
 
 /* file.c */
 extern struct inode_operations ext2_file_inode_operations;
-extern struct file_operations ext2_file_operations;
-extern struct file_operations ext2_xip_file_operations;
+extern const struct file_operations ext2_file_operations;
+extern const struct file_operations ext2_xip_file_operations;
 
 /* inode.c */
 extern struct address_space_operations ext2_aops;
diff --git a/fs/ext2/file.c b/fs/ext2/file.c
index a484412fc782..23e2c7ccec1d 100644
--- a/fs/ext2/file.c
+++ b/fs/ext2/file.c
@@ -39,7 +39,7 @@ static int ext2_release_file (struct inode * inode, struct file * filp)
  * We have mostly NULL's here: the current defaults are ok for
  * the ext2 filesystem.
  */
-struct file_operations ext2_file_operations = {
+const struct file_operations ext2_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
@@ -53,10 +53,12 @@ struct file_operations ext2_file_operations = {
 	.readv		= generic_file_readv,
 	.writev		= generic_file_writev,
 	.sendfile	= generic_file_sendfile,
+	.splice_read	= generic_file_splice_read,
+	.splice_write	= generic_file_splice_write,
 };
 
 #ifdef CONFIG_EXT2_FS_XIP
-struct file_operations ext2_xip_file_operations = {
+const struct file_operations ext2_xip_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= xip_file_read,
 	.write		= xip_file_write,
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c
index a717837f272e..04af9c45dce2 100644
--- a/fs/ext2/inode.c
+++ b/fs/ext2/inode.c
@@ -667,18 +667,6 @@ static sector_t ext2_bmap(struct address_space *mapping, sector_t block)
 	return generic_block_bmap(mapping,block,ext2_get_block);
 }
 
-static int
-ext2_get_blocks(struct inode *inode, sector_t iblock, unsigned long max_blocks,
-			struct buffer_head *bh_result, int create)
-{
-	int ret;
-
-	ret = ext2_get_block(inode, iblock, bh_result, create);
-	if (ret == 0)
-		bh_result->b_size = (1 << inode->i_blkbits);
-	return ret;
-}
-
 static ssize_t
 ext2_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 			loff_t offset, unsigned long nr_segs)
@@ -687,7 +675,7 @@ ext2_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
 	struct inode *inode = file->f_mapping->host;
 
 	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
-				offset, nr_segs, ext2_get_blocks, NULL);
+				offset, nr_segs, ext2_get_block, NULL);
 }
 
 static int
diff --git a/fs/ext2/namei.c b/fs/ext2/namei.c
index ad1432a2a62e..4ca824985321 100644
--- a/fs/ext2/namei.c
+++ b/fs/ext2/namei.c
@@ -36,22 +36,6 @@
 #include "acl.h"
 #include "xip.h"
 
-/*
- * Couple of helper functions - make the code slightly cleaner.
- */
-
-static inline void ext2_inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void ext2_dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static inline int ext2_add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = ext2_add_link(dentry, inode);
@@ -59,7 +43,7 @@ static inline int ext2_add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -201,7 +185,7 @@ out:
 	return err;
 
 out_fail:
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput (inode);
 	goto out;
 }
@@ -215,7 +199,7 @@ static int ext2_link (struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	ext2_inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
 	return ext2_add_nondir(dentry, inode);
@@ -229,7 +213,7 @@ static int ext2_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	if (dir->i_nlink >= EXT2_LINK_MAX)
 		goto out;
 
-	ext2_inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = ext2_new_inode (dir, S_IFDIR | mode);
 	err = PTR_ERR(inode);
@@ -243,7 +227,7 @@ static int ext2_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	else
 		inode->i_mapping->a_ops = &ext2_aops;
 
-	ext2_inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = ext2_make_empty(inode, dir);
 	if (err)
@@ -258,11 +242,11 @@ out:
 	return err;
 
 out_fail:
-	ext2_dec_count(inode);
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 out_dir:
-	ext2_dec_count(dir);
+	inode_dec_link_count(dir);
 	goto out;
 }
 
@@ -282,7 +266,7 @@ static int ext2_unlink(struct inode * dir, struct dentry *dentry)
 		goto out;
 
 	inode->i_ctime = dir->i_ctime;
-	ext2_dec_count(inode);
+	inode_dec_link_count(inode);
 	err = 0;
 out:
 	return err;
@@ -297,8 +281,8 @@ static int ext2_rmdir (struct inode * dir, struct dentry *dentry)
 		err = ext2_unlink(dir, dentry);
 		if (!err) {
 			inode->i_size = 0;
-			ext2_dec_count(inode);
-			ext2_dec_count(dir);
+			inode_dec_link_count(inode);
+			inode_dec_link_count(dir);
 		}
 	}
 	return err;
@@ -338,41 +322,41 @@ static int ext2_rename (struct inode * old_dir, struct dentry * old_dentry,
 		new_de = ext2_find_entry (new_dir, new_dentry, &new_page);
 		if (!new_de)
 			goto out_dir;
-		ext2_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		ext2_set_link(new_dir, new_de, new_page, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		ext2_dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= EXT2_LINK_MAX)
 				goto out_dir;
 		}
-		ext2_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = ext2_add_link(new_dentry, old_inode);
 		if (err) {
-			ext2_dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			ext2_inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	/*
 	 * Like most other Unix systems, set the ctime for inodes on a
  	 * rename.
-	 * ext2_dec_count() will mark the inode dirty.
+	 * inode_dec_link_count() will mark the inode dirty.
 	 */
 	old_inode->i_ctime = CURRENT_TIME_SEC;
 
 	ext2_delete_entry (old_de, old_page);
-	ext2_dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		ext2_set_link(old_inode, dir_de, dir_page, new_dir);
-		ext2_dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	return 0;
 
diff --git a/fs/ext2/super.c b/fs/ext2/super.c
index cb6f9bd658de..7e30bae174ed 100644
--- a/fs/ext2/super.c
+++ b/fs/ext2/super.c
@@ -175,7 +175,8 @@ static int init_inodecache(void)
 {
 	ext2_inode_cachep = kmem_cache_create("ext2_inode_cache",
 					     sizeof(struct ext2_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (ext2_inode_cachep == NULL)
 		return -ENOMEM;
@@ -210,8 +211,6 @@ static int ext2_show_options(struct seq_file *seq, struct vfsmount *vfs)
 
 	if (sbi->s_mount_opt & EXT2_MOUNT_GRPID)
 		seq_puts(seq, ",grpid");
-	else
-		seq_puts(seq, ",nogrpid");
 
 #if defined(CONFIG_QUOTA)
 	if (sbi->s_mount_opt & EXT2_MOUNT_USRQUOTA)
@@ -258,7 +257,6 @@ static struct super_operations ext2_sops = {
  * systems, but can be improved upon.
  * Currently only get_parent is required.
  */
-struct dentry *ext2_get_parent(struct dentry *child);
 static struct export_operations ext2_export_ops = {
 	.get_parent = ext2_get_parent,
 };
diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c
index 6250fcdf14a1..77927d6938f6 100644
--- a/fs/ext3/balloc.c
+++ b/fs/ext3/balloc.c
@@ -653,9 +653,11 @@ claim_block(spinlock_t *lock, int block, struct buffer_head *bh)
  */
 static int
 ext3_try_to_allocate(struct super_block *sb, handle_t *handle, int group,
-	struct buffer_head *bitmap_bh, int goal, struct ext3_reserve_window *my_rsv)
+			struct buffer_head *bitmap_bh, int goal,
+			unsigned long *count, struct ext3_reserve_window *my_rsv)
 {
 	int group_first_block, start, end;
+	unsigned long num = 0;
 
 	/* we do allocation within the reservation window if we have a window */
 	if (my_rsv) {
@@ -713,8 +715,18 @@ repeat:
 			goto fail_access;
 		goto repeat;
 	}
-	return goal;
+	num++;
+	goal++;
+	while (num < *count && goal < end
+		&& ext3_test_allocatable(goal, bitmap_bh)
+		&& claim_block(sb_bgl_lock(EXT3_SB(sb), group), goal, bitmap_bh)) {
+		num++;
+		goal++;
+	}
+	*count = num;
+	return goal - num;
 fail_access:
+	*count = num;
 	return -1;
 }
 
@@ -999,6 +1011,31 @@ retry:
 	goto retry;
 }
 
+static void try_to_extend_reservation(struct ext3_reserve_window_node *my_rsv,
+			struct super_block *sb, int size)
+{
+	struct ext3_reserve_window_node *next_rsv;
+	struct rb_node *next;
+	spinlock_t *rsv_lock = &EXT3_SB(sb)->s_rsv_window_lock;
+
+	if (!spin_trylock(rsv_lock))
+		return;
+
+	next = rb_next(&my_rsv->rsv_node);
+
+	if (!next)
+		my_rsv->rsv_end += size;
+	else {
+		next_rsv = list_entry(next, struct ext3_reserve_window_node, rsv_node);
+
+		if ((next_rsv->rsv_start - my_rsv->rsv_end - 1) >= size)
+			my_rsv->rsv_end += size;
+		else
+			my_rsv->rsv_end = next_rsv->rsv_start - 1;
+	}
+	spin_unlock(rsv_lock);
+}
+
 /*
  * This is the main function used to allocate a new block and its reservation
  * window.
@@ -1024,11 +1061,12 @@ static int
 ext3_try_to_allocate_with_rsv(struct super_block *sb, handle_t *handle,
 			unsigned int group, struct buffer_head *bitmap_bh,
 			int goal, struct ext3_reserve_window_node * my_rsv,
-			int *errp)
+			unsigned long *count, int *errp)
 {
 	unsigned long group_first_block;
 	int ret = 0;
 	int fatal;
+	unsigned long num = *count;
 
 	*errp = 0;
 
@@ -1051,7 +1089,8 @@ ext3_try_to_allocate_with_rsv(struct super_block *sb, handle_t *handle,
 	 * or last attempt to allocate a block with reservation turned on failed
 	 */
 	if (my_rsv == NULL ) {
-		ret = ext3_try_to_allocate(sb, handle, group, bitmap_bh, goal, NULL);
+		ret = ext3_try_to_allocate(sb, handle, group, bitmap_bh,
+						goal, count, NULL);
 		goto out;
 	}
 	/*
@@ -1081,6 +1120,8 @@ ext3_try_to_allocate_with_rsv(struct super_block *sb, handle_t *handle,
 	while (1) {
 		if (rsv_is_empty(&my_rsv->rsv_window) || (ret < 0) ||
 			!goal_in_my_reservation(&my_rsv->rsv_window, goal, group, sb)) {
+			if (my_rsv->rsv_goal_size < *count)
+				my_rsv->rsv_goal_size = *count;
 			ret = alloc_new_reservation(my_rsv, goal, sb,
 							group, bitmap_bh);
 			if (ret < 0)
@@ -1088,16 +1129,21 @@ ext3_try_to_allocate_with_rsv(struct super_block *sb, handle_t *handle,
 
 			if (!goal_in_my_reservation(&my_rsv->rsv_window, goal, group, sb))
 				goal = -1;
-		}
+		} else if (goal > 0 && (my_rsv->rsv_end-goal+1) < *count)
+			try_to_extend_reservation(my_rsv, sb,
+					*count-my_rsv->rsv_end + goal - 1);
+
 		if ((my_rsv->rsv_start >= group_first_block + EXT3_BLOCKS_PER_GROUP(sb))
 		    || (my_rsv->rsv_end < group_first_block))
 			BUG();
 		ret = ext3_try_to_allocate(sb, handle, group, bitmap_bh, goal,
-					   &my_rsv->rsv_window);
+					   &num, &my_rsv->rsv_window);
 		if (ret >= 0) {
-			my_rsv->rsv_alloc_hit++;
+			my_rsv->rsv_alloc_hit += num;
+			*count = num;
 			break;				/* succeed */
 		}
+		num = *count;
 	}
 out:
 	if (ret >= 0) {
@@ -1154,8 +1200,8 @@ int ext3_should_retry_alloc(struct super_block *sb, int *retries)
  * bitmap, and then for any free bit if that fails.
  * This function also updates quota and i_blocks field.
  */
-int ext3_new_block(handle_t *handle, struct inode *inode,
-			unsigned long goal, int *errp)
+int ext3_new_blocks(handle_t *handle, struct inode *inode,
+			unsigned long goal, unsigned long *count, int *errp)
 {
 	struct buffer_head *bitmap_bh = NULL;
 	struct buffer_head *gdp_bh;
@@ -1178,6 +1224,7 @@ int ext3_new_block(handle_t *handle, struct inode *inode,
 	static int goal_hits, goal_attempts;
 #endif
 	unsigned long ngroups;
+	unsigned long num = *count;
 
 	*errp = -ENOSPC;
 	sb = inode->i_sb;
@@ -1189,7 +1236,7 @@ int ext3_new_block(handle_t *handle, struct inode *inode,
 	/*
 	 * Check quota for allocation of this block.
 	 */
-	if (DQUOT_ALLOC_BLOCK(inode, 1)) {
+	if (DQUOT_ALLOC_BLOCK(inode, num)) {
 		*errp = -EDQUOT;
 		return 0;
 	}
@@ -1244,7 +1291,7 @@ retry:
 		if (!bitmap_bh)
 			goto io_error;
 		ret_block = ext3_try_to_allocate_with_rsv(sb, handle, group_no,
-					bitmap_bh, ret_block, my_rsv, &fatal);
+					bitmap_bh, ret_block, my_rsv, &num, &fatal);
 		if (fatal)
 			goto out;
 		if (ret_block >= 0)
@@ -1281,7 +1328,7 @@ retry:
 		if (!bitmap_bh)
 			goto io_error;
 		ret_block = ext3_try_to_allocate_with_rsv(sb, handle, group_no,
-					bitmap_bh, -1, my_rsv, &fatal);
+					bitmap_bh, -1, my_rsv, &num, &fatal);
 		if (fatal)
 			goto out;
 		if (ret_block >= 0) 
@@ -1316,13 +1363,15 @@ allocated:
 	target_block = ret_block + group_no * EXT3_BLOCKS_PER_GROUP(sb)
 				+ le32_to_cpu(es->s_first_data_block);
 
-	if (target_block == le32_to_cpu(gdp->bg_block_bitmap) ||
-	    target_block == le32_to_cpu(gdp->bg_inode_bitmap) ||
+	if (in_range(le32_to_cpu(gdp->bg_block_bitmap), target_block, num) ||
+	    in_range(le32_to_cpu(gdp->bg_inode_bitmap), target_block, num) ||
 	    in_range(target_block, le32_to_cpu(gdp->bg_inode_table),
+		      EXT3_SB(sb)->s_itb_per_group) ||
+	    in_range(target_block + num - 1, le32_to_cpu(gdp->bg_inode_table),
 		      EXT3_SB(sb)->s_itb_per_group))
 		ext3_error(sb, "ext3_new_block",
 			    "Allocating block in system zone - "
-			    "block = %u", target_block);
+			    "blocks from %u, length %lu", target_block, num);
 
 	performed_allocation = 1;
 
@@ -1341,10 +1390,14 @@ allocated:
 	jbd_lock_bh_state(bitmap_bh);
 	spin_lock(sb_bgl_lock(sbi, group_no));
 	if (buffer_jbd(bitmap_bh) && bh2jh(bitmap_bh)->b_committed_data) {
-		if (ext3_test_bit(ret_block,
-				bh2jh(bitmap_bh)->b_committed_data)) {
-			printk("%s: block was unexpectedly set in "
-				"b_committed_data\n", __FUNCTION__);
+		int i;
+
+		for (i = 0; i < num; i++) {
+			if (ext3_test_bit(ret_block,
+					bh2jh(bitmap_bh)->b_committed_data)) {
+				printk("%s: block was unexpectedly set in "
+					"b_committed_data\n", __FUNCTION__);
+			}
 		}
 	}
 	ext3_debug("found bit %d\n", ret_block);
@@ -1355,7 +1408,7 @@ allocated:
 	/* ret_block was blockgroup-relative.  Now it becomes fs-relative */
 	ret_block = target_block;
 
-	if (ret_block >= le32_to_cpu(es->s_blocks_count)) {
+	if (ret_block + num - 1 >= le32_to_cpu(es->s_blocks_count)) {
 		ext3_error(sb, "ext3_new_block",
 			    "block(%d) >= blocks count(%d) - "
 			    "block_group = %d, es == %p ", ret_block,
@@ -1373,9 +1426,9 @@ allocated:
 
 	spin_lock(sb_bgl_lock(sbi, group_no));
 	gdp->bg_free_blocks_count =
-			cpu_to_le16(le16_to_cpu(gdp->bg_free_blocks_count) - 1);
+			cpu_to_le16(le16_to_cpu(gdp->bg_free_blocks_count) - num);
 	spin_unlock(sb_bgl_lock(sbi, group_no));
-	percpu_counter_mod(&sbi->s_freeblocks_counter, -1);
+	percpu_counter_mod(&sbi->s_freeblocks_counter, -num);
 
 	BUFFER_TRACE(gdp_bh, "journal_dirty_metadata for group descriptor");
 	err = ext3_journal_dirty_metadata(handle, gdp_bh);
@@ -1388,6 +1441,8 @@ allocated:
 
 	*errp = 0;
 	brelse(bitmap_bh);
+	DQUOT_FREE_BLOCK(inode, *count-num);
+	*count = num;
 	return ret_block;
 
 io_error:
@@ -1401,11 +1456,19 @@ out:
 	 * Undo the block allocation
 	 */
 	if (!performed_allocation)
-		DQUOT_FREE_BLOCK(inode, 1);
+		DQUOT_FREE_BLOCK(inode, *count);
 	brelse(bitmap_bh);
 	return 0;
 }
 
+int ext3_new_block(handle_t *handle, struct inode *inode,
+			unsigned long goal, int *errp)
+{
+	unsigned long count = 1;
+
+	return ext3_new_blocks(handle, inode, goal, &count, errp);
+}
+
 unsigned long ext3_count_free_blocks(struct super_block *sb)
 {
 	unsigned long desc_count;
@@ -1493,12 +1556,33 @@ static int ext3_group_sparse(int group)
  */
 int ext3_bg_has_super(struct super_block *sb, int group)
 {
-	if (EXT3_HAS_RO_COMPAT_FEATURE(sb,EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER)&&
-	    !ext3_group_sparse(group))
+	if (EXT3_HAS_RO_COMPAT_FEATURE(sb,
+				EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER) &&
+			!ext3_group_sparse(group))
 		return 0;
 	return 1;
 }
 
+static unsigned long ext3_bg_num_gdb_meta(struct super_block *sb, int group)
+{
+	unsigned long metagroup = group / EXT3_DESC_PER_BLOCK(sb);
+	unsigned long first = metagroup * EXT3_DESC_PER_BLOCK(sb);
+	unsigned long last = first + EXT3_DESC_PER_BLOCK(sb) - 1;
+
+	if (group == first || group == first + 1 || group == last)
+		return 1;
+	return 0;
+}
+
+static unsigned long ext3_bg_num_gdb_nometa(struct super_block *sb, int group)
+{
+	if (EXT3_HAS_RO_COMPAT_FEATURE(sb,
+				EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER) &&
+			!ext3_group_sparse(group))
+		return 0;
+	return EXT3_SB(sb)->s_gdb_count;
+}
+
 /**
  *	ext3_bg_num_gdb - number of blocks used by the group table in group
  *	@sb: superblock for filesystem
@@ -1510,9 +1594,14 @@ int ext3_bg_has_super(struct super_block *sb, int group)
  */
 unsigned long ext3_bg_num_gdb(struct super_block *sb, int group)
 {
-	if (EXT3_HAS_RO_COMPAT_FEATURE(sb,EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER)&&
-	    !ext3_group_sparse(group))
-		return 0;
-	return EXT3_SB(sb)->s_gdb_count;
-}
+	unsigned long first_meta_bg =
+			le32_to_cpu(EXT3_SB(sb)->s_es->s_first_meta_bg);
+	unsigned long metagroup = group / EXT3_DESC_PER_BLOCK(sb);
+
+	if (!EXT3_HAS_INCOMPAT_FEATURE(sb,EXT3_FEATURE_INCOMPAT_META_BG) ||
+			metagroup < first_meta_bg)
+		return ext3_bg_num_gdb_nometa(sb,group);
 
+	return ext3_bg_num_gdb_meta(sb,group);
+
+}
diff --git a/fs/ext3/bitmap.c b/fs/ext3/bitmap.c
index cb16b4c5d5df..ce4f82b9e528 100644
--- a/fs/ext3/bitmap.c
+++ b/fs/ext3/bitmap.c
@@ -7,11 +7,11 @@
  * Universite Pierre et Marie Curie (Paris VI)
  */
 
-#ifdef EXT3FS_DEBUG
-
 #include <linux/buffer_head.h>
+#include <linux/jbd.h>
+#include <linux/ext3_fs.h>
 
-#include "ext3_fs.h"
+#ifdef EXT3FS_DEBUG
 
 static int nibblemap[] = {4, 3, 3, 2, 3, 2, 2, 1, 3, 2, 2, 1, 2, 1, 1, 0};
 
diff --git a/fs/ext3/dir.c b/fs/ext3/dir.c
index 832867aef3dc..f37528ed222e 100644
--- a/fs/ext3/dir.c
+++ b/fs/ext3/dir.c
@@ -39,7 +39,7 @@ static int ext3_dx_readdir(struct file * filp,
 static int ext3_release_dir (struct inode * inode,
 				struct file * filp);
 
-struct file_operations ext3_dir_operations = {
+const struct file_operations ext3_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= ext3_readdir,		/* we take BKL. needed?*/
@@ -95,11 +95,10 @@ static int ext3_readdir(struct file * filp,
 			 void * dirent, filldir_t filldir)
 {
 	int error = 0;
-	unsigned long offset, blk;
-	int i, num, stored;
-	struct buffer_head * bh, * tmp, * bha[16];
-	struct ext3_dir_entry_2 * de;
-	struct super_block * sb;
+	unsigned long offset;
+	int i, stored;
+	struct ext3_dir_entry_2 *de;
+	struct super_block *sb;
 	int err;
 	struct inode *inode = filp->f_dentry->d_inode;
 	int ret = 0;
@@ -124,12 +123,30 @@ static int ext3_readdir(struct file * filp,
 	}
 #endif
 	stored = 0;
-	bh = NULL;
 	offset = filp->f_pos & (sb->s_blocksize - 1);
 
 	while (!error && !stored && filp->f_pos < inode->i_size) {
-		blk = (filp->f_pos) >> EXT3_BLOCK_SIZE_BITS(sb);
-		bh = ext3_bread(NULL, inode, blk, 0, &err);
+		unsigned long blk = filp->f_pos >> EXT3_BLOCK_SIZE_BITS(sb);
+		struct buffer_head map_bh;
+		struct buffer_head *bh = NULL;
+
+		map_bh.b_state = 0;
+		err = ext3_get_blocks_handle(NULL, inode, blk, 1,
+						&map_bh, 0, 0);
+		if (err > 0) {
+			page_cache_readahead(sb->s_bdev->bd_inode->i_mapping,
+				&filp->f_ra,
+				filp,
+				map_bh.b_blocknr >>
+					(PAGE_CACHE_SHIFT - inode->i_blkbits),
+				1);
+			bh = ext3_bread(NULL, inode, blk, 0, &err);
+		}
+
+		/*
+		 * We ignore I/O errors on directories so users have a chance
+		 * of recovering data when there's a bad sector
+		 */
 		if (!bh) {
 			ext3_error (sb, "ext3_readdir",
 				"directory #%lu contains a hole at offset %lu",
@@ -138,26 +155,6 @@ static int ext3_readdir(struct file * filp,
 			continue;
 		}
 
-		/*
-		 * Do the readahead
-		 */
-		if (!offset) {
-			for (i = 16 >> (EXT3_BLOCK_SIZE_BITS(sb) - 9), num = 0;
-			     i > 0; i--) {
-				tmp = ext3_getblk (NULL, inode, ++blk, 0, &err);
-				if (tmp && !buffer_uptodate(tmp) &&
-						!buffer_locked(tmp))
-					bha[num++] = tmp;
-				else
-					brelse (tmp);
-			}
-			if (num) {
-				ll_rw_block (READA, num, bha);
-				for (i = 0; i < num; i++)
-					brelse (bha[i]);
-			}
-		}
-
 revalidate:
 		/* If the dir block has changed since the last call to
 		 * readdir(2), then we might be pointing to an invalid
diff --git a/fs/ext3/file.c b/fs/ext3/file.c
index 98e78345ead9..1efefb630ea9 100644
--- a/fs/ext3/file.c
+++ b/fs/ext3/file.c
@@ -37,9 +37,9 @@ static int ext3_release_file (struct inode * inode, struct file * filp)
 	if ((filp->f_mode & FMODE_WRITE) &&
 			(atomic_read(&inode->i_writecount) == 1))
 	{
-		down(&EXT3_I(inode)->truncate_sem);
+		mutex_lock(&EXT3_I(inode)->truncate_mutex);
 		ext3_discard_reservation(inode);
-		up(&EXT3_I(inode)->truncate_sem);
+		mutex_unlock(&EXT3_I(inode)->truncate_mutex);
 	}
 	if (is_dx(inode) && filp->private_data)
 		ext3_htree_free_dir_info(filp->private_data);
@@ -105,7 +105,7 @@ force_commit:
 	return ret;
 }
 
-struct file_operations ext3_file_operations = {
+const struct file_operations ext3_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
@@ -119,6 +119,8 @@ struct file_operations ext3_file_operations = {
 	.release	= ext3_release_file,
 	.fsync		= ext3_sync_file,
 	.sendfile	= generic_file_sendfile,
+	.splice_read	= generic_file_splice_read,
+	.splice_write	= generic_file_splice_write,
 };
 
 struct inode_operations ext3_file_inode_operations = {
diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c
index 0384e539b88f..48ae0339af17 100644
--- a/fs/ext3/inode.c
+++ b/fs/ext3/inode.c
@@ -44,16 +44,16 @@ static int ext3_writepage_trans_blocks(struct inode *inode);
 /*
  * Test whether an inode is a fast symlink.
  */
-static inline int ext3_inode_is_fast_symlink(struct inode *inode)
+static int ext3_inode_is_fast_symlink(struct inode *inode)
 {
 	int ea_blocks = EXT3_I(inode)->i_file_acl ?
 		(inode->i_sb->s_blocksize >> 9) : 0;
 
-	return (S_ISLNK(inode->i_mode) &&
-		inode->i_blocks - ea_blocks == 0);
+	return (S_ISLNK(inode->i_mode) && inode->i_blocks - ea_blocks == 0);
 }
 
-/* The ext3 forget function must perform a revoke if we are freeing data
+/*
+ * The ext3 forget function must perform a revoke if we are freeing data
  * which has been journaled.  Metadata (eg. indirect blocks) must be
  * revoked in all cases. 
  *
@@ -61,10 +61,8 @@ static inline int ext3_inode_is_fast_symlink(struct inode *inode)
  * but there may still be a record of it in the journal, and that record
  * still needs to be revoked.
  */
-
-int ext3_forget(handle_t *handle, int is_metadata,
-		       struct inode *inode, struct buffer_head *bh,
-		       int blocknr)
+int ext3_forget(handle_t *handle, int is_metadata, struct inode *inode,
+			struct buffer_head *bh, int blocknr)
 {
 	int err;
 
@@ -104,10 +102,9 @@ int ext3_forget(handle_t *handle, int is_metadata,
 }
 
 /*
- * Work out how many blocks we need to progress with the next chunk of a
+ * Work out how many blocks we need to proceed with the next chunk of a
  * truncate transaction.
  */
-
 static unsigned long blocks_for_truncate(struct inode *inode) 
 {
 	unsigned long needed;
@@ -141,7 +138,6 @@ static unsigned long blocks_for_truncate(struct inode *inode)
  * extend fails, we need to propagate the failure up and restart the
  * transaction in the top-level truncate loop. --sct 
  */
-
 static handle_t *start_transaction(struct inode *inode) 
 {
 	handle_t *result;
@@ -194,9 +190,11 @@ void ext3_delete_inode (struct inode * inode)
 
 	handle = start_transaction(inode);
 	if (IS_ERR(handle)) {
-		/* If we're going to skip the normal cleanup, we still
-		 * need to make sure that the in-core orphan linked list
-		 * is properly cleaned up. */
+		/*
+		 * If we're going to skip the normal cleanup, we still need to
+		 * make sure that the in-core orphan linked list is properly
+		 * cleaned up.
+		 */
 		ext3_orphan_del(NULL, inode);
 		goto no_delete;
 	}
@@ -235,16 +233,6 @@ no_delete:
 	clear_inode(inode);	/* We must guarantee clearing of inode... */
 }
 
-static int ext3_alloc_block (handle_t *handle,
-			struct inode * inode, unsigned long goal, int *err)
-{
-	unsigned long result;
-
-	result = ext3_new_block(handle, inode, goal, err);
-	return result;
-}
-
-
 typedef struct {
 	__le32	*p;
 	__le32	key;
@@ -257,7 +245,7 @@ static inline void add_chain(Indirect *p, struct buffer_head *bh, __le32 *v)
 	p->bh = bh;
 }
 
-static inline int verify_chain(Indirect *from, Indirect *to)
+static int verify_chain(Indirect *from, Indirect *to)
 {
 	while (from <= to && from->key == *from->p)
 		from++;
@@ -327,10 +315,10 @@ static int ext3_block_to_path(struct inode *inode,
 		offsets[n++] = i_block & (ptrs - 1);
 		final = ptrs;
 	} else {
-		ext3_warning (inode->i_sb, "ext3_block_to_path", "block > big");
+		ext3_warning(inode->i_sb, "ext3_block_to_path", "block > big");
 	}
 	if (boundary)
-		*boundary = (i_block & (ptrs - 1)) == (final - 1);
+		*boundary = final - 1 - (i_block & (ptrs - 1));
 	return n;
 }
 
@@ -419,7 +407,6 @@ no_block:
  *
  *	Caller must make sure that @ind is valid and will stay that way.
  */
-
 static unsigned long ext3_find_near(struct inode *inode, Indirect *ind)
 {
 	struct ext3_inode_info *ei = EXT3_I(inode);
@@ -429,17 +416,18 @@ static unsigned long ext3_find_near(struct inode *inode, Indirect *ind)
 	unsigned long colour;
 
 	/* Try to find previous block */
-	for (p = ind->p - 1; p >= start; p--)
+	for (p = ind->p - 1; p >= start; p--) {
 		if (*p)
 			return le32_to_cpu(*p);
+	}
 
 	/* No such thing, so let's try location of indirect block */
 	if (ind->bh)
 		return ind->bh->b_blocknr;
 
 	/*
-	 * It is going to be refered from inode itself? OK, just put it into
-	 * the same cylinder group then.
+	 * It is going to be referred to from the inode itself? OK, just put it
+	 * into the same cylinder group then.
 	 */
 	bg_start = (ei->i_block_group * EXT3_BLOCKS_PER_GROUP(inode->i_sb)) +
 		le32_to_cpu(EXT3_SB(inode->i_sb)->s_es->s_first_data_block);
@@ -463,7 +451,9 @@ static unsigned long ext3_find_near(struct inode *inode, Indirect *ind)
 static unsigned long ext3_find_goal(struct inode *inode, long block,
 		Indirect chain[4], Indirect *partial)
 {
-	struct ext3_block_alloc_info *block_i =  EXT3_I(inode)->i_block_alloc_info;
+	struct ext3_block_alloc_info *block_i;
+
+	block_i =  EXT3_I(inode)->i_block_alloc_info;
 
 	/*
 	 * try the heuristic for sequential allocation,
@@ -478,13 +468,113 @@ static unsigned long ext3_find_goal(struct inode *inode, long block,
 }
 
 /**
+ *	ext3_blks_to_allocate: Look up the block map and count the number
+ *	of direct blocks need to be allocated for the given branch.
+ *
+ * 	@branch: chain of indirect blocks
+ *	@k: number of blocks need for indirect blocks
+ *	@blks: number of data blocks to be mapped.
+ *	@blocks_to_boundary:  the offset in the indirect block
+ *
+ *	return the total number of blocks to be allocate, including the
+ *	direct and indirect blocks.
+ */
+static int ext3_blks_to_allocate(Indirect *branch, int k, unsigned long blks,
+		int blocks_to_boundary)
+{
+	unsigned long count = 0;
+
+	/*
+	 * Simple case, [t,d]Indirect block(s) has not allocated yet
+	 * then it's clear blocks on that path have not allocated
+	 */
+	if (k > 0) {
+		/* right now we don't handle cross boundary allocation */
+		if (blks < blocks_to_boundary + 1)
+			count += blks;
+		else
+			count += blocks_to_boundary + 1;
+		return count;
+	}
+
+	count++;
+	while (count < blks && count <= blocks_to_boundary &&
+		le32_to_cpu(*(branch[0].p + count)) == 0) {
+		count++;
+	}
+	return count;
+}
+
+/**
+ *	ext3_alloc_blocks: multiple allocate blocks needed for a branch
+ *	@indirect_blks: the number of blocks need to allocate for indirect
+ *			blocks
+ *
+ *	@new_blocks: on return it will store the new block numbers for
+ *	the indirect blocks(if needed) and the first direct block,
+ *	@blks:	on return it will store the total number of allocated
+ *		direct blocks
+ */
+static int ext3_alloc_blocks(handle_t *handle, struct inode *inode,
+			unsigned long goal, int indirect_blks, int blks,
+			unsigned long long new_blocks[4], int *err)
+{
+	int target, i;
+	unsigned long count = 0;
+	int index = 0;
+	unsigned long current_block = 0;
+	int ret = 0;
+
+	/*
+	 * Here we try to allocate the requested multiple blocks at once,
+	 * on a best-effort basis.
+	 * To build a branch, we should allocate blocks for
+	 * the indirect blocks(if not allocated yet), and at least
+	 * the first direct block of this branch.  That's the
+	 * minimum number of blocks need to allocate(required)
+	 */
+	target = blks + indirect_blks;
+
+	while (1) {
+		count = target;
+		/* allocating blocks for indirect blocks and direct blocks */
+		current_block = ext3_new_blocks(handle,inode,goal,&count,err);
+		if (*err)
+			goto failed_out;
+
+		target -= count;
+		/* allocate blocks for indirect blocks */
+		while (index < indirect_blks && count) {
+			new_blocks[index++] = current_block++;
+			count--;
+		}
+
+		if (count > 0)
+			break;
+	}
+
+	/* save the new block number for the first direct block */
+	new_blocks[index] = current_block;
+
+	/* total number of blocks allocated for direct blocks */
+	ret = count;
+	*err = 0;
+	return ret;
+failed_out:
+	for (i = 0; i <index; i++)
+		ext3_free_blocks(handle, inode, new_blocks[i], 1);
+	return ret;
+}
+
+/**
  *	ext3_alloc_branch - allocate and set up a chain of blocks.
  *	@inode: owner
- *	@num: depth of the chain (number of blocks to allocate)
+ *	@indirect_blks: number of allocated indirect blocks
+ *	@blks: number of allocated direct blocks
  *	@offsets: offsets (in the blocks) to store the pointers to next.
  *	@branch: place to store the chain in.
  *
- *	This function allocates @num blocks, zeroes out all but the last one,
+ *	This function allocates blocks, zeroes out all but the last one,
  *	links them into chain and (if we are synchronous) writes them to disk.
  *	In other words, it prepares a branch that can be spliced onto the
  *	inode. It stores the information about that chain in the branch[], in
@@ -501,97 +591,106 @@ static unsigned long ext3_find_goal(struct inode *inode, long block,
  *	ext3_alloc_block() (normally -ENOSPC). Otherwise we set the chain
  *	as described above and return 0.
  */
-
 static int ext3_alloc_branch(handle_t *handle, struct inode *inode,
-			     int num,
-			     unsigned long goal,
-			     int *offsets,
-			     Indirect *branch)
+			int indirect_blks, int *blks, unsigned long goal,
+			int *offsets, Indirect *branch)
 {
 	int blocksize = inode->i_sb->s_blocksize;
-	int n = 0, keys = 0;
+	int i, n = 0;
 	int err = 0;
-	int i;
-	int parent = ext3_alloc_block(handle, inode, goal, &err);
-
-	branch[0].key = cpu_to_le32(parent);
-	if (parent) {
-		for (n = 1; n < num; n++) {
-			struct buffer_head *bh;
-			/* Allocate the next block */
-			int nr = ext3_alloc_block(handle, inode, parent, &err);
-			if (!nr)
-				break;
-			branch[n].key = cpu_to_le32(nr);
+	struct buffer_head *bh;
+	int num;
+	unsigned long long new_blocks[4];
+	unsigned long long current_block;
 
-			/*
-			 * Get buffer_head for parent block, zero it out
-			 * and set the pointer to new one, then send
-			 * parent to disk.  
-			 */
-			bh = sb_getblk(inode->i_sb, parent);
-			if (!bh)
-				break;
-			keys = n+1;
-			branch[n].bh = bh;
-			lock_buffer(bh);
-			BUFFER_TRACE(bh, "call get_create_access");
-			err = ext3_journal_get_create_access(handle, bh);
-			if (err) {
-				unlock_buffer(bh);
-				brelse(bh);
-				break;
-			}
+	num = ext3_alloc_blocks(handle, inode, goal, indirect_blks,
+				*blks, new_blocks, &err);
+	if (err)
+		return err;
 
-			memset(bh->b_data, 0, blocksize);
-			branch[n].p = (__le32*) bh->b_data + offsets[n];
-			*branch[n].p = branch[n].key;
-			BUFFER_TRACE(bh, "marking uptodate");
-			set_buffer_uptodate(bh);
+	branch[0].key = cpu_to_le32(new_blocks[0]);
+	/*
+	 * metadata blocks and data blocks are allocated.
+	 */
+	for (n = 1; n <= indirect_blks;  n++) {
+		/*
+		 * Get buffer_head for parent block, zero it out
+		 * and set the pointer to new one, then send
+		 * parent to disk.
+		 */
+		bh = sb_getblk(inode->i_sb, new_blocks[n-1]);
+		branch[n].bh = bh;
+		lock_buffer(bh);
+		BUFFER_TRACE(bh, "call get_create_access");
+		err = ext3_journal_get_create_access(handle, bh);
+		if (err) {
 			unlock_buffer(bh);
+			brelse(bh);
+			goto failed;
+		}
 
-			BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
-			err = ext3_journal_dirty_metadata(handle, bh);
-			if (err)
-				break;
-
-			parent = nr;
+		memset(bh->b_data, 0, blocksize);
+		branch[n].p = (__le32 *) bh->b_data + offsets[n];
+		branch[n].key = cpu_to_le32(new_blocks[n]);
+		*branch[n].p = branch[n].key;
+		if ( n == indirect_blks) {
+			current_block = new_blocks[n];
+			/*
+			 * End of chain, update the last new metablock of
+			 * the chain to point to the new allocated
+			 * data blocks numbers
+			 */
+			for (i=1; i < num; i++)
+				*(branch[n].p + i) = cpu_to_le32(++current_block);
 		}
-	}
-	if (n == num)
-		return 0;
+		BUFFER_TRACE(bh, "marking uptodate");
+		set_buffer_uptodate(bh);
+		unlock_buffer(bh);
 
+		BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata");
+		err = ext3_journal_dirty_metadata(handle, bh);
+		if (err)
+			goto failed;
+	}
+	*blks = num;
+	return err;
+failed:
 	/* Allocation failed, free what we already allocated */
-	for (i = 1; i < keys; i++) {
+	for (i = 1; i <= n ; i++) {
 		BUFFER_TRACE(branch[i].bh, "call journal_forget");
 		ext3_journal_forget(handle, branch[i].bh);
 	}
-	for (i = 0; i < keys; i++)
-		ext3_free_blocks(handle, inode, le32_to_cpu(branch[i].key), 1);
+	for (i = 0; i <indirect_blks; i++)
+		ext3_free_blocks(handle, inode, new_blocks[i], 1);
+
+	ext3_free_blocks(handle, inode, new_blocks[i], num);
+
 	return err;
 }
 
 /**
- *	ext3_splice_branch - splice the allocated branch onto inode.
- *	@inode: owner
- *	@block: (logical) number of block we are adding
- *	@chain: chain of indirect blocks (with a missing link - see
- *		ext3_alloc_branch)
- *	@where: location of missing link
- *	@num:   number of blocks we are adding
- *
- *	This function fills the missing link and does all housekeeping needed in
- *	inode (->i_blocks, etc.). In case of success we end up with the full
- *	chain to new block and return 0.
+ * ext3_splice_branch - splice the allocated branch onto inode.
+ * @inode: owner
+ * @block: (logical) number of block we are adding
+ * @chain: chain of indirect blocks (with a missing link - see
+ *	ext3_alloc_branch)
+ * @where: location of missing link
+ * @num:   number of indirect blocks we are adding
+ * @blks:  number of direct blocks we are adding
+ *
+ * This function fills the missing link and does all housekeeping needed in
+ * inode (->i_blocks, etc.). In case of success we end up with the full
+ * chain to new block and return 0.
  */
-
-static int ext3_splice_branch(handle_t *handle, struct inode *inode, long block,
-			      Indirect chain[4], Indirect *where, int num)
+static int ext3_splice_branch(handle_t *handle, struct inode *inode,
+			long block, Indirect *where, int num, int blks)
 {
 	int i;
 	int err = 0;
-	struct ext3_block_alloc_info *block_i = EXT3_I(inode)->i_block_alloc_info;
+	struct ext3_block_alloc_info *block_i;
+	unsigned long current_block;
 
+	block_i = EXT3_I(inode)->i_block_alloc_info;
 	/*
 	 * If we're splicing into a [td]indirect block (as opposed to the
 	 * inode) then we need to get write access to the [td]indirect block
@@ -608,13 +707,24 @@ static int ext3_splice_branch(handle_t *handle, struct inode *inode, long block,
 	*where->p = where->key;
 
 	/*
+	 * Update the host buffer_head or inode to point to more just allocated
+	 * direct blocks blocks
+	 */
+	if (num == 0 && blks > 1) {
+		current_block = le32_to_cpu(where->key + 1);
+		for (i = 1; i < blks; i++)
+			*(where->p + i ) = cpu_to_le32(current_block++);
+	}
+
+	/*
 	 * update the most recently allocated logical & physical block
 	 * in i_block_alloc_info, to assist find the proper goal block for next
 	 * allocation
 	 */
 	if (block_i) {
-		block_i->last_alloc_logical_block = block;
-		block_i->last_alloc_physical_block = le32_to_cpu(where[num-1].key);
+		block_i->last_alloc_logical_block = block + blks - 1;
+		block_i->last_alloc_physical_block =
+				le32_to_cpu(where[num].key + blks - 1);
 	}
 
 	/* We are done with atomic stuff, now do the rest of housekeeping */
@@ -625,7 +735,7 @@ static int ext3_splice_branch(handle_t *handle, struct inode *inode, long block,
 	/* had we spliced it onto indirect block? */
 	if (where->bh) {
 		/*
-		 * akpm: If we spliced it onto an indirect block, we haven't
+		 * If we spliced it onto an indirect block, we haven't
 		 * altered the inode.  Note however that if it is being spliced
 		 * onto an indirect block at the very end of the file (the
 		 * file is growing) then we *will* alter the inode to reflect
@@ -647,10 +757,13 @@ static int ext3_splice_branch(handle_t *handle, struct inode *inode, long block,
 	return err;
 
 err_out:
-	for (i = 1; i < num; i++) {
+	for (i = 1; i <= num; i++) {
 		BUFFER_TRACE(where[i].bh, "call journal_forget");
 		ext3_journal_forget(handle, where[i].bh);
+		ext3_free_blocks(handle,inode,le32_to_cpu(where[i-1].key),1);
 	}
+	ext3_free_blocks(handle, inode, le32_to_cpu(where[num].key), blks);
+
 	return err;
 }
 
@@ -666,26 +779,33 @@ err_out:
  * allocations is needed - we simply release blocks and do not touch anything
  * reachable from inode.
  *
- * akpm: `handle' can be NULL if create == 0.
+ * `handle' can be NULL if create == 0.
  *
  * The BKL may not be held on entry here.  Be sure to take it early.
+ * return > 0, # of blocks mapped or allocated.
+ * return = 0, if plain lookup failed.
+ * return < 0, error case.
  */
-
-static int
-ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
-		struct buffer_head *bh_result, int create, int extend_disksize)
+int ext3_get_blocks_handle(handle_t *handle, struct inode *inode,
+		sector_t iblock, unsigned long maxblocks,
+		struct buffer_head *bh_result,
+		int create, int extend_disksize)
 {
 	int err = -EIO;
 	int offsets[4];
 	Indirect chain[4];
 	Indirect *partial;
 	unsigned long goal;
-	int left;
-	int boundary = 0;
-	const int depth = ext3_block_to_path(inode, iblock, offsets, &boundary);
+	int indirect_blks;
+	int blocks_to_boundary = 0;
+	int depth;
 	struct ext3_inode_info *ei = EXT3_I(inode);
+	int count = 0;
+	unsigned long first_block = 0;
+
 
 	J_ASSERT(handle != NULL || create == 0);
+	depth = ext3_block_to_path(inode,iblock,offsets,&blocks_to_boundary);
 
 	if (depth == 0)
 		goto out;
@@ -694,15 +814,38 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 
 	/* Simplest case - block found, no allocation needed */
 	if (!partial) {
+		first_block = chain[depth - 1].key;
 		clear_buffer_new(bh_result);
-		goto got_it;
+		count++;
+		/*map more blocks*/
+		while (count < maxblocks && count <= blocks_to_boundary) {
+			if (!verify_chain(chain, partial)) {
+				/*
+				 * Indirect block might be removed by
+				 * truncate while we were reading it.
+				 * Handling of that case: forget what we've
+				 * got now. Flag the err as EAGAIN, so it
+				 * will reread.
+				 */
+				err = -EAGAIN;
+				count = 0;
+				break;
+			}
+			if (le32_to_cpu(*(chain[depth-1].p+count) ==
+					(first_block + count)))
+				count++;
+			else
+				break;
+		}
+		if (err != -EAGAIN)
+			goto got_it;
 	}
 
 	/* Next simple case - plain lookup or failed read of indirect block */
 	if (!create || err == -EIO)
 		goto cleanup;
 
-	down(&ei->truncate_sem);
+	mutex_lock(&ei->truncate_mutex);
 
 	/*
 	 * If the indirect block is missing while we are reading
@@ -723,7 +866,8 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 		}
 		partial = ext3_get_branch(inode, depth, offsets, chain, &err);
 		if (!partial) {
-			up(&ei->truncate_sem);
+			count++;
+			mutex_unlock(&ei->truncate_mutex);
 			if (err)
 				goto cleanup;
 			clear_buffer_new(bh_result);
@@ -740,12 +884,19 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 
 	goal = ext3_find_goal(inode, iblock, chain, partial);
 
-	left = (chain + depth) - partial;
+	/* the number of blocks need to allocate for [d,t]indirect blocks */
+	indirect_blks = (chain + depth) - partial - 1;
 
 	/*
+	 * Next look up the indirect map to count the totoal number of
+	 * direct blocks to allocate for this branch.
+	 */
+	count = ext3_blks_to_allocate(partial, indirect_blks,
+					maxblocks, blocks_to_boundary);
+	/*
 	 * Block out ext3_truncate while we alter the tree
 	 */
-	err = ext3_alloc_branch(handle, inode, left, goal,
+	err = ext3_alloc_branch(handle, inode, indirect_blks, &count, goal,
 				offsets + (partial - chain), partial);
 
 	/*
@@ -756,24 +907,25 @@ ext3_get_block_handle(handle_t *handle, struct inode *inode, sector_t iblock,
 	 * may need to return -EAGAIN upwards in the worst case.  --sct
 	 */
 	if (!err)
-		err = ext3_splice_branch(handle, inode, iblock, chain,
-					 partial, left);
+		err = ext3_splice_branch(handle, inode, iblock,
+					partial, indirect_blks, count);
 	/*
-	 * i_disksize growing is protected by truncate_sem.  Don't forget to
+	 * i_disksize growing is protected by truncate_mutex.  Don't forget to
 	 * protect it if you're about to implement concurrent
 	 * ext3_get_block() -bzzz
 	*/
 	if (!err && extend_disksize && inode->i_size > ei->i_disksize)
 		ei->i_disksize = inode->i_size;
-	up(&ei->truncate_sem);
+	mutex_unlock(&ei->truncate_mutex);
 	if (err)
 		goto cleanup;
 
 	set_buffer_new(bh_result);
 got_it:
 	map_bh(bh_result, inode->i_sb, le32_to_cpu(chain[depth-1].key));
-	if (boundary)
+	if (blocks_to_boundary == 0)
 		set_buffer_boundary(bh_result);
+	err = count;
 	/* Clean up and exit */
 	partial = chain + depth - 1;	/* the whole chain */
 cleanup:
@@ -787,34 +939,21 @@ out:
 	return err;
 }
 
-static int ext3_get_block(struct inode *inode, sector_t iblock,
-			struct buffer_head *bh_result, int create)
-{
-	handle_t *handle = NULL;
-	int ret;
-
-	if (create) {
-		handle = ext3_journal_current_handle();
-		J_ASSERT(handle != 0);
-	}
-	ret = ext3_get_block_handle(handle, inode, iblock,
-				bh_result, create, 1);
-	return ret;
-}
-
 #define DIO_CREDITS (EXT3_RESERVE_TRANS_BLOCKS + 32)
 
-static int
-ext3_direct_io_get_blocks(struct inode *inode, sector_t iblock,
-		unsigned long max_blocks, struct buffer_head *bh_result,
-		int create)
+static int ext3_get_block(struct inode *inode, sector_t iblock,
+			struct buffer_head *bh_result, int create)
 {
 	handle_t *handle = journal_current_handle();
 	int ret = 0;
+	unsigned max_blocks = bh_result->b_size >> inode->i_blkbits;
 
-	if (!handle)
+	if (!create)
 		goto get_block;		/* A read */
 
+	if (max_blocks == 1)
+		goto get_block;		/* A single block get */
+
 	if (handle->h_transaction->t_state == T_LOCKED) {
 		/*
 		 * Huge direct-io writes can hold off commits for long
@@ -841,18 +980,22 @@ ext3_direct_io_get_blocks(struct inode *inode, sector_t iblock,
 	}
 
 get_block:
-	if (ret == 0)
-		ret = ext3_get_block_handle(handle, inode, iblock,
-					bh_result, create, 0);
-	bh_result->b_size = (1 << inode->i_blkbits);
+	if (ret == 0) {
+		ret = ext3_get_blocks_handle(handle, inode, iblock,
+					max_blocks, bh_result, create, 0);
+		if (ret > 0) {
+			bh_result->b_size = (ret << inode->i_blkbits);
+			ret = 0;
+		}
+	}
 	return ret;
 }
 
 /*
  * `handle' can be NULL if create is zero
  */
-struct buffer_head *ext3_getblk(handle_t *handle, struct inode * inode,
-				long block, int create, int * errp)
+struct buffer_head *ext3_getblk(handle_t *handle, struct inode *inode,
+				long block, int create, int *errp)
 {
 	struct buffer_head dummy;
 	int fatal = 0, err;
@@ -862,8 +1005,16 @@ struct buffer_head *ext3_getblk(handle_t *handle, struct inode * inode,
 	dummy.b_state = 0;
 	dummy.b_blocknr = -1000;
 	buffer_trace_init(&dummy.b_history);
-	*errp = ext3_get_block_handle(handle, inode, block, &dummy, create, 1);
-	if (!*errp && buffer_mapped(&dummy)) {
+	err = ext3_get_blocks_handle(handle, inode, block, 1,
+					&dummy, create, 1);
+	if (err == 1) {
+		err = 0;
+	} else if (err >= 0) {
+		WARN_ON(1);
+		err = -EIO;
+	}
+	*errp = err;
+	if (!err && buffer_mapped(&dummy)) {
 		struct buffer_head *bh;
 		bh = sb_getblk(inode->i_sb, dummy.b_blocknr);
 		if (!bh) {
@@ -874,17 +1025,18 @@ struct buffer_head *ext3_getblk(handle_t *handle, struct inode * inode,
 			J_ASSERT(create != 0);
 			J_ASSERT(handle != 0);
 
-			/* Now that we do not always journal data, we
-			   should keep in mind whether this should
-			   always journal the new buffer as metadata.
-			   For now, regular file writes use
-			   ext3_get_block instead, so it's not a
-			   problem. */
+			/*
+			 * Now that we do not always journal data, we should
+			 * keep in mind whether this should always journal the
+			 * new buffer as metadata.  For now, regular file
+			 * writes use ext3_get_block instead, so it's not a
+			 * problem.
+			 */
 			lock_buffer(bh);
 			BUFFER_TRACE(bh, "call get_create_access");
 			fatal = ext3_journal_get_create_access(handle, bh);
 			if (!fatal && !buffer_uptodate(bh)) {
-				memset(bh->b_data, 0, inode->i_sb->s_blocksize);
+				memset(bh->b_data,0,inode->i_sb->s_blocksize);
 				set_buffer_uptodate(bh);
 			}
 			unlock_buffer(bh);
@@ -906,7 +1058,7 @@ err:
 	return NULL;
 }
 
-struct buffer_head *ext3_bread(handle_t *handle, struct inode * inode,
+struct buffer_head *ext3_bread(handle_t *handle, struct inode *inode,
 			       int block, int create, int *err)
 {
 	struct buffer_head * bh;
@@ -982,9 +1134,8 @@ static int walk_page_buffers(	handle_t *handle,
  * is elevated.  We'll still have enough credits for the tiny quotafile
  * write.  
  */
-
-static int do_journal_get_write_access(handle_t *handle, 
-				       struct buffer_head *bh)
+static int do_journal_get_write_access(handle_t *handle,
+					struct buffer_head *bh)
 {
 	if (!buffer_mapped(bh) || buffer_freed(bh))
 		return 0;
@@ -1025,8 +1176,7 @@ out:
 	return ret;
 }
 
-int
-ext3_journal_dirty_data(handle_t *handle, struct buffer_head *bh)
+int ext3_journal_dirty_data(handle_t *handle, struct buffer_head *bh)
 {
 	int err = journal_dirty_data(handle, bh);
 	if (err)
@@ -1051,7 +1201,6 @@ static int commit_write_fn(handle_t *handle, struct buffer_head *bh)
  * ext3 never places buffers on inode->i_mapping->private_list.  metadata
  * buffers are managed internally.
  */
-
 static int ext3_ordered_commit_write(struct file *file, struct page *page,
 			     unsigned from, unsigned to)
 {
@@ -1227,7 +1376,7 @@ static int journal_dirty_data_fn(handle_t *handle, struct buffer_head *bh)
  *	ext3_file_write() -> generic_file_write() -> __alloc_pages() -> ...
  *
  * Same applies to ext3_get_block().  We will deadlock on various things like
- * lock_journal and i_truncate_sem.
+ * lock_journal and i_truncate_mutex.
  *
  * Setting PF_MEMALLOC here doesn't work - too many internal memory
  * allocations fail.
@@ -1261,7 +1410,7 @@ static int journal_dirty_data_fn(handle_t *handle, struct buffer_head *bh)
  * we don't need to open a transaction here.
  */
 static int ext3_ordered_writepage(struct page *page,
-			struct writeback_control *wbc)
+				struct writeback_control *wbc)
 {
 	struct inode *inode = page->mapping->host;
 	struct buffer_head *page_bufs;
@@ -1430,7 +1579,7 @@ ext3_readpages(struct file *file, struct address_space *mapping,
 	return mpage_readpages(mapping, pages, nr_pages, ext3_get_block);
 }
 
-static int ext3_invalidatepage(struct page *page, unsigned long offset)
+static void ext3_invalidatepage(struct page *page, unsigned long offset)
 {
 	journal_t *journal = EXT3_JOURNAL(page->mapping->host);
 
@@ -1440,7 +1589,7 @@ static int ext3_invalidatepage(struct page *page, unsigned long offset)
 	if (offset == 0)
 		ClearPageChecked(page);
 
-	return journal_invalidatepage(journal, page, offset);
+	journal_invalidatepage(journal, page, offset);
 }
 
 static int ext3_releasepage(struct page *page, gfp_t wait)
@@ -1492,11 +1641,10 @@ static ssize_t ext3_direct_IO(int rw, struct kiocb *iocb,
 
 	ret = blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov, 
 				 offset, nr_segs,
-				 ext3_direct_io_get_blocks, NULL);
+				 ext3_get_block, NULL);
 
 	/*
-	 * Reacquire the handle: ext3_direct_io_get_block() can restart the
-	 * transaction
+	 * Reacquire the handle: ext3_get_block() can restart the transaction
 	 */
 	handle = journal_current_handle();
 
@@ -1752,11 +1900,8 @@ static inline int all_zeroes(__le32 *p, __le32 *q)
  *		c) free the subtrees growing from the inode past the @chain[0].
  *			(no partially truncated stuff there).  */
 
-static Indirect *ext3_find_shared(struct inode *inode,
-				int depth,
-				int offsets[4],
-				Indirect chain[4],
-				__le32 *top)
+static Indirect *ext3_find_shared(struct inode *inode, int depth,
+			int offsets[4], Indirect chain[4], __le32 *top)
 {
 	Indirect *partial, *p;
 	int k, err;
@@ -1795,8 +1940,7 @@ static Indirect *ext3_find_shared(struct inode *inode,
 	}
 	/* Writer: end */
 
-	while(partial > p)
-	{
+	while(partial > p) {
 		brelse(partial->bh);
 		partial--;
 	}
@@ -1812,10 +1956,9 @@ no_top:
  * We release `count' blocks on disk, but (last - first) may be greater
  * than `count' because there can be holes in there.
  */
-static void
-ext3_clear_blocks(handle_t *handle, struct inode *inode, struct buffer_head *bh,
-		unsigned long block_to_free, unsigned long count,
-		__le32 *first, __le32 *last)
+static void ext3_clear_blocks(handle_t *handle, struct inode *inode,
+		struct buffer_head *bh, unsigned long block_to_free,
+		unsigned long count, __le32 *first, __le32 *last)
 {
 	__le32 *p;
 	if (try_to_extend_transaction(handle, inode)) {
@@ -2076,8 +2219,7 @@ static void ext3_free_branches(handle_t *handle, struct inode *inode,
  * that's fine - as long as they are linked from the inode, the post-crash
  * ext3_truncate() run will find them and release them.
  */
-
-void ext3_truncate(struct inode * inode)
+void ext3_truncate(struct inode *inode)
 {
 	handle_t *handle;
 	struct ext3_inode_info *ei = EXT3_I(inode);
@@ -2161,7 +2303,7 @@ void ext3_truncate(struct inode * inode)
 	 * From here we block out all ext3_get_block() callers who want to
 	 * modify the block allocation tree.
 	 */
-	down(&ei->truncate_sem);
+	mutex_lock(&ei->truncate_mutex);
 
 	if (n == 1) {		/* direct blocks */
 		ext3_free_data(handle, inode, NULL, i_data+offsets[0],
@@ -2201,39 +2343,38 @@ void ext3_truncate(struct inode * inode)
 do_indirects:
 	/* Kill the remaining (whole) subtrees */
 	switch (offsets[0]) {
-		default:
-			nr = i_data[EXT3_IND_BLOCK];
-			if (nr) {
-				ext3_free_branches(handle, inode, NULL,
-						   &nr, &nr+1, 1);
-				i_data[EXT3_IND_BLOCK] = 0;
-			}
-		case EXT3_IND_BLOCK:
-			nr = i_data[EXT3_DIND_BLOCK];
-			if (nr) {
-				ext3_free_branches(handle, inode, NULL,
-						   &nr, &nr+1, 2);
-				i_data[EXT3_DIND_BLOCK] = 0;
-			}
-		case EXT3_DIND_BLOCK:
-			nr = i_data[EXT3_TIND_BLOCK];
-			if (nr) {
-				ext3_free_branches(handle, inode, NULL,
-						   &nr, &nr+1, 3);
-				i_data[EXT3_TIND_BLOCK] = 0;
-			}
-		case EXT3_TIND_BLOCK:
-			;
+	default:
+		nr = i_data[EXT3_IND_BLOCK];
+		if (nr) {
+			ext3_free_branches(handle, inode, NULL, &nr, &nr+1, 1);
+			i_data[EXT3_IND_BLOCK] = 0;
+		}
+	case EXT3_IND_BLOCK:
+		nr = i_data[EXT3_DIND_BLOCK];
+		if (nr) {
+			ext3_free_branches(handle, inode, NULL, &nr, &nr+1, 2);
+			i_data[EXT3_DIND_BLOCK] = 0;
+		}
+	case EXT3_DIND_BLOCK:
+		nr = i_data[EXT3_TIND_BLOCK];
+		if (nr) {
+			ext3_free_branches(handle, inode, NULL, &nr, &nr+1, 3);
+			i_data[EXT3_TIND_BLOCK] = 0;
+		}
+	case EXT3_TIND_BLOCK:
+		;
 	}
 
 	ext3_discard_reservation(inode);
 
-	up(&ei->truncate_sem);
+	mutex_unlock(&ei->truncate_mutex);
 	inode->i_mtime = inode->i_ctime = CURRENT_TIME_SEC;
 	ext3_mark_inode_dirty(handle, inode);
 
-	/* In a multi-transaction truncate, we only make the final
-	 * transaction synchronous */
+	/*
+	 * In a multi-transaction truncate, we only make the final transaction
+	 * synchronous
+	 */
 	if (IS_SYNC(inode))
 		handle->h_sync = 1;
 out_stop:
@@ -2259,20 +2400,16 @@ static unsigned long ext3_get_inode_block(struct super_block *sb,
 	struct ext3_group_desc * gdp;
 
 
-	if ((ino != EXT3_ROOT_INO &&
-		ino != EXT3_JOURNAL_INO &&
-		ino != EXT3_RESIZE_INO &&
-		ino < EXT3_FIRST_INO(sb)) ||
-		ino > le32_to_cpu(
-			EXT3_SB(sb)->s_es->s_inodes_count)) {
-		ext3_error (sb, "ext3_get_inode_block",
+	if ((ino != EXT3_ROOT_INO && ino != EXT3_JOURNAL_INO &&
+		ino != EXT3_RESIZE_INO && ino < EXT3_FIRST_INO(sb)) ||
+		ino > le32_to_cpu(EXT3_SB(sb)->s_es->s_inodes_count)) {
+		ext3_error(sb, "ext3_get_inode_block",
 			    "bad inode number: %lu", ino);
 		return 0;
 	}
 	block_group = (ino - 1) / EXT3_INODES_PER_GROUP(sb);
 	if (block_group >= EXT3_SB(sb)->s_groups_count) {
-		ext3_error (sb, "ext3_get_inode_block",
-			    "group >= groups count");
+		ext3_error(sb,"ext3_get_inode_block","group >= groups count");
 		return 0;
 	}
 	smp_rmb();
@@ -2285,7 +2422,7 @@ static unsigned long ext3_get_inode_block(struct super_block *sb,
 		return 0;
 	}
 
-	gdp = (struct ext3_group_desc *) bh->b_data;
+	gdp = (struct ext3_group_desc *)bh->b_data;
 	/*
 	 * Figure out the offset within the block group inode table
 	 */
@@ -2834,7 +2971,7 @@ err_out:
 
 
 /*
- * akpm: how many blocks doth make a writepage()?
+ * How many blocks doth make a writepage()?
  *
  * With N blocks per page, it may be:
  * N data blocks
@@ -2924,8 +3061,8 @@ ext3_reserve_inode_write(handle_t *handle, struct inode *inode,
 }
 
 /*
- * akpm: What we do here is to mark the in-core inode as clean
- * with respect to inode dirtiness (it may still be data-dirty).
+ * What we do here is to mark the in-core inode as clean with respect to inode
+ * dirtiness (it may still be data-dirty).
  * This means that the in-core inode may be reaped by prune_icache
  * without having to perform any I/O.  This is a very good thing,
  * because *any* task may call prune_icache - even ones which
@@ -2957,7 +3094,7 @@ int ext3_mark_inode_dirty(handle_t *handle, struct inode *inode)
 }
 
 /*
- * akpm: ext3_dirty_inode() is called from __mark_inode_dirty()
+ * ext3_dirty_inode() is called from __mark_inode_dirty()
  *
  * We're really interested in the case where a file is being extended.
  * i_size has been changed by generic_commit_write() and we thus need
@@ -2993,7 +3130,7 @@ out:
 	return;
 }
 
-#ifdef AKPM
+#if 0
 /* 
  * Bind an inode's backing buffer_head into this transaction, to prevent
  * it from being flushed to disk early.  Unlike
@@ -3001,8 +3138,7 @@ out:
  * returns no iloc structure, so the caller needs to repeat the iloc
  * lookup to mark the inode dirty later.
  */
-static inline int
-ext3_pin_inode(handle_t *handle, struct inode *inode)
+static int ext3_pin_inode(handle_t *handle, struct inode *inode)
 {
 	struct ext3_iloc iloc;
 
diff --git a/fs/ext3/ioctl.c b/fs/ext3/ioctl.c
index 556cd5510078..aaf1da17b6d4 100644
--- a/fs/ext3/ioctl.c
+++ b/fs/ext3/ioctl.c
@@ -182,7 +182,7 @@ flags_err:
 		 * need to allocate reservation structure for this inode
 		 * before set the window size
 		 */
-		down(&ei->truncate_sem);
+		mutex_lock(&ei->truncate_mutex);
 		if (!ei->i_block_alloc_info)
 			ext3_init_block_alloc_info(inode);
 
@@ -190,7 +190,7 @@ flags_err:
 			struct ext3_reserve_window_node *rsv = &ei->i_block_alloc_info->rsv_window_node;
 			rsv->rsv_goal_size = rsv_window_size;
 		}
-		up(&ei->truncate_sem);
+		mutex_unlock(&ei->truncate_mutex);
 		return 0;
 	}
 	case EXT3_IOC_GROUP_EXTEND: {
diff --git a/fs/ext3/super.c b/fs/ext3/super.c
index 56bf76586019..f8a5266ea1ff 100644
--- a/fs/ext3/super.c
+++ b/fs/ext3/super.c
@@ -472,7 +472,7 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 #ifdef CONFIG_EXT3_FS_XATTR
 		init_rwsem(&ei->xattr_sem);
 #endif
-		init_MUTEX(&ei->truncate_sem);
+		mutex_init(&ei->truncate_mutex);
 		inode_init_once(&ei->vfs_inode);
 	}
 }
@@ -481,7 +481,8 @@ static int init_inodecache(void)
 {
 	ext3_inode_cachep = kmem_cache_create("ext3_inode_cache",
 					     sizeof(struct ext3_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (ext3_inode_cachep == NULL)
 		return -ENOMEM;
@@ -1677,12 +1678,6 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent)
 	}
 
 	if (test_opt(sb, NOBH)) {
-		if (sb->s_blocksize_bits != PAGE_CACHE_SHIFT) {
-			printk(KERN_WARNING "EXT3-fs: Ignoring nobh option "
-				"since filesystem blocksize doesn't match "
-				"pagesize\n");
-			clear_opt(sbi->s_mount_opt, NOBH);
-		}
 		if (!(test_opt(sb, DATA_FLAGS) == EXT3_MOUNT_WRITEBACK_DATA)) {
 			printk(KERN_WARNING "EXT3-fs: Ignoring nobh option - "
 				"its supported only with writeback mode\n");
@@ -2325,7 +2320,8 @@ restore_opts:
 
 static int ext3_statfs (struct super_block * sb, struct kstatfs * buf)
 {
-	struct ext3_super_block *es = EXT3_SB(sb)->s_es;
+	struct ext3_sb_info *sbi = EXT3_SB(sb);
+	struct ext3_super_block *es = sbi->s_es;
 	unsigned long overhead;
 	int i;
 
@@ -2367,12 +2363,12 @@ static int ext3_statfs (struct super_block * sb, struct kstatfs * buf)
 	buf->f_type = EXT3_SUPER_MAGIC;
 	buf->f_bsize = sb->s_blocksize;
 	buf->f_blocks = le32_to_cpu(es->s_blocks_count) - overhead;
-	buf->f_bfree = ext3_count_free_blocks (sb);
+	buf->f_bfree = percpu_counter_sum(&sbi->s_freeblocks_counter);
 	buf->f_bavail = buf->f_bfree - le32_to_cpu(es->s_r_blocks_count);
 	if (buf->f_bfree < le32_to_cpu(es->s_r_blocks_count))
 		buf->f_bavail = 0;
 	buf->f_files = le32_to_cpu(es->s_inodes_count);
-	buf->f_ffree = ext3_count_free_inodes (sb);
+	buf->f_ffree = percpu_counter_sum(&sbi->s_freeinodes_counter);
 	buf->f_namelen = EXT3_NAME_LEN;
 	return 0;
 }
@@ -2382,8 +2378,8 @@ static int ext3_statfs (struct super_block * sb, struct kstatfs * buf)
  * Process 1                         Process 2
  * ext3_create()                     quota_sync()
  *   journal_start()                   write_dquot()
- *   DQUOT_INIT()                        down(dqio_sem)
- *     down(dqio_sem)                    journal_start()
+ *   DQUOT_INIT()                        down(dqio_mutex)
+ *     down(dqio_mutex)                    journal_start()
  *
  */
 
diff --git a/fs/fat/cache.c b/fs/fat/cache.c
index 1acc941245fb..97b967b84fc6 100644
--- a/fs/fat/cache.c
+++ b/fs/fat/cache.c
@@ -49,7 +49,7 @@ int __init fat_cache_init(void)
 {
 	fat_cache_cachep = kmem_cache_create("fat_cache",
 				sizeof(struct fat_cache),
-				0, SLAB_RECLAIM_ACCOUNT,
+				0, SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
 				init_once, NULL);
 	if (fat_cache_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/fat/dir.c b/fs/fat/dir.c
index db0de5c621c7..698b85bb1dd4 100644
--- a/fs/fat/dir.c
+++ b/fs/fat/dir.c
@@ -114,7 +114,7 @@ static inline int fat_get_entry(struct inode *dir, loff_t *pos,
 }
 
 /*
- * Convert Unicode 16 to UTF8, translated Unicode, or ASCII.
+ * Convert Unicode 16 to UTF-8, translated Unicode, or ASCII.
  * If uni_xlate is enabled and we can't get a 1:1 conversion, use a
  * colon as an escape character since it is normally invalid on the vfat
  * filesystem. The following four characters are the hexadecimal digits
@@ -741,7 +741,7 @@ static int fat_dir_ioctl(struct inode * inode, struct file * filp,
 	return ret;
 }
 
-struct file_operations fat_dir_operations = {
+const struct file_operations fat_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= fat_readdir,
 	.ioctl		= fat_dir_ioctl,
diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c
index a1a9e0451217..ab171ea8e869 100644
--- a/fs/fat/fatent.c
+++ b/fs/fat/fatent.c
@@ -267,19 +267,19 @@ static struct fatent_operations fat32_ops = {
 
 static inline void lock_fat(struct msdos_sb_info *sbi)
 {
-	down(&sbi->fat_lock);
+	mutex_lock(&sbi->fat_lock);
 }
 
 static inline void unlock_fat(struct msdos_sb_info *sbi)
 {
-	up(&sbi->fat_lock);
+	mutex_unlock(&sbi->fat_lock);
 }
 
 void fat_ent_access_init(struct super_block *sb)
 {
 	struct msdos_sb_info *sbi = MSDOS_SB(sb);
 
-	init_MUTEX(&sbi->fat_lock);
+	mutex_init(&sbi->fat_lock);
 
 	switch (sbi->fat_bits) {
 	case 32:
diff --git a/fs/fat/file.c b/fs/fat/file.c
index 88aa1ae13f9f..1ee25232e6af 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -112,7 +112,7 @@ int fat_generic_ioctl(struct inode *inode, struct file *filp,
 	}
 }
 
-struct file_operations fat_file_operations = {
+const struct file_operations fat_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
diff --git a/fs/fat/inode.c b/fs/fat/inode.c
index e7f4aa7fc686..c1ce284f8a94 100644
--- a/fs/fat/inode.c
+++ b/fs/fat/inode.c
@@ -101,11 +101,11 @@ static int __fat_get_blocks(struct inode *inode, sector_t iblock,
 }
 
 static int fat_get_blocks(struct inode *inode, sector_t iblock,
-			  unsigned long max_blocks,
 			  struct buffer_head *bh_result, int create)
 {
 	struct super_block *sb = inode->i_sb;
 	int err;
+	unsigned long max_blocks = bh_result->b_size >> inode->i_blkbits;
 
 	err = __fat_get_blocks(inode, iblock, &max_blocks, bh_result, create);
 	if (err)
@@ -518,7 +518,8 @@ static int __init fat_init_inodecache(void)
 {
 	fat_inode_cachep = kmem_cache_create("fat_inode_cache",
 					     sizeof(struct msdos_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (fat_inode_cachep == NULL)
 		return -ENOMEM;
@@ -1101,7 +1102,7 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug,
 			return -EINVAL;
 		}
 	}
-	/* UTF8 doesn't provide FAT semantics */
+	/* UTF-8 doesn't provide FAT semantics */
 	if (!strcmp(opts->iocharset, "utf8")) {
 		printk(KERN_ERR "FAT: utf8 is not a recommended IO charset"
 		       " for FAT filesystems, filesystem will be case sensitive!\n");
@@ -1434,9 +1435,6 @@ out_fail:
 
 EXPORT_SYMBOL_GPL(fat_fill_super);
 
-int __init fat_cache_init(void);
-void fat_cache_destroy(void);
-
 static int __init init_fat_fs(void)
 {
 	int err;
diff --git a/fs/fcntl.c b/fs/fcntl.c
index dc4a7007f4e7..2a2479196f96 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -73,8 +73,8 @@ repeat:
 	 * orig_start..fdt->next_fd
 	 */
 	start = orig_start;
-	if (start < fdt->next_fd)
-		start = fdt->next_fd;
+	if (start < files->next_fd)
+		start = files->next_fd;
 
 	newfd = start;
 	if (start < fdt->max_fdset) {
@@ -102,9 +102,8 @@ repeat:
 	 * we reacquire the fdtable pointer and use it while holding
 	 * the lock, no one can free it during that time.
 	 */
-	fdt = files_fdtable(files);
-	if (start <= fdt->next_fd)
-		fdt->next_fd = newfd + 1;
+	if (start <= files->next_fd)
+		files->next_fd = newfd + 1;
 
 	error = newfd;
 	
@@ -413,7 +412,7 @@ out:
 
 /* Table to convert sigio signal codes into poll band bitmaps */
 
-static long band_table[NSIGPOLL] = {
+static const long band_table[NSIGPOLL] = {
 	POLLIN | POLLRDNORM,			/* POLL_IN */
 	POLLOUT | POLLWRNORM | POLLWRBAND,	/* POLL_OUT */
 	POLLIN | POLLRDNORM | POLLMSG,		/* POLL_MSG */
@@ -532,7 +531,7 @@ int send_sigurg(struct fown_struct *fown)
 }
 
 static DEFINE_RWLOCK(fasync_lock);
-static kmem_cache_t *fasync_cache;
+static kmem_cache_t *fasync_cache __read_mostly;
 
 /*
  * fasync_helper() is used by some character device drivers (mainly mice)
diff --git a/fs/fifo.c b/fs/fifo.c
index d13fcd3ec803..889f722ee36d 100644
--- a/fs/fifo.c
+++ b/fs/fifo.c
@@ -145,6 +145,6 @@ err_nocleanup:
  * is contain the open that then fills in the correct operations
  * depending on the access mode of the file...
  */
-struct file_operations def_fifo_fops = {
+const struct file_operations def_fifo_fops = {
 	.open		= fifo_open,	/* will set read or write pipe_fops */
 };
diff --git a/fs/file.c b/fs/file.c
index cea7cbea11d0..55f4e7022563 100644
--- a/fs/file.c
+++ b/fs/file.c
@@ -125,7 +125,8 @@ static void free_fdtable_rcu(struct rcu_head *rcu)
 		kmem_cache_free(files_cachep, fdt->free_files);
 		return;
 	}
-	if (fdt->max_fdset <= __FD_SETSIZE && fdt->max_fds <= NR_OPEN_DEFAULT) {
+	if (fdt->max_fdset <= EMBEDDED_FD_SET_SIZE &&
+		fdt->max_fds <= NR_OPEN_DEFAULT) {
 		/*
 		 * The fdtable was embedded
 		 */
@@ -155,8 +156,9 @@ static void free_fdtable_rcu(struct rcu_head *rcu)
 
 void free_fdtable(struct fdtable *fdt)
 {
-	if (fdt->free_files || fdt->max_fdset > __FD_SETSIZE ||
-					fdt->max_fds > NR_OPEN_DEFAULT)
+	if (fdt->free_files ||
+		fdt->max_fdset > EMBEDDED_FD_SET_SIZE ||
+		fdt->max_fds > NR_OPEN_DEFAULT)
 		call_rcu(&fdt->rcu, free_fdtable_rcu);
 }
 
@@ -199,7 +201,6 @@ static void copy_fdtable(struct fdtable *nfdt, struct fdtable *fdt)
 		       (nfdt->max_fds - fdt->max_fds) *
 					sizeof(struct file *));
 	}
-	nfdt->next_fd = fdt->next_fd;
 }
 
 /*
@@ -220,11 +221,9 @@ fd_set * alloc_fdset(int num)
 
 void free_fdset(fd_set *array, int num)
 {
-	int size = num / 8;
-
-	if (num <= __FD_SETSIZE) /* Don't free an embedded fdset */
+	if (num <= EMBEDDED_FD_SET_SIZE) /* Don't free an embedded fdset */
 		return;
-	else if (size <= PAGE_SIZE)
+	else if (num <= 8 * PAGE_SIZE)
 		kfree(array);
 	else
 		vfree(array);
@@ -237,22 +236,17 @@ static struct fdtable *alloc_fdtable(int nr)
   	fd_set *new_openset = NULL, *new_execset = NULL;
 	struct file **new_fds;
 
-	fdt = kmalloc(sizeof(*fdt), GFP_KERNEL);
+	fdt = kzalloc(sizeof(*fdt), GFP_KERNEL);
 	if (!fdt)
   		goto out;
-	memset(fdt, 0, sizeof(*fdt));
 
-	nfds = __FD_SETSIZE;
+	nfds = 8 * L1_CACHE_BYTES;
   	/* Expand to the max in easy steps */
-  	do {
-		if (nfds < (PAGE_SIZE * 8))
-			nfds = PAGE_SIZE * 8;
-		else {
-			nfds = nfds * 2;
-			if (nfds > NR_OPEN)
-				nfds = NR_OPEN;
-		}
-	} while (nfds <= nr);
+  	while (nfds <= nr) {
+		nfds = nfds * 2;
+		if (nfds > NR_OPEN)
+			nfds = NR_OPEN;
+	}
 
   	new_openset = alloc_fdset(nfds);
   	new_execset = alloc_fdset(nfds);
@@ -379,6 +373,6 @@ static void __devinit fdtable_defer_list_init(int cpu)
 void __init files_defer_init(void)
 {
 	int i;
-	for_each_cpu(i)
+	for_each_possible_cpu(i)
 		fdtable_defer_list_init(i);
 }
diff --git a/fs/file_table.c b/fs/file_table.c
index 44fabeaa9415..bcea1998b4de 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -88,6 +88,7 @@ int proc_nr_files(ctl_table *table, int write, struct file *filp,
  */
 struct file *get_empty_filp(void)
 {
+	struct task_struct *tsk;
 	static int old_max;
 	struct file * f;
 
@@ -112,13 +113,14 @@ struct file *get_empty_filp(void)
 	if (security_file_alloc(f))
 		goto fail_sec;
 
-	eventpoll_init_file(f);
+	tsk = current;
+	INIT_LIST_HEAD(&f->f_u.fu_list);
 	atomic_set(&f->f_count, 1);
-	f->f_uid = current->fsuid;
-	f->f_gid = current->fsgid;
 	rwlock_init(&f->f_owner.lock);
+	f->f_uid = tsk->fsuid;
+	f->f_gid = tsk->fsgid;
+	eventpoll_init_file(f);
 	/* f->f_version: 0 */
-	INIT_LIST_HEAD(&f->f_u.fu_list);
 	return f;
 
 over:
diff --git a/fs/freevxfs/vxfs_extern.h b/fs/freevxfs/vxfs_extern.h
index 927acf70c591..1cf1fe8466a2 100644
--- a/fs/freevxfs/vxfs_extern.h
+++ b/fs/freevxfs/vxfs_extern.h
@@ -63,7 +63,7 @@ extern void			vxfs_clear_inode(struct inode *);
 
 /* vxfs_lookup.c */
 extern struct inode_operations	vxfs_dir_inode_ops;
-extern struct file_operations	vxfs_dir_operations;
+extern const struct file_operations	vxfs_dir_operations;
 
 /* vxfs_olt.c */
 extern int			vxfs_read_olt(struct super_block *, u_long);
diff --git a/fs/freevxfs/vxfs_lookup.c b/fs/freevxfs/vxfs_lookup.c
index 554eb455722c..29cce456c7ce 100644
--- a/fs/freevxfs/vxfs_lookup.c
+++ b/fs/freevxfs/vxfs_lookup.c
@@ -56,7 +56,7 @@ struct inode_operations vxfs_dir_inode_ops = {
 	.lookup =		vxfs_lookup,
 };
 
-struct file_operations vxfs_dir_operations = {
+const struct file_operations vxfs_dir_operations = {
 	.readdir =		vxfs_readdir,
 };
 
diff --git a/fs/freevxfs/vxfs_super.c b/fs/freevxfs/vxfs_super.c
index 6aa6fbe4f8ee..b44c916d24a1 100644
--- a/fs/freevxfs/vxfs_super.c
+++ b/fs/freevxfs/vxfs_super.c
@@ -260,7 +260,7 @@ vxfs_init(void)
 {
 	vxfs_inode_cachep = kmem_cache_create("vxfs_inode",
 			sizeof(struct vxfs_inode_info), 0, 
-			SLAB_RECLAIM_ACCOUNT, NULL, NULL);
+			SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL, NULL);
 	if (vxfs_inode_cachep)
 		return register_filesystem(&vxfs_fs_type);
 	return -ENOMEM;
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index 785c7213a54f..f3fbe2d030f4 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -381,8 +381,8 @@ sync_sb_inodes(struct super_block *sb, struct writeback_control *wbc)
 			list_move(&inode->i_list, &sb->s_dirty);
 		}
 		spin_unlock(&inode_lock);
-		cond_resched();
 		iput(inode);
+		cond_resched();
 		spin_lock(&inode_lock);
 		if (wbc->nr_to_write <= 0)
 			break;
diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c
index 0c9a2ee54c91..23d1f52eb1b8 100644
--- a/fs/fuse/dev.c
+++ b/fs/fuse/dev.c
@@ -922,7 +922,7 @@ static int fuse_dev_release(struct inode *inode, struct file *file)
 	return 0;
 }
 
-struct file_operations fuse_dev_operations = {
+const struct file_operations fuse_dev_operations = {
 	.owner		= THIS_MODULE,
 	.llseek		= no_llseek,
 	.read		= fuse_dev_read,
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index c72a8a97935c..256355b80256 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -1170,7 +1170,7 @@ static struct inode_operations fuse_dir_inode_operations = {
 	.removexattr	= fuse_removexattr,
 };
 
-static struct file_operations fuse_dir_operations = {
+static const struct file_operations fuse_dir_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_read_dir,
 	.readdir	= fuse_readdir,
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 6f05379b0a0d..975f2697e866 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -12,7 +12,7 @@
 #include <linux/slab.h>
 #include <linux/kernel.h>
 
-static struct file_operations fuse_direct_io_file_operations;
+static const struct file_operations fuse_direct_io_file_operations;
 
 static int fuse_send_open(struct inode *inode, struct file *file, int isdir,
 			  struct fuse_open_out *outargp)
@@ -611,7 +611,7 @@ static int fuse_set_page_dirty(struct page *page)
 	return 0;
 }
 
-static struct file_operations fuse_file_operations = {
+static const struct file_operations fuse_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
@@ -623,7 +623,7 @@ static struct file_operations fuse_file_operations = {
 	.sendfile	= generic_file_sendfile,
 };
 
-static struct file_operations fuse_direct_io_file_operations = {
+static const struct file_operations fuse_direct_io_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= fuse_direct_read,
 	.write		= fuse_direct_write,
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 4a83adfec968..a16a04fcf41e 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -346,7 +346,7 @@ static inline u64 get_node_id(struct inode *inode)
 }
 
 /** Device operations */
-extern struct file_operations fuse_dev_operations;
+extern const struct file_operations fuse_dev_operations;
 
 /**
  * This is the single global spinlock which protects FUSE's structures
diff --git a/fs/hfs/bnode.c b/fs/hfs/bnode.c
index a7a7d77f3fd3..1e44dcfe49c4 100644
--- a/fs/hfs/bnode.c
+++ b/fs/hfs/bnode.c
@@ -306,8 +306,7 @@ void hfs_bnode_unhash(struct hfs_bnode *node)
 	for (p = &node->tree->node_hash[hfs_bnode_hash(node->this)];
 	     *p && *p != node; p = &(*p)->next_hash)
 		;
-	if (!*p)
-		BUG();
+	BUG_ON(!*p);
 	*p = node->next_hash;
 	node->tree->node_hash_cnt--;
 }
@@ -415,8 +414,7 @@ struct hfs_bnode *hfs_bnode_create(struct hfs_btree *tree, u32 num)
 	spin_lock(&tree->hash_lock);
 	node = hfs_bnode_findhash(tree, num);
 	spin_unlock(&tree->hash_lock);
-	if (node)
-		BUG();
+	BUG_ON(node);
 	node = __hfs_bnode_create(tree, num);
 	if (!node)
 		return ERR_PTR(-ENOMEM);
@@ -459,8 +457,7 @@ void hfs_bnode_put(struct hfs_bnode *node)
 
 		dprint(DBG_BNODE_REFS, "put_node(%d:%d): %d\n",
 		       node->tree->cnid, node->this, atomic_read(&node->refcnt));
-		if (!atomic_read(&node->refcnt))
-			BUG();
+		BUG_ON(!atomic_read(&node->refcnt));
 		if (!atomic_dec_and_lock(&node->refcnt, &tree->hash_lock))
 			return;
 		for (i = 0; i < tree->pages_per_bnode; i++) {
diff --git a/fs/hfs/btree.c b/fs/hfs/btree.c
index 7bb11edd1488..d20131ce4b95 100644
--- a/fs/hfs/btree.c
+++ b/fs/hfs/btree.c
@@ -36,8 +36,7 @@ struct hfs_btree *hfs_btree_open(struct super_block *sb, u32 id, btree_keycmp ke
 	tree->inode = iget_locked(sb, id);
 	if (!tree->inode)
 		goto free_tree;
-	if (!(tree->inode->i_state & I_NEW))
-		BUG();
+	BUG_ON(!(tree->inode->i_state & I_NEW));
 	{
 	struct hfs_mdb *mdb = HFS_SB(sb)->mdb;
 	HFS_I(tree->inode)->flags = 0;
diff --git a/fs/hfs/dir.c b/fs/hfs/dir.c
index 534e5a7480ef..7cd8cc03aea7 100644
--- a/fs/hfs/dir.c
+++ b/fs/hfs/dir.c
@@ -313,7 +313,7 @@ static int hfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	return res;
 }
 
-struct file_operations hfs_dir_operations = {
+const struct file_operations hfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= hfs_readdir,
 	.llseek		= generic_file_llseek,
diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h
index 18ce47ab1b71..3ed8663a8db1 100644
--- a/fs/hfs/hfs_fs.h
+++ b/fs/hfs/hfs_fs.h
@@ -169,7 +169,7 @@ extern int hfs_cat_move(u32, struct inode *, struct qstr *,
 extern void hfs_cat_build_key(struct super_block *, btree_key *, u32, struct qstr *);
 
 /* dir.c */
-extern struct file_operations hfs_dir_operations;
+extern const struct file_operations hfs_dir_operations;
 extern struct inode_operations hfs_dir_inode_operations;
 
 /* extent.c */
diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c
index 39fd85b9b916..2d4ced22201b 100644
--- a/fs/hfs/inode.c
+++ b/fs/hfs/inode.c
@@ -17,7 +17,7 @@
 #include "hfs_fs.h"
 #include "btree.h"
 
-static struct file_operations hfs_file_operations;
+static const struct file_operations hfs_file_operations;
 static struct inode_operations hfs_file_inode_operations;
 
 /*================ Variable-like macros ================*/
@@ -98,17 +98,6 @@ static int hfs_releasepage(struct page *page, gfp_t mask)
 	return res ? try_to_free_buffers(page) : 0;
 }
 
-static int hfs_get_blocks(struct inode *inode, sector_t iblock, unsigned long max_blocks,
-			  struct buffer_head *bh_result, int create)
-{
-	int ret;
-
-	ret = hfs_get_block(inode, iblock, bh_result, create);
-	if (!ret)
-		bh_result->b_size = (1 << inode->i_blkbits);
-	return ret;
-}
-
 static ssize_t hfs_direct_IO(int rw, struct kiocb *iocb,
 		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
 {
@@ -116,7 +105,7 @@ static ssize_t hfs_direct_IO(int rw, struct kiocb *iocb,
 	struct inode *inode = file->f_dentry->d_inode->i_mapping->host;
 
 	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
-				  offset, nr_segs, hfs_get_blocks, NULL);
+				  offset, nr_segs, hfs_get_block, NULL);
 }
 
 static int hfs_writepages(struct address_space *mapping,
@@ -612,7 +601,7 @@ int hfs_inode_setattr(struct dentry *dentry, struct iattr * attr)
 }
 
 
-static struct file_operations hfs_file_operations = {
+static const struct file_operations hfs_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c
index 01a6fe3a395c..1f9ece0de326 100644
--- a/fs/hfsplus/dir.c
+++ b/fs/hfsplus/dir.c
@@ -483,7 +483,7 @@ struct inode_operations hfsplus_dir_inode_operations = {
 	.rename		= hfsplus_rename,
 };
 
-struct file_operations hfsplus_dir_operations = {
+const struct file_operations hfsplus_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= hfsplus_readdir,
 	.ioctl          = hfsplus_ioctl,
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index 12ed2b7d046b..acf66dba3e01 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -93,17 +93,6 @@ static int hfsplus_releasepage(struct page *page, gfp_t mask)
 	return res ? try_to_free_buffers(page) : 0;
 }
 
-static int hfsplus_get_blocks(struct inode *inode, sector_t iblock, unsigned long max_blocks,
-			      struct buffer_head *bh_result, int create)
-{
-	int ret;
-
-	ret = hfsplus_get_block(inode, iblock, bh_result, create);
-	if (!ret)
-		bh_result->b_size = (1 << inode->i_blkbits);
-	return ret;
-}
-
 static ssize_t hfsplus_direct_IO(int rw, struct kiocb *iocb,
 		const struct iovec *iov, loff_t offset, unsigned long nr_segs)
 {
@@ -111,7 +100,7 @@ static ssize_t hfsplus_direct_IO(int rw, struct kiocb *iocb,
 	struct inode *inode = file->f_dentry->d_inode->i_mapping->host;
 
 	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
-				  offset, nr_segs, hfsplus_get_blocks, NULL);
+				  offset, nr_segs, hfsplus_get_block, NULL);
 }
 
 static int hfsplus_writepages(struct address_space *mapping,
@@ -291,7 +280,7 @@ static struct inode_operations hfsplus_file_inode_operations = {
 	.listxattr	= hfsplus_listxattr,
 };
 
-static struct file_operations hfsplus_file_operations = {
+static const struct file_operations hfsplus_file_operations = {
 	.llseek 	= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c
index b3ad0bd0312f..bf0f8e16e433 100644
--- a/fs/hostfs/hostfs_kern.c
+++ b/fs/hostfs/hostfs_kern.c
@@ -384,7 +384,7 @@ int hostfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 	return fsync_file(HOSTFS_I(dentry->d_inode)->fd, datasync);
 }
 
-static struct file_operations hostfs_file_fops = {
+static const struct file_operations hostfs_file_fops = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.sendfile	= generic_file_sendfile,
@@ -399,7 +399,7 @@ static struct file_operations hostfs_file_fops = {
 	.fsync		= hostfs_fsync,
 };
 
-static struct file_operations hostfs_dir_fops = {
+static const struct file_operations hostfs_dir_fops = {
 	.llseek		= generic_file_llseek,
 	.readdir	= hostfs_readdir,
 	.read		= generic_read_dir,
diff --git a/fs/hostfs/hostfs_user.c b/fs/hostfs/hostfs_user.c
index b97809deba66..23b7cee72123 100644
--- a/fs/hostfs/hostfs_user.c
+++ b/fs/hostfs/hostfs_user.c
@@ -360,7 +360,6 @@ int do_statfs(char *root, long *bsize_out, long long *blocks_out,
 	spare_out[2] = buf.f_spare[2];
 	spare_out[3] = buf.f_spare[3];
 	spare_out[4] = buf.f_spare[4];
-	spare_out[5] = buf.f_spare[5];
 	return(0);
 }
 
diff --git a/fs/hpfs/dir.c b/fs/hpfs/dir.c
index 5591f9623aa2..ecc9180645ae 100644
--- a/fs/hpfs/dir.c
+++ b/fs/hpfs/dir.c
@@ -310,7 +310,7 @@ struct dentry *hpfs_lookup(struct inode *dir, struct dentry *dentry, struct name
 	return ERR_PTR(-ENOENT);
 }
 
-struct file_operations hpfs_dir_ops =
+const struct file_operations hpfs_dir_ops =
 {
 	.llseek		= hpfs_dir_lseek,
 	.read		= generic_read_dir,
diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c
index 7c995ac4081b..d3b9fffe45a1 100644
--- a/fs/hpfs/file.c
+++ b/fs/hpfs/file.c
@@ -119,7 +119,7 @@ static ssize_t hpfs_file_write(struct file *file, const char __user *buf,
 	return retval;
 }
 
-struct file_operations hpfs_file_ops =
+const struct file_operations hpfs_file_ops =
 {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
diff --git a/fs/hpfs/hpfs_fn.h b/fs/hpfs/hpfs_fn.h
index 6628c3b352cb..29b7a3e55173 100644
--- a/fs/hpfs/hpfs_fn.h
+++ b/fs/hpfs/hpfs_fn.h
@@ -9,6 +9,7 @@
 //#define DBG
 //#define DEBUG_LOCKS
 
+#include <linux/mutex.h>
 #include <linux/pagemap.h>
 #include <linux/buffer_head.h>
 #include <linux/hpfs_fs.h>
@@ -57,8 +58,8 @@ struct hpfs_inode_info {
 	unsigned i_ea_uid : 1;	/* file's uid is stored in ea */
 	unsigned i_ea_gid : 1;	/* file's gid is stored in ea */
 	unsigned i_dirty : 1;
-	struct semaphore i_sem;
-	struct semaphore i_parent;
+	struct mutex i_mutex;
+	struct mutex i_parent_mutex;
 	loff_t **i_rddir_off;
 	struct inode vfs_inode;
 };
@@ -239,7 +240,7 @@ void hpfs_set_dentry_operations(struct dentry *);
 /* dir.c */
 
 struct dentry *hpfs_lookup(struct inode *, struct dentry *, struct nameidata *);
-extern struct file_operations hpfs_dir_ops;
+extern const struct file_operations hpfs_dir_ops;
 
 /* dnode.c */
 
@@ -265,7 +266,7 @@ void hpfs_set_ea(struct inode *, struct fnode *, char *, char *, int);
 /* file.c */
 
 int hpfs_file_fsync(struct file *, struct dentry *, int);
-extern struct file_operations hpfs_file_ops;
+extern const struct file_operations hpfs_file_ops;
 extern struct inode_operations hpfs_file_iops;
 extern struct address_space_operations hpfs_aops;
 
diff --git a/fs/hpfs/inode.c b/fs/hpfs/inode.c
index e3d17e9ea6c1..56f2c338c4d9 100644
--- a/fs/hpfs/inode.c
+++ b/fs/hpfs/inode.c
@@ -186,9 +186,9 @@ void hpfs_write_inode(struct inode *i)
 		kfree(hpfs_inode->i_rddir_off);
 		hpfs_inode->i_rddir_off = NULL;
 	}
-	down(&hpfs_inode->i_parent);
+	mutex_lock(&hpfs_inode->i_parent_mutex);
 	if (!i->i_nlink) {
-		up(&hpfs_inode->i_parent);
+		mutex_unlock(&hpfs_inode->i_parent_mutex);
 		return;
 	}
 	parent = iget_locked(i->i_sb, hpfs_inode->i_parent_dir);
@@ -199,14 +199,14 @@ void hpfs_write_inode(struct inode *i)
 			hpfs_read_inode(parent);
 			unlock_new_inode(parent);
 		}
-		down(&hpfs_inode->i_sem);
+		mutex_lock(&hpfs_inode->i_mutex);
 		hpfs_write_inode_nolock(i);
-		up(&hpfs_inode->i_sem);
+		mutex_unlock(&hpfs_inode->i_mutex);
 		iput(parent);
 	} else {
 		mark_inode_dirty(i);
 	}
-	up(&hpfs_inode->i_parent);
+	mutex_unlock(&hpfs_inode->i_parent_mutex);
 }
 
 void hpfs_write_inode_nolock(struct inode *i)
diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c
index 8ff8fc433fc1..a03abb12c610 100644
--- a/fs/hpfs/namei.c
+++ b/fs/hpfs/namei.c
@@ -60,7 +60,7 @@ static int hpfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	if (dee.read_only)
 		result->i_mode &= ~0222;
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail3;
@@ -101,11 +101,11 @@ static int hpfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 		hpfs_write_inode_nolock(result);
 	}
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	unlock_kernel();
 	return 0;
 bail3:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail2:
 	hpfs_brelse4(&qbh0);
@@ -168,7 +168,7 @@ static int hpfs_create(struct inode *dir, struct dentry *dentry, int mode, struc
 	result->i_data.a_ops = &hpfs_aops;
 	hpfs_i(result)->mmu_private = 0;
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail2;
@@ -193,12 +193,12 @@ static int hpfs_create(struct inode *dir, struct dentry *dentry, int mode, struc
 		hpfs_write_inode_nolock(result);
 	}
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	unlock_kernel();
 	return 0;
 
 bail2:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail1:
 	brelse(bh);
@@ -254,7 +254,7 @@ static int hpfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t
 	result->i_blocks = 1;
 	init_special_inode(result, mode, rdev);
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail2;
@@ -271,12 +271,12 @@ static int hpfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t
 
 	hpfs_write_inode_nolock(result);
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	brelse(bh);
 	unlock_kernel();
 	return 0;
 bail2:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail1:
 	brelse(bh);
@@ -333,7 +333,7 @@ static int hpfs_symlink(struct inode *dir, struct dentry *dentry, const char *sy
 	result->i_op = &page_symlink_inode_operations;
 	result->i_data.a_ops = &hpfs_symlink_aops;
 
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	r = hpfs_add_dirent(dir, (char *)name, len, &dee, 0);
 	if (r == 1)
 		goto bail2;
@@ -352,11 +352,11 @@ static int hpfs_symlink(struct inode *dir, struct dentry *dentry, const char *sy
 
 	hpfs_write_inode_nolock(result);
 	d_instantiate(dentry, result);
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	unlock_kernel();
 	return 0;
 bail2:
-	up(&hpfs_i(dir)->i_sem);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
 	iput(result);
 bail1:
 	brelse(bh);
@@ -382,8 +382,8 @@ static int hpfs_unlink(struct inode *dir, struct dentry *dentry)
 	lock_kernel();
 	hpfs_adjust_length((char *)name, &len);
 again:
-	down(&hpfs_i(inode)->i_parent);
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(inode)->i_parent_mutex);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	err = -ENOENT;
 	de = map_dirent(dir, hpfs_i(dir)->i_dno, (char *)name, len, &dno, &qbh);
 	if (!de)
@@ -410,8 +410,8 @@ again:
 		if (rep++)
 			break;
 
-		up(&hpfs_i(dir)->i_sem);
-		up(&hpfs_i(inode)->i_parent);
+		mutex_unlock(&hpfs_i(dir)->i_mutex);
+		mutex_unlock(&hpfs_i(inode)->i_parent_mutex);
 		d_drop(dentry);
 		spin_lock(&dentry->d_lock);
 		if (atomic_read(&dentry->d_count) > 1 ||
@@ -442,8 +442,8 @@ again:
 out1:
 	hpfs_brelse4(&qbh);
 out:
-	up(&hpfs_i(dir)->i_sem);
-	up(&hpfs_i(inode)->i_parent);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
+	mutex_unlock(&hpfs_i(inode)->i_parent_mutex);
 	unlock_kernel();
 	return err;
 }
@@ -463,8 +463,8 @@ static int hpfs_rmdir(struct inode *dir, struct dentry *dentry)
 
 	hpfs_adjust_length((char *)name, &len);
 	lock_kernel();
-	down(&hpfs_i(inode)->i_parent);
-	down(&hpfs_i(dir)->i_sem);
+	mutex_lock(&hpfs_i(inode)->i_parent_mutex);
+	mutex_lock(&hpfs_i(dir)->i_mutex);
 	err = -ENOENT;
 	de = map_dirent(dir, hpfs_i(dir)->i_dno, (char *)name, len, &dno, &qbh);
 	if (!de)
@@ -502,8 +502,8 @@ static int hpfs_rmdir(struct inode *dir, struct dentry *dentry)
 out1:
 	hpfs_brelse4(&qbh);
 out:
-	up(&hpfs_i(dir)->i_sem);
-	up(&hpfs_i(inode)->i_parent);
+	mutex_unlock(&hpfs_i(dir)->i_mutex);
+	mutex_unlock(&hpfs_i(inode)->i_parent_mutex);
 	unlock_kernel();
 	return err;
 }
@@ -565,12 +565,12 @@ static int hpfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 
 	lock_kernel();
 	/* order doesn't matter, due to VFS exclusion */
-	down(&hpfs_i(i)->i_parent);
+	mutex_lock(&hpfs_i(i)->i_parent_mutex);
 	if (new_inode)
-		down(&hpfs_i(new_inode)->i_parent);
-	down(&hpfs_i(old_dir)->i_sem);
+		mutex_lock(&hpfs_i(new_inode)->i_parent_mutex);
+	mutex_lock(&hpfs_i(old_dir)->i_mutex);
 	if (new_dir != old_dir)
-		down(&hpfs_i(new_dir)->i_sem);
+		mutex_lock(&hpfs_i(new_dir)->i_mutex);
 	
 	/* Erm? Moving over the empty non-busy directory is perfectly legal */
 	if (new_inode && S_ISDIR(new_inode->i_mode)) {
@@ -650,11 +650,11 @@ static int hpfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	hpfs_decide_conv(i, (char *)new_name, new_len);
 end1:
 	if (old_dir != new_dir)
-		up(&hpfs_i(new_dir)->i_sem);
-	up(&hpfs_i(old_dir)->i_sem);
-	up(&hpfs_i(i)->i_parent);
+		mutex_unlock(&hpfs_i(new_dir)->i_mutex);
+	mutex_unlock(&hpfs_i(old_dir)->i_mutex);
+	mutex_unlock(&hpfs_i(i)->i_parent_mutex);
 	if (new_inode)
-		up(&hpfs_i(new_inode)->i_parent);
+		mutex_unlock(&hpfs_i(new_inode)->i_parent_mutex);
 	unlock_kernel();
 	return err;
 }
diff --git a/fs/hpfs/super.c b/fs/hpfs/super.c
index 63e88d7e2c3b..d72d8c87c996 100644
--- a/fs/hpfs/super.c
+++ b/fs/hpfs/super.c
@@ -181,8 +181,8 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 
 	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
 	    SLAB_CTOR_CONSTRUCTOR) {
-		init_MUTEX(&ei->i_sem);
-		init_MUTEX(&ei->i_parent);
+		mutex_init(&ei->i_mutex);
+		mutex_init(&ei->i_parent_mutex);
 		inode_init_once(&ei->vfs_inode);
 	}
 }
@@ -191,7 +191,8 @@ static int init_inodecache(void)
 {
 	hpfs_inode_cachep = kmem_cache_create("hpfs_inode_cache",
 					     sizeof(struct hpfs_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (hpfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/hppfs/hppfs_kern.c b/fs/hppfs/hppfs_kern.c
index a44dc5897399..2ba20cdb5baa 100644
--- a/fs/hppfs/hppfs_kern.c
+++ b/fs/hppfs/hppfs_kern.c
@@ -558,7 +558,7 @@ static loff_t hppfs_llseek(struct file *file, loff_t off, int where)
 	return(default_llseek(file, off, where));
 }
 
-static struct file_operations hppfs_file_fops = {
+static const struct file_operations hppfs_file_fops = {
 	.owner		= NULL,
 	.llseek		= hppfs_llseek,
 	.read		= hppfs_read,
@@ -609,7 +609,7 @@ static int hppfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 	return(0);
 }
 
-static struct file_operations hppfs_dir_fops = {
+static const struct file_operations hppfs_dir_fops = {
 	.owner		= NULL,
 	.readdir	= hppfs_readdir,
 	.open		= hppfs_dir_open,
diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index b35195289945..3a5b4e923455 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -35,7 +35,7 @@
 
 static struct super_operations hugetlbfs_ops;
 static struct address_space_operations hugetlbfs_aops;
-struct file_operations hugetlbfs_file_operations;
+const struct file_operations hugetlbfs_file_operations;
 static struct inode_operations hugetlbfs_dir_inode_operations;
 static struct inode_operations hugetlbfs_inode_operations;
 
@@ -56,48 +56,10 @@ static void huge_pagevec_release(struct pagevec *pvec)
 	pagevec_reinit(pvec);
 }
 
-/*
- * huge_pages_needed tries to determine the number of new huge pages that
- * will be required to fully populate this VMA.  This will be equal to
- * the size of the VMA in huge pages minus the number of huge pages
- * (covered by this VMA) that are found in the page cache.
- *
- * Result is in bytes to be compatible with is_hugepage_mem_enough()
- */
-static unsigned long
-huge_pages_needed(struct address_space *mapping, struct vm_area_struct *vma)
-{
-	int i;
-	struct pagevec pvec;
-	unsigned long start = vma->vm_start;
-	unsigned long end = vma->vm_end;
-	unsigned long hugepages = (end - start) >> HPAGE_SHIFT;
-	pgoff_t next = vma->vm_pgoff >> (HPAGE_SHIFT - PAGE_SHIFT);
-	pgoff_t endpg = next + hugepages;
-
-	pagevec_init(&pvec, 0);
-	while (next < endpg) {
-		if (!pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE))
-			break;
-		for (i = 0; i < pagevec_count(&pvec); i++) {
-			struct page *page = pvec.pages[i];
-			if (page->index > next)
-				next = page->index;
-			if (page->index >= endpg)
-				break;
-			next++;
-			hugepages--;
-		}
-		huge_pagevec_release(&pvec);
-	}
-	return hugepages << HPAGE_SHIFT;
-}
-
 static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 {
 	struct inode *inode = file->f_dentry->d_inode;
-	struct address_space *mapping = inode->i_mapping;
-	unsigned long bytes;
+	struct hugetlbfs_inode_info *info = HUGETLBFS_I(inode);
 	loff_t len, vma_len;
 	int ret;
 
@@ -113,10 +75,6 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	if (vma->vm_end - vma->vm_start < HPAGE_SIZE)
 		return -EINVAL;
 
-	bytes = huge_pages_needed(mapping, vma);
-	if (!is_hugepage_mem_enough(bytes))
-		return -ENOMEM;
-
 	vma_len = (loff_t)(vma->vm_end - vma->vm_start);
 
 	mutex_lock(&inode->i_mutex);
@@ -129,6 +87,10 @@ static int hugetlbfs_file_mmap(struct file *file, struct vm_area_struct *vma)
 	if (!(vma->vm_flags & VM_WRITE) && len > inode->i_size)
 		goto out;
 
+	if (vma->vm_flags & VM_MAYSHARE)
+		if (hugetlb_extend_reservation(info, len >> HPAGE_SHIFT) != 0)
+			goto out;
+
 	ret = 0;
 	hugetlb_prefault_arch_hook(vma->vm_mm);
 	if (inode->i_size < len)
@@ -227,13 +189,18 @@ static void truncate_huge_page(struct page *page)
 	put_page(page);
 }
 
-static void truncate_hugepages(struct address_space *mapping, loff_t lstart)
+static void truncate_hugepages(struct inode *inode, loff_t lstart)
 {
+	struct address_space *mapping = &inode->i_data;
 	const pgoff_t start = lstart >> HPAGE_SHIFT;
 	struct pagevec pvec;
 	pgoff_t next;
 	int i;
 
+	hugetlb_truncate_reservation(HUGETLBFS_I(inode),
+				     lstart >> HPAGE_SHIFT);
+	if (!mapping->nrpages)
+		return;
 	pagevec_init(&pvec, 0);
 	next = start;
 	while (1) {
@@ -262,8 +229,7 @@ static void truncate_hugepages(struct address_space *mapping, loff_t lstart)
 
 static void hugetlbfs_delete_inode(struct inode *inode)
 {
-	if (inode->i_data.nrpages)
-		truncate_hugepages(&inode->i_data, 0);
+	truncate_hugepages(inode, 0);
 	clear_inode(inode);
 }
 
@@ -296,8 +262,7 @@ static void hugetlbfs_forget_inode(struct inode *inode)
 	inode->i_state |= I_FREEING;
 	inodes_stat.nr_inodes--;
 	spin_unlock(&inode_lock);
-	if (inode->i_data.nrpages)
-		truncate_hugepages(&inode->i_data, 0);
+	truncate_hugepages(inode, 0);
 	clear_inode(inode);
 	destroy_inode(inode);
 }
@@ -356,7 +321,7 @@ static int hugetlb_vmtruncate(struct inode *inode, loff_t offset)
 	if (!prio_tree_empty(&mapping->i_mmap))
 		hugetlb_vmtruncate_list(&mapping->i_mmap, pgoff);
 	spin_unlock(&mapping->i_mmap_lock);
-	truncate_hugepages(mapping, offset);
+	truncate_hugepages(inode, offset);
 	return 0;
 }
 
@@ -573,6 +538,7 @@ static struct inode *hugetlbfs_alloc_inode(struct super_block *sb)
 		hugetlbfs_inc_free_inodes(sbinfo);
 		return NULL;
 	}
+	p->prereserved_hpages = 0;
 	return &p->vfs_inode;
 }
 
@@ -600,7 +566,7 @@ static void init_once(void *foo, kmem_cache_t *cachep, unsigned long flags)
 		inode_init_once(&ei->vfs_inode);
 }
 
-struct file_operations hugetlbfs_file_operations = {
+const struct file_operations hugetlbfs_file_operations = {
 	.mmap			= hugetlbfs_file_mmap,
 	.fsync			= simple_sync_file,
 	.get_unmapped_area	= hugetlb_get_unmapped_area,
@@ -771,21 +737,6 @@ static struct file_system_type hugetlbfs_fs_type = {
 
 static struct vfsmount *hugetlbfs_vfsmount;
 
-/*
- * Return the next identifier for a shm file
- */
-static unsigned long hugetlbfs_counter(void)
-{
-	static DEFINE_SPINLOCK(lock);
-	static unsigned long counter;
-	unsigned long ret;
-
-	spin_lock(&lock);
-	ret = ++counter;
-	spin_unlock(&lock);
-	return ret;
-}
-
 static int can_do_hugetlb_shm(void)
 {
 	return likely(capable(CAP_IPC_LOCK) ||
@@ -801,18 +752,16 @@ struct file *hugetlb_zero_setup(size_t size)
 	struct dentry *dentry, *root;
 	struct qstr quick_string;
 	char buf[16];
+	static atomic_t counter;
 
 	if (!can_do_hugetlb_shm())
 		return ERR_PTR(-EPERM);
 
-	if (!is_hugepage_mem_enough(size))
-		return ERR_PTR(-ENOMEM);
-
 	if (!user_shm_lock(size, current->user))
 		return ERR_PTR(-ENOMEM);
 
 	root = hugetlbfs_vfsmount->mnt_root;
-	snprintf(buf, 16, "%lu", hugetlbfs_counter());
+	snprintf(buf, 16, "%u", atomic_inc_return(&counter));
 	quick_string.name = buf;
 	quick_string.len = strlen(quick_string.name);
 	quick_string.hash = 0;
@@ -831,6 +780,11 @@ struct file *hugetlb_zero_setup(size_t size)
 	if (!inode)
 		goto out_file;
 
+	error = -ENOMEM;
+	if (hugetlb_extend_reservation(HUGETLBFS_I(inode),
+				       size >> HPAGE_SHIFT) != 0)
+		goto out_inode;
+
 	d_instantiate(dentry, inode);
 	inode->i_size = size;
 	inode->i_nlink = 0;
@@ -841,6 +795,8 @@ struct file *hugetlb_zero_setup(size_t size)
 	file->f_mode = FMODE_WRITE | FMODE_READ;
 	return file;
 
+out_inode:
+	iput(inode);
 out_file:
 	put_filp(file);
 out_dentry:
diff --git a/fs/inode.c b/fs/inode.c
index d0be6159eb7f..32b7c3375021 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -56,8 +56,8 @@
 #define I_HASHBITS	i_hash_shift
 #define I_HASHMASK	i_hash_mask
 
-static unsigned int i_hash_mask;
-static unsigned int i_hash_shift;
+static unsigned int i_hash_mask __read_mostly;
+static unsigned int i_hash_shift __read_mostly;
 
 /*
  * Each inode can be on two separate lists. One is
@@ -73,7 +73,7 @@ static unsigned int i_hash_shift;
 
 LIST_HEAD(inode_in_use);
 LIST_HEAD(inode_unused);
-static struct hlist_head *inode_hashtable;
+static struct hlist_head *inode_hashtable __read_mostly;
 
 /*
  * A simple spinlock to protect the list manipulations.
@@ -84,27 +84,27 @@ static struct hlist_head *inode_hashtable;
 DEFINE_SPINLOCK(inode_lock);
 
 /*
- * iprune_sem provides exclusion between the kswapd or try_to_free_pages
+ * iprune_mutex provides exclusion between the kswapd or try_to_free_pages
  * icache shrinking path, and the umount path.  Without this exclusion,
  * by the time prune_icache calls iput for the inode whose pages it has
  * been invalidating, or by the time it calls clear_inode & destroy_inode
  * from its final dispose_list, the struct super_block they refer to
  * (for inode->i_sb->s_op) may already have been freed and reused.
  */
-DECLARE_MUTEX(iprune_sem);
+static DEFINE_MUTEX(iprune_mutex);
 
 /*
  * Statistics gathering..
  */
 struct inodes_stat_t inodes_stat;
 
-static kmem_cache_t * inode_cachep;
+static kmem_cache_t * inode_cachep __read_mostly;
 
 static struct inode *alloc_inode(struct super_block *sb)
 {
 	static struct address_space_operations empty_aops;
 	static struct inode_operations empty_iops;
-	static struct file_operations empty_fops;
+	static const struct file_operations empty_fops;
 	struct inode *inode;
 
 	if (sb->s_op->alloc_inode)
@@ -206,7 +206,7 @@ void inode_init_once(struct inode *inode)
 	i_size_ordered_init(inode);
 #ifdef CONFIG_INOTIFY
 	INIT_LIST_HEAD(&inode->inotify_watches);
-	sema_init(&inode->inotify_sem, 1);
+	mutex_init(&inode->inotify_mutex);
 #endif
 }
 
@@ -319,7 +319,7 @@ static int invalidate_list(struct list_head *head, struct list_head *dispose)
 		/*
 		 * We can reschedule here without worrying about the list's
 		 * consistency because the per-sb list of inodes must not
-		 * change during umount anymore, and because iprune_sem keeps
+		 * change during umount anymore, and because iprune_mutex keeps
 		 * shrink_icache_memory() away.
 		 */
 		cond_resched_lock(&inode_lock);
@@ -355,14 +355,14 @@ int invalidate_inodes(struct super_block * sb)
 	int busy;
 	LIST_HEAD(throw_away);
 
-	down(&iprune_sem);
+	mutex_lock(&iprune_mutex);
 	spin_lock(&inode_lock);
 	inotify_unmount_inodes(&sb->s_inodes);
 	busy = invalidate_list(&sb->s_inodes, &throw_away);
 	spin_unlock(&inode_lock);
 
 	dispose_list(&throw_away);
-	up(&iprune_sem);
+	mutex_unlock(&iprune_mutex);
 
 	return busy;
 }
@@ -377,7 +377,7 @@ int __invalidate_device(struct block_device *bdev)
 	if (sb) {
 		/*
 		 * no need to lock the super, get_super holds the
-		 * read semaphore so the filesystem cannot go away
+		 * read mutex so the filesystem cannot go away
 		 * under us (->put_super runs with the write lock
 		 * hold).
 		 */
@@ -423,7 +423,7 @@ static void prune_icache(int nr_to_scan)
 	int nr_scanned;
 	unsigned long reap = 0;
 
-	down(&iprune_sem);
+	mutex_lock(&iprune_mutex);
 	spin_lock(&inode_lock);
 	for (nr_scanned = 0; nr_scanned < nr_to_scan; nr_scanned++) {
 		struct inode *inode;
@@ -459,7 +459,7 @@ static void prune_icache(int nr_to_scan)
 	spin_unlock(&inode_lock);
 
 	dispose_list(&freeable);
-	up(&iprune_sem);
+	mutex_unlock(&iprune_mutex);
 
 	if (current_is_kswapd())
 		mod_page_state(kswapd_inodesteal, reap);
@@ -1375,8 +1375,13 @@ void __init inode_init(unsigned long mempages)
 	int loop;
 
 	/* inode slab cache */
-	inode_cachep = kmem_cache_create("inode_cache", sizeof(struct inode),
-				0, SLAB_RECLAIM_ACCOUNT|SLAB_PANIC, init_once, NULL);
+	inode_cachep = kmem_cache_create("inode_cache",
+					 sizeof(struct inode),
+					 0,
+					 (SLAB_RECLAIM_ACCOUNT|SLAB_PANIC|
+					 SLAB_MEM_SPREAD),
+					 init_once,
+					 NULL);
 	set_shrinker(DEFAULT_SEEKS, shrink_icache_memory);
 
 	/* Hash may have been set up in inode_init_early */
diff --git a/fs/inotify.c b/fs/inotify.c
index 3041503bde02..367c487c014b 100644
--- a/fs/inotify.c
+++ b/fs/inotify.c
@@ -38,26 +38,25 @@
 #include <asm/ioctls.h>
 
 static atomic_t inotify_cookie;
-static atomic_t inotify_watches;
 
-static kmem_cache_t *watch_cachep;
-static kmem_cache_t *event_cachep;
+static kmem_cache_t *watch_cachep __read_mostly;
+static kmem_cache_t *event_cachep __read_mostly;
 
-static struct vfsmount *inotify_mnt;
+static struct vfsmount *inotify_mnt __read_mostly;
 
 /* these are configurable via /proc/sys/fs/inotify/ */
-int inotify_max_user_instances;
-int inotify_max_user_watches;
-int inotify_max_queued_events;
+int inotify_max_user_instances __read_mostly;
+int inotify_max_user_watches __read_mostly;
+int inotify_max_queued_events __read_mostly;
 
 /*
  * Lock ordering:
  *
  * dentry->d_lock (used to keep d_move() away from dentry->d_parent)
- * iprune_sem (synchronize shrink_icache_memory())
+ * iprune_mutex (synchronize shrink_icache_memory())
  * 	inode_lock (protects the super_block->s_inodes list)
- * 	inode->inotify_sem (protects inode->inotify_watches and watches->i_list)
- * 		inotify_dev->sem (protects inotify_device and watches->d_list)
+ * 	inode->inotify_mutex (protects inode->inotify_watches and watches->i_list)
+ * 		inotify_dev->mutex (protects inotify_device and watches->d_list)
  */
 
 /*
@@ -79,12 +78,12 @@ int inotify_max_queued_events;
 /*
  * struct inotify_device - represents an inotify instance
  *
- * This structure is protected by the semaphore 'sem'.
+ * This structure is protected by the mutex 'mutex'.
  */
 struct inotify_device {
 	wait_queue_head_t 	wq;		/* wait queue for i/o */
 	struct idr		idr;		/* idr mapping wd -> watch */
-	struct semaphore	sem;		/* protects this bad boy */
+	struct mutex		mutex;		/* protects this bad boy */
 	struct list_head 	events;		/* list of queued events */
 	struct list_head	watches;	/* list of watches */
 	atomic_t		count;		/* reference count */
@@ -101,7 +100,7 @@ struct inotify_device {
  * device.  In read(), this list is walked and all events that can fit in the
  * buffer are returned.
  *
- * Protected by dev->sem of the device in which we are queued.
+ * Protected by dev->mutex of the device in which we are queued.
  */
 struct inotify_kernel_event {
 	struct inotify_event	event;	/* the user-space event */
@@ -112,8 +111,8 @@ struct inotify_kernel_event {
 /*
  * struct inotify_watch - represents a watch request on a specific inode
  *
- * d_list is protected by dev->sem of the associated watch->dev.
- * i_list and mask are protected by inode->inotify_sem of the associated inode.
+ * d_list is protected by dev->mutex of the associated watch->dev.
+ * i_list and mask are protected by inode->inotify_mutex of the associated inode.
  * dev, inode, and wd are never written to once the watch is created.
  */
 struct inotify_watch {
@@ -261,7 +260,7 @@ static struct inotify_kernel_event * kernel_event(s32 wd, u32 mask, u32 cookie,
 /*
  * inotify_dev_get_event - return the next event in the given dev's queue
  *
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
  */
 static inline struct inotify_kernel_event *
 inotify_dev_get_event(struct inotify_device *dev)
@@ -272,7 +271,7 @@ inotify_dev_get_event(struct inotify_device *dev)
 /*
  * inotify_dev_queue_event - add a new event to the given device
  *
- * Caller must hold dev->sem.  Can sleep (calls kernel_event()).
+ * Caller must hold dev->mutex.  Can sleep (calls kernel_event()).
  */
 static void inotify_dev_queue_event(struct inotify_device *dev,
 				    struct inotify_watch *watch, u32 mask,
@@ -315,7 +314,7 @@ static void inotify_dev_queue_event(struct inotify_device *dev,
 /*
  * remove_kevent - cleans up and ultimately frees the given kevent
  *
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
  */
 static void remove_kevent(struct inotify_device *dev,
 			  struct inotify_kernel_event *kevent)
@@ -332,7 +331,7 @@ static void remove_kevent(struct inotify_device *dev,
 /*
  * inotify_dev_event_dequeue - destroy an event on the given device
  *
- * Caller must hold dev->sem.
+ * Caller must hold dev->mutex.
  */
 static void inotify_dev_event_dequeue(struct inotify_device *dev)
 {
@@ -346,7 +345,7 @@ static void inotify_dev_event_dequeue(struct inotify_device *dev)
 /*
  * inotify_dev_get_wd - returns the next WD for use by the given dev
  *
- * Callers must hold dev->sem.  This function can sleep.
+ * Callers must hold dev->mutex.  This function can sleep.
  */
 static int inotify_dev_get_wd(struct inotify_device *dev,
 			      struct inotify_watch *watch)
@@ -381,9 +380,51 @@ static int find_inode(const char __user *dirname, struct nameidata *nd,
 }
 
 /*
+ * inotify_inode_watched - returns nonzero if there are watches on this inode
+ * and zero otherwise.  We call this lockless, we do not care if we race.
+ */
+static inline int inotify_inode_watched(struct inode *inode)
+{
+	return !list_empty(&inode->inotify_watches);
+}
+
+/*
+ * Get child dentry flag into synch with parent inode.
+ * Flag should always be clear for negative dentrys.
+ */
+static void set_dentry_child_flags(struct inode *inode, int watched)
+{
+	struct dentry *alias;
+
+	spin_lock(&dcache_lock);
+	list_for_each_entry(alias, &inode->i_dentry, d_alias) {
+		struct dentry *child;
+
+		list_for_each_entry(child, &alias->d_subdirs, d_u.d_child) {
+			if (!child->d_inode) {
+				WARN_ON(child->d_flags & DCACHE_INOTIFY_PARENT_WATCHED);
+				continue;
+			}
+			spin_lock(&child->d_lock);
+			if (watched) {
+				WARN_ON(child->d_flags &
+						DCACHE_INOTIFY_PARENT_WATCHED);
+				child->d_flags |= DCACHE_INOTIFY_PARENT_WATCHED;
+			} else {
+				WARN_ON(!(child->d_flags &
+					DCACHE_INOTIFY_PARENT_WATCHED));
+				child->d_flags&=~DCACHE_INOTIFY_PARENT_WATCHED;
+			}
+			spin_unlock(&child->d_lock);
+		}
+	}
+	spin_unlock(&dcache_lock);
+}
+
+/*
  * create_watch - creates a watch on the given device.
  *
- * Callers must hold dev->sem.  Calls inotify_dev_get_wd() so may sleep.
+ * Callers must hold dev->mutex.  Calls inotify_dev_get_wd() so may sleep.
  * Both 'dev' and 'inode' (by way of nameidata) need to be pinned.
  */
 static struct inotify_watch *create_watch(struct inotify_device *dev,
@@ -426,7 +467,6 @@ static struct inotify_watch *create_watch(struct inotify_device *dev,
 	get_inotify_watch(watch);
 
 	atomic_inc(&dev->user->inotify_watches);
-	atomic_inc(&inotify_watches);
 
 	return watch;
 }
@@ -434,7 +474,7 @@ static struct inotify_watch *create_watch(struct inotify_device *dev,
 /*
  * inotify_find_dev - find the watch associated with the given inode and dev
  *
- * Callers must hold inode->inotify_sem.
+ * Callers must hold inode->inotify_mutex.
  */
 static struct inotify_watch *inode_find_dev(struct inode *inode,
 					    struct inotify_device *dev)
@@ -458,8 +498,10 @@ static void remove_watch_no_event(struct inotify_watch *watch,
 	list_del(&watch->i_list);
 	list_del(&watch->d_list);
 
+	if (!inotify_inode_watched(watch->inode))
+		set_dentry_child_flags(watch->inode, 0);
+
 	atomic_dec(&dev->user->inotify_watches);
-	atomic_dec(&inotify_watches);
 	idr_remove(&dev->idr, watch->wd);
 	put_inotify_watch(watch);
 }
@@ -469,7 +511,7 @@ static void remove_watch_no_event(struct inotify_watch *watch,
  * the IN_IGNORED event to the given device signifying that the inode is no
  * longer watched.
  *
- * Callers must hold both inode->inotify_sem and dev->sem.  We drop a
+ * Callers must hold both inode->inotify_mutex and dev->mutex.  We drop a
  * reference to the inode before returning.
  *
  * The inode is not iput() so as to remain atomic.  If the inode needs to be
@@ -481,16 +523,39 @@ static void remove_watch(struct inotify_watch *watch,struct inotify_device *dev)
 	remove_watch_no_event(watch, dev);
 }
 
+/* Kernel API */
+
 /*
- * inotify_inode_watched - returns nonzero if there are watches on this inode
- * and zero otherwise.  We call this lockless, we do not care if we race.
+ * inotify_d_instantiate - instantiate dcache entry for inode
  */
-static inline int inotify_inode_watched(struct inode *inode)
+void inotify_d_instantiate(struct dentry *entry, struct inode *inode)
 {
-	return !list_empty(&inode->inotify_watches);
+	struct dentry *parent;
+
+	if (!inode)
+		return;
+
+	WARN_ON(entry->d_flags & DCACHE_INOTIFY_PARENT_WATCHED);
+	spin_lock(&entry->d_lock);
+	parent = entry->d_parent;
+	if (inotify_inode_watched(parent->d_inode))
+		entry->d_flags |= DCACHE_INOTIFY_PARENT_WATCHED;
+	spin_unlock(&entry->d_lock);
 }
 
-/* Kernel API */
+/*
+ * inotify_d_move - dcache entry has been moved
+ */
+void inotify_d_move(struct dentry *entry)
+{
+	struct dentry *parent;
+
+	parent = entry->d_parent;
+	if (inotify_inode_watched(parent->d_inode))
+		entry->d_flags |= DCACHE_INOTIFY_PARENT_WATCHED;
+	else
+		entry->d_flags &= ~DCACHE_INOTIFY_PARENT_WATCHED;
+}
 
 /**
  * inotify_inode_queue_event - queue an event to all watches on this inode
@@ -507,21 +572,21 @@ void inotify_inode_queue_event(struct inode *inode, u32 mask, u32 cookie,
 	if (!inotify_inode_watched(inode))
 		return;
 
-	down(&inode->inotify_sem);
+	mutex_lock(&inode->inotify_mutex);
 	list_for_each_entry_safe(watch, next, &inode->inotify_watches, i_list) {
 		u32 watch_mask = watch->mask;
 		if (watch_mask & mask) {
 			struct inotify_device *dev = watch->dev;
 			get_inotify_watch(watch);
-			down(&dev->sem);
+			mutex_lock(&dev->mutex);
 			inotify_dev_queue_event(dev, watch, mask, cookie, name);
 			if (watch_mask & IN_ONESHOT)
 				remove_watch_no_event(watch, dev);
-			up(&dev->sem);
+			mutex_unlock(&dev->mutex);
 			put_inotify_watch(watch);
 		}
 	}
-	up(&inode->inotify_sem);
+	mutex_unlock(&inode->inotify_mutex);
 }
 EXPORT_SYMBOL_GPL(inotify_inode_queue_event);
 
@@ -538,7 +603,7 @@ void inotify_dentry_parent_queue_event(struct dentry *dentry, u32 mask,
 	struct dentry *parent;
 	struct inode *inode;
 
-	if (!atomic_read (&inotify_watches))
+	if (!(dentry->d_flags & DCACHE_INOTIFY_PARENT_WATCHED))
 		return;
 
 	spin_lock(&dentry->d_lock);
@@ -569,7 +634,7 @@ EXPORT_SYMBOL_GPL(inotify_get_cookie);
  * @list: list of inodes being unmounted (sb->s_inodes)
  *
  * Called with inode_lock held, protecting the unmounting super block's list
- * of inodes, and with iprune_sem held, keeping shrink_icache_memory() at bay.
+ * of inodes, and with iprune_mutex held, keeping shrink_icache_memory() at bay.
  * We temporarily drop inode_lock, however, and CAN block.
  */
 void inotify_unmount_inodes(struct list_head *list)
@@ -618,7 +683,7 @@ void inotify_unmount_inodes(struct list_head *list)
 		 * We can safely drop inode_lock here because we hold
 		 * references on both inode and next_i.  Also no new inodes
 		 * will be added since the umount has begun.  Finally,
-		 * iprune_sem keeps shrink_icache_memory() away.
+		 * iprune_mutex keeps shrink_icache_memory() away.
 		 */
 		spin_unlock(&inode_lock);
 
@@ -626,16 +691,16 @@ void inotify_unmount_inodes(struct list_head *list)
 			iput(need_iput_tmp);
 
 		/* for each watch, send IN_UNMOUNT and then remove it */
-		down(&inode->inotify_sem);
+		mutex_lock(&inode->inotify_mutex);
 		watches = &inode->inotify_watches;
 		list_for_each_entry_safe(watch, next_w, watches, i_list) {
 			struct inotify_device *dev = watch->dev;
-			down(&dev->sem);
+			mutex_lock(&dev->mutex);
 			inotify_dev_queue_event(dev, watch, IN_UNMOUNT,0,NULL);
 			remove_watch(watch, dev);
-			up(&dev->sem);
+			mutex_unlock(&dev->mutex);
 		}
-		up(&inode->inotify_sem);
+		mutex_unlock(&inode->inotify_mutex);
 		iput(inode);		
 
 		spin_lock(&inode_lock);
@@ -651,14 +716,14 @@ void inotify_inode_is_dead(struct inode *inode)
 {
 	struct inotify_watch *watch, *next;
 
-	down(&inode->inotify_sem);
+	mutex_lock(&inode->inotify_mutex);
 	list_for_each_entry_safe(watch, next, &inode->inotify_watches, i_list) {
 		struct inotify_device *dev = watch->dev;
-		down(&dev->sem);
+		mutex_lock(&dev->mutex);
 		remove_watch(watch, dev);
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 	}
-	up(&inode->inotify_sem);
+	mutex_unlock(&inode->inotify_mutex);
 }
 EXPORT_SYMBOL_GPL(inotify_inode_is_dead);
 
@@ -670,10 +735,10 @@ static unsigned int inotify_poll(struct file *file, poll_table *wait)
 	int ret = 0;
 
 	poll_wait(file, &dev->wq, wait);
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	if (!list_empty(&dev->events))
 		ret = POLLIN | POLLRDNORM;
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
 	return ret;
 }
@@ -695,9 +760,9 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
 
 		prepare_to_wait(&dev->wq, &wait, TASK_INTERRUPTIBLE);
 
-		down(&dev->sem);
+		mutex_lock(&dev->mutex);
 		events = !list_empty(&dev->events);
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 		if (events) {
 			ret = 0;
 			break;
@@ -720,7 +785,7 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
 	if (ret)
 		return ret;
 
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	while (1) {
 		struct inotify_kernel_event *kevent;
 
@@ -750,7 +815,7 @@ static ssize_t inotify_read(struct file *file, char __user *buf,
 
 		remove_kevent(dev, kevent);
 	}
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
 	return ret;
 }
@@ -763,37 +828,37 @@ static int inotify_release(struct inode *ignored, struct file *file)
 	 * Destroy all of the watches on this device.  Unfortunately, not very
 	 * pretty.  We cannot do a simple iteration over the list, because we
 	 * do not know the inode until we iterate to the watch.  But we need to
-	 * hold inode->inotify_sem before dev->sem.  The following works.
+	 * hold inode->inotify_mutex before dev->mutex.  The following works.
 	 */
 	while (1) {
 		struct inotify_watch *watch;
 		struct list_head *watches;
 		struct inode *inode;
 
-		down(&dev->sem);
+		mutex_lock(&dev->mutex);
 		watches = &dev->watches;
 		if (list_empty(watches)) {
-			up(&dev->sem);
+			mutex_unlock(&dev->mutex);
 			break;
 		}
 		watch = list_entry(watches->next, struct inotify_watch, d_list);
 		get_inotify_watch(watch);
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 
 		inode = watch->inode;
-		down(&inode->inotify_sem);
-		down(&dev->sem);
+		mutex_lock(&inode->inotify_mutex);
+		mutex_lock(&dev->mutex);
 		remove_watch_no_event(watch, dev);
-		up(&dev->sem);
-		up(&inode->inotify_sem);
+		mutex_unlock(&dev->mutex);
+		mutex_unlock(&inode->inotify_mutex);
 		put_inotify_watch(watch);
 	}
 
 	/* destroy all of the events on this device */
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	while (!list_empty(&dev->events))
 		inotify_dev_event_dequeue(dev);
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
 	/* free this device: the put matching the get in inotify_init() */
 	put_inotify_dev(dev);
@@ -811,26 +876,26 @@ static int inotify_ignore(struct inotify_device *dev, s32 wd)
 	struct inotify_watch *watch;
 	struct inode *inode;
 
-	down(&dev->sem);
+	mutex_lock(&dev->mutex);
 	watch = idr_find(&dev->idr, wd);
 	if (unlikely(!watch)) {
-		up(&dev->sem);
+		mutex_unlock(&dev->mutex);
 		return -EINVAL;
 	}
 	get_inotify_watch(watch);
 	inode = watch->inode;
-	up(&dev->sem);
+	mutex_unlock(&dev->mutex);
 
-	down(&inode->inotify_sem);
-	down(&dev->sem);
+	mutex_lock(&inode->inotify_mutex);
+	mutex_lock(&dev->mutex);
 
 	/* make sure that we did not race */
 	watch = idr_find(&dev->idr, wd);
 	if (likely(watch))
 		remove_watch(watch, dev);
 
-	up(&dev->sem);
-	up(&inode->inotify_sem);
+	mutex_unlock(&dev->mutex);
+	mutex_unlock(&inode->inotify_mutex);
 	put_inotify_watch(watch);
 
 	return 0;
@@ -855,7 +920,7 @@ static long inotify_ioctl(struct file *file, unsigned int cmd,
 	return ret;
 }
 
-static struct file_operations inotify_fops = {
+static const struct file_operations inotify_fops = {
 	.poll           = inotify_poll,
 	.read           = inotify_read,
 	.release        = inotify_release,
@@ -905,7 +970,7 @@ asmlinkage long sys_inotify_init(void)
 	INIT_LIST_HEAD(&dev->events);
 	INIT_LIST_HEAD(&dev->watches);
 	init_waitqueue_head(&dev->wq);
-	sema_init(&dev->sem, 1);
+	mutex_init(&dev->mutex);
 	dev->event_count = 0;
 	dev->queue_size = 0;
 	dev->max_events = inotify_max_queued_events;
@@ -960,8 +1025,8 @@ asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask)
 	inode = nd.dentry->d_inode;
 	dev = filp->private_data;
 
-	down(&inode->inotify_sem);
-	down(&dev->sem);
+	mutex_lock(&inode->inotify_mutex);
+	mutex_lock(&dev->mutex);
 
 	if (mask & IN_MASK_ADD)
 		mask_add = 1;
@@ -993,13 +1058,16 @@ asmlinkage long sys_inotify_add_watch(int fd, const char __user *path, u32 mask)
 		goto out;
 	}
 
+	if (!inotify_inode_watched(inode))
+		set_dentry_child_flags(inode, 1);
+
 	/* Add the watch to the device's and the inode's list */
 	list_add(&watch->d_list, &dev->watches);
 	list_add(&watch->i_list, &inode->inotify_watches);
 	ret = watch->wd;
 out:
-	up(&dev->sem);
-	up(&inode->inotify_sem);
+	mutex_unlock(&dev->mutex);
+	mutex_unlock(&inode->inotify_mutex);
 	path_release(&nd);
 fput_and_out:
 	fput_light(filp, fput_needed);
@@ -1065,7 +1133,6 @@ static int __init inotify_setup(void)
 	inotify_max_user_watches = 8192;
 
 	atomic_set(&inotify_cookie, 0);
-	atomic_set(&inotify_watches, 0);
 
 	watch_cachep = kmem_cache_create("inotify_watch_cache",
 					 sizeof(struct inotify_watch),
diff --git a/fs/isofs/dir.c b/fs/isofs/dir.c
index 7901ac9f97ab..5440ea292c69 100644
--- a/fs/isofs/dir.c
+++ b/fs/isofs/dir.c
@@ -16,7 +16,7 @@
 
 static int isofs_readdir(struct file *, void *, filldir_t);
 
-struct file_operations isofs_dir_operations =
+const struct file_operations isofs_dir_operations =
 {
 	.read		= generic_read_dir,
 	.readdir	= isofs_readdir,
diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c
index 298f08be22d4..70adbb98bad1 100644
--- a/fs/isofs/inode.c
+++ b/fs/isofs/inode.c
@@ -87,7 +87,8 @@ static int init_inodecache(void)
 {
 	isofs_inode_cachep = kmem_cache_create("isofs_inode_cache",
 					     sizeof(struct iso_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (isofs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/isofs/isofs.h b/fs/isofs/isofs.h
index 38c75151fc66..b87ba066f5e7 100644
--- a/fs/isofs/isofs.h
+++ b/fs/isofs/isofs.h
@@ -175,18 +175,6 @@ isofs_normalize_block_and_offset(struct iso_directory_record* de,
 }
 
 extern struct inode_operations isofs_dir_inode_operations;
-extern struct file_operations isofs_dir_operations;
+extern const struct file_operations isofs_dir_operations;
 extern struct address_space_operations isofs_symlink_aops;
 extern struct export_operations isofs_export_ops;
-
-/* The following macros are used to check for memory leaks. */
-#ifdef LEAK_CHECK
-#define free_s leak_check_free_s
-#define malloc leak_check_malloc
-#define sb_bread leak_check_bread
-#define brelse leak_check_brelse
-extern void * leak_check_malloc(unsigned int size);
-extern void leak_check_free_s(void * obj, int size);
-extern struct buffer_head * leak_check_bread(struct super_block *sb, int block);
-extern void leak_check_brelse(struct buffer_head * bh);
-#endif /* LEAK_CHECK */
diff --git a/fs/isofs/joliet.c b/fs/isofs/joliet.c
index 2931de7f1a6a..81a90e170ac3 100644
--- a/fs/isofs/joliet.c
+++ b/fs/isofs/joliet.c
@@ -11,7 +11,7 @@
 #include "isofs.h"
 
 /*
- * Convert Unicode 16 to UTF8 or ASCII.
+ * Convert Unicode 16 to UTF-8 or ASCII.
  */
 static int
 uni16_to_x8(unsigned char *ascii, u16 *uni, int len, struct nls_table *nls)
diff --git a/fs/jbd/checkpoint.c b/fs/jbd/checkpoint.c
index 543ed543d1e5..3f5102b069db 100644
--- a/fs/jbd/checkpoint.c
+++ b/fs/jbd/checkpoint.c
@@ -85,7 +85,7 @@ void __log_wait_for_space(journal_t *journal)
 		if (journal->j_flags & JFS_ABORT)
 			return;
 		spin_unlock(&journal->j_state_lock);
-		down(&journal->j_checkpoint_sem);
+		mutex_lock(&journal->j_checkpoint_mutex);
 
 		/*
 		 * Test again, another process may have checkpointed while we
@@ -98,7 +98,7 @@ void __log_wait_for_space(journal_t *journal)
 			log_do_checkpoint(journal);
 			spin_lock(&journal->j_state_lock);
 		}
-		up(&journal->j_checkpoint_sem);
+		mutex_unlock(&journal->j_checkpoint_mutex);
 	}
 }
 
diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c
index e4b516ac4989..7f96b5cb6781 100644
--- a/fs/jbd/journal.c
+++ b/fs/jbd/journal.c
@@ -33,9 +33,11 @@
 #include <linux/mm.h>
 #include <linux/suspend.h>
 #include <linux/pagemap.h>
+#include <linux/kthread.h>
+#include <linux/proc_fs.h>
+
 #include <asm/uaccess.h>
 #include <asm/page.h>
-#include <linux/proc_fs.h>
 
 EXPORT_SYMBOL(journal_start);
 EXPORT_SYMBOL(journal_restart);
@@ -111,18 +113,15 @@ static void commit_timeout(unsigned long __data)
 
 static int kjournald(void *arg)
 {
-	journal_t *journal = (journal_t *) arg;
+	journal_t *journal = arg;
 	transaction_t *transaction;
-	struct timer_list timer;
 
-	daemonize("kjournald");
-
-	/* Set up an interval timer which can be used to trigger a
-           commit wakeup after the commit interval expires */
-	init_timer(&timer);
-	timer.data = (unsigned long) current;
-	timer.function = commit_timeout;
-	journal->j_commit_timer = &timer;
+	/*
+	 * Set up an interval timer which can be used to trigger a commit wakeup
+	 * after the commit interval expires
+	 */
+	setup_timer(&journal->j_commit_timer, commit_timeout,
+			(unsigned long)current);
 
 	/* Record that the journal thread is running */
 	journal->j_task = current;
@@ -146,7 +145,7 @@ loop:
 	if (journal->j_commit_sequence != journal->j_commit_request) {
 		jbd_debug(1, "OK, requests differ\n");
 		spin_unlock(&journal->j_state_lock);
-		del_timer_sync(journal->j_commit_timer);
+		del_timer_sync(&journal->j_commit_timer);
 		journal_commit_transaction(journal);
 		spin_lock(&journal->j_state_lock);
 		goto loop;
@@ -203,7 +202,7 @@ loop:
 
 end_loop:
 	spin_unlock(&journal->j_state_lock);
-	del_timer_sync(journal->j_commit_timer);
+	del_timer_sync(&journal->j_commit_timer);
 	journal->j_task = NULL;
 	wake_up(&journal->j_wait_done_commit);
 	jbd_debug(1, "Journal thread exiting.\n");
@@ -212,7 +211,7 @@ end_loop:
 
 static void journal_start_thread(journal_t *journal)
 {
-	kernel_thread(kjournald, journal, CLONE_VM|CLONE_FS|CLONE_FILES);
+	kthread_run(kjournald, journal, "kjournald");
 	wait_event(journal->j_wait_done_commit, journal->j_task != 0);
 }
 
@@ -659,8 +658,8 @@ static journal_t * journal_init_common (void)
 	init_waitqueue_head(&journal->j_wait_checkpoint);
 	init_waitqueue_head(&journal->j_wait_commit);
 	init_waitqueue_head(&journal->j_wait_updates);
-	init_MUTEX(&journal->j_barrier);
-	init_MUTEX(&journal->j_checkpoint_sem);
+	mutex_init(&journal->j_barrier);
+	mutex_init(&journal->j_checkpoint_mutex);
 	spin_lock_init(&journal->j_revoke_lock);
 	spin_lock_init(&journal->j_list_lock);
 	spin_lock_init(&journal->j_state_lock);
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c
index ca917973c2c0..c609f5034fcd 100644
--- a/fs/jbd/transaction.c
+++ b/fs/jbd/transaction.c
@@ -53,8 +53,8 @@ get_transaction(journal_t *journal, transaction_t *transaction)
 	spin_lock_init(&transaction->t_handle_lock);
 
 	/* Set up the commit timer for the new transaction. */
-	journal->j_commit_timer->expires = transaction->t_expires;
-	add_timer(journal->j_commit_timer);
+	journal->j_commit_timer.expires = transaction->t_expires;
+	add_timer(&journal->j_commit_timer);
 
 	J_ASSERT(journal->j_running_transaction == NULL);
 	journal->j_running_transaction = transaction;
@@ -455,7 +455,7 @@ void journal_lock_updates(journal_t *journal)
 	 * to make sure that we serialise special journal-locked operations
 	 * too.
 	 */
-	down(&journal->j_barrier);
+	mutex_lock(&journal->j_barrier);
 }
 
 /**
@@ -470,7 +470,7 @@ void journal_unlock_updates (journal_t *journal)
 {
 	J_ASSERT(journal->j_barrier_count != 0);
 
-	up(&journal->j_barrier);
+	mutex_unlock(&journal->j_barrier);
 	spin_lock(&journal->j_state_lock);
 	--journal->j_barrier_count;
 	spin_unlock(&journal->j_state_lock);
@@ -1873,16 +1873,15 @@ zap_buffer_unlocked:
 }
 
 /** 
- * int journal_invalidatepage() 
+ * void journal_invalidatepage()
  * @journal: journal to use for flush... 
  * @page:    page to flush
  * @offset:  length of page to invalidate.
  *
  * Reap page buffers containing data after offset in page.
  *
- * Return non-zero if the page's buffers were successfully reaped.
  */
-int journal_invalidatepage(journal_t *journal, 
+void journal_invalidatepage(journal_t *journal,
 		      struct page *page, 
 		      unsigned long offset)
 {
@@ -1893,7 +1892,7 @@ int journal_invalidatepage(journal_t *journal,
 	if (!PageLocked(page))
 		BUG();
 	if (!page_has_buffers(page))
-		return 1;
+		return;
 
 	/* We will potentially be playing with lists other than just the
 	 * data lists (especially for journaled data mode), so be
@@ -1916,11 +1915,9 @@ int journal_invalidatepage(journal_t *journal,
 	} while (bh != head);
 
 	if (!offset) {
-		if (!may_free || !try_to_free_buffers(page))
-			return 0;
-		J_ASSERT(!page_has_buffers(page));
+		if (may_free && try_to_free_buffers(page))
+			J_ASSERT(!page_has_buffers(page));
 	}
-	return 1;
 }
 
 /* 
diff --git a/fs/jffs/inode-v23.c b/fs/jffs/inode-v23.c
index fc3855a1aef3..020cc097c539 100644
--- a/fs/jffs/inode-v23.c
+++ b/fs/jffs/inode-v23.c
@@ -42,7 +42,7 @@
 #include <linux/quotaops.h>
 #include <linux/highmem.h>
 #include <linux/vfs.h>
-#include <asm/semaphore.h>
+#include <linux/mutex.h>
 #include <asm/byteorder.h>
 #include <asm/uaccess.h>
 
@@ -55,9 +55,9 @@
 static int jffs_remove(struct inode *dir, struct dentry *dentry, int type);
 
 static struct super_operations jffs_ops;
-static struct file_operations jffs_file_operations;
+static const struct file_operations jffs_file_operations;
 static struct inode_operations jffs_file_inode_operations;
-static struct file_operations jffs_dir_operations;
+static const struct file_operations jffs_dir_operations;
 static struct inode_operations jffs_dir_inode_operations;
 static struct address_space_operations jffs_address_operations;
 
@@ -203,7 +203,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 	fmc = c->fmc;
 
 	D3(printk (KERN_NOTICE "notify_change(): down biglock\n"));
-	down(&fmc->biglock);
+	mutex_lock(&fmc->biglock);
 
 	f = jffs_find_file(c, inode->i_ino);
 
@@ -211,7 +211,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 		printk("jffs_setattr(): Invalid inode number: %lu\n",
 		       inode->i_ino);
 		D3(printk (KERN_NOTICE "notify_change(): up biglock\n"));
-		up(&fmc->biglock);
+		mutex_unlock(&fmc->biglock);
 		res = -EINVAL;
 		goto out;
 	});
@@ -232,7 +232,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 	if (!(new_node = jffs_alloc_node())) {
 		D(printk("jffs_setattr(): Allocation failed!\n"));
 		D3(printk (KERN_NOTICE "notify_change(): up biglock\n"));
-		up(&fmc->biglock);
+		mutex_unlock(&fmc->biglock);
 		res = -ENOMEM;
 		goto out;
 	}
@@ -319,7 +319,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 		D(printk("jffs_notify_change(): The write failed!\n"));
 		jffs_free_node(new_node);
 		D3(printk (KERN_NOTICE "n_c(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		goto out;
 	}
 
@@ -327,7 +327,7 @@ jffs_setattr(struct dentry *dentry, struct iattr *iattr)
 
 	mark_inode_dirty(inode);
 	D3(printk (KERN_NOTICE "n_c(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 out:
 	unlock_kernel();
 	return res;
@@ -461,7 +461,7 @@ jffs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		goto jffs_rename_end;
 	}
 	D3(printk (KERN_NOTICE "rename(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	/* Create a node and initialize as much as needed.  */
 	result = -ENOMEM;
 	if (!(node = jffs_alloc_node())) {
@@ -555,7 +555,7 @@ jffs_rename(struct inode *old_dir, struct dentry *old_dentry,
 
 jffs_rename_end:
 	D3(printk (KERN_NOTICE "rename(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return result;
 } /* jffs_rename()  */
@@ -574,14 +574,14 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	int ddino;
 	lock_kernel();
 	D3(printk (KERN_NOTICE "readdir(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	D2(printk("jffs_readdir(): inode: 0x%p, filp: 0x%p\n", inode, filp));
 	if (filp->f_pos == 0) {
 		D3(printk("jffs_readdir(): \".\" %lu\n", inode->i_ino));
 		if (filldir(dirent, ".", 1, filp->f_pos, inode->i_ino, DT_DIR) < 0) {
 			D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-			up(&c->fmc->biglock);
+			mutex_unlock(&c->fmc->biglock);
 			unlock_kernel();
 			return 0;
 		}
@@ -598,7 +598,7 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		D3(printk("jffs_readdir(): \"..\" %u\n", ddino));
 		if (filldir(dirent, "..", 2, filp->f_pos, ddino, DT_DIR) < 0) {
 			D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-			up(&c->fmc->biglock);
+			mutex_unlock(&c->fmc->biglock);
 			unlock_kernel();
 			return 0;
 		}
@@ -617,7 +617,7 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		if (filldir(dirent, f->name, f->nsize,
 			    filp->f_pos , f->ino, DT_UNKNOWN) < 0) {
 		        D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-			up(&c->fmc->biglock);
+			mutex_unlock(&c->fmc->biglock);
 			unlock_kernel();
 			return 0;
 		}
@@ -627,7 +627,7 @@ jffs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		} while(f && f->deleted);
 	}
 	D3(printk (KERN_NOTICE "readdir(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return filp->f_pos;
 } /* jffs_readdir()  */
@@ -660,7 +660,7 @@ jffs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 	});
 
 	D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	r = -ENAMETOOLONG;
 	if (len > JFFS_MAX_NAME_LEN) {
@@ -683,31 +683,31 @@ jffs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 
 	if ((len == 1) && (name[0] == '.')) {
 		D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		if (!(inode = iget(dir->i_sb, d->ino))) {
 			D(printk("jffs_lookup(): . iget() ==> NULL\n"));
 			goto jffs_lookup_end_no_biglock;
 		}
 		D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-		down(&c->fmc->biglock);
+		mutex_lock(&c->fmc->biglock);
 	} else if ((len == 2) && (name[0] == '.') && (name[1] == '.')) {
 	        D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
  		if (!(inode = iget(dir->i_sb, d->pino))) {
 			D(printk("jffs_lookup(): .. iget() ==> NULL\n"));
 			goto jffs_lookup_end_no_biglock;
 		}
 		D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-		down(&c->fmc->biglock);
+		mutex_lock(&c->fmc->biglock);
 	} else if ((f = jffs_find_child(d, name, len))) {
 	        D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		if (!(inode = iget(dir->i_sb, f->ino))) {
 			D(printk("jffs_lookup(): iget() ==> NULL\n"));
 			goto jffs_lookup_end_no_biglock;
 		}
 		D3(printk (KERN_NOTICE "lookup(): down biglock\n"));
-		down(&c->fmc->biglock);
+		mutex_lock(&c->fmc->biglock);
 	} else {
 		D3(printk("jffs_lookup(): Couldn't find the file. "
 			  "f = 0x%p, name = \"%s\", d = 0x%p, d->ino = %u\n",
@@ -717,13 +717,13 @@ jffs_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 
 	d_add(dentry, inode);
 	D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return NULL;
 
 jffs_lookup_end:
 	D3(printk (KERN_NOTICE "lookup(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 
 jffs_lookup_end_no_biglock:
 	unlock_kernel();
@@ -753,7 +753,7 @@ jffs_do_readpage_nolock(struct file *file, struct page *page)
 	ClearPageError(page);
 
 	D3(printk (KERN_NOTICE "readpage(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	read_len = 0;
 	result = 0;
@@ -782,7 +782,7 @@ jffs_do_readpage_nolock(struct file *file, struct page *page)
 	kunmap(page);
 
 	D3(printk (KERN_NOTICE "readpage(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 
 	if (result) {
 	        SetPageError(page);
@@ -839,7 +839,7 @@ jffs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 
 	c = dir_f->c;
 	D3(printk (KERN_NOTICE "mkdir(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	dir_mode = S_IFDIR | (mode & (S_IRWXUGO|S_ISVTX)
 			      & ~current->fs->umask);
@@ -906,7 +906,7 @@ jffs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	result = 0;
 jffs_mkdir_end:
 	D3(printk (KERN_NOTICE "mkdir(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return result;
 } /* jffs_mkdir()  */
@@ -921,10 +921,10 @@ jffs_rmdir(struct inode *dir, struct dentry *dentry)
 	D3(printk("***jffs_rmdir()\n"));
 	D3(printk (KERN_NOTICE "rmdir(): down biglock\n"));
 	lock_kernel();
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	ret = jffs_remove(dir, dentry, S_IFDIR);
 	D3(printk (KERN_NOTICE "rmdir(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return ret;
 }
@@ -940,10 +940,10 @@ jffs_unlink(struct inode *dir, struct dentry *dentry)
 	lock_kernel();
 	D3(printk("***jffs_unlink()\n"));
 	D3(printk (KERN_NOTICE "unlink(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	ret = jffs_remove(dir, dentry, 0);
 	D3(printk (KERN_NOTICE "unlink(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return ret;
 }
@@ -1086,7 +1086,7 @@ jffs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
 	c = dir_f->c;
 
 	D3(printk (KERN_NOTICE "mknod(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	/* Create and initialize a new node.  */
 	if (!(node = jffs_alloc_node())) {
@@ -1152,7 +1152,7 @@ jffs_mknod_err:
 
 jffs_mknod_end:
 	D3(printk (KERN_NOTICE "mknod(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return result;
 } /* jffs_mknod()  */
@@ -1203,7 +1203,7 @@ jffs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
 		return -ENOMEM;
 	}
 	D3(printk (KERN_NOTICE "symlink(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	node->data_offset = 0;
 	node->removed_size = 0;
@@ -1253,7 +1253,7 @@ jffs_symlink(struct inode *dir, struct dentry *dentry, const char *symname)
 	d_instantiate(dentry, inode);
  jffs_symlink_end:
 	D3(printk (KERN_NOTICE "symlink(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return err;
 } /* jffs_symlink()  */
@@ -1306,7 +1306,7 @@ jffs_create(struct inode *dir, struct dentry *dentry, int mode,
 		return -ENOMEM;
 	}
 	D3(printk (KERN_NOTICE "create(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	node->data_offset = 0;
 	node->removed_size = 0;
@@ -1359,7 +1359,7 @@ jffs_create(struct inode *dir, struct dentry *dentry, int mode,
 	d_instantiate(dentry, inode);
  jffs_create_end:
 	D3(printk (KERN_NOTICE "create(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	unlock_kernel();
 	return err;
 } /* jffs_create()  */
@@ -1423,7 +1423,7 @@ jffs_file_write(struct file *filp, const char *buf, size_t count,
 	thiscount = min(c->fmc->max_chunk_size - sizeof(struct jffs_raw_inode), count);
 
 	D3(printk (KERN_NOTICE "file_write(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	/* Urgh. POSIX says we can do short writes if we feel like it. 
 	 * In practice, we can't. Nothing will cope. So we loop until
@@ -1511,7 +1511,7 @@ jffs_file_write(struct file *filp, const char *buf, size_t count,
 	}
  out:
 	D3(printk (KERN_NOTICE "file_write(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 
 	/* Fix things in the real inode.  */
 	if (pos > inode->i_size) {
@@ -1567,7 +1567,7 @@ jffs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 		return -EIO;
 	}
 	D3(printk (KERN_NOTICE "ioctl(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 
 	switch (cmd) {
 	case JFFS_PRINT_HASH:
@@ -1609,7 +1609,7 @@ jffs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
 		ret = -ENOTTY;
 	}
 	D3(printk (KERN_NOTICE "ioctl(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 	return ret;
 } /* jffs_ioctl()  */
 
@@ -1629,7 +1629,7 @@ static int jffs_fsync(struct file *f, struct dentry *d, int datasync)
 }
 
 
-static struct file_operations jffs_file_operations =
+static const struct file_operations jffs_file_operations =
 {
 	.open		= generic_file_open,
 	.llseek		= generic_file_llseek,
@@ -1649,7 +1649,7 @@ static struct inode_operations jffs_file_inode_operations =
 };
 
 
-static struct file_operations jffs_dir_operations =
+static const struct file_operations jffs_dir_operations =
 {
 	.readdir	= jffs_readdir,
 };
@@ -1685,12 +1685,12 @@ jffs_read_inode(struct inode *inode)
 	}
 	c = (struct jffs_control *)inode->i_sb->s_fs_info;
 	D3(printk (KERN_NOTICE "read_inode(): down biglock\n"));
-	down(&c->fmc->biglock);
+	mutex_lock(&c->fmc->biglock);
 	if (!(f = jffs_find_file(c, inode->i_ino))) {
 		D(printk("jffs_read_inode(): No such inode (%lu).\n",
 			 inode->i_ino));
 		D3(printk (KERN_NOTICE "read_inode(): up biglock\n"));
-		up(&c->fmc->biglock);
+		mutex_unlock(&c->fmc->biglock);
 		return;
 	}
 	inode->u.generic_ip = (void *)f;
@@ -1732,7 +1732,7 @@ jffs_read_inode(struct inode *inode)
 	}
 
 	D3(printk (KERN_NOTICE "read_inode(): up biglock\n"));
-	up(&c->fmc->biglock);
+	mutex_unlock(&c->fmc->biglock);
 }
 
 
@@ -1812,15 +1812,17 @@ init_jffs_fs(void)
 	}
 #endif
 	fm_cache = kmem_cache_create("jffs_fm", sizeof(struct jffs_fm),
-				     0, SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT, 
-				     NULL, NULL);
+		       0,
+		       SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
+		       NULL, NULL);
 	if (!fm_cache) {
 		return -ENOMEM;
 	}
 
 	node_cache = kmem_cache_create("jffs_node",sizeof(struct jffs_node),
-				       0, SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT, 
-				       NULL, NULL);
+		       0,
+		       SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
+		       NULL, NULL);
 	if (!node_cache) {
 		kmem_cache_destroy(fm_cache);
 		return -ENOMEM;
diff --git a/fs/jffs/intrep.c b/fs/jffs/intrep.c
index ce7b54b0b2b7..0ef207dfaf6f 100644
--- a/fs/jffs/intrep.c
+++ b/fs/jffs/intrep.c
@@ -62,7 +62,7 @@
 #include <linux/fs.h>
 #include <linux/stat.h>
 #include <linux/pagemap.h>
-#include <asm/semaphore.h>
+#include <linux/mutex.h>
 #include <asm/byteorder.h>
 #include <linux/smp_lock.h>
 #include <linux/time.h>
@@ -3416,7 +3416,7 @@ jffs_garbage_collect_thread(void *ptr)
 		D1(printk (KERN_NOTICE "jffs_garbage_collect_thread(): collecting.\n"));
 
 		D3(printk (KERN_NOTICE "g_c_thread(): down biglock\n"));
-		down(&fmc->biglock);
+		mutex_lock(&fmc->biglock);
 		
 		D1(printk("***jffs_garbage_collect_thread(): round #%u, "
 			  "fmc->dirty_size = %u\n", i++, fmc->dirty_size));
@@ -3447,6 +3447,6 @@ jffs_garbage_collect_thread(void *ptr)
 		
 	gc_end:
 		D3(printk (KERN_NOTICE "g_c_thread(): up biglock\n"));
-		up(&fmc->biglock);
+		mutex_unlock(&fmc->biglock);
 	} /* for (;;) */
 } /* jffs_garbage_collect_thread() */
diff --git a/fs/jffs/jffs_fm.c b/fs/jffs/jffs_fm.c
index 6da13b309bd1..7d8ca1aeace2 100644
--- a/fs/jffs/jffs_fm.c
+++ b/fs/jffs/jffs_fm.c
@@ -139,7 +139,7 @@ jffs_build_begin(struct jffs_control *c, int unit)
 	fmc->tail = NULL;
 	fmc->head_extra = NULL;
 	fmc->tail_extra = NULL;
-	init_MUTEX(&fmc->biglock);
+	mutex_init(&fmc->biglock);
 	return fmc;
 }
 
diff --git a/fs/jffs/jffs_fm.h b/fs/jffs/jffs_fm.h
index f64151e74122..c794d923df2a 100644
--- a/fs/jffs/jffs_fm.h
+++ b/fs/jffs/jffs_fm.h
@@ -20,10 +20,11 @@
 #ifndef __LINUX_JFFS_FM_H__
 #define __LINUX_JFFS_FM_H__
 
+#include <linux/config.h>
 #include <linux/types.h>
 #include <linux/jffs.h>
 #include <linux/mtd/mtd.h>
-#include <linux/config.h>
+#include <linux/mutex.h>
 
 /* The alignment between two nodes in the flash memory.  */
 #define JFFS_ALIGN_SIZE 4
@@ -97,7 +98,7 @@ struct jffs_fmcontrol
 	struct jffs_fm *tail;
 	struct jffs_fm *head_extra;
 	struct jffs_fm *tail_extra;
-	struct semaphore biglock;
+	struct mutex biglock;
 };
 
 /* Notice the two members head_extra and tail_extra in the jffs_control
diff --git a/fs/jffs2/compr_zlib.c b/fs/jffs2/compr_zlib.c
index 4db8be8e90cc..5c63e0cdcf4c 100644
--- a/fs/jffs2/compr_zlib.c
+++ b/fs/jffs2/compr_zlib.c
@@ -33,13 +33,14 @@
 	*/
 #define STREAM_END_SPACE 12
 
-static DECLARE_MUTEX(deflate_sem);
-static DECLARE_MUTEX(inflate_sem);
+static DEFINE_MUTEX(deflate_mutex);
+static DEFINE_MUTEX(inflate_mutex);
 static z_stream inf_strm, def_strm;
 
 #ifdef __KERNEL__ /* Linux-only */
 #include <linux/vmalloc.h>
 #include <linux/init.h>
+#include <linux/mutex.h>
 
 static int __init alloc_workspaces(void)
 {
@@ -79,11 +80,11 @@ static int jffs2_zlib_compress(unsigned char *data_in,
 	if (*dstlen <= STREAM_END_SPACE)
 		return -1;
 
-	down(&deflate_sem);
+	mutex_lock(&deflate_mutex);
 
 	if (Z_OK != zlib_deflateInit(&def_strm, 3)) {
 		printk(KERN_WARNING "deflateInit failed\n");
-		up(&deflate_sem);
+		mutex_unlock(&deflate_mutex);
 		return -1;
 	}
 
@@ -104,7 +105,7 @@ static int jffs2_zlib_compress(unsigned char *data_in,
 		if (ret != Z_OK) {
 			D1(printk(KERN_DEBUG "deflate in loop returned %d\n", ret));
 			zlib_deflateEnd(&def_strm);
-			up(&deflate_sem);
+			mutex_unlock(&deflate_mutex);
 			return -1;
 		}
 	}
@@ -133,7 +134,7 @@ static int jffs2_zlib_compress(unsigned char *data_in,
 	*sourcelen = def_strm.total_in;
 	ret = 0;
  out:
-	up(&deflate_sem);
+	mutex_unlock(&deflate_mutex);
 	return ret;
 }
 
@@ -145,7 +146,7 @@ static int jffs2_zlib_decompress(unsigned char *data_in,
 	int ret;
 	int wbits = MAX_WBITS;
 
-	down(&inflate_sem);
+	mutex_lock(&inflate_mutex);
 
 	inf_strm.next_in = data_in;
 	inf_strm.avail_in = srclen;
@@ -173,7 +174,7 @@ static int jffs2_zlib_decompress(unsigned char *data_in,
 
 	if (Z_OK != zlib_inflateInit2(&inf_strm, wbits)) {
 		printk(KERN_WARNING "inflateInit failed\n");
-		up(&inflate_sem);
+		mutex_unlock(&inflate_mutex);
 		return 1;
 	}
 
@@ -183,7 +184,7 @@ static int jffs2_zlib_decompress(unsigned char *data_in,
 		printk(KERN_NOTICE "inflate returned %d\n", ret);
 	}
 	zlib_inflateEnd(&inf_strm);
-	up(&inflate_sem);
+	mutex_unlock(&inflate_mutex);
         return 0;
 }
 
diff --git a/fs/jffs2/dir.c b/fs/jffs2/dir.c
index a7bf9cb2567f..8bc7a5018e40 100644
--- a/fs/jffs2/dir.c
+++ b/fs/jffs2/dir.c
@@ -37,7 +37,7 @@ static int jffs2_mknod (struct inode *,struct dentry *,int,dev_t);
 static int jffs2_rename (struct inode *, struct dentry *,
                         struct inode *, struct dentry *);
 
-struct file_operations jffs2_dir_operations =
+const struct file_operations jffs2_dir_operations =
 {
 	.read =		generic_read_dir,
 	.readdir =	jffs2_readdir,
diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index 935f273dc57b..9f4171213e58 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -38,7 +38,7 @@ int jffs2_fsync(struct file *filp, struct dentry *dentry, int datasync)
 	return 0;
 }
 
-struct file_operations jffs2_file_operations =
+const struct file_operations jffs2_file_operations =
 {
 	.llseek =	generic_file_llseek,
 	.open =		generic_file_open,
diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h
index 59e7a393200c..d307cf548625 100644
--- a/fs/jffs2/os-linux.h
+++ b/fs/jffs2/os-linux.h
@@ -159,11 +159,11 @@ void jffs2_stop_garbage_collect_thread(struct jffs2_sb_info *c);
 void jffs2_garbage_collect_trigger(struct jffs2_sb_info *c);
 
 /* dir.c */
-extern struct file_operations jffs2_dir_operations;
+extern const struct file_operations jffs2_dir_operations;
 extern struct inode_operations jffs2_dir_inode_operations;
 
 /* file.c */
-extern struct file_operations jffs2_file_operations;
+extern const struct file_operations jffs2_file_operations;
 extern struct inode_operations jffs2_file_inode_operations;
 extern struct address_space_operations jffs2_file_address_operations;
 int jffs2_fsync(struct file *, struct dentry *, int);
diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c
index 93883817cbd0..ffd8e84b22cc 100644
--- a/fs/jffs2/super.c
+++ b/fs/jffs2/super.c
@@ -152,7 +152,7 @@ static struct super_block *jffs2_get_sb_mtd(struct file_system_type *fs_type,
 	sb->s_op = &jffs2_super_operations;
 	sb->s_flags = flags | MS_NOATIME;
 
-	ret = jffs2_do_fill_super(sb, data, (flags&MS_VERBOSE)?1:0);
+	ret = jffs2_do_fill_super(sb, data, flags & MS_SILENT ? 1 : 0);
 
 	if (ret) {
 		/* Failure case... */
@@ -257,7 +257,7 @@ static struct super_block *jffs2_get_sb(struct file_system_type *fs_type,
 	}
 
 	if (imajor(nd.dentry->d_inode) != MTD_BLOCK_MAJOR) {
-		if (!(flags & MS_VERBOSE)) /* Yes I mean this. Strangely */
+		if (!(flags & MS_SILENT))
 			printk(KERN_NOTICE "Attempt to mount non-MTD device \"%s\" as JFFS2\n",
 			       dev_name);
 		goto out;
@@ -331,7 +331,8 @@ static int __init init_jffs2_fs(void)
 
 	jffs2_inode_cachep = kmem_cache_create("jffs2_i",
 					     sizeof(struct jffs2_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     jffs2_i_init_once, NULL);
 	if (!jffs2_inode_cachep) {
 		printk(KERN_ERR "JFFS2 error: Failed to initialise inode cache\n");
diff --git a/fs/jfs/Makefile b/fs/jfs/Makefile
index 6f1e0e95587a..3adb6395e42d 100644
--- a/fs/jfs/Makefile
+++ b/fs/jfs/Makefile
@@ -8,7 +8,8 @@ jfs-y    := super.o file.o inode.o namei.o jfs_mount.o jfs_umount.o \
 	    jfs_xtree.o jfs_imap.o jfs_debug.o jfs_dmap.o \
 	    jfs_unicode.o jfs_dtree.o jfs_inode.o \
 	    jfs_extent.o symlink.o jfs_metapage.o \
-	    jfs_logmgr.o jfs_txnmgr.o jfs_uniupr.o resize.o xattr.o
+	    jfs_logmgr.o jfs_txnmgr.o jfs_uniupr.o \
+	    resize.o xattr.o ioctl.o
 
 jfs-$(CONFIG_JFS_POSIX_ACL) += acl.o
 
diff --git a/fs/jfs/acl.c b/fs/jfs/acl.c
index 461e4934ca7c..e2281300979c 100644
--- a/fs/jfs/acl.c
+++ b/fs/jfs/acl.c
@@ -183,6 +183,9 @@ cleanup:
 		posix_acl_release(acl);
 	} else
 		inode->i_mode &= ~current->fs->umask;
+	
+	JFS_IP(inode)->mode2 = (JFS_IP(inode)->mode2 & 0xffff0000) |
+			       inode->i_mode;
 
 	return rc;
 }
@@ -207,12 +210,12 @@ static int jfs_acl_chmod(struct inode *inode)
 	rc = posix_acl_chmod_masq(clone, inode->i_mode);
 	if (!rc) {
 		tid_t tid = txBegin(inode->i_sb, 0);
-		down(&JFS_IP(inode)->commit_sem);
+		mutex_lock(&JFS_IP(inode)->commit_mutex);
 		rc = jfs_set_acl(tid, inode, ACL_TYPE_ACCESS, clone);
 		if (!rc)
 			rc = txCommit(tid, 1, &inode, 0);
 		txEnd(tid);
-		up(&JFS_IP(inode)->commit_sem);
+		mutex_unlock(&JFS_IP(inode)->commit_mutex);
 	}
 
 	posix_acl_release(clone);
diff --git a/fs/jfs/file.c b/fs/jfs/file.c
index c2c19c9ed9a4..1c9745be5ada 100644
--- a/fs/jfs/file.c
+++ b/fs/jfs/file.c
@@ -100,7 +100,7 @@ struct inode_operations jfs_file_inode_operations = {
 #endif
 };
 
-struct file_operations jfs_file_operations = {
+const struct file_operations jfs_file_operations = {
 	.open		= jfs_open,
 	.llseek		= generic_file_llseek,
 	.write		= generic_file_write,
@@ -113,4 +113,5 @@ struct file_operations jfs_file_operations = {
  	.sendfile	= generic_file_sendfile,
 	.fsync		= jfs_fsync,
 	.release	= jfs_release,
+	.ioctl		= jfs_ioctl,
 };
diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c
index 9f942ca8e4e3..04eb78f1252e 100644
--- a/fs/jfs/inode.c
+++ b/fs/jfs/inode.c
@@ -55,6 +55,7 @@ void jfs_read_inode(struct inode *inode)
 		inode->i_op = &jfs_file_inode_operations;
 		init_special_inode(inode, inode->i_mode, inode->i_rdev);
 	}
+	jfs_set_inode_flags(inode);
 }
 
 /*
@@ -89,16 +90,16 @@ int jfs_commit_inode(struct inode *inode, int wait)
 	}
 
 	tid = txBegin(inode->i_sb, COMMIT_INODE);
-	down(&JFS_IP(inode)->commit_sem);
+	mutex_lock(&JFS_IP(inode)->commit_mutex);
 
 	/*
-	 * Retest inode state after taking commit_sem
+	 * Retest inode state after taking commit_mutex
 	 */
 	if (inode->i_nlink && test_cflag(COMMIT_Dirty, inode))
 		rc = txCommit(tid, 1, &inode, wait ? COMMIT_SYNC : 0);
 
 	txEnd(tid);
-	up(&JFS_IP(inode)->commit_sem);
+	mutex_unlock(&JFS_IP(inode)->commit_mutex);
 	return rc;
 }
 
@@ -257,7 +258,8 @@ jfs_get_blocks(struct inode *ip, sector_t lblock, unsigned long max_blocks,
 static int jfs_get_block(struct inode *ip, sector_t lblock,
 			 struct buffer_head *bh_result, int create)
 {
-	return jfs_get_blocks(ip, lblock, 1, bh_result, create);
+	return jfs_get_blocks(ip, lblock, bh_result->b_size >> ip->i_blkbits,
+			bh_result, create);
 }
 
 static int jfs_writepage(struct page *page, struct writeback_control *wbc)
@@ -300,7 +302,7 @@ static ssize_t jfs_direct_IO(int rw, struct kiocb *iocb,
 	struct inode *inode = file->f_mapping->host;
 
 	return blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
-				offset, nr_segs, jfs_get_blocks, NULL);
+				offset, nr_segs, jfs_get_block, NULL);
 }
 
 struct address_space_operations jfs_aops = {
@@ -335,18 +337,18 @@ void jfs_truncate_nolock(struct inode *ip, loff_t length)
 		tid = txBegin(ip->i_sb, 0);
 
 		/*
-		 * The commit_sem cannot be taken before txBegin.
+		 * The commit_mutex cannot be taken before txBegin.
 		 * txBegin may block and there is a chance the inode
 		 * could be marked dirty and need to be committed
 		 * before txBegin unblocks
 		 */
-		down(&JFS_IP(ip)->commit_sem);
+		mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 		newsize = xtTruncate(tid, ip, length,
 				     COMMIT_TRUNCATE | COMMIT_PWMAP);
 		if (newsize < 0) {
 			txEnd(tid);
-			up(&JFS_IP(ip)->commit_sem);
+			mutex_unlock(&JFS_IP(ip)->commit_mutex);
 			break;
 		}
 
@@ -355,7 +357,7 @@ void jfs_truncate_nolock(struct inode *ip, loff_t length)
 
 		txCommit(tid, 1, &ip, 0);
 		txEnd(tid);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	} while (newsize > length);	/* Truncate isn't always atomic */
 }
 
diff --git a/fs/jfs/ioctl.c b/fs/jfs/ioctl.c
new file mode 100644
index 000000000000..67b3774820eb
--- /dev/null
+++ b/fs/jfs/ioctl.c
@@ -0,0 +1,107 @@
+/*
+ * linux/fs/jfs/ioctl.c
+ *
+ * Copyright (C) 2006 Herbert Poetzl
+ * adapted from Remy Card's ext2/ioctl.c
+ */
+
+#include <linux/fs.h>
+#include <linux/ext2_fs.h>
+#include <linux/ctype.h>
+#include <linux/capability.h>
+#include <linux/time.h>
+#include <asm/current.h>
+#include <asm/uaccess.h>
+
+#include "jfs_incore.h"
+#include "jfs_dinode.h"
+#include "jfs_inode.h"
+
+
+static struct {
+	long jfs_flag;
+	long ext2_flag;
+} jfs_map[] = {
+	{JFS_NOATIME_FL, EXT2_NOATIME_FL},
+	{JFS_DIRSYNC_FL, EXT2_DIRSYNC_FL},
+	{JFS_SYNC_FL, EXT2_SYNC_FL},
+	{JFS_SECRM_FL, EXT2_SECRM_FL},
+	{JFS_UNRM_FL, EXT2_UNRM_FL},
+	{JFS_APPEND_FL, EXT2_APPEND_FL},
+	{JFS_IMMUTABLE_FL, EXT2_IMMUTABLE_FL},
+	{0, 0},
+};
+
+static long jfs_map_ext2(unsigned long flags, int from)
+{
+	int index=0;
+	long mapped=0;
+
+	while (jfs_map[index].jfs_flag) {
+		if (from) {
+			if (jfs_map[index].ext2_flag & flags)
+				mapped |= jfs_map[index].jfs_flag;
+		} else {
+			if (jfs_map[index].jfs_flag & flags)
+				mapped |= jfs_map[index].ext2_flag;
+		}
+		index++;
+	}
+	return mapped;
+}
+
+
+int jfs_ioctl(struct inode * inode, struct file * filp, unsigned int cmd,
+		unsigned long arg)
+{
+	struct jfs_inode_info *jfs_inode = JFS_IP(inode);
+	unsigned int flags;
+
+	switch (cmd) {
+	case JFS_IOC_GETFLAGS:
+		flags = jfs_inode->mode2 & JFS_FL_USER_VISIBLE;
+		flags = jfs_map_ext2(flags, 0);
+		return put_user(flags, (int __user *) arg);
+	case JFS_IOC_SETFLAGS: {
+		unsigned int oldflags;
+
+		if (IS_RDONLY(inode))
+			return -EROFS;
+
+		if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER))
+			return -EACCES;
+
+		if (get_user(flags, (int __user *) arg))
+			return -EFAULT;
+
+		flags = jfs_map_ext2(flags, 1);
+		if (!S_ISDIR(inode->i_mode))
+			flags &= ~JFS_DIRSYNC_FL;
+
+		oldflags = jfs_inode->mode2;
+
+		/*
+		 * The IMMUTABLE and APPEND_ONLY flags can only be changed by
+		 * the relevant capability.
+		 */
+		if ((oldflags & JFS_IMMUTABLE_FL) ||
+			((flags ^ oldflags) &
+			(JFS_APPEND_FL | JFS_IMMUTABLE_FL))) {
+			if (!capable(CAP_LINUX_IMMUTABLE))
+				return -EPERM;
+		}
+
+		flags = flags & JFS_FL_USER_MODIFIABLE;
+		flags |= oldflags & ~JFS_FL_USER_MODIFIABLE;
+		jfs_inode->mode2 = flags;
+
+		jfs_set_inode_flags(inode);
+		inode->i_ctime = CURRENT_TIME_SEC;
+		mark_inode_dirty(inode);
+		return 0;
+	}
+	default:
+		return -ENOTTY;
+	}
+}
+
diff --git a/fs/jfs/jfs_debug.c b/fs/jfs/jfs_debug.c
index 4caea6b43b92..81f0e514c490 100644
--- a/fs/jfs/jfs_debug.c
+++ b/fs/jfs/jfs_debug.c
@@ -111,7 +111,7 @@ static struct {
 	{ "loglevel",	loglevel_read, loglevel_write }
 #endif
 };
-#define NPROCENT	(sizeof(Entries)/sizeof(Entries[0]))
+#define NPROCENT	ARRAY_SIZE(Entries)
 
 void jfs_proc_init(void)
 {
diff --git a/fs/jfs/jfs_dinode.h b/fs/jfs/jfs_dinode.h
index 580a3258449b..9f2572aea561 100644
--- a/fs/jfs/jfs_dinode.h
+++ b/fs/jfs/jfs_dinode.h
@@ -139,13 +139,36 @@ struct dinode {
 
 /* more extended mode bits: attributes for OS/2 */
 #define IREADONLY	0x02000000	/* no write access to file */
-#define IARCHIVE	0x40000000	/* file archive bit */
-#define ISYSTEM		0x08000000	/* system file */
 #define IHIDDEN		0x04000000	/* hidden file */
-#define IRASH		0x4E000000	/* mask for changeable attributes */
-#define INEWNAME	0x80000000	/* non-8.3 filename format */
+#define ISYSTEM		0x08000000	/* system file */
+
 #define IDIRECTORY	0x20000000	/* directory (shadow of real bit) */
+#define IARCHIVE	0x40000000	/* file archive bit */
+#define INEWNAME	0x80000000	/* non-8.3 filename format */
+
+#define IRASH		0x4E000000	/* mask for changeable attributes */
 #define ATTRSHIFT	25	/* bits to shift to move attribute
 				   specification to mode position */
 
+/* extended attributes for Linux */
+
+#define JFS_NOATIME_FL		0x00080000 /* do not update atime */
+
+#define JFS_DIRSYNC_FL		0x00100000 /* dirsync behaviour */
+#define JFS_SYNC_FL		0x00200000 /* Synchronous updates */
+#define JFS_SECRM_FL		0x00400000 /* Secure deletion */
+#define JFS_UNRM_FL		0x00800000 /* allow for undelete */
+
+#define JFS_APPEND_FL		0x01000000 /* writes to file may only append */
+#define JFS_IMMUTABLE_FL	0x02000000 /* Immutable file */
+
+#define JFS_FL_USER_VISIBLE	0x03F80000
+#define JFS_FL_USER_MODIFIABLE	0x03F80000
+#define JFS_FL_INHERIT		0x03C80000
+
+/* These are identical to EXT[23]_IOC_GETFLAGS/SETFLAGS */
+#define JFS_IOC_GETFLAGS	_IOR('f', 1, long)
+#define JFS_IOC_SETFLAGS	_IOW('f', 2, long)
+
+
 #endif /*_H_JFS_DINODE */
diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c
index 79b5404db100..c161c98954e0 100644
--- a/fs/jfs/jfs_dmap.c
+++ b/fs/jfs/jfs_dmap.c
@@ -64,9 +64,9 @@
  *	to the persistent bitmaps in dmaps) is guarded by (busy) buffers.
  */
 
-#define BMAP_LOCK_INIT(bmp)	init_MUTEX(&bmp->db_bmaplock)
-#define BMAP_LOCK(bmp)		down(&bmp->db_bmaplock)
-#define BMAP_UNLOCK(bmp)	up(&bmp->db_bmaplock)
+#define BMAP_LOCK_INIT(bmp)	mutex_init(&bmp->db_bmaplock)
+#define BMAP_LOCK(bmp)		mutex_lock(&bmp->db_bmaplock)
+#define BMAP_UNLOCK(bmp)	mutex_unlock(&bmp->db_bmaplock)
 
 /*
  * forward references
@@ -125,7 +125,7 @@ static int dbGetL2AGSize(s64 nblocks);
  * into the table, with the table elements yielding the maximum
  * binary buddy of free bits within the character.
  */
-static s8 budtab[256] = {
+static const s8 budtab[256] = {
 	3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
 	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 	2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
diff --git a/fs/jfs/jfs_dmap.h b/fs/jfs/jfs_dmap.h
index 32e25884e7e8..8b14cc8e0228 100644
--- a/fs/jfs/jfs_dmap.h
+++ b/fs/jfs/jfs_dmap.h
@@ -243,7 +243,7 @@ struct dbmap {
 struct bmap {
 	struct dbmap db_bmap;		/* on-disk aggregate map descriptor */
 	struct inode *db_ipbmap;	/* ptr to aggregate map incore inode */
-	struct semaphore db_bmaplock;	/* aggregate map lock */
+	struct mutex db_bmaplock;	/* aggregate map lock */
 	atomic_t db_active[MAXAG];	/* count of active, open files in AG */
 	u32 *db_DBmap;
 };
diff --git a/fs/jfs/jfs_dtree.c b/fs/jfs/jfs_dtree.c
index 404f33eae507..6c3f08319846 100644
--- a/fs/jfs/jfs_dtree.c
+++ b/fs/jfs/jfs_dtree.c
@@ -1005,6 +1005,9 @@ static int dtSplitUp(tid_t tid,
 
 		DT_PUTPAGE(smp);
 
+		if (!DO_INDEX(ip))
+			ip->i_size = xlen << sbi->l2bsize;
+
 		goto freeKeyName;
 	}
 
@@ -1055,7 +1058,9 @@ static int dtSplitUp(tid_t tid,
 				xaddr = addressPXD(pxd) + xlen;
 				dbFree(ip, xaddr, (s64) n);
 			}
-		}
+		} else if (!DO_INDEX(ip))
+			ip->i_size = lengthPXD(pxd) << sbi->l2bsize;
+
 
 	      extendOut:
 		DT_PUTPAGE(smp);
@@ -1098,6 +1103,9 @@ static int dtSplitUp(tid_t tid,
 		goto splitOut;
 	}
 
+	if (!DO_INDEX(ip))
+		ip->i_size += PSIZE;
+
 	/*
 	 * propagate up the router entry for the leaf page just split
 	 *
@@ -2424,6 +2432,9 @@ static int dtDeleteUp(tid_t tid, struct inode *ip,
 		break;
 	}
 
+	if (!DO_INDEX(ip))
+		ip->i_size -= PSIZE;
+
 	return 0;
 }
 
diff --git a/fs/jfs/jfs_extent.c b/fs/jfs/jfs_extent.c
index 4879603daa1c..5549378358bf 100644
--- a/fs/jfs/jfs_extent.c
+++ b/fs/jfs/jfs_extent.c
@@ -94,7 +94,7 @@ extAlloc(struct inode *ip, s64 xlen, s64 pno, xad_t * xp, boolean_t abnr)
 	txBeginAnon(ip->i_sb);
 
 	/* Avoid race with jfs_commit_inode() */
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	/* validate extent length */
 	if (xlen > MAXXLEN)
@@ -136,14 +136,14 @@ extAlloc(struct inode *ip, s64 xlen, s64 pno, xad_t * xp, boolean_t abnr)
 	 */
 	nxlen = xlen;
 	if ((rc = extBalloc(ip, hint ? hint : INOHINT(ip), &nxlen, &nxaddr))) {
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 		return (rc);
 	}
 
 	/* Allocate blocks to quota. */
 	if (DQUOT_ALLOC_BLOCK(ip, nxlen)) {
 		dbFree(ip, nxaddr, (s64) nxlen);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 		return -EDQUOT;
 	}
 
@@ -165,7 +165,7 @@ extAlloc(struct inode *ip, s64 xlen, s64 pno, xad_t * xp, boolean_t abnr)
 	if (rc) {
 		dbFree(ip, nxaddr, nxlen);
 		DQUOT_FREE_BLOCK(ip, nxlen);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 		return (rc);
 	}
 
@@ -177,7 +177,7 @@ extAlloc(struct inode *ip, s64 xlen, s64 pno, xad_t * xp, boolean_t abnr)
 
 	mark_inode_dirty(ip);
 
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	/*
 	 * COMMIT_SyncList flags an anonymous tlock on page that is on
 	 * sync list.
@@ -222,7 +222,7 @@ int extRealloc(struct inode *ip, s64 nxlen, xad_t * xp, boolean_t abnr)
 	/* This blocks if we are low on resources */
 	txBeginAnon(ip->i_sb);
 
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 	/* validate extent length */
 	if (nxlen > MAXXLEN)
 		nxlen = MAXXLEN;
@@ -258,7 +258,7 @@ int extRealloc(struct inode *ip, s64 nxlen, xad_t * xp, boolean_t abnr)
 	/* Allocat blocks to quota. */
 	if (DQUOT_ALLOC_BLOCK(ip, nxlen)) {
 		dbFree(ip, nxaddr, (s64) nxlen);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 		return -EDQUOT;
 	}
 
@@ -338,7 +338,7 @@ int extRealloc(struct inode *ip, s64 nxlen, xad_t * xp, boolean_t abnr)
 
 	mark_inode_dirty(ip);
 exit:
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	return (rc);
 }
 #endif			/* _NOTYET */
@@ -439,12 +439,12 @@ int extRecord(struct inode *ip, xad_t * xp)
 
 	txBeginAnon(ip->i_sb);
 
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	/* update the extent */
 	rc = xtUpdate(0, ip, xp);
 
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	return rc;
 }
 
diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c
index 4efa0d0eec39..ccbe60aff83d 100644
--- a/fs/jfs/jfs_imap.c
+++ b/fs/jfs/jfs_imap.c
@@ -66,14 +66,14 @@ static HLIST_HEAD(aggregate_hash);
  * imap locks
  */
 /* iag free list lock */
-#define IAGFREE_LOCK_INIT(imap)		init_MUTEX(&imap->im_freelock)
-#define IAGFREE_LOCK(imap)		down(&imap->im_freelock)
-#define IAGFREE_UNLOCK(imap)		up(&imap->im_freelock)
+#define IAGFREE_LOCK_INIT(imap)		mutex_init(&imap->im_freelock)
+#define IAGFREE_LOCK(imap)		mutex_lock(&imap->im_freelock)
+#define IAGFREE_UNLOCK(imap)		mutex_unlock(&imap->im_freelock)
 
 /* per ag iag list locks */
-#define AG_LOCK_INIT(imap,index)	init_MUTEX(&(imap->im_aglock[index]))
-#define AG_LOCK(imap,agno)		down(&imap->im_aglock[agno])
-#define AG_UNLOCK(imap,agno)		up(&imap->im_aglock[agno])
+#define AG_LOCK_INIT(imap,index)	mutex_init(&(imap->im_aglock[index]))
+#define AG_LOCK(imap,agno)		mutex_lock(&imap->im_aglock[agno])
+#define AG_UNLOCK(imap,agno)		mutex_unlock(&imap->im_aglock[agno])
 
 /*
  * forward references
@@ -1261,7 +1261,7 @@ int diFree(struct inode *ip)
 	 * to be freed by the transaction;  
 	 */
 	tid = txBegin(ipimap->i_sb, COMMIT_FORCE);
-	down(&JFS_IP(ipimap)->commit_sem);
+	mutex_lock(&JFS_IP(ipimap)->commit_mutex);
 
 	/* acquire tlock of the iag page of the freed ixad 
 	 * to force the page NOHOMEOK (even though no data is
@@ -1294,7 +1294,7 @@ int diFree(struct inode *ip)
 	rc = txCommit(tid, 1, &iplist[0], COMMIT_FORCE);
 
 	txEnd(tid);
-	up(&JFS_IP(ipimap)->commit_sem);
+	mutex_unlock(&JFS_IP(ipimap)->commit_mutex);
 
 	/* unlock the AG inode map information */
 	AG_UNLOCK(imap, agno);
@@ -2554,13 +2554,13 @@ diNewIAG(struct inomap * imap, int *iagnop, int agno, struct metapage ** mpp)
 		 * addressing structure pointing to the new iag page;
 		 */
 		tid = txBegin(sb, COMMIT_FORCE);
-		down(&JFS_IP(ipimap)->commit_sem);
+		mutex_lock(&JFS_IP(ipimap)->commit_mutex);
 
 		/* update the inode map addressing structure to point to it */
 		if ((rc =
 		     xtInsert(tid, ipimap, 0, blkno, xlen, &xaddr, 0))) {
 			txEnd(tid);
-			up(&JFS_IP(ipimap)->commit_sem);
+			mutex_unlock(&JFS_IP(ipimap)->commit_mutex);
 			/* Free the blocks allocated for the iag since it was
 			 * not successfully added to the inode map
 			 */
@@ -2626,7 +2626,7 @@ diNewIAG(struct inomap * imap, int *iagnop, int agno, struct metapage ** mpp)
 		rc = txCommit(tid, 1, &iplist[0], COMMIT_FORCE);
 
 		txEnd(tid);
-		up(&JFS_IP(ipimap)->commit_sem);
+		mutex_unlock(&JFS_IP(ipimap)->commit_mutex);
 
 		duplicateIXtree(sb, blkno, xlen, &xaddr);
 
@@ -3074,14 +3074,40 @@ static void duplicateIXtree(struct super_block *sb, s64 blkno,
 static int copy_from_dinode(struct dinode * dip, struct inode *ip)
 {
 	struct jfs_inode_info *jfs_ip = JFS_IP(ip);
+	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
 
 	jfs_ip->fileset = le32_to_cpu(dip->di_fileset);
 	jfs_ip->mode2 = le32_to_cpu(dip->di_mode);
 
 	ip->i_mode = le32_to_cpu(dip->di_mode) & 0xffff;
+	if (sbi->umask != -1) {
+		ip->i_mode = (ip->i_mode & ~0777) | (0777 & ~sbi->umask);
+		/* For directories, add x permission if r is allowed by umask */
+		if (S_ISDIR(ip->i_mode)) {
+			if (ip->i_mode & 0400)
+				ip->i_mode |= 0100;
+			if (ip->i_mode & 0040)
+				ip->i_mode |= 0010;
+			if (ip->i_mode & 0004)
+				ip->i_mode |= 0001;
+		}
+	}
 	ip->i_nlink = le32_to_cpu(dip->di_nlink);
-	ip->i_uid = le32_to_cpu(dip->di_uid);
-	ip->i_gid = le32_to_cpu(dip->di_gid);
+
+	jfs_ip->saved_uid = le32_to_cpu(dip->di_uid);
+	if (sbi->uid == -1)
+		ip->i_uid = jfs_ip->saved_uid;
+	else {
+		ip->i_uid = sbi->uid;
+	}
+
+	jfs_ip->saved_gid = le32_to_cpu(dip->di_gid);
+	if (sbi->gid == -1)
+		ip->i_gid = jfs_ip->saved_gid;
+	else {
+		ip->i_gid = sbi->gid;
+	}
+
 	ip->i_size = le64_to_cpu(dip->di_size);
 	ip->i_atime.tv_sec = le32_to_cpu(dip->di_atime.tv_sec);
 	ip->i_atime.tv_nsec = le32_to_cpu(dip->di_atime.tv_nsec);
@@ -3132,21 +3158,33 @@ static int copy_from_dinode(struct dinode * dip, struct inode *ip)
 static void copy_to_dinode(struct dinode * dip, struct inode *ip)
 {
 	struct jfs_inode_info *jfs_ip = JFS_IP(ip);
+	struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb);
 
 	dip->di_fileset = cpu_to_le32(jfs_ip->fileset);
-	dip->di_inostamp = cpu_to_le32(JFS_SBI(ip->i_sb)->inostamp);
+	dip->di_inostamp = cpu_to_le32(sbi->inostamp);
 	dip->di_number = cpu_to_le32(ip->i_ino);
 	dip->di_gen = cpu_to_le32(ip->i_generation);
 	dip->di_size = cpu_to_le64(ip->i_size);
 	dip->di_nblocks = cpu_to_le64(PBLK2LBLK(ip->i_sb, ip->i_blocks));
 	dip->di_nlink = cpu_to_le32(ip->i_nlink);
-	dip->di_uid = cpu_to_le32(ip->i_uid);
-	dip->di_gid = cpu_to_le32(ip->i_gid);
+	if (sbi->uid == -1)
+		dip->di_uid = cpu_to_le32(ip->i_uid);
+	else
+		dip->di_uid = cpu_to_le32(jfs_ip->saved_uid);
+	if (sbi->gid == -1)
+		dip->di_gid = cpu_to_le32(ip->i_gid);
+	else
+		dip->di_gid = cpu_to_le32(jfs_ip->saved_gid);
 	/*
 	 * mode2 is only needed for storing the higher order bits.
 	 * Trust i_mode for the lower order ones
 	 */
-	dip->di_mode = cpu_to_le32((jfs_ip->mode2 & 0xffff0000) | ip->i_mode);
+	if (sbi->umask == -1)
+		dip->di_mode = cpu_to_le32((jfs_ip->mode2 & 0xffff0000) |
+					   ip->i_mode);
+	else /* Leave the original permissions alone */
+		dip->di_mode = cpu_to_le32(jfs_ip->mode2);
+
 	dip->di_atime.tv_sec = cpu_to_le32(ip->i_atime.tv_sec);
 	dip->di_atime.tv_nsec = cpu_to_le32(ip->i_atime.tv_nsec);
 	dip->di_ctime.tv_sec = cpu_to_le32(ip->i_ctime.tv_sec);
diff --git a/fs/jfs/jfs_imap.h b/fs/jfs/jfs_imap.h
index 6b59adec036a..6e24465f0f98 100644
--- a/fs/jfs/jfs_imap.h
+++ b/fs/jfs/jfs_imap.h
@@ -140,8 +140,8 @@ struct dinomap {
 struct inomap {
 	struct dinomap im_imap;		/* 4096: inode allocation control */
 	struct inode *im_ipimap;	/* 4: ptr to inode for imap   */
-	struct semaphore im_freelock;	/* 4: iag free list lock      */
-	struct semaphore im_aglock[MAXAG];	/* 512: per AG locks          */
+	struct mutex im_freelock;	/* 4: iag free list lock      */
+	struct mutex im_aglock[MAXAG];	/* 512: per AG locks          */
 	u32 *im_DBGdimap;
 	atomic_t im_numinos;	/* num of backed inodes */
 	atomic_t im_numfree;	/* num of free backed inodes */
diff --git a/fs/jfs/jfs_incore.h b/fs/jfs/jfs_incore.h
index dc21a5bd54d4..54d73716ca8c 100644
--- a/fs/jfs/jfs_incore.h
+++ b/fs/jfs/jfs_incore.h
@@ -19,6 +19,7 @@
 #ifndef _H_JFS_INCORE
 #define _H_JFS_INCORE
 
+#include <linux/mutex.h>
 #include <linux/rwsem.h>
 #include <linux/slab.h>
 #include <linux/bitops.h>
@@ -37,6 +38,8 @@
 struct jfs_inode_info {
 	int	fileset;	/* fileset number (always 16)*/
 	uint	mode2;		/* jfs-specific mode		*/
+	uint	saved_uid;	/* saved for uid mount option */
+	uint	saved_gid;	/* saved for gid mount option */
 	pxd_t   ixpxd;		/* inode extent descriptor	*/
 	dxd_t	acl;		/* dxd describing acl	*/
 	dxd_t	ea;		/* dxd describing ea	*/
@@ -62,12 +65,12 @@ struct jfs_inode_info {
 	 */
 	struct rw_semaphore rdwrlock;
 	/*
-	 * commit_sem serializes transaction processing on an inode.
+	 * commit_mutex serializes transaction processing on an inode.
 	 * It must be taken after beginning a transaction (txBegin), since
 	 * dirty inodes may be committed while a new transaction on the
 	 * inode is blocked in txBegin or TxBeginAnon
 	 */
-	struct semaphore commit_sem;
+	struct mutex commit_mutex;
 	/* xattr_sem allows us to access the xattrs without taking i_mutex */
 	struct rw_semaphore xattr_sem;
 	lid_t	xtlid;		/* lid of xtree lock on directory */
@@ -169,6 +172,9 @@ struct jfs_sb_info {
 	uint		state;		/* mount/recovery state	*/
 	unsigned long	flag;		/* mount time flags */
 	uint		p_state;	/* state prior to going no integrity */
+	uint		uid;		/* uid to override on-disk uid */
+	uint		gid;		/* gid to override on-disk gid */
+	uint		umask;		/* umask to override on-disk umask */
 };
 
 /* jfs_sb_info commit_state */
diff --git a/fs/jfs/jfs_inode.c b/fs/jfs/jfs_inode.c
index 2af5efbfd06f..495df402916d 100644
--- a/fs/jfs/jfs_inode.c
+++ b/fs/jfs/jfs_inode.c
@@ -25,6 +25,26 @@
 #include "jfs_dinode.h"
 #include "jfs_debug.h"
 
+
+void jfs_set_inode_flags(struct inode *inode)
+{
+	unsigned int flags = JFS_IP(inode)->mode2;
+
+	inode->i_flags &= ~(S_IMMUTABLE | S_APPEND |
+		S_NOATIME | S_DIRSYNC | S_SYNC);
+
+	if (flags & JFS_IMMUTABLE_FL)
+		inode->i_flags |= S_IMMUTABLE;
+	if (flags & JFS_APPEND_FL)
+		inode->i_flags |= S_APPEND;
+	if (flags & JFS_NOATIME_FL)
+		inode->i_flags |= S_NOATIME;
+	if (flags & JFS_DIRSYNC_FL)
+		inode->i_flags |= S_DIRSYNC;
+	if (flags & JFS_SYNC_FL)
+		inode->i_flags |= S_SYNC;
+}
+
 /*
  * NAME:	ialloc()
  *
@@ -63,6 +83,13 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 		inode->i_gid = current->fsgid;
 
 	/*
+	 * New inodes need to save sane values on disk when
+	 * uid & gid mount options are used
+	 */
+	jfs_inode->saved_uid = inode->i_uid;
+	jfs_inode->saved_gid = inode->i_gid;
+
+	/*
 	 * Allocate inode to quota.
 	 */
 	if (DQUOT_ALLOC_INODE(inode)) {
@@ -74,10 +101,20 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 	}
 
 	inode->i_mode = mode;
-	if (S_ISDIR(mode))
-		jfs_inode->mode2 = IDIRECTORY | mode;
-	else
-		jfs_inode->mode2 = INLINEEA | ISPARSE | mode;
+	/* inherit flags from parent */
+	jfs_inode->mode2 = JFS_IP(parent)->mode2 & JFS_FL_INHERIT;
+
+	if (S_ISDIR(mode)) {
+		jfs_inode->mode2 |= IDIRECTORY;
+		jfs_inode->mode2 &= ~JFS_DIRSYNC_FL;
+	}
+	else {
+		jfs_inode->mode2 |= INLINEEA | ISPARSE;
+		if (S_ISLNK(mode))
+			jfs_inode->mode2 &= ~(JFS_IMMUTABLE_FL|JFS_APPEND_FL);
+	}
+	jfs_inode->mode2 |= mode;
+
 	inode->i_blksize = sb->s_blocksize;
 	inode->i_blocks = 0;
 	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
@@ -98,6 +135,7 @@ struct inode *ialloc(struct inode *parent, umode_t mode)
 	jfs_inode->atlhead = 0;
 	jfs_inode->atltail = 0;
 	jfs_inode->xtlid = 0;
+	jfs_set_inode_flags(inode);
 
 	jfs_info("ialloc returns inode = 0x%p\n", inode);
 
diff --git a/fs/jfs/jfs_inode.h b/fs/jfs/jfs_inode.h
index b54bac576cb3..c30072674464 100644
--- a/fs/jfs/jfs_inode.h
+++ b/fs/jfs/jfs_inode.h
@@ -20,6 +20,8 @@
 
 extern struct inode *ialloc(struct inode *, umode_t);
 extern int jfs_fsync(struct file *, struct dentry *, int);
+extern int jfs_ioctl(struct inode *, struct file *,
+			unsigned int, unsigned long);
 extern void jfs_read_inode(struct inode *);
 extern int jfs_commit_inode(struct inode *, int);
 extern int jfs_write_inode(struct inode*, int);
@@ -29,12 +31,13 @@ extern void jfs_truncate(struct inode *);
 extern void jfs_truncate_nolock(struct inode *, loff_t);
 extern void jfs_free_zero_link(struct inode *);
 extern struct dentry *jfs_get_parent(struct dentry *dentry);
+extern void jfs_set_inode_flags(struct inode *);
 
 extern struct address_space_operations jfs_aops;
 extern struct inode_operations jfs_dir_inode_operations;
-extern struct file_operations jfs_dir_operations;
+extern const struct file_operations jfs_dir_operations;
 extern struct inode_operations jfs_file_inode_operations;
-extern struct file_operations jfs_file_operations;
+extern const struct file_operations jfs_file_operations;
 extern struct inode_operations jfs_symlink_inode_operations;
 extern struct dentry_operations jfs_ci_dentry_operations;
 #endif				/* _H_JFS_INODE */
diff --git a/fs/jfs/jfs_lock.h b/fs/jfs/jfs_lock.h
index 10ad1d086685..70ac9f7d1e00 100644
--- a/fs/jfs/jfs_lock.h
+++ b/fs/jfs/jfs_lock.h
@@ -20,6 +20,7 @@
 #define _H_JFS_LOCK
 
 #include <linux/spinlock.h>
+#include <linux/mutex.h>
 #include <linux/sched.h>
 
 /*
diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c
index d27bac6acaa3..3315f0b1fbc0 100644
--- a/fs/jfs/jfs_logmgr.c
+++ b/fs/jfs/jfs_logmgr.c
@@ -64,10 +64,12 @@
 #include <linux/interrupt.h>
 #include <linux/smp_lock.h>
 #include <linux/completion.h>
+#include <linux/kthread.h>
 #include <linux/buffer_head.h>		/* for sync_blockdev() */
 #include <linux/bio.h>
 #include <linux/suspend.h>
 #include <linux/delay.h>
+#include <linux/mutex.h>
 #include "jfs_incore.h"
 #include "jfs_filsys.h"
 #include "jfs_metapage.h"
@@ -81,15 +83,14 @@
  */
 static struct lbuf *log_redrive_list;
 static DEFINE_SPINLOCK(log_redrive_lock);
-DECLARE_WAIT_QUEUE_HEAD(jfs_IO_thread_wait);
 
 
 /*
  *	log read/write serialization (per log)
  */
-#define LOG_LOCK_INIT(log)	init_MUTEX(&(log)->loglock)
-#define LOG_LOCK(log)		down(&((log)->loglock))
-#define LOG_UNLOCK(log)		up(&((log)->loglock))
+#define LOG_LOCK_INIT(log)	mutex_init(&(log)->loglock)
+#define LOG_LOCK(log)		mutex_lock(&((log)->loglock))
+#define LOG_UNLOCK(log)		mutex_unlock(&((log)->loglock))
 
 
 /*
@@ -165,7 +166,7 @@ do {						\
  */
 static LIST_HEAD(jfs_external_logs);
 static struct jfs_log *dummy_log = NULL;
-static DECLARE_MUTEX(jfs_log_sem);
+static DEFINE_MUTEX(jfs_log_mutex);
 
 /*
  * forward references
@@ -1085,31 +1086,30 @@ int lmLogOpen(struct super_block *sb)
 	if (sbi->mntflag & JFS_INLINELOG)
 		return open_inline_log(sb);
 
-	down(&jfs_log_sem);
+	mutex_lock(&jfs_log_mutex);
 	list_for_each_entry(log, &jfs_external_logs, journal_list) {
 		if (log->bdev->bd_dev == sbi->logdev) {
 			if (memcmp(log->uuid, sbi->loguuid,
 				   sizeof(log->uuid))) {
 				jfs_warn("wrong uuid on JFS journal\n");
-				up(&jfs_log_sem);
+				mutex_unlock(&jfs_log_mutex);
 				return -EINVAL;
 			}
 			/*
 			 * add file system to log active file system list
 			 */
 			if ((rc = lmLogFileSystem(log, sbi, 1))) {
-				up(&jfs_log_sem);
+				mutex_unlock(&jfs_log_mutex);
 				return rc;
 			}
 			goto journal_found;
 		}
 	}
 
-	if (!(log = kmalloc(sizeof(struct jfs_log), GFP_KERNEL))) {
-		up(&jfs_log_sem);
+	if (!(log = kzalloc(sizeof(struct jfs_log), GFP_KERNEL))) {
+		mutex_unlock(&jfs_log_mutex);
 		return -ENOMEM;
 	}
-	memset(log, 0, sizeof(struct jfs_log));
 	INIT_LIST_HEAD(&log->sb_list);
 	init_waitqueue_head(&log->syncwait);
 
@@ -1152,7 +1152,7 @@ journal_found:
 	sbi->log = log;
 	LOG_UNLOCK(log);
 
-	up(&jfs_log_sem);
+	mutex_unlock(&jfs_log_mutex);
 	return 0;
 
 	/*
@@ -1169,7 +1169,7 @@ journal_found:
 	blkdev_put(bdev);
 
       free:		/* free log descriptor */
-	up(&jfs_log_sem);
+	mutex_unlock(&jfs_log_mutex);
 	kfree(log);
 
 	jfs_warn("lmLogOpen: exit(%d)", rc);
@@ -1181,9 +1181,8 @@ static int open_inline_log(struct super_block *sb)
 	struct jfs_log *log;
 	int rc;
 
-	if (!(log = kmalloc(sizeof(struct jfs_log), GFP_KERNEL)))
+	if (!(log = kzalloc(sizeof(struct jfs_log), GFP_KERNEL)))
 		return -ENOMEM;
-	memset(log, 0, sizeof(struct jfs_log));
 	INIT_LIST_HEAD(&log->sb_list);
 	init_waitqueue_head(&log->syncwait);
 
@@ -1214,14 +1213,13 @@ static int open_dummy_log(struct super_block *sb)
 {
 	int rc;
 
-	down(&jfs_log_sem);
+	mutex_lock(&jfs_log_mutex);
 	if (!dummy_log) {
-		dummy_log = kmalloc(sizeof(struct jfs_log), GFP_KERNEL);
+		dummy_log = kzalloc(sizeof(struct jfs_log), GFP_KERNEL);
 		if (!dummy_log) {
-			up(&jfs_log_sem);
+			mutex_unlock(&jfs_log_mutex);
 			return -ENOMEM;
 		}
-		memset(dummy_log, 0, sizeof(struct jfs_log));
 		INIT_LIST_HEAD(&dummy_log->sb_list);
 		init_waitqueue_head(&dummy_log->syncwait);
 		dummy_log->no_integrity = 1;
@@ -1232,7 +1230,7 @@ static int open_dummy_log(struct super_block *sb)
 		if (rc) {
 			kfree(dummy_log);
 			dummy_log = NULL;
-			up(&jfs_log_sem);
+			mutex_unlock(&jfs_log_mutex);
 			return rc;
 		}
 	}
@@ -1241,7 +1239,7 @@ static int open_dummy_log(struct super_block *sb)
 	list_add(&JFS_SBI(sb)->log_list, &dummy_log->sb_list);
 	JFS_SBI(sb)->log = dummy_log;
 	LOG_UNLOCK(dummy_log);
-	up(&jfs_log_sem);
+	mutex_unlock(&jfs_log_mutex);
 
 	return 0;
 }
@@ -1469,7 +1467,7 @@ int lmLogClose(struct super_block *sb)
 
 	jfs_info("lmLogClose: log:0x%p", log);
 
-	down(&jfs_log_sem);
+	mutex_lock(&jfs_log_mutex);
 	LOG_LOCK(log);
 	list_del(&sbi->log_list);
 	LOG_UNLOCK(log);
@@ -1519,7 +1517,7 @@ int lmLogClose(struct super_block *sb)
 	kfree(log);
 
       out:
-	up(&jfs_log_sem);
+	mutex_unlock(&jfs_log_mutex);
 	jfs_info("lmLogClose: exit(%d)", rc);
 	return rc;
 }
@@ -1980,7 +1978,7 @@ static inline void lbmRedrive(struct lbuf *bp)
 	log_redrive_list = bp;
 	spin_unlock_irqrestore(&log_redrive_lock, flags);
 
-	wake_up(&jfs_IO_thread_wait);
+	wake_up_process(jfsIOthread);
 }
 
 
@@ -2347,13 +2345,7 @@ int jfsIOWait(void *arg)
 {
 	struct lbuf *bp;
 
-	daemonize("jfsIO");
-
-	complete(&jfsIOwait);
-
 	do {
-		DECLARE_WAITQUEUE(wq, current);
-
 		spin_lock_irq(&log_redrive_lock);
 		while ((bp = log_redrive_list) != 0) {
 			log_redrive_list = bp->l_redrive_next;
@@ -2362,21 +2354,19 @@ int jfsIOWait(void *arg)
 			lbmStartIO(bp);
 			spin_lock_irq(&log_redrive_lock);
 		}
+		spin_unlock_irq(&log_redrive_lock);
+
 		if (freezing(current)) {
-			spin_unlock_irq(&log_redrive_lock);
 			refrigerator();
 		} else {
-			add_wait_queue(&jfs_IO_thread_wait, &wq);
 			set_current_state(TASK_INTERRUPTIBLE);
-			spin_unlock_irq(&log_redrive_lock);
 			schedule();
 			current->state = TASK_RUNNING;
-			remove_wait_queue(&jfs_IO_thread_wait, &wq);
 		}
-	} while (!jfs_stop_threads);
+	} while (!kthread_should_stop());
 
 	jfs_info("jfsIOWait being killed!");
-	complete_and_exit(&jfsIOwait, 0);
+	return 0;
 }
 
 /*
diff --git a/fs/jfs/jfs_logmgr.h b/fs/jfs/jfs_logmgr.h
index e4978b5b65ee..8c6909b80014 100644
--- a/fs/jfs/jfs_logmgr.h
+++ b/fs/jfs/jfs_logmgr.h
@@ -389,7 +389,7 @@ struct jfs_log {
 	int eor;		/* 4: eor of last record in eol page */
 	struct lbuf *bp;	/* 4: current log page buffer */
 
-	struct semaphore loglock;	/* 4: log write serialization lock */
+	struct mutex loglock;	/* 4: log write serialization lock */
 
 	/* syncpt */
 	int nextsync;		/* 4: bytes to write before next syncpt */
diff --git a/fs/jfs/jfs_metapage.c b/fs/jfs/jfs_metapage.c
index 8a53981f9f27..f28696f235c4 100644
--- a/fs/jfs/jfs_metapage.c
+++ b/fs/jfs/jfs_metapage.c
@@ -104,10 +104,9 @@ static inline int insert_metapage(struct page *page, struct metapage *mp)
 	if (PagePrivate(page))
 		a = mp_anchor(page);
 	else {
-		a = kmalloc(sizeof(struct meta_anchor), GFP_NOFS);
+		a = kzalloc(sizeof(struct meta_anchor), GFP_NOFS);
 		if (!a)
 			return -ENOMEM;
-		memset(a, 0, sizeof(struct meta_anchor));
 		set_page_private(page, (unsigned long)a);
 		SetPagePrivate(page);
 		kmap(page);
@@ -221,8 +220,8 @@ int __init metapage_init(void)
 	if (metapage_cache == NULL)
 		return -ENOMEM;
 
-	metapage_mempool = mempool_create(METAPOOL_MIN_PAGES, mempool_alloc_slab,
-					  mempool_free_slab, metapage_cache);
+	metapage_mempool = mempool_create_slab_pool(METAPOOL_MIN_PAGES,
+						    metapage_cache);
 
 	if (metapage_mempool == NULL) {
 		kmem_cache_destroy(metapage_cache);
@@ -579,14 +578,13 @@ static int metapage_releasepage(struct page *page, gfp_t gfp_mask)
 	return 0;
 }
 
-static int metapage_invalidatepage(struct page *page, unsigned long offset)
+static void metapage_invalidatepage(struct page *page, unsigned long offset)
 {
 	BUG_ON(offset);
 
-	if (PageWriteback(page))
-		return 0;
+	BUG_ON(PageWriteback(page));
 
-	return metapage_releasepage(page, 0);
+	metapage_releasepage(page, 0);
 }
 
 struct address_space_operations jfs_metapage_aops = {
diff --git a/fs/jfs/jfs_superblock.h b/fs/jfs/jfs_superblock.h
index fcf781bf31cb..682cf1a68a18 100644
--- a/fs/jfs/jfs_superblock.h
+++ b/fs/jfs/jfs_superblock.h
@@ -113,12 +113,9 @@ extern int jfs_mount(struct super_block *);
 extern int jfs_mount_rw(struct super_block *, int);
 extern int jfs_umount(struct super_block *);
 extern int jfs_umount_rw(struct super_block *);
-
-extern int jfs_stop_threads;
-extern struct completion jfsIOwait;
-extern wait_queue_head_t jfs_IO_thread_wait;
-extern wait_queue_head_t jfs_commit_thread_wait;
-extern wait_queue_head_t jfs_sync_thread_wait;
 extern int jfs_extendfs(struct super_block *, s64, int);
 
+extern struct task_struct *jfsIOthread;
+extern struct task_struct *jfsSyncThread;
+
 #endif /*_H_JFS_SUPERBLOCK */
diff --git a/fs/jfs/jfs_txnmgr.c b/fs/jfs/jfs_txnmgr.c
index 2ddb6b892bcf..ac3d66948e8c 100644
--- a/fs/jfs/jfs_txnmgr.c
+++ b/fs/jfs/jfs_txnmgr.c
@@ -49,6 +49,7 @@
 #include <linux/suspend.h>
 #include <linux/module.h>
 #include <linux/moduleparam.h>
+#include <linux/kthread.h>
 #include "jfs_incore.h"
 #include "jfs_inode.h"
 #include "jfs_filsys.h"
@@ -121,8 +122,7 @@ static DEFINE_SPINLOCK(jfsTxnLock);
 #define LAZY_LOCK(flags)	spin_lock_irqsave(&TxAnchor.LazyLock, flags)
 #define LAZY_UNLOCK(flags) spin_unlock_irqrestore(&TxAnchor.LazyLock, flags)
 
-DECLARE_WAIT_QUEUE_HEAD(jfs_sync_thread_wait);
-DECLARE_WAIT_QUEUE_HEAD(jfs_commit_thread_wait);
+static DECLARE_WAIT_QUEUE_HEAD(jfs_commit_thread_wait);
 static int jfs_commit_thread_waking;
 
 /*
@@ -207,7 +207,7 @@ static lid_t txLockAlloc(void)
 	if ((++TxAnchor.tlocksInUse > TxLockHWM) && (jfs_tlocks_low == 0)) {
 		jfs_info("txLockAlloc tlocks low");
 		jfs_tlocks_low = 1;
-		wake_up(&jfs_sync_thread_wait);
+		wake_up_process(jfsSyncThread);
 	}
 
 	return lid;
@@ -2743,10 +2743,6 @@ int jfs_lazycommit(void *arg)
 	unsigned long flags;
 	struct jfs_sb_info *sbi;
 
-	daemonize("jfsCommit");
-
-	complete(&jfsIOwait);
-
 	do {
 		LAZY_LOCK(flags);
 		jfs_commit_thread_waking = 0;	/* OK to wake another thread */
@@ -2806,13 +2802,13 @@ int jfs_lazycommit(void *arg)
 			current->state = TASK_RUNNING;
 			remove_wait_queue(&jfs_commit_thread_wait, &wq);
 		}
-	} while (!jfs_stop_threads);
+	} while (!kthread_should_stop());
 
 	if (!list_empty(&TxAnchor.unlock_queue))
 		jfs_err("jfs_lazycommit being killed w/pending transactions!");
 	else
 		jfs_info("jfs_lazycommit being killed\n");
-	complete_and_exit(&jfsIOwait, 0);
+	return 0;
 }
 
 void txLazyUnlock(struct tblock * tblk)
@@ -2876,10 +2872,10 @@ restart:
 		 */
 		TXN_UNLOCK();
 		tid = txBegin(ip->i_sb, COMMIT_INODE | COMMIT_FORCE);
-		down(&jfs_ip->commit_sem);
+		mutex_lock(&jfs_ip->commit_mutex);
 		txCommit(tid, 1, &ip, 0);
 		txEnd(tid);
-		up(&jfs_ip->commit_sem);
+		mutex_unlock(&jfs_ip->commit_mutex);
 		/*
 		 * Just to be safe.  I don't know how
 		 * long we can run without blocking
@@ -2932,10 +2928,6 @@ int jfs_sync(void *arg)
 	int rc;
 	tid_t tid;
 
-	daemonize("jfsSync");
-
-	complete(&jfsIOwait);
-
 	do {
 		/*
 		 * write each inode on the anonymous inode list
@@ -2952,7 +2944,7 @@ int jfs_sync(void *arg)
 				 * Inode is being freed
 				 */
 				list_del_init(&jfs_ip->anon_inode_list);
-			} else if (! down_trylock(&jfs_ip->commit_sem)) {
+			} else if (! !mutex_trylock(&jfs_ip->commit_mutex)) {
 				/*
 				 * inode will be removed from anonymous list
 				 * when it is committed
@@ -2961,7 +2953,7 @@ int jfs_sync(void *arg)
 				tid = txBegin(ip->i_sb, COMMIT_INODE);
 				rc = txCommit(tid, 1, &ip, 0);
 				txEnd(tid);
-				up(&jfs_ip->commit_sem);
+				mutex_unlock(&jfs_ip->commit_mutex);
 
 				iput(ip);
 				/*
@@ -2971,7 +2963,7 @@ int jfs_sync(void *arg)
 				cond_resched();
 				TXN_LOCK();
 			} else {
-				/* We can't get the commit semaphore.  It may
+				/* We can't get the commit mutex.  It may
 				 * be held by a thread waiting for tlock's
 				 * so let's not block here.  Save it to
 				 * put back on the anon_list.
@@ -2996,19 +2988,15 @@ int jfs_sync(void *arg)
 			TXN_UNLOCK();
 			refrigerator();
 		} else {
-			DECLARE_WAITQUEUE(wq, current);
-
-			add_wait_queue(&jfs_sync_thread_wait, &wq);
 			set_current_state(TASK_INTERRUPTIBLE);
 			TXN_UNLOCK();
 			schedule();
 			current->state = TASK_RUNNING;
-			remove_wait_queue(&jfs_sync_thread_wait, &wq);
 		}
-	} while (!jfs_stop_threads);
+	} while (!kthread_should_stop());
 
 	jfs_info("jfs_sync being killed");
-	complete_and_exit(&jfsIOwait, 0);
+	return 0;
 }
 
 #if defined(CONFIG_PROC_FS) && defined(CONFIG_JFS_DEBUG)
diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c
index 4abbe8604302..09ea03f62277 100644
--- a/fs/jfs/namei.c
+++ b/fs/jfs/namei.c
@@ -104,8 +104,8 @@ static int jfs_create(struct inode *dip, struct dentry *dentry, int mode,
 
 	tid = txBegin(dip->i_sb, 0);
 
-	down(&JFS_IP(dip)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dip)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	rc = jfs_init_acl(tid, ip, dip);
 	if (rc)
@@ -165,8 +165,8 @@ static int jfs_create(struct inode *dip, struct dentry *dentry, int mode,
 
       out3:
 	txEnd(tid);
-	up(&JFS_IP(dip)->commit_sem);
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(dip)->commit_mutex);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
@@ -238,8 +238,8 @@ static int jfs_mkdir(struct inode *dip, struct dentry *dentry, int mode)
 
 	tid = txBegin(dip->i_sb, 0);
 
-	down(&JFS_IP(dip)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dip)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	rc = jfs_init_acl(tid, ip, dip);
 	if (rc)
@@ -300,8 +300,8 @@ static int jfs_mkdir(struct inode *dip, struct dentry *dentry, int mode)
 
       out3:
 	txEnd(tid);
-	up(&JFS_IP(dip)->commit_sem);
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(dip)->commit_mutex);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
@@ -365,8 +365,8 @@ static int jfs_rmdir(struct inode *dip, struct dentry *dentry)
 
 	tid = txBegin(dip->i_sb, 0);
 
-	down(&JFS_IP(dip)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dip)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	iplist[0] = dip;
 	iplist[1] = ip;
@@ -384,8 +384,8 @@ static int jfs_rmdir(struct inode *dip, struct dentry *dentry)
 		if (rc == -EIO)
 			txAbort(tid, 1);
 		txEnd(tid);
-		up(&JFS_IP(dip)->commit_sem);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(dip)->commit_mutex);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 
 		goto out2;
 	}
@@ -422,8 +422,8 @@ static int jfs_rmdir(struct inode *dip, struct dentry *dentry)
 
 	txEnd(tid);
 
-	up(&JFS_IP(dip)->commit_sem);
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(dip)->commit_mutex);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 
 	/*
 	 * Truncating the directory index table is not guaranteed.  It
@@ -488,8 +488,8 @@ static int jfs_unlink(struct inode *dip, struct dentry *dentry)
 
 	tid = txBegin(dip->i_sb, 0);
 
-	down(&JFS_IP(dip)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dip)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	iplist[0] = dip;
 	iplist[1] = ip;
@@ -503,8 +503,8 @@ static int jfs_unlink(struct inode *dip, struct dentry *dentry)
 		if (rc == -EIO)
 			txAbort(tid, 1);	/* Marks FS Dirty */
 		txEnd(tid);
-		up(&JFS_IP(dip)->commit_sem);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(dip)->commit_mutex);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 		IWRITE_UNLOCK(ip);
 		goto out1;
 	}
@@ -527,8 +527,8 @@ static int jfs_unlink(struct inode *dip, struct dentry *dentry)
 		if ((new_size = commitZeroLink(tid, ip)) < 0) {
 			txAbort(tid, 1);	/* Marks FS Dirty */
 			txEnd(tid);
-			up(&JFS_IP(dip)->commit_sem);
-			up(&JFS_IP(ip)->commit_sem);
+			mutex_unlock(&JFS_IP(dip)->commit_mutex);
+			mutex_unlock(&JFS_IP(ip)->commit_mutex);
 			IWRITE_UNLOCK(ip);
 			rc = new_size;
 			goto out1;
@@ -556,13 +556,13 @@ static int jfs_unlink(struct inode *dip, struct dentry *dentry)
 
 	txEnd(tid);
 
-	up(&JFS_IP(dip)->commit_sem);
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(dip)->commit_mutex);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 
 
 	while (new_size && (rc == 0)) {
 		tid = txBegin(dip->i_sb, 0);
-		down(&JFS_IP(ip)->commit_sem);
+		mutex_lock(&JFS_IP(ip)->commit_mutex);
 		new_size = xtTruncate_pmap(tid, ip, new_size);
 		if (new_size < 0) {
 			txAbort(tid, 1);	/* Marks FS Dirty */
@@ -570,7 +570,7 @@ static int jfs_unlink(struct inode *dip, struct dentry *dentry)
 		} else
 			rc = txCommit(tid, 2, &iplist[0], COMMIT_SYNC);
 		txEnd(tid);
-		up(&JFS_IP(ip)->commit_sem);
+		mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	}
 
 	if (ip->i_nlink == 0)
@@ -805,8 +805,8 @@ static int jfs_link(struct dentry *old_dentry,
 
 	tid = txBegin(ip->i_sb, 0);
 
-	down(&JFS_IP(dir)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dir)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	/*
 	 * scan parent directory for entry/freespace
@@ -847,8 +847,8 @@ static int jfs_link(struct dentry *old_dentry,
       out:
 	txEnd(tid);
 
-	up(&JFS_IP(dir)->commit_sem);
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(dir)->commit_mutex);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 
 	jfs_info("jfs_link: rc:%d", rc);
 	return rc;
@@ -916,8 +916,8 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry,
 
 	tid = txBegin(dip->i_sb, 0);
 
-	down(&JFS_IP(dip)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dip)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	rc = jfs_init_security(tid, ip, dip);
 	if (rc)
@@ -1037,8 +1037,8 @@ static int jfs_symlink(struct inode *dip, struct dentry *dentry,
 
       out3:
 	txEnd(tid);
-	up(&JFS_IP(dip)->commit_sem);
-	up(&JFS_IP(ip)->commit_sem);
+	mutex_unlock(&JFS_IP(dip)->commit_mutex);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
@@ -1141,13 +1141,13 @@ static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	 */
 	tid = txBegin(new_dir->i_sb, 0);
 
-	down(&JFS_IP(new_dir)->commit_sem);
-	down(&JFS_IP(old_ip)->commit_sem);
+	mutex_lock(&JFS_IP(new_dir)->commit_mutex);
+	mutex_lock(&JFS_IP(old_ip)->commit_mutex);
 	if (old_dir != new_dir)
-		down(&JFS_IP(old_dir)->commit_sem);
+		mutex_lock(&JFS_IP(old_dir)->commit_mutex);
 
 	if (new_ip) {
-		down(&JFS_IP(new_ip)->commit_sem);
+		mutex_lock(&JFS_IP(new_ip)->commit_mutex);
 		/*
 		 * Change existing directory entry to new inode number
 		 */
@@ -1160,10 +1160,10 @@ static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		if (S_ISDIR(new_ip->i_mode)) {
 			new_ip->i_nlink--;
 			if (new_ip->i_nlink) {
-				up(&JFS_IP(new_dir)->commit_sem);
-				up(&JFS_IP(old_ip)->commit_sem);
+				mutex_unlock(&JFS_IP(new_dir)->commit_mutex);
+				mutex_unlock(&JFS_IP(old_ip)->commit_mutex);
 				if (old_dir != new_dir)
-					up(&JFS_IP(old_dir)->commit_sem);
+					mutex_unlock(&JFS_IP(old_dir)->commit_mutex);
 				if (!S_ISDIR(old_ip->i_mode) && new_ip)
 					IWRITE_UNLOCK(new_ip);
 				jfs_error(new_ip->i_sb,
@@ -1282,16 +1282,16 @@ static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry,
       out4:
 	txEnd(tid);
 
-	up(&JFS_IP(new_dir)->commit_sem);
-	up(&JFS_IP(old_ip)->commit_sem);
+	mutex_unlock(&JFS_IP(new_dir)->commit_mutex);
+	mutex_unlock(&JFS_IP(old_ip)->commit_mutex);
 	if (old_dir != new_dir)
-		up(&JFS_IP(old_dir)->commit_sem);
+		mutex_unlock(&JFS_IP(old_dir)->commit_mutex);
 	if (new_ip)
-		up(&JFS_IP(new_ip)->commit_sem);
+		mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
 
 	while (new_size && (rc == 0)) {
 		tid = txBegin(new_ip->i_sb, 0);
-		down(&JFS_IP(new_ip)->commit_sem);
+		mutex_lock(&JFS_IP(new_ip)->commit_mutex);
 		new_size = xtTruncate_pmap(tid, new_ip, new_size);
 		if (new_size < 0) {
 			txAbort(tid, 1);
@@ -1299,7 +1299,7 @@ static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		} else
 			rc = txCommit(tid, 1, &new_ip, COMMIT_SYNC);
 		txEnd(tid);
-		up(&JFS_IP(new_ip)->commit_sem);
+		mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
 	}
 	if (new_ip && (new_ip->i_nlink == 0))
 		set_cflag(COMMIT_Nolink, new_ip);
@@ -1361,8 +1361,8 @@ static int jfs_mknod(struct inode *dir, struct dentry *dentry,
 
 	tid = txBegin(dir->i_sb, 0);
 
-	down(&JFS_IP(dir)->commit_sem);
-	down(&JFS_IP(ip)->commit_sem);
+	mutex_lock(&JFS_IP(dir)->commit_mutex);
+	mutex_lock(&JFS_IP(ip)->commit_mutex);
 
 	rc = jfs_init_acl(tid, ip, dir);
 	if (rc)
@@ -1407,8 +1407,8 @@ static int jfs_mknod(struct inode *dir, struct dentry *dentry,
 
       out3:
 	txEnd(tid);
-	up(&JFS_IP(ip)->commit_sem);
-	up(&JFS_IP(dir)->commit_sem);
+	mutex_unlock(&JFS_IP(ip)->commit_mutex);
+	mutex_unlock(&JFS_IP(dir)->commit_mutex);
 	if (rc) {
 		free_ea_wmap(ip);
 		ip->i_nlink = 0;
@@ -1519,10 +1519,11 @@ struct inode_operations jfs_dir_inode_operations = {
 #endif
 };
 
-struct file_operations jfs_dir_operations = {
+const struct file_operations jfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= jfs_readdir,
 	.fsync		= jfs_fsync,
+	.ioctl		= jfs_ioctl,
 };
 
 static int jfs_ci_hash(struct dentry *dir, struct qstr *this)
diff --git a/fs/jfs/super.c b/fs/jfs/super.c
index 8d31f1336431..db6f41d6dd60 100644
--- a/fs/jfs/super.c
+++ b/fs/jfs/super.c
@@ -25,6 +25,7 @@
 #include <linux/vfs.h>
 #include <linux/mount.h>
 #include <linux/moduleparam.h>
+#include <linux/kthread.h>
 #include <linux/posix_acl.h>
 #include <asm/uaccess.h>
 #include <linux/seq_file.h>
@@ -54,11 +55,9 @@ static int commit_threads = 0;
 module_param(commit_threads, int, 0);
 MODULE_PARM_DESC(commit_threads, "Number of commit threads");
 
-int jfs_stop_threads;
-static pid_t jfsIOthread;
-static pid_t jfsCommitThread[MAX_COMMIT_THREADS];
-static pid_t jfsSyncThread;
-DECLARE_COMPLETION(jfsIOwait);
+static struct task_struct *jfsCommitThread[MAX_COMMIT_THREADS];
+struct task_struct *jfsIOthread;
+struct task_struct *jfsSyncThread;
 
 #ifdef CONFIG_JFS_DEBUG
 int jfsloglevel = JFS_LOGLEVEL_WARN;
@@ -195,7 +194,7 @@ static void jfs_put_super(struct super_block *sb)
 enum {
 	Opt_integrity, Opt_nointegrity, Opt_iocharset, Opt_resize,
 	Opt_resize_nosize, Opt_errors, Opt_ignore, Opt_err, Opt_quota,
-	Opt_usrquota, Opt_grpquota
+	Opt_usrquota, Opt_grpquota, Opt_uid, Opt_gid, Opt_umask
 };
 
 static match_table_t tokens = {
@@ -209,6 +208,9 @@ static match_table_t tokens = {
 	{Opt_ignore, "quota"},
 	{Opt_usrquota, "usrquota"},
 	{Opt_grpquota, "grpquota"},
+	{Opt_uid, "uid=%u"},
+	{Opt_gid, "gid=%u"},
+	{Opt_umask, "umask=%u"},
 	{Opt_err, NULL}
 };
 
@@ -313,7 +315,29 @@ static int parse_options(char *options, struct super_block *sb, s64 *newLVSize,
 			       "JFS: quota operations not supported\n");
 			break;
 #endif
-
+		case Opt_uid:
+		{
+			char *uid = args[0].from;
+			sbi->uid = simple_strtoul(uid, &uid, 0);
+			break;
+		}
+		case Opt_gid:
+		{
+			char *gid = args[0].from;
+			sbi->gid = simple_strtoul(gid, &gid, 0);
+			break;
+		}
+		case Opt_umask:
+		{
+			char *umask = args[0].from;
+			sbi->umask = simple_strtoul(umask, &umask, 8);
+			if (sbi->umask & ~0777) {
+				printk(KERN_ERR
+				       "JFS: Invalid value of umask\n");
+				goto cleanup;
+			}
+			break;
+		}
 		default:
 			printk("jfs: Unrecognized mount option \"%s\" "
 					" or missing value\n", p);
@@ -396,12 +420,12 @@ static int jfs_fill_super(struct super_block *sb, void *data, int silent)
 	if (!new_valid_dev(sb->s_bdev->bd_dev))
 		return -EOVERFLOW;
 
-	sbi = kmalloc(sizeof (struct jfs_sb_info), GFP_KERNEL);
+	sbi = kzalloc(sizeof (struct jfs_sb_info), GFP_KERNEL);
 	if (!sbi)
 		return -ENOSPC;
-	memset(sbi, 0, sizeof (struct jfs_sb_info));
 	sb->s_fs_info = sbi;
 	sbi->sb = sb;
+	sbi->uid = sbi->gid = sbi->umask = -1;
 
 	/* initialize the mount flag and determine the default error handler */
 	flag = JFS_ERR_REMOUNT_RO;
@@ -564,10 +588,14 @@ static int jfs_show_options(struct seq_file *seq, struct vfsmount *vfs)
 {
 	struct jfs_sb_info *sbi = JFS_SBI(vfs->mnt_sb);
 
+	if (sbi->uid != -1)
+		seq_printf(seq, ",uid=%d", sbi->uid);
+	if (sbi->gid != -1)
+		seq_printf(seq, ",gid=%d", sbi->gid);
+	if (sbi->umask != -1)
+		seq_printf(seq, ",umask=%03o", sbi->umask);
 	if (sbi->flag & JFS_NOINTEGRITY)
 		seq_puts(seq, ",nointegrity");
-	else
-		seq_puts(seq, ",integrity");
 
 #if defined(CONFIG_QUOTA)
 	if (sbi->flag & JFS_USRQUOTA)
@@ -617,7 +645,7 @@ static void init_once(void *foo, kmem_cache_t * cachep, unsigned long flags)
 		memset(jfs_ip, 0, sizeof(struct jfs_inode_info));
 		INIT_LIST_HEAD(&jfs_ip->anon_inode_list);
 		init_rwsem(&jfs_ip->rdwrlock);
-		init_MUTEX(&jfs_ip->commit_sem);
+		mutex_init(&jfs_ip->commit_mutex);
 		init_rwsem(&jfs_ip->xattr_sem);
 		spin_lock_init(&jfs_ip->ag_lock);
 		jfs_ip->active_ag = -1;
@@ -636,7 +664,8 @@ static int __init init_jfs_fs(void)
 
 	jfs_inode_cachep =
 	    kmem_cache_create("jfs_ip", sizeof(struct jfs_inode_info), 0, 
-			    SLAB_RECLAIM_ACCOUNT, init_once, NULL);
+			    SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
+			    init_once, NULL);
 	if (jfs_inode_cachep == NULL)
 		return -ENOMEM;
 
@@ -661,12 +690,12 @@ static int __init init_jfs_fs(void)
 	/*
 	 * I/O completion thread (endio)
 	 */
-	jfsIOthread = kernel_thread(jfsIOWait, NULL, CLONE_KERNEL);
-	if (jfsIOthread < 0) {
-		jfs_err("init_jfs_fs: fork failed w/rc = %d", jfsIOthread);
+	jfsIOthread = kthread_run(jfsIOWait, NULL, "jfsIO");
+	if (IS_ERR(jfsIOthread)) {
+		rc = PTR_ERR(jfsIOthread);
+		jfs_err("init_jfs_fs: fork failed w/rc = %d", rc);
 		goto end_txmngr;
 	}
-	wait_for_completion(&jfsIOwait);	/* Wait until thread starts */
 
 	if (commit_threads < 1)
 		commit_threads = num_online_cpus();
@@ -674,24 +703,21 @@ static int __init init_jfs_fs(void)
 		commit_threads = MAX_COMMIT_THREADS;
 
 	for (i = 0; i < commit_threads; i++) {
-		jfsCommitThread[i] = kernel_thread(jfs_lazycommit, NULL,
-						   CLONE_KERNEL);
-		if (jfsCommitThread[i] < 0) {
-			jfs_err("init_jfs_fs: fork failed w/rc = %d",
-				jfsCommitThread[i]);
+		jfsCommitThread[i] = kthread_run(jfs_lazycommit, NULL, "jfsCommit");
+		if (IS_ERR(jfsCommitThread[i])) {
+			rc = PTR_ERR(jfsCommitThread[i]);
+			jfs_err("init_jfs_fs: fork failed w/rc = %d", rc);
 			commit_threads = i;
 			goto kill_committask;
 		}
-		/* Wait until thread starts */
-		wait_for_completion(&jfsIOwait);
 	}
 
-	jfsSyncThread = kernel_thread(jfs_sync, NULL, CLONE_KERNEL);
-	if (jfsSyncThread < 0) {
-		jfs_err("init_jfs_fs: fork failed w/rc = %d", jfsSyncThread);
+	jfsSyncThread = kthread_run(jfs_sync, NULL, "jfsSync");
+	if (IS_ERR(jfsSyncThread)) {
+		rc = PTR_ERR(jfsSyncThread);
+		jfs_err("init_jfs_fs: fork failed w/rc = %d", rc);
 		goto kill_committask;
 	}
-	wait_for_completion(&jfsIOwait);	/* Wait until thread starts */
 
 #ifdef PROC_FS_JFS
 	jfs_proc_init();
@@ -700,13 +726,9 @@ static int __init init_jfs_fs(void)
 	return register_filesystem(&jfs_fs_type);
 
 kill_committask:
-	jfs_stop_threads = 1;
-	wake_up_all(&jfs_commit_thread_wait);
 	for (i = 0; i < commit_threads; i++)
-		wait_for_completion(&jfsIOwait);
-
-	wake_up(&jfs_IO_thread_wait);
-	wait_for_completion(&jfsIOwait);	/* Wait for thread exit */
+		kthread_stop(jfsCommitThread[i]);
+	kthread_stop(jfsIOthread);
 end_txmngr:
 	txExit();
 free_metapage:
@@ -722,16 +744,13 @@ static void __exit exit_jfs_fs(void)
 
 	jfs_info("exit_jfs_fs called");
 
-	jfs_stop_threads = 1;
 	txExit();
 	metapage_exit();
-	wake_up(&jfs_IO_thread_wait);
-	wait_for_completion(&jfsIOwait);	/* Wait until IO thread exits */
-	wake_up_all(&jfs_commit_thread_wait);
+
+	kthread_stop(jfsIOthread);
 	for (i = 0; i < commit_threads; i++)
-		wait_for_completion(&jfsIOwait);
-	wake_up(&jfs_sync_thread_wait);
-	wait_for_completion(&jfsIOwait);	/* Wait until Sync thread exits */
+		kthread_stop(jfsCommitThread[i]);
+	kthread_stop(jfsSyncThread);
 #ifdef PROC_FS_JFS
 	jfs_proc_clean();
 #endif
diff --git a/fs/jfs/xattr.c b/fs/jfs/xattr.c
index f23048f9471f..9bc5b7c055ce 100644
--- a/fs/jfs/xattr.c
+++ b/fs/jfs/xattr.c
@@ -934,13 +934,13 @@ int jfs_setxattr(struct dentry *dentry, const char *name, const void *value,
 	}
 
 	tid = txBegin(inode->i_sb, 0);
-	down(&ji->commit_sem);
+	mutex_lock(&ji->commit_mutex);
 	rc = __jfs_setxattr(tid, dentry->d_inode, name, value, value_len,
 			    flags);
 	if (!rc)
 		rc = txCommit(tid, 1, &inode, 0);
 	txEnd(tid);
-	up(&ji->commit_sem);
+	mutex_unlock(&ji->commit_mutex);
 
 	return rc;
 }
@@ -1093,12 +1093,12 @@ int jfs_removexattr(struct dentry *dentry, const char *name)
 		return rc;
 
 	tid = txBegin(inode->i_sb, 0);
-	down(&ji->commit_sem);
+	mutex_lock(&ji->commit_mutex);
 	rc = __jfs_setxattr(tid, dentry->d_inode, name, NULL, 0, XATTR_REPLACE);
 	if (!rc)
 		rc = txCommit(tid, 1, &inode, 0);
 	txEnd(tid);
-	up(&ji->commit_sem);
+	mutex_unlock(&ji->commit_mutex);
 
 	return rc;
 }
diff --git a/fs/libfs.c b/fs/libfs.c
index 71fd08fa4103..7145ba7a48d0 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -7,6 +7,8 @@
 #include <linux/pagemap.h>
 #include <linux/mount.h>
 #include <linux/vfs.h>
+#include <linux/mutex.h>
+
 #include <asm/uaccess.h>
 
 int simple_getattr(struct vfsmount *mnt, struct dentry *dentry,
@@ -177,7 +179,7 @@ ssize_t generic_read_dir(struct file *filp, char __user *buf, size_t siz, loff_t
 	return -EISDIR;
 }
 
-struct file_operations simple_dir_operations = {
+const struct file_operations simple_dir_operations = {
 	.open		= dcache_dir_open,
 	.release	= dcache_dir_close,
 	.llseek		= dcache_dir_lseek,
@@ -530,7 +532,7 @@ struct simple_attr {
 	char set_buf[24];
 	void *data;
 	const char *fmt;	/* format for read operation */
-	struct semaphore sem;	/* protects access to these buffers */
+	struct mutex mutex;	/* protects access to these buffers */
 };
 
 /* simple_attr_open is called by an actual attribute open file operation
@@ -549,7 +551,7 @@ int simple_attr_open(struct inode *inode, struct file *file,
 	attr->set = set;
 	attr->data = inode->u.generic_ip;
 	attr->fmt = fmt;
-	init_MUTEX(&attr->sem);
+	mutex_init(&attr->mutex);
 
 	file->private_data = attr;
 
@@ -575,7 +577,7 @@ ssize_t simple_attr_read(struct file *file, char __user *buf,
 	if (!attr->get)
 		return -EACCES;
 
-	down(&attr->sem);
+	mutex_lock(&attr->mutex);
 	if (*ppos) /* continued read */
 		size = strlen(attr->get_buf);
 	else	  /* first read */
@@ -584,7 +586,7 @@ ssize_t simple_attr_read(struct file *file, char __user *buf,
 				 (unsigned long long)attr->get(attr->data));
 
 	ret = simple_read_from_buffer(buf, len, ppos, attr->get_buf, size);
-	up(&attr->sem);
+	mutex_unlock(&attr->mutex);
 	return ret;
 }
 
@@ -602,7 +604,7 @@ ssize_t simple_attr_write(struct file *file, const char __user *buf,
 	if (!attr->set)
 		return -EACCES;
 
-	down(&attr->sem);
+	mutex_lock(&attr->mutex);
 	ret = -EFAULT;
 	size = min(sizeof(attr->set_buf) - 1, len);
 	if (copy_from_user(attr->set_buf, buf, size))
@@ -613,7 +615,7 @@ ssize_t simple_attr_write(struct file *file, const char __user *buf,
 	val = simple_strtol(attr->set_buf, NULL, 0);
 	attr->set(attr->data, val);
 out:
-	up(&attr->sem);
+	mutex_unlock(&attr->mutex);
 	return ret;
 }
 
diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c
index da6354baa0b8..bce744468708 100644
--- a/fs/lockd/clntlock.c
+++ b/fs/lockd/clntlock.c
@@ -44,32 +44,25 @@ static LIST_HEAD(nlm_blocked);
 /*
  * Queue up a lock for blocking so that the GRANTED request can see it
  */
-int nlmclnt_prepare_block(struct nlm_rqst *req, struct nlm_host *host, struct file_lock *fl)
+struct nlm_wait *nlmclnt_prepare_block(struct nlm_host *host, struct file_lock *fl)
 {
 	struct nlm_wait *block;
 
-	BUG_ON(req->a_block != NULL);
 	block = kmalloc(sizeof(*block), GFP_KERNEL);
-	if (block == NULL)
-		return -ENOMEM;
-	block->b_host = host;
-	block->b_lock = fl;
-	init_waitqueue_head(&block->b_wait);
-	block->b_status = NLM_LCK_BLOCKED;
-
-	list_add(&block->b_list, &nlm_blocked);
-	req->a_block = block;
-
-	return 0;
+	if (block != NULL) {
+		block->b_host = host;
+		block->b_lock = fl;
+		init_waitqueue_head(&block->b_wait);
+		block->b_status = NLM_LCK_BLOCKED;
+		list_add(&block->b_list, &nlm_blocked);
+	}
+	return block;
 }
 
-void nlmclnt_finish_block(struct nlm_rqst *req)
+void nlmclnt_finish_block(struct nlm_wait *block)
 {
-	struct nlm_wait *block = req->a_block;
-
 	if (block == NULL)
 		return;
-	req->a_block = NULL;
 	list_del(&block->b_list);
 	kfree(block);
 }
@@ -77,15 +70,14 @@ void nlmclnt_finish_block(struct nlm_rqst *req)
 /*
  * Block on a lock
  */
-long nlmclnt_block(struct nlm_rqst *req, long timeout)
+int nlmclnt_block(struct nlm_wait *block, struct nlm_rqst *req, long timeout)
 {
-	struct nlm_wait	*block = req->a_block;
 	long ret;
 
 	/* A borken server might ask us to block even if we didn't
 	 * request it. Just say no!
 	 */
-	if (!req->a_args.block)
+	if (block == NULL)
 		return -EAGAIN;
 
 	/* Go to sleep waiting for GRANT callback. Some servers seem
@@ -99,13 +91,10 @@ long nlmclnt_block(struct nlm_rqst *req, long timeout)
 	ret = wait_event_interruptible_timeout(block->b_wait,
 			block->b_status != NLM_LCK_BLOCKED,
 			timeout);
-
-	if (block->b_status != NLM_LCK_BLOCKED) {
-		req->a_res.status = block->b_status;
-		block->b_status = NLM_LCK_BLOCKED;
-	}
-
-	return ret;
+	if (ret < 0)
+		return -ERESTARTSYS;
+	req->a_res.status = block->b_status;
+	return 0;
 }
 
 /*
@@ -125,7 +114,15 @@ u32 nlmclnt_grant(const struct sockaddr_in *addr, const struct nlm_lock *lock)
 	list_for_each_entry(block, &nlm_blocked, b_list) {
 		struct file_lock *fl_blocked = block->b_lock;
 
-		if (!nlm_compare_locks(fl_blocked, fl))
+		if (fl_blocked->fl_start != fl->fl_start)
+			continue;
+		if (fl_blocked->fl_end != fl->fl_end)
+			continue;
+		/*
+		 * Careful! The NLM server will return the 32-bit "pid" that
+		 * we put on the wire: in this case the lockowner "pid".
+		 */
+		if (fl_blocked->fl_u.nfs_fl.owner->pid != lock->svid)
 			continue;
 		if (!nlm_cmp_addr(&block->b_host->h_addr, addr))
 			continue;
@@ -147,34 +144,6 @@ u32 nlmclnt_grant(const struct sockaddr_in *addr, const struct nlm_lock *lock)
  */
 
 /*
- * Mark the locks for reclaiming.
- * FIXME: In 2.5 we don't want to iterate through any global file_lock_list.
- *        Maintain NLM lock reclaiming lists in the nlm_host instead.
- */
-static
-void nlmclnt_mark_reclaim(struct nlm_host *host)
-{
-	struct file_lock *fl;
-	struct inode *inode;
-	struct list_head *tmp;
-
-	list_for_each(tmp, &file_lock_list) {
-		fl = list_entry(tmp, struct file_lock, fl_link);
-
-		inode = fl->fl_file->f_dentry->d_inode;
-		if (inode->i_sb->s_magic != NFS_SUPER_MAGIC)
-			continue;
-		if (fl->fl_u.nfs_fl.owner == NULL)
-			continue;
-		if (fl->fl_u.nfs_fl.owner->host != host)
-			continue;
-		if (!(fl->fl_u.nfs_fl.flags & NFS_LCK_GRANTED))
-			continue;
-		fl->fl_u.nfs_fl.flags |= NFS_LCK_RECLAIM;
-	}
-}
-
-/*
  * Someone has sent us an SM_NOTIFY. Ensure we bind to the new port number,
  * that we mark locks for reclaiming, and that we bump the pseudo NSM state.
  */
@@ -186,7 +155,12 @@ void nlmclnt_prepare_reclaim(struct nlm_host *host, u32 newstate)
 	host->h_state++;
 	host->h_nextrebind = 0;
 	nlm_rebind_host(host);
-	nlmclnt_mark_reclaim(host);
+
+	/*
+	 * Mark the locks for reclaiming.
+	 */
+	list_splice_init(&host->h_granted, &host->h_reclaim);
+
 	dprintk("NLM: reclaiming locks for host %s", host->h_name);
 }
 
@@ -215,9 +189,7 @@ reclaimer(void *ptr)
 {
 	struct nlm_host	  *host = (struct nlm_host *) ptr;
 	struct nlm_wait	  *block;
-	struct list_head *tmp;
-	struct file_lock *fl;
-	struct inode *inode;
+	struct file_lock *fl, *next;
 
 	daemonize("%s-reclaim", host->h_name);
 	allow_signal(SIGKILL);
@@ -229,23 +201,13 @@ reclaimer(void *ptr)
 
 	/* First, reclaim all locks that have been marked. */
 restart:
-	list_for_each(tmp, &file_lock_list) {
-		fl = list_entry(tmp, struct file_lock, fl_link);
+	list_for_each_entry_safe(fl, next, &host->h_reclaim, fl_u.nfs_fl.list) {
+		list_del_init(&fl->fl_u.nfs_fl.list);
 
-		inode = fl->fl_file->f_dentry->d_inode;
-		if (inode->i_sb->s_magic != NFS_SUPER_MAGIC)
-			continue;
-		if (fl->fl_u.nfs_fl.owner == NULL)
-			continue;
-		if (fl->fl_u.nfs_fl.owner->host != host)
-			continue;
-		if (!(fl->fl_u.nfs_fl.flags & NFS_LCK_RECLAIM))
-			continue;
-
-		fl->fl_u.nfs_fl.flags &= ~NFS_LCK_RECLAIM;
-		nlmclnt_reclaim(host, fl);
 		if (signalled())
-			break;
+			continue;
+		if (nlmclnt_reclaim(host, fl) == 0)
+			list_add_tail(&fl->fl_u.nfs_fl.list, &host->h_granted);
 		goto restart;
 	}
 
diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
index 970b6a6aa337..f96e38155b5c 100644
--- a/fs/lockd/clntproc.c
+++ b/fs/lockd/clntproc.c
@@ -132,59 +132,18 @@ static void nlmclnt_setlockargs(struct nlm_rqst *req, struct file_lock *fl)
 	memcpy(&lock->fh, NFS_FH(fl->fl_file->f_dentry->d_inode), sizeof(struct nfs_fh));
 	lock->caller  = system_utsname.nodename;
 	lock->oh.data = req->a_owner;
-	lock->oh.len  = sprintf(req->a_owner, "%d@%s",
-				current->pid, system_utsname.nodename);
-	locks_copy_lock(&lock->fl, fl);
+	lock->oh.len  = snprintf(req->a_owner, sizeof(req->a_owner), "%u@%s",
+				(unsigned int)fl->fl_u.nfs_fl.owner->pid,
+				system_utsname.nodename);
+	lock->svid = fl->fl_u.nfs_fl.owner->pid;
+	lock->fl.fl_start = fl->fl_start;
+	lock->fl.fl_end = fl->fl_end;
+	lock->fl.fl_type = fl->fl_type;
 }
 
 static void nlmclnt_release_lockargs(struct nlm_rqst *req)
 {
-	struct file_lock *fl = &req->a_args.lock.fl;
-
-	if (fl->fl_ops && fl->fl_ops->fl_release_private)
-		fl->fl_ops->fl_release_private(fl);
-}
-
-/*
- * Initialize arguments for GRANTED call. The nlm_rqst structure
- * has been cleared already.
- */
-int
-nlmclnt_setgrantargs(struct nlm_rqst *call, struct nlm_lock *lock)
-{
-	locks_copy_lock(&call->a_args.lock.fl, &lock->fl);
-	memcpy(&call->a_args.lock.fh, &lock->fh, sizeof(call->a_args.lock.fh));
-	call->a_args.lock.caller = system_utsname.nodename;
-	call->a_args.lock.oh.len = lock->oh.len;
-
-	/* set default data area */
-	call->a_args.lock.oh.data = call->a_owner;
-
-	if (lock->oh.len > NLMCLNT_OHSIZE) {
-		void *data = kmalloc(lock->oh.len, GFP_KERNEL);
-		if (!data) {
-			nlmclnt_freegrantargs(call);
-			return 0;
-		}
-		call->a_args.lock.oh.data = (u8 *) data;
-	}
-
-	memcpy(call->a_args.lock.oh.data, lock->oh.data, lock->oh.len);
-	return 1;
-}
-
-void
-nlmclnt_freegrantargs(struct nlm_rqst *call)
-{
-	struct file_lock *fl = &call->a_args.lock.fl;
-	/*
-	 * Check whether we allocated memory for the owner.
-	 */
-	if (call->a_args.lock.oh.data != (u8 *) call->a_owner) {
-		kfree(call->a_args.lock.oh.data);
-	}
-	if (fl->fl_ops && fl->fl_ops->fl_release_private)
-		fl->fl_ops->fl_release_private(fl);
+	BUG_ON(req->a_args.lock.fl.fl_ops != NULL);
 }
 
 /*
@@ -193,9 +152,8 @@ nlmclnt_freegrantargs(struct nlm_rqst *call)
 int
 nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl)
 {
-	struct nfs_server	*nfssrv = NFS_SERVER(inode);
 	struct nlm_host		*host;
-	struct nlm_rqst		reqst, *call = &reqst;
+	struct nlm_rqst		*call;
 	sigset_t		oldset;
 	unsigned long		flags;
 	int			status, proto, vers;
@@ -209,23 +167,17 @@ nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl)
 	/* Retrieve transport protocol from NFS client */
 	proto = NFS_CLIENT(inode)->cl_xprt->prot;
 
-	if (!(host = nlmclnt_lookup_host(NFS_ADDR(inode), proto, vers)))
+	host = nlmclnt_lookup_host(NFS_ADDR(inode), proto, vers);
+	if (host == NULL)
 		return -ENOLCK;
 
-	/* Create RPC client handle if not there, and copy soft
-	 * and intr flags from NFS client. */
-	if (host->h_rpcclnt == NULL) {
-		struct rpc_clnt	*clnt;
+	call = nlm_alloc_call(host);
+	if (call == NULL)
+		return -ENOMEM;
 
-		/* Bind an rpc client to this host handle (does not
-		 * perform a portmapper lookup) */
-		if (!(clnt = nlm_bind_host(host))) {
-			status = -ENOLCK;
-			goto done;
-		}
-		clnt->cl_softrtry = nfssrv->client->cl_softrtry;
-		clnt->cl_intr = nfssrv->client->cl_intr;
-	}
+	nlmclnt_locks_init_private(fl, host);
+	/* Set up the argument struct */
+	nlmclnt_setlockargs(call, fl);
 
 	/* Keep the old signal mask */
 	spin_lock_irqsave(&current->sighand->siglock, flags);
@@ -238,26 +190,10 @@ nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl)
 	    && (current->flags & PF_EXITING)) {
 		sigfillset(&current->blocked);	/* Mask all signals */
 		recalc_sigpending();
-		spin_unlock_irqrestore(&current->sighand->siglock, flags);
 
-		call = nlmclnt_alloc_call();
-		if (!call) {
-			status = -ENOMEM;
-			goto out_restore;
-		}
 		call->a_flags = RPC_TASK_ASYNC;
-	} else {
-		spin_unlock_irqrestore(&current->sighand->siglock, flags);
-		memset(call, 0, sizeof(*call));
-		locks_init_lock(&call->a_args.lock.fl);
-		locks_init_lock(&call->a_res.lock.fl);
 	}
-	call->a_host = host;
-
-	nlmclnt_locks_init_private(fl, host);
-
-	/* Set up the argument struct */
-	nlmclnt_setlockargs(call, fl);
+	spin_unlock_irqrestore(&current->sighand->siglock, flags);
 
 	if (IS_SETLK(cmd) || IS_SETLKW(cmd)) {
 		if (fl->fl_type != F_UNLCK) {
@@ -270,41 +206,58 @@ nlmclnt_proc(struct inode *inode, int cmd, struct file_lock *fl)
 	else
 		status = -EINVAL;
 
- out_restore:
+	fl->fl_ops->fl_release_private(fl);
+	fl->fl_ops = NULL;
+
 	spin_lock_irqsave(&current->sighand->siglock, flags);
 	current->blocked = oldset;
 	recalc_sigpending();
 	spin_unlock_irqrestore(&current->sighand->siglock, flags);
 
-done:
 	dprintk("lockd: clnt proc returns %d\n", status);
-	nlm_release_host(host);
 	return status;
 }
 EXPORT_SYMBOL(nlmclnt_proc);
 
 /*
  * Allocate an NLM RPC call struct
+ *
+ * Note: the caller must hold a reference to host. In case of failure,
+ * this reference will be released.
  */
-struct nlm_rqst *
-nlmclnt_alloc_call(void)
+struct nlm_rqst *nlm_alloc_call(struct nlm_host *host)
 {
 	struct nlm_rqst	*call;
 
-	while (!signalled()) {
-		call = (struct nlm_rqst *) kmalloc(sizeof(struct nlm_rqst), GFP_KERNEL);
-		if (call) {
-			memset(call, 0, sizeof(*call));
+	for(;;) {
+		call = kzalloc(sizeof(*call), GFP_KERNEL);
+		if (call != NULL) {
 			locks_init_lock(&call->a_args.lock.fl);
 			locks_init_lock(&call->a_res.lock.fl);
+			call->a_host = host;
 			return call;
 		}
-		printk("nlmclnt_alloc_call: failed, waiting for memory\n");
+		if (signalled())
+			break;
+		printk("nlm_alloc_call: failed, waiting for memory\n");
 		schedule_timeout_interruptible(5*HZ);
 	}
+	nlm_release_host(host);
 	return NULL;
 }
 
+void nlm_release_call(struct nlm_rqst *call)
+{
+	nlm_release_host(call->a_host);
+	nlmclnt_release_lockargs(call);
+	kfree(call);
+}
+
+static void nlmclnt_rpc_release(void *data)
+{
+	return nlm_release_call(data);
+}
+
 static int nlm_wait_on_grace(wait_queue_head_t *queue)
 {
 	DEFINE_WAIT(wait);
@@ -401,57 +354,45 @@ in_grace_period:
 /*
  * Generic NLM call, async version.
  */
-int nlmsvc_async_call(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops)
+static int __nlm_async_call(struct nlm_rqst *req, u32 proc, struct rpc_message *msg, const struct rpc_call_ops *tk_ops)
 {
 	struct nlm_host	*host = req->a_host;
 	struct rpc_clnt	*clnt;
-	struct rpc_message msg = {
-		.rpc_argp	= &req->a_args,
-		.rpc_resp	= &req->a_res,
-	};
-	int		status;
+	int status = -ENOLCK;
 
 	dprintk("lockd: call procedure %d on %s (async)\n",
 			(int)proc, host->h_name);
 
 	/* If we have no RPC client yet, create one. */
-	if ((clnt = nlm_bind_host(host)) == NULL)
-		return -ENOLCK;
-	msg.rpc_proc = &clnt->cl_procinfo[proc];
+	clnt = nlm_bind_host(host);
+	if (clnt == NULL)
+		goto out_err;
+	msg->rpc_proc = &clnt->cl_procinfo[proc];
 
         /* bootstrap and kick off the async RPC call */
-        status = rpc_call_async(clnt, &msg, RPC_TASK_ASYNC, tk_ops, req);
-
+        status = rpc_call_async(clnt, msg, RPC_TASK_ASYNC, tk_ops, req);
+	if (status == 0)
+		return 0;
+out_err:
+	nlm_release_call(req);
 	return status;
 }
 
-static int nlmclnt_async_call(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops)
+int nlm_async_call(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops)
 {
-	struct nlm_host	*host = req->a_host;
-	struct rpc_clnt	*clnt;
-	struct nlm_args	*argp = &req->a_args;
-	struct nlm_res	*resp = &req->a_res;
 	struct rpc_message msg = {
-		.rpc_argp	= argp,
-		.rpc_resp	= resp,
+		.rpc_argp	= &req->a_args,
+		.rpc_resp	= &req->a_res,
 	};
-	int		status;
-
-	dprintk("lockd: call procedure %d on %s (async)\n",
-			(int)proc, host->h_name);
-
-	/* If we have no RPC client yet, create one. */
-	if ((clnt = nlm_bind_host(host)) == NULL)
-		return -ENOLCK;
-	msg.rpc_proc = &clnt->cl_procinfo[proc];
+	return __nlm_async_call(req, proc, &msg, tk_ops);
+}
 
-	/* Increment host refcount */
-	nlm_get_host(host);
-        /* bootstrap and kick off the async RPC call */
-        status = rpc_call_async(clnt, &msg, RPC_TASK_ASYNC, tk_ops, req);
-	if (status < 0)
-		nlm_release_host(host);
-	return status;
+int nlm_async_reply(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops)
+{
+	struct rpc_message msg = {
+		.rpc_argp	= &req->a_res,
+	};
+	return __nlm_async_call(req, proc, &msg, tk_ops);
 }
 
 /*
@@ -463,36 +404,41 @@ nlmclnt_test(struct nlm_rqst *req, struct file_lock *fl)
 	int	status;
 
 	status = nlmclnt_call(req, NLMPROC_TEST);
-	nlmclnt_release_lockargs(req);
 	if (status < 0)
-		return status;
+		goto out;
 
-	status = req->a_res.status;
-	if (status == NLM_LCK_GRANTED) {
-		fl->fl_type = F_UNLCK;
-	} if (status == NLM_LCK_DENIED) {
-		/*
-		 * Report the conflicting lock back to the application.
-		 */
-		locks_copy_lock(fl, &req->a_res.lock.fl);
-		fl->fl_pid = 0;
-	} else {
-		return nlm_stat_to_errno(req->a_res.status);
+	switch (req->a_res.status) {
+		case NLM_LCK_GRANTED:
+			fl->fl_type = F_UNLCK;
+			break;
+		case NLM_LCK_DENIED:
+			/*
+			 * Report the conflicting lock back to the application.
+			 */
+			fl->fl_start = req->a_res.lock.fl.fl_start;
+			fl->fl_end = req->a_res.lock.fl.fl_start;
+			fl->fl_type = req->a_res.lock.fl.fl_type;
+			fl->fl_pid = 0;
+			break;
+		default:
+			status = nlm_stat_to_errno(req->a_res.status);
 	}
-
-	return 0;
+out:
+	nlm_release_call(req);
+	return status;
 }
 
 static void nlmclnt_locks_copy_lock(struct file_lock *new, struct file_lock *fl)
 {
-	memcpy(&new->fl_u.nfs_fl, &fl->fl_u.nfs_fl, sizeof(new->fl_u.nfs_fl));
-	nlm_get_lockowner(new->fl_u.nfs_fl.owner);
+	new->fl_u.nfs_fl.state = fl->fl_u.nfs_fl.state;
+	new->fl_u.nfs_fl.owner = nlm_get_lockowner(fl->fl_u.nfs_fl.owner);
+	list_add_tail(&new->fl_u.nfs_fl.list, &fl->fl_u.nfs_fl.owner->host->h_granted);
 }
 
 static void nlmclnt_locks_release_private(struct file_lock *fl)
 {
+	list_del(&fl->fl_u.nfs_fl.list);
 	nlm_put_lockowner(fl->fl_u.nfs_fl.owner);
-	fl->fl_ops = NULL;
 }
 
 static struct file_lock_operations nlmclnt_lock_ops = {
@@ -504,8 +450,8 @@ static void nlmclnt_locks_init_private(struct file_lock *fl, struct nlm_host *ho
 {
 	BUG_ON(fl->fl_ops != NULL);
 	fl->fl_u.nfs_fl.state = 0;
-	fl->fl_u.nfs_fl.flags = 0;
 	fl->fl_u.nfs_fl.owner = nlm_find_lockowner(host, fl->fl_owner);
+	INIT_LIST_HEAD(&fl->fl_u.nfs_fl.list);
 	fl->fl_ops = &nlmclnt_lock_ops;
 }
 
@@ -552,57 +498,52 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl)
 {
 	struct nlm_host	*host = req->a_host;
 	struct nlm_res	*resp = &req->a_res;
-	long timeout;
-	int status;
+	struct nlm_wait *block = NULL;
+	int status = -ENOLCK;
 
 	if (!host->h_monitored && nsm_monitor(host) < 0) {
 		printk(KERN_NOTICE "lockd: failed to monitor %s\n",
 					host->h_name);
-		status = -ENOLCK;
 		goto out;
 	}
 
-	if (req->a_args.block) {
-		status = nlmclnt_prepare_block(req, host, fl);
-		if (status < 0)
-			goto out;
-	}
+	block = nlmclnt_prepare_block(host, fl);
 	for(;;) {
 		status = nlmclnt_call(req, NLMPROC_LOCK);
 		if (status < 0)
 			goto out_unblock;
-		if (resp->status != NLM_LCK_BLOCKED)
+		if (!req->a_args.block)
 			break;
-		/* Wait on an NLM blocking lock */
-		timeout = nlmclnt_block(req, NLMCLNT_POLL_TIMEOUT);
 		/* Did a reclaimer thread notify us of a server reboot? */
 		if (resp->status ==  NLM_LCK_DENIED_GRACE_PERIOD)
 			continue;
 		if (resp->status != NLM_LCK_BLOCKED)
 			break;
-		if (timeout >= 0)
-			continue;
-		/* We were interrupted. Send a CANCEL request to the server
+		/* Wait on an NLM blocking lock */
+		status = nlmclnt_block(block, req, NLMCLNT_POLL_TIMEOUT);
+		/* if we were interrupted. Send a CANCEL request to the server
 		 * and exit
 		 */
-		status = (int)timeout;
-		goto out_unblock;
+		if (status < 0)
+			goto out_unblock;
+		if (resp->status != NLM_LCK_BLOCKED)
+			break;
 	}
 
 	if (resp->status == NLM_LCK_GRANTED) {
 		fl->fl_u.nfs_fl.state = host->h_state;
-		fl->fl_u.nfs_fl.flags |= NFS_LCK_GRANTED;
 		fl->fl_flags |= FL_SLEEP;
+		/* Ensure the resulting lock will get added to granted list */
 		do_vfs_lock(fl);
 	}
 	status = nlm_stat_to_errno(resp->status);
 out_unblock:
-	nlmclnt_finish_block(req);
+	nlmclnt_finish_block(block);
 	/* Cancel the blocked request if it is still pending */
 	if (resp->status == NLM_LCK_BLOCKED)
 		nlmclnt_cancel(host, req->a_args.block, fl);
 out:
-	nlmclnt_release_lockargs(req);
+	nlm_release_call(req);
 	return status;
 }
 
@@ -658,10 +599,6 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl)
 	struct nlm_res	*resp = &req->a_res;
 	int		status;
 
-	/* Clean the GRANTED flag now so the lock doesn't get
-	 * reclaimed while we're stuck in the unlock call. */
-	fl->fl_u.nfs_fl.flags &= ~NFS_LCK_GRANTED;
-
 	/*
 	 * Note: the server is supposed to either grant us the unlock
 	 * request, or to deny it with NLM_LCK_DENIED_GRACE_PERIOD. In either
@@ -669,32 +606,24 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl)
 	 */
 	do_vfs_lock(fl);
 
-	if (req->a_flags & RPC_TASK_ASYNC) {
-		status = nlmclnt_async_call(req, NLMPROC_UNLOCK,
-					&nlmclnt_unlock_ops);
-		/* Hrmf... Do the unlock early since locks_remove_posix()
-		 * really expects us to free the lock synchronously */
-		if (status < 0) {
-			nlmclnt_release_lockargs(req);
-			kfree(req);
-		}
-		return status;
-	}
+	if (req->a_flags & RPC_TASK_ASYNC)
+		return nlm_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops);
 
 	status = nlmclnt_call(req, NLMPROC_UNLOCK);
-	nlmclnt_release_lockargs(req);
 	if (status < 0)
-		return status;
+		goto out;
 
+	status = 0;
 	if (resp->status == NLM_LCK_GRANTED)
-		return 0;
+		goto out;
 
 	if (resp->status != NLM_LCK_DENIED_NOLOCKS)
 		printk("lockd: unexpected unlock status: %d\n", resp->status);
-
 	/* What to do now? I'm out of my depth... */
-
-	return -ENOLCK;
+	status = -ENOLCK;
+out:
+	nlm_release_call(req);
+	return status;
 }
 
 static void nlmclnt_unlock_callback(struct rpc_task *task, void *data)
@@ -716,9 +645,6 @@ static void nlmclnt_unlock_callback(struct rpc_task *task, void *data)
 	if (status != NLM_LCK_GRANTED)
 		printk(KERN_WARNING "lockd: unexpected unlock status: %d\n", status);
 die:
-	nlm_release_host(req->a_host);
-	nlmclnt_release_lockargs(req);
-	kfree(req);
 	return;
  retry_rebind:
 	nlm_rebind_host(req->a_host);
@@ -728,6 +654,7 @@ die:
 
 static const struct rpc_call_ops nlmclnt_unlock_ops = {
 	.rpc_call_done = nlmclnt_unlock_callback,
+	.rpc_release = nlmclnt_rpc_release,
 };
 
 /*
@@ -749,20 +676,15 @@ static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl
 	recalc_sigpending();
 	spin_unlock_irqrestore(&current->sighand->siglock, flags);
 
-	req = nlmclnt_alloc_call();
+	req = nlm_alloc_call(nlm_get_host(host));
 	if (!req)
 		return -ENOMEM;
-	req->a_host  = host;
 	req->a_flags = RPC_TASK_ASYNC;
 
 	nlmclnt_setlockargs(req, fl);
 	req->a_args.block = block;
 
-	status = nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops);
-	if (status < 0) {
-		nlmclnt_release_lockargs(req);
-		kfree(req);
-	}
+	status = nlm_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops);
 
 	spin_lock_irqsave(&current->sighand->siglock, flags);
 	current->blocked = oldset;
@@ -791,6 +713,7 @@ static void nlmclnt_cancel_callback(struct rpc_task *task, void *data)
 	switch (req->a_res.status) {
 	case NLM_LCK_GRANTED:
 	case NLM_LCK_DENIED_GRACE_PERIOD:
+	case NLM_LCK_DENIED:
 		/* Everything's good */
 		break;
 	case NLM_LCK_DENIED_NOLOCKS:
@@ -802,9 +725,6 @@ static void nlmclnt_cancel_callback(struct rpc_task *task, void *data)
 	}
 
 die:
-	nlm_release_host(req->a_host);
-	nlmclnt_release_lockargs(req);
-	kfree(req);
 	return;
 
 retry_cancel:
@@ -818,6 +738,7 @@ retry_cancel:
 
 static const struct rpc_call_ops nlmclnt_cancel_ops = {
 	.rpc_call_done = nlmclnt_cancel_callback,
+	.rpc_release = nlmclnt_rpc_release,
 };
 
 /*
diff --git a/fs/lockd/host.c b/fs/lockd/host.c
index 82f7a0b1d8ae..729ac427d359 100644
--- a/fs/lockd/host.c
+++ b/fs/lockd/host.c
@@ -16,6 +16,7 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/lockd/lockd.h>
 #include <linux/lockd/sm_inter.h>
+#include <linux/mutex.h>
 
 
 #define NLMDBG_FACILITY		NLMDBG_HOSTCACHE
@@ -30,7 +31,7 @@
 static struct nlm_host *	nlm_hosts[NLM_HOST_NRHASH];
 static unsigned long		next_gc;
 static int			nrhosts;
-static DECLARE_MUTEX(nlm_host_sema);
+static DEFINE_MUTEX(nlm_host_mutex);
 
 
 static void			nlm_gc_hosts(void);
@@ -71,7 +72,7 @@ nlm_lookup_host(int server, struct sockaddr_in *sin,
 	hash = NLM_ADDRHASH(sin->sin_addr.s_addr);
 
 	/* Lock hash table */
-	down(&nlm_host_sema);
+	mutex_lock(&nlm_host_mutex);
 
 	if (time_after_eq(jiffies, next_gc))
 		nlm_gc_hosts();
@@ -91,7 +92,7 @@ nlm_lookup_host(int server, struct sockaddr_in *sin,
 				nlm_hosts[hash] = host;
 			}
 			nlm_get_host(host);
-			up(&nlm_host_sema);
+			mutex_unlock(&nlm_host_mutex);
 			return host;
 		}
 	}
@@ -123,12 +124,14 @@ nlm_lookup_host(int server, struct sockaddr_in *sin,
 	nlm_hosts[hash]    = host;
 	INIT_LIST_HEAD(&host->h_lockowners);
 	spin_lock_init(&host->h_lock);
+	INIT_LIST_HEAD(&host->h_granted);
+	INIT_LIST_HEAD(&host->h_reclaim);
 
 	if (++nrhosts > NLM_HOST_MAX)
 		next_gc = 0;
 
 nohost:
-	up(&nlm_host_sema);
+	mutex_unlock(&nlm_host_mutex);
 	return host;
 }
 
@@ -139,19 +142,19 @@ nlm_find_client(void)
 	 * and return it
 	 */
 	int hash;
-	down(&nlm_host_sema);
+	mutex_lock(&nlm_host_mutex);
 	for (hash = 0 ; hash < NLM_HOST_NRHASH; hash++) {
 		struct nlm_host *host, **hp;
 		for (hp = &nlm_hosts[hash]; (host = *hp) != 0; hp = &host->h_next) {
 			if (host->h_server &&
 			    host->h_killed == 0) {
 				nlm_get_host(host);
-				up(&nlm_host_sema);
+				mutex_unlock(&nlm_host_mutex);
 				return host;
 			}
 		}
 	}
-	up(&nlm_host_sema);
+	mutex_unlock(&nlm_host_mutex);
 	return NULL;
 }
 
@@ -191,11 +194,12 @@ nlm_bind_host(struct nlm_host *host)
 		xprt->resvport = 1;	/* NLM requires a reserved port */
 
 		/* Existing NLM servers accept AUTH_UNIX only */
-		clnt = rpc_create_client(xprt, host->h_name, &nlm_program,
+		clnt = rpc_new_client(xprt, host->h_name, &nlm_program,
 					host->h_version, RPC_AUTH_UNIX);
 		if (IS_ERR(clnt))
 			goto forgetit;
 		clnt->cl_autobind = 1;	/* turn on pmap queries */
+		clnt->cl_softrtry = 1; /* All queries are soft */
 
 		host->h_rpcclnt = clnt;
 	}
@@ -242,8 +246,12 @@ void nlm_release_host(struct nlm_host *host)
 {
 	if (host != NULL) {
 		dprintk("lockd: release host %s\n", host->h_name);
-		atomic_dec(&host->h_count);
 		BUG_ON(atomic_read(&host->h_count) < 0);
+		if (atomic_dec_and_test(&host->h_count)) {
+			BUG_ON(!list_empty(&host->h_lockowners));
+			BUG_ON(!list_empty(&host->h_granted));
+			BUG_ON(!list_empty(&host->h_reclaim));
+		}
 	}
 }
 
@@ -258,7 +266,7 @@ nlm_shutdown_hosts(void)
 	int		i;
 
 	dprintk("lockd: shutting down host module\n");
-	down(&nlm_host_sema);
+	mutex_lock(&nlm_host_mutex);
 
 	/* First, make all hosts eligible for gc */
 	dprintk("lockd: nuking all hosts...\n");
@@ -269,7 +277,7 @@ nlm_shutdown_hosts(void)
 
 	/* Then, perform a garbage collection pass */
 	nlm_gc_hosts();
-	up(&nlm_host_sema);
+	mutex_unlock(&nlm_host_mutex);
 
 	/* complain if any hosts are left */
 	if (nrhosts) {
@@ -331,7 +339,6 @@ nlm_gc_hosts(void)
 					rpc_destroy_client(host->h_rpcclnt);
 				}
 			}
-			BUG_ON(!list_empty(&host->h_lockowners));
 			kfree(host);
 			nrhosts--;
 		}
diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c
index 0edc03e67966..3fc683f46b3e 100644
--- a/fs/lockd/mon.c
+++ b/fs/lockd/mon.c
@@ -35,6 +35,10 @@ nsm_mon_unmon(struct nlm_host *host, u32 proc, struct nsm_res *res)
 	struct rpc_clnt	*clnt;
 	int		status;
 	struct nsm_args	args;
+	struct rpc_message msg = {
+		.rpc_argp	= &args,
+		.rpc_resp	= res,
+	};
 
 	clnt = nsm_create();
 	if (IS_ERR(clnt)) {
@@ -49,7 +53,8 @@ nsm_mon_unmon(struct nlm_host *host, u32 proc, struct nsm_res *res)
 	args.proc = NLMPROC_NSM_NOTIFY;
 	memset(res, 0, sizeof(*res));
 
-	status = rpc_call(clnt, proc, &args, res, 0);
+	msg.rpc_proc = &clnt->cl_procinfo[proc];
+	status = rpc_call_sync(clnt, &msg, 0);
 	if (status < 0)
 		printk(KERN_DEBUG "nsm_mon_unmon: rpc failed, status=%d\n",
 			status);
@@ -214,18 +219,22 @@ static struct rpc_procinfo	nsm_procedures[] = {
 		.p_encode	= (kxdrproc_t) xdr_encode_mon,
 		.p_decode	= (kxdrproc_t) xdr_decode_stat_res,
 		.p_bufsiz	= MAX(SM_mon_sz, SM_monres_sz) << 2,
+		.p_statidx	= SM_MON,
+		.p_name		= "MONITOR",
 	},
 [SM_UNMON] = {
 		.p_proc		= SM_UNMON,
 		.p_encode	= (kxdrproc_t) xdr_encode_unmon,
 		.p_decode	= (kxdrproc_t) xdr_decode_stat,
 		.p_bufsiz	= MAX(SM_mon_id_sz, SM_unmonres_sz) << 2,
+		.p_statidx	= SM_UNMON,
+		.p_name		= "UNMONITOR",
 	},
 };
 
 static struct rpc_version	nsm_version1 = {
-		.number		= 1, 
-		.nrprocs	= sizeof(nsm_procedures)/sizeof(nsm_procedures[0]),
+		.number		= 1,
+		.nrprocs	= ARRAY_SIZE(nsm_procedures),
 		.procs		= nsm_procedures
 };
 
@@ -238,7 +247,7 @@ static struct rpc_stat		nsm_stats;
 static struct rpc_program	nsm_program = {
 		.name		= "statd",
 		.number		= SM_PROGRAM,
-		.nrvers		= sizeof(nsm_version)/sizeof(nsm_version[0]),
+		.nrvers		= ARRAY_SIZE(nsm_version),
 		.version	= nsm_version,
 		.stats		= &nsm_stats
 };
diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index 71a30b416d1a..fd56c8872f34 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -25,6 +25,7 @@
 #include <linux/slab.h>
 #include <linux/smp.h>
 #include <linux/smp_lock.h>
+#include <linux/mutex.h>
 
 #include <linux/sunrpc/types.h>
 #include <linux/sunrpc/stats.h>
@@ -43,13 +44,13 @@ static struct svc_program	nlmsvc_program;
 struct nlmsvc_binding *		nlmsvc_ops;
 EXPORT_SYMBOL(nlmsvc_ops);
 
-static DECLARE_MUTEX(nlmsvc_sema);
+static DEFINE_MUTEX(nlmsvc_mutex);
 static unsigned int		nlmsvc_users;
 static pid_t			nlmsvc_pid;
 int				nlmsvc_grace_period;
 unsigned long			nlmsvc_timeout;
 
-static DECLARE_MUTEX_LOCKED(lockd_start);
+static DECLARE_COMPLETION(lockd_start_done);
 static DECLARE_WAIT_QUEUE_HEAD(lockd_exit);
 
 /*
@@ -112,7 +113,7 @@ lockd(struct svc_rqst *rqstp)
 	 * Let our maker know we're running.
 	 */
 	nlmsvc_pid = current->pid;
-	up(&lockd_start);
+	complete(&lockd_start_done);
 
 	daemonize("lockd");
 
@@ -215,7 +216,7 @@ lockd_up(void)
 	struct svc_serv *	serv;
 	int			error = 0;
 
-	down(&nlmsvc_sema);
+	mutex_lock(&nlmsvc_mutex);
 	/*
 	 * Unconditionally increment the user count ... this is
 	 * the number of clients who _want_ a lockd process.
@@ -263,7 +264,7 @@ lockd_up(void)
 			"lockd_up: create thread failed, error=%d\n", error);
 		goto destroy_and_out;
 	}
-	down(&lockd_start);
+	wait_for_completion(&lockd_start_done);
 
 	/*
 	 * Note: svc_serv structures have an initial use count of 1,
@@ -272,7 +273,7 @@ lockd_up(void)
 destroy_and_out:
 	svc_destroy(serv);
 out:
-	up(&nlmsvc_sema);
+	mutex_unlock(&nlmsvc_mutex);
 	return error;
 }
 EXPORT_SYMBOL(lockd_up);
@@ -285,7 +286,7 @@ lockd_down(void)
 {
 	static int warned;
 
-	down(&nlmsvc_sema);
+	mutex_lock(&nlmsvc_mutex);
 	if (nlmsvc_users) {
 		if (--nlmsvc_users)
 			goto out;
@@ -315,7 +316,7 @@ lockd_down(void)
 	recalc_sigpending();
 	spin_unlock_irq(&current->sighand->siglock);
 out:
-	up(&nlmsvc_sema);
+	mutex_unlock(&nlmsvc_mutex);
 }
 EXPORT_SYMBOL(lockd_down);
 
@@ -509,7 +510,7 @@ static struct svc_version *	nlmsvc_version[] = {
 
 static struct svc_stat		nlmsvc_stats;
 
-#define NLM_NRVERS	(sizeof(nlmsvc_version)/sizeof(nlmsvc_version[0]))
+#define NLM_NRVERS	ARRAY_SIZE(nlmsvc_version)
 static struct svc_program	nlmsvc_program = {
 	.pg_prog		= NLM_PROGRAM,		/* program number */
 	.pg_nvers		= NLM_NRVERS,		/* number of entries in nlmsvc_version */
diff --git a/fs/lockd/svc4proc.c b/fs/lockd/svc4proc.c
index b10f913aa06a..a2dd9ccb9b32 100644
--- a/fs/lockd/svc4proc.c
+++ b/fs/lockd/svc4proc.c
@@ -21,10 +21,6 @@
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 
-static u32	nlm4svc_callback(struct svc_rqst *, u32, struct nlm_res *);
-
-static const struct rpc_call_ops nlm4svc_callback_ops;
-
 /*
  * Obtain client and file from arguments
  */
@@ -234,83 +230,89 @@ nlm4svc_proc_granted(struct svc_rqst *rqstp, struct nlm_args *argp,
 }
 
 /*
+ * This is the generic lockd callback for async RPC calls
+ */
+static void nlm4svc_callback_exit(struct rpc_task *task, void *data)
+{
+	dprintk("lockd: %4d callback returned %d\n", task->tk_pid,
+			-task->tk_status);
+}
+
+static void nlm4svc_callback_release(void *data)
+{
+	nlm_release_call(data);
+}
+
+static const struct rpc_call_ops nlm4svc_callback_ops = {
+	.rpc_call_done = nlm4svc_callback_exit,
+	.rpc_release = nlm4svc_callback_release,
+};
+
+/*
  * `Async' versions of the above service routines. They aren't really,
  * because we send the callback before the reply proper. I hope this
  * doesn't break any clients.
  */
-static int
-nlm4svc_proc_test_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
-					     void	     *resp)
+static int nlm4svc_callback(struct svc_rqst *rqstp, u32 proc, struct nlm_args *argp,
+		int (*func)(struct svc_rqst *, struct nlm_args *, struct nlm_res  *))
 {
-	struct nlm_res	res;
-	u32		stat;
+	struct nlm_host	*host;
+	struct nlm_rqst	*call;
+	int stat;
 
-	dprintk("lockd: TEST_MSG      called\n");
-	memset(&res, 0, sizeof(res));
+	host = nlmsvc_lookup_host(rqstp);
+	if (host == NULL)
+		return rpc_system_err;
+
+	call = nlm_alloc_call(host);
+	if (call == NULL)
+		return rpc_system_err;
 
-	if ((stat = nlm4svc_proc_test(rqstp, argp, &res)) == 0)
-		stat = nlm4svc_callback(rqstp, NLMPROC_TEST_RES, &res);
-	return stat;
+	stat = func(rqstp, argp, &call->a_res);
+	if (stat != 0) {
+		nlm_release_call(call);
+		return stat;
+	}
+
+	call->a_flags = RPC_TASK_ASYNC;
+	if (nlm_async_reply(call, proc, &nlm4svc_callback_ops) < 0)
+		return rpc_system_err;
+	return rpc_success;
 }
 
-static int
-nlm4svc_proc_lock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+static int nlm4svc_proc_test_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
 					     void	     *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
+	dprintk("lockd: TEST_MSG      called\n");
+	return nlm4svc_callback(rqstp, NLMPROC_TEST_RES, argp, nlm4svc_proc_test);
+}
 
+static int nlm4svc_proc_lock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+					     void	     *resp)
+{
 	dprintk("lockd: LOCK_MSG      called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlm4svc_proc_lock(rqstp, argp, &res)) == 0)
-		stat = nlm4svc_callback(rqstp, NLMPROC_LOCK_RES, &res);
-	return stat;
+	return nlm4svc_callback(rqstp, NLMPROC_LOCK_RES, argp, nlm4svc_proc_lock);
 }
 
-static int
-nlm4svc_proc_cancel_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+static int nlm4svc_proc_cancel_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
 					       void	       *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
-
 	dprintk("lockd: CANCEL_MSG    called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlm4svc_proc_cancel(rqstp, argp, &res)) == 0)
-		stat = nlm4svc_callback(rqstp, NLMPROC_CANCEL_RES, &res);
-	return stat;
+	return nlm4svc_callback(rqstp, NLMPROC_CANCEL_RES, argp, nlm4svc_proc_cancel);
 }
 
-static int
-nlm4svc_proc_unlock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+static int nlm4svc_proc_unlock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
                                                void            *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
-
 	dprintk("lockd: UNLOCK_MSG    called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlm4svc_proc_unlock(rqstp, argp, &res)) == 0)
-		stat = nlm4svc_callback(rqstp, NLMPROC_UNLOCK_RES, &res);
-	return stat;
+	return nlm4svc_callback(rqstp, NLMPROC_UNLOCK_RES, argp, nlm4svc_proc_unlock);
 }
 
-static int
-nlm4svc_proc_granted_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+static int nlm4svc_proc_granted_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
                                                 void            *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
-
 	dprintk("lockd: GRANTED_MSG   called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlm4svc_proc_granted(rqstp, argp, &res)) == 0)
-		stat = nlm4svc_callback(rqstp, NLMPROC_GRANTED_RES, &res);
-	return stat;
+	return nlm4svc_callback(rqstp, NLMPROC_GRANTED_RES, argp, nlm4svc_proc_granted);
 }
 
 /*
@@ -472,55 +474,6 @@ nlm4svc_proc_granted_res(struct svc_rqst *rqstp, struct nlm_res  *argp,
 
 
 /*
- * This is the generic lockd callback for async RPC calls
- */
-static u32
-nlm4svc_callback(struct svc_rqst *rqstp, u32 proc, struct nlm_res *resp)
-{
-	struct nlm_host	*host;
-	struct nlm_rqst	*call;
-
-	if (!(call = nlmclnt_alloc_call()))
-		return rpc_system_err;
-
-	host = nlmclnt_lookup_host(&rqstp->rq_addr,
-				rqstp->rq_prot, rqstp->rq_vers);
-	if (!host) {
-		kfree(call);
-		return rpc_system_err;
-	}
-
-	call->a_flags = RPC_TASK_ASYNC;
-	call->a_host  = host;
-	memcpy(&call->a_args, resp, sizeof(*resp));
-
-	if (nlmsvc_async_call(call, proc, &nlm4svc_callback_ops) < 0)
-		goto error;
-
-	return rpc_success;
- error:
-	kfree(call);
-	nlm_release_host(host);
-	return rpc_system_err;
-}
-
-static void nlm4svc_callback_exit(struct rpc_task *task, void *data)
-{
-	struct nlm_rqst	*call = data;
-
-	if (task->tk_status < 0) {
-		dprintk("lockd: %4d callback failed (errno = %d)\n",
-					task->tk_pid, -task->tk_status);
-	}
-	nlm_release_host(call->a_host);
-	kfree(call);
-}
-
-static const struct rpc_call_ops nlm4svc_callback_ops = {
-	.rpc_call_done = nlm4svc_callback_exit,
-};
-
-/*
  * NLM Server procedures.
  */
 
diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index 9cfced65d4a2..d2b66bad7d50 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -39,9 +39,12 @@
 #define nlm_deadlock	nlm_lck_denied
 #endif
 
+static void nlmsvc_release_block(struct nlm_block *block);
 static void	nlmsvc_insert_block(struct nlm_block *block, unsigned long);
 static int	nlmsvc_remove_block(struct nlm_block *block);
 
+static int nlmsvc_setgrantargs(struct nlm_rqst *call, struct nlm_lock *lock);
+static void nlmsvc_freegrantargs(struct nlm_rqst *call);
 static const struct rpc_call_ops nlmsvc_grant_ops;
 
 /*
@@ -58,6 +61,7 @@ nlmsvc_insert_block(struct nlm_block *block, unsigned long when)
 	struct nlm_block **bp, *b;
 
 	dprintk("lockd: nlmsvc_insert_block(%p, %ld)\n", block, when);
+	kref_get(&block->b_count);
 	if (block->b_queued)
 		nlmsvc_remove_block(block);
 	bp = &nlm_blocked;
@@ -90,6 +94,7 @@ nlmsvc_remove_block(struct nlm_block *block)
 		if (b == block) {
 			*bp = block->b_next;
 			block->b_queued = 0;
+			nlmsvc_release_block(block);
 			return 1;
 		}
 	}
@@ -98,11 +103,10 @@ nlmsvc_remove_block(struct nlm_block *block)
 }
 
 /*
- * Find a block for a given lock and optionally remove it from
- * the list.
+ * Find a block for a given lock
  */
 static struct nlm_block *
-nlmsvc_lookup_block(struct nlm_file *file, struct nlm_lock *lock, int remove)
+nlmsvc_lookup_block(struct nlm_file *file, struct nlm_lock *lock)
 {
 	struct nlm_block	**head, *block;
 	struct file_lock	*fl;
@@ -112,17 +116,14 @@ nlmsvc_lookup_block(struct nlm_file *file, struct nlm_lock *lock, int remove)
 				(long long)lock->fl.fl_start,
 				(long long)lock->fl.fl_end, lock->fl.fl_type);
 	for (head = &nlm_blocked; (block = *head) != 0; head = &block->b_next) {
-		fl = &block->b_call.a_args.lock.fl;
+		fl = &block->b_call->a_args.lock.fl;
 		dprintk("lockd: check f=%p pd=%d %Ld-%Ld ty=%d cookie=%s\n",
 				block->b_file, fl->fl_pid,
 				(long long)fl->fl_start,
 				(long long)fl->fl_end, fl->fl_type,
-				nlmdbg_cookie2a(&block->b_call.a_args.cookie));
+				nlmdbg_cookie2a(&block->b_call->a_args.cookie));
 		if (block->b_file == file && nlm_compare_locks(fl, &lock->fl)) {
-			if (remove) {
-				*head = block->b_next;
-				block->b_queued = 0;
-			}
+			kref_get(&block->b_count);
 			return block;
 		}
 	}
@@ -150,11 +151,13 @@ nlmsvc_find_block(struct nlm_cookie *cookie,  struct sockaddr_in *sin)
 	for (block = nlm_blocked; block; block = block->b_next) {
 		dprintk("cookie: head of blocked queue %p, block %p\n", 
 			nlm_blocked, block);
-		if (nlm_cookie_match(&block->b_call.a_args.cookie,cookie)
+		if (nlm_cookie_match(&block->b_call->a_args.cookie,cookie)
 				&& nlm_cmp_addr(sin, &block->b_host->h_addr))
 			break;
 	}
 
+	if (block != NULL)
+		kref_get(&block->b_count);
 	return block;
 }
 
@@ -174,27 +177,30 @@ nlmsvc_create_block(struct svc_rqst *rqstp, struct nlm_file *file,
 {
 	struct nlm_block	*block;
 	struct nlm_host		*host;
-	struct nlm_rqst		*call;
+	struct nlm_rqst		*call = NULL;
 
 	/* Create host handle for callback */
-	host = nlmclnt_lookup_host(&rqstp->rq_addr,
-				rqstp->rq_prot, rqstp->rq_vers);
+	host = nlmsvc_lookup_host(rqstp);
 	if (host == NULL)
 		return NULL;
 
+	call = nlm_alloc_call(host);
+	if (call == NULL)
+		return NULL;
+
 	/* Allocate memory for block, and initialize arguments */
-	if (!(block = (struct nlm_block *) kmalloc(sizeof(*block), GFP_KERNEL)))
+	block = kzalloc(sizeof(*block), GFP_KERNEL);
+	if (block == NULL)
 		goto failed;
-	memset(block, 0, sizeof(*block));
-	locks_init_lock(&block->b_call.a_args.lock.fl);
-	locks_init_lock(&block->b_call.a_res.lock.fl);
+	kref_init(&block->b_count);
 
-	if (!nlmclnt_setgrantargs(&block->b_call, lock))
+	if (!nlmsvc_setgrantargs(call, lock))
 		goto failed_free;
 
 	/* Set notifier function for VFS, and init args */
-	block->b_call.a_args.lock.fl.fl_lmops = &nlmsvc_lock_operations;
-	block->b_call.a_args.cookie = *cookie;	/* see above */
+	call->a_args.lock.fl.fl_flags |= FL_SLEEP;
+	call->a_args.lock.fl.fl_lmops = &nlmsvc_lock_operations;
+	call->a_args.cookie = *cookie;	/* see above */
 
 	dprintk("lockd: created block %p...\n", block);
 
@@ -202,22 +208,23 @@ nlmsvc_create_block(struct svc_rqst *rqstp, struct nlm_file *file,
 	block->b_daemon = rqstp->rq_server;
 	block->b_host   = host;
 	block->b_file   = file;
+	file->f_count++;
 
 	/* Add to file's list of blocks */
 	block->b_fnext  = file->f_blocks;
 	file->f_blocks  = block;
 
 	/* Set up RPC arguments for callback */
-	call = &block->b_call;
-	call->a_host    = host;
+	block->b_call = call;
 	call->a_flags   = RPC_TASK_ASYNC;
+	call->a_block = block;
 
 	return block;
 
 failed_free:
 	kfree(block);
 failed:
-	nlm_release_host(host);
+	nlm_release_call(call);
 	return NULL;
 }
 
@@ -227,29 +234,26 @@ failed:
  * It is the caller's responsibility to check whether the file
  * can be closed hereafter.
  */
-static int
-nlmsvc_delete_block(struct nlm_block *block, int unlock)
+static int nlmsvc_unlink_block(struct nlm_block *block)
 {
-	struct file_lock	*fl = &block->b_call.a_args.lock.fl;
-	struct nlm_file		*file = block->b_file;
-	struct nlm_block	**bp;
-	int status = 0;
-
-	dprintk("lockd: deleting block %p...\n", block);
+	int status;
+	dprintk("lockd: unlinking block %p...\n", block);
 
 	/* Remove block from list */
+	status = posix_unblock_lock(block->b_file->f_file, &block->b_call->a_args.lock.fl);
 	nlmsvc_remove_block(block);
-	if (unlock)
-		status = posix_unblock_lock(file->f_file, fl);
+	return status;
+}
 
-	/* If the block is in the middle of a GRANT callback,
-	 * don't kill it yet. */
-	if (block->b_incall) {
-		nlmsvc_insert_block(block, NLM_NEVER);
-		block->b_done = 1;
-		return status;
-	}
+static void nlmsvc_free_block(struct kref *kref)
+{
+	struct nlm_block *block = container_of(kref, struct nlm_block, b_count);
+	struct nlm_file		*file = block->b_file;
+	struct nlm_block	**bp;
 
+	dprintk("lockd: freeing block %p...\n", block);
+
+	down(&file->f_sema);
 	/* Remove block from file's list of blocks */
 	for (bp = &file->f_blocks; *bp; bp = &(*bp)->b_fnext) {
 		if (*bp == block) {
@@ -257,36 +261,93 @@ nlmsvc_delete_block(struct nlm_block *block, int unlock)
 			break;
 		}
 	}
+	up(&file->f_sema);
 
-	if (block->b_host)
-		nlm_release_host(block->b_host);
-	nlmclnt_freegrantargs(&block->b_call);
+	nlmsvc_freegrantargs(block->b_call);
+	nlm_release_call(block->b_call);
+	nlm_release_file(block->b_file);
 	kfree(block);
-	return status;
+}
+
+static void nlmsvc_release_block(struct nlm_block *block)
+{
+	if (block != NULL)
+		kref_put(&block->b_count, nlmsvc_free_block);
+}
+
+static void nlmsvc_act_mark(struct nlm_host *host, struct nlm_file *file)
+{
+	struct nlm_block *block;
+
+	down(&file->f_sema);
+	for (block = file->f_blocks; block != NULL; block = block->b_fnext)
+		block->b_host->h_inuse = 1;
+	up(&file->f_sema);
+}
+
+static void nlmsvc_act_unlock(struct nlm_host *host, struct nlm_file *file)
+{
+	struct nlm_block *block;
+
+restart:
+	down(&file->f_sema);
+	for (block = file->f_blocks; block != NULL; block = block->b_fnext) {
+		if (host != NULL && host != block->b_host)
+			continue;
+		if (!block->b_queued)
+			continue;
+		kref_get(&block->b_count);
+		up(&file->f_sema);
+		nlmsvc_unlink_block(block);
+		nlmsvc_release_block(block);
+		goto restart;
+	}
+	up(&file->f_sema);
 }
 
 /*
  * Loop over all blocks and perform the action specified.
  * (NLM_ACT_CHECK handled by nlmsvc_inspect_file).
  */
-int
+void
 nlmsvc_traverse_blocks(struct nlm_host *host, struct nlm_file *file, int action)
 {
-	struct nlm_block	*block, *next;
-	/* XXX: Will everything get cleaned up if we don't unlock here? */
+	if (action == NLM_ACT_MARK)
+		nlmsvc_act_mark(host, file);
+	else
+		nlmsvc_act_unlock(host, file);
+}
 
-	down(&file->f_sema);
-	for (block = file->f_blocks; block; block = next) {
-		next = block->b_fnext;
-		if (action == NLM_ACT_MARK)
-			block->b_host->h_inuse = 1;
-		else if (action == NLM_ACT_UNLOCK) {
-			if (host == NULL || host == block->b_host)
-				nlmsvc_delete_block(block, 1);
-		}
+/*
+ * Initialize arguments for GRANTED call. The nlm_rqst structure
+ * has been cleared already.
+ */
+static int nlmsvc_setgrantargs(struct nlm_rqst *call, struct nlm_lock *lock)
+{
+	locks_copy_lock(&call->a_args.lock.fl, &lock->fl);
+	memcpy(&call->a_args.lock.fh, &lock->fh, sizeof(call->a_args.lock.fh));
+	call->a_args.lock.caller = system_utsname.nodename;
+	call->a_args.lock.oh.len = lock->oh.len;
+
+	/* set default data area */
+	call->a_args.lock.oh.data = call->a_owner;
+	call->a_args.lock.svid = lock->fl.fl_pid;
+
+	if (lock->oh.len > NLMCLNT_OHSIZE) {
+		void *data = kmalloc(lock->oh.len, GFP_KERNEL);
+		if (!data)
+			return 0;
+		call->a_args.lock.oh.data = (u8 *) data;
 	}
-	up(&file->f_sema);
-	return 0;
+
+	memcpy(call->a_args.lock.oh.data, lock->oh.data, lock->oh.len);
+	return 1;
+}
+
+static void nlmsvc_freegrantargs(struct nlm_rqst *call)
+{
+	if (call->a_args.lock.oh.data != call->a_owner)
+		kfree(call->a_args.lock.oh.data);
 }
 
 /*
@@ -297,9 +358,9 @@ u32
 nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 			struct nlm_lock *lock, int wait, struct nlm_cookie *cookie)
 {
-	struct file_lock	*conflock;
-	struct nlm_block	*block;
+	struct nlm_block	*block, *newblock = NULL;
 	int			error;
+	u32			ret;
 
 	dprintk("lockd: nlmsvc_lock(%s/%ld, ty=%d, pi=%d, %Ld-%Ld, bl=%d)\n",
 				file->f_file->f_dentry->d_inode->i_sb->s_id,
@@ -310,69 +371,65 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file,
 				wait);
 
 
-	/* Get existing block (in case client is busy-waiting) */
-	block = nlmsvc_lookup_block(file, lock, 0);
-
-	lock->fl.fl_flags |= FL_LOCKD;
-
+	lock->fl.fl_flags &= ~FL_SLEEP;
 again:
 	/* Lock file against concurrent access */
 	down(&file->f_sema);
+	/* Get existing block (in case client is busy-waiting) */
+	block = nlmsvc_lookup_block(file, lock);
+	if (block == NULL) {
+		if (newblock != NULL)
+			lock = &newblock->b_call->a_args.lock;
+	} else
+		lock = &block->b_call->a_args.lock;
 
-	if (!(conflock = posix_test_lock(file->f_file, &lock->fl))) {
-		error = posix_lock_file(file->f_file, &lock->fl);
+	error = posix_lock_file(file->f_file, &lock->fl);
+	lock->fl.fl_flags &= ~FL_SLEEP;
 
-		if (block)
-			nlmsvc_delete_block(block, 0);
-		up(&file->f_sema);
+	dprintk("lockd: posix_lock_file returned %d\n", error);
 
-		dprintk("lockd: posix_lock_file returned %d\n", -error);
-		switch(-error) {
+	switch(error) {
 		case 0:
-			return nlm_granted;
-		case EDEADLK:
-			return nlm_deadlock;
-		case EAGAIN:
-			return nlm_lck_denied;
+			ret = nlm_granted;
+			goto out;
+		case -EAGAIN:
+			break;
+		case -EDEADLK:
+			ret = nlm_deadlock;
+			goto out;
 		default:			/* includes ENOLCK */
-			return nlm_lck_denied_nolocks;
-		}
+			ret = nlm_lck_denied_nolocks;
+			goto out;
 	}
 
-	if (!wait) {
-		up(&file->f_sema);
-		return nlm_lck_denied;
-	}
+	ret = nlm_lck_denied;
+	if (!wait)
+		goto out;
 
-	if (posix_locks_deadlock(&lock->fl, conflock)) {
-		up(&file->f_sema);
-		return nlm_deadlock;
-	}
+	ret = nlm_lck_blocked;
+	if (block != NULL)
+		goto out;
 
 	/* If we don't have a block, create and initialize it. Then
 	 * retry because we may have slept in kmalloc. */
 	/* We have to release f_sema as nlmsvc_create_block may try to
 	 * to claim it while doing host garbage collection */
-	if (block == NULL) {
+	if (newblock == NULL) {
 		up(&file->f_sema);
 		dprintk("lockd: blocking on this lock (allocating).\n");
-		if (!(block = nlmsvc_create_block(rqstp, file, lock, cookie)))
+		if (!(newblock = nlmsvc_create_block(rqstp, file, lock, cookie)))
 			return nlm_lck_denied_nolocks;
 		goto again;
 	}
 
 	/* Append to list of blocked */
-	nlmsvc_insert_block(block, NLM_NEVER);
-
-	if (list_empty(&block->b_call.a_args.lock.fl.fl_block)) {
-		/* Now add block to block list of the conflicting lock
-		   if we haven't done so. */
-		dprintk("lockd: blocking on this lock.\n");
-		posix_block_lock(conflock, &block->b_call.a_args.lock.fl);
-	}
-
+	nlmsvc_insert_block(newblock, NLM_NEVER);
+out:
 	up(&file->f_sema);
-	return nlm_lck_blocked;
+	nlmsvc_release_block(newblock);
+	nlmsvc_release_block(block);
+	dprintk("lockd: nlmsvc_lock returned %u\n", ret);
+	return ret;
 }
 
 /*
@@ -382,8 +439,6 @@ u32
 nlmsvc_testlock(struct nlm_file *file, struct nlm_lock *lock,
 				       struct nlm_lock *conflock)
 {
-	struct file_lock	*fl;
-
 	dprintk("lockd: nlmsvc_testlock(%s/%ld, ty=%d, %Ld-%Ld)\n",
 				file->f_file->f_dentry->d_inode->i_sb->s_id,
 				file->f_file->f_dentry->d_inode->i_ino,
@@ -391,13 +446,14 @@ nlmsvc_testlock(struct nlm_file *file, struct nlm_lock *lock,
 				(long long)lock->fl.fl_start,
 				(long long)lock->fl.fl_end);
 
-	if ((fl = posix_test_lock(file->f_file, &lock->fl)) != NULL) {
+	if (posix_test_lock(file->f_file, &lock->fl, &conflock->fl)) {
 		dprintk("lockd: conflicting lock(ty=%d, %Ld-%Ld)\n",
-				fl->fl_type, (long long)fl->fl_start,
-				(long long)fl->fl_end);
+				conflock->fl.fl_type,
+				(long long)conflock->fl.fl_start,
+				(long long)conflock->fl.fl_end);
 		conflock->caller = "somehost";	/* FIXME */
 		conflock->oh.len = 0;		/* don't return OH info */
-		conflock->fl = *fl;
+		conflock->svid = conflock->fl.fl_pid;
 		return nlm_lck_denied;
 	}
 
@@ -453,9 +509,12 @@ nlmsvc_cancel_blocked(struct nlm_file *file, struct nlm_lock *lock)
 				(long long)lock->fl.fl_end);
 
 	down(&file->f_sema);
-	if ((block = nlmsvc_lookup_block(file, lock, 1)) != NULL)
-		status = nlmsvc_delete_block(block, 1);
+	block = nlmsvc_lookup_block(file, lock);
 	up(&file->f_sema);
+	if (block != NULL) {
+		status = nlmsvc_unlink_block(block);
+		nlmsvc_release_block(block);
+	}
 	return status ? nlm_lck_denied : nlm_granted;
 }
 
@@ -473,7 +532,7 @@ nlmsvc_notify_blocked(struct file_lock *fl)
 
 	dprintk("lockd: VFS unblock notification for block %p\n", fl);
 	for (bp = &nlm_blocked; (block = *bp) != 0; bp = &block->b_next) {
-		if (nlm_compare_locks(&block->b_call.a_args.lock.fl, fl)) {
+		if (nlm_compare_locks(&block->b_call->a_args.lock.fl, fl)) {
 			nlmsvc_insert_block(block, 0);
 			svc_wake_up(block->b_daemon);
 			return;
@@ -508,17 +567,13 @@ static void
 nlmsvc_grant_blocked(struct nlm_block *block)
 {
 	struct nlm_file		*file = block->b_file;
-	struct nlm_lock		*lock = &block->b_call.a_args.lock;
-	struct file_lock	*conflock;
+	struct nlm_lock		*lock = &block->b_call->a_args.lock;
 	int			error;
 
 	dprintk("lockd: grant blocked lock %p\n", block);
 
-	/* First thing is lock the file */
-	down(&file->f_sema);
-
 	/* Unlink block request from list */
-	nlmsvc_remove_block(block);
+	nlmsvc_unlink_block(block);
 
 	/* If b_granted is true this means we've been here before.
 	 * Just retry the grant callback, possibly refreshing the RPC
@@ -529,24 +584,21 @@ nlmsvc_grant_blocked(struct nlm_block *block)
 	}
 
 	/* Try the lock operation again */
-	if ((conflock = posix_test_lock(file->f_file, &lock->fl)) != NULL) {
-		/* Bummer, we blocked again */
+	lock->fl.fl_flags |= FL_SLEEP;
+	error = posix_lock_file(file->f_file, &lock->fl);
+	lock->fl.fl_flags &= ~FL_SLEEP;
+
+	switch (error) {
+	case 0:
+		break;
+	case -EAGAIN:
 		dprintk("lockd: lock still blocked\n");
 		nlmsvc_insert_block(block, NLM_NEVER);
-		posix_block_lock(conflock, &lock->fl);
-		up(&file->f_sema);
 		return;
-	}
-
-	/* Alright, no conflicting lock. Now lock it for real. If the
-	 * following yields an error, this is most probably due to low
-	 * memory. Retry the lock in a few seconds.
-	 */
-	if ((error = posix_lock_file(file->f_file, &lock->fl)) < 0) {
+	default:
 		printk(KERN_WARNING "lockd: unexpected error %d in %s!\n",
 				-error, __FUNCTION__);
 		nlmsvc_insert_block(block, 10 * HZ);
-		up(&file->f_sema);
 		return;
 	}
 
@@ -554,17 +606,15 @@ callback:
 	/* Lock was granted by VFS. */
 	dprintk("lockd: GRANTing blocked lock.\n");
 	block->b_granted = 1;
-	block->b_incall  = 1;
 
 	/* Schedule next grant callback in 30 seconds */
 	nlmsvc_insert_block(block, 30 * HZ);
 
 	/* Call the client */
-	nlm_get_host(block->b_call.a_host);
-	if (nlmsvc_async_call(&block->b_call, NLMPROC_GRANTED_MSG,
+	kref_get(&block->b_count);
+	if (nlm_async_call(block->b_call, NLMPROC_GRANTED_MSG,
 						&nlmsvc_grant_ops) < 0)
-		nlm_release_host(block->b_call.a_host);
-	up(&file->f_sema);
+		nlmsvc_release_block(block);
 }
 
 /*
@@ -578,20 +628,10 @@ callback:
 static void nlmsvc_grant_callback(struct rpc_task *task, void *data)
 {
 	struct nlm_rqst		*call = data;
-	struct nlm_block	*block;
+	struct nlm_block	*block = call->a_block;
 	unsigned long		timeout;
-	struct sockaddr_in	*peer_addr = RPC_PEERADDR(task->tk_client);
 
 	dprintk("lockd: GRANT_MSG RPC callback\n");
-	dprintk("callback: looking for cookie %s, host (%u.%u.%u.%u)\n",
-		nlmdbg_cookie2a(&call->a_args.cookie),
-		NIPQUAD(peer_addr->sin_addr.s_addr));
-	if (!(block = nlmsvc_find_block(&call->a_args.cookie, peer_addr))) {
-		dprintk("lockd: no block for cookie %s, host (%u.%u.%u.%u)\n",
-			nlmdbg_cookie2a(&call->a_args.cookie),
-			NIPQUAD(peer_addr->sin_addr.s_addr));
-		return;
-	}
 
 	/* Technically, we should down the file semaphore here. Since we
 	 * move the block towards the head of the queue only, no harm
@@ -608,13 +648,18 @@ static void nlmsvc_grant_callback(struct rpc_task *task, void *data)
 	}
 	nlmsvc_insert_block(block, timeout);
 	svc_wake_up(block->b_daemon);
-	block->b_incall = 0;
+}
 
-	nlm_release_host(call->a_host);
+void nlmsvc_grant_release(void *data)
+{
+	struct nlm_rqst		*call = data;
+
+	nlmsvc_release_block(call->a_block);
 }
 
 static const struct rpc_call_ops nlmsvc_grant_ops = {
 	.rpc_call_done = nlmsvc_grant_callback,
+	.rpc_release = nlmsvc_grant_release,
 };
 
 /*
@@ -634,25 +679,17 @@ nlmsvc_grant_reply(struct svc_rqst *rqstp, struct nlm_cookie *cookie, u32 status
 		return;
 	file = block->b_file;
 
-	file->f_count++;
-	down(&file->f_sema);
-	block = nlmsvc_find_block(cookie, &rqstp->rq_addr);
 	if (block) {
 		if (status == NLM_LCK_DENIED_GRACE_PERIOD) {
 			/* Try again in a couple of seconds */
 			nlmsvc_insert_block(block, 10 * HZ);
-			up(&file->f_sema);
 		} else {
 			/* Lock is now held by client, or has been rejected.
 			 * In both cases, the block should be removed. */
-			up(&file->f_sema);
-			if (status == NLM_LCK_GRANTED)
-				nlmsvc_delete_block(block, 0);
-			else
-				nlmsvc_delete_block(block, 1);
+			nlmsvc_unlink_block(block);
 		}
 	}
-	nlm_release_file(file);
+	nlmsvc_release_block(block);
 }
 
 /*
@@ -675,10 +712,12 @@ nlmsvc_retry_blocked(void)
 			break;
 		dprintk("nlmsvc_retry_blocked(%p, when=%ld, done=%d)\n",
 			block, block->b_when, block->b_done);
+		kref_get(&block->b_count);
 		if (block->b_done)
-			nlmsvc_delete_block(block, 0);
+			nlmsvc_unlink_block(block);
 		else
 			nlmsvc_grant_blocked(block);
+		nlmsvc_release_block(block);
 	}
 
 	if ((block = nlm_blocked) && block->b_when != NLM_NEVER)
diff --git a/fs/lockd/svcproc.c b/fs/lockd/svcproc.c
index 35681d9cf1fc..d210cf304e92 100644
--- a/fs/lockd/svcproc.c
+++ b/fs/lockd/svcproc.c
@@ -22,10 +22,6 @@
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 
-static u32	nlmsvc_callback(struct svc_rqst *, u32, struct nlm_res *);
-
-static const struct rpc_call_ops nlmsvc_callback_ops;
-
 #ifdef CONFIG_LOCKD_V4
 static u32
 cast_to_nlm(u32 status, u32 vers)
@@ -262,83 +258,91 @@ nlmsvc_proc_granted(struct svc_rqst *rqstp, struct nlm_args *argp,
 }
 
 /*
+ * This is the generic lockd callback for async RPC calls
+ */
+static void nlmsvc_callback_exit(struct rpc_task *task, void *data)
+{
+	dprintk("lockd: %4d callback returned %d\n", task->tk_pid,
+			-task->tk_status);
+}
+
+static void nlmsvc_callback_release(void *data)
+{
+	nlm_release_call(data);
+}
+
+static const struct rpc_call_ops nlmsvc_callback_ops = {
+	.rpc_call_done = nlmsvc_callback_exit,
+	.rpc_release = nlmsvc_callback_release,
+};
+
+/*
  * `Async' versions of the above service routines. They aren't really,
  * because we send the callback before the reply proper. I hope this
  * doesn't break any clients.
  */
-static int
-nlmsvc_proc_test_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
-					     void	     *resp)
+static int nlmsvc_callback(struct svc_rqst *rqstp, u32 proc, struct nlm_args *argp,
+		int (*func)(struct svc_rqst *, struct nlm_args *, struct nlm_res  *))
 {
-	struct nlm_res	res;
-	u32		stat;
+	struct nlm_host	*host;
+	struct nlm_rqst	*call;
+	int stat;
 
-	dprintk("lockd: TEST_MSG      called\n");
-	memset(&res, 0, sizeof(res));
+	host = nlmsvc_lookup_host(rqstp);
+	if (host == NULL)
+		return rpc_system_err;
 
-	if ((stat = nlmsvc_proc_test(rqstp, argp, &res)) == 0)
-		stat = nlmsvc_callback(rqstp, NLMPROC_TEST_RES, &res);
-	return stat;
+	call = nlm_alloc_call(host);
+	if (call == NULL)
+		return rpc_system_err;
+
+	stat = func(rqstp, argp, &call->a_res);
+	if (stat != 0) {
+		nlm_release_call(call);
+		return stat;
+	}
+
+	call->a_flags = RPC_TASK_ASYNC;
+	if (nlm_async_reply(call, proc, &nlmsvc_callback_ops) < 0)
+		return rpc_system_err;
+	return rpc_success;
 }
 
-static int
-nlmsvc_proc_lock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+static int nlmsvc_proc_test_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
 					     void	     *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
+	dprintk("lockd: TEST_MSG      called\n");
+	return nlmsvc_callback(rqstp, NLMPROC_TEST_RES, argp, nlmsvc_proc_test);
+}
 
+static int nlmsvc_proc_lock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+					     void	     *resp)
+{
 	dprintk("lockd: LOCK_MSG      called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlmsvc_proc_lock(rqstp, argp, &res)) == 0)
-		stat = nlmsvc_callback(rqstp, NLMPROC_LOCK_RES, &res);
-	return stat;
+	return nlmsvc_callback(rqstp, NLMPROC_LOCK_RES, argp, nlmsvc_proc_lock);
 }
 
-static int
-nlmsvc_proc_cancel_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
+static int nlmsvc_proc_cancel_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
 					       void	       *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
-
 	dprintk("lockd: CANCEL_MSG    called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlmsvc_proc_cancel(rqstp, argp, &res)) == 0)
-		stat = nlmsvc_callback(rqstp, NLMPROC_CANCEL_RES, &res);
-	return stat;
+	return nlmsvc_callback(rqstp, NLMPROC_CANCEL_RES, argp, nlmsvc_proc_cancel);
 }
 
 static int
 nlmsvc_proc_unlock_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
                                                void            *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
-
 	dprintk("lockd: UNLOCK_MSG    called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlmsvc_proc_unlock(rqstp, argp, &res)) == 0)
-		stat = nlmsvc_callback(rqstp, NLMPROC_UNLOCK_RES, &res);
-	return stat;
+	return nlmsvc_callback(rqstp, NLMPROC_UNLOCK_RES, argp, nlmsvc_proc_unlock);
 }
 
 static int
 nlmsvc_proc_granted_msg(struct svc_rqst *rqstp, struct nlm_args *argp,
                                                 void            *resp)
 {
-	struct nlm_res	res;
-	u32		stat;
-
 	dprintk("lockd: GRANTED_MSG   called\n");
-	memset(&res, 0, sizeof(res));
-
-	if ((stat = nlmsvc_proc_granted(rqstp, argp, &res)) == 0)
-		stat = nlmsvc_callback(rqstp, NLMPROC_GRANTED_RES, &res);
-	return stat;
+	return nlmsvc_callback(rqstp, NLMPROC_GRANTED_RES, argp, nlmsvc_proc_granted);
 }
 
 /*
@@ -497,55 +501,6 @@ nlmsvc_proc_granted_res(struct svc_rqst *rqstp, struct nlm_res  *argp,
 }
 
 /*
- * This is the generic lockd callback for async RPC calls
- */
-static u32
-nlmsvc_callback(struct svc_rqst *rqstp, u32 proc, struct nlm_res *resp)
-{
-	struct nlm_host	*host;
-	struct nlm_rqst	*call;
-
-	if (!(call = nlmclnt_alloc_call()))
-		return rpc_system_err;
-
-	host = nlmclnt_lookup_host(&rqstp->rq_addr,
-				rqstp->rq_prot, rqstp->rq_vers);
-	if (!host) {
-		kfree(call);
-		return rpc_system_err;
-	}
-
-	call->a_flags = RPC_TASK_ASYNC;
-	call->a_host  = host;
-	memcpy(&call->a_args, resp, sizeof(*resp));
-
-	if (nlmsvc_async_call(call, proc, &nlmsvc_callback_ops) < 0)
-		goto error;
-
-	return rpc_success;
- error:
-	nlm_release_host(host);
-	kfree(call);
-	return rpc_system_err;
-}
-
-static void nlmsvc_callback_exit(struct rpc_task *task, void *data)
-{
-	struct nlm_rqst	*call = data;
-
-	if (task->tk_status < 0) {
-		dprintk("lockd: %4d callback failed (errno = %d)\n",
-					task->tk_pid, -task->tk_status);
-	}
-	nlm_release_host(call->a_host);
-	kfree(call);
-}
-
-static const struct rpc_call_ops nlmsvc_callback_ops = {
-	.rpc_call_done = nlmsvc_callback_exit,
-};
-
-/*
  * NLM Server procedures.
  */
 
diff --git a/fs/lockd/svcshare.c b/fs/lockd/svcshare.c
index 4943fb7836ce..27288c83da96 100644
--- a/fs/lockd/svcshare.c
+++ b/fs/lockd/svcshare.c
@@ -88,7 +88,7 @@ nlmsvc_unshare_file(struct nlm_host *host, struct nlm_file *file,
  * Traverse all shares for a given file (and host).
  * NLM_ACT_CHECK is handled by nlmsvc_inspect_file.
  */
-int
+void
 nlmsvc_traverse_shares(struct nlm_host *host, struct nlm_file *file, int action)
 {
 	struct nlm_share	*share, **shpp;
@@ -106,6 +106,4 @@ nlmsvc_traverse_shares(struct nlm_host *host, struct nlm_file *file, int action)
 		}
 		shpp = &share->s_next;
 	}
-
-	return 0;
 }
diff --git a/fs/lockd/svcsubs.c b/fs/lockd/svcsubs.c
index 62f4a385177f..a570e5c8a930 100644
--- a/fs/lockd/svcsubs.c
+++ b/fs/lockd/svcsubs.c
@@ -11,6 +11,7 @@
 #include <linux/string.h>
 #include <linux/time.h>
 #include <linux/in.h>
+#include <linux/mutex.h>
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/nfsd/nfsfh.h>
@@ -28,7 +29,7 @@
 #define FILE_HASH_BITS		5
 #define FILE_NRHASH		(1<<FILE_HASH_BITS)
 static struct nlm_file *	nlm_files[FILE_NRHASH];
-static DECLARE_MUTEX(nlm_file_sema);
+static DEFINE_MUTEX(nlm_file_mutex);
 
 #ifdef NFSD_DEBUG
 static inline void nlm_debug_print_fh(char *msg, struct nfs_fh *f)
@@ -91,7 +92,7 @@ nlm_lookup_file(struct svc_rqst *rqstp, struct nlm_file **result,
 	hash = file_hash(f);
 
 	/* Lock file table */
-	down(&nlm_file_sema);
+	mutex_lock(&nlm_file_mutex);
 
 	for (file = nlm_files[hash]; file; file = file->f_next)
 		if (!nfs_compare_fh(&file->f_handle, f))
@@ -130,7 +131,7 @@ found:
 	nfserr = 0;
 
 out_unlock:
-	up(&nlm_file_sema);
+	mutex_unlock(&nlm_file_mutex);
 	return nfserr;
 
 out_free:
@@ -182,7 +183,7 @@ nlm_traverse_locks(struct nlm_host *host, struct nlm_file *file, int action)
 again:
 	file->f_locks = 0;
 	for (fl = inode->i_flock; fl; fl = fl->fl_next) {
-		if (!(fl->fl_flags & FL_LOCKD))
+		if (fl->fl_lmops != &nlmsvc_lock_operations)
 			continue;
 
 		/* update current lock count */
@@ -224,9 +225,8 @@ nlm_inspect_file(struct nlm_host *host, struct nlm_file *file, int action)
 		if (file->f_count || file->f_blocks || file->f_shares)
 			return 1;
 	} else {
-		if (nlmsvc_traverse_blocks(host, file, action)
-		 || nlmsvc_traverse_shares(host, file, action))
-			return 1;
+		nlmsvc_traverse_blocks(host, file, action);
+		nlmsvc_traverse_shares(host, file, action);
 	}
 	return nlm_traverse_locks(host, file, action);
 }
@@ -240,14 +240,14 @@ nlm_traverse_files(struct nlm_host *host, int action)
 	struct nlm_file	*file, **fp;
 	int		i;
 
-	down(&nlm_file_sema);
+	mutex_lock(&nlm_file_mutex);
 	for (i = 0; i < FILE_NRHASH; i++) {
 		fp = nlm_files + i;
 		while ((file = *fp) != NULL) {
 			/* Traverse locks, blocks and shares of this file
 			 * and update file->f_locks count */
 			if (nlm_inspect_file(host, file, action)) {
-				up(&nlm_file_sema);
+				mutex_unlock(&nlm_file_mutex);
 				return 1;
 			}
 
@@ -262,7 +262,7 @@ nlm_traverse_files(struct nlm_host *host, int action)
 			}
 		}
 	}
-	up(&nlm_file_sema);
+	mutex_unlock(&nlm_file_mutex);
 	return 0;
 }
 
@@ -282,7 +282,7 @@ nlm_release_file(struct nlm_file *file)
 				file, file->f_count);
 
 	/* Lock file table */
-	down(&nlm_file_sema);
+	mutex_lock(&nlm_file_mutex);
 
 	/* If there are no more locks etc, delete the file */
 	if(--file->f_count == 0) {
@@ -290,7 +290,7 @@ nlm_release_file(struct nlm_file *file)
 			nlm_delete_file(file);
 	}
 
-	up(&nlm_file_sema);
+	mutex_unlock(&nlm_file_mutex);
 }
 
 /*
diff --git a/fs/lockd/xdr.c b/fs/lockd/xdr.c
index 200fbda2c6d1..f22a3764461a 100644
--- a/fs/lockd/xdr.c
+++ b/fs/lockd/xdr.c
@@ -131,10 +131,11 @@ nlm_decode_lock(u32 *p, struct nlm_lock *lock)
 	 || !(p = nlm_decode_fh(p, &lock->fh))
 	 || !(p = nlm_decode_oh(p, &lock->oh)))
 		return NULL;
+	lock->svid  = ntohl(*p++);
 
 	locks_init_lock(fl);
 	fl->fl_owner = current->files;
-	fl->fl_pid   = ntohl(*p++);
+	fl->fl_pid   = (pid_t)lock->svid;
 	fl->fl_flags = FL_POSIX;
 	fl->fl_type  = F_RDLCK;		/* as good as anything else */
 	start = ntohl(*p++);
@@ -174,7 +175,7 @@ nlm_encode_lock(u32 *p, struct nlm_lock *lock)
 	else
 		len = loff_t_to_s32(fl->fl_end - fl->fl_start + 1);
 
-	*p++ = htonl(fl->fl_pid);
+	*p++ = htonl(lock->svid);
 	*p++ = htonl(start);
 	*p++ = htonl(len);
 
@@ -197,7 +198,7 @@ nlm_encode_testres(u32 *p, struct nlm_res *resp)
 		struct file_lock	*fl = &resp->lock.fl;
 
 		*p++ = (fl->fl_type == F_RDLCK)? xdr_zero : xdr_one;
-		*p++ = htonl(fl->fl_pid);
+		*p++ = htonl(resp->lock.svid);
 
 		/* Encode owner handle. */
 		if (!(p = xdr_encode_netobj(p, &resp->lock.oh)))
@@ -298,7 +299,8 @@ nlmsvc_decode_shareargs(struct svc_rqst *rqstp, u32 *p, nlm_args *argp)
 
 	memset(lock, 0, sizeof(*lock));
 	locks_init_lock(&lock->fl);
-	lock->fl.fl_pid = ~(u32) 0;
+	lock->svid = ~(u32) 0;
+	lock->fl.fl_pid = (pid_t)lock->svid;
 
 	if (!(p = nlm_decode_cookie(p, &argp->cookie))
 	 || !(p = xdr_decode_string_inplace(p, &lock->caller,
@@ -415,7 +417,8 @@ nlmclt_decode_testres(struct rpc_rqst *req, u32 *p, struct nlm_res *resp)
 		memset(&resp->lock, 0, sizeof(resp->lock));
 		locks_init_lock(fl);
 		excl = ntohl(*p++);
-		fl->fl_pid = ntohl(*p++);
+		resp->lock.svid = ntohl(*p++);
+		fl->fl_pid = (pid_t)resp->lock.svid;
 		if (!(p = nlm_decode_oh(p, &resp->lock.oh)))
 			return -EIO;
 
@@ -543,7 +546,9 @@ nlmclt_decode_res(struct rpc_rqst *req, u32 *p, struct nlm_res *resp)
 	.p_proc      = NLMPROC_##proc,					\
 	.p_encode    = (kxdrproc_t) nlmclt_encode_##argtype,		\
 	.p_decode    = (kxdrproc_t) nlmclt_decode_##restype,		\
-	.p_bufsiz    = MAX(NLM_##argtype##_sz, NLM_##restype##_sz) << 2	\
+	.p_bufsiz    = MAX(NLM_##argtype##_sz, NLM_##restype##_sz) << 2,	\
+	.p_statidx   = NLMPROC_##proc,					\
+	.p_name      = #proc,						\
 	}
 
 static struct rpc_procinfo	nlm_procedures[] = {
@@ -599,7 +604,7 @@ static struct rpc_stat		nlm_stats;
 struct rpc_program		nlm_program = {
 		.name		= "lockd",
 		.number		= NLM_PROGRAM,
-		.nrvers		= sizeof(nlm_versions) / sizeof(nlm_versions[0]),
+		.nrvers		= ARRAY_SIZE(nlm_versions),
 		.version	= nlm_versions,
 		.stats		= &nlm_stats,
 };
diff --git a/fs/lockd/xdr4.c b/fs/lockd/xdr4.c
index fdcf105a5303..36eb175ec335 100644
--- a/fs/lockd/xdr4.c
+++ b/fs/lockd/xdr4.c
@@ -130,10 +130,11 @@ nlm4_decode_lock(u32 *p, struct nlm_lock *lock)
 	 || !(p = nlm4_decode_fh(p, &lock->fh))
 	 || !(p = nlm4_decode_oh(p, &lock->oh)))
 		return NULL;
+	lock->svid  = ntohl(*p++);
 
 	locks_init_lock(fl);
 	fl->fl_owner = current->files;
-	fl->fl_pid   = ntohl(*p++);
+	fl->fl_pid   = (pid_t)lock->svid;
 	fl->fl_flags = FL_POSIX;
 	fl->fl_type  = F_RDLCK;		/* as good as anything else */
 	p = xdr_decode_hyper(p, &start);
@@ -167,7 +168,7 @@ nlm4_encode_lock(u32 *p, struct nlm_lock *lock)
 	 || (fl->fl_end > NLM4_OFFSET_MAX && fl->fl_end != OFFSET_MAX))
 		return NULL;
 
-	*p++ = htonl(fl->fl_pid);
+	*p++ = htonl(lock->svid);
 
 	start = loff_t_to_s64(fl->fl_start);
 	if (fl->fl_end == OFFSET_MAX)
@@ -198,7 +199,7 @@ nlm4_encode_testres(u32 *p, struct nlm_res *resp)
 		struct file_lock	*fl = &resp->lock.fl;
 
 		*p++ = (fl->fl_type == F_RDLCK)? xdr_zero : xdr_one;
-		*p++ = htonl(fl->fl_pid);
+		*p++ = htonl(resp->lock.svid);
 
 		/* Encode owner handle. */
 		if (!(p = xdr_encode_netobj(p, &resp->lock.oh)))
@@ -212,8 +213,8 @@ nlm4_encode_testres(u32 *p, struct nlm_res *resp)
 		
 		p = xdr_encode_hyper(p, start);
 		p = xdr_encode_hyper(p, len);
-		dprintk("xdr: encode_testres (status %d pid %d type %d start %Ld end %Ld)\n",
-			resp->status, fl->fl_pid, fl->fl_type,
+		dprintk("xdr: encode_testres (status %u pid %d type %d start %Ld end %Ld)\n",
+			resp->status, (int)resp->lock.svid, fl->fl_type,
 			(long long)fl->fl_start,  (long long)fl->fl_end);
 	}
 
@@ -303,7 +304,8 @@ nlm4svc_decode_shareargs(struct svc_rqst *rqstp, u32 *p, nlm_args *argp)
 
 	memset(lock, 0, sizeof(*lock));
 	locks_init_lock(&lock->fl);
-	lock->fl.fl_pid = ~(u32) 0;
+	lock->svid = ~(u32) 0;
+	lock->fl.fl_pid = (pid_t)lock->svid;
 
 	if (!(p = nlm4_decode_cookie(p, &argp->cookie))
 	 || !(p = xdr_decode_string_inplace(p, &lock->caller,
@@ -420,7 +422,8 @@ nlm4clt_decode_testres(struct rpc_rqst *req, u32 *p, struct nlm_res *resp)
 		memset(&resp->lock, 0, sizeof(resp->lock));
 		locks_init_lock(fl);
 		excl = ntohl(*p++);
-		fl->fl_pid = ntohl(*p++);
+		resp->lock.svid = ntohl(*p++);
+		fl->fl_pid = (pid_t)resp->lock.svid;
 		if (!(p = nlm4_decode_oh(p, &resp->lock.oh)))
 			return -EIO;
 
@@ -548,7 +551,9 @@ nlm4clt_decode_res(struct rpc_rqst *req, u32 *p, struct nlm_res *resp)
 	.p_proc      = NLMPROC_##proc,					\
 	.p_encode    = (kxdrproc_t) nlm4clt_encode_##argtype,		\
 	.p_decode    = (kxdrproc_t) nlm4clt_decode_##restype,		\
-	.p_bufsiz    = MAX(NLM4_##argtype##_sz, NLM4_##restype##_sz) << 2	\
+	.p_bufsiz    = MAX(NLM4_##argtype##_sz, NLM4_##restype##_sz) << 2,	\
+	.p_statidx   = NLMPROC_##proc,					\
+	.p_name      = #proc,						\
 	}
 
 static struct rpc_procinfo	nlm4_procedures[] = {
diff --git a/fs/locks.c b/fs/locks.c
index 909eab8fb1d0..4d9e71d43e7e 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -139,13 +139,10 @@ int lease_break_time = 45;
 #define for_each_lock(inode, lockp) \
 	for (lockp = &inode->i_flock; *lockp != NULL; lockp = &(*lockp)->fl_next)
 
-LIST_HEAD(file_lock_list);
-
-EXPORT_SYMBOL(file_lock_list);
-
+static LIST_HEAD(file_lock_list);
 static LIST_HEAD(blocked_list);
 
-static kmem_cache_t *filelock_cache;
+static kmem_cache_t *filelock_cache __read_mostly;
 
 /* Allocate an empty lock structure. */
 static struct file_lock *locks_alloc_lock(void)
@@ -153,6 +150,21 @@ static struct file_lock *locks_alloc_lock(void)
 	return kmem_cache_alloc(filelock_cache, SLAB_KERNEL);
 }
 
+static void locks_release_private(struct file_lock *fl)
+{
+	if (fl->fl_ops) {
+		if (fl->fl_ops->fl_release_private)
+			fl->fl_ops->fl_release_private(fl);
+		fl->fl_ops = NULL;
+	}
+	if (fl->fl_lmops) {
+		if (fl->fl_lmops->fl_release_private)
+			fl->fl_lmops->fl_release_private(fl);
+		fl->fl_lmops = NULL;
+	}
+
+}
+
 /* Free a lock which is not in use. */
 static void locks_free_lock(struct file_lock *fl)
 {
@@ -169,18 +181,7 @@ static void locks_free_lock(struct file_lock *fl)
 	if (!list_empty(&fl->fl_link))
 		panic("Attempting to free lock on active lock list");
 
-	if (fl->fl_ops) {
-		if (fl->fl_ops->fl_release_private)
-			fl->fl_ops->fl_release_private(fl);
-		fl->fl_ops = NULL;
-	}
-
-	if (fl->fl_lmops) {
-		if (fl->fl_lmops->fl_release_private)
-			fl->fl_lmops->fl_release_private(fl);
-		fl->fl_lmops = NULL;
-	}
-
+	locks_release_private(fl);
 	kmem_cache_free(filelock_cache, fl);
 }
 
@@ -218,24 +219,46 @@ static void init_once(void *foo, kmem_cache_t *cache, unsigned long flags)
 	locks_init_lock(lock);
 }
 
+static void locks_copy_private(struct file_lock *new, struct file_lock *fl)
+{
+	if (fl->fl_ops) {
+		if (fl->fl_ops->fl_copy_lock)
+			fl->fl_ops->fl_copy_lock(new, fl);
+		new->fl_ops = fl->fl_ops;
+	}
+	if (fl->fl_lmops) {
+		if (fl->fl_lmops->fl_copy_lock)
+			fl->fl_lmops->fl_copy_lock(new, fl);
+		new->fl_lmops = fl->fl_lmops;
+	}
+}
+
 /*
  * Initialize a new lock from an existing file_lock structure.
  */
-void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
+static void __locks_copy_lock(struct file_lock *new, const struct file_lock *fl)
 {
 	new->fl_owner = fl->fl_owner;
 	new->fl_pid = fl->fl_pid;
-	new->fl_file = fl->fl_file;
+	new->fl_file = NULL;
 	new->fl_flags = fl->fl_flags;
 	new->fl_type = fl->fl_type;
 	new->fl_start = fl->fl_start;
 	new->fl_end = fl->fl_end;
+	new->fl_ops = NULL;
+	new->fl_lmops = NULL;
+}
+
+void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
+{
+	locks_release_private(new);
+
+	__locks_copy_lock(new, fl);
+	new->fl_file = fl->fl_file;
 	new->fl_ops = fl->fl_ops;
 	new->fl_lmops = fl->fl_lmops;
-	if (fl->fl_ops && fl->fl_ops->fl_copy_lock)
-		fl->fl_ops->fl_copy_lock(new, fl);
-	if (fl->fl_lmops && fl->fl_lmops->fl_copy_lock)
-		fl->fl_lmops->fl_copy_lock(new, fl);
+
+	locks_copy_private(new, fl);
 }
 
 EXPORT_SYMBOL(locks_copy_lock);
@@ -510,12 +533,7 @@ static void locks_delete_block(struct file_lock *waiter)
 static void locks_insert_block(struct file_lock *blocker, 
 			       struct file_lock *waiter)
 {
-	if (!list_empty(&waiter->fl_block)) {
-		printk(KERN_ERR "locks_insert_block: removing duplicated lock "
-			"(pid=%d %Ld-%Ld type=%d)\n", waiter->fl_pid,
-			waiter->fl_start, waiter->fl_end, waiter->fl_type);
-		__locks_delete_block(waiter);
-	}
+	BUG_ON(!list_empty(&waiter->fl_block));
 	list_add_tail(&waiter->fl_block, &blocker->fl_block);
 	waiter->fl_next = blocker;
 	if (IS_POSIX(blocker))
@@ -654,8 +672,9 @@ static int locks_block_on_timeout(struct file_lock *blocker, struct file_lock *w
 	return result;
 }
 
-struct file_lock *
-posix_test_lock(struct file *filp, struct file_lock *fl)
+int
+posix_test_lock(struct file *filp, struct file_lock *fl,
+		struct file_lock *conflock)
 {
 	struct file_lock *cfl;
 
@@ -666,9 +685,13 @@ posix_test_lock(struct file *filp, struct file_lock *fl)
 		if (posix_locks_conflict(cfl, fl))
 			break;
 	}
+	if (cfl) {
+		__locks_copy_lock(conflock, cfl);
+		unlock_kernel();
+		return 1;
+	}
 	unlock_kernel();
-
-	return (cfl);
+	return 0;
 }
 
 EXPORT_SYMBOL(posix_test_lock);
@@ -769,9 +792,7 @@ out:
 	return error;
 }
 
-EXPORT_SYMBOL(posix_lock_file);
-
-static int __posix_lock_file(struct inode *inode, struct file_lock *request)
+static int __posix_lock_file_conf(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
 {
 	struct file_lock *fl;
 	struct file_lock *new_fl, *new_fl2;
@@ -795,6 +816,8 @@ static int __posix_lock_file(struct inode *inode, struct file_lock *request)
 				continue;
 			if (!posix_locks_conflict(request, fl))
 				continue;
+			if (conflock)
+				locks_copy_lock(conflock, fl);
 			error = -EAGAIN;
 			if (!(request->fl_flags & FL_SLEEP))
 				goto out;
@@ -904,7 +927,8 @@ static int __posix_lock_file(struct inode *inode, struct file_lock *request)
 				fl->fl_start = request->fl_start;
 				fl->fl_end = request->fl_end;
 				fl->fl_type = request->fl_type;
-				fl->fl_u = request->fl_u;
+				locks_release_private(fl);
+				locks_copy_private(fl, request);
 				request = fl;
 				added = 1;
 			}
@@ -963,8 +987,24 @@ static int __posix_lock_file(struct inode *inode, struct file_lock *request)
  */
 int posix_lock_file(struct file *filp, struct file_lock *fl)
 {
-	return __posix_lock_file(filp->f_dentry->d_inode, fl);
+	return __posix_lock_file_conf(filp->f_dentry->d_inode, fl, NULL);
 }
+EXPORT_SYMBOL(posix_lock_file);
+
+/**
+ * posix_lock_file_conf - Apply a POSIX-style lock to a file
+ * @filp: The file to apply the lock to
+ * @fl: The lock to be applied
+ * @conflock: Place to return a copy of the conflicting lock, if found.
+ *
+ * Except for the conflock parameter, acts just like posix_lock_file.
+ */
+int posix_lock_file_conf(struct file *filp, struct file_lock *fl,
+			struct file_lock *conflock)
+{
+	return __posix_lock_file_conf(filp->f_dentry->d_inode, fl, conflock);
+}
+EXPORT_SYMBOL(posix_lock_file_conf);
 
 /**
  * posix_lock_file_wait - Apply a POSIX-style lock to a file
@@ -980,7 +1020,7 @@ int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
 	int error;
 	might_sleep ();
 	for (;;) {
-		error = __posix_lock_file(filp->f_dentry->d_inode, fl);
+		error = posix_lock_file(filp, fl);
 		if ((error != -EAGAIN) || !(fl->fl_flags & FL_SLEEP))
 			break;
 		error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
@@ -1052,7 +1092,7 @@ int locks_mandatory_area(int read_write, struct inode *inode,
 	fl.fl_end = offset + count - 1;
 
 	for (;;) {
-		error = __posix_lock_file(inode, &fl);
+		error = __posix_lock_file_conf(inode, &fl, NULL);
 		if (error != -EAGAIN)
 			break;
 		if (!(fl.fl_flags & FL_SLEEP))
@@ -1544,7 +1584,7 @@ asmlinkage long sys_flock(unsigned int fd, unsigned int cmd)
  */
 int fcntl_getlk(struct file *filp, struct flock __user *l)
 {
-	struct file_lock *fl, file_lock;
+	struct file_lock *fl, cfl, file_lock;
 	struct flock flock;
 	int error;
 
@@ -1568,7 +1608,7 @@ int fcntl_getlk(struct file *filp, struct flock __user *l)
 		else
 		  fl = (file_lock.fl_type == F_UNLCK ? NULL : &file_lock);
 	} else {
-		fl = posix_test_lock(filp, &file_lock);
+		fl = (posix_test_lock(filp, &file_lock, &cfl) ? &cfl : NULL);
 	}
  
 	flock.l_type = F_UNLCK;
@@ -1665,7 +1705,7 @@ again:
 		error = filp->f_op->lock(filp, cmd, file_lock);
 	else {
 		for (;;) {
-			error = __posix_lock_file(inode, file_lock);
+			error = posix_lock_file(filp, file_lock);
 			if ((error != -EAGAIN) || (cmd == F_SETLK))
 				break;
 			error = wait_event_interruptible(file_lock->fl_wait,
@@ -1698,7 +1738,7 @@ out:
  */
 int fcntl_getlk64(struct file *filp, struct flock64 __user *l)
 {
-	struct file_lock *fl, file_lock;
+	struct file_lock *fl, cfl, file_lock;
 	struct flock64 flock;
 	int error;
 
@@ -1722,7 +1762,7 @@ int fcntl_getlk64(struct file *filp, struct flock64 __user *l)
 		else
 		  fl = (file_lock.fl_type == F_UNLCK ? NULL : &file_lock);
 	} else {
-		fl = posix_test_lock(filp, &file_lock);
+		fl = (posix_test_lock(filp, &file_lock, &cfl) ? &cfl : NULL);
 	}
  
 	flock.l_type = F_UNLCK;
@@ -1808,7 +1848,7 @@ again:
 		error = filp->f_op->lock(filp, cmd, file_lock);
 	else {
 		for (;;) {
-			error = __posix_lock_file(inode, file_lock);
+			error = posix_lock_file(filp, file_lock);
 			if ((error != -EAGAIN) || (cmd == F_SETLK64))
 				break;
 			error = wait_event_interruptible(file_lock->fl_wait,
@@ -1936,21 +1976,6 @@ void locks_remove_flock(struct file *filp)
 }
 
 /**
- *	posix_block_lock - blocks waiting for a file lock
- *	@blocker: the lock which is blocking
- *	@waiter: the lock which conflicts and has to wait
- *
- * lockd needs to block waiting for locks.
- */
-void
-posix_block_lock(struct file_lock *blocker, struct file_lock *waiter)
-{
-	locks_insert_block(blocker, waiter);
-}
-
-EXPORT_SYMBOL(posix_block_lock);
-
-/**
  *	posix_unblock_lock - stop waiting for a file lock
  *      @filp:   how the file was opened
  *	@waiter: the lock which was waiting
diff --git a/fs/mbcache.c b/fs/mbcache.c
index f5bbe4c97c58..e4fde1ab22cd 100644
--- a/fs/mbcache.c
+++ b/fs/mbcache.c
@@ -288,7 +288,7 @@ mb_cache_create(const char *name, struct mb_cache_op *cache_op,
 			INIT_LIST_HEAD(&cache->c_indexes_hash[m][n]);
 	}
 	cache->c_entry_cache = kmem_cache_create(name, entry_size, 0,
-		SLAB_RECLAIM_ACCOUNT, NULL, NULL);
+		SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL, NULL);
 	if (!cache->c_entry_cache)
 		goto fail;
 
@@ -311,7 +311,7 @@ fail:
 /*
  * mb_cache_shrink()
  *
- * Removes all cache entires of a device from the cache. All cache entries
+ * Removes all cache entries of a device from the cache. All cache entries
  * currently in use cannot be freed, and thus remain in the cache. All others
  * are freed.
  *
diff --git a/fs/minix/bitmap.c b/fs/minix/bitmap.c
index dc6a4e4abcdc..4a6abc49418e 100644
--- a/fs/minix/bitmap.c
+++ b/fs/minix/bitmap.c
@@ -56,7 +56,7 @@ void minix_free_block(struct inode * inode, int block)
 	unsigned int bit,zone;
 
 	if (block < sbi->s_firstdatazone || block >= sbi->s_nzones) {
-		printk("trying to free block not in datazone\n");
+		printk("Trying to free block not in datazone\n");
 		return;
 	}
 	zone = block - sbi->s_firstdatazone + 1;
@@ -124,7 +124,7 @@ minix_V1_raw_inode(struct super_block *sb, ino_t ino, struct buffer_head **bh)
 		 ino / MINIX_INODES_PER_BLOCK;
 	*bh = sb_bread(sb, block);
 	if (!*bh) {
-		printk("unable to read i-node block\n");
+		printk("Unable to read inode block\n");
 		return NULL;
 	}
 	p = (void *)(*bh)->b_data;
@@ -149,7 +149,7 @@ minix_V2_raw_inode(struct super_block *sb, ino_t ino, struct buffer_head **bh)
 		 ino / MINIX2_INODES_PER_BLOCK;
 	*bh = sb_bread(sb, block);
 	if (!*bh) {
-		printk("unable to read i-node block\n");
+		printk("Unable to read inode block\n");
 		return NULL;
 	}
 	p = (void *)(*bh)->b_data;
@@ -204,7 +204,7 @@ void minix_free_inode(struct inode * inode)
 	bh = sbi->s_imap[ino >> 13];
 	lock_kernel();
 	if (!minix_test_and_clear_bit(ino & 8191, bh->b_data))
-		printk("minix_free_inode: bit %lu already cleared.\n", ino);
+		printk("minix_free_inode: bit %lu already cleared\n", ino);
 	unlock_kernel();
 	mark_buffer_dirty(bh);
  out:
@@ -238,7 +238,7 @@ struct inode * minix_new_inode(const struct inode * dir, int * error)
 		return NULL;
 	}
 	if (minix_test_and_set_bit(j,bh->b_data)) {	/* shouldn't happen */
-		printk("new_inode: bit already set");
+		printk("new_inode: bit already set\n");
 		unlock_kernel();
 		iput(inode);
 		return NULL;
diff --git a/fs/minix/dir.c b/fs/minix/dir.c
index 732502aabc05..69224d1fe043 100644
--- a/fs/minix/dir.c
+++ b/fs/minix/dir.c
@@ -14,7 +14,7 @@ typedef struct minix_dir_entry minix_dirent;
 
 static int minix_readdir(struct file *, void *, filldir_t);
 
-struct file_operations minix_dir_operations = {
+const struct file_operations minix_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= minix_readdir,
 	.fsync		= minix_sync_file,
diff --git a/fs/minix/file.c b/fs/minix/file.c
index f1d77acb3f01..420b32882a10 100644
--- a/fs/minix/file.c
+++ b/fs/minix/file.c
@@ -15,7 +15,7 @@
  */
 int minix_sync_file(struct file *, struct dentry *, int);
 
-struct file_operations minix_file_operations = {
+const struct file_operations minix_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/minix/inode.c b/fs/minix/inode.c
index 790cc0d0e970..2dcccf1d1b7f 100644
--- a/fs/minix/inode.c
+++ b/fs/minix/inode.c
@@ -80,7 +80,8 @@ static int init_inodecache(void)
 {
 	minix_inode_cachep = kmem_cache_create("minix_inode_cache",
 					     sizeof(struct minix_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (minix_inode_cachep == NULL)
 		return -ENOMEM;
@@ -126,11 +127,11 @@ static int minix_remount (struct super_block * sb, int * flags, char * data)
 		mark_buffer_dirty(sbi->s_sbh);
 
 		if (!(sbi->s_mount_state & MINIX_VALID_FS))
-			printk ("MINIX-fs warning: remounting unchecked fs, "
-				"running fsck is recommended.\n");
+			printk("MINIX-fs warning: remounting unchecked fs, "
+				"running fsck is recommended\n");
 		else if ((sbi->s_mount_state & MINIX_ERROR_FS))
-			printk ("MINIX-fs warning: remounting fs with errors, "
-				"running fsck is recommended.\n");
+			printk("MINIX-fs warning: remounting fs with errors, "
+				"running fsck is recommended\n");
 	}
 	return 0;
 }
@@ -244,11 +245,11 @@ static int minix_fill_super(struct super_block *s, void *data, int silent)
 		mark_buffer_dirty(bh);
 	}
 	if (!(sbi->s_mount_state & MINIX_VALID_FS))
-		printk ("MINIX-fs: mounting unchecked file system, "
-			"running fsck is recommended.\n");
+		printk("MINIX-fs: mounting unchecked file system, "
+			"running fsck is recommended\n");
  	else if (sbi->s_mount_state & MINIX_ERROR_FS)
-		printk ("MINIX-fs: mounting file system with errors, "
-			"running fsck is recommended.\n");
+		printk("MINIX-fs: mounting file system with errors, "
+			"running fsck is recommended\n");
 	return 0;
 
 out_iput:
@@ -272,19 +273,19 @@ out_no_bitmap:
 
 out_no_map:
 	if (!silent)
-		printk ("MINIX-fs: can't allocate map\n");
+		printk("MINIX-fs: can't allocate map\n");
 	goto out_release;
 
 out_no_fs:
 	if (!silent)
-		printk("VFS: Can't find a Minix or Minix V2 filesystem on device "
-		       "%s.\n", s->s_id);
+		printk("VFS: Can't find a Minix or Minix V2 filesystem "
+			"on device %s\n", s->s_id);
     out_release:
 	brelse(bh);
 	goto out;
 
 out_bad_hblock:
-	printk("MINIX-fs: blocksize too small for device.\n");
+	printk("MINIX-fs: blocksize too small for device\n");
 	goto out;
 
 out_bad_sb:
@@ -523,7 +524,7 @@ int minix_sync_inode(struct inode * inode)
 		sync_dirty_buffer(bh);
 		if (buffer_req(bh) && !buffer_uptodate(bh))
 		{
-			printk ("IO error syncing minix inode [%s:%08lx]\n",
+			printk("IO error syncing minix inode [%s:%08lx]\n",
 				inode->i_sb->s_id, inode->i_ino);
 			err = -1;
 		}
diff --git a/fs/minix/itree_v1.c b/fs/minix/itree_v1.c
index ba06aef4aca1..656b1347a25b 100644
--- a/fs/minix/itree_v1.c
+++ b/fs/minix/itree_v1.c
@@ -25,9 +25,9 @@ static int block_to_path(struct inode * inode, long block, int offsets[DEPTH])
 	int n = 0;
 
 	if (block < 0) {
-		printk("minix_bmap: block<0");
+		printk("minix_bmap: block<0\n");
 	} else if (block >= (minix_sb(inode->i_sb)->s_max_size/BLOCK_SIZE)) {
-		printk("minix_bmap: block>big");
+		printk("minix_bmap: block>big\n");
 	} else if (block < 7) {
 		offsets[n++] = block;
 	} else if ((block -= 7) < 512) {
diff --git a/fs/minix/itree_v2.c b/fs/minix/itree_v2.c
index 3adc7675560f..9adcdc754e0f 100644
--- a/fs/minix/itree_v2.c
+++ b/fs/minix/itree_v2.c
@@ -25,9 +25,9 @@ static int block_to_path(struct inode * inode, long block, int offsets[DEPTH])
 	int n = 0;
 
 	if (block < 0) {
-		printk("minix_bmap: block<0");
+		printk("minix_bmap: block<0\n");
 	} else if (block >= (minix_sb(inode->i_sb)->s_max_size/BLOCK_SIZE)) {
-		printk("minix_bmap: block>big");
+		printk("minix_bmap: block>big\n");
 	} else if (block < 7) {
 		offsets[n++] = block;
 	} else if ((block -= 7) < 256) {
diff --git a/fs/minix/minix.h b/fs/minix/minix.h
index e42a8bb89001..c55b77cdcc8e 100644
--- a/fs/minix/minix.h
+++ b/fs/minix/minix.h
@@ -81,8 +81,8 @@ extern int minix_sync_file(struct file *, struct dentry *, int);
 
 extern struct inode_operations minix_file_inode_operations;
 extern struct inode_operations minix_dir_inode_operations;
-extern struct file_operations minix_file_operations;
-extern struct file_operations minix_dir_operations;
+extern const struct file_operations minix_file_operations;
+extern const struct file_operations minix_dir_operations;
 extern struct dentry_operations minix_dentry_operations;
 
 static inline struct minix_sb_info *minix_sb(struct super_block *sb)
diff --git a/fs/minix/namei.c b/fs/minix/namei.c
index b25bca5bdb57..5b6a4540a05b 100644
--- a/fs/minix/namei.c
+++ b/fs/minix/namei.c
@@ -6,18 +6,6 @@
 
 #include "minix.h"
 
-static inline void inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static int add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = minix_add_link(dentry, inode);
@@ -25,7 +13,7 @@ static int add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -125,7 +113,7 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	goto out;
 }
@@ -139,7 +127,7 @@ static int minix_link(struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 	return add_nondir(dentry, inode);
 }
@@ -152,7 +140,7 @@ static int minix_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 	if (dir->i_nlink >= minix_sb(dir->i_sb)->s_link_max)
 		goto out;
 
-	inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = minix_new_inode(dir, &err);
 	if (!inode)
@@ -163,7 +151,7 @@ static int minix_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 		inode->i_mode |= S_ISGID;
 	minix_set_inode(inode, 0);
 
-	inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = minix_make_empty(inode, dir);
 	if (err)
@@ -178,11 +166,11 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
-	dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 out_dir:
-	dec_count(dir);
+	inode_dec_link_count(dir);
 	goto out;
 }
 
@@ -202,7 +190,7 @@ static int minix_unlink(struct inode * dir, struct dentry *dentry)
 		goto end_unlink;
 
 	inode->i_ctime = dir->i_ctime;
-	dec_count(inode);
+	inode_dec_link_count(inode);
 end_unlink:
 	return err;
 }
@@ -215,8 +203,8 @@ static int minix_rmdir(struct inode * dir, struct dentry *dentry)
 	if (minix_empty_dir(inode)) {
 		err = minix_unlink(dir, dentry);
 		if (!err) {
-			dec_count(dir);
-			dec_count(inode);
+			inode_dec_link_count(dir);
+			inode_dec_link_count(inode);
 		}
 	}
 	return err;
@@ -257,34 +245,34 @@ static int minix_rename(struct inode * old_dir, struct dentry *old_dentry,
 		new_de = minix_find_entry(new_dentry, &new_page);
 		if (!new_de)
 			goto out_dir;
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		minix_set_link(new_de, new_page, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= info->s_link_max)
 				goto out_dir;
 		}
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = minix_add_link(new_dentry, old_inode);
 		if (err) {
-			dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	minix_delete_entry(old_de, old_page);
-	dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		minix_set_link(dir_de, dir_page, new_dir);
-		dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	return 0;
 
diff --git a/fs/mpage.c b/fs/mpage.c
index e431cb3878d6..9bf2eb30e6f4 100644
--- a/fs/mpage.c
+++ b/fs/mpage.c
@@ -163,9 +163,19 @@ map_buffer_to_page(struct page *page, struct buffer_head *bh, int page_block)
 	} while (page_bh != head);
 }
 
+/*
+ * This is the worker routine which does all the work of mapping the disk
+ * blocks and constructs largest possible bios, submits them for IO if the
+ * blocks are not contiguous on the disk.
+ *
+ * We pass a buffer_head back and forth and use its buffer_mapped() flag to
+ * represent the validity of its disk mapping and to decide when to do the next
+ * get_block() call.
+ */
 static struct bio *
 do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
-			sector_t *last_block_in_bio, get_block_t get_block)
+		sector_t *last_block_in_bio, struct buffer_head *map_bh,
+		unsigned long *first_logical_block, get_block_t get_block)
 {
 	struct inode *inode = page->mapping->host;
 	const unsigned blkbits = inode->i_blkbits;
@@ -173,33 +183,72 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 	const unsigned blocksize = 1 << blkbits;
 	sector_t block_in_file;
 	sector_t last_block;
+	sector_t last_block_in_file;
 	sector_t blocks[MAX_BUF_PER_PAGE];
 	unsigned page_block;
 	unsigned first_hole = blocks_per_page;
 	struct block_device *bdev = NULL;
-	struct buffer_head bh;
 	int length;
 	int fully_mapped = 1;
+	unsigned nblocks;
+	unsigned relative_block;
 
 	if (page_has_buffers(page))
 		goto confused;
 
 	block_in_file = (sector_t)page->index << (PAGE_CACHE_SHIFT - blkbits);
-	last_block = (i_size_read(inode) + blocksize - 1) >> blkbits;
+	last_block = block_in_file + nr_pages * blocks_per_page;
+	last_block_in_file = (i_size_read(inode) + blocksize - 1) >> blkbits;
+	if (last_block > last_block_in_file)
+		last_block = last_block_in_file;
+	page_block = 0;
+
+	/*
+	 * Map blocks using the result from the previous get_blocks call first.
+	 */
+	nblocks = map_bh->b_size >> blkbits;
+	if (buffer_mapped(map_bh) && block_in_file > *first_logical_block &&
+			block_in_file < (*first_logical_block + nblocks)) {
+		unsigned map_offset = block_in_file - *first_logical_block;
+		unsigned last = nblocks - map_offset;
+
+		for (relative_block = 0; ; relative_block++) {
+			if (relative_block == last) {
+				clear_buffer_mapped(map_bh);
+				break;
+			}
+			if (page_block == blocks_per_page)
+				break;
+			blocks[page_block] = map_bh->b_blocknr + map_offset +
+						relative_block;
+			page_block++;
+			block_in_file++;
+		}
+		bdev = map_bh->b_bdev;
+	}
+
+	/*
+	 * Then do more get_blocks calls until we are done with this page.
+	 */
+	map_bh->b_page = page;
+	while (page_block < blocks_per_page) {
+		map_bh->b_state = 0;
+		map_bh->b_size = 0;
 
-	bh.b_page = page;
-	for (page_block = 0; page_block < blocks_per_page;
-				page_block++, block_in_file++) {
-		bh.b_state = 0;
 		if (block_in_file < last_block) {
-			if (get_block(inode, block_in_file, &bh, 0))
+			map_bh->b_size = (last_block-block_in_file) << blkbits;
+			if (get_block(inode, block_in_file, map_bh, 0))
 				goto confused;
+			*first_logical_block = block_in_file;
 		}
 
-		if (!buffer_mapped(&bh)) {
+		if (!buffer_mapped(map_bh)) {
 			fully_mapped = 0;
 			if (first_hole == blocks_per_page)
 				first_hole = page_block;
+			page_block++;
+			block_in_file++;
+			clear_buffer_mapped(map_bh);
 			continue;
 		}
 
@@ -209,8 +258,8 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 		 * we just collected from get_block into the page's buffers
 		 * so readpage doesn't have to repeat the get_block call
 		 */
-		if (buffer_uptodate(&bh)) {
-			map_buffer_to_page(page, &bh, page_block);
+		if (buffer_uptodate(map_bh)) {
+			map_buffer_to_page(page, map_bh, page_block);
 			goto confused;
 		}
 	
@@ -218,10 +267,20 @@ do_mpage_readpage(struct bio *bio, struct page *page, unsigned nr_pages,
 			goto confused;		/* hole -> non-hole */
 
 		/* Contiguous blocks? */
-		if (page_block && blocks[page_block-1] != bh.b_blocknr-1)
+		if (page_block && blocks[page_block-1] != map_bh->b_blocknr-1)
 			goto confused;
-		blocks[page_block] = bh.b_blocknr;
-		bdev = bh.b_bdev;
+		nblocks = map_bh->b_size >> blkbits;
+		for (relative_block = 0; ; relative_block++) {
+			if (relative_block == nblocks) {
+				clear_buffer_mapped(map_bh);
+				break;
+			} else if (page_block == blocks_per_page)
+				break;
+			blocks[page_block] = map_bh->b_blocknr+relative_block;
+			page_block++;
+			block_in_file++;
+		}
+		bdev = map_bh->b_bdev;
 	}
 
 	if (first_hole != blocks_per_page) {
@@ -260,7 +319,7 @@ alloc_new:
 		goto alloc_new;
 	}
 
-	if (buffer_boundary(&bh) || (first_hole != blocks_per_page))
+	if (buffer_boundary(map_bh) || (first_hole != blocks_per_page))
 		bio = mpage_bio_submit(READ, bio);
 	else
 		*last_block_in_bio = blocks[blocks_per_page - 1];
@@ -331,7 +390,10 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 	unsigned page_idx;
 	sector_t last_block_in_bio = 0;
 	struct pagevec lru_pvec;
+	struct buffer_head map_bh;
+	unsigned long first_logical_block = 0;
 
+	clear_buffer_mapped(&map_bh);
 	pagevec_init(&lru_pvec, 0);
 	for (page_idx = 0; page_idx < nr_pages; page_idx++) {
 		struct page *page = list_entry(pages->prev, struct page, lru);
@@ -342,7 +404,9 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
 					page->index, GFP_KERNEL)) {
 			bio = do_mpage_readpage(bio, page,
 					nr_pages - page_idx,
-					&last_block_in_bio, get_block);
+					&last_block_in_bio, &map_bh,
+					&first_logical_block,
+					get_block);
 			if (!pagevec_add(&lru_pvec, page))
 				__pagevec_lru_add(&lru_pvec);
 		} else {
@@ -364,9 +428,12 @@ int mpage_readpage(struct page *page, get_block_t get_block)
 {
 	struct bio *bio = NULL;
 	sector_t last_block_in_bio = 0;
+	struct buffer_head map_bh;
+	unsigned long first_logical_block = 0;
 
-	bio = do_mpage_readpage(bio, page, 1,
-			&last_block_in_bio, get_block);
+	clear_buffer_mapped(&map_bh);
+	bio = do_mpage_readpage(bio, page, 1, &last_block_in_bio,
+			&map_bh, &first_logical_block, get_block);
 	if (bio)
 		mpage_bio_submit(READ, bio);
 	return 0;
@@ -472,6 +539,7 @@ __mpage_writepage(struct bio *bio, struct page *page, get_block_t get_block,
 	for (page_block = 0; page_block < blocks_per_page; ) {
 
 		map_bh.b_state = 0;
+		map_bh.b_size = 1 << blkbits;
 		if (get_block(inode, block_in_file, &map_bh, 1))
 			goto confused;
 		if (buffer_new(&map_bh))
diff --git a/fs/namei.c b/fs/namei.c
index 8dc2b038d5d9..22f6e8d16aa8 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -104,7 +104,7 @@
  */
 /*
  * [Sep 2001 AV] Single-semaphore locking scheme (kudos to David Holland)
- * implemented.  Let's see if raised priority of ->s_vfs_rename_sem gives
+ * implemented.  Let's see if raised priority of ->s_vfs_rename_mutex gives
  * any extra contention...
  */
 
@@ -546,6 +546,22 @@ struct path {
 	struct dentry *dentry;
 };
 
+static inline void dput_path(struct path *path, struct nameidata *nd)
+{
+	dput(path->dentry);
+	if (path->mnt != nd->mnt)
+		mntput(path->mnt);
+}
+
+static inline void path_to_nameidata(struct path *path, struct nameidata *nd)
+{
+	dput(nd->dentry);
+	if (nd->mnt != path->mnt)
+		mntput(nd->mnt);
+	nd->mnt = path->mnt;
+	nd->dentry = path->dentry;
+}
+
 static __always_inline int __do_follow_link(struct path *path, struct nameidata *nd)
 {
 	int error;
@@ -555,8 +571,11 @@ static __always_inline int __do_follow_link(struct path *path, struct nameidata
 	touch_atime(path->mnt, dentry);
 	nd_set_link(nd, NULL);
 
-	if (path->mnt == nd->mnt)
-		mntget(path->mnt);
+	if (path->mnt != nd->mnt) {
+		path_to_nameidata(path, nd);
+		dget(dentry);
+	}
+	mntget(path->mnt);
 	cookie = dentry->d_inode->i_op->follow_link(dentry, nd);
 	error = PTR_ERR(cookie);
 	if (!IS_ERR(cookie)) {
@@ -573,22 +592,6 @@ static __always_inline int __do_follow_link(struct path *path, struct nameidata
 	return error;
 }
 
-static inline void dput_path(struct path *path, struct nameidata *nd)
-{
-	dput(path->dentry);
-	if (path->mnt != nd->mnt)
-		mntput(path->mnt);
-}
-
-static inline void path_to_nameidata(struct path *path, struct nameidata *nd)
-{
-	dput(nd->dentry);
-	if (nd->mnt != path->mnt)
-		mntput(nd->mnt);
-	nd->mnt = path->mnt;
-	nd->dentry = path->dentry;
-}
-
 /*
  * This limits recursive symlink follows to 8, while
  * limiting consecutive symlinks to 40.
@@ -1353,6 +1356,7 @@ static int may_delete(struct inode *dir,struct dentry *victim,int isdir)
 		return -ENOENT;
 
 	BUG_ON(victim->d_parent->d_inode != dir);
+	audit_inode_child(victim->d_name.name, victim->d_inode, dir->i_ino);
 
 	error = permission(dir,MAY_WRITE | MAY_EXEC, NULL);
 	if (error)
@@ -1422,7 +1426,7 @@ struct dentry *lock_rename(struct dentry *p1, struct dentry *p2)
 		return NULL;
 	}
 
-	down(&p1->d_inode->i_sb->s_vfs_rename_sem);
+	mutex_lock(&p1->d_inode->i_sb->s_vfs_rename_mutex);
 
 	for (p = p1; p->d_parent != p; p = p->d_parent) {
 		if (p->d_parent == p2) {
@@ -1450,7 +1454,7 @@ void unlock_rename(struct dentry *p1, struct dentry *p2)
 	mutex_unlock(&p1->d_inode->i_mutex);
 	if (p1 != p2) {
 		mutex_unlock(&p2->d_inode->i_mutex);
-		up(&p1->d_inode->i_sb->s_vfs_rename_sem);
+		mutex_unlock(&p1->d_inode->i_sb->s_vfs_rename_mutex);
 	}
 }
 
@@ -1472,7 +1476,7 @@ int vfs_create(struct inode *dir, struct dentry *dentry, int mode,
 	DQUOT_INIT(dir);
 	error = dir->i_op->create(dir, dentry, mode, nd);
 	if (!error)
-		fsnotify_create(dir, dentry->d_name.name);
+		fsnotify_create(dir, dentry);
 	return error;
 }
 
@@ -1628,6 +1632,12 @@ do_last:
 		goto exit;
 	}
 
+	if (IS_ERR(nd->intent.open.file)) {
+		mutex_unlock(&dir->d_inode->i_mutex);
+		error = PTR_ERR(nd->intent.open.file);
+		goto exit_dput;
+	}
+
 	/* Negative dentry, just create the file */
 	if (!path.dentry->d_inode) {
 		if (!IS_POSIXACL(dir->d_inode))
@@ -1793,7 +1803,7 @@ int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev)
 	DQUOT_INIT(dir);
 	error = dir->i_op->mknod(dir, dentry, mode, dev);
 	if (!error)
-		fsnotify_create(dir, dentry->d_name.name);
+		fsnotify_create(dir, dentry);
 	return error;
 }
 
@@ -1870,7 +1880,7 @@ int vfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	DQUOT_INIT(dir);
 	error = dir->i_op->mkdir(dir, dentry, mode);
 	if (!error)
-		fsnotify_mkdir(dir, dentry->d_name.name);
+		fsnotify_mkdir(dir, dentry);
 	return error;
 }
 
@@ -2133,7 +2143,7 @@ int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname, i
 	DQUOT_INIT(dir);
 	error = dir->i_op->symlink(dir, dentry, oldname);
 	if (!error)
-		fsnotify_create(dir, dentry->d_name.name);
+		fsnotify_create(dir, dentry);
 	return error;
 }
 
@@ -2210,7 +2220,7 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de
 	error = dir->i_op->link(old_dentry, dir, new_dentry);
 	mutex_unlock(&old_dentry->d_inode->i_mutex);
 	if (!error)
-		fsnotify_create(dir, new_dentry->d_name.name);
+		fsnotify_create(dir, new_dentry);
 	return error;
 }
 
@@ -2277,17 +2287,17 @@ asmlinkage long sys_link(const char __user *oldname, const char __user *newname)
  *	a) we can get into loop creation. Check is done in is_subdir().
  *	b) race potential - two innocent renames can create a loop together.
  *	   That's where 4.4 screws up. Current fix: serialization on
- *	   sb->s_vfs_rename_sem. We might be more accurate, but that's another
+ *	   sb->s_vfs_rename_mutex. We might be more accurate, but that's another
  *	   story.
  *	c) we have to lock _three_ objects - parents and victim (if it exists).
  *	   And that - after we got ->i_mutex on parents (until then we don't know
  *	   whether the target exists).  Solution: try to be smart with locking
  *	   order for inodes.  We rely on the fact that tree topology may change
- *	   only under ->s_vfs_rename_sem _and_ that parent of the object we
+ *	   only under ->s_vfs_rename_mutex _and_ that parent of the object we
  *	   move will be locked.  Thus we can rank directories by the tree
  *	   (ancestors first) and rank all non-directories after them.
  *	   That works since everybody except rename does "lock parent, lookup,
- *	   lock child" and rename is under ->s_vfs_rename_sem.
+ *	   lock child" and rename is under ->s_vfs_rename_mutex.
  *	   HOWEVER, it relies on the assumption that any object with ->lookup()
  *	   has no more than 1 dentry.  If "hybrid" objects will ever appear,
  *	   we'd better make sure that there's no link(2) for them.
@@ -2621,16 +2631,27 @@ int __page_symlink(struct inode *inode, const char *symname, int len,
 	int err = -ENOMEM;
 	char *kaddr;
 
+retry:
 	page = find_or_create_page(mapping, 0, gfp_mask);
 	if (!page)
 		goto fail;
 	err = mapping->a_ops->prepare_write(NULL, page, 0, len-1);
+	if (err == AOP_TRUNCATED_PAGE) {
+		page_cache_release(page);
+		goto retry;
+	}
 	if (err)
 		goto fail_map;
 	kaddr = kmap_atomic(page, KM_USER0);
 	memcpy(kaddr, symname, len-1);
 	kunmap_atomic(kaddr, KM_USER0);
-	mapping->a_ops->commit_write(NULL, page, 0, len-1);
+	err = mapping->a_ops->commit_write(NULL, page, 0, len-1);
+	if (err == AOP_TRUNCATED_PAGE) {
+		page_cache_release(page);
+		goto retry;
+	}
+	if (err)
+		goto fail_map;
 	/*
 	 * Notice that we are _not_ going to block here - end of page is
 	 * unmapped, so this will only try to map the rest of page, see
@@ -2640,7 +2661,8 @@ int __page_symlink(struct inode *inode, const char *symname, int len,
 	 */
 	if (!PageUptodate(page)) {
 		err = mapping->a_ops->readpage(NULL, page);
-		wait_on_page_locked(page);
+		if (err != AOP_TRUNCATED_PAGE)
+			wait_on_page_locked(page);
 	} else {
 		unlock_page(page);
 	}
diff --git a/fs/namespace.c b/fs/namespace.c
index 39c81a8d6316..bf478addb852 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -43,9 +43,9 @@ __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock);
 
 static int event;
 
-static struct list_head *mount_hashtable;
+static struct list_head *mount_hashtable __read_mostly;
 static int hash_mask __read_mostly, hash_bits __read_mostly;
-static kmem_cache_t *mnt_cache;
+static kmem_cache_t *mnt_cache __read_mostly;
 static struct rw_semaphore namespace_sem;
 
 /* /sys/fs */
@@ -399,6 +399,44 @@ struct seq_operations mounts_op = {
 	.show	= show_vfsmnt
 };
 
+static int show_vfsstat(struct seq_file *m, void *v)
+{
+	struct vfsmount *mnt = v;
+	int err = 0;
+
+	/* device */
+	if (mnt->mnt_devname) {
+		seq_puts(m, "device ");
+		mangle(m, mnt->mnt_devname);
+	} else
+		seq_puts(m, "no device");
+
+	/* mount point */
+	seq_puts(m, " mounted on ");
+	seq_path(m, mnt, mnt->mnt_root, " \t\n\\");
+	seq_putc(m, ' ');
+
+	/* file system type */
+	seq_puts(m, "with fstype ");
+	mangle(m, mnt->mnt_sb->s_type->name);
+
+	/* optional statistics */
+	if (mnt->mnt_sb->s_op->show_stats) {
+		seq_putc(m, ' ');
+		err = mnt->mnt_sb->s_op->show_stats(m, mnt);
+	}
+
+	seq_putc(m, '\n');
+	return err;
+}
+
+struct seq_operations mountstats_op = {
+	.start	= m_start,
+	.next	= m_next,
+	.stop	= m_stop,
+	.show	= show_vfsstat,
+};
+
 /**
  * may_umount_tree - check if a mount tree is busy
  * @mnt: root of mount tree
@@ -421,9 +459,9 @@ int may_umount_tree(struct vfsmount *mnt)
 	spin_unlock(&vfsmount_lock);
 
 	if (actual_refs > minimum_refs)
-		return -EBUSY;
+		return 0;
 
-	return 0;
+	return 1;
 }
 
 EXPORT_SYMBOL(may_umount_tree);
@@ -443,10 +481,10 @@ EXPORT_SYMBOL(may_umount_tree);
  */
 int may_umount(struct vfsmount *mnt)
 {
-	int ret = 0;
+	int ret = 1;
 	spin_lock(&vfsmount_lock);
 	if (propagate_mount_busy(mnt, 2))
-		ret = -EBUSY;
+		ret = 0;
 	spin_unlock(&vfsmount_lock);
 	return ret;
 }
diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c
index cfd76f431dc0..f0860c602d8b 100644
--- a/fs/ncpfs/dir.c
+++ b/fs/ncpfs/dir.c
@@ -49,7 +49,7 @@ extern int ncp_symlink(struct inode *, struct dentry *, const char *);
 #define ncp_symlink NULL
 #endif
 		      
-struct file_operations ncp_dir_operations =
+const struct file_operations ncp_dir_operations =
 {
 	.read		= generic_read_dir,
 	.readdir	= ncp_readdir,
diff --git a/fs/ncpfs/file.c b/fs/ncpfs/file.c
index 973b444d6914..e6b7c67cf057 100644
--- a/fs/ncpfs/file.c
+++ b/fs/ncpfs/file.c
@@ -46,7 +46,7 @@ int ncp_make_open(struct inode *inode, int right)
 		NCP_FINFO(inode)->volNumber, 
 		NCP_FINFO(inode)->dirEntNum);
 	error = -EACCES;
-	down(&NCP_FINFO(inode)->open_sem);
+	mutex_lock(&NCP_FINFO(inode)->open_mutex);
 	if (!atomic_read(&NCP_FINFO(inode)->opened)) {
 		struct ncp_entry_info finfo;
 		int result;
@@ -93,7 +93,7 @@ int ncp_make_open(struct inode *inode, int right)
 	}
 
 out_unlock:
-	up(&NCP_FINFO(inode)->open_sem);
+	mutex_unlock(&NCP_FINFO(inode)->open_mutex);
 out:
 	return error;
 }
@@ -283,7 +283,7 @@ static int ncp_release(struct inode *inode, struct file *file) {
 	return 0;
 }
 
-struct file_operations ncp_file_operations =
+const struct file_operations ncp_file_operations =
 {
 	.llseek		= remote_llseek,
 	.read		= ncp_file_read,
diff --git a/fs/ncpfs/inode.c b/fs/ncpfs/inode.c
index d277a58bd128..a1f3e972c6ef 100644
--- a/fs/ncpfs/inode.c
+++ b/fs/ncpfs/inode.c
@@ -63,7 +63,7 @@ static void init_once(void * foo, kmem_cache_t * cachep, unsigned long flags)
 
 	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
 	    SLAB_CTOR_CONSTRUCTOR) {
-		init_MUTEX(&ei->open_sem);
+		mutex_init(&ei->open_mutex);
 		inode_init_once(&ei->vfs_inode);
 	}
 }
@@ -72,7 +72,8 @@ static int init_inodecache(void)
 {
 	ncp_inode_cachep = kmem_cache_create("ncp_inode_cache",
 					     sizeof(struct ncp_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (ncp_inode_cachep == NULL)
 		return -ENOMEM;
@@ -520,7 +521,7 @@ static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent)
 	}
 
 /*	server->lock = 0;	*/
-	init_MUTEX(&server->sem);
+	mutex_init(&server->mutex);
 	server->packet = NULL;
 /*	server->buffer_size = 0;	*/
 /*	server->conn_status = 0;	*/
@@ -557,7 +558,7 @@ static int ncp_fill_super(struct super_block *sb, void *raw_data, int silent)
 	server->dentry_ttl = 0;	/* no caching */
 
 	INIT_LIST_HEAD(&server->tx.requests);
-	init_MUTEX(&server->rcv.creq_sem);
+	mutex_init(&server->rcv.creq_mutex);
 	server->tx.creq		= NULL;
 	server->rcv.creq	= NULL;
 	server->data_ready	= sock->sk->sk_data_ready;
diff --git a/fs/ncpfs/ncplib_kernel.c b/fs/ncpfs/ncplib_kernel.c
index c755e1848a42..d9ebf6439f59 100644
--- a/fs/ncpfs/ncplib_kernel.c
+++ b/fs/ncpfs/ncplib_kernel.c
@@ -291,7 +291,7 @@ ncp_make_closed(struct inode *inode)
 	int err;
 
 	err = 0;
-	down(&NCP_FINFO(inode)->open_sem);	
+	mutex_lock(&NCP_FINFO(inode)->open_mutex);
 	if (atomic_read(&NCP_FINFO(inode)->opened) == 1) {
 		atomic_set(&NCP_FINFO(inode)->opened, 0);
 		err = ncp_close_file(NCP_SERVER(inode), NCP_FINFO(inode)->file_handle);
@@ -301,7 +301,7 @@ ncp_make_closed(struct inode *inode)
 				NCP_FINFO(inode)->volNumber,
 				NCP_FINFO(inode)->dirEntNum, err);
 	}
-	up(&NCP_FINFO(inode)->open_sem);
+	mutex_unlock(&NCP_FINFO(inode)->open_mutex);
 	return err;
 }
 
diff --git a/fs/ncpfs/sock.c b/fs/ncpfs/sock.c
index 6593a5ca88ba..8783eb7ec641 100644
--- a/fs/ncpfs/sock.c
+++ b/fs/ncpfs/sock.c
@@ -171,9 +171,9 @@ static inline void __ncp_abort_request(struct ncp_server *server, struct ncp_req
 
 static inline void ncp_abort_request(struct ncp_server *server, struct ncp_request_reply *req, int err)
 {
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncp_abort_request(server, req, err);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 static inline void __ncptcp_abort(struct ncp_server *server)
@@ -303,20 +303,20 @@ static inline void __ncp_start_request(struct ncp_server *server, struct ncp_req
 
 static int ncp_add_request(struct ncp_server *server, struct ncp_request_reply *req)
 {
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	if (!ncp_conn_valid(server)) {
-		up(&server->rcv.creq_sem);
+		mutex_unlock(&server->rcv.creq_mutex);
 		printk(KERN_ERR "ncpfs: tcp: Server died\n");
 		return -EIO;
 	}
 	if (server->tx.creq || server->rcv.creq) {
 		req->status = RQ_QUEUED;
 		list_add_tail(&req->req, &server->tx.requests);
-		up(&server->rcv.creq_sem);
+		mutex_unlock(&server->rcv.creq_mutex);
 		return 0;
 	}
 	__ncp_start_request(server, req);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 	return 0;
 }
 
@@ -400,7 +400,7 @@ void ncpdgram_rcv_proc(void *s)
 				info_server(server, 0, server->unexpected_packet.data, result);
 				continue;
 			}
-			down(&server->rcv.creq_sem);		
+			mutex_lock(&server->rcv.creq_mutex);
 			req = server->rcv.creq;
 			if (req && (req->tx_type == NCP_ALLOC_SLOT_REQUEST || (server->sequence == reply.sequence && 
 					server->connection == get_conn_number(&reply)))) {
@@ -430,11 +430,11 @@ void ncpdgram_rcv_proc(void *s)
 				     	server->rcv.creq = NULL;
 					ncp_finish_request(req, result);
 					__ncp_next_request(server);
-					up(&server->rcv.creq_sem);
+					mutex_unlock(&server->rcv.creq_mutex);
 					continue;
 				}
 			}
-			up(&server->rcv.creq_sem);
+			mutex_unlock(&server->rcv.creq_mutex);
 		}
 drop:;		
 		_recv(sock, &reply, sizeof(reply), MSG_DONTWAIT);
@@ -472,9 +472,9 @@ static void __ncpdgram_timeout_proc(struct ncp_server *server)
 void ncpdgram_timeout_proc(void *s)
 {
 	struct ncp_server *server = s;
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncpdgram_timeout_proc(server);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 static inline void ncp_init_req(struct ncp_request_reply* req)
@@ -657,18 +657,18 @@ void ncp_tcp_rcv_proc(void *s)
 {
 	struct ncp_server *server = s;
 
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncptcp_rcv_proc(server);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 void ncp_tcp_tx_proc(void *s)
 {
 	struct ncp_server *server = s;
 	
-	down(&server->rcv.creq_sem);
+	mutex_lock(&server->rcv.creq_mutex);
 	__ncptcp_try_send(server);
-	up(&server->rcv.creq_sem);
+	mutex_unlock(&server->rcv.creq_mutex);
 }
 
 static int do_ncp_rpc_call(struct ncp_server *server, int size,
@@ -833,7 +833,7 @@ int ncp_disconnect(struct ncp_server *server)
 
 void ncp_lock_server(struct ncp_server *server)
 {
-	down(&server->sem);
+	mutex_lock(&server->mutex);
 	if (server->lock)
 		printk(KERN_WARNING "ncp_lock_server: was locked!\n");
 	server->lock = 1;
@@ -846,5 +846,5 @@ void ncp_unlock_server(struct ncp_server *server)
 		return;
 	}
 	server->lock = 0;
-	up(&server->sem);
+	mutex_unlock(&server->mutex);
 }
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index fcd97406a778..90c95adc8c1b 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -14,6 +14,7 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/svcsock.h>
 #include <linux/nfs_fs.h>
+#include <linux/mutex.h>
 
 #include <net/inet_sock.h>
 
@@ -31,7 +32,7 @@ struct nfs_callback_data {
 };
 
 static struct nfs_callback_data nfs_callback_info;
-static DECLARE_MUTEX(nfs_callback_sema);
+static DEFINE_MUTEX(nfs_callback_mutex);
 static struct svc_program nfs4_callback_program;
 
 unsigned int nfs_callback_set_tcpport;
@@ -55,7 +56,12 @@ static void nfs_callback_svc(struct svc_rqst *rqstp)
 
 	complete(&nfs_callback_info.started);
 
-	while (nfs_callback_info.users != 0 || !signalled()) {
+	for(;;) {
+		if (signalled()) {
+			if (nfs_callback_info.users == 0)
+				break;
+			flush_signals(current);
+		}
 		/*
 		 * Listen for a request on the socket
 		 */
@@ -73,6 +79,7 @@ static void nfs_callback_svc(struct svc_rqst *rqstp)
 		svc_process(serv, rqstp);
 	}
 
+	svc_exit_thread(rqstp);
 	nfs_callback_info.pid = 0;
 	complete(&nfs_callback_info.stopped);
 	unlock_kernel();
@@ -89,7 +96,7 @@ int nfs_callback_up(void)
 	int ret = 0;
 
 	lock_kernel();
-	down(&nfs_callback_sema);
+	mutex_lock(&nfs_callback_mutex);
 	if (nfs_callback_info.users++ || nfs_callback_info.pid != 0)
 		goto out;
 	init_completion(&nfs_callback_info.started);
@@ -115,7 +122,7 @@ int nfs_callback_up(void)
 	nfs_callback_info.serv = serv;
 	wait_for_completion(&nfs_callback_info.started);
 out:
-	up(&nfs_callback_sema);
+	mutex_unlock(&nfs_callback_mutex);
 	unlock_kernel();
 	return ret;
 out_destroy:
@@ -133,13 +140,15 @@ int nfs_callback_down(void)
 	int ret = 0;
 
 	lock_kernel();
-	down(&nfs_callback_sema);
-	if (--nfs_callback_info.users || nfs_callback_info.pid == 0)
-		goto out;
-	kill_proc(nfs_callback_info.pid, SIGKILL, 1);
-	wait_for_completion(&nfs_callback_info.stopped);
-out:
-	up(&nfs_callback_sema);
+	mutex_lock(&nfs_callback_mutex);
+	nfs_callback_info.users--;
+	do {
+		if (nfs_callback_info.users != 0 || nfs_callback_info.pid == 0)
+			break;
+		if (kill_proc(nfs_callback_info.pid, SIGKILL, 1) < 0)
+			break;
+	} while (wait_for_completion_timeout(&nfs_callback_info.stopped, 5*HZ) == 0);
+	mutex_unlock(&nfs_callback_mutex);
 	unlock_kernel();
 	return ret;
 }
diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c
index 7c33b9a81a94..05c38cf40b69 100644
--- a/fs/nfs/callback_xdr.c
+++ b/fs/nfs/callback_xdr.c
@@ -330,7 +330,7 @@ static unsigned encode_op_hdr(struct xdr_stream *xdr, uint32_t op, uint32_t res)
 
 static unsigned encode_getattr_res(struct svc_rqst *rqstp, struct xdr_stream *xdr, const struct cb_getattrres *res)
 {
-	uint32_t *savep;
+	uint32_t *savep = NULL;
 	unsigned status = res->status;
 	
 	if (unlikely(status != 0))
@@ -358,23 +358,26 @@ static unsigned process_op(struct svc_rqst *rqstp,
 		struct xdr_stream *xdr_in, void *argp,
 		struct xdr_stream *xdr_out, void *resp)
 {
-	struct callback_op *op;
-	unsigned int op_nr;
+	struct callback_op *op = &callback_ops[0];
+	unsigned int op_nr = OP_CB_ILLEGAL;
 	unsigned int status = 0;
 	long maxlen;
 	unsigned res;
 
 	dprintk("%s: start\n", __FUNCTION__);
 	status = decode_op_hdr(xdr_in, &op_nr);
-	if (unlikely(status != 0)) {
-		op_nr = OP_CB_ILLEGAL;
-		op = &callback_ops[0];
-	} else if (unlikely(op_nr != OP_CB_GETATTR && op_nr != OP_CB_RECALL)) {
-		op_nr = OP_CB_ILLEGAL;
-		op = &callback_ops[0];
-		status = htonl(NFS4ERR_OP_ILLEGAL);
-	} else
-		op = &callback_ops[op_nr];
+	if (likely(status == 0)) {
+		switch (op_nr) {
+			case OP_CB_GETATTR:
+			case OP_CB_RECALL:
+				op = &callback_ops[op_nr];
+				break;
+			default:
+				op_nr = OP_CB_ILLEGAL;
+				op = &callback_ops[0];
+				status = htonl(NFS4ERR_OP_ILLEGAL);
+		}
+	}
 
 	maxlen = xdr_out->end - xdr_out->p;
 	if (maxlen > 0 && maxlen < PAGE_SIZE) {
@@ -416,6 +419,7 @@ static int nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *resp
 	decode_compound_hdr_arg(&xdr_in, &hdr_arg);
 	hdr_res.taglen = hdr_arg.taglen;
 	hdr_res.tag = hdr_arg.tag;
+	hdr_res.nops = NULL;
 	encode_compound_hdr_res(&xdr_out, &hdr_res);
 
 	for (;;) {
diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
index c6f07c1c71e6..d3be923d4e43 100644
--- a/fs/nfs/delegation.c
+++ b/fs/nfs/delegation.c
@@ -421,3 +421,22 @@ void nfs_delegation_reap_unclaimed(struct nfs4_client *clp)
 		nfs_free_delegation(delegation);
 	}
 }
+
+int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode)
+{
+	struct nfs4_client *clp = NFS_SERVER(inode)->nfs4_state;
+	struct nfs_inode *nfsi = NFS_I(inode);
+	struct nfs_delegation *delegation;
+	int res = 0;
+
+	if (nfsi->delegation_state == 0)
+		return 0;
+	spin_lock(&clp->cl_lock);
+	delegation = nfsi->delegation;
+	if (delegation != NULL) {
+		memcpy(dst->data, delegation->stateid.data, sizeof(dst->data));
+		res = 1;
+	}
+	spin_unlock(&clp->cl_lock);
+	return res;
+}
diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h
index 7a0b2bfce771..3858694652fa 100644
--- a/fs/nfs/delegation.h
+++ b/fs/nfs/delegation.h
@@ -41,6 +41,7 @@ void nfs_delegation_reap_unclaimed(struct nfs4_client *clp);
 int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid);
 int nfs4_open_delegation_recall(struct dentry *dentry, struct nfs4_state *state);
 int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl);
+int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode);
 
 static inline int nfs_have_delegation(struct inode *inode, int flags)
 {
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index a1554bead692..a23f34894167 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -34,6 +34,7 @@
 
 #include "nfs4_fs.h"
 #include "delegation.h"
+#include "iostat.h"
 
 #define NFS_PARANOIA 1
 /* #define NFS_DEBUG_VERBOSE 1 */
@@ -53,7 +54,7 @@ static int nfs_rename(struct inode *, struct dentry *,
 static int nfs_fsync_dir(struct file *, struct dentry *, int);
 static loff_t nfs_llseek_dir(struct file *, loff_t, int);
 
-struct file_operations nfs_dir_operations = {
+const struct file_operations nfs_dir_operations = {
 	.llseek		= nfs_llseek_dir,
 	.read		= generic_read_dir,
 	.readdir	= nfs_readdir,
@@ -129,6 +130,9 @@ nfs_opendir(struct inode *inode, struct file *filp)
 {
 	int res = 0;
 
+	dfprintk(VFS, "NFS: opendir(%s/%ld)\n",
+			inode->i_sb->s_id, inode->i_ino);
+
 	lock_kernel();
 	/* Call generic open code in order to cache credentials */
 	if (!res)
@@ -172,7 +176,9 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
 	unsigned long	timestamp;
 	int		error;
 
-	dfprintk(VFS, "NFS: nfs_readdir_filler() reading cookie %Lu into page %lu.\n", (long long)desc->entry->cookie, page->index);
+	dfprintk(DIRCACHE, "NFS: %s: reading cookie %Lu into page %lu\n",
+			__FUNCTION__, (long long)desc->entry->cookie,
+			page->index);
 
  again:
 	timestamp = jiffies;
@@ -244,7 +250,8 @@ int find_dirent(nfs_readdir_descriptor_t *desc)
 			status;
 
 	while((status = dir_decode(desc)) == 0) {
-		dfprintk(VFS, "NFS: found cookie %Lu\n", (unsigned long long)entry->cookie);
+		dfprintk(DIRCACHE, "NFS: %s: examining cookie %Lu\n",
+				__FUNCTION__, (unsigned long long)entry->cookie);
 		if (entry->prev_cookie == *desc->dir_cookie)
 			break;
 		if (loop_count++ > 200) {
@@ -252,7 +259,6 @@ int find_dirent(nfs_readdir_descriptor_t *desc)
 			schedule();
 		}
 	}
-	dfprintk(VFS, "NFS: find_dirent() returns %d\n", status);
 	return status;
 }
 
@@ -276,7 +282,8 @@ int find_dirent_index(nfs_readdir_descriptor_t *desc)
 		if (status)
 			break;
 
-		dfprintk(VFS, "NFS: found cookie %Lu at index %Ld\n", (unsigned long long)entry->cookie, desc->current_index);
+		dfprintk(DIRCACHE, "NFS: found cookie %Lu at index %Ld\n",
+				(unsigned long long)entry->cookie, desc->current_index);
 
 		if (desc->file->f_pos == desc->current_index) {
 			*desc->dir_cookie = entry->cookie;
@@ -288,7 +295,6 @@ int find_dirent_index(nfs_readdir_descriptor_t *desc)
 			schedule();
 		}
 	}
-	dfprintk(VFS, "NFS: find_dirent_index() returns %d\n", status);
 	return status;
 }
 
@@ -303,7 +309,9 @@ int find_dirent_page(nfs_readdir_descriptor_t *desc)
 	struct page	*page;
 	int		status;
 
-	dfprintk(VFS, "NFS: find_dirent_page() searching directory page %ld\n", desc->page_index);
+	dfprintk(DIRCACHE, "NFS: %s: searching page %ld for target %Lu\n",
+			__FUNCTION__, desc->page_index,
+			(long long) *desc->dir_cookie);
 
 	page = read_cache_page(inode->i_mapping, desc->page_index,
 			       (filler_t *)nfs_readdir_filler, desc);
@@ -324,7 +332,7 @@ int find_dirent_page(nfs_readdir_descriptor_t *desc)
 	if (status < 0)
 		dir_page_release(desc);
  out:
-	dfprintk(VFS, "NFS: find_dirent_page() returns %d\n", status);
+	dfprintk(DIRCACHE, "NFS: %s: returns %d\n", __FUNCTION__, status);
 	return status;
  read_error:
 	page_cache_release(page);
@@ -346,13 +354,15 @@ int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
 
 	/* Always search-by-index from the beginning of the cache */
 	if (*desc->dir_cookie == 0) {
-		dfprintk(VFS, "NFS: readdir_search_pagecache() searching for offset %Ld\n", (long long)desc->file->f_pos);
+		dfprintk(DIRCACHE, "NFS: readdir_search_pagecache() searching for offset %Ld\n",
+				(long long)desc->file->f_pos);
 		desc->page_index = 0;
 		desc->entry->cookie = desc->entry->prev_cookie = 0;
 		desc->entry->eof = 0;
 		desc->current_index = 0;
 	} else
-		dfprintk(VFS, "NFS: readdir_search_pagecache() searching for cookie %Lu\n", (unsigned long long)*desc->dir_cookie);
+		dfprintk(DIRCACHE, "NFS: readdir_search_pagecache() searching for cookie %Lu\n",
+				(unsigned long long)*desc->dir_cookie);
 
 	for (;;) {
 		res = find_dirent_page(desc);
@@ -365,7 +375,8 @@ int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
 			schedule();
 		}
 	}
-	dfprintk(VFS, "NFS: readdir_search_pagecache() returned %d\n", res);
+
+	dfprintk(DIRCACHE, "NFS: %s: returns %d\n", __FUNCTION__, res);
 	return res;
 }
 
@@ -390,7 +401,8 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc, void *dirent,
 	int		loop_count = 0,
 			res;
 
-	dfprintk(VFS, "NFS: nfs_do_filldir() filling starting @ cookie %Lu\n", (long long)entry->cookie);
+	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling starting @ cookie %Lu\n",
+			(unsigned long long)entry->cookie);
 
 	for(;;) {
 		unsigned d_type = DT_UNKNOWN;
@@ -427,7 +439,8 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc, void *dirent,
 	dir_page_release(desc);
 	if (dentry != NULL)
 		dput(dentry);
-	dfprintk(VFS, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n", (unsigned long long)*desc->dir_cookie, res);
+	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n",
+			(unsigned long long)*desc->dir_cookie, res);
 	return res;
 }
 
@@ -453,7 +466,8 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	struct page	*page = NULL;
 	int		status;
 
-	dfprintk(VFS, "NFS: uncached_readdir() searching for cookie %Lu\n", (unsigned long long)*desc->dir_cookie);
+	dfprintk(DIRCACHE, "NFS: uncached_readdir() searching for cookie %Lu\n",
+			(unsigned long long)*desc->dir_cookie);
 
 	page = alloc_page(GFP_HIGHUSER);
 	if (!page) {
@@ -485,7 +499,8 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc, void *dirent,
 	desc->entry->cookie = desc->entry->prev_cookie = 0;
 	desc->entry->eof = 0;
  out:
-	dfprintk(VFS, "NFS: uncached_readdir() returns %d\n", status);
+	dfprintk(DIRCACHE, "NFS: %s: returns %d\n",
+			__FUNCTION__, status);
 	return status;
  out_release:
 	dir_page_release(desc);
@@ -507,6 +522,11 @@ static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	struct nfs_fattr fattr;
 	long		res;
 
+	dfprintk(VFS, "NFS: readdir(%s/%s) starting at cookie %Lu\n",
+			dentry->d_parent->d_name.name, dentry->d_name.name,
+			(long long)filp->f_pos);
+	nfs_inc_stats(inode, NFSIOS_VFSGETDENTS);
+
 	lock_kernel();
 
 	res = nfs_revalidate_inode(NFS_SERVER(inode), inode);
@@ -566,9 +586,12 @@ static int nfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		}
 	}
 	unlock_kernel();
-	if (res < 0)
-		return res;
-	return 0;
+	if (res > 0)
+		res = 0;
+	dfprintk(VFS, "NFS: readdir(%s/%s) returns %ld\n",
+			dentry->d_parent->d_name.name, dentry->d_name.name,
+			res);
+	return res;
 }
 
 loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int origin)
@@ -599,6 +622,10 @@ out:
  */
 int nfs_fsync_dir(struct file *filp, struct dentry *dentry, int datasync)
 {
+	dfprintk(VFS, "NFS: fsync_dir(%s/%s) datasync %d\n",
+			dentry->d_parent->d_name.name, dentry->d_name.name,
+			datasync);
+
 	return 0;
 }
 
@@ -713,6 +740,7 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
 	parent = dget_parent(dentry);
 	lock_kernel();
 	dir = parent->d_inode;
+	nfs_inc_stats(dir, NFSIOS_DENTRYREVALIDATE);
 	inode = dentry->d_inode;
 
 	if (!inode) {
@@ -722,8 +750,9 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
 	}
 
 	if (is_bad_inode(inode)) {
-		dfprintk(VFS, "nfs_lookup_validate: %s/%s has dud inode\n",
-			dentry->d_parent->d_name.name, dentry->d_name.name);
+		dfprintk(LOOKUPCACHE, "%s: %s/%s has dud inode\n",
+				__FUNCTION__, dentry->d_parent->d_name.name,
+				dentry->d_name.name);
 		goto out_bad;
 	}
 
@@ -755,6 +784,9 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
  out_valid:
 	unlock_kernel();
 	dput(parent);
+	dfprintk(LOOKUPCACHE, "NFS: %s(%s/%s) is valid\n",
+			__FUNCTION__, dentry->d_parent->d_name.name,
+			dentry->d_name.name);
 	return 1;
 out_zap_parent:
 	nfs_zap_caches(dir);
@@ -771,6 +803,9 @@ out_zap_parent:
 	d_drop(dentry);
 	unlock_kernel();
 	dput(parent);
+	dfprintk(LOOKUPCACHE, "NFS: %s(%s/%s) is invalid\n",
+			__FUNCTION__, dentry->d_parent->d_name.name,
+			dentry->d_name.name);
 	return 0;
 }
 
@@ -844,6 +879,7 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru
 
 	dfprintk(VFS, "NFS: lookup(%s/%s)\n",
 		dentry->d_parent->d_name.name, dentry->d_name.name);
+	nfs_inc_stats(dir, NFSIOS_VFSLOOKUP);
 
 	res = ERR_PTR(-ENAMETOOLONG);
 	if (dentry->d_name.len > NFS_SERVER(dir)->namelen)
@@ -865,9 +901,9 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru
 		res = ERR_PTR(error);
 		goto out_unlock;
 	}
-	res = ERR_PTR(-EACCES);
 	inode = nfs_fhget(dentry->d_sb, &fhandle, &fattr);
-	if (!inode)
+	res = (struct dentry *)inode;
+	if (IS_ERR(res))
 		goto out_unlock;
 no_entry:
 	res = d_add_unique(dentry, inode);
@@ -912,6 +948,9 @@ static struct dentry *nfs_atomic_lookup(struct inode *dir, struct dentry *dentry
 	struct dentry *res = NULL;
 	int error;
 
+	dfprintk(VFS, "NFS: atomic_lookup(%s/%ld), %s\n",
+			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
+
 	/* Check that we are indeed trying to open this file */
 	if (!is_atomic_open(dir, nd))
 		goto no_open;
@@ -1057,7 +1096,7 @@ static struct dentry *nfs_readdir_lookup(nfs_readdir_descriptor_t *desc)
 		return NULL;
 	dentry->d_op = NFS_PROTO(dir)->dentry_ops;
 	inode = nfs_fhget(dentry->d_sb, entry->fh, entry->fattr);
-	if (!inode) {
+	if (IS_ERR(inode)) {
 		dput(dentry);
 		return NULL;
 	}
@@ -1095,9 +1134,9 @@ int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
 		if (error < 0)
 			goto out_err;
 	}
-	error = -ENOMEM;
 	inode = nfs_fhget(dentry->d_sb, fhandle, fattr);
-	if (inode == NULL)
+	error = PTR_ERR(inode);
+	if (IS_ERR(inode))
 		goto out_err;
 	d_instantiate(dentry, inode);
 	return 0;
@@ -1119,8 +1158,8 @@ static int nfs_create(struct inode *dir, struct dentry *dentry, int mode,
 	int error;
 	int open_flags = 0;
 
-	dfprintk(VFS, "NFS: create(%s/%ld, %s\n", dir->i_sb->s_id, 
-		dir->i_ino, dentry->d_name.name);
+	dfprintk(VFS, "NFS: create(%s/%ld), %s\n",
+			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
 
 	attr.ia_mode = mode;
 	attr.ia_valid = ATTR_MODE;
@@ -1153,8 +1192,8 @@ nfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t rdev)
 	struct iattr attr;
 	int status;
 
-	dfprintk(VFS, "NFS: mknod(%s/%ld, %s\n", dir->i_sb->s_id,
-		dir->i_ino, dentry->d_name.name);
+	dfprintk(VFS, "NFS: mknod(%s/%ld), %s\n",
+			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
 
 	if (!new_valid_dev(rdev))
 		return -EINVAL;
@@ -1186,8 +1225,8 @@ static int nfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
 	struct iattr attr;
 	int error;
 
-	dfprintk(VFS, "NFS: mkdir(%s/%ld, %s\n", dir->i_sb->s_id,
-		dir->i_ino, dentry->d_name.name);
+	dfprintk(VFS, "NFS: mkdir(%s/%ld), %s\n",
+			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
 
 	attr.ia_valid = ATTR_MODE;
 	attr.ia_mode = mode | S_IFDIR;
@@ -1212,8 +1251,8 @@ static int nfs_rmdir(struct inode *dir, struct dentry *dentry)
 {
 	int error;
 
-	dfprintk(VFS, "NFS: rmdir(%s/%ld, %s\n", dir->i_sb->s_id,
-		dir->i_ino, dentry->d_name.name);
+	dfprintk(VFS, "NFS: rmdir(%s/%ld), %s\n",
+			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
 
 	lock_kernel();
 	nfs_begin_data_update(dir);
@@ -1241,6 +1280,7 @@ static int nfs_sillyrename(struct inode *dir, struct dentry *dentry)
 	dfprintk(VFS, "NFS: silly-rename(%s/%s, ct=%d)\n",
 		dentry->d_parent->d_name.name, dentry->d_name.name, 
 		atomic_read(&dentry->d_count));
+	nfs_inc_stats(dir, NFSIOS_SILLYRENAME);
 
 #ifdef NFS_PARANOIA
 if (!dentry->d_inode)
@@ -1268,8 +1308,8 @@ dentry->d_parent->d_name.name, dentry->d_name.name);
 		sillycounter++;
 		sprintf(suffix, "%*.*x", countersize, countersize, sillycounter);
 
-		dfprintk(VFS, "trying to rename %s to %s\n",
-			 dentry->d_name.name, silly);
+		dfprintk(VFS, "NFS: trying to rename %s to %s\n",
+				dentry->d_name.name, silly);
 		
 		sdentry = lookup_one_len(silly, dentry->d_parent, slen);
 		/*
@@ -1640,6 +1680,8 @@ int nfs_permission(struct inode *inode, int mask, struct nameidata *nd)
 	struct rpc_cred *cred;
 	int res = 0;
 
+	nfs_inc_stats(inode, NFSIOS_VFSACCESS);
+
 	if (mask == 0)
 		goto out;
 	/* Is this sys_access() ? */
@@ -1679,13 +1721,15 @@ force_lookup:
 		res = PTR_ERR(cred);
 	unlock_kernel();
 out:
+	dfprintk(VFS, "NFS: permission(%s/%ld), mask=0x%x, res=%d\n",
+		inode->i_sb->s_id, inode->i_ino, mask, res);
 	return res;
 out_notsup:
 	res = nfs_revalidate_inode(NFS_SERVER(inode), inode);
 	if (res == 0)
 		res = generic_permission(inode, mask, NULL);
 	unlock_kernel();
-	return res;
+	goto out;
 }
 
 /*
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 4e9b3a1b36c5..0f583cb16ddb 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -7,11 +7,11 @@
  *
  * There are important applications whose performance or correctness
  * depends on uncached access to file data.  Database clusters
- * (multiple copies of the same instance running on separate hosts) 
+ * (multiple copies of the same instance running on separate hosts)
  * implement their own cache coherency protocol that subsumes file
- * system cache protocols.  Applications that process datasets 
- * considerably larger than the client's memory do not always benefit 
- * from a local cache.  A streaming video server, for instance, has no 
+ * system cache protocols.  Applications that process datasets
+ * considerably larger than the client's memory do not always benefit
+ * from a local cache.  A streaming video server, for instance, has no
  * need to cache the contents of a file.
  *
  * When an application requests uncached I/O, all read and write requests
@@ -34,6 +34,7 @@
  * 08 Jun 2003	Port to 2.5 APIs  --cel
  * 31 Mar 2004	Handle direct I/O without VFS support  --cel
  * 15 Sep 2004	Parallel async reads  --cel
+ * 04 May 2005	support O_DIRECT with aio  --cel
  *
  */
 
@@ -54,10 +55,10 @@
 #include <asm/uaccess.h>
 #include <asm/atomic.h>
 
+#include "iostat.h"
+
 #define NFSDBG_FACILITY		NFSDBG_VFS
-#define MAX_DIRECTIO_SIZE	(4096UL << PAGE_SHIFT)
 
-static void nfs_free_user_pages(struct page **pages, int npages, int do_dirty);
 static kmem_cache_t *nfs_direct_cachep;
 
 /*
@@ -65,38 +66,78 @@ static kmem_cache_t *nfs_direct_cachep;
  */
 struct nfs_direct_req {
 	struct kref		kref;		/* release manager */
-	struct list_head	list;		/* nfs_read_data structs */
-	wait_queue_head_t	wait;		/* wait for i/o completion */
+
+	/* I/O parameters */
+	struct list_head	list,		/* nfs_read/write_data structs */
+				rewrite_list;	/* saved nfs_write_data structs */
+	struct nfs_open_context	*ctx;		/* file open context info */
+	struct kiocb *		iocb;		/* controlling i/o request */
+	struct inode *		inode;		/* target file of i/o */
+	unsigned long		user_addr;	/* location of user's buffer */
+	size_t			user_count;	/* total bytes to move */
+	loff_t			pos;		/* starting offset in file */
 	struct page **		pages;		/* pages in our buffer */
 	unsigned int		npages;		/* count of pages */
-	atomic_t		complete,	/* i/os we're waiting for */
-				count,		/* bytes actually processed */
+
+	/* completion state */
+	spinlock_t		lock;		/* protect completion state */
+	int			outstanding;	/* i/os we're waiting for */
+	ssize_t			count,		/* bytes actually processed */
 				error;		/* any reported error */
+	struct completion	completion;	/* wait for i/o completion */
+
+	/* commit state */
+	struct nfs_write_data *	commit_data;	/* special write_data for commits */
+	int			flags;
+#define NFS_ODIRECT_DO_COMMIT		(1)	/* an unstable reply was received */
+#define NFS_ODIRECT_RESCHED_WRITES	(2)	/* write verification failed */
+	struct nfs_writeverf	verf;		/* unstable write verifier */
 };
 
+static void nfs_direct_write_schedule(struct nfs_direct_req *dreq, int sync);
+static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode);
 
 /**
- * nfs_get_user_pages - find and set up pages underlying user's buffer
- * rw: direction (read or write)
- * user_addr: starting address of this segment of user's buffer
- * count: size of this segment
- * @pages: returned array of page struct pointers underlying user's buffer
+ * nfs_direct_IO - NFS address space operation for direct I/O
+ * @rw: direction (read or write)
+ * @iocb: target I/O control block
+ * @iov: array of vectors that define I/O buffer
+ * @pos: offset in file to begin the operation
+ * @nr_segs: size of iovec array
+ *
+ * The presence of this routine in the address space ops vector means
+ * the NFS client supports direct I/O.  However, we shunt off direct
+ * read and write requests before the VFS gets them, so this method
+ * should never be called.
  */
-static inline int
-nfs_get_user_pages(int rw, unsigned long user_addr, size_t size,
-		struct page ***pages)
+ssize_t nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov, loff_t pos, unsigned long nr_segs)
+{
+	struct dentry *dentry = iocb->ki_filp->f_dentry;
+
+	dprintk("NFS: nfs_direct_IO (%s) off/no(%Ld/%lu) EINVAL\n",
+			dentry->d_name.name, (long long) pos, nr_segs);
+
+	return -EINVAL;
+}
+
+static void nfs_free_user_pages(struct page **pages, int npages, int do_dirty)
+{
+	int i;
+	for (i = 0; i < npages; i++) {
+		struct page *page = pages[i];
+		if (do_dirty && !PageCompound(page))
+			set_page_dirty_lock(page);
+		page_cache_release(page);
+	}
+	kfree(pages);
+}
+
+static inline int nfs_get_user_pages(int rw, unsigned long user_addr, size_t size, struct page ***pages)
 {
 	int result = -ENOMEM;
 	unsigned long page_count;
 	size_t array_size;
 
-	/* set an arbitrary limit to prevent type overflow */
-	/* XXX: this can probably be as large as INT_MAX */
-	if (size > MAX_DIRECTIO_SIZE) {
-		*pages = NULL;
-		return -EFBIG;
-	}
-
 	page_count = (user_addr + size + PAGE_SIZE - 1) >> PAGE_SHIFT;
 	page_count -= user_addr >> PAGE_SHIFT;
 
@@ -108,75 +149,117 @@ nfs_get_user_pages(int rw, unsigned long user_addr, size_t size,
 					page_count, (rw == READ), 0,
 					*pages, NULL);
 		up_read(&current->mm->mmap_sem);
-		/*
-		 * If we got fewer pages than expected from get_user_pages(),
-		 * the user buffer runs off the end of a mapping; return EFAULT.
-		 */
-		if (result >= 0 && result < page_count) {
-			nfs_free_user_pages(*pages, result, 0);
+		if (result != page_count) {
+			/*
+			 * If we got fewer pages than expected from
+			 * get_user_pages(), the user buffer runs off the
+			 * end of a mapping; return EFAULT.
+			 */
+			if (result >= 0) {
+				nfs_free_user_pages(*pages, result, 0);
+				result = -EFAULT;
+			} else
+				kfree(*pages);
 			*pages = NULL;
-			result = -EFAULT;
 		}
 	}
 	return result;
 }
 
-/**
- * nfs_free_user_pages - tear down page struct array
- * @pages: array of page struct pointers underlying target buffer
- * @npages: number of pages in the array
- * @do_dirty: dirty the pages as we release them
- */
-static void
-nfs_free_user_pages(struct page **pages, int npages, int do_dirty)
+static inline struct nfs_direct_req *nfs_direct_req_alloc(void)
 {
-	int i;
-	for (i = 0; i < npages; i++) {
-		struct page *page = pages[i];
-		if (do_dirty && !PageCompound(page))
-			set_page_dirty_lock(page);
-		page_cache_release(page);
-	}
-	kfree(pages);
+	struct nfs_direct_req *dreq;
+
+	dreq = kmem_cache_alloc(nfs_direct_cachep, SLAB_KERNEL);
+	if (!dreq)
+		return NULL;
+
+	kref_init(&dreq->kref);
+	init_completion(&dreq->completion);
+	INIT_LIST_HEAD(&dreq->list);
+	INIT_LIST_HEAD(&dreq->rewrite_list);
+	dreq->iocb = NULL;
+	dreq->ctx = NULL;
+	spin_lock_init(&dreq->lock);
+	dreq->outstanding = 0;
+	dreq->count = 0;
+	dreq->error = 0;
+	dreq->flags = 0;
+
+	return dreq;
 }
 
-/**
- * nfs_direct_req_release - release  nfs_direct_req structure for direct read
- * @kref: kref object embedded in an nfs_direct_req structure
- *
- */
 static void nfs_direct_req_release(struct kref *kref)
 {
 	struct nfs_direct_req *dreq = container_of(kref, struct nfs_direct_req, kref);
+
+	if (dreq->ctx != NULL)
+		put_nfs_open_context(dreq->ctx);
 	kmem_cache_free(nfs_direct_cachep, dreq);
 }
 
-/**
- * nfs_direct_read_alloc - allocate nfs_read_data structures for direct read
- * @count: count of bytes for the read request
- * @rsize: local rsize setting
+/*
+ * Collects and returns the final error value/byte-count.
+ */
+static ssize_t nfs_direct_wait(struct nfs_direct_req *dreq)
+{
+	ssize_t result = -EIOCBQUEUED;
+
+	/* Async requests don't wait here */
+	if (dreq->iocb)
+		goto out;
+
+	result = wait_for_completion_interruptible(&dreq->completion);
+
+	if (!result)
+		result = dreq->error;
+	if (!result)
+		result = dreq->count;
+
+out:
+	kref_put(&dreq->kref, nfs_direct_req_release);
+	return (ssize_t) result;
+}
+
+/*
+ * We must hold a reference to all the pages in this direct read request
+ * until the RPCs complete.  This could be long *after* we are woken up in
+ * nfs_direct_wait (for instance, if someone hits ^C on a slow server).
  *
+ * In addition, synchronous I/O uses a stack-allocated iocb.  Thus we
+ * can't trust the iocb is still valid here if this is a synchronous
+ * request.  If the waiter is woken prematurely, the iocb is long gone.
+ */
+static void nfs_direct_complete(struct nfs_direct_req *dreq)
+{
+	nfs_free_user_pages(dreq->pages, dreq->npages, 1);
+
+	if (dreq->iocb) {
+		long res = (long) dreq->error;
+		if (!res)
+			res = (long) dreq->count;
+		aio_complete(dreq->iocb, res, 0);
+	}
+	complete_all(&dreq->completion);
+
+	kref_put(&dreq->kref, nfs_direct_req_release);
+}
+
+/*
  * Note we also set the number of requests we have in the dreq when we are
  * done.  This prevents races with I/O completion so we will always wait
  * until all requests have been dispatched and completed.
  */
-static struct nfs_direct_req *nfs_direct_read_alloc(size_t nbytes, unsigned int rsize)
+static struct nfs_direct_req *nfs_direct_read_alloc(size_t nbytes, size_t rsize)
 {
 	struct list_head *list;
 	struct nfs_direct_req *dreq;
-	unsigned int reads = 0;
 	unsigned int rpages = (rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
 
-	dreq = kmem_cache_alloc(nfs_direct_cachep, SLAB_KERNEL);
+	dreq = nfs_direct_req_alloc();
 	if (!dreq)
 		return NULL;
 
-	kref_init(&dreq->kref);
-	init_waitqueue_head(&dreq->wait);
-	INIT_LIST_HEAD(&dreq->list);
-	atomic_set(&dreq->count, 0);
-	atomic_set(&dreq->error, 0);
-
 	list = &dreq->list;
 	for(;;) {
 		struct nfs_read_data *data = nfs_readdata_alloc(rpages);
@@ -196,72 +279,70 @@ static struct nfs_direct_req *nfs_direct_read_alloc(size_t nbytes, unsigned int
 		list_add(&data->pages, list);
 
 		data->req = (struct nfs_page *) dreq;
-		reads++;
+		dreq->outstanding++;
 		if (nbytes <= rsize)
 			break;
 		nbytes -= rsize;
 	}
 	kref_get(&dreq->kref);
-	atomic_set(&dreq->complete, reads);
 	return dreq;
 }
 
-/**
- * nfs_direct_read_result - handle a read reply for a direct read request
- * @data: address of NFS READ operation control block
- * @status: status of this NFS READ operation
- *
- * We must hold a reference to all the pages in this direct read request
- * until the RPCs complete.  This could be long *after* we are woken up in
- * nfs_direct_read_wait (for instance, if someone hits ^C on a slow server).
- */
-static void nfs_direct_read_result(struct nfs_read_data *data, int status)
+static void nfs_direct_read_result(struct rpc_task *task, void *calldata)
 {
+	struct nfs_read_data *data = calldata;
 	struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req;
 
-	if (likely(status >= 0))
-		atomic_add(data->res.count, &dreq->count);
+	if (nfs_readpage_result(task, data) != 0)
+		return;
+
+	spin_lock(&dreq->lock);
+
+	if (likely(task->tk_status >= 0))
+		dreq->count += data->res.count;
 	else
-		atomic_set(&dreq->error, status);
+		dreq->error = task->tk_status;
 
-	if (unlikely(atomic_dec_and_test(&dreq->complete))) {
-		nfs_free_user_pages(dreq->pages, dreq->npages, 1);
-		wake_up(&dreq->wait);
-		kref_put(&dreq->kref, nfs_direct_req_release);
+	if (--dreq->outstanding) {
+		spin_unlock(&dreq->lock);
+		return;
 	}
+
+	spin_unlock(&dreq->lock);
+	nfs_direct_complete(dreq);
 }
 
-/**
- * nfs_direct_read_schedule - dispatch NFS READ operations for a direct read
- * @dreq: address of nfs_direct_req struct for this request
- * @inode: target inode
- * @ctx: target file open context
- * @user_addr: starting address of this segment of user's buffer
- * @count: size of this segment
- * @file_offset: offset in file to begin the operation
- *
+static const struct rpc_call_ops nfs_read_direct_ops = {
+	.rpc_call_done = nfs_direct_read_result,
+	.rpc_release = nfs_readdata_release,
+};
+
+/*
  * For each nfs_read_data struct that was allocated on the list, dispatch
  * an NFS READ operation
  */
-static void nfs_direct_read_schedule(struct nfs_direct_req *dreq,
-		struct inode *inode, struct nfs_open_context *ctx,
-		unsigned long user_addr, size_t count, loff_t file_offset)
+static void nfs_direct_read_schedule(struct nfs_direct_req *dreq)
 {
+	struct nfs_open_context *ctx = dreq->ctx;
+	struct inode *inode = ctx->dentry->d_inode;
 	struct list_head *list = &dreq->list;
 	struct page **pages = dreq->pages;
+	size_t count = dreq->user_count;
+	loff_t pos = dreq->pos;
+	size_t rsize = NFS_SERVER(inode)->rsize;
 	unsigned int curpage, pgbase;
-	unsigned int rsize = NFS_SERVER(inode)->rsize;
 
 	curpage = 0;
-	pgbase = user_addr & ~PAGE_MASK;
+	pgbase = dreq->user_addr & ~PAGE_MASK;
 	do {
 		struct nfs_read_data *data;
-		unsigned int bytes;
+		size_t bytes;
 
 		bytes = rsize;
 		if (count < rsize)
 			bytes = count;
 
+		BUG_ON(list_empty(list));
 		data = list_entry(list->next, struct nfs_read_data, pages);
 		list_del_init(&data->pages);
 
@@ -269,7 +350,7 @@ static void nfs_direct_read_schedule(struct nfs_direct_req *dreq,
 		data->cred = ctx->cred;
 		data->args.fh = NFS_FH(inode);
 		data->args.context = ctx;
-		data->args.offset = file_offset;
+		data->args.offset = pos;
 		data->args.pgbase = pgbase;
 		data->args.pages = &pages[curpage];
 		data->args.count = bytes;
@@ -277,77 +358,38 @@ static void nfs_direct_read_schedule(struct nfs_direct_req *dreq,
 		data->res.eof = 0;
 		data->res.count = bytes;
 
+		rpc_init_task(&data->task, NFS_CLIENT(inode), RPC_TASK_ASYNC,
+				&nfs_read_direct_ops, data);
 		NFS_PROTO(inode)->read_setup(data);
 
 		data->task.tk_cookie = (unsigned long) inode;
-		data->complete = nfs_direct_read_result;
 
 		lock_kernel();
 		rpc_execute(&data->task);
 		unlock_kernel();
 
-		dfprintk(VFS, "NFS: %4d initiated direct read call (req %s/%Ld, %u bytes @ offset %Lu)\n",
+		dfprintk(VFS, "NFS: %5u initiated direct read call (req %s/%Ld, %zu bytes @ offset %Lu)\n",
 				data->task.tk_pid,
 				inode->i_sb->s_id,
 				(long long)NFS_FILEID(inode),
 				bytes,
 				(unsigned long long)data->args.offset);
 
-		file_offset += bytes;
+		pos += bytes;
 		pgbase += bytes;
 		curpage += pgbase >> PAGE_SHIFT;
 		pgbase &= ~PAGE_MASK;
 
 		count -= bytes;
 	} while (count != 0);
+	BUG_ON(!list_empty(list));
 }
 
-/**
- * nfs_direct_read_wait - wait for I/O completion for direct reads
- * @dreq: request on which we are to wait
- * @intr: whether or not this wait can be interrupted
- *
- * Collects and returns the final error value/byte-count.
- */
-static ssize_t nfs_direct_read_wait(struct nfs_direct_req *dreq, int intr)
-{
-	int result = 0;
-
-	if (intr) {
-		result = wait_event_interruptible(dreq->wait,
-					(atomic_read(&dreq->complete) == 0));
-	} else {
-		wait_event(dreq->wait, (atomic_read(&dreq->complete) == 0));
-	}
-
-	if (!result)
-		result = atomic_read(&dreq->error);
-	if (!result)
-		result = atomic_read(&dreq->count);
-
-	kref_put(&dreq->kref, nfs_direct_req_release);
-	return (ssize_t) result;
-}
-
-/**
- * nfs_direct_read_seg - Read in one iov segment.  Generate separate
- *                        read RPCs for each "rsize" bytes.
- * @inode: target inode
- * @ctx: target file open context
- * @user_addr: starting address of this segment of user's buffer
- * @count: size of this segment
- * @file_offset: offset in file to begin the operation
- * @pages: array of addresses of page structs defining user's buffer
- * @nr_pages: number of pages in the array
- *
- */
-static ssize_t nfs_direct_read_seg(struct inode *inode,
-		struct nfs_open_context *ctx, unsigned long user_addr,
-		size_t count, loff_t file_offset, struct page **pages,
-		unsigned int nr_pages)
+static ssize_t nfs_direct_read(struct kiocb *iocb, unsigned long user_addr, size_t count, loff_t pos, struct page **pages, unsigned int nr_pages)
 {
 	ssize_t result;
 	sigset_t oldset;
+	struct inode *inode = iocb->ki_filp->f_mapping->host;
 	struct rpc_clnt *clnt = NFS_CLIENT(inode);
 	struct nfs_direct_req *dreq;
 
@@ -355,284 +397,350 @@ static ssize_t nfs_direct_read_seg(struct inode *inode,
 	if (!dreq)
 		return -ENOMEM;
 
+	dreq->user_addr = user_addr;
+	dreq->user_count = count;
+	dreq->pos = pos;
 	dreq->pages = pages;
 	dreq->npages = nr_pages;
+	dreq->inode = inode;
+	dreq->ctx = get_nfs_open_context((struct nfs_open_context *)iocb->ki_filp->private_data);
+	if (!is_sync_kiocb(iocb))
+		dreq->iocb = iocb;
 
+	nfs_add_stats(inode, NFSIOS_DIRECTREADBYTES, count);
 	rpc_clnt_sigmask(clnt, &oldset);
-	nfs_direct_read_schedule(dreq, inode, ctx, user_addr, count,
-				 file_offset);
-	result = nfs_direct_read_wait(dreq, clnt->cl_intr);
+	nfs_direct_read_schedule(dreq);
+	result = nfs_direct_wait(dreq);
 	rpc_clnt_sigunmask(clnt, &oldset);
 
 	return result;
 }
 
-/**
- * nfs_direct_read - For each iov segment, map the user's buffer
- *                   then generate read RPCs.
- * @inode: target inode
- * @ctx: target file open context
- * @iov: array of vectors that define I/O buffer
- * file_offset: offset in file to begin the operation
- * nr_segs: size of iovec array
- *
- * We've already pushed out any non-direct writes so that this read
- * will see them when we read from the server.
- */
-static ssize_t
-nfs_direct_read(struct inode *inode, struct nfs_open_context *ctx,
-		const struct iovec *iov, loff_t file_offset,
-		unsigned long nr_segs)
+static void nfs_direct_free_writedata(struct nfs_direct_req *dreq)
 {
-	ssize_t tot_bytes = 0;
-	unsigned long seg = 0;
-
-	while ((seg < nr_segs) && (tot_bytes >= 0)) {
-		ssize_t result;
-		int page_count;
-		struct page **pages;
-		const struct iovec *vec = &iov[seg++];
-		unsigned long user_addr = (unsigned long) vec->iov_base;
-		size_t size = vec->iov_len;
-
-                page_count = nfs_get_user_pages(READ, user_addr, size, &pages);
-                if (page_count < 0) {
-                        nfs_free_user_pages(pages, 0, 0);
-			if (tot_bytes > 0)
-				break;
-                        return page_count;
-                }
+	list_splice_init(&dreq->rewrite_list, &dreq->list);
+	while (!list_empty(&dreq->list)) {
+		struct nfs_write_data *data = list_entry(dreq->list.next, struct nfs_write_data, pages);
+		list_del(&data->pages);
+		nfs_writedata_release(data);
+	}
+}
 
-		result = nfs_direct_read_seg(inode, ctx, user_addr, size,
-				file_offset, pages, page_count);
+#if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
+static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq)
+{
+	struct list_head *pos;
 
-		if (result <= 0) {
-			if (tot_bytes > 0)
-				break;
-			return result;
-		}
-		tot_bytes += result;
-		file_offset += result;
-		if (result < size)
-			break;
+	list_splice_init(&dreq->rewrite_list, &dreq->list);
+	list_for_each(pos, &dreq->list)
+		dreq->outstanding++;
+	dreq->count = 0;
+
+	nfs_direct_write_schedule(dreq, FLUSH_STABLE);
+}
+
+static void nfs_direct_commit_result(struct rpc_task *task, void *calldata)
+{
+	struct nfs_write_data *data = calldata;
+	struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req;
+
+	/* Call the NFS version-specific code */
+	if (NFS_PROTO(data->inode)->commit_done(task, data) != 0)
+		return;
+	if (unlikely(task->tk_status < 0)) {
+		dreq->error = task->tk_status;
+		dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
+	}
+	if (memcmp(&dreq->verf, &data->verf, sizeof(data->verf))) {
+		dprintk("NFS: %5u commit verify failed\n", task->tk_pid);
+		dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
 	}
 
-	return tot_bytes;
+	dprintk("NFS: %5u commit returned %d\n", task->tk_pid, task->tk_status);
+	nfs_direct_write_complete(dreq, data->inode);
 }
 
-/**
- * nfs_direct_write_seg - Write out one iov segment.  Generate separate
- *                        write RPCs for each "wsize" bytes, then commit.
- * @inode: target inode
- * @ctx: target file open context
- * user_addr: starting address of this segment of user's buffer
- * count: size of this segment
- * file_offset: offset in file to begin the operation
- * @pages: array of addresses of page structs defining user's buffer
- * nr_pages: size of pages array
- */
-static ssize_t nfs_direct_write_seg(struct inode *inode,
-		struct nfs_open_context *ctx, unsigned long user_addr,
-		size_t count, loff_t file_offset, struct page **pages,
-		int nr_pages)
+static const struct rpc_call_ops nfs_commit_direct_ops = {
+	.rpc_call_done = nfs_direct_commit_result,
+	.rpc_release = nfs_commit_release,
+};
+
+static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq)
 {
-	const unsigned int wsize = NFS_SERVER(inode)->wsize;
-	size_t request;
-	int curpage, need_commit;
-	ssize_t result, tot_bytes;
-	struct nfs_writeverf first_verf;
-	struct nfs_write_data *wdata;
-
-	wdata = nfs_writedata_alloc(NFS_SERVER(inode)->wpages);
-	if (!wdata)
-		return -ENOMEM;
+	struct nfs_write_data *data = dreq->commit_data;
+	struct rpc_task *task = &data->task;
 
-	wdata->inode = inode;
-	wdata->cred = ctx->cred;
-	wdata->args.fh = NFS_FH(inode);
-	wdata->args.context = ctx;
-	wdata->args.stable = NFS_UNSTABLE;
-	if (IS_SYNC(inode) || NFS_PROTO(inode)->version == 2 || count <= wsize)
-		wdata->args.stable = NFS_FILE_SYNC;
-	wdata->res.fattr = &wdata->fattr;
-	wdata->res.verf = &wdata->verf;
+	data->inode = dreq->inode;
+	data->cred = dreq->ctx->cred;
 
-	nfs_begin_data_update(inode);
-retry:
-	need_commit = 0;
-	tot_bytes = 0;
-	curpage = 0;
-	request = count;
-	wdata->args.pgbase = user_addr & ~PAGE_MASK;
-	wdata->args.offset = file_offset;
-	do {
-		wdata->args.count = request;
-		if (wdata->args.count > wsize)
-			wdata->args.count = wsize;
-		wdata->args.pages = &pages[curpage];
+	data->args.fh = NFS_FH(data->inode);
+	data->args.offset = dreq->pos;
+	data->args.count = dreq->user_count;
+	data->res.count = 0;
+	data->res.fattr = &data->fattr;
+	data->res.verf = &data->verf;
 
-		dprintk("NFS: direct write: c=%u o=%Ld ua=%lu, pb=%u, cp=%u\n",
-			wdata->args.count, (long long) wdata->args.offset,
-			user_addr + tot_bytes, wdata->args.pgbase, curpage);
+	rpc_init_task(&data->task, NFS_CLIENT(dreq->inode), RPC_TASK_ASYNC,
+				&nfs_commit_direct_ops, data);
+	NFS_PROTO(data->inode)->commit_setup(data, 0);
 
-		lock_kernel();
-		result = NFS_PROTO(inode)->write(wdata);
-		unlock_kernel();
+	data->task.tk_priority = RPC_PRIORITY_NORMAL;
+	data->task.tk_cookie = (unsigned long)data->inode;
+	/* Note: task.tk_ops->rpc_release will free dreq->commit_data */
+	dreq->commit_data = NULL;
 
-		if (result <= 0) {
-			if (tot_bytes > 0)
-				break;
-			goto out;
-		}
+	dprintk("NFS: %5u initiated commit call\n", task->tk_pid);
 
-		if (tot_bytes == 0)
-			memcpy(&first_verf.verifier, &wdata->verf.verifier,
-						sizeof(first_verf.verifier));
-		if (wdata->verf.committed != NFS_FILE_SYNC) {
-			need_commit = 1;
-			if (memcmp(&first_verf.verifier, &wdata->verf.verifier,
-					sizeof(first_verf.verifier)))
-				goto sync_retry;
-		}
+	lock_kernel();
+	rpc_execute(&data->task);
+	unlock_kernel();
+}
 
-		tot_bytes += result;
+static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode)
+{
+	int flags = dreq->flags;
 
-		/* in case of a short write: stop now, let the app recover */
-		if (result < wdata->args.count)
+	dreq->flags = 0;
+	switch (flags) {
+		case NFS_ODIRECT_DO_COMMIT:
+			nfs_direct_commit_schedule(dreq);
 			break;
+		case NFS_ODIRECT_RESCHED_WRITES:
+			nfs_direct_write_reschedule(dreq);
+			break;
+		default:
+			nfs_end_data_update(inode);
+			if (dreq->commit_data != NULL)
+				nfs_commit_free(dreq->commit_data);
+			nfs_direct_free_writedata(dreq);
+			nfs_direct_complete(dreq);
+	}
+}
 
-		wdata->args.offset += result;
-		wdata->args.pgbase += result;
-		curpage += wdata->args.pgbase >> PAGE_SHIFT;
-		wdata->args.pgbase &= ~PAGE_MASK;
-		request -= result;
-	} while (request != 0);
+static void nfs_alloc_commit_data(struct nfs_direct_req *dreq)
+{
+	dreq->commit_data = nfs_commit_alloc(0);
+	if (dreq->commit_data != NULL)
+		dreq->commit_data->req = (struct nfs_page *) dreq;
+}
+#else
+static inline void nfs_alloc_commit_data(struct nfs_direct_req *dreq)
+{
+	dreq->commit_data = NULL;
+}
 
-	/*
-	 * Commit data written so far, even in the event of an error
-	 */
-	if (need_commit) {
-		wdata->args.count = tot_bytes;
-		wdata->args.offset = file_offset;
+static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode *inode)
+{
+	nfs_end_data_update(inode);
+	nfs_direct_free_writedata(dreq);
+	nfs_direct_complete(dreq);
+}
+#endif
 
-		lock_kernel();
-		result = NFS_PROTO(inode)->commit(wdata);
-		unlock_kernel();
+static struct nfs_direct_req *nfs_direct_write_alloc(size_t nbytes, size_t wsize)
+{
+	struct list_head *list;
+	struct nfs_direct_req *dreq;
+	unsigned int wpages = (wsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+
+	dreq = nfs_direct_req_alloc();
+	if (!dreq)
+		return NULL;
+
+	list = &dreq->list;
+	for(;;) {
+		struct nfs_write_data *data = nfs_writedata_alloc(wpages);
 
-		if (result < 0 || memcmp(&first_verf.verifier,
-					 &wdata->verf.verifier,
-					 sizeof(first_verf.verifier)) != 0)
-			goto sync_retry;
+		if (unlikely(!data)) {
+			while (!list_empty(list)) {
+				data = list_entry(list->next,
+						  struct nfs_write_data, pages);
+				list_del(&data->pages);
+				nfs_writedata_free(data);
+			}
+			kref_put(&dreq->kref, nfs_direct_req_release);
+			return NULL;
+		}
+
+		INIT_LIST_HEAD(&data->pages);
+		list_add(&data->pages, list);
+
+		data->req = (struct nfs_page *) dreq;
+		dreq->outstanding++;
+		if (nbytes <= wsize)
+			break;
+		nbytes -= wsize;
 	}
-	result = tot_bytes;
 
-out:
-	nfs_end_data_update(inode);
-	nfs_writedata_free(wdata);
-	return result;
+	nfs_alloc_commit_data(dreq);
 
-sync_retry:
-	wdata->args.stable = NFS_FILE_SYNC;
-	goto retry;
+	kref_get(&dreq->kref);
+	return dreq;
 }
 
-/**
- * nfs_direct_write - For each iov segment, map the user's buffer
- *                    then generate write and commit RPCs.
- * @inode: target inode
- * @ctx: target file open context
- * @iov: array of vectors that define I/O buffer
- * file_offset: offset in file to begin the operation
- * nr_segs: size of iovec array
- *
- * Upon return, generic_file_direct_IO invalidates any cached pages
- * that non-direct readers might access, so they will pick up these
- * writes immediately.
- */
-static ssize_t nfs_direct_write(struct inode *inode,
-		struct nfs_open_context *ctx, const struct iovec *iov,
-		loff_t file_offset, unsigned long nr_segs)
+static void nfs_direct_write_result(struct rpc_task *task, void *calldata)
 {
-	ssize_t tot_bytes = 0;
-	unsigned long seg = 0;
-
-	while ((seg < nr_segs) && (tot_bytes >= 0)) {
-		ssize_t result;
-		int page_count;
-		struct page **pages;
-		const struct iovec *vec = &iov[seg++];
-		unsigned long user_addr = (unsigned long) vec->iov_base;
-		size_t size = vec->iov_len;
-
-                page_count = nfs_get_user_pages(WRITE, user_addr, size, &pages);
-                if (page_count < 0) {
-                        nfs_free_user_pages(pages, 0, 0);
-			if (tot_bytes > 0)
-				break;
-                        return page_count;
-                }
+	struct nfs_write_data *data = calldata;
+	struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req;
+	int status = task->tk_status;
+
+	if (nfs_writeback_done(task, data) != 0)
+		return;
+
+	spin_lock(&dreq->lock);
 
-		result = nfs_direct_write_seg(inode, ctx, user_addr, size,
-				file_offset, pages, page_count);
-		nfs_free_user_pages(pages, page_count, 0);
+	if (likely(status >= 0))
+		dreq->count += data->res.count;
+	else
+		dreq->error = task->tk_status;
 
-		if (result <= 0) {
-			if (tot_bytes > 0)
+	if (data->res.verf->committed != NFS_FILE_SYNC) {
+		switch (dreq->flags) {
+			case 0:
+				memcpy(&dreq->verf, &data->verf, sizeof(dreq->verf));
+				dreq->flags = NFS_ODIRECT_DO_COMMIT;
 				break;
-			return result;
+			case NFS_ODIRECT_DO_COMMIT:
+				if (memcmp(&dreq->verf, &data->verf, sizeof(dreq->verf))) {
+					dprintk("NFS: %5u write verify failed\n", task->tk_pid);
+					dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
+				}
 		}
-		tot_bytes += result;
-		file_offset += result;
-		if (result < size)
-			break;
 	}
-	return tot_bytes;
+	/* In case we have to resend */
+	data->args.stable = NFS_FILE_SYNC;
+
+	spin_unlock(&dreq->lock);
 }
 
-/**
- * nfs_direct_IO - NFS address space operation for direct I/O
- * rw: direction (read or write)
- * @iocb: target I/O control block
- * @iov: array of vectors that define I/O buffer
- * file_offset: offset in file to begin the operation
- * nr_segs: size of iovec array
- *
+/*
+ * NB: Return the value of the first error return code.  Subsequent
+ *     errors after the first one are ignored.
  */
-ssize_t
-nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
-		loff_t file_offset, unsigned long nr_segs)
+static void nfs_direct_write_release(void *calldata)
 {
-	ssize_t result = -EINVAL;
-	struct file *file = iocb->ki_filp;
-	struct nfs_open_context *ctx;
-	struct dentry *dentry = file->f_dentry;
-	struct inode *inode = dentry->d_inode;
+	struct nfs_write_data *data = calldata;
+	struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req;
 
-	/*
-	 * No support for async yet
-	 */
-	if (!is_sync_kiocb(iocb))
-		return result;
-
-	ctx = (struct nfs_open_context *)file->private_data;
-	switch (rw) {
-	case READ:
-		dprintk("NFS: direct_IO(read) (%s) off/no(%Lu/%lu)\n",
-				dentry->d_name.name, file_offset, nr_segs);
-
-		result = nfs_direct_read(inode, ctx, iov,
-						file_offset, nr_segs);
-		break;
-	case WRITE:
-		dprintk("NFS: direct_IO(write) (%s) off/no(%Lu/%lu)\n",
-				dentry->d_name.name, file_offset, nr_segs);
-
-		result = nfs_direct_write(inode, ctx, iov,
-						file_offset, nr_segs);
-		break;
-	default:
-		break;
+	spin_lock(&dreq->lock);
+	if (--dreq->outstanding) {
+		spin_unlock(&dreq->lock);
+		return;
 	}
+	spin_unlock(&dreq->lock);
+
+	nfs_direct_write_complete(dreq, data->inode);
+}
+
+static const struct rpc_call_ops nfs_write_direct_ops = {
+	.rpc_call_done = nfs_direct_write_result,
+	.rpc_release = nfs_direct_write_release,
+};
+
+/*
+ * For each nfs_write_data struct that was allocated on the list, dispatch
+ * an NFS WRITE operation
+ */
+static void nfs_direct_write_schedule(struct nfs_direct_req *dreq, int sync)
+{
+	struct nfs_open_context *ctx = dreq->ctx;
+	struct inode *inode = ctx->dentry->d_inode;
+	struct list_head *list = &dreq->list;
+	struct page **pages = dreq->pages;
+	size_t count = dreq->user_count;
+	loff_t pos = dreq->pos;
+	size_t wsize = NFS_SERVER(inode)->wsize;
+	unsigned int curpage, pgbase;
+
+	curpage = 0;
+	pgbase = dreq->user_addr & ~PAGE_MASK;
+	do {
+		struct nfs_write_data *data;
+		size_t bytes;
+
+		bytes = wsize;
+		if (count < wsize)
+			bytes = count;
+
+		BUG_ON(list_empty(list));
+		data = list_entry(list->next, struct nfs_write_data, pages);
+		list_move_tail(&data->pages, &dreq->rewrite_list);
+
+		data->inode = inode;
+		data->cred = ctx->cred;
+		data->args.fh = NFS_FH(inode);
+		data->args.context = ctx;
+		data->args.offset = pos;
+		data->args.pgbase = pgbase;
+		data->args.pages = &pages[curpage];
+		data->args.count = bytes;
+		data->res.fattr = &data->fattr;
+		data->res.count = bytes;
+		data->res.verf = &data->verf;
+
+		rpc_init_task(&data->task, NFS_CLIENT(inode), RPC_TASK_ASYNC,
+				&nfs_write_direct_ops, data);
+		NFS_PROTO(inode)->write_setup(data, sync);
+
+		data->task.tk_priority = RPC_PRIORITY_NORMAL;
+		data->task.tk_cookie = (unsigned long) inode;
+
+		lock_kernel();
+		rpc_execute(&data->task);
+		unlock_kernel();
+
+		dfprintk(VFS, "NFS: %5u initiated direct write call (req %s/%Ld, %zu bytes @ offset %Lu)\n",
+				data->task.tk_pid,
+				inode->i_sb->s_id,
+				(long long)NFS_FILEID(inode),
+				bytes,
+				(unsigned long long)data->args.offset);
+
+		pos += bytes;
+		pgbase += bytes;
+		curpage += pgbase >> PAGE_SHIFT;
+		pgbase &= ~PAGE_MASK;
+
+		count -= bytes;
+	} while (count != 0);
+	BUG_ON(!list_empty(list));
+}
+
+static ssize_t nfs_direct_write(struct kiocb *iocb, unsigned long user_addr, size_t count, loff_t pos, struct page **pages, int nr_pages)
+{
+	ssize_t result;
+	sigset_t oldset;
+	struct inode *inode = iocb->ki_filp->f_mapping->host;
+	struct rpc_clnt *clnt = NFS_CLIENT(inode);
+	struct nfs_direct_req *dreq;
+	size_t wsize = NFS_SERVER(inode)->wsize;
+	int sync = 0;
+
+	dreq = nfs_direct_write_alloc(count, wsize);
+	if (!dreq)
+		return -ENOMEM;
+	if (dreq->commit_data == NULL || count < wsize)
+		sync = FLUSH_STABLE;
+
+	dreq->user_addr = user_addr;
+	dreq->user_count = count;
+	dreq->pos = pos;
+	dreq->pages = pages;
+	dreq->npages = nr_pages;
+	dreq->inode = inode;
+	dreq->ctx = get_nfs_open_context((struct nfs_open_context *)iocb->ki_filp->private_data);
+	if (!is_sync_kiocb(iocb))
+		dreq->iocb = iocb;
+
+	nfs_add_stats(inode, NFSIOS_DIRECTWRITTENBYTES, count);
+
+	nfs_begin_data_update(inode);
+
+	rpc_clnt_sigmask(clnt, &oldset);
+	nfs_direct_write_schedule(dreq, sync);
+	result = nfs_direct_wait(dreq);
+	rpc_clnt_sigunmask(clnt, &oldset);
+
 	return result;
 }
 
@@ -640,49 +748,40 @@ nfs_direct_IO(int rw, struct kiocb *iocb, const struct iovec *iov,
  * nfs_file_direct_read - file direct read operation for NFS files
  * @iocb: target I/O control block
  * @buf: user's buffer into which to read data
- * count: number of bytes to read
- * pos: byte offset in file where reading starts
+ * @count: number of bytes to read
+ * @pos: byte offset in file where reading starts
  *
  * We use this function for direct reads instead of calling
  * generic_file_aio_read() in order to avoid gfar's check to see if
  * the request starts before the end of the file.  For that check
  * to work, we must generate a GETATTR before each direct read, and
  * even then there is a window between the GETATTR and the subsequent
- * READ where the file size could change.  So our preference is simply
+ * READ where the file size could change.  Our preference is simply
  * to do all reads the application wants, and the server will take
  * care of managing the end of file boundary.
- * 
+ *
  * This function also eliminates unnecessarily updating the file's
  * atime locally, as the NFS server sets the file's atime, and this
  * client must read the updated atime from the server back into its
  * cache.
  */
-ssize_t
-nfs_file_direct_read(struct kiocb *iocb, char __user *buf, size_t count, loff_t pos)
+ssize_t nfs_file_direct_read(struct kiocb *iocb, char __user *buf, size_t count, loff_t pos)
 {
 	ssize_t retval = -EINVAL;
-	loff_t *ppos = &iocb->ki_pos;
+	int page_count;
+	struct page **pages;
 	struct file *file = iocb->ki_filp;
-	struct nfs_open_context *ctx =
-			(struct nfs_open_context *) file->private_data;
 	struct address_space *mapping = file->f_mapping;
-	struct inode *inode = mapping->host;
-	struct iovec iov = {
-		.iov_base = buf,
-		.iov_len = count,
-	};
 
 	dprintk("nfs: direct read(%s/%s, %lu@%Ld)\n",
 		file->f_dentry->d_parent->d_name.name,
 		file->f_dentry->d_name.name,
 		(unsigned long) count, (long long) pos);
 
-	if (!is_sync_kiocb(iocb))
-		goto out;
 	if (count < 0)
 		goto out;
 	retval = -EFAULT;
-	if (!access_ok(VERIFY_WRITE, iov.iov_base, iov.iov_len))
+	if (!access_ok(VERIFY_WRITE, buf, count))
 		goto out;
 	retval = 0;
 	if (!count)
@@ -692,9 +791,16 @@ nfs_file_direct_read(struct kiocb *iocb, char __user *buf, size_t count, loff_t
 	if (retval)
 		goto out;
 
-	retval = nfs_direct_read(inode, ctx, &iov, pos, 1);
+	retval = nfs_get_user_pages(READ, (unsigned long) buf,
+						count, &pages);
+	if (retval < 0)
+		goto out;
+	page_count = retval;
+
+	retval = nfs_direct_read(iocb, (unsigned long) buf, count, pos,
+						pages, page_count);
 	if (retval > 0)
-		*ppos = pos + retval;
+		iocb->ki_pos = pos + retval;
 
 out:
 	return retval;
@@ -704,8 +810,8 @@ out:
  * nfs_file_direct_write - file direct write operation for NFS files
  * @iocb: target I/O control block
  * @buf: user's buffer from which to write data
- * count: number of bytes to write
- * pos: byte offset in file where writing starts
+ * @count: number of bytes to write
+ * @pos: byte offset in file where writing starts
  *
  * We use this function for direct writes instead of calling
  * generic_file_aio_write() in order to avoid taking the inode
@@ -725,28 +831,19 @@ out:
  * Note that O_APPEND is not supported for NFS direct writes, as there
  * is no atomic O_APPEND write facility in the NFS protocol.
  */
-ssize_t
-nfs_file_direct_write(struct kiocb *iocb, const char __user *buf, size_t count, loff_t pos)
+ssize_t nfs_file_direct_write(struct kiocb *iocb, const char __user *buf, size_t count, loff_t pos)
 {
 	ssize_t retval;
+	int page_count;
+	struct page **pages;
 	struct file *file = iocb->ki_filp;
-	struct nfs_open_context *ctx =
-			(struct nfs_open_context *) file->private_data;
 	struct address_space *mapping = file->f_mapping;
-	struct inode *inode = mapping->host;
-	struct iovec iov = {
-		.iov_base = (char __user *)buf,
-	};
 
 	dfprintk(VFS, "nfs: direct write(%s/%s, %lu@%Ld)\n",
 		file->f_dentry->d_parent->d_name.name,
 		file->f_dentry->d_name.name,
 		(unsigned long) count, (long long) pos);
 
-	retval = -EINVAL;
-	if (!is_sync_kiocb(iocb))
-		goto out;
-
 	retval = generic_write_checks(file, &pos, &count, 0);
 	if (retval)
 		goto out;
@@ -757,19 +854,35 @@ nfs_file_direct_write(struct kiocb *iocb, const char __user *buf, size_t count,
 	retval = 0;
 	if (!count)
 		goto out;
-	iov.iov_len = count,
 
 	retval = -EFAULT;
-	if (!access_ok(VERIFY_READ, iov.iov_base, iov.iov_len))
+	if (!access_ok(VERIFY_READ, buf, count))
 		goto out;
 
 	retval = nfs_sync_mapping(mapping);
 	if (retval)
 		goto out;
 
-	retval = nfs_direct_write(inode, ctx, &iov, pos, 1);
+	retval = nfs_get_user_pages(WRITE, (unsigned long) buf,
+						count, &pages);
+	if (retval < 0)
+		goto out;
+	page_count = retval;
+
+	retval = nfs_direct_write(iocb, (unsigned long) buf, count,
+					pos, pages, page_count);
+
+	/*
+	 * XXX: nfs_end_data_update() already ensures this file's
+	 *      cached data is subsequently invalidated.  Do we really
+	 *      need to call invalidate_inode_pages2() again here?
+	 *
+	 *      For aio writes, this invalidation will almost certainly
+	 *      occur before the writes complete.  Kind of racey.
+	 */
 	if (mapping->nrpages)
 		invalidate_inode_pages2(mapping);
+
 	if (retval > 0)
 		iocb->ki_pos = pos + retval;
 
@@ -777,11 +890,16 @@ out:
 	return retval;
 }
 
+/**
+ * nfs_init_directcache - create a slab cache for nfs_direct_req structures
+ *
+ */
 int nfs_init_directcache(void)
 {
 	nfs_direct_cachep = kmem_cache_create("nfs_direct_cache",
 						sizeof(struct nfs_direct_req),
-						0, SLAB_RECLAIM_ACCOUNT,
+						0, (SLAB_RECLAIM_ACCOUNT|
+							SLAB_MEM_SPREAD),
 						NULL, NULL);
 	if (nfs_direct_cachep == NULL)
 		return -ENOMEM;
@@ -789,6 +907,10 @@ int nfs_init_directcache(void)
 	return 0;
 }
 
+/**
+ * nfs_init_directcache - destroy the slab cache for nfs_direct_req structures
+ *
+ */
 void nfs_destroy_directcache(void)
 {
 	if (kmem_cache_destroy(nfs_direct_cachep))
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 7a79fbe9f539..f1df2c8d9259 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -32,6 +32,7 @@
 #include <asm/system.h>
 
 #include "delegation.h"
+#include "iostat.h"
 
 #define NFSDBG_FACILITY		NFSDBG_FILE
 
@@ -48,7 +49,7 @@ static int nfs_check_flags(int flags);
 static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl);
 static int nfs_flock(struct file *filp, int cmd, struct file_lock *fl);
 
-struct file_operations nfs_file_operations = {
+const struct file_operations nfs_file_operations = {
 	.llseek		= nfs_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
@@ -102,18 +103,15 @@ static int nfs_check_flags(int flags)
 static int
 nfs_file_open(struct inode *inode, struct file *filp)
 {
-	struct nfs_server *server = NFS_SERVER(inode);
-	int (*open)(struct inode *, struct file *);
 	int res;
 
 	res = nfs_check_flags(filp->f_flags);
 	if (res)
 		return res;
 
+	nfs_inc_stats(inode, NFSIOS_VFSOPEN);
 	lock_kernel();
-	/* Do NFSv4 open() call */
-	if ((open = server->rpc_ops->file_open) != NULL)
-		res = open(inode, filp);
+	res = NFS_SERVER(inode)->rpc_ops->file_open(inode, filp);
 	unlock_kernel();
 	return res;
 }
@@ -124,6 +122,7 @@ nfs_file_release(struct inode *inode, struct file *filp)
 	/* Ensure that dirty pages are flushed out with the right creds */
 	if (filp->f_mode & FMODE_WRITE)
 		filemap_fdatawrite(filp->f_mapping);
+	nfs_inc_stats(inode, NFSIOS_VFSRELEASE);
 	return NFS_PROTO(inode)->file_release(inode, filp);
 }
 
@@ -199,6 +198,7 @@ nfs_file_flush(struct file *file)
 
 	if ((file->f_mode & FMODE_WRITE) == 0)
 		return 0;
+	nfs_inc_stats(inode, NFSIOS_VFSFLUSH);
 	lock_kernel();
 	/* Ensure that data+attribute caches are up to date after close() */
 	status = nfs_wb_all(inode);
@@ -229,6 +229,7 @@ nfs_file_read(struct kiocb *iocb, char __user * buf, size_t count, loff_t pos)
 		(unsigned long) count, (unsigned long) pos);
 
 	result = nfs_revalidate_file(inode, iocb->ki_filp);
+	nfs_add_stats(inode, NFSIOS_NORMALREADBYTES, count);
 	if (!result)
 		result = generic_file_aio_read(iocb, buf, count, pos);
 	return result;
@@ -282,6 +283,7 @@ nfs_fsync(struct file *file, struct dentry *dentry, int datasync)
 
 	dfprintk(VFS, "nfs: fsync(%s/%ld)\n", inode->i_sb->s_id, inode->i_ino);
 
+	nfs_inc_stats(inode, NFSIOS_VFSFSYNC);
 	lock_kernel();
 	status = nfs_wb_all(inode);
 	if (!status) {
@@ -316,6 +318,16 @@ static int nfs_commit_write(struct file *file, struct page *page, unsigned offse
 	return status;
 }
 
+static void nfs_invalidate_page(struct page *page, unsigned long offset)
+{
+	/* FIXME: we really should cancel any unstarted writes on this page */
+}
+
+static int nfs_release_page(struct page *page, gfp_t gfp)
+{
+	return !nfs_wb_page(page->mapping->host, page);
+}
+
 struct address_space_operations nfs_file_aops = {
 	.readpage = nfs_readpage,
 	.readpages = nfs_readpages,
@@ -324,6 +336,8 @@ struct address_space_operations nfs_file_aops = {
 	.writepages = nfs_writepages,
 	.prepare_write = nfs_prepare_write,
 	.commit_write = nfs_commit_write,
+	.invalidatepage = nfs_invalidate_page,
+	.releasepage = nfs_release_page,
 #ifdef CONFIG_NFS_DIRECTIO
 	.direct_IO = nfs_direct_IO,
 #endif
@@ -365,6 +379,7 @@ nfs_file_write(struct kiocb *iocb, const char __user *buf, size_t count, loff_t
 	if (!count)
 		goto out;
 
+	nfs_add_stats(inode, NFSIOS_NORMALWRITTENBYTES, count);
 	result = generic_file_aio_write(iocb, buf, count, pos);
 out:
 	return result;
@@ -376,15 +391,17 @@ out_swapfile:
 
 static int do_getlk(struct file *filp, int cmd, struct file_lock *fl)
 {
-	struct file_lock *cfl;
+	struct file_lock cfl;
 	struct inode *inode = filp->f_mapping->host;
 	int status = 0;
 
 	lock_kernel();
 	/* Try local locking first */
-	cfl = posix_test_lock(filp, fl);
-	if (cfl != NULL) {
-		locks_copy_lock(fl, cfl);
+	if (posix_test_lock(filp, fl, &cfl)) {
+		fl->fl_start = cfl.fl_start;
+		fl->fl_end = cfl.fl_end;
+		fl->fl_type = cfl.fl_type;
+		fl->fl_pid = cfl.fl_pid;
 		goto out;
 	}
 
@@ -425,10 +442,8 @@ static int do_vfs_lock(struct file *file, struct file_lock *fl)
 static int do_unlk(struct file *filp, int cmd, struct file_lock *fl)
 {
 	struct inode *inode = filp->f_mapping->host;
-	sigset_t oldset;
 	int status;
 
-	rpc_clnt_sigmask(NFS_CLIENT(inode), &oldset);
 	/*
 	 * Flush all pending writes before doing anything
 	 * with locks..
@@ -446,17 +461,14 @@ static int do_unlk(struct file *filp, int cmd, struct file_lock *fl)
 	else
 		status = do_vfs_lock(filp, fl);
 	unlock_kernel();
-	rpc_clnt_sigunmask(NFS_CLIENT(inode), &oldset);
 	return status;
 }
 
 static int do_setlk(struct file *filp, int cmd, struct file_lock *fl)
 {
 	struct inode *inode = filp->f_mapping->host;
-	sigset_t oldset;
 	int status;
 
-	rpc_clnt_sigmask(NFS_CLIENT(inode), &oldset);
 	/*
 	 * Flush all pending writes before doing anything
 	 * with locks..
@@ -489,7 +501,6 @@ static int do_setlk(struct file *filp, int cmd, struct file_lock *fl)
 	nfs_sync_mapping(filp->f_mapping);
 	nfs_zap_caches(inode);
 out:
-	rpc_clnt_sigunmask(NFS_CLIENT(inode), &oldset);
 	return status;
 }
 
@@ -504,9 +515,7 @@ static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl)
 			inode->i_sb->s_id, inode->i_ino,
 			fl->fl_type, fl->fl_flags,
 			(long long)fl->fl_start, (long long)fl->fl_end);
-
-	if (!inode)
-		return -EINVAL;
+	nfs_inc_stats(inode, NFSIOS_VFSLOCK);
 
 	/* No mandatory locks over NFS */
 	if ((inode->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID &&
@@ -531,9 +540,6 @@ static int nfs_flock(struct file *filp, int cmd, struct file_lock *fl)
 			inode->i_sb->s_id, inode->i_ino,
 			fl->fl_type, fl->fl_flags);
 
-	if (!inode)
-		return -EINVAL;
-
 	/*
 	 * No BSD flocks over NFS allowed.
 	 * Note: we could try to fake a POSIX lock request here by
diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c
index 821edd30333b..3fab5b0cfc5a 100644
--- a/fs/nfs/idmap.c
+++ b/fs/nfs/idmap.c
@@ -35,6 +35,7 @@
  */
 
 #include <linux/module.h>
+#include <linux/mutex.h>
 #include <linux/init.h>
 #include <linux/types.h>
 #include <linux/slab.h>
@@ -74,8 +75,8 @@ struct idmap {
 	struct dentry        *idmap_dentry;
 	wait_queue_head_t     idmap_wq;
 	struct idmap_msg      idmap_im;
-	struct semaphore      idmap_lock;    /* Serializes upcalls */
-	struct semaphore      idmap_im_lock; /* Protects the hashtable */
+	struct mutex          idmap_lock;    /* Serializes upcalls */
+	struct mutex          idmap_im_lock; /* Protects the hashtable */
 	struct idmap_hashtable idmap_user_hash;
 	struct idmap_hashtable idmap_group_hash;
 };
@@ -101,11 +102,9 @@ nfs_idmap_new(struct nfs4_client *clp)
 
 	if (clp->cl_idmap != NULL)
 		return;
-        if ((idmap = kmalloc(sizeof(*idmap), GFP_KERNEL)) == NULL)
+        if ((idmap = kzalloc(sizeof(*idmap), GFP_KERNEL)) == NULL)
                 return;
 
-	memset(idmap, 0, sizeof(*idmap));
-
 	snprintf(idmap->idmap_path, sizeof(idmap->idmap_path),
 	    "%s/idmap", clp->cl_rpcclient->cl_pathname);
 
@@ -116,8 +115,8 @@ nfs_idmap_new(struct nfs4_client *clp)
 		return;
 	}
 
-        init_MUTEX(&idmap->idmap_lock);
-        init_MUTEX(&idmap->idmap_im_lock);
+        mutex_init(&idmap->idmap_lock);
+        mutex_init(&idmap->idmap_im_lock);
 	init_waitqueue_head(&idmap->idmap_wq);
 	idmap->idmap_user_hash.h_type = IDMAP_TYPE_USER;
 	idmap->idmap_group_hash.h_type = IDMAP_TYPE_GROUP;
@@ -132,6 +131,8 @@ nfs_idmap_delete(struct nfs4_client *clp)
 
 	if (!idmap)
 		return;
+	dput(idmap->idmap_dentry);
+	idmap->idmap_dentry = NULL;
 	rpc_unlink(idmap->idmap_path);
 	clp->cl_idmap = NULL;
 	kfree(idmap);
@@ -232,8 +233,8 @@ nfs_idmap_id(struct idmap *idmap, struct idmap_hashtable *h,
 	if (namelen >= IDMAP_NAMESZ)
 		return -EINVAL;
 
-	down(&idmap->idmap_lock);
-	down(&idmap->idmap_im_lock);
+	mutex_lock(&idmap->idmap_lock);
+	mutex_lock(&idmap->idmap_im_lock);
 
 	he = idmap_lookup_name(h, name, namelen);
 	if (he != NULL) {
@@ -259,11 +260,11 @@ nfs_idmap_id(struct idmap *idmap, struct idmap_hashtable *h,
 	}
 
 	set_current_state(TASK_UNINTERRUPTIBLE);
-	up(&idmap->idmap_im_lock);
+	mutex_unlock(&idmap->idmap_im_lock);
 	schedule();
 	current->state = TASK_RUNNING;
 	remove_wait_queue(&idmap->idmap_wq, &wq);
-	down(&idmap->idmap_im_lock);
+	mutex_lock(&idmap->idmap_im_lock);
 
 	if (im->im_status & IDMAP_STATUS_SUCCESS) {
 		*id = im->im_id;
@@ -272,8 +273,8 @@ nfs_idmap_id(struct idmap *idmap, struct idmap_hashtable *h,
 
  out:
 	memset(im, 0, sizeof(*im));
-	up(&idmap->idmap_im_lock);
-	up(&idmap->idmap_lock);
+	mutex_unlock(&idmap->idmap_im_lock);
+	mutex_unlock(&idmap->idmap_lock);
 	return (ret);
 }
 
@@ -293,8 +294,8 @@ nfs_idmap_name(struct idmap *idmap, struct idmap_hashtable *h,
 
 	im = &idmap->idmap_im;
 
-	down(&idmap->idmap_lock);
-	down(&idmap->idmap_im_lock);
+	mutex_lock(&idmap->idmap_lock);
+	mutex_lock(&idmap->idmap_im_lock);
 
 	he = idmap_lookup_id(h, id);
 	if (he != 0) {
@@ -320,11 +321,11 @@ nfs_idmap_name(struct idmap *idmap, struct idmap_hashtable *h,
 	}
 
 	set_current_state(TASK_UNINTERRUPTIBLE);
-	up(&idmap->idmap_im_lock);
+	mutex_unlock(&idmap->idmap_im_lock);
 	schedule();
 	current->state = TASK_RUNNING;
 	remove_wait_queue(&idmap->idmap_wq, &wq);
-	down(&idmap->idmap_im_lock);
+	mutex_lock(&idmap->idmap_im_lock);
 
 	if (im->im_status & IDMAP_STATUS_SUCCESS) {
 		if ((len = strnlen(im->im_name, IDMAP_NAMESZ)) == 0)
@@ -335,8 +336,8 @@ nfs_idmap_name(struct idmap *idmap, struct idmap_hashtable *h,
 
  out:
 	memset(im, 0, sizeof(*im));
-	up(&idmap->idmap_im_lock);
-	up(&idmap->idmap_lock);
+	mutex_unlock(&idmap->idmap_im_lock);
+	mutex_unlock(&idmap->idmap_lock);
 	return ret;
 }
 
@@ -380,7 +381,7 @@ idmap_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
         if (copy_from_user(&im_in, src, mlen) != 0)
 		return (-EFAULT);
 
-	down(&idmap->idmap_im_lock);
+	mutex_lock(&idmap->idmap_im_lock);
 
 	ret = mlen;
 	im->im_status = im_in.im_status;
@@ -440,7 +441,7 @@ idmap_pipe_downcall(struct file *filp, const char __user *src, size_t mlen)
 		idmap_update_entry(he, im_in.im_name, namelen_in, im_in.im_id);
 	ret = mlen;
 out:
-	up(&idmap->idmap_im_lock);
+	mutex_unlock(&idmap->idmap_im_lock);
 	return ret;
 }
 
@@ -452,10 +453,10 @@ idmap_pipe_destroy_msg(struct rpc_pipe_msg *msg)
 
 	if (msg->errno >= 0)
 		return;
-	down(&idmap->idmap_im_lock);
+	mutex_lock(&idmap->idmap_im_lock);
 	im->im_status = IDMAP_STATUS_LOOKUPFAIL;
 	wake_up(&idmap->idmap_wq);
-	up(&idmap->idmap_im_lock);
+	mutex_unlock(&idmap->idmap_im_lock);
 }
 
 /* 
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index a77ee95b7efb..2f7656b911b6 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -26,6 +26,7 @@
 #include <linux/unistd.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/stats.h>
+#include <linux/sunrpc/metrics.h>
 #include <linux/nfs_fs.h>
 #include <linux/nfs_mount.h>
 #include <linux/nfs4_mount.h>
@@ -42,6 +43,7 @@
 #include "nfs4_fs.h"
 #include "callback.h"
 #include "delegation.h"
+#include "iostat.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
 #define NFS_PARANOIA 1
@@ -65,6 +67,7 @@ static void nfs_clear_inode(struct inode *);
 static void nfs_umount_begin(struct super_block *);
 static int  nfs_statfs(struct super_block *, struct kstatfs *);
 static int  nfs_show_options(struct seq_file *, struct vfsmount *);
+static int  nfs_show_stats(struct seq_file *, struct vfsmount *);
 static void nfs_zap_acl_cache(struct inode *);
 
 static struct rpc_program	nfs_program;
@@ -78,6 +81,7 @@ static struct super_operations nfs_sops = {
 	.clear_inode	= nfs_clear_inode,
 	.umount_begin	= nfs_umount_begin,
 	.show_options	= nfs_show_options,
+	.show_stats	= nfs_show_stats,
 };
 
 /*
@@ -103,7 +107,7 @@ static struct rpc_version *	nfs_version[] = {
 static struct rpc_program	nfs_program = {
 	.name			= "nfs",
 	.number			= NFS_PROGRAM,
-	.nrvers			= sizeof(nfs_version) / sizeof(nfs_version[0]),
+	.nrvers			= ARRAY_SIZE(nfs_version),
 	.version		= nfs_version,
 	.stats			= &nfs_rpcstat,
 	.pipe_dir_name		= "/nfs",
@@ -118,7 +122,7 @@ static struct rpc_version *	nfsacl_version[] = {
 struct rpc_program		nfsacl_program = {
 	.name =			"nfsacl",
 	.number =		NFS_ACL_PROGRAM,
-	.nrvers =		sizeof(nfsacl_version) / sizeof(nfsacl_version[0]),
+	.nrvers =		ARRAY_SIZE(nfsacl_version),
 	.version =		nfsacl_version,
 	.stats =		&nfsacl_rpcstat,
 };
@@ -133,7 +137,7 @@ nfs_fattr_to_ino_t(struct nfs_fattr *fattr)
 static int
 nfs_write_inode(struct inode *inode, int sync)
 {
-	int flags = sync ? FLUSH_WAIT : 0;
+	int flags = sync ? FLUSH_SYNC : 0;
 	int ret;
 
 	ret = nfs_commit_inode(inode, flags);
@@ -237,7 +241,6 @@ static struct inode *
 nfs_get_root(struct super_block *sb, struct nfs_fh *rootfh, struct nfs_fsinfo *fsinfo)
 {
 	struct nfs_server	*server = NFS_SB(sb);
-	struct inode *rooti;
 	int			error;
 
 	error = server->rpc_ops->getroot(server, rootfh, fsinfo);
@@ -246,10 +249,7 @@ nfs_get_root(struct super_block *sb, struct nfs_fh *rootfh, struct nfs_fsinfo *f
 		return ERR_PTR(error);
 	}
 
-	rooti = nfs_fhget(sb, rootfh, fsinfo->fattr);
-	if (!rooti)
-		return ERR_PTR(-ENOMEM);
-	return rooti;
+	return nfs_fhget(sb, rootfh, fsinfo->fattr);
 }
 
 /*
@@ -277,6 +277,10 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor)
 
 	sb->s_magic      = NFS_SUPER_MAGIC;
 
+	server->io_stats = nfs_alloc_iostats();
+	if (server->io_stats == NULL)
+		return -ENOMEM;
+
 	root_inode = nfs_get_root(sb, &server->fh, &fsinfo);
 	/* Did getting the root inode fail? */
 	if (IS_ERR(root_inode)) {
@@ -290,6 +294,9 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor)
 	}
 	sb->s_root->d_op = server->rpc_ops->dentry_ops;
 
+	/* mount time stamp, in seconds */
+	server->mount_time = jiffies;
+
 	/* Get some general file system info */
 	if (server->namelen == 0 &&
 	    server->rpc_ops->pathconf(server, &server->fh, &pathinfo) >= 0)
@@ -396,6 +403,9 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data)
 
 	nfs_init_timeout_values(&timeparms, proto, data->timeo, data->retrans);
 
+	server->retrans_timeo = timeparms.to_initval;
+	server->retrans_count = timeparms.to_retries;
+
 	/* create transport and client */
 	xprt = xprt_create_proto(proto, &server->addr, &timeparms);
 	if (IS_ERR(xprt)) {
@@ -579,7 +589,7 @@ nfs_statfs(struct super_block *sb, struct kstatfs *buf)
 
 }
 
-static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
+static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, int showdefaults)
 {
 	static struct proc_nfs_info {
 		int flag;
@@ -588,28 +598,26 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
 	} nfs_info[] = {
 		{ NFS_MOUNT_SOFT, ",soft", ",hard" },
 		{ NFS_MOUNT_INTR, ",intr", "" },
-		{ NFS_MOUNT_POSIX, ",posix", "" },
 		{ NFS_MOUNT_NOCTO, ",nocto", "" },
 		{ NFS_MOUNT_NOAC, ",noac", "" },
-		{ NFS_MOUNT_NONLM, ",nolock", ",lock" },
+		{ NFS_MOUNT_NONLM, ",nolock", "" },
 		{ NFS_MOUNT_NOACL, ",noacl", "" },
 		{ 0, NULL, NULL }
 	};
 	struct proc_nfs_info *nfs_infop;
-	struct nfs_server *nfss = NFS_SB(mnt->mnt_sb);
 	char buf[12];
 	char *proto;
 
-	seq_printf(m, ",v%d", nfss->rpc_ops->version);
+	seq_printf(m, ",vers=%d", nfss->rpc_ops->version);
 	seq_printf(m, ",rsize=%d", nfss->rsize);
 	seq_printf(m, ",wsize=%d", nfss->wsize);
-	if (nfss->acregmin != 3*HZ)
+	if (nfss->acregmin != 3*HZ || showdefaults)
 		seq_printf(m, ",acregmin=%d", nfss->acregmin/HZ);
-	if (nfss->acregmax != 60*HZ)
+	if (nfss->acregmax != 60*HZ || showdefaults)
 		seq_printf(m, ",acregmax=%d", nfss->acregmax/HZ);
-	if (nfss->acdirmin != 30*HZ)
+	if (nfss->acdirmin != 30*HZ || showdefaults)
 		seq_printf(m, ",acdirmin=%d", nfss->acdirmin/HZ);
-	if (nfss->acdirmax != 60*HZ)
+	if (nfss->acdirmax != 60*HZ || showdefaults)
 		seq_printf(m, ",acdirmax=%d", nfss->acdirmax/HZ);
 	for (nfs_infop = nfs_info; nfs_infop->flag; nfs_infop++) {
 		if (nfss->flags & nfs_infop->flag)
@@ -629,8 +637,96 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
 			proto = buf;
 	}
 	seq_printf(m, ",proto=%s", proto);
+	seq_printf(m, ",timeo=%lu", 10U * nfss->retrans_timeo / HZ);
+	seq_printf(m, ",retrans=%u", nfss->retrans_count);
+}
+
+static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
+{
+	struct nfs_server *nfss = NFS_SB(mnt->mnt_sb);
+
+	nfs_show_mount_options(m, nfss, 0);
+
 	seq_puts(m, ",addr=");
 	seq_escape(m, nfss->hostname, " \t\n\\");
+
+	return 0;
+}
+
+static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt)
+{
+	int i, cpu;
+	struct nfs_server *nfss = NFS_SB(mnt->mnt_sb);
+	struct rpc_auth *auth = nfss->client->cl_auth;
+	struct nfs_iostats totals = { };
+
+	seq_printf(m, "statvers=%s", NFS_IOSTAT_VERS);
+
+	/*
+	 * Display all mount option settings
+	 */
+	seq_printf(m, "\n\topts:\t");
+	seq_puts(m, mnt->mnt_sb->s_flags & MS_RDONLY ? "ro" : "rw");
+	seq_puts(m, mnt->mnt_sb->s_flags & MS_SYNCHRONOUS ? ",sync" : "");
+	seq_puts(m, mnt->mnt_sb->s_flags & MS_NOATIME ? ",noatime" : "");
+	seq_puts(m, mnt->mnt_sb->s_flags & MS_NODIRATIME ? ",nodiratime" : "");
+	nfs_show_mount_options(m, nfss, 1);
+
+	seq_printf(m, "\n\tage:\t%lu", (jiffies - nfss->mount_time) / HZ);
+
+	seq_printf(m, "\n\tcaps:\t");
+	seq_printf(m, "caps=0x%x", nfss->caps);
+	seq_printf(m, ",wtmult=%d", nfss->wtmult);
+	seq_printf(m, ",dtsize=%d", nfss->dtsize);
+	seq_printf(m, ",bsize=%d", nfss->bsize);
+	seq_printf(m, ",namelen=%d", nfss->namelen);
+
+#ifdef CONFIG_NFS_V4
+	if (nfss->rpc_ops->version == 4) {
+		seq_printf(m, "\n\tnfsv4:\t");
+		seq_printf(m, "bm0=0x%x", nfss->attr_bitmask[0]);
+		seq_printf(m, ",bm1=0x%x", nfss->attr_bitmask[1]);
+		seq_printf(m, ",acl=0x%x", nfss->acl_bitmask);
+	}
+#endif
+
+	/*
+	 * Display security flavor in effect for this mount
+	 */
+	seq_printf(m, "\n\tsec:\tflavor=%d", auth->au_ops->au_flavor);
+	if (auth->au_flavor)
+		seq_printf(m, ",pseudoflavor=%d", auth->au_flavor);
+
+	/*
+	 * Display superblock I/O counters
+	 */
+	for (cpu = 0; cpu < NR_CPUS; cpu++) {
+		struct nfs_iostats *stats;
+
+		if (!cpu_possible(cpu))
+			continue;
+
+		preempt_disable();
+		stats = per_cpu_ptr(nfss->io_stats, cpu);
+
+		for (i = 0; i < __NFSIOS_COUNTSMAX; i++)
+			totals.events[i] += stats->events[i];
+		for (i = 0; i < __NFSIOS_BYTESMAX; i++)
+			totals.bytes[i] += stats->bytes[i];
+
+		preempt_enable();
+	}
+
+	seq_printf(m, "\n\tevents:\t");
+	for (i = 0; i < __NFSIOS_COUNTSMAX; i++)
+		seq_printf(m, "%lu ", totals.events[i]);
+	seq_printf(m, "\n\tbytes:\t");
+	for (i = 0; i < __NFSIOS_BYTESMAX; i++)
+		seq_printf(m, "%Lu ", totals.bytes[i]);
+	seq_printf(m, "\n");
+
+	rpc_print_iostats(m, nfss->client);
+
 	return 0;
 }
 
@@ -660,6 +756,8 @@ static void nfs_zap_caches_locked(struct inode *inode)
 	struct nfs_inode *nfsi = NFS_I(inode);
 	int mode = inode->i_mode;
 
+	nfs_inc_stats(inode, NFSIOS_ATTRINVALIDATE);
+
 	NFS_ATTRTIMEO(inode) = NFS_MINATTRTIMEO(inode);
 	NFS_ATTRTIMEO_UPDATE(inode) = jiffies;
 
@@ -751,7 +849,7 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
 		.fh	= fh,
 		.fattr	= fattr
 	};
-	struct inode *inode = NULL;
+	struct inode *inode = ERR_PTR(-ENOENT);
 	unsigned long hash;
 
 	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
@@ -764,8 +862,11 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr)
 
 	hash = nfs_fattr_to_ino_t(fattr);
 
-	if (!(inode = iget5_locked(sb, hash, nfs_find_actor, nfs_init_locked, &desc)))
+	inode = iget5_locked(sb, hash, nfs_find_actor, nfs_init_locked, &desc);
+	if (inode == NULL) {
+		inode = ERR_PTR(-ENOMEM);
 		goto out_no_inode;
+	}
 
 	if (inode->i_state & I_NEW) {
 		struct nfs_inode *nfsi = NFS_I(inode);
@@ -834,7 +935,7 @@ out:
 	return inode;
 
 out_no_inode:
-	printk("nfs_fhget: iget failed\n");
+	dprintk("nfs_fhget: iget failed with error %ld\n", PTR_ERR(inode));
 	goto out;
 }
 
@@ -847,6 +948,8 @@ nfs_setattr(struct dentry *dentry, struct iattr *attr)
 	struct nfs_fattr fattr;
 	int error;
 
+	nfs_inc_stats(inode, NFSIOS_VFSSETATTR);
+
 	if (attr->ia_valid & ATTR_SIZE) {
 		if (!S_ISREG(inode->i_mode) || attr->ia_size == i_size_read(inode))
 			attr->ia_valid &= ~ATTR_SIZE;
@@ -859,11 +962,9 @@ nfs_setattr(struct dentry *dentry, struct iattr *attr)
 
 	lock_kernel();
 	nfs_begin_data_update(inode);
-	/* Write all dirty data if we're changing file permissions or size */
-	if ((attr->ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID|ATTR_SIZE)) != 0) {
-		filemap_write_and_wait(inode->i_mapping);
-		nfs_wb_all(inode);
-	}
+	/* Write all dirty data */
+	filemap_write_and_wait(inode->i_mapping);
+	nfs_wb_all(inode);
 	/*
 	 * Return any delegations if we're going to change ACLs
 	 */
@@ -902,6 +1003,7 @@ void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr)
 		spin_unlock(&inode->i_lock);
 	}
 	if ((attr->ia_valid & ATTR_SIZE) != 0) {
+		nfs_inc_stats(inode, NFSIOS_SETATTRTRUNC);
 		inode->i_size = attr->ia_size;
 		vmtruncate(inode, attr->ia_size);
 	}
@@ -949,7 +1051,7 @@ int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
 	int err;
 
 	/* Flush out writes to the server in order to update c/mtime */
-	nfs_sync_inode(inode, 0, 0, FLUSH_WAIT|FLUSH_NOCOMMIT);
+	nfs_sync_inode_wait(inode, 0, 0, FLUSH_NOCOMMIT);
 
 	/*
 	 * We may force a getattr if the user cares about atime.
@@ -973,7 +1075,7 @@ int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
 	return err;
 }
 
-struct nfs_open_context *alloc_nfs_open_context(struct dentry *dentry, struct rpc_cred *cred)
+static struct nfs_open_context *alloc_nfs_open_context(struct vfsmount *mnt, struct dentry *dentry, struct rpc_cred *cred)
 {
 	struct nfs_open_context *ctx;
 
@@ -981,6 +1083,7 @@ struct nfs_open_context *alloc_nfs_open_context(struct dentry *dentry, struct rp
 	if (ctx != NULL) {
 		atomic_set(&ctx->count, 1);
 		ctx->dentry = dget(dentry);
+		ctx->vfsmnt = mntget(mnt);
 		ctx->cred = get_rpccred(cred);
 		ctx->state = NULL;
 		ctx->lockowner = current->files;
@@ -1011,6 +1114,7 @@ void put_nfs_open_context(struct nfs_open_context *ctx)
 		if (ctx->cred != NULL)
 			put_rpccred(ctx->cred);
 		dput(ctx->dentry);
+		mntput(ctx->vfsmnt);
 		kfree(ctx);
 	}
 }
@@ -1019,7 +1123,7 @@ void put_nfs_open_context(struct nfs_open_context *ctx)
  * Ensure that mmap has a recent RPC credential for use when writing out
  * shared pages
  */
-void nfs_file_set_open_context(struct file *filp, struct nfs_open_context *ctx)
+static void nfs_file_set_open_context(struct file *filp, struct nfs_open_context *ctx)
 {
 	struct inode *inode = filp->f_dentry->d_inode;
 	struct nfs_inode *nfsi = NFS_I(inode);
@@ -1051,7 +1155,7 @@ struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_c
 	return ctx;
 }
 
-void nfs_file_clear_open_context(struct file *filp)
+static void nfs_file_clear_open_context(struct file *filp)
 {
 	struct inode *inode = filp->f_dentry->d_inode;
 	struct nfs_open_context *ctx = (struct nfs_open_context *)filp->private_data;
@@ -1076,7 +1180,7 @@ int nfs_open(struct inode *inode, struct file *filp)
 	cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0);
 	if (IS_ERR(cred))
 		return PTR_ERR(cred);
-	ctx = alloc_nfs_open_context(filp->f_dentry, cred);
+	ctx = alloc_nfs_open_context(filp->f_vfsmnt, filp->f_dentry, cred);
 	put_rpccred(cred);
 	if (ctx == NULL)
 		return -ENOMEM;
@@ -1185,6 +1289,7 @@ int nfs_attribute_timeout(struct inode *inode)
  */
 int nfs_revalidate_inode(struct nfs_server *server, struct inode *inode)
 {
+	nfs_inc_stats(inode, NFSIOS_INODEREVALIDATE);
 	if (!(NFS_I(inode)->cache_validity & (NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA))
 			&& !nfs_attribute_timeout(inode))
 		return NFS_STALE(inode) ? -ESTALE : 0;
@@ -1201,6 +1306,7 @@ void nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping)
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	if (nfsi->cache_validity & NFS_INO_INVALID_DATA) {
+		nfs_inc_stats(inode, NFSIOS_DATAINVALIDATE);
 		if (S_ISREG(inode->i_mode))
 			nfs_sync_mapping(mapping);
 		invalidate_inode_pages2(mapping);
@@ -1299,39 +1405,37 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
 		return 0;
 
+	/* Has the inode gone and changed behind our back? */
+	if (nfsi->fileid != fattr->fileid
+			|| (inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT)) {
+		return -EIO;
+	}
+
 	/* Are we in the process of updating data on the server? */
 	data_unstable = nfs_caches_unstable(inode);
 
 	/* Do atomic weak cache consistency updates */
 	nfs_wcc_update_inode(inode, fattr);
 
-	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
-			nfsi->change_attr != fattr->change_attr) {
+	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0) {
+		if (nfsi->change_attr == fattr->change_attr)
+			goto out;
 		nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
 		if (!data_unstable)
 			nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE;
 	}
 
-	/* Has the inode gone and changed behind our back? */
-	if (nfsi->fileid != fattr->fileid
-			|| (inode->i_mode & S_IFMT) != (fattr->mode & S_IFMT)) {
-		return -EIO;
-	}
-
-	cur_size = i_size_read(inode);
- 	new_isize = nfs_size_to_loff_t(fattr->size);
-
 	/* Verify a few of the more important attributes */
 	if (!timespec_equal(&inode->i_mtime, &fattr->mtime)) {
 		nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
 		if (!data_unstable)
 			nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE;
 	}
-	if (cur_size != new_isize) {
-		nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
-		if (nfsi->npages == 0)
-			nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE;
-	}
+
+	cur_size = i_size_read(inode);
+ 	new_isize = nfs_size_to_loff_t(fattr->size);
+	if (cur_size != new_isize && nfsi->npages == 0)
+		nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
 
 	/* Have any file permissions changed? */
 	if ((inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO)
@@ -1343,6 +1447,7 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 	if (inode->i_nlink != fattr->nlink)
 		nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
 
+out:
 	if (!timespec_equal(&inode->i_atime, &fattr->atime))
 		nfsi->cache_validity |= NFS_INO_INVALID_ATIME;
 
@@ -1481,15 +1586,6 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		nfsi->cache_change_attribute = jiffies;
 	}
 
-	if ((fattr->valid & NFS_ATTR_FATTR_V4)
-	    && nfsi->change_attr != fattr->change_attr) {
-		dprintk("NFS: change_attr change on server for file %s/%ld\n",
-		       inode->i_sb->s_id, inode->i_ino);
-		nfsi->change_attr = fattr->change_attr;
-		invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
-		nfsi->cache_change_attribute = jiffies;
-	}
-
 	/* If ctime has changed we should definitely clear access+acl caches */
 	if (!timespec_equal(&inode->i_ctime, &fattr->ctime)) {
 		invalid |= NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
@@ -1519,8 +1615,20 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
  		inode->i_blksize = fattr->du.nfs2.blocksize;
  	}
 
+	if ((fattr->valid & NFS_ATTR_FATTR_V4)) {
+		if (nfsi->change_attr != fattr->change_attr) {
+			dprintk("NFS: change_attr change on server for file %s/%ld\n",
+					inode->i_sb->s_id, inode->i_ino);
+			nfsi->change_attr = fattr->change_attr;
+			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
+			nfsi->cache_change_attribute = jiffies;
+		} else
+			invalid &= ~(NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA);
+	}
+
 	/* Update attrtimeo value if we're out of the unstable period */
 	if (invalid & NFS_INO_INVALID_ATTR) {
+		nfs_inc_stats(inode, NFSIOS_ATTRINVALIDATE);
 		nfsi->attrtimeo = NFS_MINATTRTIMEO(inode);
 		nfsi->attrtimeo_timestamp = jiffies;
 	} else if (time_after(jiffies, nfsi->attrtimeo_timestamp+nfsi->attrtimeo)) {
@@ -1637,10 +1745,9 @@ static struct super_block *nfs_get_sb(struct file_system_type *fs_type,
 #endif /* CONFIG_NFS_V3 */
 
 	s = ERR_PTR(-ENOMEM);
-	server = kmalloc(sizeof(struct nfs_server), GFP_KERNEL);
+	server = kzalloc(sizeof(struct nfs_server), GFP_KERNEL);
 	if (!server)
 		goto out_err;
-	memset(server, 0, sizeof(struct nfs_server));
 	/* Zero out the NFS state stuff */
 	init_nfsv4_state(server);
 	server->client = server->client_sys = server->client_acl = ERR_PTR(-EINVAL);
@@ -1679,7 +1786,7 @@ static struct super_block *nfs_get_sb(struct file_system_type *fs_type,
 
 	s->s_flags = flags;
 
-	error = nfs_fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
+	error = nfs_fill_super(s, data, flags & MS_SILENT ? 1 : 0);
 	if (error) {
 		up_write(&s->s_umount);
 		deactivate_super(s);
@@ -1712,6 +1819,7 @@ static void nfs_kill_super(struct super_block *s)
 
 	rpciod_down();		/* release rpciod */
 
+	nfs_free_iostats(server->io_stats);
 	kfree(server->hostname);
 	kfree(server);
 }
@@ -1738,6 +1846,7 @@ static struct super_operations nfs4_sops = {
 	.clear_inode	= nfs4_clear_inode,
 	.umount_begin	= nfs_umount_begin,
 	.show_options	= nfs_show_options,
+	.show_stats	= nfs_show_stats,
 };
 
 /*
@@ -1800,6 +1909,9 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data,
 
 	nfs_init_timeout_values(&timeparms, data->proto, data->timeo, data->retrans);
 
+	server->retrans_timeo = timeparms.to_initval;
+	server->retrans_count = timeparms.to_retries;
+
 	clp = nfs4_get_client(&server->addr.sin_addr);
 	if (!clp) {
 		dprintk("%s: failed to create NFS4 client.\n", __FUNCTION__);
@@ -1941,10 +2053,9 @@ static struct super_block *nfs4_get_sb(struct file_system_type *fs_type,
 		return ERR_PTR(-EINVAL);
 	}
 
-	server = kmalloc(sizeof(struct nfs_server), GFP_KERNEL);
+	server = kzalloc(sizeof(struct nfs_server), GFP_KERNEL);
 	if (!server)
 		return ERR_PTR(-ENOMEM);
-	memset(server, 0, sizeof(struct nfs_server));
 	/* Zero out the NFS state stuff */
 	init_nfsv4_state(server);
 	server->client = server->client_sys = server->client_acl = ERR_PTR(-EINVAL);
@@ -1996,7 +2107,7 @@ static struct super_block *nfs4_get_sb(struct file_system_type *fs_type,
 
 	s->s_flags = flags;
 
-	error = nfs4_fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
+	error = nfs4_fill_super(s, data, flags & MS_SILENT ? 1 : 0);
 	if (error) {
 		up_write(&s->s_umount);
 		deactivate_super(s);
@@ -2024,10 +2135,12 @@ static void nfs4_kill_super(struct super_block *sb)
 
 	if (server->client != NULL && !IS_ERR(server->client))
 		rpc_shutdown_client(server->client);
-	rpciod_down();		/* release rpciod */
 
 	destroy_nfsv4_state(server);
 
+	rpciod_down();
+
+	nfs_free_iostats(server->io_stats);
 	kfree(server->hostname);
 	kfree(server);
 }
@@ -2163,7 +2276,8 @@ static int nfs_init_inodecache(void)
 {
 	nfs_inode_cachep = kmem_cache_create("nfs_inode_cache",
 					     sizeof(struct nfs_inode),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (nfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/nfs/iostat.h b/fs/nfs/iostat.h
new file mode 100644
index 000000000000..6350ecbde589
--- /dev/null
+++ b/fs/nfs/iostat.h
@@ -0,0 +1,164 @@
+/*
+ *  linux/fs/nfs/iostat.h
+ *
+ *  Declarations for NFS client per-mount statistics
+ *
+ *  Copyright (C) 2005, 2006 Chuck Lever <cel@netapp.com>
+ *
+ *  NFS client per-mount statistics provide information about the health of
+ *  the NFS client and the health of each NFS mount point.  Generally these
+ *  are not for detailed problem diagnosis, but simply to indicate that there
+ *  is a problem.
+ *
+ *  These counters are not meant to be human-readable, but are meant to be
+ *  integrated into system monitoring tools such as "sar" and "iostat".  As
+ *  such, the counters are sampled by the tools over time, and are never
+ *  zeroed after a file system is mounted.  Moving averages can be computed
+ *  by the tools by taking the difference between two instantaneous samples
+ *  and dividing that by the time between the samples.
+ */
+
+#ifndef _NFS_IOSTAT
+#define _NFS_IOSTAT
+
+#define NFS_IOSTAT_VERS		"1.0"
+
+/*
+ * NFS byte counters
+ *
+ * 1.  SERVER - the number of payload bytes read from or written to the
+ *     server by the NFS client via an NFS READ or WRITE request.
+ *
+ * 2.  NORMAL - the number of bytes read or written by applications via
+ *     the read(2) and write(2) system call interfaces.
+ *
+ * 3.  DIRECT - the number of bytes read or written from files opened
+ *     with the O_DIRECT flag.
+ *
+ * These counters give a view of the data throughput into and out of the NFS
+ * client.  Comparing the number of bytes requested by an application with the
+ * number of bytes the client requests from the server can provide an
+ * indication of client efficiency (per-op, cache hits, etc).
+ *
+ * These counters can also help characterize which access methods are in
+ * use.  DIRECT by itself shows whether there is any O_DIRECT traffic.
+ * NORMAL + DIRECT shows how much data is going through the system call
+ * interface.  A large amount of SERVER traffic without much NORMAL or
+ * DIRECT traffic shows that applications are using mapped files.
+ *
+ * NFS page counters
+ *
+ * These count the number of pages read or written via nfs_readpage(),
+ * nfs_readpages(), or their write equivalents.
+ */
+enum nfs_stat_bytecounters {
+	NFSIOS_NORMALREADBYTES = 0,
+	NFSIOS_NORMALWRITTENBYTES,
+	NFSIOS_DIRECTREADBYTES,
+	NFSIOS_DIRECTWRITTENBYTES,
+	NFSIOS_SERVERREADBYTES,
+	NFSIOS_SERVERWRITTENBYTES,
+	NFSIOS_READPAGES,
+	NFSIOS_WRITEPAGES,
+	__NFSIOS_BYTESMAX,
+};
+
+/*
+ * NFS event counters
+ *
+ * These counters provide a low-overhead way of monitoring client activity
+ * without enabling NFS trace debugging.  The counters show the rate at
+ * which VFS requests are made, and how often the client invalidates its
+ * data and attribute caches.  This allows system administrators to monitor
+ * such things as how close-to-open is working, and answer questions such
+ * as "why are there so many GETATTR requests on the wire?"
+ *
+ * They also count anamolous events such as short reads and writes, silly
+ * renames due to close-after-delete, and operations that change the size
+ * of a file (such operations can often be the source of data corruption
+ * if applications aren't using file locking properly).
+ */
+enum nfs_stat_eventcounters {
+	NFSIOS_INODEREVALIDATE = 0,
+	NFSIOS_DENTRYREVALIDATE,
+	NFSIOS_DATAINVALIDATE,
+	NFSIOS_ATTRINVALIDATE,
+	NFSIOS_VFSOPEN,
+	NFSIOS_VFSLOOKUP,
+	NFSIOS_VFSACCESS,
+	NFSIOS_VFSUPDATEPAGE,
+	NFSIOS_VFSREADPAGE,
+	NFSIOS_VFSREADPAGES,
+	NFSIOS_VFSWRITEPAGE,
+	NFSIOS_VFSWRITEPAGES,
+	NFSIOS_VFSGETDENTS,
+	NFSIOS_VFSSETATTR,
+	NFSIOS_VFSFLUSH,
+	NFSIOS_VFSFSYNC,
+	NFSIOS_VFSLOCK,
+	NFSIOS_VFSRELEASE,
+	NFSIOS_CONGESTIONWAIT,
+	NFSIOS_SETATTRTRUNC,
+	NFSIOS_EXTENDWRITE,
+	NFSIOS_SILLYRENAME,
+	NFSIOS_SHORTREAD,
+	NFSIOS_SHORTWRITE,
+	NFSIOS_DELAY,
+	__NFSIOS_COUNTSMAX,
+};
+
+#ifdef __KERNEL__
+
+#include <linux/percpu.h>
+#include <linux/cache.h>
+
+struct nfs_iostats {
+	unsigned long long	bytes[__NFSIOS_BYTESMAX];
+	unsigned long		events[__NFSIOS_COUNTSMAX];
+} ____cacheline_aligned;
+
+static inline void nfs_inc_server_stats(struct nfs_server *server, enum nfs_stat_eventcounters stat)
+{
+	struct nfs_iostats *iostats;
+	int cpu;
+
+	cpu = get_cpu();
+	iostats = per_cpu_ptr(server->io_stats, cpu);
+	iostats->events[stat] ++;
+	put_cpu_no_resched();
+}
+
+static inline void nfs_inc_stats(struct inode *inode, enum nfs_stat_eventcounters stat)
+{
+	nfs_inc_server_stats(NFS_SERVER(inode), stat);
+}
+
+static inline void nfs_add_server_stats(struct nfs_server *server, enum nfs_stat_bytecounters stat, unsigned long addend)
+{
+	struct nfs_iostats *iostats;
+	int cpu;
+
+	cpu = get_cpu();
+	iostats = per_cpu_ptr(server->io_stats, cpu);
+	iostats->bytes[stat] += addend;
+	put_cpu_no_resched();
+}
+
+static inline void nfs_add_stats(struct inode *inode, enum nfs_stat_bytecounters stat, unsigned long addend)
+{
+	nfs_add_server_stats(NFS_SERVER(inode), stat, addend);
+}
+
+static inline struct nfs_iostats *nfs_alloc_iostats(void)
+{
+	return alloc_percpu(struct nfs_iostats);
+}
+
+static inline void nfs_free_iostats(struct nfs_iostats *stats)
+{
+	if (stats != NULL)
+		free_percpu(stats);
+}
+
+#endif
+#endif
diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c
index db99b8f678f8..445abb4d4214 100644
--- a/fs/nfs/mount_clnt.c
+++ b/fs/nfs/mount_clnt.c
@@ -49,9 +49,12 @@ nfsroot_mount(struct sockaddr_in *addr, char *path, struct nfs_fh *fh,
 	struct mnt_fhstatus	result = {
 		.fh		= fh
 	};
+	struct rpc_message msg	= {
+		.rpc_argp	= path,
+		.rpc_resp	= &result,
+	};
 	char			hostname[32];
 	int			status;
-	int			call;
 
 	dprintk("NFS:      nfs_mount(%08x:%s)\n",
 			(unsigned)ntohl(addr->sin_addr.s_addr), path);
@@ -61,8 +64,12 @@ nfsroot_mount(struct sockaddr_in *addr, char *path, struct nfs_fh *fh,
 	if (IS_ERR(mnt_clnt))
 		return PTR_ERR(mnt_clnt);
 
-	call = (version == NFS_MNT3_VERSION) ? MOUNTPROC3_MNT : MNTPROC_MNT;
-	status = rpc_call(mnt_clnt, call, path, &result, 0);
+	if (version == NFS_MNT3_VERSION)
+		msg.rpc_proc = &mnt_clnt->cl_procinfo[MOUNTPROC3_MNT];
+	else
+		msg.rpc_proc = &mnt_clnt->cl_procinfo[MNTPROC_MNT];
+
+	status = rpc_call_sync(mnt_clnt, &msg, 0);
 	return status < 0? status : (result.status? -EACCES : 0);
 }
 
@@ -137,6 +144,8 @@ static struct rpc_procinfo	mnt_procedures[] = {
 	  .p_encode		= (kxdrproc_t) xdr_encode_dirpath,	
 	  .p_decode		= (kxdrproc_t) xdr_decode_fhstatus,
 	  .p_bufsiz		= MNT_dirpath_sz << 2,
+	  .p_statidx		= MNTPROC_MNT,
+	  .p_name		= "MOUNT",
 	},
 };
 
@@ -146,6 +155,8 @@ static struct rpc_procinfo mnt3_procedures[] = {
 	  .p_encode		= (kxdrproc_t) xdr_encode_dirpath,
 	  .p_decode		= (kxdrproc_t) xdr_decode_fhstatus3,
 	  .p_bufsiz		= MNT_dirpath_sz << 2,
+	  .p_statidx		= MOUNTPROC3_MNT,
+	  .p_name		= "MOUNT",
 	},
 };
 
@@ -174,7 +185,7 @@ static struct rpc_stat		mnt_stats;
 static struct rpc_program	mnt_program = {
 	.name		= "mount",
 	.number		= NFS_MNT_PROGRAM,
-	.nrvers		= sizeof(mnt_version)/sizeof(mnt_version[0]),
+	.nrvers		= ARRAY_SIZE(mnt_version),
 	.version	= mnt_version,
 	.stats		= &mnt_stats,
 };
diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c
index 7fc0560c89c9..f0015fa876e1 100644
--- a/fs/nfs/nfs2xdr.c
+++ b/fs/nfs/nfs2xdr.c
@@ -682,7 +682,9 @@ nfs_stat_to_errno(int stat)
 	.p_encode   =  (kxdrproc_t) nfs_xdr_##argtype,			\
 	.p_decode   =  (kxdrproc_t) nfs_xdr_##restype,			\
 	.p_bufsiz   =  MAX(NFS_##argtype##_sz,NFS_##restype##_sz) << 2,	\
-	.p_timer    =  timer						\
+	.p_timer    =  timer,						\
+	.p_statidx  =  NFSPROC_##proc,					\
+	.p_name     =  #proc,						\
 	}
 struct rpc_procinfo	nfs_procedures[] = {
     PROC(GETATTR,	fhandle,	attrstat, 1),
@@ -704,6 +706,6 @@ struct rpc_procinfo	nfs_procedures[] = {
 
 struct rpc_version		nfs_version2 = {
 	.number			= 2,
-	.nrprocs		= sizeof(nfs_procedures)/sizeof(nfs_procedures[0]),
+	.nrprocs		= ARRAY_SIZE(nfs_procedures),
 	.procs			= nfs_procedures
 };
diff --git a/fs/nfs/nfs3acl.c b/fs/nfs/nfs3acl.c
index 6a5bbc0ae941..33287879bd23 100644
--- a/fs/nfs/nfs3acl.c
+++ b/fs/nfs/nfs3acl.c
@@ -190,6 +190,10 @@ struct posix_acl *nfs3_proc_getacl(struct inode *inode, int type)
 	struct nfs3_getaclres res = {
 		.fattr =	&fattr,
 	};
+	struct rpc_message msg = {
+		.rpc_argp	= &args,
+		.rpc_resp	= &res,
+	};
 	struct posix_acl *acl;
 	int status, count;
 
@@ -218,8 +222,8 @@ struct posix_acl *nfs3_proc_getacl(struct inode *inode, int type)
 		return NULL;
 
 	dprintk("NFS call getacl\n");
-	status = rpc_call(server->client_acl, ACLPROC3_GETACL,
-			  &args, &res, 0);
+	msg.rpc_proc = &server->client_acl->cl_procinfo[ACLPROC3_GETACL];
+	status = rpc_call_sync(server->client_acl, &msg, 0);
 	dprintk("NFS reply getacl: %d\n", status);
 
 	/* pages may have been allocated at the xdr layer. */
@@ -286,6 +290,10 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 		.acl_access = acl,
 		.pages = pages,
 	};
+	struct rpc_message msg = {
+		.rpc_argp	= &args,
+		.rpc_resp	= &fattr,
+	};
 	int status, count;
 
 	status = -EOPNOTSUPP;
@@ -306,8 +314,8 @@ static int nfs3_proc_setacls(struct inode *inode, struct posix_acl *acl,
 
 	dprintk("NFS call setacl\n");
 	nfs_begin_data_update(inode);
-	status = rpc_call(server->client_acl, ACLPROC3_SETACL,
-			  &args, &fattr, 0);
+	msg.rpc_proc = &server->client_acl->cl_procinfo[ACLPROC3_SETACL];
+	status = rpc_call_sync(server->client_acl, &msg, 0);
 	spin_lock(&inode->i_lock);
 	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ACCESS;
 	spin_unlock(&inode->i_lock);
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index ed67567f0556..cf186f0d2b3b 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -19,6 +19,8 @@
 #include <linux/smp_lock.h>
 #include <linux/nfs_mount.h>
 
+#include "iostat.h"
+
 #define NFSDBG_FACILITY		NFSDBG_PROC
 
 extern struct rpc_procinfo nfs3_procedures[];
@@ -41,27 +43,14 @@ nfs3_rpc_wrapper(struct rpc_clnt *clnt, struct rpc_message *msg, int flags)
 	return res;
 }
 
-static inline int
-nfs3_rpc_call_wrapper(struct rpc_clnt *clnt, u32 proc, void *argp, void *resp, int flags)
-{
-	struct rpc_message msg = {
-		.rpc_proc	= &clnt->cl_procinfo[proc],
-		.rpc_argp	= argp,
-		.rpc_resp	= resp,
-	};
-	return nfs3_rpc_wrapper(clnt, &msg, flags);
-}
-
-#define rpc_call(clnt, proc, argp, resp, flags) \
-		nfs3_rpc_call_wrapper(clnt, proc, argp, resp, flags)
-#define rpc_call_sync(clnt, msg, flags) \
-		nfs3_rpc_wrapper(clnt, msg, flags)
+#define rpc_call_sync(clnt, msg, flags)	nfs3_rpc_wrapper(clnt, msg, flags)
 
 static int
-nfs3_async_handle_jukebox(struct rpc_task *task)
+nfs3_async_handle_jukebox(struct rpc_task *task, struct inode *inode)
 {
 	if (task->tk_status != -EJUKEBOX)
 		return 0;
+	nfs_inc_stats(inode, NFSIOS_DELAY);
 	task->tk_status = 0;
 	rpc_restart_call(task);
 	rpc_delay(task, NFS_JUKEBOX_RETRY_TIME);
@@ -72,14 +61,21 @@ static int
 do_proc_get_root(struct rpc_clnt *client, struct nfs_fh *fhandle,
 		 struct nfs_fsinfo *info)
 {
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_FSINFO],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= info,
+	};
 	int	status;
 
 	dprintk("%s: call  fsinfo\n", __FUNCTION__);
 	nfs_fattr_init(info->fattr);
-	status = rpc_call(client, NFS3PROC_FSINFO, fhandle, info, 0);
+	status = rpc_call_sync(client, &msg, 0);
 	dprintk("%s: reply fsinfo: %d\n", __FUNCTION__, status);
 	if (!(info->fattr->valid & NFS_ATTR_FATTR)) {
-		status = rpc_call(client, NFS3PROC_GETATTR, fhandle, info->fattr, 0);
+		msg.rpc_proc = &nfs3_procedures[NFS3PROC_GETATTR];
+		msg.rpc_resp = info->fattr;
+		status = rpc_call_sync(client, &msg, 0);
 		dprintk("%s: reply getattr: %d\n", __FUNCTION__, status);
 	}
 	return status;
@@ -107,12 +103,16 @@ static int
 nfs3_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle,
 		struct nfs_fattr *fattr)
 {
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_GETATTR],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= fattr,
+	};
 	int	status;
 
 	dprintk("NFS call  getattr\n");
 	nfs_fattr_init(fattr);
-	status = rpc_call(server->client, NFS3PROC_GETATTR,
-			  fhandle, fattr, 0);
+	status = rpc_call_sync(server->client, &msg, 0);
 	dprintk("NFS reply getattr: %d\n", status);
 	return status;
 }
@@ -126,11 +126,16 @@ nfs3_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 		.fh		= NFS_FH(inode),
 		.sattr		= sattr,
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_SETATTR],
+		.rpc_argp	= &arg,
+		.rpc_resp	= fattr,
+	};
 	int	status;
 
 	dprintk("NFS call  setattr\n");
 	nfs_fattr_init(fattr);
-	status = rpc_call(NFS_CLIENT(inode), NFS3PROC_SETATTR, &arg, fattr, 0);
+	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
 	if (status == 0)
 		nfs_setattr_update_inode(inode, sattr);
 	dprintk("NFS reply setattr: %d\n", status);
@@ -152,15 +157,23 @@ nfs3_proc_lookup(struct inode *dir, struct qstr *name,
 		.fh		= fhandle,
 		.fattr		= fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_LOOKUP],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	dprintk("NFS call  lookup %s\n", name->name);
 	nfs_fattr_init(&dir_attr);
 	nfs_fattr_init(fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_LOOKUP, &arg, &res, 0);
-	if (status >= 0 && !(fattr->valid & NFS_ATTR_FATTR))
-		status = rpc_call(NFS_CLIENT(dir), NFS3PROC_GETATTR,
-			 fhandle, fattr, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
+	if (status >= 0 && !(fattr->valid & NFS_ATTR_FATTR)) {
+		msg.rpc_proc = &nfs3_procedures[NFS3PROC_GETATTR];
+		msg.rpc_argp = fhandle;
+		msg.rpc_resp = fattr;
+		status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
+	}
 	dprintk("NFS reply lookup: %d\n", status);
 	if (status >= 0)
 		status = nfs_refresh_inode(dir, &dir_attr);
@@ -180,7 +193,7 @@ static int nfs3_proc_access(struct inode *inode, struct nfs_access_entry *entry)
 		.rpc_proc	= &nfs3_procedures[NFS3PROC_ACCESS],
 		.rpc_argp	= &arg,
 		.rpc_resp	= &res,
-		.rpc_cred	= entry->cred
+		.rpc_cred	= entry->cred,
 	};
 	int mode = entry->mask;
 	int status;
@@ -226,12 +239,16 @@ static int nfs3_proc_readlink(struct inode *inode, struct page *page,
 		.pglen		= pglen,
 		.pages		= &page
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_READLINK],
+		.rpc_argp	= &args,
+		.rpc_resp	= &fattr,
+	};
 	int			status;
 
 	dprintk("NFS call  readlink\n");
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(inode), NFS3PROC_READLINK,
-			  &args, &fattr, 0);
+	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
 	nfs_refresh_inode(inode, &fattr);
 	dprintk("NFS reply readlink: %d\n", status);
 	return status;
@@ -327,6 +344,11 @@ nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_CREATE],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	mode_t mode = sattr->ia_mode;
 	int status;
 
@@ -343,8 +365,8 @@ nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 again:
 	nfs_fattr_init(&dir_attr);
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_CREATE, &arg, &res, 0);
-	nfs_post_op_update_inode(dir, &dir_attr);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
+	nfs_refresh_inode(dir, &dir_attr);
 
 	/* If the server doesn't support the exclusive creation semantics,
 	 * try again with simple 'guarded' mode. */
@@ -447,7 +469,7 @@ nfs3_proc_unlink_done(struct dentry *dir, struct rpc_task *task)
 	struct rpc_message *msg = &task->tk_msg;
 	struct nfs_fattr	*dir_attr;
 
-	if (nfs3_async_handle_jukebox(task))
+	if (nfs3_async_handle_jukebox(task, dir->d_inode))
 		return 1;
 	if (msg->rpc_argp) {
 		dir_attr = (struct nfs_fattr*)msg->rpc_resp;
@@ -474,12 +496,17 @@ nfs3_proc_rename(struct inode *old_dir, struct qstr *old_name,
 		.fromattr	= &old_dir_attr,
 		.toattr		= &new_dir_attr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_RENAME],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	dprintk("NFS call  rename %s -> %s\n", old_name->name, new_name->name);
 	nfs_fattr_init(&old_dir_attr);
 	nfs_fattr_init(&new_dir_attr);
-	status = rpc_call(NFS_CLIENT(old_dir), NFS3PROC_RENAME, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(old_dir), &msg, 0);
 	nfs_post_op_update_inode(old_dir, &old_dir_attr);
 	nfs_post_op_update_inode(new_dir, &new_dir_attr);
 	dprintk("NFS reply rename: %d\n", status);
@@ -500,12 +527,17 @@ nfs3_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
 		.dir_attr	= &dir_attr,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_LINK],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	dprintk("NFS call  link %s\n", name->name);
 	nfs_fattr_init(&dir_attr);
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(inode), NFS3PROC_LINK, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
 	nfs_post_op_update_inode(dir, &dir_attr);
 	nfs_post_op_update_inode(inode, &fattr);
 	dprintk("NFS reply link: %d\n", status);
@@ -531,6 +563,11 @@ nfs3_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path,
 		.fh		= fhandle,
 		.fattr		= fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_SYMLINK],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	if (path->len > NFS3_MAXPATHLEN)
@@ -538,7 +575,7 @@ nfs3_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path,
 	dprintk("NFS call  symlink %s -> %s\n", name->name, path->name);
 	nfs_fattr_init(&dir_attr);
 	nfs_fattr_init(fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_SYMLINK, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_post_op_update_inode(dir, &dir_attr);
 	dprintk("NFS reply symlink: %d\n", status);
 	return status;
@@ -560,6 +597,11 @@ nfs3_proc_mkdir(struct inode *dir, struct dentry *dentry, struct iattr *sattr)
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_MKDIR],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int mode = sattr->ia_mode;
 	int status;
 
@@ -569,7 +611,7 @@ nfs3_proc_mkdir(struct inode *dir, struct dentry *dentry, struct iattr *sattr)
 
 	nfs_fattr_init(&dir_attr);
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_MKDIR, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_post_op_update_inode(dir, &dir_attr);
 	if (status != 0)
 		goto out;
@@ -591,11 +633,16 @@ nfs3_proc_rmdir(struct inode *dir, struct qstr *name)
 		.name		= name->name,
 		.len		= name->len
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_RMDIR],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &dir_attr,
+	};
 	int			status;
 
 	dprintk("NFS call  rmdir %s\n", name->name);
 	nfs_fattr_init(&dir_attr);
-	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_RMDIR, &arg, &dir_attr, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_post_op_update_inode(dir, &dir_attr);
 	dprintk("NFS reply rmdir: %d\n", status);
 	return status;
@@ -672,6 +719,11 @@ nfs3_proc_mknod(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		.fh		= &fh,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_MKNOD],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	mode_t mode = sattr->ia_mode;
 	int status;
 
@@ -690,7 +742,7 @@ nfs3_proc_mknod(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 
 	nfs_fattr_init(&dir_attr);
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFS3PROC_MKNOD, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_post_op_update_inode(dir, &dir_attr);
 	if (status != 0)
 		goto out;
@@ -707,11 +759,16 @@ static int
 nfs3_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle,
 		 struct nfs_fsstat *stat)
 {
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_FSSTAT],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= stat,
+	};
 	int	status;
 
 	dprintk("NFS call  fsstat\n");
 	nfs_fattr_init(stat->fattr);
-	status = rpc_call(server->client, NFS3PROC_FSSTAT, fhandle, stat, 0);
+	status = rpc_call_sync(server->client, &msg, 0);
 	dprintk("NFS reply statfs: %d\n", status);
 	return status;
 }
@@ -720,11 +777,16 @@ static int
 nfs3_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
 		 struct nfs_fsinfo *info)
 {
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_FSINFO],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= info,
+	};
 	int	status;
 
 	dprintk("NFS call  fsinfo\n");
 	nfs_fattr_init(info->fattr);
-	status = rpc_call(server->client_sys, NFS3PROC_FSINFO, fhandle, info, 0);
+	status = rpc_call_sync(server->client_sys, &msg, 0);
 	dprintk("NFS reply fsinfo: %d\n", status);
 	return status;
 }
@@ -733,40 +795,34 @@ static int
 nfs3_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 		   struct nfs_pathconf *info)
 {
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs3_procedures[NFS3PROC_PATHCONF],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= info,
+	};
 	int	status;
 
 	dprintk("NFS call  pathconf\n");
 	nfs_fattr_init(info->fattr);
-	status = rpc_call(server->client, NFS3PROC_PATHCONF, fhandle, info, 0);
+	status = rpc_call_sync(server->client, &msg, 0);
 	dprintk("NFS reply pathconf: %d\n", status);
 	return status;
 }
 
 extern u32 *nfs3_decode_dirent(u32 *, struct nfs_entry *, int);
 
-static void nfs3_read_done(struct rpc_task *task, void *calldata)
+static int nfs3_read_done(struct rpc_task *task, struct nfs_read_data *data)
 {
-	struct nfs_read_data *data = calldata;
-
-	if (nfs3_async_handle_jukebox(task))
-		return;
+	if (nfs3_async_handle_jukebox(task, data->inode))
+		return -EAGAIN;
 	/* Call back common NFS readpage processing */
 	if (task->tk_status >= 0)
 		nfs_refresh_inode(data->inode, &data->fattr);
-	nfs_readpage_result(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs3_read_ops = {
-	.rpc_call_done = nfs3_read_done,
-	.rpc_release = nfs_readdata_release,
-};
-
-static void
-nfs3_proc_read_setup(struct nfs_read_data *data)
+static void nfs3_proc_read_setup(struct nfs_read_data *data)
 {
-	struct rpc_task		*task = &data->task;
-	struct inode		*inode = data->inode;
-	int			flags;
 	struct rpc_message	msg = {
 		.rpc_proc	= &nfs3_procedures[NFS3PROC_READ],
 		.rpc_argp	= &data->args,
@@ -774,37 +830,20 @@ nfs3_proc_read_setup(struct nfs_read_data *data)
 		.rpc_cred	= data->cred,
 	};
 
-	/* N.B. Do we need to test? Never called for swapfile inode */
-	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
-
-	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs3_read_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
-static void nfs3_write_done(struct rpc_task *task, void *calldata)
+static int nfs3_write_done(struct rpc_task *task, struct nfs_write_data *data)
 {
-	struct nfs_write_data *data = calldata;
-
-	if (nfs3_async_handle_jukebox(task))
-		return;
+	if (nfs3_async_handle_jukebox(task, data->inode))
+		return -EAGAIN;
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(data->inode, data->res.fattr);
-	nfs_writeback_done(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs3_write_ops = {
-	.rpc_call_done = nfs3_write_done,
-	.rpc_release = nfs_writedata_release,
-};
-
-static void
-nfs3_proc_write_setup(struct nfs_write_data *data, int how)
+static void nfs3_proc_write_setup(struct nfs_write_data *data, int how)
 {
-	struct rpc_task		*task = &data->task;
-	struct inode		*inode = data->inode;
-	int			stable;
-	int			flags;
 	struct rpc_message	msg = {
 		.rpc_proc	= &nfs3_procedures[NFS3PROC_WRITE],
 		.rpc_argp	= &data->args,
@@ -812,45 +851,28 @@ nfs3_proc_write_setup(struct nfs_write_data *data, int how)
 		.rpc_cred	= data->cred,
 	};
 
+	data->args.stable = NFS_UNSTABLE;
 	if (how & FLUSH_STABLE) {
-		if (!NFS_I(inode)->ncommit)
-			stable = NFS_FILE_SYNC;
-		else
-			stable = NFS_DATA_SYNC;
-	} else
-		stable = NFS_UNSTABLE;
-	data->args.stable = stable;
-
-	/* Set the initial flags for the task.  */
-	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
+		data->args.stable = NFS_FILE_SYNC;
+		if (NFS_I(data->inode)->ncommit)
+			data->args.stable = NFS_DATA_SYNC;
+	}
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs3_write_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
-static void nfs3_commit_done(struct rpc_task *task, void *calldata)
+static int nfs3_commit_done(struct rpc_task *task, struct nfs_write_data *data)
 {
-	struct nfs_write_data *data = calldata;
-
-	if (nfs3_async_handle_jukebox(task))
-		return;
+	if (nfs3_async_handle_jukebox(task, data->inode))
+		return -EAGAIN;
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(data->inode, data->res.fattr);
-	nfs_commit_done(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs3_commit_ops = {
-	.rpc_call_done = nfs3_commit_done,
-	.rpc_release = nfs_commit_release,
-};
-
-static void
-nfs3_proc_commit_setup(struct nfs_write_data *data, int how)
+static void nfs3_proc_commit_setup(struct nfs_write_data *data, int how)
 {
-	struct rpc_task		*task = &data->task;
-	struct inode		*inode = data->inode;
-	int			flags;
 	struct rpc_message	msg = {
 		.rpc_proc	= &nfs3_procedures[NFS3PROC_COMMIT],
 		.rpc_argp	= &data->args,
@@ -858,12 +880,7 @@ nfs3_proc_commit_setup(struct nfs_write_data *data, int how)
 		.rpc_cred	= data->cred,
 	};
 
-	/* Set the initial flags for the task.  */
-	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
-
-	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs3_commit_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
 static int
@@ -902,8 +919,11 @@ struct nfs_rpc_ops	nfs_v3_clientops = {
 	.pathconf	= nfs3_proc_pathconf,
 	.decode_dirent	= nfs3_decode_dirent,
 	.read_setup	= nfs3_proc_read_setup,
+	.read_done	= nfs3_read_done,
 	.write_setup	= nfs3_proc_write_setup,
+	.write_done	= nfs3_write_done,
 	.commit_setup	= nfs3_proc_commit_setup,
+	.commit_done	= nfs3_commit_done,
 	.file_open	= nfs_open,
 	.file_release	= nfs_release,
 	.lock		= nfs3_proc_lock,
diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c
index b6c0b5012bce..ec233619687e 100644
--- a/fs/nfs/nfs3xdr.c
+++ b/fs/nfs/nfs3xdr.c
@@ -1109,7 +1109,9 @@ nfs3_xdr_setaclres(struct rpc_rqst *req, u32 *p, struct nfs_fattr *fattr)
 	.p_encode    = (kxdrproc_t) nfs3_xdr_##argtype,			\
 	.p_decode    = (kxdrproc_t) nfs3_xdr_##restype,			\
 	.p_bufsiz    = MAX(NFS3_##argtype##_sz,NFS3_##restype##_sz) << 2,	\
-	.p_timer     = timer						\
+	.p_timer     = timer,						\
+	.p_statidx   = NFS3PROC_##proc,					\
+	.p_name      = #proc,						\
 	}
 
 struct rpc_procinfo	nfs3_procedures[] = {
@@ -1138,7 +1140,7 @@ struct rpc_procinfo	nfs3_procedures[] = {
 
 struct rpc_version		nfs_version3 = {
 	.number			= 3,
-	.nrprocs		= sizeof(nfs3_procedures)/sizeof(nfs3_procedures[0]),
+	.nrprocs		= ARRAY_SIZE(nfs3_procedures),
 	.procs			= nfs3_procedures
 };
 
@@ -1150,6 +1152,7 @@ static struct rpc_procinfo	nfs3_acl_procedures[] = {
 		.p_decode = (kxdrproc_t) nfs3_xdr_getaclres,
 		.p_bufsiz = MAX(ACL3_getaclargs_sz, ACL3_getaclres_sz) << 2,
 		.p_timer = 1,
+		.p_name = "GETACL",
 	},
 	[ACLPROC3_SETACL] = {
 		.p_proc = ACLPROC3_SETACL,
@@ -1157,6 +1160,7 @@ static struct rpc_procinfo	nfs3_acl_procedures[] = {
 		.p_decode = (kxdrproc_t) nfs3_xdr_setaclres,
 		.p_bufsiz = MAX(ACL3_setaclargs_sz, ACL3_setaclres_sz) << 2,
 		.p_timer = 0,
+		.p_name = "SETACL",
 	},
 };
 
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index f8c0066e02e1..47ece1dd3c67 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -51,6 +51,7 @@
 
 #include "nfs4_fs.h"
 #include "delegation.h"
+#include "iostat.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PROC
 
@@ -335,7 +336,7 @@ static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data
 	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
 		goto out;
 	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
-	if (inode == NULL)
+	if (IS_ERR(inode))
 		goto out;
 	state = nfs4_get_open_state(inode, data->owner);
 	if (state == NULL)
@@ -604,11 +605,14 @@ static int _nfs4_proc_open_confirm(struct nfs4_opendata *data)
 	int status;
 
 	atomic_inc(&data->count);
+	/*
+	 * If rpc_run_task() ends up calling ->rpc_release(), we
+	 * want to ensure that it takes the 'error' code path.
+	 */
+	data->rpc_status = -ENOMEM;
 	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_open_confirm_ops, data);
-	if (IS_ERR(task)) {
-		nfs4_opendata_free(data);
+	if (IS_ERR(task))
 		return PTR_ERR(task);
-	}
 	status = nfs4_wait_for_completion_rpc_task(task);
 	if (status != 0) {
 		data->cancelled = 1;
@@ -707,11 +711,14 @@ static int _nfs4_proc_open(struct nfs4_opendata *data)
 	int status;
 
 	atomic_inc(&data->count);
+	/*
+	 * If rpc_run_task() ends up calling ->rpc_release(), we
+	 * want to ensure that it takes the 'error' code path.
+	 */
+	data->rpc_status = -ENOMEM;
 	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_open_ops, data);
-	if (IS_ERR(task)) {
-		nfs4_opendata_free(data);
+	if (IS_ERR(task))
 		return PTR_ERR(task);
-	}
 	status = nfs4_wait_for_completion_rpc_task(task);
 	if (status != 0) {
 		data->cancelled = 1;
@@ -908,7 +915,7 @@ out_put_state_owner:
 static struct nfs4_state *nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred *cred)
 {
 	struct nfs4_exception exception = { };
-	struct nfs4_state *res;
+	struct nfs4_state *res = ERR_PTR(-EIO);
 	int err;
 
 	do {
@@ -1017,12 +1024,12 @@ static struct nfs4_state *nfs4_do_open(struct inode *dir, struct dentry *dentry,
 	return res;
 }
 
-static int _nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
-                struct nfs_fh *fhandle, struct iattr *sattr,
-                struct nfs4_state *state)
+static int _nfs4_do_setattr(struct inode *inode, struct nfs_fattr *fattr,
+                struct iattr *sattr, struct nfs4_state *state)
 {
+	struct nfs_server *server = NFS_SERVER(inode);
         struct nfs_setattrargs  arg = {
-                .fh             = fhandle,
+                .fh             = NFS_FH(inode),
                 .iap            = sattr,
 		.server		= server,
 		.bitmask = server->attr_bitmask,
@@ -1041,7 +1048,9 @@ static int _nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
 
 	nfs_fattr_init(fattr);
 
-	if (state != NULL) {
+	if (nfs4_copy_delegation_stateid(&arg.stateid, inode)) {
+		/* Use that stateid */
+	} else if (state != NULL) {
 		msg.rpc_cred = state->owner->so_cred;
 		nfs4_copy_stateid(&arg.stateid, state, current->files);
 	} else
@@ -1053,16 +1062,15 @@ static int _nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
 	return status;
 }
 
-static int nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
-                struct nfs_fh *fhandle, struct iattr *sattr,
-                struct nfs4_state *state)
+static int nfs4_do_setattr(struct inode *inode, struct nfs_fattr *fattr,
+                struct iattr *sattr, struct nfs4_state *state)
 {
+	struct nfs_server *server = NFS_SERVER(inode);
 	struct nfs4_exception exception = { };
 	int err;
 	do {
 		err = nfs4_handle_exception(server,
-				_nfs4_do_setattr(server, fattr, fhandle, sattr,
-					state),
+				_nfs4_do_setattr(inode, fattr, sattr, state),
 				&exception);
 	} while (exception.retry);
 	return err;
@@ -1503,8 +1511,7 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 	if (ctx != NULL)
 		state = ctx->state;
 
-	status = nfs4_do_setattr(NFS_SERVER(inode), fattr,
-			NFS_FH(inode), sattr, state);
+	status = nfs4_do_setattr(inode, fattr, sattr, state);
 	if (status == 0)
 		nfs_setattr_update_inode(inode, sattr);
 	if (ctx != NULL)
@@ -1823,8 +1830,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 	d_instantiate(dentry, igrab(state->inode));
 	if (flags & O_EXCL) {
 		struct nfs_fattr fattr;
-		status = nfs4_do_setattr(NFS_SERVER(dir), &fattr,
-		                     NFS_FH(state->inode), sattr, state);
+		status = nfs4_do_setattr(state->inode, &fattr, sattr, state);
 		if (status == 0)
 			nfs_setattr_update_inode(state->inode, sattr);
 	}
@@ -2344,75 +2350,50 @@ static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 	return err;
 }
 
-static void nfs4_read_done(struct rpc_task *task, void *calldata)
+static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
 {
-	struct nfs_read_data *data = calldata;
-	struct inode *inode = data->inode;
+	struct nfs_server *server = NFS_SERVER(data->inode);
 
-	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
+	if (nfs4_async_handle_error(task, server) == -EAGAIN) {
 		rpc_restart_call(task);
-		return;
+		return -EAGAIN;
 	}
 	if (task->tk_status > 0)
-		renew_lease(NFS_SERVER(inode), data->timestamp);
-	/* Call back common NFS readpage processing */
-	nfs_readpage_result(task, calldata);
+		renew_lease(server, data->timestamp);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs4_read_ops = {
-	.rpc_call_done = nfs4_read_done,
-	.rpc_release = nfs_readdata_release,
-};
-
-static void
-nfs4_proc_read_setup(struct nfs_read_data *data)
+static void nfs4_proc_read_setup(struct nfs_read_data *data)
 {
-	struct rpc_task	*task = &data->task;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ],
 		.rpc_argp = &data->args,
 		.rpc_resp = &data->res,
 		.rpc_cred = data->cred,
 	};
-	struct inode *inode = data->inode;
-	int flags;
 
 	data->timestamp   = jiffies;
 
-	/* N.B. Do we need to test? Never called for swapfile inode */
-	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
-
-	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs4_read_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
-static void nfs4_write_done(struct rpc_task *task, void *calldata)
+static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
 {
-	struct nfs_write_data *data = calldata;
 	struct inode *inode = data->inode;
 	
 	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
 		rpc_restart_call(task);
-		return;
+		return -EAGAIN;
 	}
 	if (task->tk_status >= 0) {
 		renew_lease(NFS_SERVER(inode), data->timestamp);
 		nfs_post_op_update_inode(inode, data->res.fattr);
 	}
-	/* Call back common NFS writeback processing */
-	nfs_writeback_done(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs4_write_ops = {
-	.rpc_call_done = nfs4_write_done,
-	.rpc_release = nfs_writedata_release,
-};
-
-static void
-nfs4_proc_write_setup(struct nfs_write_data *data, int how)
+static void nfs4_proc_write_setup(struct nfs_write_data *data, int how)
 {
-	struct rpc_task	*task = &data->task;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_WRITE],
 		.rpc_argp = &data->args,
@@ -2422,7 +2403,6 @@ nfs4_proc_write_setup(struct nfs_write_data *data, int how)
 	struct inode *inode = data->inode;
 	struct nfs_server *server = NFS_SERVER(inode);
 	int stable;
-	int flags;
 	
 	if (how & FLUSH_STABLE) {
 		if (!NFS_I(inode)->ncommit)
@@ -2437,57 +2417,37 @@ nfs4_proc_write_setup(struct nfs_write_data *data, int how)
 
 	data->timestamp   = jiffies;
 
-	/* Set the initial flags for the task.  */
-	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
-
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs4_write_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
-static void nfs4_commit_done(struct rpc_task *task, void *calldata)
+static int nfs4_commit_done(struct rpc_task *task, struct nfs_write_data *data)
 {
-	struct nfs_write_data *data = calldata;
 	struct inode *inode = data->inode;
 	
 	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
 		rpc_restart_call(task);
-		return;
+		return -EAGAIN;
 	}
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(inode, data->res.fattr);
-	/* Call back common NFS writeback processing */
-	nfs_commit_done(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs4_commit_ops = {
-	.rpc_call_done = nfs4_commit_done,
-	.rpc_release = nfs_commit_release,
-};
-
-static void
-nfs4_proc_commit_setup(struct nfs_write_data *data, int how)
+static void nfs4_proc_commit_setup(struct nfs_write_data *data, int how)
 {
-	struct rpc_task	*task = &data->task;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_COMMIT],
 		.rpc_argp = &data->args,
 		.rpc_resp = &data->res,
 		.rpc_cred = data->cred,
 	};	
-	struct inode *inode = data->inode;
-	struct nfs_server *server = NFS_SERVER(inode);
-	int flags;
+	struct nfs_server *server = NFS_SERVER(data->inode);
 	
 	data->args.bitmask = server->attr_bitmask;
 	data->res.server = server;
 
-	/* Set the initial flags for the task.  */
-	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
-
-	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs4_commit_ops, data);
-	rpc_call_setup(task, &msg, 0);	
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
 /*
@@ -2755,8 +2715,10 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
 				rpc_wake_up_task(task);
 			task->tk_status = 0;
 			return -EAGAIN;
-		case -NFS4ERR_GRACE:
 		case -NFS4ERR_DELAY:
+			nfs_inc_server_stats((struct nfs_server *) server,
+						NFSIOS_DELAY);
+		case -NFS4ERR_GRACE:
 			rpc_delay(task, NFS4_POLL_RETRY_MAX);
 			task->tk_status = 0;
 			return -EAGAIN;
@@ -2893,8 +2855,7 @@ int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short p
 	return status;
 }
 
-int
-nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred)
+static int _nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred)
 {
 	struct nfs_fsinfo fsinfo;
 	struct rpc_message msg = {
@@ -2918,6 +2879,24 @@ nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred)
 	return status;
 }
 
+int nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred)
+{
+	long timeout;
+	int err;
+	do {
+		err = _nfs4_proc_setclientid_confirm(clp, cred);
+		switch (err) {
+			case 0:
+				return err;
+			case -NFS4ERR_RESOURCE:
+				/* The IBM lawyers misread another document! */
+			case -NFS4ERR_DELAY:
+				err = nfs4_delay(clp->cl_rpcclient, &timeout);
+		}
+	} while (err == 0);
+	return err;
+}
+
 struct nfs4_delegreturndata {
 	struct nfs4_delegreturnargs args;
 	struct nfs4_delegreturnres res;
@@ -2958,7 +2937,7 @@ static void nfs4_delegreturn_release(void *calldata)
 	kfree(calldata);
 }
 
-const static struct rpc_call_ops nfs4_delegreturn_ops = {
+static const struct rpc_call_ops nfs4_delegreturn_ops = {
 	.rpc_call_prepare = nfs4_delegreturn_prepare,
 	.rpc_call_done = nfs4_delegreturn_done,
 	.rpc_release = nfs4_delegreturn_release,
@@ -2986,10 +2965,8 @@ static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, co
 	data->rpc_status = 0;
 
 	task = rpc_run_task(NFS_CLIENT(inode), RPC_TASK_ASYNC, &nfs4_delegreturn_ops, data);
-	if (IS_ERR(task)) {
-		nfs4_delegreturn_release(data);
+	if (IS_ERR(task))
 		return PTR_ERR(task);
-	}
 	status = nfs4_wait_for_completion_rpc_task(task);
 	if (status == 0) {
 		status = data->rpc_status;
@@ -3209,7 +3186,6 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
 		struct nfs_seqid *seqid)
 {
 	struct nfs4_unlockdata *data;
-	struct rpc_task *task;
 
 	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
 	if (data == NULL) {
@@ -3219,10 +3195,7 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
 
 	/* Unlock _before_ we do the RPC call */
 	do_vfs_lock(fl->fl_file, fl);
-	task = rpc_run_task(NFS_CLIENT(lsp->ls_state->inode), RPC_TASK_ASYNC, &nfs4_locku_ops, data);
-	if (IS_ERR(task))
-		nfs4_locku_release_calldata(data);
-	return task;
+	return rpc_run_task(NFS_CLIENT(lsp->ls_state->inode), RPC_TASK_ASYNC, &nfs4_locku_ops, data);
 }
 
 static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
@@ -3403,10 +3376,8 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *f
 		data->arg.reclaim = 1;
 	task = rpc_run_task(NFS_CLIENT(state->inode), RPC_TASK_ASYNC,
 			&nfs4_lock_ops, data);
-	if (IS_ERR(task)) {
-		nfs4_lock_release(data);
+	if (IS_ERR(task))
 		return PTR_ERR(task);
-	}
 	ret = nfs4_wait_for_completion_rpc_task(task);
 	if (ret == 0) {
 		ret = data->rpc_status;
@@ -3588,6 +3559,8 @@ ssize_t nfs4_listxattr(struct dentry *dentry, char *buf, size_t buflen)
 {
 	size_t len = strlen(XATTR_NAME_NFSV4_ACL) + 1;
 
+	if (!nfs4_server_supports_acls(NFS_SERVER(dentry->d_inode)))
+		return 0;
 	if (buf && buflen < len)
 		return -ERANGE;
 	if (buf)
@@ -3644,8 +3617,11 @@ struct nfs_rpc_ops	nfs_v4_clientops = {
 	.pathconf	= nfs4_proc_pathconf,
 	.decode_dirent	= nfs4_decode_dirent,
 	.read_setup	= nfs4_proc_read_setup,
+	.read_done	= nfs4_read_done,
 	.write_setup	= nfs4_proc_write_setup,
+	.write_done	= nfs4_write_done,
 	.commit_setup	= nfs4_proc_commit_setup,
+	.commit_done	= nfs4_commit_done,
 	.file_open      = nfs_open,
 	.file_release   = nfs_release,
 	.lock		= nfs4_proc_lock,
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index afad0255e7db..96e5b82c153b 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -977,6 +977,7 @@ out:
 out_error:
 	printk(KERN_WARNING "Error: state recovery failed on NFSv4 server %u.%u.%u.%u with error %d\n",
 				NIPQUAD(clp->cl_addr.s_addr), -status);
+	set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
 	goto out;
 }
 
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 4bbf5ef57785..7c5d70efe720 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -4344,6 +4344,8 @@ nfs_stat_to_errno(int stat)
 	.p_encode = (kxdrproc_t) nfs4_xdr_##argtype,		\
 	.p_decode = (kxdrproc_t) nfs4_xdr_##restype,		\
 	.p_bufsiz = MAX(NFS4_##argtype##_sz,NFS4_##restype##_sz) << 2,	\
+	.p_statidx = NFSPROC4_CLNT_##proc,			\
+	.p_name   = #proc,					\
     }
 
 struct rpc_procinfo	nfs4_procedures[] = {
@@ -4384,7 +4386,7 @@ struct rpc_procinfo	nfs4_procedures[] = {
 
 struct rpc_version		nfs_version4 = {
 	.number			= 4,
-	.nrprocs		= sizeof(nfs4_procedures)/sizeof(nfs4_procedures[0]),
+	.nrprocs		= ARRAY_SIZE(nfs4_procedures),
 	.procs			= nfs4_procedures
 };
 
diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c
index d53857b148e2..106aca388ebc 100644
--- a/fs/nfs/pagelist.c
+++ b/fs/nfs/pagelist.c
@@ -85,6 +85,9 @@ nfs_create_request(struct nfs_open_context *ctx, struct inode *inode,
 	atomic_set(&req->wb_complete, 0);
 	req->wb_index	= page->index;
 	page_cache_get(page);
+	BUG_ON(PagePrivate(page));
+	BUG_ON(!PageLocked(page));
+	BUG_ON(page->mapping->host != inode);
 	req->wb_offset  = offset;
 	req->wb_pgbase	= offset;
 	req->wb_bytes   = count;
@@ -132,9 +135,11 @@ void nfs_clear_page_writeback(struct nfs_page *req)
 {
 	struct nfs_inode *nfsi = NFS_I(req->wb_context->dentry->d_inode);
 
-	spin_lock(&nfsi->req_lock);
-	radix_tree_tag_clear(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_WRITEBACK);
-	spin_unlock(&nfsi->req_lock);
+	if (req->wb_page != NULL) {
+		spin_lock(&nfsi->req_lock);
+		radix_tree_tag_clear(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_WRITEBACK);
+		spin_unlock(&nfsi->req_lock);
+	}
 	nfs_unlock_request(req);
 }
 
@@ -147,8 +152,9 @@ void nfs_clear_page_writeback(struct nfs_page *req)
  */
 void nfs_clear_request(struct nfs_page *req)
 {
-	if (req->wb_page) {
-		page_cache_release(req->wb_page);
+	struct page *page = req->wb_page;
+	if (page != NULL) {
+		page_cache_release(page);
 		req->wb_page = NULL;
 	}
 }
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index f5150d71c03d..9dd85cac2df0 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -58,16 +58,23 @@ nfs_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
 {
 	struct nfs_fattr *fattr = info->fattr;
 	struct nfs2_fsstat fsinfo;
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_GETATTR],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= fattr,
+	};
 	int status;
 
 	dprintk("%s: call getattr\n", __FUNCTION__);
 	nfs_fattr_init(fattr);
-	status = rpc_call(server->client_sys, NFSPROC_GETATTR, fhandle, fattr, 0);
+	status = rpc_call_sync(server->client_sys, &msg, 0);
 	dprintk("%s: reply getattr: %d\n", __FUNCTION__, status);
 	if (status)
 		return status;
 	dprintk("%s: call statfs\n", __FUNCTION__);
-	status = rpc_call(server->client_sys, NFSPROC_STATFS, fhandle, &fsinfo, 0);
+	msg.rpc_proc = &nfs_procedures[NFSPROC_STATFS];
+	msg.rpc_resp = &fsinfo;
+	status = rpc_call_sync(server->client_sys, &msg, 0);
 	dprintk("%s: reply statfs: %d\n", __FUNCTION__, status);
 	if (status)
 		return status;
@@ -90,12 +97,16 @@ static int
 nfs_proc_getattr(struct nfs_server *server, struct nfs_fh *fhandle,
 		struct nfs_fattr *fattr)
 {
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_GETATTR],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= fattr,
+	};
 	int	status;
 
 	dprintk("NFS call  getattr\n");
 	nfs_fattr_init(fattr);
-	status = rpc_call(server->client, NFSPROC_GETATTR,
-				fhandle, fattr, 0);
+	status = rpc_call_sync(server->client, &msg, 0);
 	dprintk("NFS reply getattr: %d\n", status);
 	return status;
 }
@@ -109,6 +120,11 @@ nfs_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 		.fh	= NFS_FH(inode),
 		.sattr	= sattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_SETATTR],
+		.rpc_argp	= &arg,
+		.rpc_resp	= fattr,
+	};
 	int	status;
 
 	/* Mask out the non-modebit related stuff from attr->ia_mode */
@@ -116,7 +132,7 @@ nfs_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 
 	dprintk("NFS call  setattr\n");
 	nfs_fattr_init(fattr);
-	status = rpc_call(NFS_CLIENT(inode), NFSPROC_SETATTR, &arg, fattr, 0);
+	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
 	if (status == 0)
 		nfs_setattr_update_inode(inode, sattr);
 	dprintk("NFS reply setattr: %d\n", status);
@@ -136,11 +152,16 @@ nfs_proc_lookup(struct inode *dir, struct qstr *name,
 		.fh		= fhandle,
 		.fattr		= fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_LOOKUP],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	dprintk("NFS call  lookup %s\n", name->name);
 	nfs_fattr_init(fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFSPROC_LOOKUP, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	dprintk("NFS reply lookup: %d\n", status);
 	return status;
 }
@@ -154,10 +175,14 @@ static int nfs_proc_readlink(struct inode *inode, struct page *page,
 		.pglen		= pglen,
 		.pages		= &page
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_READLINK],
+		.rpc_argp	= &args,
+	};
 	int			status;
 
 	dprintk("NFS call  readlink\n");
-	status = rpc_call(NFS_CLIENT(inode), NFSPROC_READLINK, &args, NULL, 0);
+	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
 	dprintk("NFS reply readlink: %d\n", status);
 	return status;
 }
@@ -233,11 +258,16 @@ nfs_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_CREATE],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	nfs_fattr_init(&fattr);
 	dprintk("NFS call  create %s\n", dentry->d_name.name);
-	status = rpc_call(NFS_CLIENT(dir), NFSPROC_CREATE, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	if (status == 0)
 		status = nfs_instantiate(dentry, &fhandle, &fattr);
 	dprintk("NFS reply create: %d\n", status);
@@ -263,6 +293,11 @@ nfs_proc_mknod(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_CREATE],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int status, mode;
 
 	dprintk("NFS call  mknod %s\n", dentry->d_name.name);
@@ -277,13 +312,13 @@ nfs_proc_mknod(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 	}
 
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFSPROC_CREATE, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_mark_for_revalidate(dir);
 
 	if (status == -EINVAL && S_ISFIFO(mode)) {
 		sattr->ia_mode = mode;
 		nfs_fattr_init(&fattr);
-		status = rpc_call(NFS_CLIENT(dir), NFSPROC_CREATE, &arg, &res, 0);
+		status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	}
 	if (status == 0)
 		status = nfs_instantiate(dentry, &fhandle, &fattr);
@@ -302,8 +337,6 @@ nfs_proc_remove(struct inode *dir, struct qstr *name)
 	struct rpc_message	msg = { 
 		.rpc_proc	= &nfs_procedures[NFSPROC_REMOVE],
 		.rpc_argp	= &arg,
-		.rpc_resp	= NULL,
-		.rpc_cred	= NULL
 	};
 	int			status;
 
@@ -355,10 +388,14 @@ nfs_proc_rename(struct inode *old_dir, struct qstr *old_name,
 		.toname		= new_name->name,
 		.tolen		= new_name->len
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_RENAME],
+		.rpc_argp	= &arg,
+	};
 	int			status;
 
 	dprintk("NFS call  rename %s -> %s\n", old_name->name, new_name->name);
-	status = rpc_call(NFS_CLIENT(old_dir), NFSPROC_RENAME, &arg, NULL, 0);
+	status = rpc_call_sync(NFS_CLIENT(old_dir), &msg, 0);
 	nfs_mark_for_revalidate(old_dir);
 	nfs_mark_for_revalidate(new_dir);
 	dprintk("NFS reply rename: %d\n", status);
@@ -374,10 +411,14 @@ nfs_proc_link(struct inode *inode, struct inode *dir, struct qstr *name)
 		.toname		= name->name,
 		.tolen		= name->len
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_LINK],
+		.rpc_argp	= &arg,
+	};
 	int			status;
 
 	dprintk("NFS call  link %s\n", name->name);
-	status = rpc_call(NFS_CLIENT(inode), NFSPROC_LINK, &arg, NULL, 0);
+	status = rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
 	nfs_mark_for_revalidate(inode);
 	nfs_mark_for_revalidate(dir);
 	dprintk("NFS reply link: %d\n", status);
@@ -397,6 +438,10 @@ nfs_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path,
 		.tolen		= path->len,
 		.sattr		= sattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_SYMLINK],
+		.rpc_argp	= &arg,
+	};
 	int			status;
 
 	if (path->len > NFS2_MAXPATHLEN)
@@ -404,7 +449,7 @@ nfs_proc_symlink(struct inode *dir, struct qstr *name, struct qstr *path,
 	dprintk("NFS call  symlink %s -> %s\n", name->name, path->name);
 	nfs_fattr_init(fattr);
 	fhandle->size = 0;
-	status = rpc_call(NFS_CLIENT(dir), NFSPROC_SYMLINK, &arg, NULL, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_mark_for_revalidate(dir);
 	dprintk("NFS reply symlink: %d\n", status);
 	return status;
@@ -425,11 +470,16 @@ nfs_proc_mkdir(struct inode *dir, struct dentry *dentry, struct iattr *sattr)
 		.fh		= &fhandle,
 		.fattr		= &fattr
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_MKDIR],
+		.rpc_argp	= &arg,
+		.rpc_resp	= &res,
+	};
 	int			status;
 
 	dprintk("NFS call  mkdir %s\n", dentry->d_name.name);
 	nfs_fattr_init(&fattr);
-	status = rpc_call(NFS_CLIENT(dir), NFSPROC_MKDIR, &arg, &res, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_mark_for_revalidate(dir);
 	if (status == 0)
 		status = nfs_instantiate(dentry, &fhandle, &fattr);
@@ -445,10 +495,14 @@ nfs_proc_rmdir(struct inode *dir, struct qstr *name)
 		.name		= name->name,
 		.len		= name->len
 	};
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_RMDIR],
+		.rpc_argp	= &arg,
+	};
 	int			status;
 
 	dprintk("NFS call  rmdir %s\n", name->name);
-	status = rpc_call(NFS_CLIENT(dir), NFSPROC_RMDIR, &arg, NULL, 0);
+	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_mark_for_revalidate(dir);
 	dprintk("NFS reply rmdir: %d\n", status);
 	return status;
@@ -470,13 +524,12 @@ nfs_proc_readdir(struct dentry *dentry, struct rpc_cred *cred,
 		.fh		= NFS_FH(dir),
 		.cookie		= cookie,
 		.count		= count,
-		.pages		= &page
+		.pages		= &page,
 	};
 	struct rpc_message	msg = {
 		.rpc_proc	= &nfs_procedures[NFSPROC_READDIR],
 		.rpc_argp	= &arg,
-		.rpc_resp	= NULL,
-		.rpc_cred	= cred
+		.rpc_cred	= cred,
 	};
 	int			status;
 
@@ -495,11 +548,16 @@ nfs_proc_statfs(struct nfs_server *server, struct nfs_fh *fhandle,
 			struct nfs_fsstat *stat)
 {
 	struct nfs2_fsstat fsinfo;
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_STATFS],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= &fsinfo,
+	};
 	int	status;
 
 	dprintk("NFS call  statfs\n");
 	nfs_fattr_init(stat->fattr);
-	status = rpc_call(server->client, NFSPROC_STATFS, fhandle, &fsinfo, 0);
+	status = rpc_call_sync(server->client, &msg, 0);
 	dprintk("NFS reply statfs: %d\n", status);
 	if (status)
 		goto out;
@@ -518,11 +576,16 @@ nfs_proc_fsinfo(struct nfs_server *server, struct nfs_fh *fhandle,
 			struct nfs_fsinfo *info)
 {
 	struct nfs2_fsstat fsinfo;
+	struct rpc_message msg = {
+		.rpc_proc	= &nfs_procedures[NFSPROC_STATFS],
+		.rpc_argp	= fhandle,
+		.rpc_resp	= &fsinfo,
+	};
 	int	status;
 
 	dprintk("NFS call  fsinfo\n");
 	nfs_fattr_init(info->fattr);
-	status = rpc_call(server->client, NFSPROC_STATFS, fhandle, &fsinfo, 0);
+	status = rpc_call_sync(server->client, &msg, 0);
 	dprintk("NFS reply fsinfo: %d\n", status);
 	if (status)
 		goto out;
@@ -550,10 +613,8 @@ nfs_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 
 extern u32 * nfs_decode_dirent(u32 *, struct nfs_entry *, int);
 
-static void nfs_read_done(struct rpc_task *task, void *calldata)
+static int nfs_read_done(struct rpc_task *task, struct nfs_read_data *data)
 {
-	struct nfs_read_data *data = calldata;
-
 	if (task->tk_status >= 0) {
 		nfs_refresh_inode(data->inode, data->res.fattr);
 		/* Emulate the eof flag, which isn't normally needed in NFSv2
@@ -562,20 +623,11 @@ static void nfs_read_done(struct rpc_task *task, void *calldata)
 		if (data->args.offset + data->args.count >= data->res.fattr->size)
 			data->res.eof = 1;
 	}
-	nfs_readpage_result(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs_read_ops = {
-	.rpc_call_done = nfs_read_done,
-	.rpc_release = nfs_readdata_release,
-};
-
-static void
-nfs_proc_read_setup(struct nfs_read_data *data)
+static void nfs_proc_read_setup(struct nfs_read_data *data)
 {
-	struct rpc_task		*task = &data->task;
-	struct inode		*inode = data->inode;
-	int			flags;
 	struct rpc_message	msg = {
 		.rpc_proc	= &nfs_procedures[NFSPROC_READ],
 		.rpc_argp	= &data->args,
@@ -583,34 +635,18 @@ nfs_proc_read_setup(struct nfs_read_data *data)
 		.rpc_cred	= data->cred,
 	};
 
-	/* N.B. Do we need to test? Never called for swapfile inode */
-	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
-
-	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs_read_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
-static void nfs_write_done(struct rpc_task *task, void *calldata)
+static int nfs_write_done(struct rpc_task *task, struct nfs_write_data *data)
 {
-	struct nfs_write_data *data = calldata;
-
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(data->inode, data->res.fattr);
-	nfs_writeback_done(task, calldata);
+	return 0;
 }
 
-static const struct rpc_call_ops nfs_write_ops = {
-	.rpc_call_done = nfs_write_done,
-	.rpc_release = nfs_writedata_release,
-};
-
-static void
-nfs_proc_write_setup(struct nfs_write_data *data, int how)
+static void nfs_proc_write_setup(struct nfs_write_data *data, int how)
 {
-	struct rpc_task		*task = &data->task;
-	struct inode		*inode = data->inode;
-	int			flags;
 	struct rpc_message	msg = {
 		.rpc_proc	= &nfs_procedures[NFSPROC_WRITE],
 		.rpc_argp	= &data->args,
@@ -621,12 +657,8 @@ nfs_proc_write_setup(struct nfs_write_data *data, int how)
 	/* Note: NFSv2 ignores @stable and always uses NFS_FILE_SYNC */
 	data->args.stable = NFS_FILE_SYNC;
 
-	/* Set the initial flags for the task.  */
-	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
-
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs_write_ops, data);
-	rpc_call_setup(task, &msg, 0);
+	rpc_call_setup(&data->task, &msg, 0);
 }
 
 static void
@@ -672,7 +704,9 @@ struct nfs_rpc_ops	nfs_v2_clientops = {
 	.pathconf	= nfs_proc_pathconf,
 	.decode_dirent	= nfs_decode_dirent,
 	.read_setup	= nfs_proc_read_setup,
+	.read_done	= nfs_read_done,
 	.write_setup	= nfs_proc_write_setup,
+	.write_done	= nfs_write_done,
 	.commit_setup	= nfs_proc_commit_setup,
 	.file_open	= nfs_open,
 	.file_release	= nfs_release,
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 05eb43fadf8e..624ca7146b6b 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -31,17 +31,49 @@
 
 #include <asm/system.h>
 
+#include "iostat.h"
+
 #define NFSDBG_FACILITY		NFSDBG_PAGECACHE
 
 static int nfs_pagein_one(struct list_head *, struct inode *);
-static void nfs_readpage_result_partial(struct nfs_read_data *, int);
-static void nfs_readpage_result_full(struct nfs_read_data *, int);
+static const struct rpc_call_ops nfs_read_partial_ops;
+static const struct rpc_call_ops nfs_read_full_ops;
 
 static kmem_cache_t *nfs_rdata_cachep;
-mempool_t *nfs_rdata_mempool;
+static mempool_t *nfs_rdata_mempool;
 
 #define MIN_POOL_READ	(32)
 
+struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount)
+{
+	struct nfs_read_data *p = mempool_alloc(nfs_rdata_mempool, SLAB_NOFS);
+
+	if (p) {
+		memset(p, 0, sizeof(*p));
+		INIT_LIST_HEAD(&p->pages);
+		if (pagecount < NFS_PAGEVEC_SIZE)
+			p->pagevec = &p->page_array[0];
+		else {
+			size_t size = ++pagecount * sizeof(struct page *);
+			p->pagevec = kmalloc(size, GFP_NOFS);
+			if (p->pagevec) {
+				memset(p->pagevec, 0, size);
+			} else {
+				mempool_free(p, nfs_rdata_mempool);
+				p = NULL;
+			}
+		}
+	}
+	return p;
+}
+
+void nfs_readdata_free(struct nfs_read_data *p)
+{
+	if (p && (p->pagevec != &p->page_array[0]))
+		kfree(p->pagevec);
+	mempool_free(p, nfs_rdata_mempool);
+}
+
 void nfs_readdata_release(void *data)
 {
         nfs_readdata_free(data);
@@ -133,6 +165,8 @@ static int nfs_readpage_sync(struct nfs_open_context *ctx, struct inode *inode,
 		}
 		count -= result;
 		rdata->args.pgbase += result;
+		nfs_add_stats(inode, NFSIOS_SERVERREADBYTES, result);
+
 		/* Note: result == 0 should only happen if we're caching
 		 * a write that extends the file and punches a hole.
 		 */
@@ -196,9 +230,11 @@ static void nfs_readpage_release(struct nfs_page *req)
  * Set up the NFS read request struct
  */
 static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
+		const struct rpc_call_ops *call_ops,
 		unsigned int count, unsigned int offset)
 {
 	struct inode		*inode;
+	int flags;
 
 	data->req	  = req;
 	data->inode	  = inode = req->wb_context->dentry->d_inode;
@@ -216,6 +252,9 @@ static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
 	data->res.eof     = 0;
 	nfs_fattr_init(&data->fattr);
 
+	/* Set up the initial task struct. */
+	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
+	rpc_init_task(&data->task, NFS_CLIENT(inode), flags, call_ops, data);
 	NFS_PROTO(inode)->read_setup(data);
 
 	data->task.tk_cookie = (unsigned long)inode;
@@ -303,14 +342,15 @@ static int nfs_pagein_multi(struct list_head *head, struct inode *inode)
 		list_del_init(&data->pages);
 
 		data->pagevec[0] = page;
-		data->complete = nfs_readpage_result_partial;
 
 		if (nbytes > rsize) {
-			nfs_read_rpcsetup(req, data, rsize, offset);
+			nfs_read_rpcsetup(req, data, &nfs_read_partial_ops,
+					rsize, offset);
 			offset += rsize;
 			nbytes -= rsize;
 		} else {
-			nfs_read_rpcsetup(req, data, nbytes, offset);
+			nfs_read_rpcsetup(req, data, &nfs_read_partial_ops,
+					nbytes, offset);
 			nbytes = 0;
 		}
 		nfs_execute_read(data);
@@ -356,8 +396,7 @@ static int nfs_pagein_one(struct list_head *head, struct inode *inode)
 	}
 	req = nfs_list_entry(data->pages.next);
 
-	data->complete = nfs_readpage_result_full;
-	nfs_read_rpcsetup(req, data, count, 0);
+	nfs_read_rpcsetup(req, data, &nfs_read_full_ops, count, 0);
 
 	nfs_execute_read(data);
 	return 0;
@@ -391,12 +430,15 @@ nfs_pagein_list(struct list_head *head, int rpages)
 /*
  * Handle a read reply that fills part of a page.
  */
-static void nfs_readpage_result_partial(struct nfs_read_data *data, int status)
+static void nfs_readpage_result_partial(struct rpc_task *task, void *calldata)
 {
+	struct nfs_read_data *data = calldata;
 	struct nfs_page *req = data->req;
 	struct page *page = req->wb_page;
  
-	if (status >= 0) {
+	if (nfs_readpage_result(task, data) != 0)
+		return;
+	if (task->tk_status >= 0) {
 		unsigned int request = data->args.count;
 		unsigned int result = data->res.count;
 
@@ -415,20 +457,28 @@ static void nfs_readpage_result_partial(struct nfs_read_data *data, int status)
 	}
 }
 
+static const struct rpc_call_ops nfs_read_partial_ops = {
+	.rpc_call_done = nfs_readpage_result_partial,
+	.rpc_release = nfs_readdata_release,
+};
+
 /*
  * This is the callback from RPC telling us whether a reply was
  * received or some error occurred (timeout or socket shutdown).
  */
-static void nfs_readpage_result_full(struct nfs_read_data *data, int status)
+static void nfs_readpage_result_full(struct rpc_task *task, void *calldata)
 {
+	struct nfs_read_data *data = calldata;
 	unsigned int count = data->res.count;
 
+	if (nfs_readpage_result(task, data) != 0)
+		return;
 	while (!list_empty(&data->pages)) {
 		struct nfs_page *req = nfs_list_entry(data->pages.next);
 		struct page *page = req->wb_page;
 		nfs_list_remove_request(req);
 
-		if (status >= 0) {
+		if (task->tk_status >= 0) {
 			if (count < PAGE_CACHE_SIZE) {
 				if (count < req->wb_bytes)
 					memclear_highpage_flush(page,
@@ -444,22 +494,33 @@ static void nfs_readpage_result_full(struct nfs_read_data *data, int status)
 	}
 }
 
+static const struct rpc_call_ops nfs_read_full_ops = {
+	.rpc_call_done = nfs_readpage_result_full,
+	.rpc_release = nfs_readdata_release,
+};
+
 /*
  * This is the callback from RPC telling us whether a reply was
  * received or some error occurred (timeout or socket shutdown).
  */
-void nfs_readpage_result(struct rpc_task *task, void *calldata)
+int nfs_readpage_result(struct rpc_task *task, struct nfs_read_data *data)
 {
-	struct nfs_read_data *data = calldata;
 	struct nfs_readargs *argp = &data->args;
 	struct nfs_readres *resp = &data->res;
-	int status = task->tk_status;
+	int status;
 
 	dprintk("NFS: %4d nfs_readpage_result, (status %d)\n",
-		task->tk_pid, status);
+		task->tk_pid, task->tk_status);
+
+	status = NFS_PROTO(data->inode)->read_done(task, data);
+	if (status != 0)
+		return status;
+
+	nfs_add_stats(data->inode, NFSIOS_SERVERREADBYTES, resp->count);
 
 	/* Is this a short read? */
 	if (task->tk_status >= 0 && resp->count < argp->count && !resp->eof) {
+		nfs_inc_stats(data->inode, NFSIOS_SHORTREAD);
 		/* Has the server at least made some progress? */
 		if (resp->count != 0) {
 			/* Yes, so retry the read at the end of the data */
@@ -467,14 +528,14 @@ void nfs_readpage_result(struct rpc_task *task, void *calldata)
 			argp->pgbase += resp->count;
 			argp->count -= resp->count;
 			rpc_restart_call(task);
-			return;
+			return -EAGAIN;
 		}
 		task->tk_status = -EIO;
 	}
 	spin_lock(&data->inode->i_lock);
 	NFS_I(data->inode)->cache_validity |= NFS_INO_INVALID_ATIME;
 	spin_unlock(&data->inode->i_lock);
-	data->complete(data, status);
+	return 0;
 }
 
 /*
@@ -491,6 +552,9 @@ int nfs_readpage(struct file *file, struct page *page)
 
 	dprintk("NFS: nfs_readpage (%p %ld@%lu)\n",
 		page, PAGE_CACHE_SIZE, page->index);
+	nfs_inc_stats(inode, NFSIOS_VFSREADPAGE);
+	nfs_add_stats(inode, NFSIOS_READPAGES, 1);
+
 	/*
 	 * Try to flush any pending writes to the file..
 	 *
@@ -570,6 +634,7 @@ int nfs_readpages(struct file *filp, struct address_space *mapping,
 			inode->i_sb->s_id,
 			(long long)NFS_FILEID(inode),
 			nr_pages);
+	nfs_inc_stats(inode, NFSIOS_VFSREADPAGES);
 
 	if (filp == NULL) {
 		desc.ctx = nfs_find_open_context(inode, NULL, FMODE_READ);
@@ -582,6 +647,7 @@ int nfs_readpages(struct file *filp, struct address_space *mapping,
 	if (!list_empty(&head)) {
 		int err = nfs_pagein_list(&head, server->rpages);
 		if (!ret)
+			nfs_add_stats(inode, NFSIOS_READPAGES, err);
 			ret = err;
 	}
 	put_nfs_open_context(desc.ctx);
@@ -597,10 +663,8 @@ int nfs_init_readpagecache(void)
 	if (nfs_rdata_cachep == NULL)
 		return -ENOMEM;
 
-	nfs_rdata_mempool = mempool_create(MIN_POOL_READ,
-					   mempool_alloc_slab,
-					   mempool_free_slab,
-					   nfs_rdata_cachep);
+	nfs_rdata_mempool = mempool_create_slab_pool(MIN_POOL_READ,
+						     nfs_rdata_cachep);
 	if (nfs_rdata_mempool == NULL)
 		return -ENOMEM;
 
diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c
index a65c7b53d558..0e28189c2151 100644
--- a/fs/nfs/unlink.c
+++ b/fs/nfs/unlink.c
@@ -163,10 +163,9 @@ nfs_async_unlink(struct dentry *dentry)
 	struct rpc_clnt	*clnt = NFS_CLIENT(dir->d_inode);
 	int		status = -ENOMEM;
 
-	data = kmalloc(sizeof(*data), GFP_KERNEL);
+	data = kzalloc(sizeof(*data), GFP_KERNEL);
 	if (!data)
 		goto out;
-	memset(data, 0, sizeof(*data));
 
 	data->cred = rpcauth_lookupcred(clnt->cl_auth, 0);
 	if (IS_ERR(data->cred)) {
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 9449b6835509..4cfada2cc09f 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -63,6 +63,7 @@
 #include <linux/smp_lock.h>
 
 #include "delegation.h"
+#include "iostat.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PAGECACHE
 
@@ -76,20 +77,21 @@ static struct nfs_page * nfs_update_request(struct nfs_open_context*,
 					    struct inode *,
 					    struct page *,
 					    unsigned int, unsigned int);
-static void nfs_writeback_done_partial(struct nfs_write_data *, int);
-static void nfs_writeback_done_full(struct nfs_write_data *, int);
 static int nfs_wait_on_write_congestion(struct address_space *, int);
 static int nfs_wait_on_requests(struct inode *, unsigned long, unsigned int);
 static int nfs_flush_inode(struct inode *inode, unsigned long idx_start,
 			   unsigned int npages, int how);
+static const struct rpc_call_ops nfs_write_partial_ops;
+static const struct rpc_call_ops nfs_write_full_ops;
+static const struct rpc_call_ops nfs_commit_ops;
 
 static kmem_cache_t *nfs_wdata_cachep;
-mempool_t *nfs_wdata_mempool;
+static mempool_t *nfs_wdata_mempool;
 static mempool_t *nfs_commit_mempool;
 
 static DECLARE_WAIT_QUEUE_HEAD(nfs_write_congestion);
 
-static inline struct nfs_write_data *nfs_commit_alloc(unsigned int pagecount)
+struct nfs_write_data *nfs_commit_alloc(unsigned int pagecount)
 {
 	struct nfs_write_data *p = mempool_alloc(nfs_commit_mempool, SLAB_NOFS);
 
@@ -100,11 +102,39 @@ static inline struct nfs_write_data *nfs_commit_alloc(unsigned int pagecount)
 			p->pagevec = &p->page_array[0];
 		else {
 			size_t size = ++pagecount * sizeof(struct page *);
+			p->pagevec = kzalloc(size, GFP_NOFS);
+			if (!p->pagevec) {
+				mempool_free(p, nfs_commit_mempool);
+				p = NULL;
+			}
+		}
+	}
+	return p;
+}
+
+void nfs_commit_free(struct nfs_write_data *p)
+{
+	if (p && (p->pagevec != &p->page_array[0]))
+		kfree(p->pagevec);
+	mempool_free(p, nfs_commit_mempool);
+}
+
+struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount)
+{
+	struct nfs_write_data *p = mempool_alloc(nfs_wdata_mempool, SLAB_NOFS);
+
+	if (p) {
+		memset(p, 0, sizeof(*p));
+		INIT_LIST_HEAD(&p->pages);
+		if (pagecount < NFS_PAGEVEC_SIZE)
+			p->pagevec = &p->page_array[0];
+		else {
+			size_t size = ++pagecount * sizeof(struct page *);
 			p->pagevec = kmalloc(size, GFP_NOFS);
 			if (p->pagevec) {
 				memset(p->pagevec, 0, size);
 			} else {
-				mempool_free(p, nfs_commit_mempool);
+				mempool_free(p, nfs_wdata_mempool);
 				p = NULL;
 			}
 		}
@@ -112,11 +142,11 @@ static inline struct nfs_write_data *nfs_commit_alloc(unsigned int pagecount)
 	return p;
 }
 
-static inline void nfs_commit_free(struct nfs_write_data *p)
+void nfs_writedata_free(struct nfs_write_data *p)
 {
 	if (p && (p->pagevec != &p->page_array[0]))
 		kfree(p->pagevec);
-	mempool_free(p, nfs_commit_mempool);
+	mempool_free(p, nfs_wdata_mempool);
 }
 
 void nfs_writedata_release(void *wdata)
@@ -136,6 +166,7 @@ static void nfs_grow_file(struct page *page, unsigned int offset, unsigned int c
 	end = ((loff_t)page->index << PAGE_CACHE_SHIFT) + ((loff_t)offset+count);
 	if (i_size >= end)
 		return;
+	nfs_inc_stats(inode, NFSIOS_EXTENDWRITE);
 	i_size_write(inode, end);
 }
 
@@ -225,6 +256,7 @@ static int nfs_writepage_sync(struct nfs_open_context *ctx, struct inode *inode,
 	        wdata->args.pgbase += result;
 		written += result;
 		count -= result;
+		nfs_add_stats(inode, NFSIOS_SERVERWRITTENBYTES, result);
 	} while (count);
 	/* Update file length */
 	nfs_grow_file(page, offset, written);
@@ -281,6 +313,9 @@ int nfs_writepage(struct page *page, struct writeback_control *wbc)
 	int priority = wb_priority(wbc);
 	int err;
 
+	nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE);
+	nfs_add_stats(inode, NFSIOS_WRITEPAGES, 1);
+
 	/*
 	 * Note: We need to ensure that we have a reference to the inode
 	 *       if we are to do asynchronous writes. If not, waiting
@@ -345,6 +380,8 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc)
 	struct inode *inode = mapping->host;
 	int err;
 
+	nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGES);
+
 	err = generic_writepages(mapping, wbc);
 	if (err)
 		return err;
@@ -356,6 +393,7 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc)
 	err = nfs_flush_inode(inode, 0, 0, wb_priority(wbc));
 	if (err < 0)
 		goto out;
+	nfs_add_stats(inode, NFSIOS_WRITEPAGES, err);
 	wbc->nr_to_write -= err;
 	if (!wbc->nonblocking && wbc->sync_mode == WB_SYNC_ALL) {
 		err = nfs_wait_on_requests(inode, 0, 0);
@@ -391,6 +429,7 @@ static int nfs_inode_add_request(struct inode *inode, struct nfs_page *req)
 		if (nfs_have_delegation(inode, FMODE_WRITE))
 			nfsi->change_attr++;
 	}
+	SetPagePrivate(req->wb_page);
 	nfsi->npages++;
 	atomic_inc(&req->wb_count);
 	return 0;
@@ -407,6 +446,7 @@ static void nfs_inode_remove_request(struct nfs_page *req)
 	BUG_ON (!NFS_WBACK_BUSY(req));
 
 	spin_lock(&nfsi->req_lock);
+	ClearPagePrivate(req->wb_page);
 	radix_tree_delete(&nfsi->nfs_page_tree, req->wb_index);
 	nfsi->npages--;
 	if (!nfsi->npages) {
@@ -499,8 +539,7 @@ nfs_mark_request_commit(struct nfs_page *req)
  *
  * Interruptible by signals only if mounted with intr flag.
  */
-static int
-nfs_wait_on_requests(struct inode *inode, unsigned long idx_start, unsigned int npages)
+static int nfs_wait_on_requests_locked(struct inode *inode, unsigned long idx_start, unsigned int npages)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	struct nfs_page *req;
@@ -513,7 +552,6 @@ nfs_wait_on_requests(struct inode *inode, unsigned long idx_start, unsigned int
 	else
 		idx_end = idx_start + npages - 1;
 
-	spin_lock(&nfsi->req_lock);
 	next = idx_start;
 	while (radix_tree_gang_lookup_tag(&nfsi->nfs_page_tree, (void **)&req, next, 1, NFS_PAGE_TAG_WRITEBACK)) {
 		if (req->wb_index > idx_end)
@@ -526,15 +564,25 @@ nfs_wait_on_requests(struct inode *inode, unsigned long idx_start, unsigned int
 		spin_unlock(&nfsi->req_lock);
 		error = nfs_wait_on_request(req);
 		nfs_release_request(req);
+		spin_lock(&nfsi->req_lock);
 		if (error < 0)
 			return error;
-		spin_lock(&nfsi->req_lock);
 		res++;
 	}
-	spin_unlock(&nfsi->req_lock);
 	return res;
 }
 
+static int nfs_wait_on_requests(struct inode *inode, unsigned long idx_start, unsigned int npages)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+	int ret;
+
+	spin_lock(&nfsi->req_lock);
+	ret = nfs_wait_on_requests_locked(inode, idx_start, npages);
+	spin_unlock(&nfsi->req_lock);
+	return ret;
+}
+
 /*
  * nfs_scan_dirty - Scan an inode for dirty requests
  * @inode: NFS inode to scan
@@ -586,6 +634,11 @@ nfs_scan_commit(struct inode *inode, struct list_head *dst, unsigned long idx_st
 	}
 	return res;
 }
+#else
+static inline int nfs_scan_commit(struct inode *inode, struct list_head *dst, unsigned long idx_start, unsigned int npages)
+{
+	return 0;
+}
 #endif
 
 static int nfs_wait_on_write_congestion(struct address_space *mapping, int intr)
@@ -598,6 +651,9 @@ static int nfs_wait_on_write_congestion(struct address_space *mapping, int intr)
 
 	if (!bdi_write_congested(bdi))
 		return 0;
+
+	nfs_inc_stats(mapping->host, NFSIOS_CONGESTIONWAIT);
+
 	if (intr) {
 		struct rpc_clnt *clnt = NFS_CLIENT(mapping->host);
 		sigset_t oldset;
@@ -653,8 +709,11 @@ static struct nfs_page * nfs_update_request(struct nfs_open_context* ctx,
 				spin_unlock(&nfsi->req_lock);
 				error = nfs_wait_on_request(req);
 				nfs_release_request(req);
-				if (error < 0)
+				if (error < 0) {
+					if (new)
+						nfs_release_request(new);
 					return ERR_PTR(error);
+				}
 				continue;
 			}
 			spin_unlock(&nfsi->req_lock);
@@ -748,6 +807,8 @@ int nfs_updatepage(struct file *file, struct page *page,
 	struct nfs_page	*req;
 	int		status = 0;
 
+	nfs_inc_stats(inode, NFSIOS_VFSUPDATEPAGE);
+
 	dprintk("NFS:      nfs_updatepage(%s/%s %d@%Ld)\n",
 		file->f_dentry->d_parent->d_name.name,
 		file->f_dentry->d_name.name, count,
@@ -857,10 +918,12 @@ static inline int flush_task_priority(int how)
  */
 static void nfs_write_rpcsetup(struct nfs_page *req,
 		struct nfs_write_data *data,
+		const struct rpc_call_ops *call_ops,
 		unsigned int count, unsigned int offset,
 		int how)
 {
 	struct inode		*inode;
+	int flags;
 
 	/* Set up the RPC argument and reply structs
 	 * NB: take care not to mess about with data->commit et al. */
@@ -881,6 +944,9 @@ static void nfs_write_rpcsetup(struct nfs_page *req,
 	data->res.verf    = &data->verf;
 	nfs_fattr_init(&data->fattr);
 
+	/* Set up the initial task struct.  */
+	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
+	rpc_init_task(&data->task, NFS_CLIENT(inode), flags, call_ops, data);
 	NFS_PROTO(inode)->write_setup(data, how);
 
 	data->task.tk_priority = flush_task_priority(how);
@@ -910,7 +976,7 @@ static void nfs_execute_write(struct nfs_write_data *data)
  * Generate multiple small requests to write out a single
  * contiguous dirty area on one page.
  */
-static int nfs_flush_multi(struct list_head *head, struct inode *inode, int how)
+static int nfs_flush_multi(struct inode *inode, struct list_head *head, int how)
 {
 	struct nfs_page *req = nfs_list_entry(head->next);
 	struct page *page = req->wb_page;
@@ -944,14 +1010,15 @@ static int nfs_flush_multi(struct list_head *head, struct inode *inode, int how)
 		list_del_init(&data->pages);
 
 		data->pagevec[0] = page;
-		data->complete = nfs_writeback_done_partial;
 
 		if (nbytes > wsize) {
-			nfs_write_rpcsetup(req, data, wsize, offset, how);
+			nfs_write_rpcsetup(req, data, &nfs_write_partial_ops,
+					wsize, offset, how);
 			offset += wsize;
 			nbytes -= wsize;
 		} else {
-			nfs_write_rpcsetup(req, data, nbytes, offset, how);
+			nfs_write_rpcsetup(req, data, &nfs_write_partial_ops,
+					nbytes, offset, how);
 			nbytes = 0;
 		}
 		nfs_execute_write(data);
@@ -978,16 +1045,13 @@ out_bad:
  * This is the case if nfs_updatepage detects a conflicting request
  * that has been written but not committed.
  */
-static int nfs_flush_one(struct list_head *head, struct inode *inode, int how)
+static int nfs_flush_one(struct inode *inode, struct list_head *head, int how)
 {
 	struct nfs_page		*req;
 	struct page		**pages;
 	struct nfs_write_data	*data;
 	unsigned int		count;
 
-	if (NFS_SERVER(inode)->wsize < PAGE_CACHE_SIZE)
-		return nfs_flush_multi(head, inode, how);
-
 	data = nfs_writedata_alloc(NFS_SERVER(inode)->wpages);
 	if (!data)
 		goto out_bad;
@@ -1005,9 +1069,8 @@ static int nfs_flush_one(struct list_head *head, struct inode *inode, int how)
 	}
 	req = nfs_list_entry(data->pages.next);
 
-	data->complete = nfs_writeback_done_full;
 	/* Set up the argument struct */
-	nfs_write_rpcsetup(req, data, count, 0, how);
+	nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, how);
 
 	nfs_execute_write(data);
 	return 0;
@@ -1021,24 +1084,32 @@ static int nfs_flush_one(struct list_head *head, struct inode *inode, int how)
 	return -ENOMEM;
 }
 
-static int
-nfs_flush_list(struct list_head *head, int wpages, int how)
+static int nfs_flush_list(struct inode *inode, struct list_head *head, int npages, int how)
 {
 	LIST_HEAD(one_request);
-	struct nfs_page		*req;
-	int			error = 0;
-	unsigned int		pages = 0;
+	int (*flush_one)(struct inode *, struct list_head *, int);
+	struct nfs_page	*req;
+	int wpages = NFS_SERVER(inode)->wpages;
+	int wsize = NFS_SERVER(inode)->wsize;
+	int error;
 
-	while (!list_empty(head)) {
-		pages += nfs_coalesce_requests(head, &one_request, wpages);
+	flush_one = nfs_flush_one;
+	if (wsize < PAGE_CACHE_SIZE)
+		flush_one = nfs_flush_multi;
+	/* For single writes, FLUSH_STABLE is more efficient */
+	if (npages <= wpages && npages == NFS_I(inode)->npages
+			&& nfs_list_entry(head->next)->wb_bytes <= wsize)
+		how |= FLUSH_STABLE;
+
+	do {
+		nfs_coalesce_requests(head, &one_request, wpages);
 		req = nfs_list_entry(one_request.next);
-		error = nfs_flush_one(&one_request, req->wb_context->dentry->d_inode, how);
+		error = flush_one(inode, &one_request, how);
 		if (error < 0)
-			break;
-	}
-	if (error >= 0)
-		return pages;
-
+			goto out_err;
+	} while (!list_empty(head));
+	return 0;
+out_err:
 	while (!list_empty(head)) {
 		req = nfs_list_entry(head->next);
 		nfs_list_remove_request(req);
@@ -1051,8 +1122,9 @@ nfs_flush_list(struct list_head *head, int wpages, int how)
 /*
  * Handle a write reply that flushed part of a page.
  */
-static void nfs_writeback_done_partial(struct nfs_write_data *data, int status)
+static void nfs_writeback_done_partial(struct rpc_task *task, void *calldata)
 {
+	struct nfs_write_data	*data = calldata;
 	struct nfs_page		*req = data->req;
 	struct page		*page = req->wb_page;
 
@@ -1062,11 +1134,14 @@ static void nfs_writeback_done_partial(struct nfs_write_data *data, int status)
 		req->wb_bytes,
 		(long long)req_offset(req));
 
-	if (status < 0) {
+	if (nfs_writeback_done(task, data) != 0)
+		return;
+
+	if (task->tk_status < 0) {
 		ClearPageUptodate(page);
 		SetPageError(page);
-		req->wb_context->error = status;
-		dprintk(", error = %d\n", status);
+		req->wb_context->error = task->tk_status;
+		dprintk(", error = %d\n", task->tk_status);
 	} else {
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
 		if (data->verf.committed < NFS_FILE_SYNC) {
@@ -1087,6 +1162,11 @@ static void nfs_writeback_done_partial(struct nfs_write_data *data, int status)
 		nfs_writepage_release(req);
 }
 
+static const struct rpc_call_ops nfs_write_partial_ops = {
+	.rpc_call_done = nfs_writeback_done_partial,
+	.rpc_release = nfs_writedata_release,
+};
+
 /*
  * Handle a write reply that flushes a whole page.
  *
@@ -1094,11 +1174,15 @@ static void nfs_writeback_done_partial(struct nfs_write_data *data, int status)
  *	  writebacks since the page->count is kept > 1 for as long
  *	  as the page has a write request pending.
  */
-static void nfs_writeback_done_full(struct nfs_write_data *data, int status)
+static void nfs_writeback_done_full(struct rpc_task *task, void *calldata)
 {
+	struct nfs_write_data	*data = calldata;
 	struct nfs_page		*req;
 	struct page		*page;
 
+	if (nfs_writeback_done(task, data) != 0)
+		return;
+
 	/* Update attributes as result of writeback. */
 	while (!list_empty(&data->pages)) {
 		req = nfs_list_entry(data->pages.next);
@@ -1111,13 +1195,13 @@ static void nfs_writeback_done_full(struct nfs_write_data *data, int status)
 			req->wb_bytes,
 			(long long)req_offset(req));
 
-		if (status < 0) {
+		if (task->tk_status < 0) {
 			ClearPageUptodate(page);
 			SetPageError(page);
-			req->wb_context->error = status;
+			req->wb_context->error = task->tk_status;
 			end_page_writeback(page);
 			nfs_inode_remove_request(req);
-			dprintk(", error = %d\n", status);
+			dprintk(", error = %d\n", task->tk_status);
 			goto next;
 		}
 		end_page_writeback(page);
@@ -1139,18 +1223,30 @@ static void nfs_writeback_done_full(struct nfs_write_data *data, int status)
 	}
 }
 
+static const struct rpc_call_ops nfs_write_full_ops = {
+	.rpc_call_done = nfs_writeback_done_full,
+	.rpc_release = nfs_writedata_release,
+};
+
+
 /*
  * This function is called when the WRITE call is complete.
  */
-void nfs_writeback_done(struct rpc_task *task, void *calldata)
+int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 {
-	struct nfs_write_data	*data = calldata;
 	struct nfs_writeargs	*argp = &data->args;
 	struct nfs_writeres	*resp = &data->res;
+	int status;
 
 	dprintk("NFS: %4d nfs_writeback_done (status %d)\n",
 		task->tk_pid, task->tk_status);
 
+	/* Call the NFS version-specific code */
+	status = NFS_PROTO(data->inode)->write_done(task, data);
+	if (status != 0)
+		return status;
+	nfs_add_stats(data->inode, NFSIOS_SERVERWRITTENBYTES, resp->count);
+
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
 	if (resp->verf->committed < argp->stable && task->tk_status >= 0) {
 		/* We tried a write call, but the server did not
@@ -1176,6 +1272,8 @@ void nfs_writeback_done(struct rpc_task *task, void *calldata)
 	if (task->tk_status >= 0 && resp->count < argp->count) {
 		static unsigned long    complain;
 
+		nfs_inc_stats(data->inode, NFSIOS_SHORTWRITE);
+
 		/* Has the server at least made some progress? */
 		if (resp->count != 0) {
 			/* Was this an NFSv2 write or an NFSv3 stable write? */
@@ -1191,7 +1289,7 @@ void nfs_writeback_done(struct rpc_task *task, void *calldata)
 				argp->stable = NFS_FILE_SYNC;
 			}
 			rpc_restart_call(task);
-			return;
+			return -EAGAIN;
 		}
 		if (time_before(complain, jiffies)) {
 			printk(KERN_WARNING
@@ -1202,11 +1300,7 @@ void nfs_writeback_done(struct rpc_task *task, void *calldata)
 		/* Can't do anything about it except throw an error. */
 		task->tk_status = -EIO;
 	}
-
-	/*
-	 * Process the nfs_page list
-	 */
-	data->complete(data, task->tk_status);
+	return 0;
 }
 
 
@@ -1220,10 +1314,12 @@ void nfs_commit_release(void *wdata)
  * Set up the argument/result storage required for the RPC call.
  */
 static void nfs_commit_rpcsetup(struct list_head *head,
-		struct nfs_write_data *data, int how)
+		struct nfs_write_data *data,
+		int how)
 {
 	struct nfs_page		*first;
 	struct inode		*inode;
+	int flags;
 
 	/* Set up the RPC argument and reply structs
 	 * NB: take care not to mess about with data->commit et al. */
@@ -1243,7 +1339,10 @@ static void nfs_commit_rpcsetup(struct list_head *head,
 	data->res.fattr   = &data->fattr;
 	data->res.verf    = &data->verf;
 	nfs_fattr_init(&data->fattr);
-	
+
+	/* Set up the initial task struct.  */
+	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
+	rpc_init_task(&data->task, NFS_CLIENT(inode), flags, &nfs_commit_ops, data);
 	NFS_PROTO(inode)->commit_setup(data, how);
 
 	data->task.tk_priority = flush_task_priority(how);
@@ -1284,7 +1383,7 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how)
 /*
  * COMMIT call returned
  */
-void nfs_commit_done(struct rpc_task *task, void *calldata)
+static void nfs_commit_done(struct rpc_task *task, void *calldata)
 {
 	struct nfs_write_data	*data = calldata;
 	struct nfs_page		*req;
@@ -1293,6 +1392,10 @@ void nfs_commit_done(struct rpc_task *task, void *calldata)
         dprintk("NFS: %4d nfs_commit_done (status %d)\n",
                                 task->tk_pid, task->tk_status);
 
+	/* Call the NFS version-specific code */
+	if (NFS_PROTO(data->inode)->commit_done(task, data) != 0)
+		return;
+
 	while (!list_empty(&data->pages)) {
 		req = nfs_list_entry(data->pages.next);
 		nfs_list_remove_request(req);
@@ -1326,6 +1429,16 @@ void nfs_commit_done(struct rpc_task *task, void *calldata)
 	}
 	sub_page_state(nr_unstable,res);
 }
+
+static const struct rpc_call_ops nfs_commit_ops = {
+	.rpc_call_done = nfs_commit_done,
+	.rpc_release = nfs_commit_release,
+};
+#else
+static inline int nfs_commit_list(struct inode *inode, struct list_head *head, int how)
+{
+	return 0;
+}
 #endif
 
 static int nfs_flush_inode(struct inode *inode, unsigned long idx_start,
@@ -1333,24 +1446,16 @@ static int nfs_flush_inode(struct inode *inode, unsigned long idx_start,
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	LIST_HEAD(head);
-	int			res,
-				error = 0;
+	int res;
 
 	spin_lock(&nfsi->req_lock);
 	res = nfs_scan_dirty(inode, &head, idx_start, npages);
 	spin_unlock(&nfsi->req_lock);
 	if (res) {
-		struct nfs_server *server = NFS_SERVER(inode);
-
-		/* For single writes, FLUSH_STABLE is more efficient */
-		if (res == nfsi->npages && nfsi->npages <= server->wpages) {
-			if (res > 1 || nfs_list_entry(head.next)->wb_bytes <= server->wsize)
-				how |= FLUSH_STABLE;
-		}
-		error = nfs_flush_list(&head, server->wpages, how);
+		int error = nfs_flush_list(inode, &head, res, how);
+		if (error < 0)
+			return error;
 	}
-	if (error < 0)
-		return error;
 	return res;
 }
 
@@ -1359,14 +1464,13 @@ int nfs_commit_inode(struct inode *inode, int how)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	LIST_HEAD(head);
-	int			res,
-				error = 0;
+	int res;
 
 	spin_lock(&nfsi->req_lock);
 	res = nfs_scan_commit(inode, &head, 0, 0);
 	spin_unlock(&nfsi->req_lock);
 	if (res) {
-		error = nfs_commit_list(inode, &head, how);
+		int error = nfs_commit_list(inode, &head, how);
 		if (error < 0)
 			return error;
 	}
@@ -1374,28 +1478,38 @@ int nfs_commit_inode(struct inode *inode, int how)
 }
 #endif
 
-int nfs_sync_inode(struct inode *inode, unsigned long idx_start,
-		  unsigned int npages, int how)
+int nfs_sync_inode_wait(struct inode *inode, unsigned long idx_start,
+		unsigned int npages, int how)
 {
+	struct nfs_inode *nfsi = NFS_I(inode);
+	LIST_HEAD(head);
 	int nocommit = how & FLUSH_NOCOMMIT;
-	int wait = how & FLUSH_WAIT;
-	int error;
-
-	how &= ~(FLUSH_WAIT|FLUSH_NOCOMMIT);
+	int pages, ret;
 
+	how &= ~FLUSH_NOCOMMIT;
+	spin_lock(&nfsi->req_lock);
 	do {
-		if (wait) {
-			error = nfs_wait_on_requests(inode, idx_start, npages);
-			if (error != 0)
-				continue;
-		}
-		error = nfs_flush_inode(inode, idx_start, npages, how);
-		if (error != 0)
+		ret = nfs_wait_on_requests_locked(inode, idx_start, npages);
+		if (ret != 0)
 			continue;
-		if (!nocommit)
-			error = nfs_commit_inode(inode, how);
-	} while (error > 0);
-	return error;
+		pages = nfs_scan_dirty(inode, &head, idx_start, npages);
+		if (pages != 0) {
+			spin_unlock(&nfsi->req_lock);
+			ret = nfs_flush_list(inode, &head, pages, how);
+			spin_lock(&nfsi->req_lock);
+			continue;
+		}
+		if (nocommit)
+			break;
+		pages = nfs_scan_commit(inode, &head, 0, 0);
+		if (pages == 0)
+			break;
+		spin_unlock(&nfsi->req_lock);
+		ret = nfs_commit_list(inode, &head, how);
+		spin_lock(&nfsi->req_lock);
+	} while (ret >= 0);
+	spin_unlock(&nfsi->req_lock);
+	return ret;
 }
 
 int nfs_init_writepagecache(void)
@@ -1407,17 +1521,13 @@ int nfs_init_writepagecache(void)
 	if (nfs_wdata_cachep == NULL)
 		return -ENOMEM;
 
-	nfs_wdata_mempool = mempool_create(MIN_POOL_WRITE,
-					   mempool_alloc_slab,
-					   mempool_free_slab,
-					   nfs_wdata_cachep);
+	nfs_wdata_mempool = mempool_create_slab_pool(MIN_POOL_WRITE,
+						     nfs_wdata_cachep);
 	if (nfs_wdata_mempool == NULL)
 		return -ENOMEM;
 
-	nfs_commit_mempool = mempool_create(MIN_POOL_COMMIT,
-					   mempool_alloc_slab,
-					   mempool_free_slab,
-					   nfs_wdata_cachep);
+	nfs_commit_mempool = mempool_create_slab_pool(MIN_POOL_COMMIT,
+						      nfs_wdata_cachep);
 	if (nfs_commit_mempool == NULL)
 		return -ENOMEM;
 
diff --git a/fs/nfsctl.c b/fs/nfsctl.c
index 1c72c7f85ddc..a5a18d4aca40 100644
--- a/fs/nfsctl.c
+++ b/fs/nfsctl.c
@@ -101,7 +101,7 @@ asmlinkage sys_nfsservctl(int cmd, struct nfsctl_arg __user *arg, void __user *r
 	if (version != NFSCTL_VERSION)
 		return -EINVAL;
 
-	if (cmd < 0 || cmd >= sizeof(map)/sizeof(map[0]) || !map[cmd].name)
+	if (cmd < 0 || cmd >= ARRAY_SIZE(map) || !map[cmd].name)
 		return -EINVAL;
 
 	file = do_open(map[cmd].name, map[cmd].rsize ? O_RDWR : O_WRONLY);	
diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c
index 417ec02df44f..c340be0a3f59 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -57,27 +57,17 @@ static int		exp_verify_string(char *cp, int max);
 #define	EXPKEY_HASHMASK		(EXPKEY_HASHMAX -1)
 static struct cache_head *expkey_table[EXPKEY_HASHMAX];
 
-static inline int svc_expkey_hash(struct svc_expkey *item)
+static void expkey_put(struct kref *ref)
 {
-	int hash = item->ek_fsidtype;
-	char * cp = (char*)item->ek_fsid;
-	int len = key_len(item->ek_fsidtype);
+	struct svc_expkey *key = container_of(ref, struct svc_expkey, h.ref);
 
-	hash ^= hash_mem(cp, len, EXPKEY_HASHBITS);
-	hash ^= hash_ptr(item->ek_client, EXPKEY_HASHBITS);
-	return hash & EXPKEY_HASHMASK;
-}
-
-void expkey_put(struct cache_head *item, struct cache_detail *cd)
-{
-	if (cache_put(item, cd)) {
-		struct svc_expkey *key = container_of(item, struct svc_expkey, h);
-		if (test_bit(CACHE_VALID, &item->flags) &&
-		    !test_bit(CACHE_NEGATIVE, &item->flags))
-			exp_put(key->ek_export);
-		auth_domain_put(key->ek_client);
-		kfree(key);
+	if (test_bit(CACHE_VALID, &key->h.flags) &&
+	    !test_bit(CACHE_NEGATIVE, &key->h.flags)) {
+		dput(key->ek_dentry);
+		mntput(key->ek_mnt);
 	}
+	auth_domain_put(key->ek_client);
+	kfree(key);
 }
 
 static void expkey_request(struct cache_detail *cd,
@@ -95,7 +85,10 @@ static void expkey_request(struct cache_detail *cd,
 	(*bpp)[-1] = '\n';
 }
 
-static struct svc_expkey *svc_expkey_lookup(struct svc_expkey *, int);
+static struct svc_expkey *svc_expkey_update(struct svc_expkey *new, struct svc_expkey *old);
+static struct svc_expkey *svc_expkey_lookup(struct svc_expkey *);
+static struct cache_detail svc_expkey_cache;
+
 static int expkey_parse(struct cache_detail *cd, char *mesg, int mlen)
 {
 	/* client fsidtype fsid [path] */
@@ -106,6 +99,7 @@ static int expkey_parse(struct cache_detail *cd, char *mesg, int mlen)
 	int fsidtype;
 	char *ep;
 	struct svc_expkey key;
+	struct svc_expkey *ek;
 
 	if (mesg[mlen-1] != '\n')
 		return -EINVAL;
@@ -150,40 +144,38 @@ static int expkey_parse(struct cache_detail *cd, char *mesg, int mlen)
 	key.ek_fsidtype = fsidtype;
 	memcpy(key.ek_fsid, buf, len);
 
+	ek = svc_expkey_lookup(&key);
+	err = -ENOMEM;
+	if (!ek)
+		goto out;
+
 	/* now we want a pathname, or empty meaning NEGATIVE  */
+	err = -EINVAL;
 	if ((len=qword_get(&mesg, buf, PAGE_SIZE)) < 0)
 		goto out;
 	dprintk("Path seems to be <%s>\n", buf);
 	err = 0;
 	if (len == 0) {
-		struct svc_expkey *ek;
 		set_bit(CACHE_NEGATIVE, &key.h.flags);
-		ek = svc_expkey_lookup(&key, 1);
+		ek = svc_expkey_update(&key, ek);
 		if (ek)
-			expkey_put(&ek->h, &svc_expkey_cache);
+			cache_put(&ek->h, &svc_expkey_cache);
+		else err = -ENOMEM;
 	} else {
 		struct nameidata nd;
-		struct svc_expkey *ek;
-		struct svc_export *exp;
 		err = path_lookup(buf, 0, &nd);
 		if (err)
 			goto out;
 
 		dprintk("Found the path %s\n", buf);
-		exp = exp_get_by_name(dom, nd.mnt, nd.dentry, NULL);
-
-		err = -ENOENT;
-		if (!exp)
-			goto out_nd;
-		key.ek_export = exp;
-		dprintk("And found export\n");
+		key.ek_mnt = nd.mnt;
+		key.ek_dentry = nd.dentry;
 		
-		ek = svc_expkey_lookup(&key, 1);
+		ek = svc_expkey_update(&key, ek);
 		if (ek)
-			expkey_put(&ek->h, &svc_expkey_cache);
-		exp_put(exp);
-		err = 0;
-	out_nd:
+			cache_put(&ek->h, &svc_expkey_cache);
+		else
+			err = -ENOMEM;
 		path_release(&nd);
 	}
 	cache_flush();
@@ -214,35 +206,31 @@ static int expkey_show(struct seq_file *m,
 	if (test_bit(CACHE_VALID, &h->flags) && 
 	    !test_bit(CACHE_NEGATIVE, &h->flags)) {
 		seq_printf(m, " ");
-		seq_path(m, ek->ek_export->ex_mnt, ek->ek_export->ex_dentry, "\\ \t\n");
+		seq_path(m, ek->ek_mnt, ek->ek_dentry, "\\ \t\n");
 	}
 	seq_printf(m, "\n");
 	return 0;
 }
-	
-struct cache_detail svc_expkey_cache = {
-	.owner		= THIS_MODULE,
-	.hash_size	= EXPKEY_HASHMAX,
-	.hash_table	= expkey_table,
-	.name		= "nfsd.fh",
-	.cache_put	= expkey_put,
-	.cache_request	= expkey_request,
-	.cache_parse	= expkey_parse,
-	.cache_show	= expkey_show,
-};
 
-static inline int svc_expkey_match (struct svc_expkey *a, struct svc_expkey *b)
+static inline int expkey_match (struct cache_head *a, struct cache_head *b)
 {
-	if (a->ek_fsidtype != b->ek_fsidtype ||
-	    a->ek_client != b->ek_client ||
-	    memcmp(a->ek_fsid, b->ek_fsid, key_len(a->ek_fsidtype)) != 0)
+	struct svc_expkey *orig = container_of(a, struct svc_expkey, h);
+	struct svc_expkey *new = container_of(b, struct svc_expkey, h);
+
+	if (orig->ek_fsidtype != new->ek_fsidtype ||
+	    orig->ek_client != new->ek_client ||
+	    memcmp(orig->ek_fsid, new->ek_fsid, key_len(orig->ek_fsidtype)) != 0)
 		return 0;
 	return 1;
 }
 
-static inline void svc_expkey_init(struct svc_expkey *new, struct svc_expkey *item)
+static inline void expkey_init(struct cache_head *cnew,
+				   struct cache_head *citem)
 {
-	cache_get(&item->ek_client->h);
+	struct svc_expkey *new = container_of(cnew, struct svc_expkey, h);
+	struct svc_expkey *item = container_of(citem, struct svc_expkey, h);
+
+	kref_get(&item->ek_client->ref);
 	new->ek_client = item->ek_client;
 	new->ek_fsidtype = item->ek_fsidtype;
 	new->ek_fsid[0] = item->ek_fsid[0];
@@ -250,39 +238,94 @@ static inline void svc_expkey_init(struct svc_expkey *new, struct svc_expkey *it
 	new->ek_fsid[2] = item->ek_fsid[2];
 }
 
-static inline void svc_expkey_update(struct svc_expkey *new, struct svc_expkey *item)
+static inline void expkey_update(struct cache_head *cnew,
+				   struct cache_head *citem)
+{
+	struct svc_expkey *new = container_of(cnew, struct svc_expkey, h);
+	struct svc_expkey *item = container_of(citem, struct svc_expkey, h);
+
+	new->ek_mnt = mntget(item->ek_mnt);
+	new->ek_dentry = dget(item->ek_dentry);
+}
+
+static struct cache_head *expkey_alloc(void)
 {
-	cache_get(&item->ek_export->h);
-	new->ek_export = item->ek_export;
+	struct svc_expkey *i = kmalloc(sizeof(*i), GFP_KERNEL);
+	if (i)
+		return &i->h;
+	else
+		return NULL;
 }
 
-static DefineSimpleCacheLookup(svc_expkey,0) /* no inplace updates */
+static struct cache_detail svc_expkey_cache = {
+	.owner		= THIS_MODULE,
+	.hash_size	= EXPKEY_HASHMAX,
+	.hash_table	= expkey_table,
+	.name		= "nfsd.fh",
+	.cache_put	= expkey_put,
+	.cache_request	= expkey_request,
+	.cache_parse	= expkey_parse,
+	.cache_show	= expkey_show,
+	.match		= expkey_match,
+	.init		= expkey_init,
+	.update       	= expkey_update,
+	.alloc		= expkey_alloc,
+};
 
-#define	EXPORT_HASHBITS		8
-#define	EXPORT_HASHMAX		(1<< EXPORT_HASHBITS)
-#define	EXPORT_HASHMASK		(EXPORT_HASHMAX -1)
+static struct svc_expkey *
+svc_expkey_lookup(struct svc_expkey *item)
+{
+	struct cache_head *ch;
+	int hash = item->ek_fsidtype;
+	char * cp = (char*)item->ek_fsid;
+	int len = key_len(item->ek_fsidtype);
 
-static struct cache_head *export_table[EXPORT_HASHMAX];
+	hash ^= hash_mem(cp, len, EXPKEY_HASHBITS);
+	hash ^= hash_ptr(item->ek_client, EXPKEY_HASHBITS);
+	hash &= EXPKEY_HASHMASK;
 
-static inline int svc_export_hash(struct svc_export *item)
+	ch = sunrpc_cache_lookup(&svc_expkey_cache, &item->h,
+				 hash);
+	if (ch)
+		return container_of(ch, struct svc_expkey, h);
+	else
+		return NULL;
+}
+
+static struct svc_expkey *
+svc_expkey_update(struct svc_expkey *new, struct svc_expkey *old)
 {
-	int rv;
+	struct cache_head *ch;
+	int hash = new->ek_fsidtype;
+	char * cp = (char*)new->ek_fsid;
+	int len = key_len(new->ek_fsidtype);
 
-	rv = hash_ptr(item->ex_client, EXPORT_HASHBITS);
-	rv ^= hash_ptr(item->ex_dentry, EXPORT_HASHBITS);
-	rv ^= hash_ptr(item->ex_mnt, EXPORT_HASHBITS);
-	return rv;
+	hash ^= hash_mem(cp, len, EXPKEY_HASHBITS);
+	hash ^= hash_ptr(new->ek_client, EXPKEY_HASHBITS);
+	hash &= EXPKEY_HASHMASK;
+
+	ch = sunrpc_cache_update(&svc_expkey_cache, &new->h,
+				 &old->h, hash);
+	if (ch)
+		return container_of(ch, struct svc_expkey, h);
+	else
+		return NULL;
 }
 
-void svc_export_put(struct cache_head *item, struct cache_detail *cd)
+
+#define	EXPORT_HASHBITS		8
+#define	EXPORT_HASHMAX		(1<< EXPORT_HASHBITS)
+#define	EXPORT_HASHMASK		(EXPORT_HASHMAX -1)
+
+static struct cache_head *export_table[EXPORT_HASHMAX];
+
+static void svc_export_put(struct kref *ref)
 {
-	if (cache_put(item, cd)) {
-		struct svc_export *exp = container_of(item, struct svc_export, h);
-		dput(exp->ex_dentry);
-		mntput(exp->ex_mnt);
-		auth_domain_put(exp->ex_client);
-		kfree(exp);
-	}
+	struct svc_export *exp = container_of(ref, struct svc_export, h.ref);
+	dput(exp->ex_dentry);
+	mntput(exp->ex_mnt);
+	auth_domain_put(exp->ex_client);
+	kfree(exp);
 }
 
 static void svc_export_request(struct cache_detail *cd,
@@ -304,7 +347,9 @@ static void svc_export_request(struct cache_detail *cd,
 	(*bpp)[-1] = '\n';
 }
 
-static struct svc_export *svc_export_lookup(struct svc_export *, int);
+static struct svc_export *svc_export_update(struct svc_export *new,
+					    struct svc_export *old);
+static struct svc_export *svc_export_lookup(struct svc_export *);
 
 static int check_export(struct inode *inode, int flags)
 {
@@ -417,11 +462,16 @@ static int svc_export_parse(struct cache_detail *cd, char *mesg, int mlen)
 		if (err) goto out;
 	}
 
-	expp = svc_export_lookup(&exp, 1);
+	expp = svc_export_lookup(&exp);
 	if (expp)
-		exp_put(expp);
-	err = 0;
+		expp = svc_export_update(&exp, expp);
+	else
+		err = -ENOMEM;
 	cache_flush();
+	if (expp == NULL)
+		err = -ENOMEM;
+	else
+		exp_put(expp);
  out:
 	if (nd.dentry)
 		path_release(&nd);
@@ -455,6 +505,46 @@ static int svc_export_show(struct seq_file *m,
 	seq_puts(m, ")\n");
 	return 0;
 }
+static int svc_export_match(struct cache_head *a, struct cache_head *b)
+{
+	struct svc_export *orig = container_of(a, struct svc_export, h);
+	struct svc_export *new = container_of(b, struct svc_export, h);
+	return orig->ex_client == new->ex_client &&
+		orig->ex_dentry == new->ex_dentry &&
+		orig->ex_mnt == new->ex_mnt;
+}
+
+static void svc_export_init(struct cache_head *cnew, struct cache_head *citem)
+{
+	struct svc_export *new = container_of(cnew, struct svc_export, h);
+	struct svc_export *item = container_of(citem, struct svc_export, h);
+
+	kref_get(&item->ex_client->ref);
+	new->ex_client = item->ex_client;
+	new->ex_dentry = dget(item->ex_dentry);
+	new->ex_mnt = mntget(item->ex_mnt);
+}
+
+static void export_update(struct cache_head *cnew, struct cache_head *citem)
+{
+	struct svc_export *new = container_of(cnew, struct svc_export, h);
+	struct svc_export *item = container_of(citem, struct svc_export, h);
+
+	new->ex_flags = item->ex_flags;
+	new->ex_anon_uid = item->ex_anon_uid;
+	new->ex_anon_gid = item->ex_anon_gid;
+	new->ex_fsid = item->ex_fsid;
+}
+
+static struct cache_head *svc_export_alloc(void)
+{
+	struct svc_export *i = kmalloc(sizeof(*i), GFP_KERNEL);
+	if (i)
+		return &i->h;
+	else
+		return NULL;
+}
+
 struct cache_detail svc_export_cache = {
 	.owner		= THIS_MODULE,
 	.hash_size	= EXPORT_HASHMAX,
@@ -464,34 +554,49 @@ struct cache_detail svc_export_cache = {
 	.cache_request	= svc_export_request,
 	.cache_parse	= svc_export_parse,
 	.cache_show	= svc_export_show,
+	.match		= svc_export_match,
+	.init		= svc_export_init,
+	.update		= export_update,
+	.alloc		= svc_export_alloc,
 };
 
-static inline int svc_export_match(struct svc_export *a, struct svc_export *b)
+static struct svc_export *
+svc_export_lookup(struct svc_export *exp)
 {
-	return a->ex_client == b->ex_client &&
-		a->ex_dentry == b->ex_dentry &&
-		a->ex_mnt == b->ex_mnt;
-}
-static inline void svc_export_init(struct svc_export *new, struct svc_export *item)
-{
-	cache_get(&item->ex_client->h);
-	new->ex_client = item->ex_client;
-	new->ex_dentry = dget(item->ex_dentry);
-	new->ex_mnt = mntget(item->ex_mnt);
+	struct cache_head *ch;
+	int hash;
+	hash = hash_ptr(exp->ex_client, EXPORT_HASHBITS);
+	hash ^= hash_ptr(exp->ex_dentry, EXPORT_HASHBITS);
+	hash ^= hash_ptr(exp->ex_mnt, EXPORT_HASHBITS);
+
+	ch = sunrpc_cache_lookup(&svc_export_cache, &exp->h,
+				 hash);
+	if (ch)
+		return container_of(ch, struct svc_export, h);
+	else
+		return NULL;
 }
 
-static inline void svc_export_update(struct svc_export *new, struct svc_export *item)
+static struct svc_export *
+svc_export_update(struct svc_export *new, struct svc_export *old)
 {
-	new->ex_flags = item->ex_flags;
-	new->ex_anon_uid = item->ex_anon_uid;
-	new->ex_anon_gid = item->ex_anon_gid;
-	new->ex_fsid = item->ex_fsid;
+	struct cache_head *ch;
+	int hash;
+	hash = hash_ptr(old->ex_client, EXPORT_HASHBITS);
+	hash ^= hash_ptr(old->ex_dentry, EXPORT_HASHBITS);
+	hash ^= hash_ptr(old->ex_mnt, EXPORT_HASHBITS);
+
+	ch = sunrpc_cache_update(&svc_export_cache, &new->h,
+				 &old->h,
+				 hash);
+	if (ch)
+		return container_of(ch, struct svc_export, h);
+	else
+		return NULL;
 }
 
-static DefineSimpleCacheLookup(svc_export,1) /* allow inplace updates */
 
-
-struct svc_expkey *
+static struct svc_expkey *
 exp_find_key(svc_client *clp, int fsid_type, u32 *fsidv, struct cache_req *reqp)
 {
 	struct svc_expkey key, *ek;
@@ -504,7 +609,7 @@ exp_find_key(svc_client *clp, int fsid_type, u32 *fsidv, struct cache_req *reqp)
 	key.ek_fsidtype = fsid_type;
 	memcpy(key.ek_fsid, fsidv, key_len(fsid_type));
 
-	ek = svc_expkey_lookup(&key, 0);
+	ek = svc_expkey_lookup(&key);
 	if (ek != NULL)
 		if ((err = cache_check(&svc_expkey_cache, &ek->h, reqp)))
 			ek = ERR_PTR(err);
@@ -519,13 +624,16 @@ static int exp_set_key(svc_client *clp, int fsid_type, u32 *fsidv,
 	key.ek_client = clp;
 	key.ek_fsidtype = fsid_type;
 	memcpy(key.ek_fsid, fsidv, key_len(fsid_type));
-	key.ek_export = exp;
+	key.ek_mnt = exp->ex_mnt;
+	key.ek_dentry = exp->ex_dentry;
 	key.h.expiry_time = NEVER;
 	key.h.flags = 0;
 
-	ek = svc_expkey_lookup(&key, 1);
+	ek = svc_expkey_lookup(&key);
+	if (ek)
+		ek = svc_expkey_update(&key,ek);
 	if (ek) {
-		expkey_put(&ek->h, &svc_expkey_cache);
+		cache_put(&ek->h, &svc_expkey_cache);
 		return 0;
 	}
 	return -ENOMEM;
@@ -573,7 +681,7 @@ exp_get_by_name(svc_client *clp, struct vfsmount *mnt, struct dentry *dentry,
 	key.ex_mnt = mnt;
 	key.ex_dentry = dentry;
 
-	exp = svc_export_lookup(&key, 0);
+	exp = svc_export_lookup(&key);
 	if (exp != NULL) 
 		switch (cache_check(&svc_export_cache, &exp->h, reqp)) {
 		case 0: break;
@@ -654,7 +762,7 @@ static void exp_fsid_unhash(struct svc_export *exp)
 	ek = exp_get_fsid_key(exp->ex_client, exp->ex_fsid);
 	if (ek && !IS_ERR(ek)) {
 		ek->h.expiry_time = get_seconds()-1;
-		expkey_put(&ek->h, &svc_expkey_cache);
+		cache_put(&ek->h, &svc_expkey_cache);
 	}
 	svc_expkey_cache.nextcheck = get_seconds();
 }
@@ -692,7 +800,7 @@ static void exp_unhash(struct svc_export *exp)
 	ek = exp_get_key(exp->ex_client, inode->i_sb->s_dev, inode->i_ino);
 	if (ek && !IS_ERR(ek)) {
 		ek->h.expiry_time = get_seconds()-1;
-		expkey_put(&ek->h, &svc_expkey_cache);
+		cache_put(&ek->h, &svc_expkey_cache);
 	}
 	svc_expkey_cache.nextcheck = get_seconds();
 }
@@ -741,8 +849,8 @@ exp_export(struct nfsctl_export *nxp)
 	if ((nxp->ex_flags & NFSEXP_FSID) &&
 	    (fsid_key = exp_get_fsid_key(clp, nxp->ex_dev)) &&
 	    !IS_ERR(fsid_key) &&
-	    fsid_key->ek_export &&
-	    fsid_key->ek_export != exp)
+	    fsid_key->ek_mnt &&
+	    (fsid_key->ek_mnt != nd.mnt || fsid_key->ek_dentry != nd.dentry) )
 		goto finish;
 
 	if (exp) {
@@ -775,13 +883,13 @@ exp_export(struct nfsctl_export *nxp)
 	new.ex_anon_gid = nxp->ex_anon_gid;
 	new.ex_fsid = nxp->ex_dev;
 
-	exp = svc_export_lookup(&new, 1);
+	exp = svc_export_lookup(&new);
+	if (exp)
+		exp = svc_export_update(&new, exp);
 
-	if (exp == NULL)
+	if (!exp)
 		goto finish;
 
-	err = 0;
-
 	if (exp_hash(clp, exp) ||
 	    exp_fsid_hash(clp, exp)) {
 		/* failed to create at least one index */
@@ -794,7 +902,7 @@ finish:
 	if (exp)
 		exp_put(exp);
 	if (fsid_key && !IS_ERR(fsid_key))
-		expkey_put(&fsid_key->h, &svc_expkey_cache);
+		cache_put(&fsid_key->h, &svc_expkey_cache);
 	if (clp)
 		auth_domain_put(clp);
 	path_release(&nd);
@@ -912,6 +1020,24 @@ out:
 	return err;
 }
 
+struct svc_export *
+exp_find(struct auth_domain *clp, int fsid_type, u32 *fsidv,
+	 struct cache_req *reqp)
+{
+	struct svc_export *exp;
+	struct svc_expkey *ek = exp_find_key(clp, fsid_type, fsidv, reqp);
+	if (!ek || IS_ERR(ek))
+		return ERR_PTR(PTR_ERR(ek));
+
+	exp = exp_get_by_name(clp, ek->ek_mnt, ek->ek_dentry, reqp);
+	cache_put(&ek->h, &svc_expkey_cache);
+
+	if (!exp || IS_ERR(exp))
+		return ERR_PTR(PTR_ERR(exp));
+	return exp;
+}
+
+
 /*
  * Called when we need the filehandle for the root of the pseudofs,
  * for a given NFSv4 client.   The root is defined to be the
@@ -922,6 +1048,7 @@ exp_pseudoroot(struct auth_domain *clp, struct svc_fh *fhp,
 	       struct cache_req *creq)
 {
 	struct svc_expkey *fsid_key;
+	struct svc_export *exp;
 	int rv;
 	u32 fsidv[2];
 
@@ -933,9 +1060,15 @@ exp_pseudoroot(struct auth_domain *clp, struct svc_fh *fhp,
 	if (!fsid_key || IS_ERR(fsid_key))
 		return nfserr_perm;
 
-	rv = fh_compose(fhp, fsid_key->ek_export, 
-			  fsid_key->ek_export->ex_dentry, NULL);
-	expkey_put(&fsid_key->h, &svc_expkey_cache);
+	exp = exp_get_by_name(clp, fsid_key->ek_mnt, fsid_key->ek_dentry, creq);
+	if (exp == NULL)
+		rv = nfserr_perm;
+	else if (IS_ERR(exp))
+		rv = nfserrno(PTR_ERR(exp));
+	else
+		rv = fh_compose(fhp, exp,
+				fsid_key->ek_dentry, NULL);
+	cache_put(&fsid_key->h, &svc_expkey_cache);
 	return rv;
 }
 
@@ -1054,7 +1187,7 @@ static int e_show(struct seq_file *m, void *p)
 	cache_get(&exp->h);
 	if (cache_check(&svc_export_cache, &exp->h, NULL))
 		return 0;
-	if (cache_put(&exp->h, &svc_export_cache)) BUG();
+	cache_put(&exp->h, &svc_export_cache);
 	return svc_export_show(m, &svc_export_cache, cp);
 }
 
@@ -1129,7 +1262,6 @@ exp_delclient(struct nfsctl_client *ncp)
 	 */
 	if (dom) {
 		err = auth_unix_forget_old(dom);
-		dom->h.expiry_time = get_seconds();
 		auth_domain_put(dom);
 	}
 
diff --git a/fs/nfsd/nfs4acl.c b/fs/nfsd/nfs4acl.c
index 4a2105552ac4..7391f4aabedb 100644
--- a/fs/nfsd/nfs4acl.c
+++ b/fs/nfsd/nfs4acl.c
@@ -907,7 +907,7 @@ nfs4_acl_get_whotype(char *p, u32 len)
 {
 	int i;
 
-	for (i=0; i < sizeof(s2t_map) / sizeof(*s2t_map); i++) {
+	for (i = 0; i < ARRAY_SIZE(s2t_map); i++) {
 		if (s2t_map[i].stringlen == len &&
 				0 == memcmp(s2t_map[i].string, p, len))
 			return s2t_map[i].type;
@@ -920,7 +920,7 @@ nfs4_acl_write_who(int who, char *p)
 {
 	int i;
 
-	for (i=0; i < sizeof(s2t_map) / sizeof(*s2t_map); i++) {
+	for (i = 0; i < ARRAY_SIZE(s2t_map); i++) {
 		if (s2t_map[i].type == who) {
 			memcpy(p, s2t_map[i].string, s2t_map[i].stringlen);
 			return s2t_map[i].stringlen;
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index d828662d737d..c872bd07fc10 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -326,6 +326,8 @@ out:
         .p_encode = (kxdrproc_t) nfs4_xdr_##argtype,                    \
         .p_decode = (kxdrproc_t) nfs4_xdr_##restype,                    \
         .p_bufsiz = MAX(NFS4_##argtype##_sz,NFS4_##restype##_sz) << 2,  \
+        .p_statidx = NFSPROC4_CB_##call,				\
+	.p_name   = #proc,                                              \
 }
 
 static struct rpc_procinfo     nfs4_cb_procedures[] = {
@@ -335,7 +337,7 @@ static struct rpc_procinfo     nfs4_cb_procedures[] = {
 
 static struct rpc_version       nfs_cb_version4 = {
         .number                 = 1,
-        .nrprocs                = sizeof(nfs4_cb_procedures)/sizeof(nfs4_cb_procedures[0]),
+        .nrprocs                = ARRAY_SIZE(nfs4_cb_procedures),
         .procs                  = nfs4_cb_procedures
 };
 
@@ -411,7 +413,7 @@ nfsd4_probe_callback(struct nfs4_client *clp)
 	/* Initialize rpc_program */
 	program->name = "nfs4_cb";
 	program->number = cb->cb_prog;
-	program->nrvers = sizeof(nfs_cb_version)/sizeof(nfs_cb_version[0]);
+	program->nrvers = ARRAY_SIZE(nfs_cb_version);
 	program->version = nfs_cb_version;
 	program->stats = stat;
 
diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c
index 13369650cdf9..4b6aa60dfceb 100644
--- a/fs/nfsd/nfs4idmap.c
+++ b/fs/nfsd/nfs4idmap.c
@@ -76,21 +76,18 @@ struct ent {
 	char              authname[IDMAP_NAMESZ];
 };
 
-#define DefineSimpleCacheLookupMap(STRUCT, FUNC)			\
-        DefineCacheLookup(struct STRUCT, h, FUNC##_lookup,		\
-        (struct STRUCT *item, int set), /*no setup */,			\
-	& FUNC##_cache, FUNC##_hash(item), FUNC##_match(item, tmp),	\
-	STRUCT##_init(new, item), STRUCT##_update(tmp, item), 0)
-
 /* Common entry handling */
 
 #define ENT_HASHBITS          8
 #define ENT_HASHMAX           (1 << ENT_HASHBITS)
 #define ENT_HASHMASK          (ENT_HASHMAX - 1)
 
-static inline void
-ent_init(struct ent *new, struct ent *itm)
+static void
+ent_init(struct cache_head *cnew, struct cache_head *citm)
 {
+	struct ent *new = container_of(cnew, struct ent, h);
+	struct ent *itm = container_of(citm, struct ent, h);
+
 	new->id = itm->id;
 	new->type = itm->type;
 
@@ -98,19 +95,21 @@ ent_init(struct ent *new, struct ent *itm)
 	strlcpy(new->authname, itm->authname, sizeof(new->name));
 }
 
-static inline void
-ent_update(struct ent *new, struct ent *itm)
+static void
+ent_put(struct kref *ref)
 {
-	ent_init(new, itm);
+	struct ent *map = container_of(ref, struct ent, h.ref);
+	kfree(map);
 }
 
-static void
-ent_put(struct cache_head *ch, struct cache_detail *cd)
+static struct cache_head *
+ent_alloc(void)
 {
-	if (cache_put(ch, cd)) {
-		struct ent *map = container_of(ch, struct ent, h);
-		kfree(map);
-	}
+	struct ent *e = kmalloc(sizeof(*e), GFP_KERNEL);
+	if (e)
+		return &e->h;
+	else
+		return NULL;
 }
 
 /*
@@ -149,9 +148,12 @@ idtoname_request(struct cache_detail *cd, struct cache_head *ch, char **bpp,
 	(*bpp)[-1] = '\n';
 }
 
-static inline int
-idtoname_match(struct ent *a, struct ent *b)
+static int
+idtoname_match(struct cache_head *ca, struct cache_head *cb)
 {
+	struct ent *a = container_of(ca, struct ent, h);
+	struct ent *b = container_of(cb, struct ent, h);
+
 	return (a->id == b->id && a->type == b->type &&
 	    strcmp(a->authname, b->authname) == 0);
 }
@@ -184,7 +186,8 @@ warn_no_idmapd(struct cache_detail *detail)
 
 
 static int         idtoname_parse(struct cache_detail *, char *, int);
-static struct ent *idtoname_lookup(struct ent *, int);
+static struct ent *idtoname_lookup(struct ent *);
+static struct ent *idtoname_update(struct ent *, struct ent *);
 
 static struct cache_detail idtoname_cache = {
 	.owner		= THIS_MODULE,
@@ -196,6 +199,10 @@ static struct cache_detail idtoname_cache = {
 	.cache_parse	= idtoname_parse,
 	.cache_show	= idtoname_show,
 	.warn_no_listener = warn_no_idmapd,
+	.match		= idtoname_match,
+	.init		= ent_init,
+	.update		= ent_init,
+	.alloc		= ent_alloc,
 };
 
 int
@@ -238,6 +245,11 @@ idtoname_parse(struct cache_detail *cd, char *buf, int buflen)
 	if (ent.h.expiry_time == 0)
 		goto out;
 
+	error = -ENOMEM;
+	res = idtoname_lookup(&ent);
+	if (!res)
+		goto out;
+
 	/* Name */
 	error = qword_get(&buf, buf1, PAGE_SIZE);
 	if (error == -EINVAL)
@@ -252,10 +264,11 @@ idtoname_parse(struct cache_detail *cd, char *buf, int buflen)
 		memcpy(ent.name, buf1, sizeof(ent.name));
 	}
 	error = -ENOMEM;
-	if ((res = idtoname_lookup(&ent, 1)) == NULL)
+	res = idtoname_update(&ent, res);
+	if (res == NULL)
 		goto out;
 
-	ent_put(&res->h, &idtoname_cache);
+	cache_put(&res->h, &idtoname_cache);
 
 	error = 0;
 out:
@@ -264,7 +277,31 @@ out:
 	return error;
 }
 
-static DefineSimpleCacheLookupMap(ent, idtoname);
+
+static struct ent *
+idtoname_lookup(struct ent *item)
+{
+	struct cache_head *ch = sunrpc_cache_lookup(&idtoname_cache,
+						    &item->h,
+						    idtoname_hash(item));
+	if (ch)
+		return container_of(ch, struct ent, h);
+	else
+		return NULL;
+}
+
+static struct ent *
+idtoname_update(struct ent *new, struct ent *old)
+{
+	struct cache_head *ch = sunrpc_cache_update(&idtoname_cache,
+						    &new->h, &old->h,
+						    idtoname_hash(new));
+	if (ch)
+		return container_of(ch, struct ent, h);
+	else
+		return NULL;
+}
+
 
 /*
  * Name -> ID cache
@@ -291,9 +328,12 @@ nametoid_request(struct cache_detail *cd, struct cache_head *ch, char **bpp,
 	(*bpp)[-1] = '\n';
 }
 
-static inline int
-nametoid_match(struct ent *a, struct ent *b)
+static int
+nametoid_match(struct cache_head *ca, struct cache_head *cb)
 {
+	struct ent *a = container_of(ca, struct ent, h);
+	struct ent *b = container_of(cb, struct ent, h);
+
 	return (a->type == b->type && strcmp(a->name, b->name) == 0 &&
 	    strcmp(a->authname, b->authname) == 0);
 }
@@ -317,7 +357,8 @@ nametoid_show(struct seq_file *m, struct cache_detail *cd, struct cache_head *h)
 	return 0;
 }
 
-static struct ent *nametoid_lookup(struct ent *, int);
+static struct ent *nametoid_lookup(struct ent *);
+static struct ent *nametoid_update(struct ent *, struct ent *);
 static int         nametoid_parse(struct cache_detail *, char *, int);
 
 static struct cache_detail nametoid_cache = {
@@ -330,6 +371,10 @@ static struct cache_detail nametoid_cache = {
 	.cache_parse	= nametoid_parse,
 	.cache_show	= nametoid_show,
 	.warn_no_listener = warn_no_idmapd,
+	.match		= nametoid_match,
+	.init		= ent_init,
+	.update		= ent_init,
+	.alloc		= ent_alloc,
 };
 
 static int
@@ -379,10 +424,14 @@ nametoid_parse(struct cache_detail *cd, char *buf, int buflen)
 		set_bit(CACHE_NEGATIVE, &ent.h.flags);
 
 	error = -ENOMEM;
-	if ((res = nametoid_lookup(&ent, 1)) == NULL)
+	res = nametoid_lookup(&ent);
+	if (res == NULL)
+		goto out;
+	res = nametoid_update(&ent, res);
+	if (res == NULL)
 		goto out;
 
-	ent_put(&res->h, &nametoid_cache);
+	cache_put(&res->h, &nametoid_cache);
 	error = 0;
 out:
 	kfree(buf1);
@@ -390,7 +439,30 @@ out:
 	return (error);
 }
 
-static DefineSimpleCacheLookupMap(ent, nametoid);
+
+static struct ent *
+nametoid_lookup(struct ent *item)
+{
+	struct cache_head *ch = sunrpc_cache_lookup(&nametoid_cache,
+						    &item->h,
+						    nametoid_hash(item));
+	if (ch)
+		return container_of(ch, struct ent, h);
+	else
+		return NULL;
+}
+
+static struct ent *
+nametoid_update(struct ent *new, struct ent *old)
+{
+	struct cache_head *ch = sunrpc_cache_update(&nametoid_cache,
+						    &new->h, &old->h,
+						    nametoid_hash(new));
+	if (ch)
+		return container_of(ch, struct ent, h);
+	else
+		return NULL;
+}
 
 /*
  * Exported API
@@ -458,24 +530,24 @@ idmap_defer(struct cache_req *req)
 }
 
 static inline int
-do_idmap_lookup(struct ent *(*lookup_fn)(struct ent *, int), struct ent *key,
+do_idmap_lookup(struct ent *(*lookup_fn)(struct ent *), struct ent *key,
 		struct cache_detail *detail, struct ent **item,
 		struct idmap_defer_req *mdr)
 {
-	*item = lookup_fn(key, 0);
+	*item = lookup_fn(key);
 	if (!*item)
 		return -ENOMEM;
 	return cache_check(detail, &(*item)->h, &mdr->req);
 }
 
 static inline int
-do_idmap_lookup_nowait(struct ent *(*lookup_fn)(struct ent *, int),
+do_idmap_lookup_nowait(struct ent *(*lookup_fn)(struct ent *),
 			struct ent *key, struct cache_detail *detail,
 			struct ent **item)
 {
 	int ret = -ENOMEM;
 
-	*item = lookup_fn(key, 0);
+	*item = lookup_fn(key);
 	if (!*item)
 		goto out_err;
 	ret = -ETIMEDOUT;
@@ -488,7 +560,7 @@ do_idmap_lookup_nowait(struct ent *(*lookup_fn)(struct ent *, int),
 		goto out_put;
 	return 0;
 out_put:
-	ent_put(&(*item)->h, detail);
+	cache_put(&(*item)->h, detail);
 out_err:
 	*item = NULL;
 	return ret;
@@ -496,7 +568,7 @@ out_err:
 
 static int
 idmap_lookup(struct svc_rqst *rqstp,
-		struct ent *(*lookup_fn)(struct ent *, int), struct ent *key,
+		struct ent *(*lookup_fn)(struct ent *), struct ent *key,
 		struct cache_detail *detail, struct ent **item)
 {
 	struct idmap_defer_req *mdr;
@@ -539,7 +611,7 @@ idmap_name_to_id(struct svc_rqst *rqstp, int type, const char *name, u32 namelen
 	if (ret)
 		return ret;
 	*id = item->id;
-	ent_put(&item->h, &nametoid_cache);
+	cache_put(&item->h, &nametoid_cache);
 	return 0;
 }
 
@@ -561,7 +633,7 @@ idmap_id_to_name(struct svc_rqst *rqstp, int type, uid_t id, char *name)
 	ret = strlen(item->name);
 	BUG_ON(ret > IDMAP_NAMESZ);
 	memcpy(name, item->name, ret);
-	ent_put(&item->h, &idtoname_cache);
+	cache_put(&item->h, &idtoname_cache);
 	return ret;
 }
 
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 1143cfb64549..47ec112b266c 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -49,6 +49,7 @@
 #include <linux/nfsd/state.h>
 #include <linux/nfsd/xdr4.h>
 #include <linux/namei.h>
+#include <linux/mutex.h>
 
 #define NFSDDBG_FACILITY                NFSDDBG_PROC
 
@@ -77,11 +78,11 @@ static void nfs4_set_recdir(char *recdir);
 
 /* Locking:
  *
- * client_sema: 
+ * client_mutex:
  * 	protects clientid_hashtbl[], clientstr_hashtbl[],
  * 	unconfstr_hashtbl[], uncofid_hashtbl[].
  */
-static DECLARE_MUTEX(client_sema);
+static DEFINE_MUTEX(client_mutex);
 
 static kmem_cache_t *stateowner_slab = NULL;
 static kmem_cache_t *file_slab = NULL;
@@ -91,13 +92,13 @@ static kmem_cache_t *deleg_slab = NULL;
 void
 nfs4_lock_state(void)
 {
-	down(&client_sema);
+	mutex_lock(&client_mutex);
 }
 
 void
 nfs4_unlock_state(void)
 {
-	up(&client_sema);
+	mutex_unlock(&client_mutex);
 }
 
 static inline u32
@@ -2639,7 +2640,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_lock
 	struct nfs4_stateid *lock_stp;
 	struct file *filp;
 	struct file_lock file_lock;
-	struct file_lock *conflock;
+	struct file_lock conflock;
 	int status = 0;
 	unsigned int strhashval;
 
@@ -2749,37 +2750,31 @@ nfsd4_lock(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_lock
 	* Note: locks.c uses the BKL to protect the inode's lock list.
 	*/
 
-	status = posix_lock_file(filp, &file_lock);
-	dprintk("NFSD: nfsd4_lock: posix_lock_file status %d\n",status);
+	/* XXX?: Just to divert the locks_release_private at the start of
+	 * locks_copy_lock: */
+	conflock.fl_ops = NULL;
+	conflock.fl_lmops = NULL;
+	status = posix_lock_file_conf(filp, &file_lock, &conflock);
+	dprintk("NFSD: nfsd4_lock: posix_lock_file_conf status %d\n",status);
 	switch (-status) {
 	case 0: /* success! */
 		update_stateid(&lock_stp->st_stateid);
 		memcpy(&lock->lk_resp_stateid, &lock_stp->st_stateid, 
 				sizeof(stateid_t));
-		goto out;
-	case (EAGAIN):
-		goto conflicting_lock;
+		break;
+	case (EAGAIN):		/* conflock holds conflicting lock */
+		status = nfserr_denied;
+		dprintk("NFSD: nfsd4_lock: conflicting lock found!\n");
+		nfs4_set_lock_denied(&conflock, &lock->lk_denied);
+		break;
 	case (EDEADLK):
 		status = nfserr_deadlock;
-		dprintk("NFSD: nfsd4_lock: posix_lock_file() failed! status %d\n",status);
-		goto out;
+		break;
 	default:        
-		status = nfserrno(status);
-		dprintk("NFSD: nfsd4_lock: posix_lock_file() failed! status %d\n",status);
-		goto out;
-	}
-
-conflicting_lock:
-	dprintk("NFSD: nfsd4_lock: conflicting lock found!\n");
-	status = nfserr_denied;
-	/* XXX There is a race here. Future patch needed to provide 
-	 * an atomic posix_lock_and_test_file
-	 */
-	if (!(conflock = posix_test_lock(filp, &file_lock))) {
-		status = nfserr_serverfault;
-		goto out;
+		dprintk("NFSD: nfsd4_lock: posix_lock_file_conf() failed! status %d\n",status);
+		status = nfserr_resource;
+		break;
 	}
-	nfs4_set_lock_denied(conflock, &lock->lk_denied);
 out:
 	if (status && lock->lk_is_new && lock_sop)
 		release_stateowner(lock_sop);
@@ -2800,7 +2795,7 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_lock
 	struct inode *inode;
 	struct file file;
 	struct file_lock file_lock;
-	struct file_lock *conflicting_lock;
+	struct file_lock conflock;
 	int status;
 
 	if (nfs4_in_grace())
@@ -2864,10 +2859,9 @@ nfsd4_lockt(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_lock
 	file.f_dentry = current_fh->fh_dentry;
 
 	status = nfs_ok;
-	conflicting_lock = posix_test_lock(&file, &file_lock);
-	if (conflicting_lock) {
+	if (posix_test_lock(&file, &file_lock, &conflock)) {
 		status = nfserr_denied;
-		nfs4_set_lock_denied(conflicting_lock, &lockt->lt_denied);
+		nfs4_set_lock_denied(&conflock, &lockt->lt_denied);
 	}
 out:
 	nfs4_unlock_state();
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index 69d3501173a8..03857fd81126 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -992,7 +992,7 @@ nfsd4_decode_compound(struct nfsd4_compoundargs *argp)
 	if (argp->opcnt > 100)
 		goto xdr_error;
 
-	if (argp->opcnt > sizeof(argp->iops)/sizeof(argp->iops[0])) {
+	if (argp->opcnt > ARRAY_SIZE(argp->iops)) {
 		argp->ops = kmalloc(argp->opcnt * sizeof(*argp->ops), GFP_KERNEL);
 		if (!argp->ops) {
 			argp->ops = argp->iops;
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index a0871b3efeb7..3ef017b3b5bd 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -105,7 +105,7 @@ static ssize_t nfsctl_transaction_write(struct file *file, const char __user *bu
 	char *data;
 	ssize_t rv;
 
-	if (ino >= sizeof(write_op)/sizeof(write_op[0]) || !write_op[ino])
+	if (ino >= ARRAY_SIZE(write_op) || !write_op[ino])
 		return -EINVAL;
 
 	data = simple_transaction_get(file, buf, size);
@@ -134,7 +134,7 @@ static ssize_t nfsctl_transaction_read(struct file *file, char __user *buf, size
 	return simple_transaction_read(file, buf, size, pos);
 }
 
-static struct file_operations transaction_ops = {
+static const struct file_operations transaction_ops = {
 	.write		= nfsctl_transaction_write,
 	.read		= nfsctl_transaction_read,
 	.release	= simple_transaction_release,
@@ -146,7 +146,7 @@ static int exports_open(struct inode *inode, struct file *file)
 	return seq_open(file, &nfs_exports_op);
 }
 
-static struct file_operations exports_operations = {
+static const struct file_operations exports_operations = {
 	.open		= exports_open,
 	.read		= seq_read,
 	.llseek		= seq_lseek,
diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c
index 7a3e397b4ed3..3f2ec2e6d06c 100644
--- a/fs/nfsd/nfsfh.c
+++ b/fs/nfsd/nfsfh.c
@@ -506,7 +506,7 @@ fh_put(struct svc_fh *fhp)
 		nfsd_nr_put++;
 	}
 	if (exp) {
-		svc_export_put(&exp->h, &svc_export_cache);
+		cache_put(&exp->h, &svc_export_cache);
 		fhp->fh_export = NULL;
 	}
 	return;
diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index 1d163b616915..3790727e5dfd 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -72,7 +72,7 @@ static struct svc_version *	nfsd_acl_version[] = {
 };
 
 #define NFSD_ACL_MINVERS            2
-#define NFSD_ACL_NRVERS		(sizeof(nfsd_acl_version)/sizeof(nfsd_acl_version[0]))
+#define NFSD_ACL_NRVERS		ARRAY_SIZE(nfsd_acl_version)
 static struct svc_version *nfsd_acl_versions[NFSD_ACL_NRVERS];
 
 static struct svc_program	nfsd_acl_program = {
@@ -101,7 +101,7 @@ static struct svc_version *	nfsd_version[] = {
 };
 
 #define NFSD_MINVERS    	2
-#define NFSD_NRVERS		(sizeof(nfsd_version)/sizeof(nfsd_version[0]))
+#define NFSD_NRVERS		ARRAY_SIZE(nfsd_version)
 static struct svc_version *nfsd_versions[NFSD_NRVERS];
 
 struct svc_program		nfsd_program = {
diff --git a/fs/nfsd/stats.c b/fs/nfsd/stats.c
index 1cf955bcc526..57265d563804 100644
--- a/fs/nfsd/stats.c
+++ b/fs/nfsd/stats.c
@@ -80,7 +80,7 @@ static int nfsd_proc_open(struct inode *inode, struct file *file)
 	return single_open(file, nfsd_proc_show, NULL);
 }
 
-static struct file_operations nfsd_proc_fops = {
+static const struct file_operations nfsd_proc_fops = {
 	.owner = THIS_MODULE,
 	.open = nfsd_proc_open,
 	.read  = seq_read,
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 5320e5afaddb..31018333dc38 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -706,7 +706,7 @@ nfsd_close(struct file *filp)
  * after it.
  */
 static inline int nfsd_dosync(struct file *filp, struct dentry *dp,
-			      struct file_operations *fop)
+			      const struct file_operations *fop)
 {
 	struct inode *inode = dp->d_inode;
 	int (*fsync) (struct file *, struct dentry *, int);
diff --git a/fs/nls/Kconfig b/fs/nls/Kconfig
index 0ab8f00bdbb2..976ecccd6f56 100644
--- a/fs/nls/Kconfig
+++ b/fs/nls/Kconfig
@@ -491,7 +491,7 @@ config NLS_KOI8_U
 	  (koi8-u) and Belarusian (koi8-ru) character sets.
 
 config NLS_UTF8
-	tristate "NLS UTF8"
+	tristate "NLS UTF-8"
 	depends on NLS
 	help
 	  If you want to display filenames with native language characters
diff --git a/fs/nls/nls_euc-jp.c b/fs/nls/nls_euc-jp.c
index 80f108ae6661..06640c3e4021 100644
--- a/fs/nls/nls_euc-jp.c
+++ b/fs/nls/nls_euc-jp.c
@@ -268,8 +268,6 @@ static unsigned char euc2sjisibm_g3upper_map[][2] = {
 	{0xFC, 0x4B},
 };
 
-#define MAP_ELEMENT_OF(map)	(sizeof(map) / sizeof(map[0]))
-
 static inline int sjisibm2euc(unsigned char *euc, const unsigned char sjis_hi,
 			      const unsigned char sjis_lo);
 static inline int euc2sjisibm_jisx0212(unsigned char *sjis, const unsigned char euc_hi,
@@ -310,7 +308,7 @@ static inline int euc2sjisibm_jisx0212(unsigned char *sjis, const unsigned char
 	unsigned short euc;
 
 	min_index = 0;
-	max_index = MAP_ELEMENT_OF(euc2sjisibm_jisx0212_map) - 1;
+	max_index = ARRAY_SIZE(euc2sjisibm_jisx0212_map) - 1;
 	euc = (euc_hi << 8) | euc_lo;
 
 	while (min_index <= max_index) {
@@ -339,7 +337,7 @@ static inline int euc2sjisibm_g3upper(unsigned char *sjis, const unsigned char e
 	else
 		index = ((euc_hi << 8) | euc_lo) - 0xF4A1 + 12;
 
-	if ((index < 0) || (index >= MAP_ELEMENT_OF(euc2sjisibm_g3upper_map)))
+	if ((index < 0) || (index >= ARRAY_SIZE(euc2sjisibm_g3upper_map)))
 		return 0;
 
 	sjis[0] = euc2sjisibm_g3upper_map[index][0];
diff --git a/fs/ntfs/ChangeLog b/fs/ntfs/ChangeLog
index 9d8ffa89e2c2..35cc4b1d60f7 100644
--- a/fs/ntfs/ChangeLog
+++ b/fs/ntfs/ChangeLog
@@ -16,8 +16,34 @@ ToDo/Notes:
 	  inode having been discarded already.  Whether this can actually ever
 	  happen is unclear however so it is worth waiting until someone hits
 	  the problem.
-	- Enable the code for setting the NT4 compatibility flag when we start
-	  making NTFS 1.2 specific modifications.
+
+2.1.27 - Various bug fixes and cleanups.
+
+	- Fix two compiler warnings on Alpha.  Thanks to Andrew Morton for
+	  reporting them.
+	- Fix an (innocent) off-by-one error in the runlist code.
+	- Fix a buggette in an "should be impossible" case handling where we
+	  continued the attribute lookup loop instead of aborting it.
+	- Use buffer_migrate_page() for the ->migratepage function of all ntfs
+	  address space operations.
+	- Fix comparison of $MFT and $MFTMirr to not bail out when there are
+	  unused, invalid mft records which are the same in both $MFT and
+	  $MFTMirr.
+	- Add support for sparse files which have a compression unit of 0.
+	- Remove all the make_bad_inode() calls.  This should only be called
+	  from read inode and new inode code paths.
+	- Limit name length in fs/ntfs/unistr.c::ntfs_nlstoucs() to maximum
+	  allowed by NTFS, i.e. 255 Unicode characters, not including the
+	  terminating NULL (which is not stored on disk).
+	- Improve comments on file attribute flags in fs/ntfs/layout.h.
+	- Fix a bug in fs/ntfs/inode.c::ntfs_read_locked_index_inode() where we
+	  forgot to update a temporary variable so loading index inodes which
+	  have an index allocation attribute failed.
+	- Add a missing call to flush_dcache_mft_record_page() in
+	  fs/ntfs/inode.c::ntfs_write_inode().
+	- Handle the recently introduced -ENAMETOOLONG return value from
+	  fs/ntfs/unistr.c::ntfs_nlstoucs() in fs/ntfs/namei.c::ntfs_lookup().
+	- Semaphore to mutex conversion.  (Ingo Molnar)
 
 2.1.26 - Minor bug fixes and updates.
 
diff --git a/fs/ntfs/Makefile b/fs/ntfs/Makefile
index d95fac7fdeb6..e27b4eacffbf 100644
--- a/fs/ntfs/Makefile
+++ b/fs/ntfs/Makefile
@@ -6,7 +6,7 @@ ntfs-objs := aops.o attrib.o collate.o compress.o debug.o dir.o file.o \
 	     index.o inode.o mft.o mst.o namei.o runlist.o super.o sysctl.o \
 	     unistr.o upcase.o
 
-EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.26\"
+EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.27\"
 
 ifeq ($(CONFIG_NTFS_DEBUG),y)
 EXTRA_CFLAGS += -DDEBUG
diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
index 7e361da770b3..580412d330cb 100644
--- a/fs/ntfs/aops.c
+++ b/fs/ntfs/aops.c
@@ -22,6 +22,7 @@
  */
 
 #include <linux/errno.h>
+#include <linux/fs.h>
 #include <linux/mm.h>
 #include <linux/pagemap.h>
 #include <linux/swap.h>
@@ -1277,18 +1278,18 @@ unm_done:
 		
 		tni = locked_nis[nr_locked_nis];
 		/* Get the base inode. */
-		down(&tni->extent_lock);
+		mutex_lock(&tni->extent_lock);
 		if (tni->nr_extents >= 0)
 			base_tni = tni;
 		else {
 			base_tni = tni->ext.base_ntfs_ino;
 			BUG_ON(!base_tni);
 		}
-		up(&tni->extent_lock);
+		mutex_unlock(&tni->extent_lock);
 		ntfs_debug("Unlocking %s inode 0x%lx.",
 				tni == base_tni ? "base" : "extent",
 				tni->mft_no);
-		up(&tni->mrec_lock);
+		mutex_unlock(&tni->mrec_lock);
 		atomic_dec(&tni->count);
 		iput(VFS_I(base_tni));
 	}
@@ -1529,7 +1530,6 @@ err_out:
 				"error %i.", err);
 		SetPageError(page);
 		NVolSetErrors(ni->vol);
-		make_bad_inode(vi);
 	}
 	unlock_page(page);
 	if (ctx)
@@ -1551,6 +1551,9 @@ struct address_space_operations ntfs_aops = {
 #ifdef NTFS_RW
 	.writepage	= ntfs_writepage,	/* Write dirty page to disk. */
 #endif /* NTFS_RW */
+	.migratepage	= buffer_migrate_page,	/* Move a page cache page from
+						   one physical page to an
+						   other. */
 };
 
 /**
@@ -1567,6 +1570,9 @@ struct address_space_operations ntfs_mst_aops = {
 						   without touching the buffers
 						   belonging to the page. */
 #endif /* NTFS_RW */
+	.migratepage	= buffer_migrate_page,	/* Move a page cache page from
+						   one physical page to an
+						   other. */
 };
 
 #ifdef NTFS_RW
diff --git a/fs/ntfs/attrib.c b/fs/ntfs/attrib.c
index 9480a0526cd3..1663f5c3c6aa 100644
--- a/fs/ntfs/attrib.c
+++ b/fs/ntfs/attrib.c
@@ -1,7 +1,7 @@
 /**
  * attrib.c - NTFS attribute operations.  Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  * Copyright (c) 2002 Richard Russon
  *
  * This program/include file is free software; you can redistribute it and/or
@@ -1048,7 +1048,7 @@ do_next_attr_loop:
 				le32_to_cpu(ctx->mrec->bytes_allocated))
 			break;
 		if (a->type == AT_END)
-			continue;
+			break;
 		if (!a->length)
 			break;
 		if (al_entry->instance != a->instance)
@@ -1695,7 +1695,9 @@ int ntfs_attr_make_non_resident(ntfs_inode *ni, const u32 data_size)
 			a->data.non_resident.initialized_size =
 			cpu_to_sle64(attr_size);
 	if (NInoSparse(ni) || NInoCompressed(ni)) {
-		a->data.non_resident.compression_unit = 4;
+		a->data.non_resident.compression_unit = 0;
+		if (NInoCompressed(ni) || vol->major_ver < 3)
+			a->data.non_resident.compression_unit = 4;
 		a->data.non_resident.compressed_size =
 				a->data.non_resident.allocated_size;
 	} else
@@ -1714,13 +1716,20 @@ int ntfs_attr_make_non_resident(ntfs_inode *ni, const u32 data_size)
 	ni->allocated_size = new_size;
 	if (NInoSparse(ni) || NInoCompressed(ni)) {
 		ni->itype.compressed.size = ni->allocated_size;
-		ni->itype.compressed.block_size = 1U <<
-				(a->data.non_resident.compression_unit +
-				vol->cluster_size_bits);
-		ni->itype.compressed.block_size_bits =
-				ffs(ni->itype.compressed.block_size) - 1;
-		ni->itype.compressed.block_clusters = 1U <<
-				a->data.non_resident.compression_unit;
+		if (a->data.non_resident.compression_unit) {
+			ni->itype.compressed.block_size = 1U << (a->data.
+					non_resident.compression_unit +
+					vol->cluster_size_bits);
+			ni->itype.compressed.block_size_bits =
+					ffs(ni->itype.compressed.block_size) -
+					1;
+			ni->itype.compressed.block_clusters = 1U <<
+					a->data.non_resident.compression_unit;
+		} else {
+			ni->itype.compressed.block_size = 0;
+			ni->itype.compressed.block_size_bits = 0;
+			ni->itype.compressed.block_clusters = 0;
+		}
 		vi->i_blocks = ni->itype.compressed.size >> 9;
 	} else
 		vi->i_blocks = ni->allocated_size >> 9;
@@ -2429,16 +2438,12 @@ undo_alloc:
 				"chkdsk to recover.", IS_ERR(m) ?
 				"restore attribute search context" :
 				"truncate attribute runlist");
-		make_bad_inode(vi);
-		make_bad_inode(VFS_I(base_ni));
 		NVolSetErrors(vol);
 	} else if (mp_rebuilt) {
 		if (ntfs_attr_record_resize(m, a, attr_len)) {
 			ntfs_error(vol->sb, "Failed to restore attribute "
 					"record in error code path.  Run "
 					"chkdsk to recover.");
-			make_bad_inode(vi);
-			make_bad_inode(VFS_I(base_ni));
 			NVolSetErrors(vol);
 		} else /* if (success) */ {
 			if (ntfs_mapping_pairs_build(vol, (u8*)a + le16_to_cpu(
@@ -2451,8 +2456,6 @@ undo_alloc:
 						"mapping pairs array in error "
 						"code path.  Run chkdsk to "
 						"recover.");
-				make_bad_inode(vi);
-				make_bad_inode(VFS_I(base_ni));
 				NVolSetErrors(vol);
 			}
 			flush_dcache_mft_record_page(ctx->ntfs_ino);
diff --git a/fs/ntfs/compress.c b/fs/ntfs/compress.c
index 25d24106f893..68a607ff9fd3 100644
--- a/fs/ntfs/compress.c
+++ b/fs/ntfs/compress.c
@@ -67,7 +67,7 @@ static DEFINE_SPINLOCK(ntfs_cb_lock);
 /**
  * allocate_compression_buffers - allocate the decompression buffers
  *
- * Caller has to hold the ntfs_lock semaphore.
+ * Caller has to hold the ntfs_lock mutex.
  *
  * Return 0 on success or -ENOMEM if the allocations failed.
  */
@@ -84,7 +84,7 @@ int allocate_compression_buffers(void)
 /**
  * free_compression_buffers - free the decompression buffers
  *
- * Caller has to hold the ntfs_lock semaphore.
+ * Caller has to hold the ntfs_lock mutex.
  */
 void free_compression_buffers(void)
 {
diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
index b0690d4c8906..d1e2c6f9f05e 100644
--- a/fs/ntfs/dir.c
+++ b/fs/ntfs/dir.c
@@ -1136,7 +1136,7 @@ static int ntfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	if (fpos == 1) {
 		ntfs_debug("Calling filldir for .. with len 2, fpos 0x1, "
 				"inode 0x%lx, DT_DIR.",
-				parent_ino(filp->f_dentry));
+				(unsigned long)parent_ino(filp->f_dentry));
 		rc = filldir(dirent, "..", 2, fpos,
 				parent_ino(filp->f_dentry), DT_DIR);
 		if (rc)
@@ -1553,7 +1553,7 @@ static int ntfs_dir_fsync(struct file *filp, struct dentry *dentry,
 
 #endif /* NTFS_RW */
 
-struct file_operations ntfs_dir_ops = {
+const struct file_operations ntfs_dir_ops = {
 	.llseek		= generic_file_llseek,	/* Seek inside directory. */
 	.read		= generic_read_dir,	/* Return -EISDIR. */
 	.readdir	= ntfs_readdir,		/* Read directory contents. */
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index 5027d3d1b3fe..c63a83e8da98 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -943,7 +943,8 @@ rl_not_mapped_enoent:
 		}
 		ni->runlist.rl = rl;
 		status.runlist_merged = 1;
-		ntfs_debug("Allocated cluster, lcn 0x%llx.", lcn);
+		ntfs_debug("Allocated cluster, lcn 0x%llx.",
+				(unsigned long long)lcn);
 		/* Map and lock the mft record and get the attribute record. */
 		if (!NInoAttr(ni))
 			base_ni = ni;
@@ -1206,8 +1207,6 @@ rl_not_mapped_enoent:
 					"attribute runlist in error code "
 					"path.  Run chkdsk to recover the "
 					"lost cluster.");
-			make_bad_inode(vi);
-			make_bad_inode(VFS_I(base_ni));
 			NVolSetErrors(vol);
 		} else /* if (success) */ {
 			status.runlist_merged = 0;
@@ -1238,8 +1237,6 @@ rl_not_mapped_enoent:
 			ntfs_error(vol->sb, "Failed to restore attribute "
 					"record in error code path.  Run "
 					"chkdsk to recover.");
-			make_bad_inode(vi);
-			make_bad_inode(VFS_I(base_ni));
 			NVolSetErrors(vol);
 		} else /* if (success) */ {
 			if (ntfs_mapping_pairs_build(vol, (u8*)a +
@@ -1252,8 +1249,6 @@ rl_not_mapped_enoent:
 						"mapping pairs array in error "
 						"code path.  Run chkdsk to "
 						"recover.");
-				make_bad_inode(vi);
-				make_bad_inode(VFS_I(base_ni));
 				NVolSetErrors(vol);
 			}
 			flush_dcache_mft_record_page(ctx->ntfs_ino);
@@ -1622,11 +1617,8 @@ err_out:
 		unmap_mft_record(base_ni);
 	ntfs_error(vi->i_sb, "Failed to update initialized_size/i_size (error "
 			"code %i).", err);
-	if (err != -ENOMEM) {
+	if (err != -ENOMEM)
 		NVolSetErrors(ni->vol);
-		make_bad_inode(VFS_I(base_ni));
-		make_bad_inode(vi);
-	}
 	return err;
 }
 
@@ -1801,8 +1793,6 @@ err_out:
 		ntfs_error(vi->i_sb, "Resident attribute commit write failed "
 				"with error %i.", err);
 		NVolSetErrors(ni->vol);
-		make_bad_inode(VFS_I(base_ni));
-		make_bad_inode(vi);
 	}
 	if (ctx)
 		ntfs_attr_put_search_ctx(ctx);
@@ -2304,7 +2294,7 @@ static int ntfs_file_fsync(struct file *filp, struct dentry *dentry,
 
 #endif /* NTFS_RW */
 
-struct file_operations ntfs_file_ops = {
+const struct file_operations ntfs_file_ops = {
 	.llseek		= generic_file_llseek,	 /* Seek inside file. */
 	.read		= generic_file_read,	 /* Read from file. */
 	.aio_read	= generic_file_aio_read, /* Async read from file. */
@@ -2347,6 +2337,6 @@ struct inode_operations ntfs_file_inode_ops = {
 #endif /* NTFS_RW */
 };
 
-struct file_operations ntfs_empty_file_ops = {};
+const struct file_operations ntfs_empty_file_ops = {};
 
 struct inode_operations ntfs_empty_inode_ops = {};
diff --git a/fs/ntfs/inode.c b/fs/ntfs/inode.c
index 55263b7de9c0..4c86b7e1d1eb 100644
--- a/fs/ntfs/inode.c
+++ b/fs/ntfs/inode.c
@@ -1,7 +1,7 @@
 /**
  * inode.c - NTFS kernel inode handling. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  *
  * This program/include file is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License as published
@@ -19,13 +19,19 @@
  * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
-#include <linux/pagemap.h>
 #include <linux/buffer_head.h>
-#include <linux/smp_lock.h>
-#include <linux/quotaops.h>
+#include <linux/fs.h>
+#include <linux/mm.h>
 #include <linux/mount.h>
+#include <linux/mutex.h>
+#include <linux/pagemap.h>
+#include <linux/quotaops.h>
+#include <linux/slab.h>
+#include <linux/smp_lock.h>
 
 #include "aops.h"
+#include "attrib.h"
+#include "bitmap.h"
 #include "dir.h"
 #include "debug.h"
 #include "inode.h"
@@ -382,7 +388,7 @@ void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni)
 	atomic_set(&ni->count, 1);
 	ni->vol = NTFS_SB(sb);
 	ntfs_init_runlist(&ni->runlist);
-	init_MUTEX(&ni->mrec_lock);
+	mutex_init(&ni->mrec_lock);
 	ni->page = NULL;
 	ni->page_ofs = 0;
 	ni->attr_list_size = 0;
@@ -394,7 +400,7 @@ void __ntfs_init_inode(struct super_block *sb, ntfs_inode *ni)
 	ni->itype.index.collation_rule = 0;
 	ni->itype.index.block_size_bits = 0;
 	ni->itype.index.vcn_size_bits = 0;
-	init_MUTEX(&ni->extent_lock);
+	mutex_init(&ni->extent_lock);
 	ni->nr_extents = 0;
 	ni->ext.base_ntfs_ino = NULL;
 }
@@ -1064,10 +1070,10 @@ skip_large_dir_stuff:
 		if (a->non_resident) {
 			NInoSetNonResident(ni);
 			if (NInoCompressed(ni) || NInoSparse(ni)) {
-				if (a->data.non_resident.compression_unit !=
-						4) {
+				if (NInoCompressed(ni) && a->data.non_resident.
+						compression_unit != 4) {
 					ntfs_error(vi->i_sb, "Found "
-							"nonstandard "
+							"non-standard "
 							"compression unit (%u "
 							"instead of 4).  "
 							"Cannot handle this.",
@@ -1076,16 +1082,26 @@ skip_large_dir_stuff:
 					err = -EOPNOTSUPP;
 					goto unm_err_out;
 				}
-				ni->itype.compressed.block_clusters = 1U <<
-						a->data.non_resident.
-						compression_unit;
-				ni->itype.compressed.block_size = 1U << (
-						a->data.non_resident.
-						compression_unit +
-						vol->cluster_size_bits);
-				ni->itype.compressed.block_size_bits = ffs(
-						ni->itype.compressed.
-						block_size) - 1;
+				if (a->data.non_resident.compression_unit) {
+					ni->itype.compressed.block_size = 1U <<
+							(a->data.non_resident.
+							compression_unit +
+							vol->cluster_size_bits);
+					ni->itype.compressed.block_size_bits =
+							ffs(ni->itype.
+							compressed.
+							block_size) - 1;
+					ni->itype.compressed.block_clusters =
+							1U << a->data.
+							non_resident.
+							compression_unit;
+				} else {
+					ni->itype.compressed.block_size = 0;
+					ni->itype.compressed.block_size_bits =
+							0;
+					ni->itype.compressed.block_clusters =
+							0;
+				}
 				ni->itype.compressed.size = sle64_to_cpu(
 						a->data.non_resident.
 						compressed_size);
@@ -1338,8 +1354,9 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 			goto unm_err_out;
 		}
 		if (NInoCompressed(ni) || NInoSparse(ni)) {
-			if (a->data.non_resident.compression_unit != 4) {
-				ntfs_error(vi->i_sb, "Found nonstandard "
+			if (NInoCompressed(ni) && a->data.non_resident.
+					compression_unit != 4) {
+				ntfs_error(vi->i_sb, "Found non-standard "
 						"compression unit (%u instead "
 						"of 4).  Cannot handle this.",
 						a->data.non_resident.
@@ -1347,13 +1364,22 @@ static int ntfs_read_locked_attr_inode(struct inode *base_vi, struct inode *vi)
 				err = -EOPNOTSUPP;
 				goto unm_err_out;
 			}
-			ni->itype.compressed.block_clusters = 1U <<
-					a->data.non_resident.compression_unit;
-			ni->itype.compressed.block_size = 1U << (
-					a->data.non_resident.compression_unit +
-					vol->cluster_size_bits);
-			ni->itype.compressed.block_size_bits = ffs(
-					ni->itype.compressed.block_size) - 1;
+			if (a->data.non_resident.compression_unit) {
+				ni->itype.compressed.block_size = 1U <<
+						(a->data.non_resident.
+						compression_unit +
+						vol->cluster_size_bits);
+				ni->itype.compressed.block_size_bits =
+						ffs(ni->itype.compressed.
+						block_size) - 1;
+				ni->itype.compressed.block_clusters = 1U <<
+						a->data.non_resident.
+						compression_unit;
+			} else {
+				ni->itype.compressed.block_size = 0;
+				ni->itype.compressed.block_size_bits = 0;
+				ni->itype.compressed.block_clusters = 0;
+			}
 			ni->itype.compressed.size = sle64_to_cpu(
 					a->data.non_resident.compressed_size);
 		}
@@ -1406,7 +1432,6 @@ err_out:
 			"Run chkdsk.", err, vi->i_ino, ni->type, ni->name_len,
 			base_vi->i_ino);
 	make_bad_inode(vi);
-	make_bad_inode(base_vi);
 	if (err != -ENOMEM)
 		NVolSetErrors(vol);
 	return err;
@@ -1591,6 +1616,7 @@ static int ntfs_read_locked_index_inode(struct inode *base_vi, struct inode *vi)
 					"$INDEX_ALLOCATION attribute.");
 		goto unm_err_out;
 	}
+	a = ctx->attr;
 	if (!a->non_resident) {
 		ntfs_error(vi->i_sb, "$INDEX_ALLOCATION attribute is "
 				"resident.");
@@ -2823,11 +2849,8 @@ done:
 old_bad_out:
 	old_size = -1;
 bad_out:
-	if (err != -ENOMEM && err != -EOPNOTSUPP) {
-		make_bad_inode(vi);
-		make_bad_inode(VFS_I(base_ni));
+	if (err != -ENOMEM && err != -EOPNOTSUPP)
 		NVolSetErrors(vol);
-	}
 	if (err != -EOPNOTSUPP)
 		NInoSetTruncateFailed(ni);
 	else if (old_size >= 0)
@@ -2842,11 +2865,8 @@ out:
 	ntfs_debug("Failed.  Returning error code %i.", err);
 	return err;
 conv_err_out:
-	if (err != -ENOMEM && err != -EOPNOTSUPP) {
-		make_bad_inode(vi);
-		make_bad_inode(VFS_I(base_ni));
+	if (err != -ENOMEM && err != -EOPNOTSUPP)
 		NVolSetErrors(vol);
-	}
 	if (err != -EOPNOTSUPP)
 		NInoSetTruncateFailed(ni);
 	else
@@ -3044,15 +3064,18 @@ int ntfs_write_inode(struct inode *vi, int sync)
 	 * record will be cleaned and written out to disk below, i.e. before
 	 * this function returns.
 	 */
-	if (modified && !NInoTestSetDirty(ctx->ntfs_ino))
-		mark_ntfs_record_dirty(ctx->ntfs_ino->page,
-				ctx->ntfs_ino->page_ofs);
+	if (modified) {
+		flush_dcache_mft_record_page(ctx->ntfs_ino);
+		if (!NInoTestSetDirty(ctx->ntfs_ino))
+			mark_ntfs_record_dirty(ctx->ntfs_ino->page,
+					ctx->ntfs_ino->page_ofs);
+	}
 	ntfs_attr_put_search_ctx(ctx);
 	/* Now the access times are updated, write the base mft record. */
 	if (NInoDirty(ni))
 		err = write_mft_record(ni, m, sync);
 	/* Write all attached extent mft records. */
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	if (ni->nr_extents > 0) {
 		ntfs_inode **extent_nis = ni->ext.extent_ntfs_inos;
 		int i;
@@ -3079,7 +3102,7 @@ int ntfs_write_inode(struct inode *vi, int sync)
 			}
 		}
 	}
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 	unmap_mft_record(ni);
 	if (unlikely(err))
 		goto err_out;
@@ -3094,9 +3117,7 @@ err_out:
 				"retries later.");
 		mark_inode_dirty(vi);
 	} else {
-		ntfs_error(vi->i_sb, "Failed (error code %i):  Marking inode "
-				"as bad.  You should run chkdsk.", -err);
-		make_bad_inode(vi);
+		ntfs_error(vi->i_sb, "Failed (error %i):  Run chkdsk.", -err);
 		NVolSetErrors(ni->vol);
 	}
 	return err;
diff --git a/fs/ntfs/inode.h b/fs/ntfs/inode.h
index 3de5c0231966..f088291e017c 100644
--- a/fs/ntfs/inode.h
+++ b/fs/ntfs/inode.h
@@ -24,12 +24,13 @@
 #ifndef _LINUX_NTFS_INODE_H
 #define _LINUX_NTFS_INODE_H
 
-#include <linux/mm.h>
+#include <asm/atomic.h>
+
 #include <linux/fs.h>
-#include <linux/seq_file.h>
 #include <linux/list.h>
-#include <asm/atomic.h>
-#include <asm/semaphore.h>
+#include <linux/mm.h>
+#include <linux/mutex.h>
+#include <linux/seq_file.h>
 
 #include "layout.h"
 #include "volume.h"
@@ -81,7 +82,7 @@ struct _ntfs_inode {
 	 * The following fields are only valid for real inodes and extent
 	 * inodes.
 	 */
-	struct semaphore mrec_lock; /* Lock for serializing access to the
+	struct mutex mrec_lock;	/* Lock for serializing access to the
 				   mft record belonging to this inode. */
 	struct page *page;	/* The page containing the mft record of the
 				   inode. This should only be touched by the
@@ -119,7 +120,7 @@ struct _ntfs_inode {
 			u8 block_clusters;	/* Number of clusters per cb. */
 		} compressed;
 	} itype;
-	struct semaphore extent_lock;	/* Lock for accessing/modifying the
+	struct mutex extent_lock;	/* Lock for accessing/modifying the
 					   below . */
 	s32 nr_extents;	/* For a base mft record, the number of attached extent
 			   inodes (0 if none), for extent records and for fake
diff --git a/fs/ntfs/layout.h b/fs/ntfs/layout.h
index bb408d4dcbb0..d34b93cb8b48 100644
--- a/fs/ntfs/layout.h
+++ b/fs/ntfs/layout.h
@@ -769,7 +769,7 @@ typedef struct {
 				compressed.  (This effectively limits the
 				compression unit size to be a power of two
 				clusters.)  WinNT4 only uses a value of 4.
-				Sparse files also have this set to 4. */
+				Sparse files have this set to 0 on XPSP2. */
 /* 35*/			u8 reserved[5];		/* Align to 8-byte boundary. */
 /* The sizes below are only used when lowest_vcn is zero, as otherwise it would
    be difficult to keep them up-to-date.*/
@@ -801,13 +801,16 @@ typedef struct {
 typedef ATTR_RECORD ATTR_REC;
 
 /*
- * File attribute flags (32-bit).
+ * File attribute flags (32-bit) appearing in the file_attributes fields of the
+ * STANDARD_INFORMATION attribute of MFT_RECORDs and the FILENAME_ATTR
+ * attributes of MFT_RECORDs and directory index entries.
+ *
+ * All of the below flags appear in the directory index entries but only some
+ * appear in the STANDARD_INFORMATION attribute whilst only some others appear
+ * in the FILENAME_ATTR attribute of MFT_RECORDs.  Unless otherwise stated the
+ * flags appear in all of the above.
  */
 enum {
-	/*
-	 * The following flags are only present in the STANDARD_INFORMATION
-	 * attribute (in the field file_attributes).
-	 */
 	FILE_ATTR_READONLY		= const_cpu_to_le32(0x00000001),
 	FILE_ATTR_HIDDEN		= const_cpu_to_le32(0x00000002),
 	FILE_ATTR_SYSTEM		= const_cpu_to_le32(0x00000004),
@@ -839,18 +842,14 @@ enum {
 	   F_A_COMPRESSED, and F_A_ENCRYPTED and preserves the rest.  This mask
 	   is used to to obtain all flags that are valid for setting. */
 	/*
-	 * The following flag is only present in the FILE_NAME attribute (in
-	 * the field file_attributes).
+	 * The flag FILE_ATTR_DUP_FILENAME_INDEX_PRESENT is present in all
+	 * FILENAME_ATTR attributes but not in the STANDARD_INFORMATION
+	 * attribute of an mft record.
 	 */
 	FILE_ATTR_DUP_FILE_NAME_INDEX_PRESENT	= const_cpu_to_le32(0x10000000),
 	/* Note, this is a copy of the corresponding bit from the mft record,
 	   telling us whether this is a directory or not, i.e. whether it has
 	   an index root attribute or not. */
-	/*
-	 * The following flag is present both in the STANDARD_INFORMATION
-	 * attribute and in the FILE_NAME attribute (in the field
-	 * file_attributes).
-	 */
 	FILE_ATTR_DUP_VIEW_INDEX_PRESENT	= const_cpu_to_le32(0x20000000),
 	/* Note, this is a copy of the corresponding bit from the mft record,
 	   telling us whether this file has a view index present (eg. object id
@@ -891,7 +890,7 @@ typedef struct {
 					   Windows this is only updated when
 					   accessed if some time delta has
 					   passed since the last update. Also,
-					   last access times updates can be
+					   last access time updates can be
 					   disabled altogether for speed. */
 /* 32*/	FILE_ATTR_FLAGS file_attributes; /* Flags describing the file. */
 /* 36*/	union {
@@ -1076,16 +1075,21 @@ typedef struct {
 /* 20*/	sle64 last_access_time;		/* Time this mft record was last
 					   accessed. */
 /* 28*/	sle64 allocated_size;		/* Byte size of on-disk allocated space
-					   for the data attribute.  So for
-					   normal $DATA, this is the
+					   for the unnamed data attribute.  So
+					   for normal $DATA, this is the
 					   allocated_size from the unnamed
 					   $DATA attribute and for compressed
 					   and/or sparse $DATA, this is the
 					   compressed_size from the unnamed
-					   $DATA attribute.  NOTE: This is a
-					   multiple of the cluster size. */
-/* 30*/	sle64 data_size;		/* Byte size of actual data in data
-					   attribute. */
+					   $DATA attribute.  For a directory or
+					   other inode without an unnamed $DATA
+					   attribute, this is always 0.  NOTE:
+					   This is a multiple of the cluster
+					   size. */
+/* 30*/	sle64 data_size;		/* Byte size of actual data in unnamed
+					   data attribute.  For a directory or
+					   other inode without an unnamed $DATA
+					   attribute, this is always 0. */
 /* 38*/	FILE_ATTR_FLAGS file_attributes;	/* Flags describing the file. */
 /* 3c*/	union {
 	/* 3c*/	struct {
diff --git a/fs/ntfs/logfile.c b/fs/ntfs/logfile.c
index 0fd70295cca6..4af2ad1193ec 100644
--- a/fs/ntfs/logfile.c
+++ b/fs/ntfs/logfile.c
@@ -515,10 +515,10 @@ BOOL ntfs_check_logfile(struct inode *log_vi, RESTART_PAGE_HEADER **rp)
 		log_page_size = PAGE_CACHE_SIZE;
 	log_page_mask = log_page_size - 1;
 	/*
-	 * Use generic_ffs() instead of ffs() to enable the compiler to
+	 * Use ntfs_ffs() instead of ffs() to enable the compiler to
 	 * optimize log_page_size and log_page_bits into constants.
 	 */
-	log_page_bits = generic_ffs(log_page_size) - 1;
+	log_page_bits = ntfs_ffs(log_page_size) - 1;
 	size &= ~(s64)(log_page_size - 1);
 	/*
 	 * Ensure the log file is big enough to store at least the two restart
diff --git a/fs/ntfs/mft.c b/fs/ntfs/mft.c
index 6499aafc2258..2438c00ec0ce 100644
--- a/fs/ntfs/mft.c
+++ b/fs/ntfs/mft.c
@@ -93,6 +93,7 @@ static inline MFT_RECORD *map_mft_record_page(ntfs_inode *ni)
 				"Run chkdsk.", ni->mft_no);
 		ntfs_unmap_page(page);
 		page = ERR_PTR(-EIO);
+		NVolSetErrors(vol);
 	}
 err_out:
 	ni->page = NULL;
@@ -104,8 +105,8 @@ err_out:
  * map_mft_record - map, pin and lock an mft record
  * @ni:		ntfs inode whose MFT record to map
  *
- * First, take the mrec_lock semaphore. We might now be sleeping, while waiting
- * for the semaphore if it was already locked by someone else.
+ * First, take the mrec_lock mutex.  We might now be sleeping, while waiting
+ * for the mutex if it was already locked by someone else.
  *
  * The page of the record is mapped using map_mft_record_page() before being
  * returned to the caller.
@@ -135,9 +136,9 @@ err_out:
  * So that code will end up having to own the mrec_lock of all mft
  * records/inodes present in the page before I/O can proceed. In that case we
  * wouldn't need to bother with PG_locked and PG_uptodate as nobody will be
- * accessing anything without owning the mrec_lock semaphore. But we do need
- * to use them because of the read_cache_page() invocation and the code becomes
- * so much simpler this way that it is well worth it.
+ * accessing anything without owning the mrec_lock mutex.  But we do need to
+ * use them because of the read_cache_page() invocation and the code becomes so
+ * much simpler this way that it is well worth it.
  *
  * The mft record is now ours and we return a pointer to it. You need to check
  * the returned pointer with IS_ERR() and if that is true, PTR_ERR() will return
@@ -160,13 +161,13 @@ MFT_RECORD *map_mft_record(ntfs_inode *ni)
 	atomic_inc(&ni->count);
 
 	/* Serialize access to this mft record. */
-	down(&ni->mrec_lock);
+	mutex_lock(&ni->mrec_lock);
 
 	m = map_mft_record_page(ni);
 	if (likely(!IS_ERR(m)))
 		return m;
 
-	up(&ni->mrec_lock);
+	mutex_unlock(&ni->mrec_lock);
 	atomic_dec(&ni->count);
 	ntfs_error(ni->vol->sb, "Failed with error code %lu.", -PTR_ERR(m));
 	return m;
@@ -217,7 +218,7 @@ void unmap_mft_record(ntfs_inode *ni)
 	ntfs_debug("Entering for mft_no 0x%lx.", ni->mft_no);
 
 	unmap_mft_record_page(ni);
-	up(&ni->mrec_lock);
+	mutex_unlock(&ni->mrec_lock);
 	atomic_dec(&ni->count);
 	/*
 	 * If pure ntfs_inode, i.e. no vfs inode attached, we leave it to
@@ -261,7 +262,7 @@ MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
 	 * in which case just return it. If not found, add it to the base
 	 * inode before returning it.
 	 */
-	down(&base_ni->extent_lock);
+	mutex_lock(&base_ni->extent_lock);
 	if (base_ni->nr_extents > 0) {
 		extent_nis = base_ni->ext.extent_ntfs_inos;
 		for (i = 0; i < base_ni->nr_extents; i++) {
@@ -274,7 +275,7 @@ MFT_RECORD *map_extent_mft_record(ntfs_inode *base_ni, MFT_REF mref,
 		}
 	}
 	if (likely(ni != NULL)) {
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		atomic_dec(&base_ni->count);
 		/* We found the record; just have to map and return it. */
 		m = map_mft_record(ni);
@@ -301,7 +302,7 @@ map_err_out:
 	/* Record wasn't there. Get a new ntfs inode and initialize it. */
 	ni = ntfs_new_extent_inode(base_ni->vol->sb, mft_no);
 	if (unlikely(!ni)) {
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		atomic_dec(&base_ni->count);
 		return ERR_PTR(-ENOMEM);
 	}
@@ -312,7 +313,7 @@ map_err_out:
 	/* Now map the record. */
 	m = map_mft_record(ni);
 	if (IS_ERR(m)) {
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		atomic_dec(&base_ni->count);
 		ntfs_clear_extent_inode(ni);
 		goto map_err_out;
@@ -347,14 +348,14 @@ map_err_out:
 		base_ni->ext.extent_ntfs_inos = tmp;
 	}
 	base_ni->ext.extent_ntfs_inos[base_ni->nr_extents++] = ni;
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 	atomic_dec(&base_ni->count);
 	ntfs_debug("Done 2.");
 	*ntfs_ino = ni;
 	return m;
 unm_err_out:
 	unmap_mft_record(ni);
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 	atomic_dec(&base_ni->count);
 	/*
 	 * If the extent inode was not attached to the base inode we need to
@@ -399,12 +400,12 @@ void __mark_mft_record_dirty(ntfs_inode *ni)
 	BUG_ON(NInoAttr(ni));
 	mark_ntfs_record_dirty(ni->page, ni->page_ofs);
 	/* Determine the base vfs inode and mark it dirty, too. */
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	if (likely(ni->nr_extents >= 0))
 		base_ni = ni;
 	else
 		base_ni = ni->ext.base_ntfs_ino;
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 	__mark_inode_dirty(VFS_I(base_ni), I_DIRTY_SYNC | I_DIRTY_DATASYNC);
 }
 
@@ -650,10 +651,7 @@ err_out:
  * fs/ntfs/aops.c::mark_ntfs_record_dirty().
  *
  * On success, clean the mft record and return 0.  On error, leave the mft
- * record dirty and return -errno.  The caller should call make_bad_inode() on
- * the base inode to ensure no more access happens to this inode.  We do not do
- * it here as the caller may want to finish writing other extent mft records
- * first to minimize on-disk metadata inconsistencies.
+ * record dirty and return -errno.
  *
  * NOTE:  We always perform synchronous i/o and ignore the @sync parameter.
  * However, if the mft record has a counterpart in the mft mirror and @sync is
@@ -983,7 +981,7 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 		}
 		ntfs_debug("Inode 0x%lx is not dirty.", mft_no);
 		/* The inode is not dirty, try to take the mft record lock. */
-		if (unlikely(down_trylock(&ni->mrec_lock))) {
+		if (unlikely(!mutex_trylock(&ni->mrec_lock))) {
 			ntfs_debug("Mft record 0x%lx is already locked, do "
 					"not write it.", mft_no);
 			atomic_dec(&ni->count);
@@ -1043,13 +1041,13 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 	 * corresponding to this extent mft record attached.
 	 */
 	ni = NTFS_I(vi);
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	if (ni->nr_extents <= 0) {
 		/*
 		 * The base inode has no attached extent inodes, write this
 		 * extent mft record.
 		 */
-		up(&ni->extent_lock);
+		mutex_unlock(&ni->extent_lock);
 		iput(vi);
 		ntfs_debug("Base inode 0x%lx has no attached extent inodes, "
 				"write the extent record.", na.mft_no);
@@ -1072,7 +1070,7 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 	 * extent mft record.
 	 */
 	if (!eni) {
-		up(&ni->extent_lock);
+		mutex_unlock(&ni->extent_lock);
 		iput(vi);
 		ntfs_debug("Extent inode 0x%lx is not attached to its base "
 				"inode 0x%lx, write the extent record.",
@@ -1083,12 +1081,12 @@ BOOL ntfs_may_write_mft_record(ntfs_volume *vol, const unsigned long mft_no,
 			mft_no, na.mft_no);
 	/* Take a reference to the extent ntfs inode. */
 	atomic_inc(&eni->count);
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 	/*
 	 * Found the extent inode coresponding to this extent mft record.
 	 * Try to take the mft record lock.
 	 */
-	if (unlikely(down_trylock(&eni->mrec_lock))) {
+	if (unlikely(!mutex_trylock(&eni->mrec_lock))) {
 		atomic_dec(&eni->count);
 		iput(vi);
 		ntfs_debug("Extent mft record 0x%lx is already locked, do "
@@ -2672,7 +2670,7 @@ mft_rec_already_initialized:
 			ni->name_len = 4;
 
 			ni->itype.index.block_size = 4096;
-			ni->itype.index.block_size_bits = generic_ffs(4096) - 1;
+			ni->itype.index.block_size_bits = ntfs_ffs(4096) - 1;
 			ni->itype.index.collation_rule = COLLATION_FILE_NAME;
 			if (vol->cluster_size <= ni->itype.index.block_size) {
 				ni->itype.index.vcn_size = vol->cluster_size;
@@ -2711,7 +2709,7 @@ mft_rec_already_initialized:
 		 * have its page mapped and it is very easy to do.
 		 */
 		atomic_inc(&ni->count);
-		down(&ni->mrec_lock);
+		mutex_lock(&ni->mrec_lock);
 		ni->page = page;
 		ni->page_ofs = ofs;
 		/*
@@ -2798,22 +2796,22 @@ int ntfs_extent_mft_record_free(ntfs_inode *ni, MFT_RECORD *m)
 	BUG_ON(NInoAttr(ni));
 	BUG_ON(ni->nr_extents != -1);
 
-	down(&ni->extent_lock);
+	mutex_lock(&ni->extent_lock);
 	base_ni = ni->ext.base_ntfs_ino;
-	up(&ni->extent_lock);
+	mutex_unlock(&ni->extent_lock);
 
 	BUG_ON(base_ni->nr_extents <= 0);
 
 	ntfs_debug("Entering for extent inode 0x%lx, base inode 0x%lx.\n",
 			mft_no, base_ni->mft_no);
 
-	down(&base_ni->extent_lock);
+	mutex_lock(&base_ni->extent_lock);
 
 	/* Make sure we are holding the only reference to the extent inode. */
 	if (atomic_read(&ni->count) > 2) {
 		ntfs_error(vol->sb, "Tried to free busy extent inode 0x%lx, "
 				"not freeing.", base_ni->mft_no);
-		up(&base_ni->extent_lock);
+		mutex_unlock(&base_ni->extent_lock);
 		return -EBUSY;
 	}
 
@@ -2831,7 +2829,7 @@ int ntfs_extent_mft_record_free(ntfs_inode *ni, MFT_RECORD *m)
 		break;
 	}
 
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 
 	if (unlikely(err)) {
 		ntfs_error(vol->sb, "Extent inode 0x%lx is not attached to "
@@ -2890,7 +2888,7 @@ rollback_error:
 	return 0;
 rollback:
 	/* Rollback what we did... */
-	down(&base_ni->extent_lock);
+	mutex_lock(&base_ni->extent_lock);
 	extent_nis = base_ni->ext.extent_ntfs_inos;
 	if (!(base_ni->nr_extents & 3)) {
 		int new_size = (base_ni->nr_extents + 4) * sizeof(ntfs_inode*);
@@ -2899,7 +2897,7 @@ rollback:
 		if (unlikely(!extent_nis)) {
 			ntfs_error(vol->sb, "Failed to allocate internal "
 					"buffer during rollback.%s", es);
-			up(&base_ni->extent_lock);
+			mutex_unlock(&base_ni->extent_lock);
 			NVolSetErrors(vol);
 			goto rollback_error;
 		}
@@ -2914,7 +2912,7 @@ rollback:
 	m->flags |= MFT_RECORD_IN_USE;
 	m->sequence_number = old_seq_no;
 	extent_nis[base_ni->nr_extents++] = ni;
-	up(&base_ni->extent_lock);
+	mutex_unlock(&base_ni->extent_lock);
 	mark_mft_record_dirty(ni);
 	return err;
 }
diff --git a/fs/ntfs/mft.h b/fs/ntfs/mft.h
index 407de2cef1d6..639cd1bab08b 100644
--- a/fs/ntfs/mft.h
+++ b/fs/ntfs/mft.h
@@ -97,10 +97,7 @@ extern int write_mft_record_nolock(ntfs_inode *ni, MFT_RECORD *m, int sync);
  * uptodate.
  *
  * On success, clean the mft record and return 0.  On error, leave the mft
- * record dirty and return -errno.  The caller should call make_bad_inode() on
- * the base inode to ensure no more access happens to this inode.  We do not do
- * it here as the caller may want to finish writing other extent mft records
- * first to minimize on-disk metadata inconsistencies.
+ * record dirty and return -errno.
  */
 static inline int write_mft_record(ntfs_inode *ni, MFT_RECORD *m, int sync)
 {
diff --git a/fs/ntfs/namei.c b/fs/ntfs/namei.c
index 5ea9eb93af62..eddb2247cec5 100644
--- a/fs/ntfs/namei.c
+++ b/fs/ntfs/namei.c
@@ -2,7 +2,7 @@
  * namei.c - NTFS kernel directory inode operations. Part of the Linux-NTFS
  *	     project.
  *
- * Copyright (c) 2001-2004 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  *
  * This program/include file is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License as published
@@ -115,7 +115,9 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 	uname_len = ntfs_nlstoucs(vol, dent->d_name.name, dent->d_name.len,
 			&uname);
 	if (uname_len < 0) {
-		ntfs_error(vol->sb, "Failed to convert name to Unicode.");
+		if (uname_len != -ENAMETOOLONG)
+			ntfs_error(vol->sb, "Failed to convert name to "
+					"Unicode.");
 		return ERR_PTR(uname_len);
 	}
 	mref = ntfs_lookup_inode_by_name(NTFS_I(dir_ino), uname, uname_len,
@@ -157,7 +159,7 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 		/* Return the error code. */
 		return (struct dentry *)dent_inode;
 	}
-	/* It is guaranteed that name is no longer allocated at this point. */
+	/* It is guaranteed that @name is no longer allocated at this point. */
 	if (MREF_ERR(mref) == -ENOENT) {
 		ntfs_debug("Entry was not found, adding negative dentry.");
 		/* The dcache will handle negative entries. */
@@ -168,7 +170,6 @@ static struct dentry *ntfs_lookup(struct inode *dir_ino, struct dentry *dent,
 	ntfs_error(vol->sb, "ntfs_lookup_ino_by_name() failed with error "
 			"code %i.", -MREF_ERR(mref));
 	return ERR_PTR(MREF_ERR(mref));
-
 	// TODO: Consider moving this lot to a separate function! (AIA)
 handle_name:
    {
diff --git a/fs/ntfs/ntfs.h b/fs/ntfs/ntfs.h
index 653d2a5c4899..bf7b3d7c0930 100644
--- a/fs/ntfs/ntfs.h
+++ b/fs/ntfs/ntfs.h
@@ -60,13 +60,13 @@ extern struct kmem_cache *ntfs_index_ctx_cache;
 extern struct address_space_operations ntfs_aops;
 extern struct address_space_operations ntfs_mst_aops;
 
-extern struct  file_operations ntfs_file_ops;
+extern const struct  file_operations ntfs_file_ops;
 extern struct inode_operations ntfs_file_inode_ops;
 
-extern struct  file_operations ntfs_dir_ops;
+extern const struct  file_operations ntfs_dir_ops;
 extern struct inode_operations ntfs_dir_inode_ops;
 
-extern struct  file_operations ntfs_empty_file_ops;
+extern const struct  file_operations ntfs_empty_file_ops;
 extern struct inode_operations ntfs_empty_inode_ops;
 
 extern struct export_operations ntfs_export_ops;
@@ -91,7 +91,7 @@ extern void free_compression_buffers(void);
 
 /* From fs/ntfs/super.c */
 #define default_upcase_len 0x10000
-extern struct semaphore ntfs_lock;
+extern struct mutex ntfs_lock;
 
 typedef struct {
 	int val;
@@ -132,4 +132,33 @@ extern int ntfs_ucstonls(const ntfs_volume *vol, const ntfschar *ins,
 /* From fs/ntfs/upcase.c */
 extern ntfschar *generate_default_upcase(void);
 
+static inline int ntfs_ffs(int x)
+{
+	int r = 1;
+
+	if (!x)
+		return 0;
+	if (!(x & 0xffff)) {
+		x >>= 16;
+		r += 16;
+	}
+	if (!(x & 0xff)) {
+		x >>= 8;
+		r += 8;
+	}
+	if (!(x & 0xf)) {
+		x >>= 4;
+		r += 4;
+	}
+	if (!(x & 3)) {
+		x >>= 2;
+		r += 2;
+	}
+	if (!(x & 1)) {
+		x >>= 1;
+		r += 1;
+	}
+	return r;
+}
+
 #endif /* _LINUX_NTFS_H */
diff --git a/fs/ntfs/runlist.c b/fs/ntfs/runlist.c
index 061b5ff6b73c..eb52b801512b 100644
--- a/fs/ntfs/runlist.c
+++ b/fs/ntfs/runlist.c
@@ -381,6 +381,7 @@ static inline runlist_element *ntfs_rl_insert(runlist_element *dst,
 static inline runlist_element *ntfs_rl_replace(runlist_element *dst,
 		int dsize, runlist_element *src, int ssize, int loc)
 {
+	signed delta;
 	BOOL left = FALSE;	/* Left end of @src needs merging. */
 	BOOL right = FALSE;	/* Right end of @src needs merging. */
 	int tail;		/* Start of tail of @dst. */
@@ -396,11 +397,14 @@ static inline runlist_element *ntfs_rl_replace(runlist_element *dst,
 		left = ntfs_are_rl_mergeable(dst + loc - 1, src);
 	/*
 	 * Allocate some space.  We will need less if the left, right, or both
-	 * ends get merged.
+	 * ends get merged.  The -1 accounts for the run being replaced.
 	 */
-	dst = ntfs_rl_realloc(dst, dsize, dsize + ssize - left - right);
-	if (IS_ERR(dst))
-		return dst;
+	delta = ssize - 1 - left - right;
+	if (delta > 0) {
+		dst = ntfs_rl_realloc(dst, dsize, dsize + delta);
+		if (IS_ERR(dst))
+			return dst;
+	}
 	/*
 	 * We are guaranteed to succeed from here so can start modifying the
 	 * original runlists.
diff --git a/fs/ntfs/super.c b/fs/ntfs/super.c
index 368a8ec10668..27833f6df49f 100644
--- a/fs/ntfs/super.c
+++ b/fs/ntfs/super.c
@@ -1099,26 +1099,38 @@ static BOOL check_mft_mirror(ntfs_volume *vol)
 			kmirr = page_address(mirr_page);
 			++index;
 		}
-		/* Make sure the record is ok. */
-		if (ntfs_is_baad_recordp((le32*)kmft)) {
-			ntfs_error(sb, "Incomplete multi sector transfer "
-					"detected in mft record %i.", i);
+		/* Do not check the record if it is not in use. */
+		if (((MFT_RECORD*)kmft)->flags & MFT_RECORD_IN_USE) {
+			/* Make sure the record is ok. */
+			if (ntfs_is_baad_recordp((le32*)kmft)) {
+				ntfs_error(sb, "Incomplete multi sector "
+						"transfer detected in mft "
+						"record %i.", i);
 mm_unmap_out:
-			ntfs_unmap_page(mirr_page);
+				ntfs_unmap_page(mirr_page);
 mft_unmap_out:
-			ntfs_unmap_page(mft_page);
-			return FALSE;
+				ntfs_unmap_page(mft_page);
+				return FALSE;
+			}
 		}
-		if (ntfs_is_baad_recordp((le32*)kmirr)) {
-			ntfs_error(sb, "Incomplete multi sector transfer "
-					"detected in mft mirror record %i.", i);
-			goto mm_unmap_out;
+		/* Do not check the mirror record if it is not in use. */
+		if (((MFT_RECORD*)kmirr)->flags & MFT_RECORD_IN_USE) {
+			if (ntfs_is_baad_recordp((le32*)kmirr)) {
+				ntfs_error(sb, "Incomplete multi sector "
+						"transfer detected in mft "
+						"mirror record %i.", i);
+				goto mm_unmap_out;
+			}
 		}
 		/* Get the amount of data in the current record. */
 		bytes = le32_to_cpu(((MFT_RECORD*)kmft)->bytes_in_use);
-		if (!bytes || bytes > vol->mft_record_size) {
+		if (bytes < sizeof(MFT_RECORD_OLD) ||
+				bytes > vol->mft_record_size ||
+				ntfs_is_baad_recordp((le32*)kmft)) {
 			bytes = le32_to_cpu(((MFT_RECORD*)kmirr)->bytes_in_use);
-			if (!bytes || bytes > vol->mft_record_size)
+			if (bytes < sizeof(MFT_RECORD_OLD) ||
+					bytes > vol->mft_record_size ||
+					ntfs_is_baad_recordp((le32*)kmirr))
 				bytes = vol->mft_record_size;
 		}
 		/* Compare the two records. */
@@ -1665,11 +1677,11 @@ read_partial_upcase_page:
 	ntfs_debug("Read %llu bytes from $UpCase (expected %zu bytes).",
 			i_size, 64 * 1024 * sizeof(ntfschar));
 	iput(ino);
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (!default_upcase) {
 		ntfs_debug("Using volume specified $UpCase since default is "
 				"not present.");
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		return TRUE;
 	}
 	max = default_upcase_len;
@@ -1683,12 +1695,12 @@ read_partial_upcase_page:
 		vol->upcase = default_upcase;
 		vol->upcase_len = max;
 		ntfs_nr_upcase_users++;
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		ntfs_debug("Volume specified $UpCase matches default. Using "
 				"default.");
 		return TRUE;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	ntfs_debug("Using volume specified $UpCase since it does not match "
 			"the default.");
 	return TRUE;
@@ -1697,17 +1709,17 @@ iput_upcase_failed:
 	ntfs_free(vol->upcase);
 	vol->upcase = NULL;
 upcase_failed:
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (default_upcase) {
 		vol->upcase = default_upcase;
 		vol->upcase_len = default_upcase_len;
 		ntfs_nr_upcase_users++;
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		ntfs_error(sb, "Failed to load $UpCase from the volume. Using "
 				"default.");
 		return TRUE;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	ntfs_error(sb, "Failed to initialize upcase table.");
 	return FALSE;
 }
@@ -2183,12 +2195,12 @@ iput_attrdef_err_out:
 iput_upcase_err_out:
 #endif /* NTFS_RW */
 	vol->upcase_len = 0;
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (vol->upcase == default_upcase) {
 		ntfs_nr_upcase_users--;
 		vol->upcase = NULL;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	if (vol->upcase) {
 		ntfs_free(vol->upcase);
 		vol->upcase = NULL;
@@ -2393,7 +2405,7 @@ static void ntfs_put_super(struct super_block *sb)
 	 * Destroy the global default upcase table if necessary.  Also decrease
 	 * the number of upcase users if we are a user.
 	 */
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (vol->upcase == default_upcase) {
 		ntfs_nr_upcase_users--;
 		vol->upcase = NULL;
@@ -2404,7 +2416,7 @@ static void ntfs_put_super(struct super_block *sb)
 	}
 	if (vol->cluster_size <= 4096 && !--ntfs_nr_compression_users)
 		free_compression_buffers();
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	if (vol->upcase) {
 		ntfs_free(vol->upcase);
 		vol->upcase = NULL;
@@ -2878,7 +2890,7 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 			ntfs_error(sb, "Failed to load essential metadata.");
 		goto iput_tmp_ino_err_out_now;
 	}
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	/*
 	 * The current mount is a compression user if the cluster size is
 	 * less than or equal 4kiB.
@@ -2889,7 +2901,7 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 			ntfs_error(NULL, "Failed to allocate buffers "
 					"for compression engine.");
 			ntfs_nr_compression_users--;
-			up(&ntfs_lock);
+			mutex_unlock(&ntfs_lock);
 			goto iput_tmp_ino_err_out_now;
 		}
 	}
@@ -2901,7 +2913,7 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 	if (!default_upcase)
 		default_upcase = generate_default_upcase();
 	ntfs_nr_upcase_users++;
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	/*
 	 * From now on, ignore @silent parameter. If we fail below this line,
 	 * it will be due to a corrupt fs or a system error, so we report it.
@@ -2919,12 +2931,12 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 		atomic_inc(&vol->root_ino->i_count);
 		ntfs_debug("Exiting, status successful.");
 		/* Release the default upcase if it has no users. */
-		down(&ntfs_lock);
+		mutex_lock(&ntfs_lock);
 		if (!--ntfs_nr_upcase_users && default_upcase) {
 			ntfs_free(default_upcase);
 			default_upcase = NULL;
 		}
-		up(&ntfs_lock);
+		mutex_unlock(&ntfs_lock);
 		sb->s_export_op = &ntfs_export_ops;
 		lock_kernel();
 		return 0;
@@ -2992,12 +3004,12 @@ static int ntfs_fill_super(struct super_block *sb, void *opt, const int silent)
 		vol->attrdef = NULL;
 	}
 	vol->upcase_len = 0;
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (vol->upcase == default_upcase) {
 		ntfs_nr_upcase_users--;
 		vol->upcase = NULL;
 	}
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 	if (vol->upcase) {
 		ntfs_free(vol->upcase);
 		vol->upcase = NULL;
@@ -3012,14 +3024,14 @@ unl_upcase_iput_tmp_ino_err_out_now:
 	 * Decrease the number of upcase users and destroy the global default
 	 * upcase table if necessary.
 	 */
-	down(&ntfs_lock);
+	mutex_lock(&ntfs_lock);
 	if (!--ntfs_nr_upcase_users && default_upcase) {
 		ntfs_free(default_upcase);
 		default_upcase = NULL;
 	}
 	if (vol->cluster_size <= 4096 && !--ntfs_nr_compression_users)
 		free_compression_buffers();
-	up(&ntfs_lock);
+	mutex_unlock(&ntfs_lock);
 iput_tmp_ino_err_out_now:
 	iput(tmp_ino);
 	if (vol->mft_ino && vol->mft_ino != tmp_ino)
@@ -3078,8 +3090,8 @@ static void ntfs_big_inode_init_once(void *foo, struct kmem_cache *cachep,
 struct kmem_cache *ntfs_attr_ctx_cache;
 struct kmem_cache *ntfs_index_ctx_cache;
 
-/* Driver wide semaphore. */
-DECLARE_MUTEX(ntfs_lock);
+/* Driver wide mutex. */
+DEFINE_MUTEX(ntfs_lock);
 
 static struct super_block *ntfs_get_sb(struct file_system_type *fs_type,
 	int flags, const char *dev_name, void *data)
@@ -3151,7 +3163,7 @@ static int __init init_ntfs_fs(void)
 
 	ntfs_inode_cache = kmem_cache_create(ntfs_inode_cache_name,
 			sizeof(ntfs_inode), 0,
-			SLAB_RECLAIM_ACCOUNT, NULL, NULL);
+			SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD, NULL, NULL);
 	if (!ntfs_inode_cache) {
 		printk(KERN_CRIT "NTFS: Failed to create %s!\n",
 				ntfs_inode_cache_name);
@@ -3160,7 +3172,7 @@ static int __init init_ntfs_fs(void)
 
 	ntfs_big_inode_cache = kmem_cache_create(ntfs_big_inode_cache_name,
 			sizeof(big_ntfs_inode), 0,
-			SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT,
+			SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
 			ntfs_big_inode_init_once, NULL);
 	if (!ntfs_big_inode_cache) {
 		printk(KERN_CRIT "NTFS: Failed to create %s!\n",
@@ -3234,7 +3246,7 @@ static void __exit exit_ntfs_fs(void)
 }
 
 MODULE_AUTHOR("Anton Altaparmakov <aia21@cantab.net>");
-MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2005 Anton Altaparmakov");
+MODULE_DESCRIPTION("NTFS 1.2/3.x driver - Copyright (c) 2001-2006 Anton Altaparmakov");
 MODULE_VERSION(NTFS_VERSION);
 MODULE_LICENSE("GPL");
 #ifdef DEBUG
diff --git a/fs/ntfs/unistr.c b/fs/ntfs/unistr.c
index 0ea887fc859c..b123c0fa6bf6 100644
--- a/fs/ntfs/unistr.c
+++ b/fs/ntfs/unistr.c
@@ -1,7 +1,7 @@
 /*
  * unistr.c - NTFS Unicode string handling. Part of the Linux-NTFS project.
  *
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
  *
  * This program/include file is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public License as published
@@ -19,6 +19,8 @@
  * Foundation,Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
+#include <linux/slab.h>
+
 #include "types.h"
 #include "debug.h"
 #include "ntfs.h"
@@ -242,7 +244,7 @@ int ntfs_file_compare_values(FILE_NAME_ATTR *file_name_attr1,
  * map dictates, into a little endian, 2-byte Unicode string.
  *
  * This function allocates the string and the caller is responsible for
- * calling kmem_cache_free(ntfs_name_cache, @outs); when finished with it.
+ * calling kmem_cache_free(ntfs_name_cache, *@outs); when finished with it.
  *
  * On success the function returns the number of Unicode characters written to
  * the output string *@outs (>= 0), not counting the terminating Unicode NULL
@@ -262,37 +264,48 @@ int ntfs_nlstoucs(const ntfs_volume *vol, const char *ins,
 	wchar_t wc;
 	int i, o, wc_len;
 
-	/* We don't trust outside sources. */
-	if (ins) {
+	/* We do not trust outside sources. */
+	if (likely(ins)) {
 		ucs = kmem_cache_alloc(ntfs_name_cache, SLAB_NOFS);
-		if (ucs) {
+		if (likely(ucs)) {
 			for (i = o = 0; i < ins_len; i += wc_len) {
 				wc_len = nls->char2uni(ins + i, ins_len - i,
 						&wc);
-				if (wc_len >= 0) {
-					if (wc) {
+				if (likely(wc_len >= 0 &&
+						o < NTFS_MAX_NAME_LEN)) {
+					if (likely(wc)) {
 						ucs[o++] = cpu_to_le16(wc);
 						continue;
-					} /* else (!wc) */
+					} /* else if (!wc) */
 					break;
-				} /* else (wc_len < 0) */
-				goto conversion_err;
+				} /* else if (wc_len < 0 ||
+						o >= NTFS_MAX_NAME_LEN) */
+				goto name_err;
 			}
 			ucs[o] = 0;
 			*outs = ucs;
 			return o;
-		} /* else (!ucs) */
-		ntfs_error(vol->sb, "Failed to allocate name from "
-				"ntfs_name_cache!");
+		} /* else if (!ucs) */
+		ntfs_error(vol->sb, "Failed to allocate buffer for converted "
+				"name from ntfs_name_cache.");
 		return -ENOMEM;
-	} /* else (!ins) */
-	ntfs_error(NULL, "Received NULL pointer.");
+	} /* else if (!ins) */
+	ntfs_error(vol->sb, "Received NULL pointer.");
 	return -EINVAL;
-conversion_err:
-	ntfs_error(vol->sb, "Name using character set %s contains characters "
-			"that cannot be converted to Unicode.", nls->charset);
+name_err:
 	kmem_cache_free(ntfs_name_cache, ucs);
-	return -EILSEQ;
+	if (wc_len < 0) {
+		ntfs_error(vol->sb, "Name using character set %s contains "
+				"characters that cannot be converted to "
+				"Unicode.", nls->charset);
+		i = -EILSEQ;
+	} else /* if (o >= NTFS_MAX_NAME_LEN) */ {
+		ntfs_error(vol->sb, "Name is too long (maximum length for a "
+				"name on NTFS is %d Unicode characters.",
+				NTFS_MAX_NAME_LEN);
+		i = -ENAMETOOLONG;
+	}
+	return i;
 }
 
 /**
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 6b9812db3779..edaab05a93e0 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -566,9 +566,8 @@ static int ocfs2_do_insert_extent(struct ocfs2_super *osb,
 			next_free = le16_to_cpu(el->l_next_free_rec);
 			if (next_free == 0) {
 				ocfs2_error(inode->i_sb,
-					    "Dinode %"MLFu64" has a bad "
-					    "extent list",
-					    OCFS2_I(inode)->ip_blkno);
+					    "Dinode %llu has a bad extent list",
+					    (unsigned long long)OCFS2_I(inode)->ip_blkno);
 				status = -EIO;
 				goto bail;
 			}
@@ -611,9 +610,8 @@ static int ocfs2_do_insert_extent(struct ocfs2_super *osb,
 		next_free = le16_to_cpu(el->l_next_free_rec);
 		if (next_free == 0) {
 			ocfs2_error(inode->i_sb,
-				    "Dinode %"MLFu64" has a bad "
-				    "extent list",
-				    OCFS2_I(inode)->ip_blkno);
+				    "Dinode %llu has a bad extent list",
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			status = -EIO;
 			goto bail;
 		}
@@ -652,8 +650,9 @@ static int ocfs2_do_insert_extent(struct ocfs2_super *osb,
 		/* having an empty extent at eof is legal. */
 		if (el->l_recs[i].e_cpos != fe->i_clusters) {
 			ocfs2_error(inode->i_sb,
-				    "Dinode %"MLFu64" trailing extent is bad: "
+				    "Dinode %llu trailing extent is bad: "
 				    "cpos (%u) != number of clusters (%u)",
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno,
 				    le32_to_cpu(el->l_recs[i].e_cpos),
 				    le32_to_cpu(fe->i_clusters));
 			status = -EIO;
@@ -747,19 +746,19 @@ static int ocfs2_find_branch_target(struct ocfs2_super *osb,
 
 	while(le16_to_cpu(el->l_tree_depth) > 1) {
 		if (le16_to_cpu(el->l_next_free_rec) == 0) {
-			ocfs2_error(inode->i_sb, "Dinode %"MLFu64" has empty "
+			ocfs2_error(inode->i_sb, "Dinode %llu has empty "
 				    "extent list (next_free_rec == 0)",
-				    OCFS2_I(inode)->ip_blkno);
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			status = -EIO;
 			goto bail;
 		}
 		i = le16_to_cpu(el->l_next_free_rec) - 1;
 		blkno = le64_to_cpu(el->l_recs[i].e_blkno);
 		if (!blkno) {
-			ocfs2_error(inode->i_sb, "Dinode %"MLFu64" has extent "
+			ocfs2_error(inode->i_sb, "Dinode %llu has extent "
 				    "list where extent # %d has no physical "
 				    "block start",
-				    OCFS2_I(inode)->ip_blkno, i);
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno, i);
 			status = -EIO;
 			goto bail;
 		}
@@ -826,9 +825,9 @@ int ocfs2_insert_extent(struct ocfs2_super *osb,
 
 	mlog_entry_void();
 
-	mlog(0, "add %u clusters starting at block %"MLFu64" to "
-		"inode %"MLFu64"\n",
-	     new_clusters, start_blk, OCFS2_I(inode)->ip_blkno);
+	mlog(0, "add %u clusters starting at block %llu to inode %llu\n",
+	     new_clusters, (unsigned long long)start_blk,
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	fe = (struct ocfs2_dinode *) fe_bh->b_data;
 	el = &fe->id2.i_list;
@@ -963,8 +962,8 @@ static int ocfs2_truncate_log_append(struct ocfs2_super *osb,
 	struct ocfs2_dinode *di;
 	struct ocfs2_truncate_log *tl;
 
-	mlog_entry("start_blk = %"MLFu64", num_clusters = %u\n", start_blk,
-		   num_clusters);
+	mlog_entry("start_blk = %llu, num_clusters = %u\n",
+		   (unsigned long long)start_blk, num_clusters);
 
 	BUG_ON(mutex_trylock(&tl_inode->i_mutex));
 
@@ -981,8 +980,9 @@ static int ocfs2_truncate_log_append(struct ocfs2_super *osb,
 	tl_count = le16_to_cpu(tl->tl_count);
 	mlog_bug_on_msg(tl_count > ocfs2_truncate_recs_per_inode(osb->sb) ||
 			tl_count == 0,
-			"Truncate record count on #%"MLFu64" invalid ("
-			"wanted %u, actual %u\n", OCFS2_I(tl_inode)->ip_blkno,
+			"Truncate record count on #%llu invalid "
+			"wanted %u, actual %u\n",
+			(unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
 			ocfs2_truncate_recs_per_inode(osb->sb),
 			le16_to_cpu(tl->tl_count));
 
@@ -1002,8 +1002,8 @@ static int ocfs2_truncate_log_append(struct ocfs2_super *osb,
 	}
 
 	mlog(0, "Log truncate of %u clusters starting at cluster %u to "
-	     "%"MLFu64" (index = %d)\n", num_clusters, start_cluster,
-	     OCFS2_I(tl_inode)->ip_blkno, index);
+	     "%llu (index = %d)\n", num_clusters, start_cluster,
+	     (unsigned long long)OCFS2_I(tl_inode)->ip_blkno, index);
 
 	if (ocfs2_truncate_log_can_coalesce(tl, start_cluster)) {
 		/*
@@ -1134,8 +1134,8 @@ static int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
 	}
 
 	num_to_flush = le16_to_cpu(tl->tl_used);
-	mlog(0, "Flush %u records from truncate log #%"MLFu64"\n",
-	     num_to_flush, OCFS2_I(tl_inode)->ip_blkno);
+	mlog(0, "Flush %u records from truncate log #%llu\n",
+	     num_to_flush, (unsigned long long)OCFS2_I(tl_inode)->ip_blkno);
 	if (!num_to_flush) {
 		status = 0;
 		goto bail;
@@ -1360,8 +1360,8 @@ int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
 
 	tl = &tl_copy->id2.i_dealloc;
 	num_recs = le16_to_cpu(tl->tl_used);
-	mlog(0, "cleanup %u records from %"MLFu64"\n", num_recs,
-	     tl_copy->i_blkno);
+	mlog(0, "cleanup %u records from %llu\n", num_recs,
+	     (unsigned long long)tl_copy->i_blkno);
 
 	mutex_lock(&tl_inode->i_mutex);
 	for(i = 0; i < num_recs; i++) {
@@ -1529,7 +1529,8 @@ static int ocfs2_find_new_last_ext_blk(struct ocfs2_super *osb,
 
 	*new_last_eb = bh;
 	get_bh(*new_last_eb);
-	mlog(0, "returning block %"MLFu64"\n", le64_to_cpu(eb->h_blkno));
+	mlog(0, "returning block %llu\n",
+	     (unsigned long long)le64_to_cpu(eb->h_blkno));
 bail:
 	if (bh)
 		brelse(bh);
@@ -1646,8 +1647,8 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 
 	/* if our tree depth > 0, update all the tree blocks below us. */
 	while (depth) {
-		mlog(0, "traveling tree (depth = %d, next_eb = %"MLFu64")\n",
-		     depth,  next_eb);
+		mlog(0, "traveling tree (depth = %d, next_eb = %llu)\n",
+		     depth,  (unsigned long long)next_eb);
 		status = ocfs2_read_block(osb, next_eb, &eb_bh,
 					  OCFS2_BH_CACHED, inode);
 		if (status < 0) {
@@ -1674,12 +1675,12 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 
 		i = le16_to_cpu(el->l_next_free_rec) - 1;
 
-		mlog(0, "extent block %"MLFu64", before: record %d: "
-		     "(%u, %u, %"MLFu64"), next = %u\n",
-		     le64_to_cpu(eb->h_blkno), i,
+		mlog(0, "extent block %llu, before: record %d: "
+		     "(%u, %u, %llu), next = %u\n",
+		     (unsigned long long)le64_to_cpu(eb->h_blkno), i,
 		     le32_to_cpu(el->l_recs[i].e_cpos),
 		     le32_to_cpu(el->l_recs[i].e_clusters),
-		     le64_to_cpu(el->l_recs[i].e_blkno),
+		     (unsigned long long)le64_to_cpu(el->l_recs[i].e_blkno),
 		     le16_to_cpu(el->l_next_free_rec));
 
 		BUG_ON(le32_to_cpu(el->l_recs[i].e_clusters) < clusters_to_del);
@@ -1697,12 +1698,12 @@ static int ocfs2_do_truncate(struct ocfs2_super *osb,
 			BUG_ON(!el->l_next_free_rec);
 			le16_add_cpu(&el->l_next_free_rec, -1);
 		}
-		mlog(0, "extent block %"MLFu64", after: record %d: "
-		     "(%u, %u, %"MLFu64"), next = %u\n",
-		     le64_to_cpu(eb->h_blkno), i,
+		mlog(0, "extent block %llu, after: record %d: "
+		     "(%u, %u, %llu), next = %u\n",
+		     (unsigned long long)le64_to_cpu(eb->h_blkno), i,
 		     le32_to_cpu(el->l_recs[i].e_cpos),
 		     le32_to_cpu(el->l_recs[i].e_clusters),
-		     le64_to_cpu(el->l_recs[i].e_blkno),
+		     (unsigned long long)le64_to_cpu(el->l_recs[i].e_blkno),
 		     le16_to_cpu(el->l_next_free_rec));
 
 		status = ocfs2_journal_dirty(handle, eb_bh);
@@ -1792,10 +1793,10 @@ int ocfs2_commit_truncate(struct ocfs2_super *osb,
 	last_eb = le64_to_cpu(fe->i_last_eb_blk);
 start:
 	mlog(0, "ocfs2_commit_truncate: fe->i_clusters = %u, "
-	     "last_eb = %"MLFu64", fe->i_last_eb_blk = %"MLFu64", "
+	     "last_eb = %llu, fe->i_last_eb_blk = %llu, "
 	     "fe->id2.i_list.l_tree_depth = %u last_eb_bh = %p\n",
-	     le32_to_cpu(fe->i_clusters), last_eb,
-	     le64_to_cpu(fe->i_last_eb_blk),
+	     le32_to_cpu(fe->i_clusters), (unsigned long long)last_eb,
+	     (unsigned long long)le64_to_cpu(fe->i_last_eb_blk),
 	     le16_to_cpu(fe->id2.i_list.l_tree_depth), last_eb_bh);
 
 	if (last_eb != le64_to_cpu(fe->i_last_eb_blk)) {
@@ -1934,16 +1935,17 @@ int ocfs2_prepare_truncate(struct ocfs2_super *osb,
 	fe = (struct ocfs2_dinode *) fe_bh->b_data;
 
 	mlog(0, "fe->i_clusters = %u, new_i_clusters = %u, fe->i_size ="
-	     "%"MLFu64"\n", fe->i_clusters, new_i_clusters, fe->i_size);
+	     "%llu\n", fe->i_clusters, new_i_clusters,
+	     (unsigned long long)fe->i_size);
 
 	if (le32_to_cpu(fe->i_clusters) <= new_i_clusters) {
-		ocfs2_error(inode->i_sb, "Dinode %"MLFu64" has cluster count "
-			    "%u and size %"MLFu64" whereas struct inode has "
+		ocfs2_error(inode->i_sb, "Dinode %llu has cluster count "
+			    "%u and size %llu whereas struct inode has "
 			    "cluster count %u and size %llu which caused an "
 			    "invalid truncate to %u clusters.",
-			    le64_to_cpu(fe->i_blkno),
+			    (unsigned long long)le64_to_cpu(fe->i_blkno),
 			    le32_to_cpu(fe->i_clusters),
-			    le64_to_cpu(fe->i_size),
+			    (unsigned long long)le64_to_cpu(fe->i_size),
 			    OCFS2_I(inode)->ip_clusters, i_size_read(inode),
 			    new_i_clusters);
 		mlog_meta_lvb(ML_ERROR, &OCFS2_I(inode)->ip_meta_lockres);
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 8f4467a930a5..0d858d0b25be 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -74,8 +74,8 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
 	fe = (struct ocfs2_dinode *) bh->b_data;
 
 	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		mlog(ML_ERROR, "Invalid dinode #%"MLFu64": signature = %.*s\n",
-		     fe->i_blkno, 7, fe->i_signature);
+		mlog(ML_ERROR, "Invalid dinode #%llu: signature = %.*s\n",
+		     (unsigned long long)fe->i_blkno, 7, fe->i_signature);
 		goto bail;
 	}
 
@@ -162,8 +162,8 @@ static int ocfs2_get_block(struct inode *inode, sector_t iblock,
 					  NULL);
 	if (err) {
 		mlog(ML_ERROR, "Error %d from get_blocks(0x%p, %llu, 1, "
-		     "%"MLFu64", NULL)\n", err, inode,
-		     (unsigned long long)iblock, p_blkno);
+		     "%llu, NULL)\n", err, inode, (unsigned long long)iblock,
+		     (unsigned long long)p_blkno);
 		goto bail;
 	}
 
@@ -171,13 +171,15 @@ static int ocfs2_get_block(struct inode *inode, sector_t iblock,
 
 	if (bh_result->b_blocknr == 0) {
 		err = -EIO;
-		mlog(ML_ERROR, "iblock = %llu p_blkno = %"MLFu64" "
-		     "blkno=(%"MLFu64")\n", (unsigned long long)iblock,
-		     p_blkno, OCFS2_I(inode)->ip_blkno);
+		mlog(ML_ERROR, "iblock = %llu p_blkno = %llu blkno=(%llu)\n",
+		     (unsigned long long)iblock,
+		     (unsigned long long)p_blkno,
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 	}
 
 	past_eof = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
-	mlog(0, "Inode %lu, past_eof = %"MLFu64"\n", inode->i_ino, past_eof);
+	mlog(0, "Inode %lu, past_eof = %llu\n", inode->i_ino,
+	     (unsigned long long)past_eof);
 
 	if (create && (iblock >= past_eof))
 		set_buffer_new(bh_result);
@@ -538,7 +540,6 @@ bail:
  * 					fs_count, map_bh, dio->rw == WRITE);
  */
 static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
-				     unsigned long max_blocks,
 				     struct buffer_head *bh_result, int create)
 {
 	int ret;
@@ -546,6 +547,7 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
 	u64 p_blkno;
 	int contig_blocks;
 	unsigned char blocksize_bits;
+	unsigned long max_blocks = bh_result->b_size >> inode->i_blkbits;
 
 	if (!inode || !bh_result) {
 		mlog(ML_ERROR, "inode or bh_result is null\n");
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index bae3d7548bea..9a24adf9be6e 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -97,8 +97,8 @@ int ocfs2_read_blocks(struct ocfs2_super *osb, u64 block, int nr,
 	int i, ignore_cache = 0;
 	struct buffer_head *bh;
 
-	mlog_entry("(block=(%"MLFu64"), nr=(%d), flags=%d, inode=%p)\n",
-		   block, nr, flags, inode);
+	mlog_entry("(block=(%llu), nr=(%d), flags=%d, inode=%p)\n",
+		   (unsigned long long)block, nr, flags, inode);
 
 	if (osb == NULL || osb->sb == NULL || bhs == NULL) {
 		status = -EINVAL;
@@ -143,9 +143,9 @@ int ocfs2_read_blocks(struct ocfs2_super *osb, u64 block, int nr,
 		if (flags & OCFS2_BH_CACHED &&
 		    !ocfs2_buffer_uptodate(inode, bh)) {
 			mlog(ML_UPTODATE,
-			     "bh (%llu), inode %"MLFu64" not uptodate\n",
+			     "bh (%llu), inode %llu not uptodate\n",
 			     (unsigned long long)bh->b_blocknr,
-			     OCFS2_I(inode)->ip_blkno);
+			     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			ignore_cache = 1;
 		}
 
@@ -222,7 +222,8 @@ int ocfs2_read_blocks(struct ocfs2_super *osb, u64 block, int nr,
 	if (inode)
 		mutex_unlock(&OCFS2_I(inode)->ip_io_mutex);
 
-	mlog(ML_BH_IO, "block=(%"MLFu64"), nr=(%d), cached=%s\n", block, nr,
+	mlog(ML_BH_IO, "block=(%llu), nr=(%d), cached=%s\n", 
+	     (unsigned long long)block, nr,
 	     (!(flags & OCFS2_BH_CACHED) || ignore_cache) ? "no" : "yes");
 
 bail:
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index d08971d29b63..bff0f0d06867 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -449,11 +449,11 @@ static u32 o2hb_compute_block_crc_le(struct o2hb_region *reg,
 
 static void o2hb_dump_slot(struct o2hb_disk_heartbeat_block *hb_block)
 {
-	mlog(ML_ERROR, "Dump slot information: seq = 0x%"MLFx64", node = %u, "
-	     "cksum = 0x%x, generation 0x%"MLFx64"\n",
-	     le64_to_cpu(hb_block->hb_seq), hb_block->hb_node,
-	     le32_to_cpu(hb_block->hb_cksum),
-	     le64_to_cpu(hb_block->hb_generation));
+	mlog(ML_ERROR, "Dump slot information: seq = 0x%llx, node = %u, "
+	     "cksum = 0x%x, generation 0x%llx\n",
+	     (long long)le64_to_cpu(hb_block->hb_seq),
+	     hb_block->hb_node, le32_to_cpu(hb_block->hb_cksum),
+	     (long long)le64_to_cpu(hb_block->hb_generation));
 }
 
 static int o2hb_verify_crc(struct o2hb_region *reg,
@@ -516,8 +516,9 @@ static inline void o2hb_prepare_block(struct o2hb_region *reg,
 	hb_block->hb_cksum = cpu_to_le32(o2hb_compute_block_crc_le(reg,
 								   hb_block));
 
-	mlog(ML_HB_BIO, "our node generation = 0x%"MLFx64", cksum = 0x%x\n",
-	     cpu_to_le64(generation), le32_to_cpu(hb_block->hb_cksum));
+	mlog(ML_HB_BIO, "our node generation = 0x%llx, cksum = 0x%x\n",
+	     (long long)cpu_to_le64(generation),
+	     le32_to_cpu(hb_block->hb_cksum));
 }
 
 static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
@@ -686,19 +687,20 @@ static int o2hb_check_slot(struct o2hb_region *reg,
 	if (slot->ds_last_generation != le64_to_cpu(hb_block->hb_generation)) {
 		gen_changed = 1;
 		slot->ds_equal_samples = 0;
-		mlog(ML_HEARTBEAT, "Node %d changed generation (0x%"MLFx64" "
-		     "to 0x%"MLFx64")\n", slot->ds_node_num,
-		     slot->ds_last_generation,
-		     le64_to_cpu(hb_block->hb_generation));
+		mlog(ML_HEARTBEAT, "Node %d changed generation (0x%llx "
+		     "to 0x%llx)\n", slot->ds_node_num,
+		     (long long)slot->ds_last_generation,
+		     (long long)le64_to_cpu(hb_block->hb_generation));
 	}
 
 	slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
 
-	mlog(ML_HEARTBEAT, "Slot %d gen 0x%"MLFx64" cksum 0x%x "
-	     "seq %"MLFu64" last %"MLFu64" changed %u equal %u\n",
-	     slot->ds_node_num, slot->ds_last_generation,
-	     le32_to_cpu(hb_block->hb_cksum), le64_to_cpu(hb_block->hb_seq), 
-	     slot->ds_last_time, slot->ds_changed_samples,
+	mlog(ML_HEARTBEAT, "Slot %d gen 0x%llx cksum 0x%x "
+	     "seq %llu last %llu changed %u equal %u\n",
+	     slot->ds_node_num, (long long)slot->ds_last_generation,
+	     le32_to_cpu(hb_block->hb_cksum),
+	     (unsigned long long)le64_to_cpu(hb_block->hb_seq), 
+	     (unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
 	     slot->ds_equal_samples);
 
 	spin_lock(&o2hb_live_lock);
@@ -708,8 +710,8 @@ fire_callbacks:
 	 * changes at any time during their dead time */
 	if (list_empty(&slot->ds_live_item) &&
 	    slot->ds_changed_samples >= O2HB_LIVE_THRESHOLD) {
-		mlog(ML_HEARTBEAT, "Node %d (id 0x%"MLFx64") joined my "
-		     "region\n", slot->ds_node_num, slot->ds_last_generation);
+		mlog(ML_HEARTBEAT, "Node %d (id 0x%llx) joined my region\n",
+		     slot->ds_node_num, (long long)slot->ds_last_generation);
 
 		/* first on the list generates a callback */
 		if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 2cadc3009c83..73edad782537 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -256,16 +256,6 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits;
 	}								\
 } while (0)
 
-#if (BITS_PER_LONG == 32) || defined(CONFIG_X86_64) || (defined(CONFIG_UML_X86) && defined(CONFIG_64BIT))
-#define MLFi64 "lld"
-#define MLFu64 "llu"
-#define MLFx64 "llx"
-#else
-#define MLFi64 "ld"
-#define MLFu64 "lu"
-#define MLFx64 "lx"
-#endif
-
 #include <linux/kobject.h>
 #include <linux/sysfs.h>
 int mlog_sys_init(struct subsystem *o2cb_subsys);
diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c
index bd85182e97bc..1a01380e3878 100644
--- a/fs/ocfs2/dcache.c
+++ b/fs/ocfs2/dcache.c
@@ -64,15 +64,16 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
 		/* did we or someone else delete this inode? */
 		if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_DELETED) {
 			spin_unlock(&OCFS2_I(inode)->ip_lock);
-			mlog(0, "inode (%"MLFu64") deleted, returning false\n",
-			     OCFS2_I(inode)->ip_blkno);
+			mlog(0, "inode (%llu) deleted, returning false\n",
+			     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			goto bail;
 		}
 		spin_unlock(&OCFS2_I(inode)->ip_lock);
 
 		if (!inode->i_nlink) {
-			mlog(0, "Inode %"MLFu64" orphaned, returning false "
-			     "dir = %d\n", OCFS2_I(inode)->ip_blkno,
+			mlog(0, "Inode %llu orphaned, returning false "
+			     "dir = %d\n",
+			     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 			     S_ISDIR(inode->i_mode));
 			goto bail;
 		}
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 57158fa75d91..ae47f450792f 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -83,7 +83,8 @@ int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir)
 	struct super_block * sb = inode->i_sb;
 	int have_disk_lock = 0;
 
-	mlog_entry("dirino=%"MLFu64"\n", OCFS2_I(inode)->ip_blkno);
+	mlog_entry("dirino=%llu\n",
+		   (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	stored = 0;
 	bh = NULL;
@@ -104,9 +105,9 @@ int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir)
 		blk = (filp->f_pos) >> sb->s_blocksize_bits;
 		bh = ocfs2_bread(inode, blk, &err, 0);
 		if (!bh) {
-			mlog(ML_ERROR, "directory #%"MLFu64" contains a hole "
-				       "at offset %lld\n",
-			     OCFS2_I(inode)->ip_blkno,
+			mlog(ML_ERROR,
+			     "directory #%llu contains a hole at offset %lld\n",
+			     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 			     filp->f_pos);
 			filp->f_pos += sb->s_blocksize - offset;
 			continue;
@@ -214,9 +215,9 @@ int ocfs2_find_files_on_disk(const char *name,
 	int status = -ENOENT;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
-	mlog_entry("(osb=%p, parent=%"MLFu64", name='%.*s', blkno=%p, "
-		   "inode=%p)\n",
-		   osb, OCFS2_I(inode)->ip_blkno, namelen, name, blkno, inode);
+	mlog_entry("(osb=%p, parent=%llu, name='%.*s', blkno=%p, inode=%p)\n",
+		   osb, (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		   namelen, name, blkno, inode);
 
 	*dirent_bh = ocfs2_find_entry(name, namelen, inode, dirent);
 	if (!*dirent_bh || !*dirent) {
@@ -255,8 +256,8 @@ int ocfs2_check_dir_for_entry(struct inode *dir,
 	struct buffer_head *dirent_bh = NULL;
 	struct ocfs2_dir_entry *dirent = NULL;
 
-	mlog_entry("dir %"MLFu64", name '%.*s'\n", OCFS2_I(dir)->ip_blkno,
-		   namelen, name);
+	mlog_entry("dir %llu, name '%.*s'\n",
+		   (unsigned long long)OCFS2_I(dir)->ip_blkno, namelen, name);
 
 	ret = -EEXIST;
 	dirent_bh = ocfs2_find_entry(name, namelen, dir, &dirent);
@@ -287,9 +288,8 @@ int ocfs2_empty_dir(struct inode *inode)
 	if ((i_size_read(inode) <
 	     (OCFS2_DIR_REC_LEN(1) + OCFS2_DIR_REC_LEN(2))) ||
 	    !(bh = ocfs2_bread(inode, 0, &err, 0))) {
-	    	mlog(ML_ERROR, "bad directory (dir #%"MLFu64") - "
-			       "no data block\n",
-		     OCFS2_I(inode)->ip_blkno);
+	    	mlog(ML_ERROR, "bad directory (dir #%llu) - no data block\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		return 1;
 	}
 
@@ -300,9 +300,8 @@ int ocfs2_empty_dir(struct inode *inode)
 			!le64_to_cpu(de1->inode) ||
 			strcmp(".", de->name) ||
 			strcmp("..", de1->name)) {
-	    	mlog(ML_ERROR, "bad directory (dir #%"MLFu64") - "
-			       "no `.' or `..'\n",
-		     OCFS2_I(inode)->ip_blkno);
+	    	mlog(ML_ERROR, "bad directory (dir #%llu) - no `.' or `..'\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		brelse(bh);
 		return 1;
 	}
@@ -314,9 +313,8 @@ int ocfs2_empty_dir(struct inode *inode)
 			bh = ocfs2_bread(inode,
 					 offset >> sb->s_blocksize_bits, &err, 0);
 			if (!bh) {
-				mlog(ML_ERROR, "directory #%"MLFu64" contains "
-					       "a hole at offset %lu\n",
-				     OCFS2_I(inode)->ip_blkno, offset);
+				mlog(ML_ERROR, "dir %llu has a hole at %lu\n",
+				     (unsigned long long)OCFS2_I(inode)->ip_blkno, offset);
 				offset += sb->s_blocksize;
 				continue;
 			}
@@ -406,8 +404,8 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb,
 	mlog_entry_void();
 
 	dir_i_size = i_size_read(dir);
-	mlog(0, "extending dir %"MLFu64" (i_size = %lld)\n",
-	     OCFS2_I(dir)->ip_blkno, dir_i_size);
+	mlog(0, "extending dir %llu (i_size = %lld)\n",
+	     (unsigned long long)OCFS2_I(dir)->ip_blkno, dir_i_size);
 
 	handle = ocfs2_alloc_handle(osb);
 	if (handle == NULL) {
@@ -531,8 +529,8 @@ int ocfs2_prepare_dir_for_insert(struct ocfs2_super *osb,
 
 	mlog_entry_void();
 
-	mlog(0, "getting ready to insert namelen %d into dir %"MLFu64"\n",
-	     namelen, OCFS2_I(dir)->ip_blkno);
+	mlog(0, "getting ready to insert namelen %d into dir %llu\n",
+	     namelen, (unsigned long long)OCFS2_I(dir)->ip_blkno);
 
 	BUG_ON(!S_ISDIR(dir->i_mode));
 	fe = (struct ocfs2_dinode *) parent_fe_bh->b_data;
diff --git a/fs/ocfs2/dlm/dlmast.c b/fs/ocfs2/dlm/dlmast.c
index 8d17d28ef91c..355593dd8ef8 100644
--- a/fs/ocfs2/dlm/dlmast.c
+++ b/fs/ocfs2/dlm/dlmast.c
@@ -307,8 +307,11 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data)
 
 	if (past->type != DLM_AST &&
 	    past->type != DLM_BAST) {
-		mlog(ML_ERROR, "Unknown ast type! %d, cookie=%"MLFu64", "
-		     "name=%.*s\n", past->type, cookie, locklen, name);
+		mlog(ML_ERROR, "Unknown ast type! %d, cookie=%u:%llu"
+		     "name=%.*s\n", past->type, 
+		     dlm_get_lock_cookie_node(cookie),
+		     dlm_get_lock_cookie_seq(cookie),
+		     locklen, name);
 		ret = DLM_IVLOCKID;
 		goto leave;
 	}
@@ -316,9 +319,11 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data)
 	res = dlm_lookup_lockres(dlm, name, locklen);
 	if (!res) {
 		mlog(ML_ERROR, "got %sast for unknown lockres! "
-			       "cookie=%"MLFu64", name=%.*s, namelen=%u\n",
+			       "cookie=%u:%llu, name=%.*s, namelen=%u\n",
 		     past->type == DLM_AST ? "" : "b",
-		     cookie, locklen, name, locklen);
+		     dlm_get_lock_cookie_node(cookie),
+		     dlm_get_lock_cookie_seq(cookie),
+		     locklen, name, locklen);
 		ret = DLM_IVLOCKID;
 		goto leave;
 	}
@@ -360,9 +365,12 @@ int dlm_proxy_ast_handler(struct o2net_msg *msg, u32 len, void *data)
 			goto do_ast;
 	}
 
-	mlog(ML_ERROR, "got %sast for unknown lock!  cookie=%"MLFu64", "
-		       "name=%.*s, namelen=%u\n",
-             past->type == DLM_AST ? "" : "b", cookie, locklen, name, locklen);
+	mlog(ML_ERROR, "got %sast for unknown lock!  cookie=%u:%llu, "
+		       "name=%.*s, namelen=%u\n", 
+		       past->type == DLM_AST ? "" : "b", 
+		       dlm_get_lock_cookie_node(cookie),
+		       dlm_get_lock_cookie_seq(cookie),
+		       locklen, name, locklen);
 
 	ret = DLM_NORMAL;
 unlock_out:
diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h
index 9c772583744a..88cc43df18f1 100644
--- a/fs/ocfs2/dlm/dlmcommon.h
+++ b/fs/ocfs2/dlm/dlmcommon.h
@@ -630,6 +630,21 @@ __dlm_lockres_state_to_status(struct dlm_lock_resource *res)
 	return status;
 }
 
+static inline u8 dlm_get_lock_cookie_node(u64 cookie)
+{
+	u8 ret;
+	cookie >>= 56;
+	ret = (u8)(cookie & 0xffULL);
+	return ret;
+}
+
+static inline unsigned long long dlm_get_lock_cookie_seq(u64 cookie)
+{
+	unsigned long long ret;
+	ret = ((unsigned long long)cookie) & 0x00ffffffffffffffULL;
+	return ret;
+}
+
 struct dlm_lock * dlm_new_lock(int type, u8 node, u64 cookie,
 			       struct dlm_lockstatus *lksb);
 void dlm_lock_get(struct dlm_lock *lock);
@@ -658,6 +673,7 @@ void dlm_complete_thread(struct dlm_ctxt *dlm);
 int dlm_launch_recovery_thread(struct dlm_ctxt *dlm);
 void dlm_complete_recovery_thread(struct dlm_ctxt *dlm);
 void dlm_wait_for_recovery(struct dlm_ctxt *dlm);
+void dlm_kick_recovery_thread(struct dlm_ctxt *dlm);
 int dlm_is_node_dead(struct dlm_ctxt *dlm, u8 node);
 int dlm_wait_for_node_death(struct dlm_ctxt *dlm, u8 node, int timeout);
 
@@ -762,6 +778,11 @@ int dlm_request_all_locks_handler(struct o2net_msg *msg, u32 len, void *data);
 int dlm_reco_data_done_handler(struct o2net_msg *msg, u32 len, void *data);
 int dlm_begin_reco_handler(struct o2net_msg *msg, u32 len, void *data);
 int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data);
+int dlm_do_master_requery(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
+			  u8 nodenum, u8 *real_master);
+int dlm_lockres_master_requery(struct dlm_ctxt *dlm,
+			       struct dlm_lock_resource *res, u8 *real_master);
+
 
 int dlm_dispatch_assert_master(struct dlm_ctxt *dlm,
 			       struct dlm_lock_resource *res,
diff --git a/fs/ocfs2/dlm/dlmconvert.c b/fs/ocfs2/dlm/dlmconvert.c
index f66e2d818ccd..8285228d9e37 100644
--- a/fs/ocfs2/dlm/dlmconvert.c
+++ b/fs/ocfs2/dlm/dlmconvert.c
@@ -284,8 +284,10 @@ enum dlm_status dlmconvert_remote(struct dlm_ctxt *dlm,
 	if (lock->ml.convert_type != LKM_IVMODE) {
 		__dlm_print_one_lock_resource(res);
 		mlog(ML_ERROR, "converting a remote lock that is already "
-		     "converting! (cookie=%"MLFu64", conv=%d)\n",
-		     lock->ml.cookie, lock->ml.convert_type);
+		     "converting! (cookie=%u:%llu, conv=%d)\n",
+		     dlm_get_lock_cookie_node(lock->ml.cookie),
+		     dlm_get_lock_cookie_seq(lock->ml.cookie),
+		     lock->ml.convert_type);
 		status = DLM_DENIED;
 		goto bail;
 	}
@@ -513,8 +515,9 @@ int dlm_convert_lock_handler(struct o2net_msg *msg, u32 len, void *data)
 leave:
 	if (!lock)
 		mlog(ML_ERROR, "did not find lock to convert on grant queue! "
-			       "cookie=%"MLFu64"\n",
-		     cnv->cookie);
+			       "cookie=%u:%llu\n",
+			       dlm_get_lock_cookie_node(cnv->cookie),
+			       dlm_get_lock_cookie_seq(cnv->cookie));
 	else
 		dlm_lock_put(lock);
 
diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c
index 54f61b76ab51..c7eae5d3324e 100644
--- a/fs/ocfs2/dlm/dlmdebug.c
+++ b/fs/ocfs2/dlm/dlmdebug.c
@@ -72,8 +72,10 @@ void __dlm_print_one_lock_resource(struct dlm_lock_resource *res)
 		lock = list_entry(iter2, struct dlm_lock, list);
 		spin_lock(&lock->spinlock);
 		mlog(ML_NOTICE, "    type=%d, conv=%d, node=%u, "
-		       "cookie=%"MLFu64", ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", 
-		       lock->ml.type, lock->ml.convert_type, lock->ml.node, lock->ml.cookie, 
+		       "cookie=%u:%llu, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", 
+		       lock->ml.type, lock->ml.convert_type, lock->ml.node, 
+		       dlm_get_lock_cookie_node(lock->ml.cookie), 
+		       dlm_get_lock_cookie_seq(lock->ml.cookie), 
 		       list_empty(&lock->ast_list) ? 'y' : 'n',
 		       lock->ast_pending ? 'y' : 'n',
 		       list_empty(&lock->bast_list) ? 'y' : 'n',
@@ -85,8 +87,10 @@ void __dlm_print_one_lock_resource(struct dlm_lock_resource *res)
 		lock = list_entry(iter2, struct dlm_lock, list);
 		spin_lock(&lock->spinlock);
 		mlog(ML_NOTICE, "    type=%d, conv=%d, node=%u, "
-		       "cookie=%"MLFu64", ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", 
-		       lock->ml.type, lock->ml.convert_type, lock->ml.node, lock->ml.cookie, 
+		       "cookie=%u:%llu, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", 
+		       lock->ml.type, lock->ml.convert_type, lock->ml.node, 
+		       dlm_get_lock_cookie_node(lock->ml.cookie), 
+		       dlm_get_lock_cookie_seq(lock->ml.cookie), 
 		       list_empty(&lock->ast_list) ? 'y' : 'n',
 		       lock->ast_pending ? 'y' : 'n',
 		       list_empty(&lock->bast_list) ? 'y' : 'n',
@@ -98,8 +102,10 @@ void __dlm_print_one_lock_resource(struct dlm_lock_resource *res)
 		lock = list_entry(iter2, struct dlm_lock, list);
 		spin_lock(&lock->spinlock);
 		mlog(ML_NOTICE, "    type=%d, conv=%d, node=%u, "
-		       "cookie=%"MLFu64", ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", 
-		       lock->ml.type, lock->ml.convert_type, lock->ml.node, lock->ml.cookie, 
+		       "cookie=%u:%llu, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", 
+		       lock->ml.type, lock->ml.convert_type, lock->ml.node, 
+		       dlm_get_lock_cookie_node(lock->ml.cookie), 
+		       dlm_get_lock_cookie_seq(lock->ml.cookie), 
 		       list_empty(&lock->ast_list) ? 'y' : 'n',
 		       lock->ast_pending ? 'y' : 'n',
 		       list_empty(&lock->bast_list) ? 'y' : 'n',
diff --git a/fs/ocfs2/dlm/dlmfs.c b/fs/ocfs2/dlm/dlmfs.c
index dd2d24dc25e0..7e88e24b3471 100644
--- a/fs/ocfs2/dlm/dlmfs.c
+++ b/fs/ocfs2/dlm/dlmfs.c
@@ -596,7 +596,8 @@ static int __init init_dlmfs_fs(void)
 
 	dlmfs_inode_cache = kmem_cache_create("dlmfs_inode_cache",
 				sizeof(struct dlmfs_inode_private),
-				0, SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT,
+				0, (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
+					SLAB_MEM_SPREAD),
 				dlmfs_init_once, NULL);
 	if (!dlmfs_inode_cache)
 		return -ENOMEM;
diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c
index 671d4ff222cc..6fea28318d6d 100644
--- a/fs/ocfs2/dlm/dlmlock.c
+++ b/fs/ocfs2/dlm/dlmlock.c
@@ -141,13 +141,23 @@ static enum dlm_status dlmlock_master(struct dlm_ctxt *dlm,
 					  res->lockname.len)) {
 			kick_thread = 1;
 			call_ast = 1;
+		} else {
+			mlog(0, "%s: returning DLM_NORMAL to "
+			     "node %u for reco lock\n", dlm->name,
+			     lock->ml.node);
 		}
 	} else {
 		/* for NOQUEUE request, unless we get the
 		 * lock right away, return DLM_NOTQUEUED */
-		if (flags & LKM_NOQUEUE)
+		if (flags & LKM_NOQUEUE) {
 			status = DLM_NOTQUEUED;
-		else {
+			if (dlm_is_recovery_lock(res->lockname.name,
+						 res->lockname.len)) {
+				mlog(0, "%s: returning NOTQUEUED to "
+				     "node %u for reco lock\n", dlm->name,
+				     lock->ml.node);
+			}
+		} else {
 			dlm_lock_get(lock);
 			list_add_tail(&lock->list, &res->blocked);
 			kick_thread = 1;
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 847dd3cc4cf5..940be4c13b1f 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -239,6 +239,8 @@ static void dlm_remove_nonlocal_locks(struct dlm_ctxt *dlm,
 static int dlm_mark_lockres_migrating(struct dlm_ctxt *dlm,
 				       struct dlm_lock_resource *res,
 				       u8 target);
+static int dlm_pre_master_reco_lockres(struct dlm_ctxt *dlm,
+				       struct dlm_lock_resource *res);
 
 
 int dlm_is_host_down(int errno)
@@ -677,6 +679,7 @@ struct dlm_lock_resource * dlm_get_lock_resource(struct dlm_ctxt *dlm,
 	struct dlm_node_iter iter;
 	unsigned int namelen;
 	int tries = 0;
+	int bit, wait_on_recovery = 0;
 
 	BUG_ON(!lockid);
 
@@ -762,6 +765,18 @@ lookup:
 		dlm_init_mle(mle, DLM_MLE_MASTER, dlm, res, NULL, 0);
 		set_bit(dlm->node_num, mle->maybe_map);
 		list_add(&mle->list, &dlm->master_list);
+
+		/* still holding the dlm spinlock, check the recovery map
+		 * to see if there are any nodes that still need to be 
+		 * considered.  these will not appear in the mle nodemap
+		 * but they might own this lockres.  wait on them. */
+		bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0);
+		if (bit < O2NM_MAX_NODES) {
+			mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to"
+			     "recover before lock mastery can begin\n",
+			     dlm->name, namelen, (char *)lockid, bit);
+			wait_on_recovery = 1;
+		}
 	}
 
 	/* at this point there is either a DLM_MLE_BLOCK or a
@@ -779,6 +794,39 @@ lookup:
 	spin_unlock(&dlm->master_lock);
 	spin_unlock(&dlm->spinlock);
 
+	while (wait_on_recovery) {
+		/* any cluster changes that occurred after dropping the
+		 * dlm spinlock would be detectable be a change on the mle,
+		 * so we only need to clear out the recovery map once. */
+		if (dlm_is_recovery_lock(lockid, namelen)) {
+			mlog(ML_NOTICE, "%s: recovery map is not empty, but "
+			     "must master $RECOVERY lock now\n", dlm->name);
+			if (!dlm_pre_master_reco_lockres(dlm, res))
+				wait_on_recovery = 0;
+			else {
+				mlog(0, "%s: waiting 500ms for heartbeat state "
+				    "change\n", dlm->name);
+				msleep(500);
+			}
+			continue;
+		} 
+
+		dlm_kick_recovery_thread(dlm);
+		msleep(100);
+		dlm_wait_for_recovery(dlm);
+
+		spin_lock(&dlm->spinlock);
+		bit = find_next_bit(dlm->recovery_map, O2NM_MAX_NODES, 0);
+		if (bit < O2NM_MAX_NODES) {
+			mlog(ML_NOTICE, "%s:%.*s: at least one node (%d) to"
+			     "recover before lock mastery can begin\n",
+			     dlm->name, namelen, (char *)lockid, bit);
+			wait_on_recovery = 1;
+		} else
+			wait_on_recovery = 0;
+		spin_unlock(&dlm->spinlock);
+	}
+
 	/* must wait for lock to be mastered elsewhere */
 	if (blocked)
 		goto wait;
@@ -792,7 +840,15 @@ redo_request:
 			mlog_errno(ret);
 		if (mle->master != O2NM_MAX_NODES) {
 			/* found a master ! */
-			break;
+			if (mle->master <= nodenum)
+				break;
+			/* if our master request has not reached the master
+			 * yet, keep going until it does.  this is how the
+			 * master will know that asserts are needed back to
+			 * the lower nodes. */
+			mlog(0, "%s:%.*s: requests only up to %u but master "
+			     "is %u, keep going\n", dlm->name, namelen,
+			     lockid, nodenum, mle->master);
 		}
 	}
 
@@ -860,7 +916,19 @@ recheck:
 	/* check if another node has already become the owner */
 	spin_lock(&res->spinlock);
 	if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN) {
+		mlog(0, "%s:%.*s: owner is suddenly %u\n", dlm->name,
+		     res->lockname.len, res->lockname.name, res->owner);
 		spin_unlock(&res->spinlock);
+		/* this will cause the master to re-assert across
+		 * the whole cluster, freeing up mles */
+		ret = dlm_do_master_request(mle, res->owner);
+		if (ret < 0) {
+			/* give recovery a chance to run */
+			mlog(ML_ERROR, "link to %u went down?: %d\n", res->owner, ret);
+			msleep(500);
+			goto recheck;
+		}
+		ret = 0;
 		goto leave;
 	}
 	spin_unlock(&res->spinlock);
@@ -1244,13 +1312,14 @@ int dlm_master_request_handler(struct o2net_msg *msg, u32 len, void *data)
 {
 	u8 response = DLM_MASTER_RESP_MAYBE;
 	struct dlm_ctxt *dlm = data;
-	struct dlm_lock_resource *res;
+	struct dlm_lock_resource *res = NULL;
 	struct dlm_master_request *request = (struct dlm_master_request *) msg->buf;
 	struct dlm_master_list_entry *mle = NULL, *tmpmle = NULL;
 	char *name;
 	unsigned int namelen;
 	int found, ret;
 	int set_maybe;
+	int dispatch_assert = 0;
 
 	if (!dlm_grab(dlm))
 		return DLM_MASTER_RESP_NO;
@@ -1287,7 +1356,6 @@ way_up_top:
 		}
 
 		if (res->owner == dlm->node_num) {
-			u32 flags = DLM_ASSERT_MASTER_MLE_CLEANUP;
 			spin_unlock(&res->spinlock);
 			// mlog(0, "this node is the master\n");
 			response = DLM_MASTER_RESP_YES;
@@ -1300,16 +1368,7 @@ way_up_top:
 			 * caused all nodes up to this one to
 			 * create mles.  this node now needs to
 			 * go back and clean those up. */
-			mlog(0, "%u is the owner of %.*s, cleaning everyone else\n",
-			     dlm->node_num, res->lockname.len, res->lockname.name);
-			ret = dlm_dispatch_assert_master(dlm, res, 1,
-							 request->node_idx,
-							 flags);
-			if (ret < 0) {
-				mlog(ML_ERROR, "failed to dispatch assert "
-				     "master work\n");
-				response = DLM_MASTER_RESP_ERROR;
-			}
+			dispatch_assert = 1;
 			goto send_response;
 		} else if (res->owner != DLM_LOCK_RES_OWNER_UNKNOWN) {
 			spin_unlock(&res->spinlock);
@@ -1357,9 +1416,13 @@ way_up_top:
 			}
 		} else if (tmpmle->master != DLM_LOCK_RES_OWNER_UNKNOWN) {
 			set_maybe = 0;
-			if (tmpmle->master == dlm->node_num)
+			if (tmpmle->master == dlm->node_num) {
 				response = DLM_MASTER_RESP_YES;
-			else
+				/* this node will be the owner.
+				 * go back and clean the mles on any
+				 * other nodes */
+				dispatch_assert = 1;
+			} else
 				response = DLM_MASTER_RESP_NO;
 		} else {
 			// mlog(0, "this node is attempting to "
@@ -1398,8 +1461,8 @@ way_up_top:
 			mle = (struct dlm_master_list_entry *)
 				kmem_cache_alloc(dlm_mle_cache, GFP_KERNEL);
 			if (!mle) {
-				// bad bad bad... this sucks.
 				response = DLM_MASTER_RESP_ERROR;
+				mlog_errno(-ENOMEM);
 				goto send_response;
 			}
 			spin_lock(&dlm->spinlock);
@@ -1418,25 +1481,19 @@ way_up_top:
 		// mlog(0, "mle was found\n");
 		set_maybe = 1;
 		spin_lock(&tmpmle->spinlock);
+		if (tmpmle->master == dlm->node_num) {
+			mlog(ML_ERROR, "no lockres, but an mle with this node as master!\n");
+			BUG();
+		}
 		if (tmpmle->type == DLM_MLE_BLOCK)
 			response = DLM_MASTER_RESP_NO;
 		else if (tmpmle->type == DLM_MLE_MIGRATION) {
 			mlog(0, "migration mle was found (%u->%u)\n",
 			     tmpmle->master, tmpmle->new_master);
-			if (tmpmle->master == dlm->node_num) {
-				mlog(ML_ERROR, "no lockres, but migration mle "
-				     "says that this node is master!\n");
-				BUG();
-			}
 			/* real master can respond on its own */
 			response = DLM_MASTER_RESP_NO;
-		} else {
-			if (tmpmle->master == dlm->node_num) {
-				response = DLM_MASTER_RESP_YES;
-				set_maybe = 0;
-			} else
-				response = DLM_MASTER_RESP_MAYBE;
-		}
+		} else
+			response = DLM_MASTER_RESP_MAYBE;
 		if (set_maybe)
 			set_bit(request->node_idx, tmpmle->maybe_map);
 		spin_unlock(&tmpmle->spinlock);
@@ -1449,6 +1506,24 @@ way_up_top:
 		dlm_put_mle(tmpmle);
 	}
 send_response:
+
+	if (dispatch_assert) {
+		if (response != DLM_MASTER_RESP_YES)
+			mlog(ML_ERROR, "invalid response %d\n", response);
+		if (!res) {
+			mlog(ML_ERROR, "bad lockres while trying to assert!\n");
+			BUG();
+		}
+		mlog(0, "%u is the owner of %.*s, cleaning everyone else\n",
+			     dlm->node_num, res->lockname.len, res->lockname.name);
+		ret = dlm_dispatch_assert_master(dlm, res, 0, request->node_idx, 
+						 DLM_ASSERT_MASTER_MLE_CLEANUP);
+		if (ret < 0) {
+			mlog(ML_ERROR, "failed to dispatch assert master work\n");
+			response = DLM_MASTER_RESP_ERROR;
+		}
+	}
+
 	dlm_put(dlm);
 	return response;
 }
@@ -1471,8 +1546,11 @@ static int dlm_do_assert_master(struct dlm_ctxt *dlm, const char *lockname,
 	int to, tmpret;
 	struct dlm_node_iter iter;
 	int ret = 0;
+	int reassert;
 
 	BUG_ON(namelen > O2NM_MAX_NAME_LEN);
+again:
+	reassert = 0;
 
 	/* note that if this nodemap is empty, it returns 0 */
 	dlm_node_iter_init(nodemap, &iter);
@@ -1504,9 +1582,17 @@ static int dlm_do_assert_master(struct dlm_ctxt *dlm, const char *lockname,
 			     "got %d.\n", namelen, lockname, to, r);
 			dlm_dump_lock_resources(dlm);
 			BUG();
+		} else if (r == EAGAIN) {
+			mlog(0, "%.*s: node %u create mles on other "
+			     "nodes and requests a re-assert\n", 
+			     namelen, lockname, to);
+			reassert = 1;
 		}
 	}
 
+	if (reassert)
+		goto again;
+
 	return ret;
 }
 
@@ -1528,6 +1614,8 @@ int dlm_assert_master_handler(struct o2net_msg *msg, u32 len, void *data)
 	char *name;
 	unsigned int namelen;
 	u32 flags;
+	int master_request = 0;
+	int ret = 0;
 
 	if (!dlm_grab(dlm))
 		return 0;
@@ -1642,11 +1730,22 @@ ok:
 	// mlog(0, "woo!  got an assert_master from node %u!\n",
 	// 	     assert->node_idx);
 	if (mle) {
-		int extra_ref;
+		int extra_ref = 0;
+		int nn = -1;
 		
 		spin_lock(&mle->spinlock);
-		extra_ref = !!(mle->type == DLM_MLE_BLOCK
-			       || mle->type == DLM_MLE_MIGRATION);
+		if (mle->type == DLM_MLE_BLOCK || mle->type == DLM_MLE_MIGRATION)
+			extra_ref = 1;
+		else {
+			/* MASTER mle: if any bits set in the response map
+			 * then the calling node needs to re-assert to clear
+			 * up nodes that this node contacted */
+			while ((nn = find_next_bit (mle->response_map, O2NM_MAX_NODES, 
+						    nn+1)) < O2NM_MAX_NODES) {
+				if (nn != dlm->node_num && nn != assert->node_idx)
+					master_request = 1;
+			}
+		}
 		mle->master = assert->node_idx;
 		atomic_set(&mle->woken, 1);
 		wake_up(&mle->wq);
@@ -1677,10 +1776,15 @@ ok:
 	}
 
 done:
+	ret = 0;
 	if (res)
 		dlm_lockres_put(res);
 	dlm_put(dlm);
-	return 0;
+	if (master_request) {
+		mlog(0, "need to tell master to reassert\n");
+		ret = EAGAIN;  // positive. negative would shoot down the node.
+	}
+	return ret;
 
 kill:
 	/* kill the caller! */
@@ -1713,6 +1817,10 @@ int dlm_dispatch_assert_master(struct dlm_ctxt *dlm,
 	item->u.am.request_from = request_from;
 	item->u.am.flags = flags;
 
+	if (ignore_higher) 
+		mlog(0, "IGNORE HIGHER: %.*s\n", res->lockname.len, 
+		     res->lockname.name);
+		
 	spin_lock(&dlm->work_lock);
 	list_add_tail(&item->list, &dlm->work_list);
 	spin_unlock(&dlm->work_lock);
@@ -1775,6 +1883,61 @@ static void dlm_assert_master_worker(struct dlm_work_item *item, void *data)
 	mlog(0, "finished with dlm_assert_master_worker\n");
 }
 
+/* SPECIAL CASE for the $RECOVERY lock used by the recovery thread.
+ * We cannot wait for node recovery to complete to begin mastering this
+ * lockres because this lockres is used to kick off recovery! ;-)
+ * So, do a pre-check on all living nodes to see if any of those nodes
+ * think that $RECOVERY is currently mastered by a dead node.  If so,
+ * we wait a short time to allow that node to get notified by its own
+ * heartbeat stack, then check again.  All $RECOVERY lock resources
+ * mastered by dead nodes are purged when the hearbeat callback is 
+ * fired, so we can know for sure that it is safe to continue once
+ * the node returns a live node or no node.  */
+static int dlm_pre_master_reco_lockres(struct dlm_ctxt *dlm,
+				       struct dlm_lock_resource *res)
+{
+	struct dlm_node_iter iter;
+	int nodenum;
+	int ret = 0;
+	u8 master = DLM_LOCK_RES_OWNER_UNKNOWN;
+
+	spin_lock(&dlm->spinlock);
+	dlm_node_iter_init(dlm->domain_map, &iter);
+	spin_unlock(&dlm->spinlock);
+
+	while ((nodenum = dlm_node_iter_next(&iter)) >= 0) {
+		/* do not send to self */
+		if (nodenum == dlm->node_num)
+			continue;
+		ret = dlm_do_master_requery(dlm, res, nodenum, &master);
+		if (ret < 0) {
+			mlog_errno(ret);
+			if (!dlm_is_host_down(ret))
+				BUG();
+			/* host is down, so answer for that node would be
+			 * DLM_LOCK_RES_OWNER_UNKNOWN.  continue. */
+		}
+
+		if (master != DLM_LOCK_RES_OWNER_UNKNOWN) {
+			/* check to see if this master is in the recovery map */
+			spin_lock(&dlm->spinlock);
+			if (test_bit(master, dlm->recovery_map)) {
+				mlog(ML_NOTICE, "%s: node %u has not seen "
+				     "node %u go down yet, and thinks the "
+				     "dead node is mastering the recovery "
+				     "lock.  must wait.\n", dlm->name,
+				     nodenum, master);
+				ret = -EAGAIN;
+			}
+			spin_unlock(&dlm->spinlock);
+			mlog(0, "%s: reco lock master is %u\n", dlm->name, 
+			     master);
+			break;
+		}
+	}
+	return ret;
+}
+
 
 /*
  * DLM_MIGRATE_LOCKRES
diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c
index 1e232000f3f7..805cbabac051 100644
--- a/fs/ocfs2/dlm/dlmrecovery.c
+++ b/fs/ocfs2/dlm/dlmrecovery.c
@@ -58,7 +58,7 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node);
 static int dlm_recovery_thread(void *data);
 void dlm_complete_recovery_thread(struct dlm_ctxt *dlm);
 int dlm_launch_recovery_thread(struct dlm_ctxt *dlm);
-static void dlm_kick_recovery_thread(struct dlm_ctxt *dlm);
+void dlm_kick_recovery_thread(struct dlm_ctxt *dlm);
 static int dlm_do_recovery(struct dlm_ctxt *dlm);
 
 static int dlm_pick_recovery_master(struct dlm_ctxt *dlm);
@@ -78,15 +78,9 @@ static int dlm_send_mig_lockres_msg(struct dlm_ctxt *dlm,
 				    u8 send_to,
 				    struct dlm_lock_resource *res,
 				    int total_locks);
-static int dlm_lockres_master_requery(struct dlm_ctxt *dlm,
-				      struct dlm_lock_resource *res,
-				      u8 *real_master);
 static int dlm_process_recovery_data(struct dlm_ctxt *dlm,
 				     struct dlm_lock_resource *res,
 				     struct dlm_migratable_lockres *mres);
-static int dlm_do_master_requery(struct dlm_ctxt *dlm,
-				 struct dlm_lock_resource *res,
-				 u8 nodenum, u8 *real_master);
 static int dlm_send_finalize_reco_message(struct dlm_ctxt *dlm);
 static int dlm_send_all_done_msg(struct dlm_ctxt *dlm,
 				 u8 dead_node, u8 send_to);
@@ -165,7 +159,7 @@ void dlm_dispatch_work(void *data)
  * RECOVERY THREAD
  */
 
-static void dlm_kick_recovery_thread(struct dlm_ctxt *dlm)
+void dlm_kick_recovery_thread(struct dlm_ctxt *dlm)
 {
 	/* wake the recovery thread
 	 * this will wake the reco thread in one of three places
@@ -750,10 +744,12 @@ static void dlm_request_all_locks_worker(struct dlm_work_item *item, void *data)
 		     dlm->name, dlm->reco.dead_node, dlm->reco.new_master,
 		     dead_node, reco_master);
 		mlog(ML_ERROR, "%s: name=%.*s master=%u locks=%u/%u flags=%u "
-		     "entry[0]={c=%"MLFu64",l=%u,f=%u,t=%d,ct=%d,hb=%d,n=%u}\n",
+		     "entry[0]={c=%u:%llu,l=%u,f=%u,t=%d,ct=%d,hb=%d,n=%u}\n",
 		     dlm->name, mres->lockname_len, mres->lockname, mres->master,
 		     mres->num_locks, mres->total_locks, mres->flags,
-		     mres->ml[0].cookie, mres->ml[0].list, mres->ml[0].flags,
+		     dlm_get_lock_cookie_node(mres->ml[0].cookie),
+		     dlm_get_lock_cookie_seq(mres->ml[0].cookie),
+		     mres->ml[0].list, mres->ml[0].flags,
 		     mres->ml[0].type, mres->ml[0].convert_type,
 		     mres->ml[0].highest_blocked, mres->ml[0].node);
 		BUG();
@@ -1316,9 +1312,8 @@ leave:
 
 
 
-static int dlm_lockres_master_requery(struct dlm_ctxt *dlm,
-				      struct dlm_lock_resource *res,
-				      u8 *real_master)
+int dlm_lockres_master_requery(struct dlm_ctxt *dlm,
+			       struct dlm_lock_resource *res, u8 *real_master)
 {
 	struct dlm_node_iter iter;
 	int nodenum;
@@ -1360,8 +1355,10 @@ static int dlm_lockres_master_requery(struct dlm_ctxt *dlm,
 		ret = dlm_do_master_requery(dlm, res, nodenum, real_master);
 		if (ret < 0) {
 			mlog_errno(ret);
-			BUG();
-			/* TODO: need to figure a way to restart this */
+			if (!dlm_is_host_down(ret))
+				BUG();
+			/* host is down, so answer for that node would be
+			 * DLM_LOCK_RES_OWNER_UNKNOWN.  continue. */
 		}
 		if (*real_master != DLM_LOCK_RES_OWNER_UNKNOWN) {
 			mlog(0, "lock master is %u\n", *real_master);
@@ -1372,9 +1369,8 @@ static int dlm_lockres_master_requery(struct dlm_ctxt *dlm,
 }
 
 
-static int dlm_do_master_requery(struct dlm_ctxt *dlm,
-				 struct dlm_lock_resource *res,
-				 u8 nodenum, u8 *real_master)
+int dlm_do_master_requery(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
+			  u8 nodenum, u8 *real_master)
 {
 	int ret = -EINVAL;
 	struct dlm_master_requery req;
@@ -1519,9 +1515,11 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm,
 			/* lock is always created locally first, and
 			 * destroyed locally last.  it must be on the list */
 			if (!lock) {
+				u64 c = ml->cookie;
 				mlog(ML_ERROR, "could not find local lock "
-					       "with cookie %"MLFu64"!\n",
-				     ml->cookie);
+					       "with cookie %u:%llu!\n",
+					       dlm_get_lock_cookie_node(c),
+					       dlm_get_lock_cookie_seq(c));
 				BUG();
 			}
 			BUG_ON(lock->ml.node != ml->node);
@@ -1739,6 +1737,13 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm,
 				} else
 					continue;
 
+				if (!list_empty(&res->recovering)) {
+					mlog(0, "%s:%.*s: lockres was "
+					     "marked RECOVERING, owner=%u\n",
+					     dlm->name, res->lockname.len,
+					     res->lockname.name, res->owner);
+					list_del_init(&res->recovering);
+				}
 				spin_lock(&res->spinlock);
 				dlm_change_lockres_owner(dlm, res, new_master);
 				res->state &= ~DLM_LOCK_RES_RECOVERING;
@@ -2258,7 +2263,10 @@ int dlm_begin_reco_handler(struct o2net_msg *msg, u32 len, void *data)
 			mlog(0, "%u not in domain/live_nodes map "
 			     "so setting it in reco map manually\n",
 			     br->dead_node);
-		set_bit(br->dead_node, dlm->recovery_map);
+		/* force the recovery cleanup in __dlm_hb_node_down
+		 * both of these will be cleared in a moment */
+		set_bit(br->dead_node, dlm->domain_map);
+		set_bit(br->dead_node, dlm->live_nodes_map);
 		__dlm_hb_node_down(dlm, br->dead_node);
 	}
 	spin_unlock(&dlm->spinlock);
diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c
index c95f08d2e925..7b1a27542674 100644
--- a/fs/ocfs2/dlm/dlmunlock.c
+++ b/fs/ocfs2/dlm/dlmunlock.c
@@ -244,8 +244,10 @@ leave:
 	if (actions & DLM_UNLOCK_FREE_LOCK) {
 		/* this should always be coupled with list removal */
 		BUG_ON(!(actions & DLM_UNLOCK_REMOVE_LOCK));
-		mlog(0, "lock %"MLFu64" should be gone now! refs=%d\n",
-		     lock->ml.cookie, atomic_read(&lock->lock_refs.refcount)-1);
+		mlog(0, "lock %u:%llu should be gone now! refs=%d\n",
+		     dlm_get_lock_cookie_node(lock->ml.cookie),
+		     dlm_get_lock_cookie_seq(lock->ml.cookie),
+		     atomic_read(&lock->lock_refs.refcount)-1);
 		dlm_lock_put(lock);
 	}
 	if (actions & DLM_UNLOCK_CALL_AST)
@@ -493,8 +495,9 @@ int dlm_unlock_lock_handler(struct o2net_msg *msg, u32 len, void *data)
 not_found:
 	if (!found)
 		mlog(ML_ERROR, "failed to find lock to unlock! "
-			       "cookie=%"MLFu64"\n",
-		     unlock->cookie);
+			       "cookie=%u:%llu\n",
+			       dlm_get_lock_cookie_node(unlock->cookie),
+			       dlm_get_lock_cookie_seq(unlock->cookie));
 	else {
 		/* send the lksb->status back to the other node */
 		status = lksb->status;
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index e971ec2f8407..64cd52860c87 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -231,9 +231,9 @@ static void ocfs2_build_lock_name(enum ocfs2_lock_type type,
 
 	BUG_ON(type >= OCFS2_NUM_LOCK_TYPES);
 
-	len = snprintf(name, OCFS2_LOCK_ID_MAX_LEN, "%c%s%016"MLFx64"%08x",
-		       ocfs2_lock_type_char(type), OCFS2_LOCK_ID_PAD, blkno,
-		       generation);
+	len = snprintf(name, OCFS2_LOCK_ID_MAX_LEN, "%c%s%016llx%08x",
+		       ocfs2_lock_type_char(type), OCFS2_LOCK_ID_PAD,
+		       (long long)blkno, generation);
 
 	BUG_ON(len != (OCFS2_LOCK_ID_MAX_LEN - 1));
 
@@ -533,8 +533,8 @@ static void ocfs2_inode_ast_func(void *opaque)
 
 	inode = ocfs2_lock_res_inode(lockres);
 
-	mlog(0, "AST fired for inode %"MLFu64", l_action = %u, type = %s\n",
-	     OCFS2_I(inode)->ip_blkno, lockres->l_action,
+	mlog(0, "AST fired for inode %llu, l_action = %u, type = %s\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, lockres->l_action,
 	     ocfs2_lock_type_string(lockres->l_type));
 
 	BUG_ON(!ocfs2_is_inode_lock(lockres));
@@ -544,8 +544,8 @@ static void ocfs2_inode_ast_func(void *opaque)
 	lksb = &(lockres->l_lksb);
 	if (lksb->status != DLM_NORMAL) {
 		mlog(ML_ERROR, "ocfs2_inode_ast_func: lksb status value of %u "
-		     "on inode %"MLFu64"\n", lksb->status,
-		     OCFS2_I(inode)->ip_blkno);
+		     "on inode %llu\n", lksb->status,
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		spin_unlock_irqrestore(&lockres->l_lock, flags);
 		mlog_exit_void();
 		return;
@@ -646,10 +646,9 @@ static void ocfs2_inode_bast_func(void *opaque, int level)
 	inode = ocfs2_lock_res_inode(lockres);
 	osb = OCFS2_SB(inode->i_sb);
 
-	mlog(0, "BAST fired for inode %"MLFu64", blocking = %d, level = %d "
-	     "type = %s\n", OCFS2_I(inode)->ip_blkno, level,
-	     lockres->l_level,
-	     ocfs2_lock_type_string(lockres->l_type));
+	mlog(0, "BAST fired for inode %llu, blocking %d, level %d type %s\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, level,
+	     lockres->l_level, ocfs2_lock_type_string(lockres->l_type));
 
 	ocfs2_generic_bast_func(osb, lockres, level);
 
@@ -1104,7 +1103,7 @@ int ocfs2_create_new_inode_locks(struct inode *inode)
 
 	mlog_entry_void();
 
-	mlog(0, "Inode %"MLFu64"\n", OCFS2_I(inode)->ip_blkno);
+	mlog(0, "Inode %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	/* NOTE: That we don't increment any of the holder counts, nor
 	 * do we add anything to a journal handle. Since this is
@@ -1149,8 +1148,8 @@ int ocfs2_rw_lock(struct inode *inode, int write)
 
 	mlog_entry_void();
 
-	mlog(0, "inode %"MLFu64" take %s RW lock\n",
-	     OCFS2_I(inode)->ip_blkno,
+	mlog(0, "inode %llu take %s RW lock\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
 	lockres = &OCFS2_I(inode)->ip_rw_lockres;
@@ -1173,8 +1172,8 @@ void ocfs2_rw_unlock(struct inode *inode, int write)
 
 	mlog_entry_void();
 
-	mlog(0, "inode %"MLFu64" drop %s RW lock\n",
-	     OCFS2_I(inode)->ip_blkno,
+	mlog(0, "inode %llu drop %s RW lock\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
 	ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
@@ -1193,8 +1192,8 @@ int ocfs2_data_lock_full(struct inode *inode,
 
 	mlog_entry_void();
 
-	mlog(0, "inode %"MLFu64" take %s DATA lock\n",
-	     OCFS2_I(inode)->ip_blkno,
+	mlog(0, "inode %llu take %s DATA lock\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
 	/* We'll allow faking a readonly data lock for
@@ -1278,8 +1277,8 @@ void ocfs2_data_unlock(struct inode *inode,
 
 	mlog_entry_void();
 
-	mlog(0, "inode %"MLFu64" drop %s DATA lock\n",
-	     OCFS2_I(inode)->ip_blkno,
+	mlog(0, "inode %llu drop %s DATA lock\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     write ? "EXMODE" : "PRMODE");
 
 	if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)))
@@ -1462,9 +1461,9 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 
 	spin_lock(&oi->ip_lock);
 	if (oi->ip_flags & OCFS2_INODE_DELETED) {
-		mlog(0, "Orphaned inode %"MLFu64" was deleted while we "
+		mlog(0, "Orphaned inode %llu was deleted while we "
 		     "were waiting on a lock. ip_flags = 0x%x\n",
-		     oi->ip_blkno, oi->ip_flags);
+		     (unsigned long long)oi->ip_blkno, oi->ip_flags);
 		spin_unlock(&oi->ip_lock);
 		status = -ENOENT;
 		goto bail;
@@ -1485,8 +1484,8 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 	ocfs2_extent_map_trunc(inode, 0);
 
 	if (ocfs2_meta_lvb_is_trustable(lockres)) {
-		mlog(0, "Trusting LVB on inode %"MLFu64"\n",
-		     oi->ip_blkno);
+		mlog(0, "Trusting LVB on inode %llu\n",
+		     (unsigned long long)oi->ip_blkno);
 		ocfs2_refresh_inode_from_lvb(inode);
 	} else {
 		/* Boo, we have to go to disk. */
@@ -1514,15 +1513,16 @@ static int ocfs2_meta_lock_update(struct inode *inode,
 		}
 		mlog_bug_on_msg(inode->i_generation !=
 				le32_to_cpu(fe->i_generation),
-				"Invalid dinode %"MLFu64" disk generation: %u "
+				"Invalid dinode %llu disk generation: %u "
 				"inode->i_generation: %u\n",
-				oi->ip_blkno, le32_to_cpu(fe->i_generation),
+				(unsigned long long)oi->ip_blkno,
+				le32_to_cpu(fe->i_generation),
 				inode->i_generation);
 		mlog_bug_on_msg(le64_to_cpu(fe->i_dtime) ||
 				!(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL)),
-				"Stale dinode %"MLFu64" dtime: %"MLFu64" "
-				"flags: 0x%x\n", oi->ip_blkno,
-				le64_to_cpu(fe->i_dtime),
+				"Stale dinode %llu dtime: %llu flags: 0x%x\n",
+				(unsigned long long)oi->ip_blkno,
+				(unsigned long long)le64_to_cpu(fe->i_dtime),
 				le32_to_cpu(fe->i_flags));
 
 		ocfs2_refresh_inode(inode, fe);
@@ -1581,8 +1581,8 @@ int ocfs2_meta_lock_full(struct inode *inode,
 
 	mlog_entry_void();
 
-	mlog(0, "inode %"MLFu64", take %s META lock\n",
-	     OCFS2_I(inode)->ip_blkno,
+	mlog(0, "inode %llu, take %s META lock\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     ex ? "EXMODE" : "PRMODE");
 
 	status = 0;
@@ -1716,8 +1716,8 @@ void ocfs2_meta_unlock(struct inode *inode,
 
 	mlog_entry_void();
 
-	mlog(0, "inode %"MLFu64" drop %s META lock\n",
-	     OCFS2_I(inode)->ip_blkno,
+	mlog(0, "inode %llu drop %s META lock\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno,
 	     ex ? "EXMODE" : "PRMODE");
 
 	if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)))
@@ -2017,7 +2017,7 @@ out:
 	return ret;
 }
 
-static struct file_operations ocfs2_dlm_debug_fops = {
+static const struct file_operations ocfs2_dlm_debug_fops = {
 	.open =		ocfs2_dlm_debug_open,
 	.release =	ocfs2_dlm_debug_release,
 	.read =		seq_read,
@@ -2686,8 +2686,8 @@ static void ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres,
 	mapping = inode->i_mapping;
 
 	if (filemap_fdatawrite(mapping)) {
-		mlog(ML_ERROR, "Could not sync inode %"MLFu64" for downconvert!",
-		     OCFS2_I(inode)->ip_blkno);
+		mlog(ML_ERROR, "Could not sync inode %llu for downconvert!",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 	}
 	sync_mapping_buffers(mapping);
 	if (blocking == LKM_EXMODE) {
@@ -2717,7 +2717,8 @@ int ocfs2_unblock_data(struct ocfs2_lock_res *lockres,
 	inode = ocfs2_lock_res_inode(lockres);
 	osb = OCFS2_SB(inode->i_sb);
 
-	mlog(0, "unblock inode %"MLFu64"\n", OCFS2_I(inode)->ip_blkno);
+	mlog(0, "unblock inode %llu\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	status = ocfs2_generic_unblock_lock(osb,
 					    lockres,
@@ -2726,8 +2727,8 @@ int ocfs2_unblock_data(struct ocfs2_lock_res *lockres,
 	if (status < 0)
 		mlog_errno(status);
 
-	mlog(0, "inode %"MLFu64", requeue = %d\n",
-	     OCFS2_I(inode)->ip_blkno, *requeue);
+	mlog(0, "inode %llu, requeue = %d\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, *requeue);
 
 	mlog_exit(status);
 	return status;
@@ -2767,14 +2768,15 @@ int ocfs2_unblock_meta(struct ocfs2_lock_res *lockres,
 
        	inode = ocfs2_lock_res_inode(lockres);
 
-	mlog(0, "unblock inode %"MLFu64"\n", OCFS2_I(inode)->ip_blkno);
+	mlog(0, "unblock inode %llu\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	status = ocfs2_do_unblock_meta(inode, requeue);
 	if (status < 0)
 		mlog_errno(status);
 
-	mlog(0, "inode %"MLFu64", requeue = %d\n",
-	     OCFS2_I(inode)->ip_blkno, *requeue);
+	mlog(0, "inode %llu, requeue = %d\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, *requeue);
 
 	mlog_exit(status);
 	return status;
@@ -2893,12 +2895,13 @@ void ocfs2_dump_meta_lvb_info(u64 level,
 	     lockres->l_name, function, line);
 	mlog(level, "version: %u, clusters: %u\n",
 	     be32_to_cpu(lvb->lvb_version), be32_to_cpu(lvb->lvb_iclusters));
-	mlog(level, "size: %"MLFu64", uid %u, gid %u, mode 0x%x\n",
-	     be64_to_cpu(lvb->lvb_isize), be32_to_cpu(lvb->lvb_iuid),
-	     be32_to_cpu(lvb->lvb_igid), be16_to_cpu(lvb->lvb_imode));
-	mlog(level, "nlink %u, atime_packed 0x%"MLFx64", "
-	     "ctime_packed 0x%"MLFx64", mtime_packed 0x%"MLFx64"\n",
-	     be16_to_cpu(lvb->lvb_inlink), be64_to_cpu(lvb->lvb_iatime_packed),
-	     be64_to_cpu(lvb->lvb_ictime_packed),
-	     be64_to_cpu(lvb->lvb_imtime_packed));
+	mlog(level, "size: %llu, uid %u, gid %u, mode 0x%x\n",
+	     (unsigned long long)be64_to_cpu(lvb->lvb_isize),
+	     be32_to_cpu(lvb->lvb_iuid), be32_to_cpu(lvb->lvb_igid),
+	     be16_to_cpu(lvb->lvb_imode));
+	mlog(level, "nlink %u, atime_packed 0x%llx, ctime_packed 0x%llx, "
+	     "mtime_packed 0x%llx\n", be16_to_cpu(lvb->lvb_inlink),
+	     (long long)be64_to_cpu(lvb->lvb_iatime_packed),
+	     (long long)be64_to_cpu(lvb->lvb_ictime_packed),
+	     (long long)be64_to_cpu(lvb->lvb_imtime_packed));
 }
diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c
index 5810160d92a8..ec55ab3c1214 100644
--- a/fs/ocfs2/export.c
+++ b/fs/ocfs2/export.c
@@ -95,8 +95,8 @@ static struct dentry *ocfs2_get_parent(struct dentry *child)
 	mlog_entry("(0x%p, '%.*s')\n", child,
 		   child->d_name.len, child->d_name.name);
 
-	mlog(0, "find parent of directory %"MLFu64"\n",
-	     OCFS2_I(dir)->ip_blkno);
+	mlog(0, "find parent of directory %llu\n",
+	     (unsigned long long)OCFS2_I(dir)->ip_blkno);
 
 	status = ocfs2_meta_lock(dir, NULL, NULL, 0);
 	if (status < 0) {
@@ -115,7 +115,8 @@ static struct dentry *ocfs2_get_parent(struct dentry *child)
 
 	inode = ocfs2_iget(OCFS2_SB(dir->i_sb), blkno);
 	if (IS_ERR(inode)) {
-		mlog(ML_ERROR, "Unable to create inode %"MLFu64"\n", blkno);
+		mlog(ML_ERROR, "Unable to create inode %llu\n",
+		     (unsigned long long)blkno);
 		parent = ERR_PTR(-EACCES);
 		goto bail_unlock;
 	}
@@ -160,8 +161,8 @@ static int ocfs2_encode_fh(struct dentry *dentry, __be32 *fh, int *max_len,
 	blkno = OCFS2_I(inode)->ip_blkno;
 	generation = inode->i_generation;
 
-	mlog(0, "Encoding fh: blkno: %"MLFu64", generation: %u\n",
-	     blkno, generation);
+	mlog(0, "Encoding fh: blkno: %llu, generation: %u\n",
+	     (unsigned long long)blkno, generation);
 
 	len = 3;
 	fh[0] = cpu_to_le32((u32)(blkno >> 32));
@@ -186,8 +187,8 @@ static int ocfs2_encode_fh(struct dentry *dentry, __be32 *fh, int *max_len,
 		len = 6;
 		type = 2;
 
-		mlog(0, "Encoding parent: blkno: %"MLFu64", generation: %u\n",
-		     blkno, generation);
+		mlog(0, "Encoding parent: blkno: %llu, generation: %u\n",
+		     (unsigned long long)blkno, generation);
 	}
 	
 	*max_len = len;
@@ -220,16 +221,17 @@ static struct dentry *ocfs2_decode_fh(struct super_block *sb, __be32 *fh,
 		parent.ih_blkno |= (u64)le32_to_cpu(fh[4]);
 		parent.ih_generation = le32_to_cpu(fh[5]);
 
-		mlog(0, "Decoding parent: blkno: %"MLFu64", generation: %u\n",
-		     parent.ih_blkno, parent.ih_generation);
+		mlog(0, "Decoding parent: blkno: %llu, generation: %u\n",
+		     (unsigned long long)parent.ih_blkno,
+		     parent.ih_generation);
 	}
 
 	handle.ih_blkno = (u64)le32_to_cpu(fh[0]) << 32;
 	handle.ih_blkno |= (u64)le32_to_cpu(fh[1]);
 	handle.ih_generation = le32_to_cpu(fh[2]);
 
-	mlog(0, "Encoding fh: blkno: %"MLFu64", generation: %u\n",
-	     handle.ih_blkno, handle.ih_generation);
+	mlog(0, "Encoding fh: blkno: %llu, generation: %u\n",
+	     (unsigned long long)handle.ih_blkno, handle.ih_generation);
 
 	ret = ocfs2_export_ops.find_exported_dentry(sb, &handle, &parent,
 						    acceptable, context);
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index e6f207eebab4..4601fc256f11 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -182,10 +182,10 @@ static int ocfs2_extent_map_find_leaf(struct inode *inode,
 			if (rec_end > OCFS2_I(inode)->ip_clusters) {
 				mlog_errno(ret);
 				ocfs2_error(inode->i_sb,
-					    "Extent %d at e_blkno %"MLFu64" of inode %"MLFu64" goes past ip_clusters of %u\n",
+					    "Extent %d at e_blkno %llu of inode %llu goes past ip_clusters of %u\n",
 					    i,
-					    le64_to_cpu(rec->e_blkno),
-					    OCFS2_I(inode)->ip_blkno,
+					    (unsigned long long)le64_to_cpu(rec->e_blkno),
+					    (unsigned long long)OCFS2_I(inode)->ip_blkno,
 					    OCFS2_I(inode)->ip_clusters);
 				goto out_free;
 			}
@@ -233,11 +233,11 @@ static int ocfs2_extent_map_find_leaf(struct inode *inode,
 			if (blkno) {
 				mlog_errno(ret);
 				ocfs2_error(inode->i_sb,
-					    "Multiple extents for (cpos = %u, clusters = %u) on inode %"MLFu64"; e_blkno %"MLFu64" and rec %d at e_blkno %"MLFu64"\n",
+					    "Multiple extents for (cpos = %u, clusters = %u) on inode %llu; e_blkno %llu and rec %d at e_blkno %llu\n",
 					    cpos, clusters,
-					    OCFS2_I(inode)->ip_blkno,
-					    blkno, i,
-					    le64_to_cpu(rec->e_blkno));
+					    (unsigned long long)OCFS2_I(inode)->ip_blkno,
+					    (unsigned long long)blkno, i,
+					    (unsigned long long)le64_to_cpu(rec->e_blkno));
 				goto out_free;
 			}
 
@@ -251,9 +251,9 @@ static int ocfs2_extent_map_find_leaf(struct inode *inode,
 		ret = -EBADR;
 		if (!blkno) {
 			ocfs2_error(inode->i_sb,
-				    "No record found for (cpos = %u, clusters = %u) on inode %"MLFu64"\n",
+				    "No record found for (cpos = %u, clusters = %u) on inode %llu\n",
 				    cpos, clusters,
-				    OCFS2_I(inode)->ip_blkno);
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			mlog_errno(ret);
 			goto out_free;
 		}
@@ -288,10 +288,10 @@ static int ocfs2_extent_map_find_leaf(struct inode *inode,
 			ret = -EBADR;
 			mlog_errno(ret);
 			ocfs2_error(inode->i_sb,
-				    "Extent %d at e_blkno %"MLFu64" of inode %"MLFu64" goes past ip_clusters of %u\n",
+				    "Extent %d at e_blkno %llu of inode %llu goes past ip_clusters of %u\n",
 				    i,
-				    le64_to_cpu(rec->e_blkno),
-				    OCFS2_I(inode)->ip_blkno,
+				    (unsigned long long)le64_to_cpu(rec->e_blkno),
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno,
 				    OCFS2_I(inode)->ip_clusters);
 			return ret;
 		}
@@ -557,9 +557,9 @@ static int ocfs2_extent_map_insert(struct inode *inode,
 			ret = -EBADR;
 			mlog_errno(ret);
 			ocfs2_error(inode->i_sb,
-				    "Zero e_clusters on non-tail extent record at e_blkno %"MLFu64" on inode %"MLFu64"\n",
-				    le64_to_cpu(rec->e_blkno),
-				    OCFS2_I(inode)->ip_blkno);
+				    "Zero e_clusters on non-tail extent record at e_blkno %llu on inode %llu\n",
+				    (unsigned long long)le64_to_cpu(rec->e_blkno),
+				    (unsigned long long)OCFS2_I(inode)->ip_blkno);
 			return ret;
 		}
 
@@ -660,10 +660,10 @@ int ocfs2_extent_map_append(struct inode *inode,
 	mlog_bug_on_msg((le32_to_cpu(rec->e_cpos) +
 			 le32_to_cpu(rec->e_clusters)) !=
 			(em->em_clusters + new_clusters),
-			"Inode %"MLFu64":\n"
+			"Inode %llu:\n"
 			"rec->e_cpos = %u + rec->e_clusters = %u = %u\n"
 			"em->em_clusters = %u + new_clusters = %u = %u\n",
-			OCFS2_I(inode)->ip_blkno,
+			(unsigned long long)OCFS2_I(inode)->ip_blkno,
 			le32_to_cpu(rec->e_cpos), le32_to_cpu(rec->e_clusters),
 			le32_to_cpu(rec->e_cpos) + le32_to_cpu(rec->e_clusters),
 			em->em_clusters, new_clusters,
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 8a4048b55fdc..34e903a6a46b 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -220,8 +220,9 @@ static int ocfs2_truncate_file(struct inode *inode,
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_truncate_context *tc = NULL;
 
-	mlog_entry("(inode = %"MLFu64", new_i_size = %"MLFu64"\n",
-		   OCFS2_I(inode)->ip_blkno, new_i_size);
+	mlog_entry("(inode = %llu, new_i_size = %llu\n",
+		   (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		   (unsigned long long)new_i_size);
 
 	truncate_inode_pages(inode->i_mapping, new_i_size);
 
@@ -233,23 +234,26 @@ static int ocfs2_truncate_file(struct inode *inode,
 	}
 
 	mlog_bug_on_msg(le64_to_cpu(fe->i_size) != i_size_read(inode),
-			"Inode %"MLFu64", inode i_size = %lld != di "
-			"i_size = %"MLFu64", i_flags = 0x%x\n",
-			OCFS2_I(inode)->ip_blkno,
+			"Inode %llu, inode i_size = %lld != di "
+			"i_size = %llu, i_flags = 0x%x\n",
+			(unsigned long long)OCFS2_I(inode)->ip_blkno,
 			i_size_read(inode),
-			le64_to_cpu(fe->i_size), le32_to_cpu(fe->i_flags));
+			(unsigned long long)le64_to_cpu(fe->i_size),
+			le32_to_cpu(fe->i_flags));
 
 	if (new_i_size > le64_to_cpu(fe->i_size)) {
-		mlog(0, "asked to truncate file with size (%"MLFu64") "
-		     "to size (%"MLFu64")!\n",
-		     le64_to_cpu(fe->i_size), new_i_size);
+		mlog(0, "asked to truncate file with size (%llu) to size (%llu)!\n",
+		     (unsigned long long)le64_to_cpu(fe->i_size),
+		     (unsigned long long)new_i_size);
 		status = -EINVAL;
 		mlog_errno(status);
 		goto bail;
 	}
 
-	mlog(0, "inode %"MLFu64", i_size = %"MLFu64", new_i_size = %"MLFu64"\n",
-	     le64_to_cpu(fe->i_blkno), le64_to_cpu(fe->i_size), new_i_size);
+	mlog(0, "inode %llu, i_size = %llu, new_i_size = %llu\n",
+	     (unsigned long long)le64_to_cpu(fe->i_blkno),
+	     (unsigned long long)le64_to_cpu(fe->i_size),
+	     (unsigned long long)new_i_size);
 
 	/* lets handle the simple truncate cases before doing any more
 	 * cluster locking. */
@@ -378,8 +382,8 @@ int ocfs2_do_extend_allocation(struct ocfs2_super *osb,
 	}
 
 	block = ocfs2_clusters_to_blocks(osb->sb, bit_off);
-	mlog(0, "Allocating %u clusters at block %u for inode %"MLFu64"\n",
-	     num_bits, bit_off, OCFS2_I(inode)->ip_blkno);
+	mlog(0, "Allocating %u clusters at block %u for inode %llu\n",
+	     num_bits, bit_off, (unsigned long long)OCFS2_I(inode)->ip_blkno);
 	status = ocfs2_insert_extent(osb, handle, inode, fe_bh, block,
 				     num_bits, meta_ac);
 	if (status < 0) {
@@ -449,9 +453,9 @@ static int ocfs2_extend_allocation(struct inode *inode,
 restart_all:
 	BUG_ON(le32_to_cpu(fe->i_clusters) != OCFS2_I(inode)->ip_clusters);
 
-	mlog(0, "extend inode %"MLFu64", i_size = %lld, fe->i_clusters = %u, "
+	mlog(0, "extend inode %llu, i_size = %lld, fe->i_clusters = %u, "
 	     "clusters_to_add = %u\n",
-	     OCFS2_I(inode)->ip_blkno, i_size_read(inode),
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, i_size_read(inode),
 	     fe->i_clusters, clusters_to_add);
 
 	handle = ocfs2_alloc_handle(osb);
@@ -569,8 +573,8 @@ restarted_transaction:
 		}
 	}
 
-	mlog(0, "fe: i_clusters = %u, i_size=%"MLFu64"\n",
-	     fe->i_clusters, fe->i_size);
+	mlog(0, "fe: i_clusters = %u, i_size=%llu\n",
+	     fe->i_clusters, (unsigned long long)fe->i_size);
 	mlog(0, "inode: ip_clusters=%u, i_size=%lld\n",
 	     OCFS2_I(inode)->ip_clusters, i_size_read(inode));
 
@@ -865,8 +869,8 @@ static int ocfs2_write_remove_suid(struct inode *inode)
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	struct ocfs2_dinode *di;
 
-	mlog_entry("(Inode %"MLFu64", mode 0%o)\n", oi->ip_blkno,
-		   inode->i_mode);
+	mlog_entry("(Inode %llu, mode 0%o)\n",
+		   (unsigned long long)oi->ip_blkno, inode->i_mode);
 
 	handle = ocfs2_start_trans(osb, NULL, OCFS2_INODE_UPDATE_CREDITS);
 	if (handle == NULL) {
@@ -1172,7 +1176,7 @@ struct inode_operations ocfs2_special_file_iops = {
 	.getattr	= ocfs2_getattr,
 };
 
-struct file_operations ocfs2_fops = {
+const struct file_operations ocfs2_fops = {
 	.read		= do_sync_read,
 	.write		= do_sync_write,
 	.sendfile	= generic_file_sendfile,
@@ -1184,7 +1188,7 @@ struct file_operations ocfs2_fops = {
 	.aio_write	= ocfs2_file_aio_write,
 };
 
-struct file_operations ocfs2_dops = {
+const struct file_operations ocfs2_dops = {
 	.read		= generic_read_dir,
 	.readdir	= ocfs2_readdir,
 	.fsync		= ocfs2_sync_file,
diff --git a/fs/ocfs2/file.h b/fs/ocfs2/file.h
index a5ea33b24060..740c9e7ca599 100644
--- a/fs/ocfs2/file.h
+++ b/fs/ocfs2/file.h
@@ -26,8 +26,8 @@
 #ifndef OCFS2_FILE_H
 #define OCFS2_FILE_H
 
-extern struct file_operations ocfs2_fops;
-extern struct file_operations ocfs2_dops;
+extern const struct file_operations ocfs2_fops;
+extern const struct file_operations ocfs2_dops;
 extern struct inode_operations ocfs2_file_iops;
 extern struct inode_operations ocfs2_special_file_iops;
 struct ocfs2_alloc_context;
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 315472a5c192..327a5b7b86ed 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -95,7 +95,7 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno)
 	struct super_block *sb = osb->sb;
 	struct ocfs2_find_inode_args args;
 
-	mlog_entry("(blkno = %"MLFu64")\n", blkno);
+	mlog_entry("(blkno = %llu)\n", (unsigned long long)blkno);
 
 	/* Ok. By now we've either got the offsets passed to us by the
 	 * caller, or we just pulled them off the bh. Lets do some
@@ -134,8 +134,8 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno)
 
 bail:
 	if (!IS_ERR(inode)) {
-		mlog(0, "returning inode with number %"MLFu64"\n",
-		     OCFS2_I(inode)->ip_blkno);
+		mlog(0, "returning inode with number %llu\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		mlog_exit_ptr(inode);
 	} else
 		mlog_errno(PTR_ERR(inode));
@@ -219,7 +219,8 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 	struct ocfs2_super *osb;
 	int status = -EINVAL;
 
-	mlog_entry("(0x%p, size:%"MLFu64")\n", inode, fe->i_size);
+	mlog_entry("(0x%p, size:%llu)\n", inode,
+		   (unsigned long long)fe->i_size);
 
 	sb = inode->i_sb;
 	osb = OCFS2_SB(sb);
@@ -228,9 +229,10 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 	 * today.  change if needed. */
 	if (!OCFS2_IS_VALID_DINODE(fe) ||
 	    !(fe->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
-		mlog(ML_ERROR, "Invalid dinode: i_ino=%lu, i_blkno=%"MLFu64", "
+		mlog(ML_ERROR, "Invalid dinode: i_ino=%lu, i_blkno=%llu, "
 		     "signature = %.*s, flags = 0x%x\n",
-		     inode->i_ino, le64_to_cpu(fe->i_blkno), 7,
+		     inode->i_ino,
+		     (unsigned long long)le64_to_cpu(fe->i_blkno), 7,
 		     fe->i_signature, le32_to_cpu(fe->i_flags));
 		goto bail;
 	}
@@ -268,8 +270,9 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 
 	if (OCFS2_I(inode)->ip_blkno != le64_to_cpu(fe->i_blkno))
 		mlog(ML_ERROR,
-		     "ip_blkno %"MLFu64" != i_blkno %"MLFu64"!\n",
-		     OCFS2_I(inode)->ip_blkno, fe->i_blkno);
+		     "ip_blkno %llu != i_blkno %llu!\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		     (unsigned long long)fe->i_blkno);
 
 	OCFS2_I(inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
 	OCFS2_I(inode)->ip_orphaned_slot = OCFS2_INVALID_SLOT;
@@ -278,8 +281,8 @@ int ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
 		inode->i_ino = ino_from_blkno(inode->i_sb,
 			       le64_to_cpu(fe->i_blkno));
 
-	mlog(0, "blkno = %"MLFu64", ino = %lu, create_ino = %s\n",
-	     fe->i_blkno, inode->i_ino, create_ino ? "true" : "false");
+	mlog(0, "blkno = %llu, ino = %lu, create_ino = %s\n",
+	     (unsigned long long)fe->i_blkno, inode->i_ino, create_ino ? "true" : "false");
 
 	inode->i_nlink = le16_to_cpu(fe->i_links_count);
 
@@ -371,8 +374,8 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 
 	fe = (struct ocfs2_dinode *) bh->b_data;
 	if (!OCFS2_IS_VALID_DINODE(fe)) {
-		mlog(ML_ERROR, "Invalid dinode #%"MLFu64": signature = %.*s\n",
-		     fe->i_blkno, 7, fe->i_signature);
+		mlog(ML_ERROR, "Invalid dinode #%llu: signature = %.*s\n",
+		     (unsigned long long)fe->i_blkno, 7, fe->i_signature);
 		make_bad_inode(inode);
 		goto bail;
 	}
@@ -386,8 +389,8 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 
 	status = -EINVAL;
 	if (ocfs2_populate_inode(inode, fe, 0) < 0) {
-		mlog(ML_ERROR, "populate inode failed! i_blkno=%"MLFu64", "
-		     "i_ino=%lu\n", fe->i_blkno, inode->i_ino);
+		mlog(ML_ERROR, "populate failed! i_blkno=%llu, i_ino=%lu\n",
+		     (unsigned long long)fe->i_blkno, inode->i_ino);
 		make_bad_inode(inode);
 		goto bail;
 	}
@@ -675,8 +678,8 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode)
 	 * never get here as system file inodes should always have a
 	 * positive link count. */
 	if (oi->ip_flags & OCFS2_INODE_SYSTEM_FILE) {
-		mlog(ML_ERROR, "Skipping delete of system file %"MLFu64".\n",
-		     oi->ip_blkno);
+		mlog(ML_ERROR, "Skipping delete of system file %llu\n",
+		     (unsigned long long)oi->ip_blkno);
 		goto bail_unlock;
 	}
 
@@ -715,16 +718,16 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
 	 * ocfs2_delete_inode, another node might have asked to delete
 	 * the inode. Recheck our flags to catch this. */
 	if (!ocfs2_inode_is_valid_to_delete(inode)) {
-		mlog(0, "Skipping delete of %"MLFu64" because flags changed\n",
-		     oi->ip_blkno);
+		mlog(0, "Skipping delete of %llu because flags changed\n",
+		     (unsigned long long)oi->ip_blkno);
 		goto bail;
 	}
 
 	/* Now that we have an up to date inode, we can double check
 	 * the link count. */
 	if (inode->i_nlink) {
-		mlog(0, "Skipping delete of %"MLFu64" because nlink = %u\n",
-		     oi->ip_blkno, inode->i_nlink);
+		mlog(0, "Skipping delete of %llu because nlink = %u\n",
+		     (unsigned long long)oi->ip_blkno, inode->i_nlink);
 		goto bail;
 	}
 
@@ -734,9 +737,11 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
 		/* for lack of a better error? */
 		status = -EEXIST;
 		mlog(ML_ERROR,
-		     "Inode %"MLFu64" (on-disk %"MLFu64") not orphaned! "
+		     "Inode %llu (on-disk %llu) not orphaned! "
 		     "Disk flags  0x%x, inode flags 0x%x\n",
-		     oi->ip_blkno, di->i_blkno, di->i_flags, oi->ip_flags);
+		     (unsigned long long)oi->ip_blkno,
+		     (unsigned long long)di->i_blkno, di->i_flags,
+		     oi->ip_flags);
 		goto bail;
 	}
 
@@ -753,8 +758,8 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
 	 * disk and let them worry about deleting it. */
 	if (status == -EBUSY) {
 		status = 0;
-		mlog(0, "Skipping delete of %"MLFu64" because it is in use on"
-		     "other nodes\n", oi->ip_blkno);
+		mlog(0, "Skipping delete of %llu because it is in use on"
+		     "other nodes\n", (unsigned long long)oi->ip_blkno);
 		goto bail;
 	}
 	if (status < 0) {
@@ -768,13 +773,13 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
 		 * into. This may happen during node death and
 		 * recovery knows how to clean it up so we can safely
 		 * ignore this inode for now on. */
-		mlog(0, "Nobody knew where inode %"MLFu64" was orphaned!\n",
-		     oi->ip_blkno);
+		mlog(0, "Nobody knew where inode %llu was orphaned!\n",
+		     (unsigned long long)oi->ip_blkno);
 	} else {
 		*wipe = 1;
 
-		mlog(0, "Inode %"MLFu64" is ok to wipe from orphan dir %d\n",
-		     oi->ip_blkno, oi->ip_orphaned_slot);
+		mlog(0, "Inode %llu is ok to wipe from orphan dir %d\n",
+		     (unsigned long long)oi->ip_blkno, oi->ip_orphaned_slot);
 	}
 	spin_unlock(&oi->ip_lock);
 
@@ -788,8 +793,8 @@ bail:
 static void ocfs2_cleanup_delete_inode(struct inode *inode,
 				       int sync_data)
 {
-	mlog(0, "Cleanup inode %"MLFu64", sync = %d\n",
-	     OCFS2_I(inode)->ip_blkno, sync_data);
+	mlog(0, "Cleanup inode %llu, sync = %d\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data);
 	if (sync_data)
 		write_inode_now(inode, 1);
 	truncate_inode_pages(&inode->i_data, 0);
@@ -897,8 +902,8 @@ void ocfs2_clear_inode(struct inode *inode)
 	if (!inode)
 		goto bail;
 
-	mlog(0, "Clearing inode: %"MLFu64", nlink = %u\n",
-	     OCFS2_I(inode)->ip_blkno, inode->i_nlink);
+	mlog(0, "Clearing inode: %llu, nlink = %u\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_nlink);
 
 	mlog_bug_on_msg(OCFS2_SB(inode->i_sb) == NULL,
 			"Inode=%lu\n", inode->i_ino);
@@ -919,8 +924,8 @@ void ocfs2_clear_inode(struct inode *inode)
 		ocfs2_checkpoint_inode(inode);
 
 	mlog_bug_on_msg(!list_empty(&oi->ip_io_markers),
-			"Clear inode of %"MLFu64", inode has io markers\n",
-			oi->ip_blkno);
+			"Clear inode of %llu, inode has io markers\n",
+			(unsigned long long)oi->ip_blkno);
 
 	ocfs2_extent_map_drop(inode, 0);
 	ocfs2_extent_map_init(inode);
@@ -936,20 +941,20 @@ void ocfs2_clear_inode(struct inode *inode)
 	ocfs2_metadata_cache_purge(inode);
 
 	mlog_bug_on_msg(oi->ip_metadata_cache.ci_num_cached,
-			"Clear inode of %"MLFu64", inode has %u cache items\n",
-			oi->ip_blkno, oi->ip_metadata_cache.ci_num_cached);
+			"Clear inode of %llu, inode has %u cache items\n",
+			(unsigned long long)oi->ip_blkno, oi->ip_metadata_cache.ci_num_cached);
 
 	mlog_bug_on_msg(!(oi->ip_flags & OCFS2_INODE_CACHE_INLINE),
-			"Clear inode of %"MLFu64", inode has a bad flag\n",
-			oi->ip_blkno);
+			"Clear inode of %llu, inode has a bad flag\n",
+			(unsigned long long)oi->ip_blkno);
 
 	mlog_bug_on_msg(spin_is_locked(&oi->ip_lock),
-			"Clear inode of %"MLFu64", inode is locked\n",
-			oi->ip_blkno);
+			"Clear inode of %llu, inode is locked\n",
+			(unsigned long long)oi->ip_blkno);
 
 	mlog_bug_on_msg(!mutex_trylock(&oi->ip_io_mutex),
-			"Clear inode of %"MLFu64", io_mutex is locked\n",
-			oi->ip_blkno);
+			"Clear inode of %llu, io_mutex is locked\n",
+			(unsigned long long)oi->ip_blkno);
 	mutex_unlock(&oi->ip_io_mutex);
 
 	/*
@@ -957,19 +962,19 @@ void ocfs2_clear_inode(struct inode *inode)
 	 * kernel 1, world 0
 	 */
 	mlog_bug_on_msg(!down_write_trylock(&oi->ip_alloc_sem),
-			"Clear inode of %"MLFu64", alloc_sem is locked\n",
-			oi->ip_blkno);
+			"Clear inode of %llu, alloc_sem is locked\n",
+			(unsigned long long)oi->ip_blkno);
 	up_write(&oi->ip_alloc_sem);
 
 	mlog_bug_on_msg(oi->ip_open_count,
-			"Clear inode of %"MLFu64" has open count %d\n",
-			oi->ip_blkno, oi->ip_open_count);
+			"Clear inode of %llu has open count %d\n",
+			(unsigned long long)oi->ip_blkno, oi->ip_open_count);
 	mlog_bug_on_msg(!list_empty(&oi->ip_handle_list),
-			"Clear inode of %"MLFu64" has non empty handle list\n",
-			oi->ip_blkno);
+			"Clear inode of %llu has non empty handle list\n",
+			(unsigned long long)oi->ip_blkno);
 	mlog_bug_on_msg(oi->ip_handle,
-			"Clear inode of %"MLFu64" has non empty handle pointer\n",
-			oi->ip_blkno);
+			"Clear inode of %llu has non empty handle pointer\n",
+			(unsigned long long)oi->ip_blkno);
 
 	/* Clear all other flags. */
 	oi->ip_flags = OCFS2_INODE_CACHE_INLINE;
@@ -991,8 +996,8 @@ void ocfs2_drop_inode(struct inode *inode)
 
 	mlog_entry_void();
 
-	mlog(0, "Drop inode %"MLFu64", nlink = %u, ip_flags = 0x%x\n",
-	     oi->ip_blkno, inode->i_nlink, oi->ip_flags);
+	mlog(0, "Drop inode %llu, nlink = %u, ip_flags = 0x%x\n",
+	     (unsigned long long)oi->ip_blkno, inode->i_nlink, oi->ip_flags);
 
 	/* Testing ip_orphaned_slot here wouldn't work because we may
 	 * not have gotten a delete_inode vote from any other nodes
@@ -1069,8 +1074,8 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
 	struct inode *inode = dentry->d_inode;
 	int status = 0;
 
-	mlog_entry("(inode = 0x%p, ino = %"MLFu64")\n", inode,
-		   inode ? OCFS2_I(inode)->ip_blkno : 0ULL);
+	mlog_entry("(inode = 0x%p, ino = %llu)\n", inode,
+		   inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL);
 
 	if (!inode) {
 		mlog(0, "eep, no inode!\n");
@@ -1114,7 +1119,8 @@ int ocfs2_mark_inode_dirty(struct ocfs2_journal_handle *handle,
 	int status;
 	struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
 
-	mlog_entry("(inode %"MLFu64")\n", OCFS2_I(inode)->ip_blkno);
+	mlog_entry("(inode %llu)\n",
+		   (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	status = ocfs2_journal_access(handle, inode, bh,
 				      OCFS2_JOURNAL_ACCESS_WRITE);
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 4be801f4559b..6a610ae53583 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -377,7 +377,7 @@ int ocfs2_journal_access(struct ocfs2_journal_handle *handle,
 	BUG_ON(!bh);
 	BUG_ON(!(handle->flags & OCFS2_HANDLE_STARTED));
 
-	mlog_entry("bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %hu\n",
+	mlog_entry("bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
 		   (unsigned long long)bh->b_blocknr, type,
 		   (type == OCFS2_JOURNAL_ACCESS_CREATE) ?
 		   "OCFS2_JOURNAL_ACCESS_CREATE" :
@@ -503,8 +503,8 @@ static void ocfs2_handle_cleanup_locks(struct ocfs2_journal *journal,
 		ocfs2_meta_unlock(inode, 1);
 		if (atomic_read(&inode->i_count) == 1)
 			mlog(ML_ERROR,
-			     "Inode %"MLFu64", I'm doing a last iput for!",
-			     OCFS2_I(inode)->ip_blkno);
+			     "Inode %llu, I'm doing a last iput for!",
+			     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		iput(inode);
 		kmem_cache_free(ocfs2_lock_cache, lock);
 	}
@@ -582,7 +582,8 @@ int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
 	}
 
 	mlog(0, "inode->i_size = %lld\n", inode->i_size);
-	mlog(0, "inode->i_blocks = %lu\n", inode->i_blocks);
+	mlog(0, "inode->i_blocks = %llu\n",
+			(unsigned long long)inode->i_blocks);
 	mlog(0, "inode->ip_clusters = %u\n", OCFS2_I(inode)->ip_clusters);
 
 	/* call the kernels journal init function now */
@@ -640,8 +641,9 @@ static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
 		/* This is called from startup/shutdown which will
 		 * handle the errors in a specific manner, so no need
 		 * to call ocfs2_error() here. */
-		mlog(ML_ERROR, "Journal dinode %"MLFu64"  has invalid "
-		     "signature: %.*s", fe->i_blkno, 7, fe->i_signature);
+		mlog(ML_ERROR, "Journal dinode %llu  has invalid "
+		     "signature: %.*s", (unsigned long long)fe->i_blkno, 7,
+		     fe->i_signature);
 		status = -EIO;
 		goto out;
 	}
@@ -849,8 +851,9 @@ static int ocfs2_force_read_journal(struct inode *inode)
 
 	memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL);
 
-	mlog(0, "Force reading %lu blocks\n",
-	     (inode->i_blocks >> (inode->i_sb->s_blocksize_bits - 9)));
+	mlog(0, "Force reading %llu blocks\n",
+		(unsigned long long)(inode->i_blocks >>
+			(inode->i_sb->s_blocksize_bits - 9)));
 
 	v_blkno = 0;
 	while (v_blkno <
@@ -934,8 +937,8 @@ void ocfs2_complete_recovery(void *data)
 
 		la_dinode = item->lri_la_dinode;
 		if (la_dinode) {
-			mlog(0, "Clean up local alloc %"MLFu64"\n",
-			     la_dinode->i_blkno);
+			mlog(0, "Clean up local alloc %llu\n",
+			     (unsigned long long)la_dinode->i_blkno);
 
 			ret = ocfs2_complete_local_alloc_recovery(osb,
 								  la_dinode);
@@ -947,8 +950,8 @@ void ocfs2_complete_recovery(void *data)
 
 		tl_dinode = item->lri_tl_dinode;
 		if (tl_dinode) {
-			mlog(0, "Clean up truncate log %"MLFu64"\n",
-			     tl_dinode->i_blkno);
+			mlog(0, "Clean up truncate log %llu\n",
+			     (unsigned long long)tl_dinode->i_blkno);
 
 			ret = ocfs2_complete_truncate_log_recovery(osb,
 								   tl_dinode);
@@ -1473,11 +1476,11 @@ static int ocfs2_queue_orphans(struct ocfs2_super *osb,
 			if (de->file_type > OCFS2_FT_MAX) {
 				mlog(ML_ERROR,
 				     "block %llu contains invalid de: "
-				     "inode = %"MLFu64", rec_len = %u, "
+				     "inode = %llu, rec_len = %u, "
 				     "name_len = %u, file_type = %u, "
 				     "name='%.*s'\n",
 				     (unsigned long long)bh->b_blocknr,
-				     le64_to_cpu(de->inode),
+				     (unsigned long long)le64_to_cpu(de->inode),
 				     le16_to_cpu(de->rec_len),
 				     de->name_len,
 				     de->file_type,
@@ -1494,8 +1497,8 @@ static int ocfs2_queue_orphans(struct ocfs2_super *osb,
 			if (IS_ERR(iter))
 				continue;
 
-			mlog(0, "queue orphan %"MLFu64"\n",
-			     OCFS2_I(iter)->ip_blkno);
+			mlog(0, "queue orphan %llu\n",
+			     (unsigned long long)OCFS2_I(iter)->ip_blkno);
 			/* No locking is required for the next_orphan
 			 * queue as there is only ever a single
 			 * process doing orphan recovery. */
@@ -1588,7 +1591,7 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb,
 
 	while (inode) {
 		oi = OCFS2_I(inode);
-		mlog(0, "iput orphan %"MLFu64"\n", oi->ip_blkno);
+		mlog(0, "iput orphan %llu\n", (unsigned long long)oi->ip_blkno);
 
 		iter = oi->ip_next_orphan;
 
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index 149b35181666..0d1973ea32b0 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -143,8 +143,8 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb)
 
 	if (!(le32_to_cpu(alloc->i_flags) &
 	    (OCFS2_LOCAL_ALLOC_FL|OCFS2_BITMAP_FL))) {
-		mlog(ML_ERROR, "Invalid local alloc inode, %"MLFu64"\n",
-		     OCFS2_I(inode)->ip_blkno);
+		mlog(ML_ERROR, "Invalid local alloc inode, %llu\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		status = -EINVAL;
 		goto bail;
 	}
@@ -493,9 +493,9 @@ int ocfs2_reserve_local_alloc_bits(struct ocfs2_super *osb,
 
 	if (le32_to_cpu(alloc->id1.bitmap1.i_used) !=
 	    ocfs2_local_alloc_count_bits(alloc)) {
-		ocfs2_error(osb->sb, "local alloc inode %"MLFu64" says it has "
+		ocfs2_error(osb->sb, "local alloc inode %llu says it has "
 			    "%u free bits, but a count shows %u",
-			    le64_to_cpu(alloc->i_blkno),
+			    (unsigned long long)le64_to_cpu(alloc->i_blkno),
 			    le32_to_cpu(alloc->id1.bitmap1.i_used),
 			    ocfs2_local_alloc_count_bits(alloc));
 		status = -EIO;
@@ -753,10 +753,11 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
 				ocfs2_clusters_to_blocks(osb->sb,
 							 start - count);
 
-			mlog(0, "freeing %u bits starting at local "
-			     "alloc bit %u (la_start_blk = %"MLFu64", "
-			     "blkno = %"MLFu64")\n", count, start - count,
-			     la_start_blk, blkno);
+			mlog(0, "freeing %u bits starting at local alloc bit "
+			     "%u (la_start_blk = %llu, blkno = %llu)\n",
+			     count, start - count,
+			     (unsigned long long)la_start_blk,
+			     (unsigned long long)blkno);
 
 			status = ocfs2_free_clusters(handle, main_bm_inode,
 						     main_bm_bh, blkno, count);
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index f6b77ff1d2bf..0673862c8bdd 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -161,8 +161,8 @@ static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry,
 		goto bail;
 	}
 
-	mlog(0, "find name %.*s in directory %"MLFu64"\n", dentry->d_name.len,
-	     dentry->d_name.name, OCFS2_I(dir)->ip_blkno);
+	mlog(0, "find name %.*s in directory %llu\n", dentry->d_name.len,
+	     dentry->d_name.name, (unsigned long long)OCFS2_I(dir)->ip_blkno);
 
 	status = ocfs2_meta_lock(dir, NULL, NULL, 0);
 	if (status < 0) {
@@ -180,7 +180,8 @@ static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry,
 
 	inode = ocfs2_iget(OCFS2_SB(dir->i_sb), blkno);
 	if (IS_ERR(inode)) {
-		mlog(ML_ERROR, "Unable to create inode %"MLFu64"\n", blkno);
+		mlog(ML_ERROR, "Unable to create inode %llu\n",
+		     (unsigned long long)blkno);
 		ret = ERR_PTR(-EACCES);
 		goto bail_unlock;
 	}
@@ -310,8 +311,8 @@ static int ocfs2_mknod(struct inode *dir,
 	osb = OCFS2_SB(dir->i_sb);
 
 	if (S_ISDIR(mode) && (dir->i_nlink >= OCFS2_LINK_MAX)) {
-		mlog(ML_ERROR, "inode %"MLFu64" has i_nlink of %u\n",
-		     OCFS2_I(dir)->ip_blkno, dir->i_nlink);
+		mlog(ML_ERROR, "inode %llu has i_nlink of %u\n",
+		     (unsigned long long)OCFS2_I(dir)->ip_blkno, dir->i_nlink);
 		status = -EMLINK;
 		goto leave;
 	}
@@ -562,9 +563,9 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb,
 
 	if (ocfs2_populate_inode(inode, fe, 1) < 0) {
 		mlog(ML_ERROR, "populate inode failed! bh->b_blocknr=%llu, "
-		     "i_blkno=%"MLFu64", i_ino=%lu\n",
+		     "i_blkno=%llu, i_ino=%lu\n",
 		     (unsigned long long) (*new_fe_bh)->b_blocknr,
-		     fe->i_blkno, inode->i_ino);
+		     (unsigned long long)fe->i_blkno, inode->i_ino);
 		BUG();
 	}
 
@@ -765,7 +766,7 @@ static int ocfs2_unlink(struct inode *dir,
 
 	BUG_ON(dentry->d_parent->d_inode != dir);
 
-	mlog(0, "ino = %"MLFu64"\n", OCFS2_I(inode)->ip_blkno);
+	mlog(0, "ino = %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 	if (inode == osb->root_inode) {
 		mlog(0, "Cannot delete the root directory\n");
@@ -799,9 +800,9 @@ static int ocfs2_unlink(struct inode *dir,
 	if (OCFS2_I(inode)->ip_blkno != blkno) {
 		status = -ENOENT;
 
-		mlog(0, "ip_blkno (%"MLFu64") != dirent blkno (%"MLFu64") "
-		     "ip_flags = %x\n", OCFS2_I(inode)->ip_blkno, blkno,
-		     OCFS2_I(inode)->ip_flags);
+		mlog(0, "ip_blkno %llu != dirent blkno %llu ip_flags = %x\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		     (unsigned long long)blkno, OCFS2_I(inode)->ip_flags);
 		goto leave;
 	}
 
@@ -946,8 +947,9 @@ static int ocfs2_double_lock(struct ocfs2_super *osb,
 	struct buffer_head **tmpbh;
 	struct inode *tmpinode;
 
-	mlog_entry("(inode1 = %"MLFu64", inode2 = %"MLFu64")\n",
-		   oi1->ip_blkno, oi2->ip_blkno);
+	mlog_entry("(inode1 = %llu, inode2 = %llu)\n",
+		   (unsigned long long)oi1->ip_blkno,
+		   (unsigned long long)oi2->ip_blkno);
 
 	BUG_ON(!handle);
 
@@ -1187,9 +1189,9 @@ static int ocfs2_rename(struct inode *old_dir,
 		if (OCFS2_I(new_inode)->ip_blkno != newfe_blkno) {
 			status = -EACCES;
 
-			mlog(0, "Inode blkno (%"MLFu64") and dir (%"MLFu64") "
-			     "disagree. ip_flags = %x\n",
-			     OCFS2_I(new_inode)->ip_blkno, newfe_blkno,
+			mlog(0, "Inode %llu and dir %llu disagree. flags = %x\n",
+			     (unsigned long long)OCFS2_I(new_inode)->ip_blkno,
+			     (unsigned long long)newfe_blkno,
 			     OCFS2_I(new_inode)->ip_flags);
 			goto bail;
 		}
@@ -1215,9 +1217,9 @@ static int ocfs2_rename(struct inode *old_dir,
 
 		newfe = (struct ocfs2_dinode *) newfe_bh->b_data;
 
-		mlog(0, "aha rename over existing... new_de=%p "
-		     "new_blkno=%"MLFu64" newfebh=%p bhblocknr=%llu\n",
-		     new_de, newfe_blkno, newfe_bh, newfe_bh ?
+		mlog(0, "aha rename over existing... new_de=%p new_blkno=%llu "
+		     "newfebh=%p bhblocknr=%llu\n", new_de,
+		     (unsigned long long)newfe_blkno, newfe_bh, newfe_bh ?
 		     (unsigned long long)newfe_bh->b_blocknr : 0ULL);
 
 		if (S_ISDIR(new_inode->i_mode) || (new_inode->i_nlink == 1)) {
@@ -1354,8 +1356,8 @@ static int ocfs2_rename(struct inode *old_dir,
 		if (new_dir_nlink != new_dir->i_nlink) {
 			if (!new_dir_bh) {
 				mlog(ML_ERROR, "need to change nlink for new "
-				     "dir %"MLFu64" from %d to %d but bh is "
-				     "NULL\n", OCFS2_I(new_dir)->ip_blkno,
+				     "dir %llu from %d to %d but bh is NULL\n",
+				     (unsigned long long)OCFS2_I(new_dir)->ip_blkno,
 				     (int)new_dir_nlink, new_dir->i_nlink);
 			} else {
 				struct ocfs2_dinode *fe;
@@ -1372,10 +1374,9 @@ static int ocfs2_rename(struct inode *old_dir,
 	if (old_dir_nlink != old_dir->i_nlink) {
 		if (!old_dir_bh) {
 			mlog(ML_ERROR, "need to change nlink for old dir "
-			     "%"MLFu64" from %d to %d but bh is NULL!\n",
-			     OCFS2_I(old_dir)->ip_blkno,
-			     (int)old_dir_nlink,
-			     old_dir->i_nlink);
+			     "%llu from %d to %d but bh is NULL!\n",
+			     (unsigned long long)OCFS2_I(old_dir)->ip_blkno,
+			     (int)old_dir_nlink, old_dir->i_nlink);
 		} else {
 			struct ocfs2_dinode *fe;
 			status = ocfs2_journal_access(handle, old_dir,
@@ -1443,8 +1444,9 @@ static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
 	 * write i_size + 1 bytes. */
 	blocks = (bytes_left + sb->s_blocksize - 1) >> sb->s_blocksize_bits;
 
-	mlog_entry("i_blocks = %lu, i_size = %llu, blocks = %d\n",
-		       inode->i_blocks, i_size_read(inode), blocks);
+	mlog_entry("i_blocks = %llu, i_size = %llu, blocks = %d\n",
+			(unsigned long long)inode->i_blocks,
+			i_size_read(inode), blocks);
 
 	/* Sanity check -- make sure we're going to fit. */
 	if (bytes_left >
@@ -1634,9 +1636,9 @@ static int ocfs2_symlink(struct inode *dir,
 						    NULL);
 		if (status < 0) {
 			if (status != -ENOSPC && status != -EINTR) {
-				mlog(ML_ERROR, "Failed to extend file to "
-					       "%"MLFu64"\n",
-				     newsize);
+				mlog(ML_ERROR,
+				     "Failed to extend file to %llu\n",
+				     (unsigned long long)newsize);
 				mlog_errno(status);
 				status = -ENOSPC;
 			}
@@ -1716,10 +1718,11 @@ int ocfs2_check_dir_entry(struct inode * dir,
 		error_msg = "directory entry across blocks";
 
 	if (error_msg != NULL)
-		mlog(ML_ERROR, "bad entry in directory #%"MLFu64": %s - "
-		     "offset=%lu, inode=%"MLFu64", rec_len=%d, name_len=%d\n",
-		     OCFS2_I(dir)->ip_blkno, error_msg, offset,
-		     le64_to_cpu(de->inode), rlen, de->name_len);
+		mlog(ML_ERROR, "bad entry in directory #%llu: %s - "
+		     "offset=%lu, inode=%llu, rec_len=%d, name_len=%d\n",
+		     (unsigned long long)OCFS2_I(dir)->ip_blkno, error_msg,
+		     offset, (unsigned long long)le64_to_cpu(de->inode), rlen,
+		     de->name_len);
 	return error_msg == NULL ? 1 : 0;
 }
 
@@ -2021,8 +2024,8 @@ static int ocfs2_blkno_stringify(u64 blkno, char *name)
 
 	mlog_entry_void();
 
-	namelen = snprintf(name, OCFS2_ORPHAN_NAMELEN + 1, "%016"MLFx64,
-			   blkno);
+	namelen = snprintf(name, OCFS2_ORPHAN_NAMELEN + 1, "%016llx",
+			   (long long)blkno);
 	if (namelen <= 0) {
 		if (namelen)
 			status = namelen;
@@ -2167,8 +2170,8 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
 	OCFS2_I(inode)->ip_orphaned_slot = osb->slot_num;
 	spin_unlock(&OCFS2_I(inode)->ip_lock);
 
-	mlog(0, "Inode %"MLFu64" orphaned in slot %d\n",
-	     OCFS2_I(inode)->ip_blkno, osb->slot_num);
+	mlog(0, "Inode %llu orphaned in slot %d\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, osb->slot_num);
 
 leave:
 	if (orphan_dir_inode)
@@ -2202,8 +2205,9 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
 		goto leave;
 	}
 
-	mlog(0, "removing '%s' from orphan dir %"MLFu64" (namelen=%d)\n",
-	     name, OCFS2_I(orphan_dir_inode)->ip_blkno, OCFS2_ORPHAN_NAMELEN);
+	mlog(0, "removing '%s' from orphan dir %llu (namelen=%d)\n",
+	     name, (unsigned long long)OCFS2_I(orphan_dir_inode)->ip_blkno,
+	     OCFS2_ORPHAN_NAMELEN);
 
 	/* find it's spot in the orphan directory */
 	target_de_bh = ocfs2_find_entry(name, OCFS2_ORPHAN_NAMELEN,
diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h
index e89de9b6e491..da1093039c01 100644
--- a/fs/ocfs2/ocfs2.h
+++ b/fs/ocfs2/ocfs2.h
@@ -357,8 +357,8 @@ static inline int ocfs2_is_soft_readonly(struct ocfs2_super *osb)
 #define OCFS2_RO_ON_INVALID_DINODE(__sb, __di)	do {			\
 	typeof(__di) ____di = (__di);					\
 	ocfs2_error((__sb), 						\
-		"Dinode # %"MLFu64" has bad signature %.*s",		\
-		(____di)->i_blkno, 7,					\
+		"Dinode # %llu has bad signature %.*s",			\
+		(unsigned long long)(____di)->i_blkno, 7,		\
 		(____di)->i_signature);					\
 } while (0);
 
@@ -368,8 +368,8 @@ static inline int ocfs2_is_soft_readonly(struct ocfs2_super *osb)
 #define OCFS2_RO_ON_INVALID_EXTENT_BLOCK(__sb, __eb)	do {		\
 	typeof(__eb) ____eb = (__eb);					\
 	ocfs2_error((__sb), 						\
-		"Extent Block # %"MLFu64" has bad signature %.*s",	\
-		(____eb)->h_blkno, 7,					\
+		"Extent Block # %llu has bad signature %.*s",		\
+		(unsigned long long)(____eb)->h_blkno, 7,		\
 		(____eb)->h_signature);					\
 } while (0);
 
@@ -379,8 +379,8 @@ static inline int ocfs2_is_soft_readonly(struct ocfs2_super *osb)
 #define OCFS2_RO_ON_INVALID_GROUP_DESC(__sb, __gd)	do {		\
 	typeof(__gd) ____gd = (__gd);					\
 		ocfs2_error((__sb),					\
-		"Group Descriptor # %"MLFu64" has bad signature %.*s",	\
-		(____gd)->bg_blkno, 7,					\
+		"Group Descriptor # %llu has bad signature %.*s",	\
+		(unsigned long long)(____gd)->bg_blkno, 7,		\
 		(____gd)->bg_signature);				\
 } while (0);
 
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index c46c164aefbb..195523090c87 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -157,8 +157,9 @@ static int ocfs2_block_group_fill(struct ocfs2_journal_handle *handle,
 	mlog_entry_void();
 
 	if (((unsigned long long) bg_bh->b_blocknr) != group_blkno) {
-		ocfs2_error(alloc_inode->i_sb, "group block (%"MLFu64") "
-			    "!= b_blocknr (%llu)", group_blkno,
+		ocfs2_error(alloc_inode->i_sb, "group block (%llu) != "
+			    "b_blocknr (%llu)",
+			    (unsigned long long)group_blkno,
 			    (unsigned long long) bg_bh->b_blocknr);
 		status = -EIO;
 		goto bail;
@@ -280,8 +281,8 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
 
 	/* setup the group */
 	bg_blkno = ocfs2_clusters_to_blocks(osb->sb, bit_off);
-	mlog(0, "new descriptor, record %u, at block %"MLFu64"\n",
-	     alloc_rec, bg_blkno);
+	mlog(0, "new descriptor, record %u, at block %llu\n",
+	     alloc_rec, (unsigned long long)bg_blkno);
 
 	bg_bh = sb_getblk(osb->sb, bg_blkno);
 	if (!bg_bh) {
@@ -382,8 +383,8 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
 		goto bail;
 	}
 	if (!(fe->i_flags & cpu_to_le32(OCFS2_CHAIN_FL))) {
-		ocfs2_error(alloc_inode->i_sb, "Invalid chain allocator "
-			    "# %"MLFu64, le64_to_cpu(fe->i_blkno));
+		ocfs2_error(alloc_inode->i_sb, "Invalid chain allocator %llu",
+			    (unsigned long long)le64_to_cpu(fe->i_blkno));
 		status = -EIO;
 		goto bail;
 	}
@@ -829,9 +830,10 @@ static int ocfs2_relink_block_group(struct ocfs2_journal_handle *handle,
 		goto out;
 	}
 
-	mlog(0, "In suballoc %"MLFu64", chain %u, move group %"MLFu64" to "
-	     "top, prev = %"MLFu64"\n",
-	     fe->i_blkno, chain, bg->bg_blkno, prev_bg->bg_blkno);
+	mlog(0, "Suballoc %llu, chain %u, move group %llu to top, prev = %llu\n",
+	     (unsigned long long)fe->i_blkno, chain,
+	     (unsigned long long)bg->bg_blkno,
+	     (unsigned long long)prev_bg->bg_blkno);
 
 	fe_ptr = le64_to_cpu(fe->id2.i_chain.cl_recs[chain].c_blkno);
 	bg_ptr = le64_to_cpu(bg->bg_next_group);
@@ -974,8 +976,9 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 	struct ocfs2_group_desc *bg;
 
 	chain = ac->ac_chain;
-	mlog(0, "trying to alloc %u bits from chain %u, inode %"MLFu64"\n",
-	     bits_wanted, chain, OCFS2_I(alloc_inode)->ip_blkno);
+	mlog(0, "trying to alloc %u bits from chain %u, inode %llu\n",
+	     bits_wanted, chain,
+	     (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno);
 
 	status = ocfs2_read_block(OCFS2_SB(alloc_inode->i_sb),
 				  le64_to_cpu(cl->cl_recs[chain].c_blkno),
@@ -1027,8 +1030,8 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 		goto bail;
 	}
 
-	mlog(0, "alloc succeeds: we give %u bits from block group %"MLFu64"\n",
-	     tmp_bits, bg->bg_blkno);
+	mlog(0, "alloc succeeds: we give %u bits from block group %llu\n",
+	     tmp_bits, (unsigned long long)bg->bg_blkno);
 
 	*num_bits = tmp_bits;
 
@@ -1092,8 +1095,8 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
 		goto bail;
 	}
 
-	mlog(0, "Allocated %u bits from suballocator %"MLFu64"\n",
-	     *num_bits, fe->i_blkno);
+	mlog(0, "Allocated %u bits from suballocator %llu\n", *num_bits,
+	     (unsigned long long)fe->i_blkno);
 
 	*bg_blkno = le64_to_cpu(bg->bg_blkno);
 bail:
@@ -1134,9 +1137,9 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_super *osb,
 	}
 	if (le32_to_cpu(fe->id1.bitmap1.i_used) >=
 	    le32_to_cpu(fe->id1.bitmap1.i_total)) {
-		ocfs2_error(osb->sb, "Chain allocator dinode %"MLFu64" has %u"
-			    "used bits but only %u total.",
-			    le64_to_cpu(fe->i_blkno),
+		ocfs2_error(osb->sb, "Chain allocator dinode %llu has %u used "
+			    "bits but only %u total.",
+			    (unsigned long long)le64_to_cpu(fe->i_blkno),
 			    le32_to_cpu(fe->id1.bitmap1.i_used),
 			    le32_to_cpu(fe->id1.bitmap1.i_total));
 		status = -EIO;
@@ -1479,10 +1482,9 @@ static int ocfs2_free_suballoc_bits(struct ocfs2_journal_handle *handle,
 	}
 	BUG_ON((count + start_bit) > ocfs2_bits_per_group(cl));
 
-	mlog(0, "suballocator %"MLFu64": freeing %u bits from group %"MLFu64
-	     ", starting at %u\n",
-	     OCFS2_I(alloc_inode)->ip_blkno, count, bg_blkno,
-	     start_bit);
+	mlog(0, "%llu: freeing %u bits from group %llu, starting at %u\n",
+	     (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno, count,
+	     (unsigned long long)bg_blkno, start_bit);
 
 	status = ocfs2_read_block(osb, bg_blkno, &group_bh, OCFS2_BH_CACHED,
 				  alloc_inode);
@@ -1592,10 +1594,10 @@ int ocfs2_free_clusters(struct ocfs2_journal_handle *handle,
 	ocfs2_block_to_cluster_group(bitmap_inode, start_blk, &bg_blkno,
 				     &bg_start_bit);
 
-	mlog(0, "want to free %u clusters starting at block %"MLFu64"\n",
-	     num_clusters, start_blk);
-	mlog(0, "bg_blkno = %"MLFu64", bg_start_bit = %u\n",
-	     bg_blkno, bg_start_bit);
+	mlog(0, "want to free %u clusters starting at block %llu\n",
+	     num_clusters, (unsigned long long)start_blk);
+	mlog(0, "bg_blkno = %llu, bg_start_bit = %u\n",
+	     (unsigned long long)bg_blkno, bg_start_bit);
 
 	status = ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
 					  bg_start_bit, bg_blkno,
@@ -1616,18 +1618,22 @@ static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg)
 	printk("bg_free_bits_count: %u\n", bg->bg_free_bits_count);
 	printk("bg_chain:           %u\n", bg->bg_chain);
 	printk("bg_generation:      %u\n", le32_to_cpu(bg->bg_generation));
-	printk("bg_next_group:      %"MLFu64"\n", bg->bg_next_group);
-	printk("bg_parent_dinode:   %"MLFu64"\n", bg->bg_parent_dinode);
-	printk("bg_blkno:           %"MLFu64"\n", bg->bg_blkno);
+	printk("bg_next_group:      %llu\n",
+	       (unsigned long long)bg->bg_next_group);
+	printk("bg_parent_dinode:   %llu\n",
+	       (unsigned long long)bg->bg_parent_dinode);
+	printk("bg_blkno:           %llu\n",
+	       (unsigned long long)bg->bg_blkno);
 }
 
 static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe)
 {
 	int i;
 
-	printk("Suballoc Inode %"MLFu64":\n", fe->i_blkno);
+	printk("Suballoc Inode %llu:\n", (unsigned long long)fe->i_blkno);
 	printk("i_signature:                  %s\n", fe->i_signature);
-	printk("i_size:                       %"MLFu64"\n", fe->i_size);
+	printk("i_size:                       %llu\n",
+	       (unsigned long long)fe->i_size);
 	printk("i_clusters:                   %u\n", fe->i_clusters);
 	printk("i_generation:                 %u\n",
 	       le32_to_cpu(fe->i_generation));
@@ -1645,7 +1651,7 @@ static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe)
 		       fe->id2.i_chain.cl_recs[i].c_free);
 		printk("fe->id2.i_chain.cl_recs[%d].c_total: %u\n", i,
 		       fe->id2.i_chain.cl_recs[i].c_total);
-		printk("fe->id2.i_chain.cl_recs[%d].c_blkno: %"MLFu64"\n", i,
-		       fe->id2.i_chain.cl_recs[i].c_blkno);
+		printk("fe->id2.i_chain.cl_recs[%d].c_blkno: %llu\n", i,
+		       (unsigned long long)fe->id2.i_chain.cl_recs[i].c_blkno);
 	}
 }
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 8dd3aafec499..949b3dac30f1 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -950,16 +950,18 @@ static void ocfs2_inode_init_once(void *data,
 static int ocfs2_initialize_mem_caches(void)
 {
 	ocfs2_inode_cachep = kmem_cache_create("ocfs2_inode_cache",
-					       sizeof(struct ocfs2_inode_info),
-					       0, SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT,
-					       ocfs2_inode_init_once, NULL);
+				       sizeof(struct ocfs2_inode_info),
+				       0,
+				       (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
+				       ocfs2_inode_init_once, NULL);
 	if (!ocfs2_inode_cachep)
 		return -ENOMEM;
 
 	ocfs2_lock_cache = kmem_cache_create("ocfs2_lock",
 					     sizeof(struct ocfs2_journal_lock),
 					     0,
-					     SLAB_NO_REAP|SLAB_HWCACHE_ALIGN,
+					     SLAB_HWCACHE_ALIGN,
 					     NULL, NULL);
 	if (!ocfs2_lock_cache)
 		return -ENOMEM;
@@ -1426,8 +1428,9 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	osb->fs_generation = le32_to_cpu(di->i_fs_generation);
 	mlog(0, "vol_label: %s\n", osb->vol_label);
 	mlog(0, "uuid: %s\n", osb->uuid_str);
-	mlog(0, "root_blkno=%"MLFu64", system_dir_blkno=%"MLFu64"\n",
-	     osb->root_blkno, osb->system_dir_blkno);
+	mlog(0, "root_blkno=%llu, system_dir_blkno=%llu\n",
+	     (unsigned long long)osb->root_blkno,
+	     (unsigned long long)osb->system_dir_blkno);
 
 	osb->osb_dlm_debug = ocfs2_new_dlm_debug();
 	if (!osb->osb_dlm_debug) {
@@ -1470,8 +1473,8 @@ static int ocfs2_initialize_super(struct super_block *sb,
 	osb->bitmap_cpg = le16_to_cpu(di->id2.i_chain.cl_cpg);
 	osb->num_clusters = le32_to_cpu(di->id1.bitmap1.i_total);
 	brelse(bitmap_bh);
-	mlog(0, "cluster bitmap inode: %"MLFu64", clusters per group: %u\n",
-	     osb->bitmap_blkno, osb->bitmap_cpg);
+	mlog(0, "cluster bitmap inode: %llu, clusters per group: %u\n",
+	     (unsigned long long)osb->bitmap_blkno, osb->bitmap_cpg);
 
 	status = ocfs2_init_slot_info(osb);
 	if (status < 0) {
@@ -1529,8 +1532,9 @@ static int ocfs2_verify_volume(struct ocfs2_dinode *di,
 			     OCFS2_MINOR_REV_LEVEL);
 		} else if (bh->b_blocknr != le64_to_cpu(di->i_blkno)) {
 			mlog(ML_ERROR, "bad block number on superblock: "
-			     "found %"MLFu64", should be %llu\n",
-			     di->i_blkno, (unsigned long long)bh->b_blocknr);
+			     "found %llu, should be %llu\n",
+			     (unsigned long long)di->i_blkno,
+			     (unsigned long long)bh->b_blocknr);
 		} else if (le32_to_cpu(di->id2.i_super.s_clustersize_bits) < 12 ||
 			    le32_to_cpu(di->id2.i_super.s_clustersize_bits) > 20) {
 			mlog(ML_ERROR, "bad cluster size found: %u\n",
diff --git a/fs/ocfs2/super.h b/fs/ocfs2/super.h
index c564177dfbdc..783f5270f2a1 100644
--- a/fs/ocfs2/super.h
+++ b/fs/ocfs2/super.h
@@ -33,12 +33,16 @@ int ocfs2_publish_get_mount_state(struct ocfs2_super *osb,
 
 void __ocfs2_error(struct super_block *sb,
 		   const char *function,
-		   const char *fmt, ...);
+		   const char *fmt, ...)
+	__attribute__ ((format (printf, 3, 4)));
+
 #define ocfs2_error(sb, fmt, args...) __ocfs2_error(sb, __PRETTY_FUNCTION__, fmt, ##args)
 
 void __ocfs2_abort(struct super_block *sb,
 		   const char *function,
-		   const char *fmt, ...);
+		   const char *fmt, ...)
+	__attribute__ ((format (printf, 3, 4)));
+
 #define ocfs2_abort(sb, fmt, args...) __ocfs2_abort(sb, __PRETTY_FUNCTION__, fmt, ##args)
 
 #endif /* OCFS2_SUPER_H */
diff --git a/fs/ocfs2/uptodate.c b/fs/ocfs2/uptodate.c
index 300b5bedfb21..04a684dfdd96 100644
--- a/fs/ocfs2/uptodate.c
+++ b/fs/ocfs2/uptodate.c
@@ -119,8 +119,8 @@ void ocfs2_metadata_cache_purge(struct inode *inode)
 	tree = !(oi->ip_flags & OCFS2_INODE_CACHE_INLINE);
 	to_purge = ci->ci_num_cached;
 
-	mlog(0, "Purge %u %s items from Inode %"MLFu64"\n", to_purge,
-	     tree ? "array" : "tree", oi->ip_blkno);
+	mlog(0, "Purge %u %s items from Inode %llu\n", to_purge,
+	     tree ? "array" : "tree", (unsigned long long)oi->ip_blkno);
 
 	/* If we're a tree, save off the root so that we can safely
 	 * initialize the cache. We do the work to free tree members
@@ -136,8 +136,8 @@ void ocfs2_metadata_cache_purge(struct inode *inode)
 	 * easily detect counting errors. Unfortunately, this is only
 	 * meaningful for trees. */
 	if (tree && purged != to_purge)
-		mlog(ML_ERROR, "Inode %"MLFu64", count = %u, purged = %u\n",
-		     oi->ip_blkno, to_purge, purged);
+		mlog(ML_ERROR, "Inode %llu, count = %u, purged = %u\n",
+		     (unsigned long long)oi->ip_blkno, to_purge, purged);
 }
 
 /* Returns the index in the cache array, -1 if not found.
@@ -186,8 +186,9 @@ static int ocfs2_buffer_cached(struct ocfs2_inode_info *oi,
 
 	spin_lock(&oi->ip_lock);
 
-	mlog(0, "Inode %"MLFu64", query block %llu (inline = %u)\n",
-	     oi->ip_blkno, (unsigned long long) bh->b_blocknr,
+	mlog(0, "Inode %llu, query block %llu (inline = %u)\n",
+	     (unsigned long long)oi->ip_blkno,
+	     (unsigned long long) bh->b_blocknr,
 	     !!(oi->ip_flags & OCFS2_INODE_CACHE_INLINE));
 
 	if (oi->ip_flags & OCFS2_INODE_CACHE_INLINE)
@@ -293,12 +294,12 @@ static void ocfs2_expand_cache(struct ocfs2_inode_info *oi,
 	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 
 	mlog_bug_on_msg(ci->ci_num_cached != OCFS2_INODE_MAX_CACHE_ARRAY,
-			"Inode %"MLFu64", num cached = %u, should be %u\n",
-			oi->ip_blkno, ci->ci_num_cached,
+			"Inode %llu, num cached = %u, should be %u\n",
+			(unsigned long long)oi->ip_blkno, ci->ci_num_cached,
 			OCFS2_INODE_MAX_CACHE_ARRAY);
 	mlog_bug_on_msg(!(oi->ip_flags & OCFS2_INODE_CACHE_INLINE),
-			"Inode %"MLFu64" not marked as inline anymore!\n",
-			oi->ip_blkno);
+			"Inode %llu not marked as inline anymore!\n",
+			(unsigned long long)oi->ip_blkno);
 	assert_spin_locked(&oi->ip_lock);
 
 	/* Be careful to initialize the tree members *first* because
@@ -316,8 +317,8 @@ static void ocfs2_expand_cache(struct ocfs2_inode_info *oi,
 		tree[i] = NULL;
 	}
 
-	mlog(0, "Expanded %"MLFu64" to a tree cache: flags 0x%x, num = %u\n",
-	     oi->ip_blkno, oi->ip_flags, ci->ci_num_cached);
+	mlog(0, "Expanded %llu to a tree cache: flags 0x%x, num = %u\n",
+	     (unsigned long long)oi->ip_blkno, oi->ip_flags, ci->ci_num_cached);
 }
 
 /* Slow path function - memory allocation is necessary. See the
@@ -332,8 +333,9 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_inode_info *oi,
 	struct ocfs2_meta_cache_item *tree[OCFS2_INODE_MAX_CACHE_ARRAY] =
 		{ NULL, };
 
-	mlog(0, "Inode %"MLFu64", block %llu, expand = %d\n",
-	     oi->ip_blkno, (unsigned long long) block, expand_tree);
+	mlog(0, "Inode %llu, block %llu, expand = %d\n",
+	     (unsigned long long)oi->ip_blkno,
+	     (unsigned long long)block, expand_tree);
 
 	new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_KERNEL);
 	if (!new) {
@@ -414,8 +416,9 @@ void ocfs2_set_buffer_uptodate(struct inode *inode,
 	if (ocfs2_buffer_cached(oi, bh))
 		return;
 
-	mlog(0, "Inode %"MLFu64", inserting block %llu\n", oi->ip_blkno,
-	     (unsigned long long) bh->b_blocknr);
+	mlog(0, "Inode %llu, inserting block %llu\n",
+	     (unsigned long long)oi->ip_blkno,
+	     (unsigned long long)bh->b_blocknr);
 
 	/* No need to recheck under spinlock - insertion is guarded by
 	 * ip_io_mutex */
@@ -504,8 +507,9 @@ void ocfs2_remove_from_cache(struct inode *inode,
 	struct ocfs2_caching_info *ci = &oi->ip_metadata_cache;
 
 	spin_lock(&oi->ip_lock);
-	mlog(0, "Inode %"MLFu64", remove %llu, items = %u, array = %u\n",
-	     oi->ip_blkno, (unsigned long long) block, ci->ci_num_cached,
+	mlog(0, "Inode %llu, remove %llu, items = %u, array = %u\n",
+	     (unsigned long long)oi->ip_blkno,
+	     (unsigned long long) block, ci->ci_num_cached,
 	     oi->ip_flags & OCFS2_INODE_CACHE_INLINE);
 
 	if (oi->ip_flags & OCFS2_INODE_CACHE_INLINE) {
diff --git a/fs/ocfs2/vote.c b/fs/ocfs2/vote.c
index 021978e0576b..53049a204197 100644
--- a/fs/ocfs2/vote.c
+++ b/fs/ocfs2/vote.c
@@ -190,20 +190,21 @@ static int ocfs2_process_delete_request(struct inode *inode,
 				OCFS2_INVALID_SLOT &&
 				OCFS2_I(inode)->ip_orphaned_slot !=
 				(*orphaned_slot),
-				"Inode %"MLFu64": This node thinks it's "
+				"Inode %llu: This node thinks it's "
 				"orphaned in slot %d, messaged it's in %d\n",
-				OCFS2_I(inode)->ip_blkno,
+				(unsigned long long)OCFS2_I(inode)->ip_blkno,
 				OCFS2_I(inode)->ip_orphaned_slot,
 				*orphaned_slot);
 
-		mlog(0, "Setting orphaned slot for inode %"MLFu64" to %d\n",
-		     OCFS2_I(inode)->ip_blkno, *orphaned_slot);
+		mlog(0, "Setting orphaned slot for inode %llu to %d\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		     *orphaned_slot);
 
 		OCFS2_I(inode)->ip_orphaned_slot = *orphaned_slot;
 	} else {
-		mlog(0, "Sending back orphaned slot %d for inode %"MLFu64"\n",
+		mlog(0, "Sending back orphaned slot %d for inode %llu\n",
 		     OCFS2_I(inode)->ip_orphaned_slot,
-		     OCFS2_I(inode)->ip_blkno);
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 
 		*orphaned_slot = OCFS2_I(inode)->ip_orphaned_slot;
 	}
@@ -226,8 +227,8 @@ static int ocfs2_process_delete_request(struct inode *inode,
 	}
 
 	if (filemap_fdatawrite(inode->i_mapping)) {
-		mlog(ML_ERROR, "Could not sync inode %"MLFu64" for delete!\n",
-		     OCFS2_I(inode)->ip_blkno);
+		mlog(ML_ERROR, "Could not sync inode %llu for delete!\n",
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno);
 		goto done;
 	}
 	sync_mapping_buffers(inode->i_mapping);
@@ -302,8 +303,8 @@ static void ocfs2_process_dentry_request(struct inode *inode,
 	struct list_head *p;
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 
-	mlog(0, "parent %"MLFu64", namelen = %u, name = %.*s\n", parent_blkno,
-	     namelen, namelen, name);
+	mlog(0, "parent %llu, namelen = %u, name = %.*s\n",
+	     (unsigned long long)parent_blkno, namelen, namelen, name);
 
 	spin_lock(&dcache_lock);
 
@@ -370,9 +371,10 @@ static void ocfs2_process_vote(struct ocfs2_super *osb,
 	if (request == OCFS2_VOTE_REQ_DELETE)
 		orphaned_slot = be32_to_cpu(msg->md1.v_orphaned_slot);
 
-	mlog(0, "processing vote: request = %u, blkno = %"MLFu64", "
+	mlog(0, "processing vote: request = %u, blkno = %llu, "
 	     "generation = %u, node_num = %u, priv1 = %u\n", request,
-	     blkno, generation, node_num, be32_to_cpu(msg->md1.v_generic1));
+	     (unsigned long long)blkno, generation, node_num,
+	     be32_to_cpu(msg->md1.v_generic1));
 
 	if (!ocfs2_is_valid_vote_request(request)) {
 		mlog(ML_ERROR, "Invalid vote request %d from node %u\n",
@@ -419,11 +421,12 @@ static void ocfs2_process_vote(struct ocfs2_super *osb,
 	 * we had not found an inode in the first place. */
 	if (inode->i_generation != generation) {
 		mlog(0, "generation passed %u != inode generation = %u, "
-		     "ip_flags = %x, ip_blkno = %"MLFu64", msg %"MLFu64", "
-		     "i_count = %u, message type = %u\n",
-		     generation, inode->i_generation, OCFS2_I(inode)->ip_flags,
-		     OCFS2_I(inode)->ip_blkno, blkno,
-		     atomic_read(&inode->i_count), request);
+		     "ip_flags = %x, ip_blkno = %llu, msg %llu, i_count = %u, "
+		     "message type = %u\n", generation, inode->i_generation,
+		     OCFS2_I(inode)->ip_flags,
+		     (unsigned long long)OCFS2_I(inode)->ip_blkno,
+		     (unsigned long long)blkno, atomic_read(&inode->i_count),
+		     request);
 		iput(inode);
 		inode = NULL;
 		goto respond;
@@ -830,8 +833,9 @@ static void ocfs2_delete_response_cb(void *priv,
 
 	orphaned_slot = be32_to_cpu(resp->r_orphaned_slot);
 	node = be32_to_cpu(resp->r_hdr.h_node_num);
-	mlog(0, "node %d tells us that inode %"MLFu64" is orphaned in slot "
-	     "%d\n", node, OCFS2_I(inode)->ip_blkno, orphaned_slot);
+	mlog(0, "node %d tells us that inode %llu is orphaned in slot %d\n",
+	     node, (unsigned long long)OCFS2_I(inode)->ip_blkno,
+	     orphaned_slot);
 
 	/* The other node may not actually know which slot the inode
 	 * is orphaned in. */
@@ -845,9 +849,9 @@ static void ocfs2_delete_response_cb(void *priv,
 	spin_lock(&OCFS2_I(inode)->ip_lock);
 	mlog_bug_on_msg(OCFS2_I(inode)->ip_orphaned_slot != orphaned_slot &&
 			OCFS2_I(inode)->ip_orphaned_slot
-			!= OCFS2_INVALID_SLOT, "Inode %"MLFu64": Node %d "
-			"says it's orphaned in slot %d, we think it's in %d\n",
-			OCFS2_I(inode)->ip_blkno,
+			!= OCFS2_INVALID_SLOT, "Inode %llu: Node %d says it's "
+			"orphaned in slot %d, we think it's in %d\n",
+			(unsigned long long)OCFS2_I(inode)->ip_blkno,
 			be32_to_cpu(resp->r_hdr.h_node_num),
 			orphaned_slot, OCFS2_I(inode)->ip_orphaned_slot);
 
@@ -869,8 +873,8 @@ int ocfs2_request_delete_vote(struct inode *inode)
 	delete_cb.rc_cb = ocfs2_delete_response_cb;
 	delete_cb.rc_priv = inode;
 
-	mlog(0, "Inode %"MLFu64", we start thinking orphaned slot is %d\n",
-	     OCFS2_I(inode)->ip_blkno, orphaned_slot);
+	mlog(0, "Inode %llu, we start thinking orphaned slot is %d\n",
+	     (unsigned long long)OCFS2_I(inode)->ip_blkno, orphaned_slot);
 
 	status = -ENOMEM;
 	request = ocfs2_new_vote_request(osb, OCFS2_I(inode)->ip_blkno,
@@ -895,8 +899,8 @@ static void ocfs2_setup_unlink_vote(struct ocfs2_vote_msg *request,
 	 * d_delete against it. Parent directory block and full name
 	 * should suffice. */
 
-	mlog(0, "unlink/rename request: parent: %"MLFu64" name: %.*s\n",
-	     OCFS2_I(parent)->ip_blkno, dentry->d_name.len,
+	mlog(0, "unlink/rename request: parent: %llu name: %.*s\n",
+	     (unsigned long long)OCFS2_I(parent)->ip_blkno, dentry->d_name.len,
 	     dentry->d_name.name);
 
 	request->v_unlink_parent = cpu_to_be64(OCFS2_I(parent)->ip_blkno);
@@ -1082,7 +1086,8 @@ static int ocfs2_handle_response_message(struct o2net_msg *msg,
 	mlog(0, "received response message:\n");
 	mlog(0, "h_response_id = %u\n", response_id);
 	mlog(0, "h_request = %u\n", be32_to_cpu(resp->r_hdr.h_request));
-	mlog(0, "h_blkno = %"MLFu64"\n", be64_to_cpu(resp->r_hdr.h_blkno));
+	mlog(0, "h_blkno = %llu\n",
+	     (unsigned long long)be64_to_cpu(resp->r_hdr.h_blkno));
 	mlog(0, "h_generation = %u\n", be32_to_cpu(resp->r_hdr.h_generation));
 	mlog(0, "h_node_num = %u\n", node_num);
 	mlog(0, "r_response = %d\n", response_status);
@@ -1138,8 +1143,8 @@ static int ocfs2_handle_vote_message(struct o2net_msg *msg,
 	mlog(0, "h_response_id = %u\n",
 	     be32_to_cpu(work->w_msg.v_hdr.h_response_id));
 	mlog(0, "h_request = %u\n", be32_to_cpu(work->w_msg.v_hdr.h_request));
-	mlog(0, "h_blkno = %"MLFu64"\n",
-	     be64_to_cpu(work->w_msg.v_hdr.h_blkno));
+	mlog(0, "h_blkno = %llu\n",
+	     (unsigned long long)be64_to_cpu(work->w_msg.v_hdr.h_blkno));
 	mlog(0, "h_generation = %u\n",
 	     be32_to_cpu(work->w_msg.v_hdr.h_generation));
 	mlog(0, "h_node_num = %u\n",
diff --git a/fs/open.c b/fs/open.c
index 70e0230d8e77..c32c89d6d8db 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -27,6 +27,7 @@
 #include <linux/pagemap.h>
 #include <linux/syscalls.h>
 #include <linux/rcupdate.h>
+#include <linux/audit.h>
 
 #include <asm/unistd.h>
 
@@ -626,6 +627,8 @@ asmlinkage long sys_fchmod(unsigned int fd, mode_t mode)
 	dentry = file->f_dentry;
 	inode = dentry->d_inode;
 
+	audit_inode(NULL, inode, 0);
+
 	err = -EROFS;
 	if (IS_RDONLY(inode))
 		goto out_putf;
@@ -775,7 +778,10 @@ asmlinkage long sys_fchown(unsigned int fd, uid_t user, gid_t group)
 
 	file = fget(fd);
 	if (file) {
-		error = chown_common(file->f_dentry, user, group);
+		struct dentry * dentry;
+		dentry = file->f_dentry;
+		audit_inode(NULL, dentry->d_inode, 0);
+		error = chown_common(dentry, user, group);
 		fput(file);
 	}
 	return error;
@@ -890,6 +896,10 @@ EXPORT_SYMBOL(filp_open);
  * a fully instantiated struct file to the caller.
  * This function is meant to be called from within a filesystem's
  * lookup method.
+ * Beware of calling it for non-regular files! Those ->open methods might block
+ * (e.g. in fifo_open), leaving you with parent locked (and in case of fifo,
+ * leading to a deadlock, as nobody can open that fifo anymore, because
+ * another process to open fifo will block on locked parent when doing lookup).
  * Note that in case of error, nd->intent.open.file is destroyed, but the
  * path information remains valid.
  * If the open callback is set to NULL, then the standard f_op->open()
@@ -973,7 +983,7 @@ repeat:
 	fdt = files_fdtable(files);
  	fd = find_next_zero_bit(fdt->open_fds->fds_bits,
 				fdt->max_fdset,
-				fdt->next_fd);
+				files->next_fd);
 
 	/*
 	 * N.B. For clone tasks sharing a files structure, this test
@@ -998,7 +1008,7 @@ repeat:
 
 	FD_SET(fd, fdt->open_fds);
 	FD_CLR(fd, fdt->close_on_exec);
-	fdt->next_fd = fd + 1;
+	files->next_fd = fd + 1;
 #if 1
 	/* Sanity check */
 	if (fdt->fd[fd] != NULL) {
@@ -1019,8 +1029,8 @@ static void __put_unused_fd(struct files_struct *files, unsigned int fd)
 {
 	struct fdtable *fdt = files_fdtable(files);
 	__FD_CLR(fd, fdt->open_fds);
-	if (fd < fdt->next_fd)
-		fdt->next_fd = fd;
+	if (fd < files->next_fd)
+		files->next_fd = fd;
 }
 
 void fastcall put_unused_fd(unsigned int fd)
diff --git a/fs/openpromfs/inode.c b/fs/openpromfs/inode.c
index aeb0106890e4..0f14276a2e51 100644
--- a/fs/openpromfs/inode.c
+++ b/fs/openpromfs/inode.c
@@ -581,17 +581,17 @@ int property_release (struct inode *inode, struct file *filp)
 	return 0;
 }
 
-static struct file_operations openpromfs_prop_ops = {
+static const struct file_operations openpromfs_prop_ops = {
 	.read		= property_read,
 	.write		= property_write,
 	.release	= property_release,
 };
 
-static struct file_operations openpromfs_nodenum_ops = {
+static const struct file_operations openpromfs_nodenum_ops = {
 	.read		= nodenum_read,
 };
 
-static struct file_operations openprom_operations = {
+static const struct file_operations openprom_operations = {
 	.read		= generic_read_dir,
 	.readdir	= openpromfs_readdir,
 };
diff --git a/fs/partitions/check.c b/fs/partitions/check.c
index f924f459bdb8..af0cb4b9e784 100644
--- a/fs/partitions/check.c
+++ b/fs/partitions/check.c
@@ -297,6 +297,25 @@ struct kobj_type ktype_part = {
 	.sysfs_ops	= &part_sysfs_ops,
 };
 
+static inline void partition_sysfs_add_subdir(struct hd_struct *p)
+{
+	struct kobject *k;
+
+	k = kobject_get(&p->kobj);
+	p->holder_dir = kobject_add_dir(k, "holders");
+	kobject_put(k);
+}
+
+static inline void disk_sysfs_add_subdirs(struct gendisk *disk)
+{
+	struct kobject *k;
+
+	k = kobject_get(&disk->kobj);
+	disk->holder_dir = kobject_add_dir(k, "holders");
+	disk->slave_dir = kobject_add_dir(k, "slaves");
+	kobject_put(k);
+}
+
 void delete_partition(struct gendisk *disk, int part)
 {
 	struct hd_struct *p = disk->part[part-1];
@@ -310,6 +329,8 @@ void delete_partition(struct gendisk *disk, int part)
 	p->ios[0] = p->ios[1] = 0;
 	p->sectors[0] = p->sectors[1] = 0;
 	devfs_remove("%s/part%d", disk->devfs_name, part);
+	if (p->holder_dir)
+		kobject_unregister(p->holder_dir);
 	kobject_unregister(&p->kobj);
 }
 
@@ -337,6 +358,7 @@ void add_partition(struct gendisk *disk, int part, sector_t start, sector_t len)
 	p->kobj.parent = &disk->kobj;
 	p->kobj.ktype = &ktype_part;
 	kobject_register(&p->kobj);
+	partition_sysfs_add_subdir(p);
 	disk->part[part-1] = p;
 }
 
@@ -383,6 +405,7 @@ void register_disk(struct gendisk *disk)
 	if ((err = kobject_add(&disk->kobj)))
 		return;
 	disk_sysfs_symlinks(disk);
+ 	disk_sysfs_add_subdirs(disk);
 	kobject_uevent(&disk->kobj, KOBJ_ADD);
 
 	/* No minors to use for partitions */
@@ -483,6 +506,10 @@ void del_gendisk(struct gendisk *disk)
 
 	devfs_remove_disk(disk);
 
+	if (disk->holder_dir)
+		kobject_unregister(disk->holder_dir);
+	if (disk->slave_dir)
+		kobject_unregister(disk->slave_dir);
 	if (disk->driverfs_dev) {
 		char *disk_name = make_block_name(disk);
 		sysfs_remove_link(&disk->kobj, "device");
diff --git a/fs/partitions/devfs.c b/fs/partitions/devfs.c
index 87f50444fd39..3f0a780c9cec 100644
--- a/fs/partitions/devfs.c
+++ b/fs/partitions/devfs.c
@@ -6,7 +6,7 @@
 #include <linux/vmalloc.h>
 #include <linux/genhd.h>
 #include <linux/bitops.h>
-#include <asm/semaphore.h>
+#include <linux/mutex.h>
 
 
 struct unique_numspace {
@@ -16,7 +16,7 @@ struct unique_numspace {
 	struct semaphore  mutex;
 };
 
-static DECLARE_MUTEX(numspace_mutex);
+static DEFINE_MUTEX(numspace_mutex);
 
 static int expand_numspace(struct unique_numspace *s)
 {
@@ -48,7 +48,7 @@ static int alloc_unique_number(struct unique_numspace *s)
 {
 	int rval = 0;
 
-	down(&numspace_mutex);
+	mutex_lock(&numspace_mutex);
 	if (s->num_free < 1)
 		rval = expand_numspace(s);
 	if (!rval) {
@@ -56,7 +56,7 @@ static int alloc_unique_number(struct unique_numspace *s)
 		--s->num_free;
 		__set_bit(rval, s->bits);
 	}
-	up(&numspace_mutex);
+	mutex_unlock(&numspace_mutex);
 
 	return rval;
 }
@@ -66,11 +66,11 @@ static void dealloc_unique_number(struct unique_numspace *s, int number)
 	int old_val;
 
 	if (number >= 0) {
-		down(&numspace_mutex);
+		mutex_lock(&numspace_mutex);
 		old_val = __test_and_clear_bit(number, s->bits);
 		if (old_val)
 			++s->num_free;
-		up(&numspace_mutex);
+		mutex_unlock(&numspace_mutex);
 	}
 }
 
diff --git a/fs/partitions/ibm.c b/fs/partitions/ibm.c
index 1e4a93835fed..830c55d86ab1 100644
--- a/fs/partitions/ibm.c
+++ b/fs/partitions/ibm.c
@@ -1,15 +1,9 @@
 /*
- * File...........: linux/fs/partitions/ibm.c      
+ * File...........: linux/fs/partitions/ibm.c
  * Author(s)......: Holger Smolinski <Holger.Smolinski@de.ibm.com>
  *                  Volker Sameske <sameske@de.ibm.com>
  * Bugreports.to..: <Linux390@de.ibm.com>
  * (C) IBM Corporation, IBM Deutschland Entwicklung GmbH, 1999,2000
-
- * History of changes (starts July 2000)
- * 07/10/00 Fixed detection of CMS formatted disks     
- * 02/13/00 VTOC partition support added
- * 12/27/01 fixed PL030593 (CMS reserved minidisk not detected on 64 bit)
- * 07/24/03 no longer using contents of freed page for CMS label recognition (BZ3611)
  */
 
 #include <linux/config.h>
@@ -25,7 +19,7 @@
 #include "ibm.h"
 
 /*
- * compute the block number from a 
+ * compute the block number from a
  * cyl-cyl-head-head structure
  */
 static inline int
@@ -34,9 +28,8 @@ cchh2blk (struct vtoc_cchh *ptr, struct hd_geometry *geo) {
 	       ptr->hh * geo->sectors;
 }
 
-
 /*
- * compute the block number from a 
+ * compute the block number from a
  * cyl-cyl-head-head-block structure
  */
 static inline int
@@ -48,7 +41,7 @@ cchhb2blk (struct vtoc_cchhb *ptr, struct hd_geometry *geo) {
 
 /*
  */
-int 
+int
 ibm_partition(struct parsed_partitions *state, struct block_device *bdev)
 {
 	int blocksize, offset, size;
@@ -77,7 +70,7 @@ ibm_partition(struct parsed_partitions *state, struct block_device *bdev)
 		goto out_nogeo;
 	if ((label = kmalloc(sizeof(union label_t), GFP_KERNEL)) == NULL)
 		goto out_nolab;
-	
+
 	if (ioctl_by_bdev(bdev, BIODASDINFO, (unsigned long)info) != 0 ||
 	    ioctl_by_bdev(bdev, HDIO_GETGEO, (unsigned long)geo) != 0)
 		goto out_noioctl;
@@ -154,13 +147,13 @@ ibm_partition(struct parsed_partitions *state, struct block_device *bdev)
 
 			/* OK, we got valid partition data */
 		        offset = cchh2blk(&f1.DS1EXT1.llimit, geo);
-			size  = cchh2blk(&f1.DS1EXT1.ulimit, geo) - 
+			size  = cchh2blk(&f1.DS1EXT1.ulimit, geo) -
 				offset + geo->sectors;
 			if (counter >= state->limit)
 				break;
-			put_partition(state, counter + 1, 
-					 offset * (blocksize >> 9),
-					 size * (blocksize >> 9));
+			put_partition(state, counter + 1,
+				      offset * (blocksize >> 9),
+				      size * (blocksize >> 9));
 			counter++;
 			blk++;
 		}
@@ -175,7 +168,7 @@ ibm_partition(struct parsed_partitions *state, struct block_device *bdev)
 		offset = (info->label_block + 1);
 		size = i_size >> 9;
 		put_partition(state, 1, offset*(blocksize >> 9),
-				 size-offset*(blocksize >> 9));
+			      size-offset*(blocksize >> 9));
 	}
 
 	printk("\n");
@@ -183,7 +176,7 @@ ibm_partition(struct parsed_partitions *state, struct block_device *bdev)
 	kfree(geo);
 	kfree(info);
 	return 1;
-	
+
 out_readerr:
 out_noioctl:
 	kfree(label);
diff --git a/fs/partitions/mac.c b/fs/partitions/mac.c
index bb22cdd0cb14..813292f21210 100644
--- a/fs/partitions/mac.c
+++ b/fs/partitions/mac.c
@@ -12,6 +12,7 @@
 #include "mac.h"
 
 #ifdef CONFIG_PPC_PMAC
+#include <asm/machdep.h>
 extern void note_bootable_part(dev_t dev, int part, int goodness);
 #endif
 
@@ -79,7 +80,7 @@ int mac_partition(struct parsed_partitions *state, struct block_device *bdev)
 		 * If this is the first bootable partition, tell the
 		 * setup code, in case it wants to make this the root.
 		 */
-		if (_machine == _MACH_Pmac) {
+		if (machine_is(powermac)) {
 			int goodness = 0;
 
 			mac_fix_string(part->processor, 16);
diff --git a/fs/pipe.c b/fs/pipe.c
index 8aada8e426f4..109a102c150d 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -15,6 +15,7 @@
 #include <linux/pipe_fs_i.h>
 #include <linux/uio.h>
 #include <linux/highmem.h>
+#include <linux/pagemap.h>
 
 #include <asm/uaccess.h>
 #include <asm/ioctls.h>
@@ -94,11 +95,20 @@ static void anon_pipe_buf_release(struct pipe_inode_info *info, struct pipe_buff
 {
 	struct page *page = buf->page;
 
-	if (info->tmp_page) {
-		__free_page(page);
+	/*
+	 * If nobody else uses this page, and we don't already have a
+	 * temporary page, let's keep track of it as a one-deep
+	 * allocation cache
+	 */
+	if (page_count(page) == 1 && !info->tmp_page) {
+		info->tmp_page = page;
 		return;
 	}
-	info->tmp_page = page;
+
+	/*
+	 * Otherwise just release our reference to it
+	 */
+	page_cache_release(page);
 }
 
 static void *anon_pipe_buf_map(struct file *file, struct pipe_inode_info *info, struct pipe_buffer *buf)
@@ -111,11 +121,19 @@ static void anon_pipe_buf_unmap(struct pipe_inode_info *info, struct pipe_buffer
 	kunmap(buf->page);
 }
 
+static int anon_pipe_buf_steal(struct pipe_inode_info *info,
+			       struct pipe_buffer *buf)
+{
+	buf->stolen = 1;
+	return 0;
+}
+
 static struct pipe_buf_operations anon_pipe_buf_ops = {
 	.can_merge = 1,
 	.map = anon_pipe_buf_map,
 	.unmap = anon_pipe_buf_unmap,
 	.release = anon_pipe_buf_release,
+	.steal = anon_pipe_buf_steal,
 };
 
 static ssize_t
@@ -152,6 +170,11 @@ pipe_readv(struct file *filp, const struct iovec *_iov,
 				chars = total_len;
 
 			addr = ops->map(filp, info, buf);
+			if (IS_ERR(addr)) {
+				if (!ret)
+					ret = PTR_ERR(addr);
+				break;
+			}
 			error = pipe_iov_copy_to_user(iov, addr + buf->offset, chars);
 			ops->unmap(info, buf);
 			if (unlikely(error)) {
@@ -254,8 +277,16 @@ pipe_writev(struct file *filp, const struct iovec *_iov,
 		struct pipe_buf_operations *ops = buf->ops;
 		int offset = buf->offset + buf->len;
 		if (ops->can_merge && offset + chars <= PAGE_SIZE) {
-			void *addr = ops->map(filp, info, buf);
-			int error = pipe_iov_copy_from_user(offset + addr, iov, chars);
+			void *addr;
+			int error;
+
+			addr = ops->map(filp, info, buf);
+			if (IS_ERR(addr)) {
+				error = PTR_ERR(addr);
+				goto out;
+			}
+			error = pipe_iov_copy_from_user(offset + addr, iov,
+							chars);
 			ops->unmap(info, buf);
 			ret = error;
 			do_wakeup = 1;
@@ -568,7 +599,7 @@ pipe_rdwr_open(struct inode *inode, struct file *filp)
  * The file_operations structs are not static because they
  * are also used in linux/fs/fifo.c to do operations on FIFOs.
  */
-struct file_operations read_fifo_fops = {
+const struct file_operations read_fifo_fops = {
 	.llseek		= no_llseek,
 	.read		= pipe_read,
 	.readv		= pipe_readv,
@@ -580,7 +611,7 @@ struct file_operations read_fifo_fops = {
 	.fasync		= pipe_read_fasync,
 };
 
-struct file_operations write_fifo_fops = {
+const struct file_operations write_fifo_fops = {
 	.llseek		= no_llseek,
 	.read		= bad_pipe_r,
 	.write		= pipe_write,
@@ -592,7 +623,7 @@ struct file_operations write_fifo_fops = {
 	.fasync		= pipe_write_fasync,
 };
 
-struct file_operations rdwr_fifo_fops = {
+const struct file_operations rdwr_fifo_fops = {
 	.llseek		= no_llseek,
 	.read		= pipe_read,
 	.readv		= pipe_readv,
@@ -662,10 +693,9 @@ struct inode* pipe_new(struct inode* inode)
 {
 	struct pipe_inode_info *info;
 
-	info = kmalloc(sizeof(struct pipe_inode_info), GFP_KERNEL);
+	info = kzalloc(sizeof(struct pipe_inode_info), GFP_KERNEL);
 	if (!info)
 		goto fail_page;
-	memset(info, 0, sizeof(*info));
 	inode->i_pipe = info;
 
 	init_waitqueue_head(PIPE_WAIT(*inode));
@@ -676,7 +706,7 @@ fail_page:
 	return NULL;
 }
 
-static struct vfsmount *pipe_mnt;
+static struct vfsmount *pipe_mnt __read_mostly;
 static int pipefs_delete_dentry(struct dentry *dentry)
 {
 	return 1;
diff --git a/fs/pnode.c b/fs/pnode.c
index f1871f773f64..37b568ed0e05 100644
--- a/fs/pnode.c
+++ b/fs/pnode.c
@@ -130,7 +130,7 @@ static struct vfsmount *get_source(struct vfsmount *dest,
 {
 	struct vfsmount *p_last_src = NULL;
 	struct vfsmount *p_last_dest = NULL;
-	*type = CL_PROPAGATION;;
+	*type = CL_PROPAGATION;
 
 	if (IS_MNT_SHARED(dest))
 		*type |= CL_MAKE_SHARED;
diff --git a/fs/proc/array.c b/fs/proc/array.c
index 7eb1bd7f800c..7a76ad570230 100644
--- a/fs/proc/array.c
+++ b/fs/proc/array.c
@@ -330,7 +330,6 @@ static int do_task_stat(struct task_struct *task, char * buffer, int whole)
 	unsigned long  min_flt = 0,  maj_flt = 0;
 	cputime_t cutime, cstime, utime, stime;
 	unsigned long rsslim = 0;
-	DEFINE_KTIME(it_real_value);
 	struct task_struct *t;
 	char tcomm[sizeof(task->comm)];
 
@@ -386,7 +385,6 @@ static int do_task_stat(struct task_struct *task, char * buffer, int whole)
 			utime = cputime_add(utime, task->signal->utime);
 			stime = cputime_add(stime, task->signal->stime);
 		}
-		it_real_value = task->signal->real_timer.expires;
 	}
 	ppid = pid_alive(task) ? task->group_leader->real_parent->tgid : 0;
 	read_unlock(&tasklist_lock);
@@ -413,7 +411,7 @@ static int do_task_stat(struct task_struct *task, char * buffer, int whole)
 	start_time = nsec_to_clock_t(start_time);
 
 	res = sprintf(buffer,"%d (%s) %c %d %d %d %d %d %lu %lu \
-%lu %lu %lu %lu %lu %ld %ld %ld %ld %d %ld %llu %lu %ld %lu %lu %lu %lu %lu \
+%lu %lu %lu %lu %lu %ld %ld %ld %ld %d 0 %llu %lu %ld %lu %lu %lu %lu %lu \
 %lu %lu %lu %lu %lu %lu %lu %lu %d %d %lu %lu\n",
 		task->pid,
 		tcomm,
@@ -435,7 +433,6 @@ static int do_task_stat(struct task_struct *task, char * buffer, int whole)
 		priority,
 		nice,
 		num_threads,
-		(long) ktime_to_clock_t(it_real_value),
 		start_time,
 		vsize,
 		mm ? get_mm_rss(mm) : 0,
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 20feb7568deb..8f1f49ceebec 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -104,6 +104,7 @@ enum pid_directory_inos {
 	PROC_TGID_MAPS,
 	PROC_TGID_NUMA_MAPS,
 	PROC_TGID_MOUNTS,
+	PROC_TGID_MOUNTSTATS,
 	PROC_TGID_WCHAN,
 #ifdef CONFIG_MMU
 	PROC_TGID_SMAPS,
@@ -144,6 +145,7 @@ enum pid_directory_inos {
 	PROC_TID_MAPS,
 	PROC_TID_NUMA_MAPS,
 	PROC_TID_MOUNTS,
+	PROC_TID_MOUNTSTATS,
 	PROC_TID_WCHAN,
 #ifdef CONFIG_MMU
 	PROC_TID_SMAPS,
@@ -201,6 +203,7 @@ static struct pid_entry tgid_base_stuff[] = {
 	E(PROC_TGID_ROOT,      "root",    S_IFLNK|S_IRWXUGO),
 	E(PROC_TGID_EXE,       "exe",     S_IFLNK|S_IRWXUGO),
 	E(PROC_TGID_MOUNTS,    "mounts",  S_IFREG|S_IRUGO),
+	E(PROC_TGID_MOUNTSTATS, "mountstats", S_IFREG|S_IRUSR),
 #ifdef CONFIG_MMU
 	E(PROC_TGID_SMAPS,     "smaps",   S_IFREG|S_IRUGO),
 #endif
@@ -732,6 +735,38 @@ static struct file_operations proc_mounts_operations = {
 	.poll		= mounts_poll,
 };
 
+extern struct seq_operations mountstats_op;
+static int mountstats_open(struct inode *inode, struct file *file)
+{
+	struct task_struct *task = proc_task(inode);
+	int ret = seq_open(file, &mountstats_op);
+
+	if (!ret) {
+		struct seq_file *m = file->private_data;
+		struct namespace *namespace;
+		task_lock(task);
+		namespace = task->namespace;
+		if (namespace)
+			get_namespace(namespace);
+		task_unlock(task);
+
+		if (namespace)
+			m->private = namespace;
+		else {
+			seq_release(inode, file);
+			ret = -EINVAL;
+		}
+	}
+	return ret;
+}
+
+static struct file_operations proc_mountstats_operations = {
+	.open		= mountstats_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= mounts_release,
+};
+
 #define PROC_BLOCK_SIZE	(3*1024)		/* 4K page size but our output routines use some slack for overruns */
 
 static ssize_t proc_info_read(struct file * file, char __user * buf,
@@ -1730,6 +1765,10 @@ static struct dentry *proc_pident_lookup(struct inode *dir,
 			inode->i_fop = &proc_smaps_operations;
 			break;
 #endif
+		case PROC_TID_MOUNTSTATS:
+		case PROC_TGID_MOUNTSTATS:
+			inode->i_fop = &proc_mountstats_operations;
+			break;
 #ifdef CONFIG_SECURITY
 		case PROC_TID_ATTR:
 			inode->i_nlink = 2;
diff --git a/fs/proc/generic.c b/fs/proc/generic.c
index 20e5c4509a43..4ba03009cf72 100644
--- a/fs/proc/generic.c
+++ b/fs/proc/generic.c
@@ -19,6 +19,7 @@
 #include <linux/idr.h>
 #include <linux/namei.h>
 #include <linux/bitops.h>
+#include <linux/spinlock.h>
 #include <asm/uaccess.h>
 
 #include "internal.h"
@@ -29,6 +30,8 @@ static ssize_t proc_file_write(struct file *file, const char __user *buffer,
 			       size_t count, loff_t *ppos);
 static loff_t proc_file_lseek(struct file *, loff_t, int);
 
+DEFINE_SPINLOCK(proc_subdir_lock);
+
 int proc_match(int len, const char *name, struct proc_dir_entry *de)
 {
 	if (de->namelen != len)
@@ -277,7 +280,9 @@ static int xlate_proc_name(const char *name,
 	const char     		*cp = name, *next;
 	struct proc_dir_entry	*de;
 	int			len;
+	int 			rtn = 0;
 
+	spin_lock(&proc_subdir_lock);
 	de = &proc_root;
 	while (1) {
 		next = strchr(cp, '/');
@@ -289,13 +294,17 @@ static int xlate_proc_name(const char *name,
 			if (proc_match(len, cp, de))
 				break;
 		}
-		if (!de)
-			return -ENOENT;
+		if (!de) {
+			rtn = -ENOENT;
+			goto out;
+		}
 		cp += len + 1;
 	}
 	*residual = cp;
 	*ret = de;
-	return 0;
+out:
+	spin_unlock(&proc_subdir_lock);
+	return rtn;
 }
 
 static DEFINE_IDR(proc_inum_idr);
@@ -380,6 +389,7 @@ struct dentry *proc_lookup(struct inode * dir, struct dentry *dentry, struct nam
 	int error = -ENOENT;
 
 	lock_kernel();
+	spin_lock(&proc_subdir_lock);
 	de = PDE(dir);
 	if (de) {
 		for (de = de->subdir; de ; de = de->next) {
@@ -388,12 +398,15 @@ struct dentry *proc_lookup(struct inode * dir, struct dentry *dentry, struct nam
 			if (!memcmp(dentry->d_name.name, de->name, de->namelen)) {
 				unsigned int ino = de->low_ino;
 
+				spin_unlock(&proc_subdir_lock);
 				error = -EINVAL;
 				inode = proc_get_inode(dir->i_sb, ino, de);
+				spin_lock(&proc_subdir_lock);
 				break;
 			}
 		}
 	}
+	spin_unlock(&proc_subdir_lock);
 	unlock_kernel();
 
 	if (inode) {
@@ -447,11 +460,13 @@ int proc_readdir(struct file * filp,
 			filp->f_pos++;
 			/* fall through */
 		default:
+			spin_lock(&proc_subdir_lock);
 			de = de->subdir;
 			i -= 2;
 			for (;;) {
 				if (!de) {
 					ret = 1;
+					spin_unlock(&proc_subdir_lock);
 					goto out;
 				}
 				if (!i)
@@ -461,12 +476,16 @@ int proc_readdir(struct file * filp,
 			}
 
 			do {
+				/* filldir passes info to user space */
+				spin_unlock(&proc_subdir_lock);
 				if (filldir(dirent, de->name, de->namelen, filp->f_pos,
 					    de->low_ino, de->mode >> 12) < 0)
 					goto out;
+				spin_lock(&proc_subdir_lock);
 				filp->f_pos++;
 				de = de->next;
 			} while (de);
+			spin_unlock(&proc_subdir_lock);
 	}
 	ret = 1;
 out:	unlock_kernel();
@@ -500,9 +519,13 @@ static int proc_register(struct proc_dir_entry * dir, struct proc_dir_entry * dp
 	if (i == 0)
 		return -EAGAIN;
 	dp->low_ino = i;
+
+	spin_lock(&proc_subdir_lock);
 	dp->next = dir->subdir;
 	dp->parent = dir;
 	dir->subdir = dp;
+	spin_unlock(&proc_subdir_lock);
+
 	if (S_ISDIR(dp->mode)) {
 		if (dp->proc_iops == NULL) {
 			dp->proc_fops = &proc_dir_operations;
@@ -537,7 +560,7 @@ static void proc_kill_inodes(struct proc_dir_entry *de)
 		struct file * filp = list_entry(p, struct file, f_u.fu_list);
 		struct dentry * dentry = filp->f_dentry;
 		struct inode * inode;
-		struct file_operations *fops;
+		const struct file_operations *fops;
 
 		if (dentry->d_op != &proc_dentry_operations)
 			continue;
@@ -694,6 +717,8 @@ void remove_proc_entry(const char *name, struct proc_dir_entry *parent)
 	if (!parent && xlate_proc_name(name, &parent, &fn) != 0)
 		goto out;
 	len = strlen(fn);
+
+	spin_lock(&proc_subdir_lock);
 	for (p = &parent->subdir; *p; p=&(*p)->next ) {
 		if (!proc_match(len, fn, *p))
 			continue;
@@ -714,6 +739,7 @@ void remove_proc_entry(const char *name, struct proc_dir_entry *parent)
 		}
 		break;
 	}
+	spin_unlock(&proc_subdir_lock);
 out:
 	return;
 }
diff --git a/fs/proc/inode.c b/fs/proc/inode.c
index 075d3e945602..722b9c463111 100644
--- a/fs/proc/inode.c
+++ b/fs/proc/inode.c
@@ -121,7 +121,8 @@ int __init proc_init_inodecache(void)
 {
 	proc_inode_cachep = kmem_cache_create("proc_inode_cache",
 					     sizeof(struct proc_inode),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (proc_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/proc/internal.h b/fs/proc/internal.h
index 95a1cf32b838..0502f17b860d 100644
--- a/fs/proc/internal.h
+++ b/fs/proc/internal.h
@@ -30,7 +30,7 @@ do {						\
 
 #endif
 
-extern void create_seq_entry(char *name, mode_t mode, struct file_operations *f);
+extern void create_seq_entry(char *name, mode_t mode, const struct file_operations *f);
 extern int proc_exe_link(struct inode *, struct dentry **, struct vfsmount **);
 extern int proc_tid_stat(struct task_struct *,  char *);
 extern int proc_tgid_stat(struct task_struct *, char *);
diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c
index adc2cd95169a..17f6e8fa1397 100644
--- a/fs/proc/kcore.c
+++ b/fs/proc/kcore.c
@@ -31,7 +31,7 @@ static int open_kcore(struct inode * inode, struct file * filp)
 
 static ssize_t read_kcore(struct file *, char __user *, size_t, loff_t *);
 
-struct file_operations proc_kcore_operations = {
+const struct file_operations proc_kcore_operations = {
 	.read		= read_kcore,
 	.open		= open_kcore,
 };
diff --git a/fs/proc/kmsg.c b/fs/proc/kmsg.c
index 10d37bf25206..ff3b90b56e9d 100644
--- a/fs/proc/kmsg.c
+++ b/fs/proc/kmsg.c
@@ -47,7 +47,7 @@ static unsigned int kmsg_poll(struct file *file, poll_table *wait)
 }
 
 
-struct file_operations proc_kmsg_operations = {
+const struct file_operations proc_kmsg_operations = {
 	.read		= kmsg_read,
 	.poll		= kmsg_poll,
 	.open		= kmsg_open,
diff --git a/fs/proc/proc_devtree.c b/fs/proc/proc_devtree.c
index 9bdd077d6f55..abdf068bc27f 100644
--- a/fs/proc/proc_devtree.c
+++ b/fs/proc/proc_devtree.c
@@ -52,7 +52,8 @@ static int property_read_proc(char *page, char **start, off_t off,
  * Add a property to a node
  */
 static struct proc_dir_entry *
-__proc_device_tree_add_prop(struct proc_dir_entry *de, struct property *pp)
+__proc_device_tree_add_prop(struct proc_dir_entry *de, struct property *pp,
+		const char *name)
 {
 	struct proc_dir_entry *ent;
 
@@ -60,14 +61,14 @@ __proc_device_tree_add_prop(struct proc_dir_entry *de, struct property *pp)
 	 * Unfortunately proc_register puts each new entry
 	 * at the beginning of the list.  So we rearrange them.
 	 */
-	ent = create_proc_read_entry(pp->name,
-				     strncmp(pp->name, "security-", 9)
+	ent = create_proc_read_entry(name,
+				     strncmp(name, "security-", 9)
 				     ? S_IRUGO : S_IRUSR, de,
 				     property_read_proc, pp);
 	if (ent == NULL)
 		return NULL;
 
-	if (!strncmp(pp->name, "security-", 9))
+	if (!strncmp(name, "security-", 9))
 		ent->size = 0; /* don't leak number of password chars */
 	else
 		ent->size = pp->length;
@@ -78,7 +79,7 @@ __proc_device_tree_add_prop(struct proc_dir_entry *de, struct property *pp)
 
 void proc_device_tree_add_prop(struct proc_dir_entry *pde, struct property *prop)
 {
-	__proc_device_tree_add_prop(pde, prop);
+	__proc_device_tree_add_prop(pde, prop, prop->name);
 }
 
 void proc_device_tree_remove_prop(struct proc_dir_entry *pde,
@@ -106,6 +107,69 @@ void proc_device_tree_update_prop(struct proc_dir_entry *pde,
 }
 
 /*
+ * Various dodgy firmware might give us nodes and/or properties with
+ * conflicting names. That's generally ok, except for exporting via /proc,
+ * so munge names here to ensure they're unique.
+ */
+
+static int duplicate_name(struct proc_dir_entry *de, const char *name)
+{
+	struct proc_dir_entry *ent;
+	int found = 0;
+
+	spin_lock(&proc_subdir_lock);
+
+	for (ent = de->subdir; ent != NULL; ent = ent->next) {
+		if (strcmp(ent->name, name) == 0) {
+			found = 1;
+			break;
+		}
+	}
+
+	spin_unlock(&proc_subdir_lock);
+
+	return found;
+}
+
+static const char *fixup_name(struct device_node *np, struct proc_dir_entry *de,
+		const char *name)
+{
+	char *fixed_name;
+	int fixup_len = strlen(name) + 2 + 1; /* name + #x + \0 */
+	int i = 1, size;
+
+realloc:
+	fixed_name = kmalloc(fixup_len, GFP_KERNEL);
+	if (fixed_name == NULL) {
+		printk(KERN_ERR "device-tree: Out of memory trying to fixup "
+				"name \"%s\"\n", name);
+		return name;
+	}
+
+retry:
+	size = snprintf(fixed_name, fixup_len, "%s#%d", name, i);
+	size++; /* account for NULL */
+
+	if (size > fixup_len) {
+		/* We ran out of space, free and reallocate. */
+		kfree(fixed_name);
+		fixup_len = size;
+		goto realloc;
+	}
+
+	if (duplicate_name(de, fixed_name)) {
+		/* Multiple duplicates. Retry with a different offset. */
+		i++;
+		goto retry;
+	}
+
+	printk(KERN_WARNING "device-tree: Duplicate name in %s, "
+			"renamed to \"%s\"\n", np->full_name, fixed_name);
+
+	return fixed_name;
+}
+
+/*
  * Process a node, adding entries for its children and its properties.
  */
 void proc_device_tree_add_node(struct device_node *np,
@@ -118,35 +182,30 @@ void proc_device_tree_add_node(struct device_node *np,
 
 	set_node_proc_entry(np, de);
 	for (child = NULL; (child = of_get_next_child(np, child));) {
+		/* Use everything after the last slash, or the full name */
 		p = strrchr(child->full_name, '/');
 		if (!p)
 			p = child->full_name;
 		else
 			++p;
+
+		if (duplicate_name(de, p))
+			p = fixup_name(np, de, p);
+
 		ent = proc_mkdir(p, de);
 		if (ent == 0)
 			break;
 		proc_device_tree_add_node(child, ent);
 	}
 	of_node_put(child);
+
 	for (pp = np->properties; pp != 0; pp = pp->next) {
-		/*
-		 * Yet another Apple device-tree bogosity: on some machines,
-		 * they have properties & nodes with the same name. Those
-		 * properties are quite unimportant for us though, thus we
-		 * simply "skip" them here, but we do have to check.
-		 */
-		for (ent = de->subdir; ent != NULL; ent = ent->next)
-			if (!strcmp(ent->name, pp->name))
-				break;
-		if (ent != NULL) {
-			printk(KERN_WARNING "device-tree: property \"%s\" name"
-			       " conflicts with node in %s\n", pp->name,
-			       np->full_name);
-			continue;
-		}
+		p = pp->name;
+
+		if (duplicate_name(de, p))
+			p = fixup_name(np, de, p);
 
-		ent = __proc_device_tree_add_prop(de, pp);
+		ent = __proc_device_tree_add_prop(de, pp, p);
 		if (ent == 0)
 			break;
 	}
diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c
index 1d24fead51a6..ef5a3323f4b5 100644
--- a/fs/proc/proc_misc.c
+++ b/fs/proc/proc_misc.c
@@ -312,7 +312,7 @@ static void *devinfo_next(struct seq_file *f, void *v, loff_t *pos)
 		case BLK_HDR:
 			info->state = BLK_LIST;
 			(*pos)++;
-			break;
+			/*fallthrough*/
 		case BLK_LIST:
 			if (get_blkdev_info(info->blkdev,&idummy,&ndummy)) {
 				/*
@@ -485,6 +485,40 @@ static struct file_operations proc_slabinfo_operations = {
 	.llseek		= seq_lseek,
 	.release	= seq_release,
 };
+
+#ifdef CONFIG_DEBUG_SLAB_LEAK
+extern struct seq_operations slabstats_op;
+static int slabstats_open(struct inode *inode, struct file *file)
+{
+	unsigned long *n = kzalloc(PAGE_SIZE, GFP_KERNEL);
+	int ret = -ENOMEM;
+	if (n) {
+		ret = seq_open(file, &slabstats_op);
+		if (!ret) {
+			struct seq_file *m = file->private_data;
+			*n = PAGE_SIZE / (2 * sizeof(unsigned long));
+			m->private = n;
+			n = NULL;
+		}
+		kfree(n);
+	}
+	return ret;
+}
+
+static int slabstats_release(struct inode *inode, struct file *file)
+{
+	struct seq_file *m = file->private_data;
+	kfree(m->private);
+	return seq_release(inode, file);
+}
+
+static struct file_operations proc_slabstats_operations = {
+	.open		= slabstats_open,
+	.read		= seq_read,
+	.llseek		= seq_lseek,
+	.release	= slabstats_release,
+};
+#endif
 #endif
 
 static int show_stat(struct seq_file *p, void *v)
@@ -500,7 +534,7 @@ static int show_stat(struct seq_file *p, void *v)
 	if (wall_to_monotonic.tv_nsec)
 		--jif;
 
-	for_each_cpu(i) {
+	for_each_possible_cpu(i) {
 		int j;
 
 		user = cputime64_add(user, kstat_cpu(i).cpustat.user);
@@ -697,7 +731,7 @@ static struct file_operations proc_sysrq_trigger_operations = {
 
 struct proc_dir_entry *proc_root_kcore;
 
-void create_seq_entry(char *name, mode_t mode, struct file_operations *f)
+void create_seq_entry(char *name, mode_t mode, const struct file_operations *f)
 {
 	struct proc_dir_entry *entry;
 	entry = create_proc_entry(name, mode, NULL);
@@ -744,6 +778,9 @@ void __init proc_misc_init(void)
 	create_seq_entry("interrupts", 0, &proc_interrupts_operations);
 #ifdef CONFIG_SLAB
 	create_seq_entry("slabinfo",S_IWUSR|S_IRUGO,&proc_slabinfo_operations);
+#ifdef CONFIG_DEBUG_SLAB_LEAK
+	create_seq_entry("slab_allocators", 0 ,&proc_slabstats_operations);
+#endif
 #endif
 	create_seq_entry("buddyinfo",S_IRUGO, &fragmentation_file_operations);
 	create_seq_entry("vmstat",S_IRUGO, &proc_vmstat_file_operations);
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 4063fb32f78c..7efa73d44c9a 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -172,7 +172,7 @@ static int open_vmcore(struct inode *inode, struct file *filp)
 	return 0;
 }
 
-struct file_operations proc_vmcore_operations = {
+const struct file_operations proc_vmcore_operations = {
 	.read		= read_vmcore,
 	.open		= open_vmcore,
 };
diff --git a/fs/qnx4/dir.c b/fs/qnx4/dir.c
index 7a8f5595c26f..9031948fefd0 100644
--- a/fs/qnx4/dir.c
+++ b/fs/qnx4/dir.c
@@ -81,7 +81,7 @@ out:
 	return 0;
 }
 
-struct file_operations qnx4_dir_operations =
+const struct file_operations qnx4_dir_operations =
 {
 	.read		= generic_read_dir,
 	.readdir	= qnx4_readdir,
diff --git a/fs/qnx4/file.c b/fs/qnx4/file.c
index b471315e24ef..62af4b1348bd 100644
--- a/fs/qnx4/file.c
+++ b/fs/qnx4/file.c
@@ -12,17 +12,14 @@
  * 27-06-1998 by Frank Denis : file overwriting.
  */
 
-#include <linux/config.h>
-#include <linux/types.h>
 #include <linux/fs.h>
-#include <linux/time.h>
 #include <linux/qnx4_fs.h>
 
 /*
  * We have mostly NULL's here: the current defaults are ok for
  * the qnx4 filesystem.
  */
-struct file_operations qnx4_file_operations =
+const struct file_operations qnx4_file_operations =
 {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
diff --git a/fs/qnx4/inode.c b/fs/qnx4/inode.c
index 80f32911c0cb..2ecd46f85e9f 100644
--- a/fs/qnx4/inode.c
+++ b/fs/qnx4/inode.c
@@ -546,7 +546,8 @@ static int init_inodecache(void)
 {
 	qnx4_inode_cachep = kmem_cache_create("qnx4_inode_cache",
 					     sizeof(struct qnx4_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (qnx4_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/quota.c b/fs/quota.c
index ba9e0bf32f67..d6a2be826e29 100644
--- a/fs/quota.c
+++ b/fs/quota.c
@@ -170,10 +170,10 @@ static void quota_sync_sb(struct super_block *sb, int type)
 
 	/* Now when everything is written we can discard the pagecache so
 	 * that userspace sees the changes. We need i_mutex and so we could
-	 * not do it inside dqonoff_sem. Moreover we need to be carefull
+	 * not do it inside dqonoff_mutex. Moreover we need to be carefull
 	 * about races with quotaoff() (that is the reason why we have own
 	 * reference to inode). */
-	down(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_lock(&sb_dqopt(sb)->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		discard[cnt] = NULL;
 		if (type != -1 && cnt != type)
@@ -182,7 +182,7 @@ static void quota_sync_sb(struct super_block *sb, int type)
 			continue;
 		discard[cnt] = igrab(sb_dqopt(sb)->files[cnt]);
 	}
-	up(&sb_dqopt(sb)->dqonoff_sem);
+	mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex);
 	for (cnt = 0; cnt < MAXQUOTAS; cnt++) {
 		if (discard[cnt]) {
 			mutex_lock(&discard[cnt]->i_mutex);
diff --git a/fs/quota_v2.c b/fs/quota_v2.c
index b4199ec3ece4..c519a583e681 100644
--- a/fs/quota_v2.c
+++ b/fs/quota_v2.c
@@ -394,7 +394,7 @@ static int v2_write_dquot(struct dquot *dquot)
 	ssize_t ret;
 	struct v2_disk_dqblk ddquot, empty;
 
-	/* dq_off is guarded by dqio_sem */
+	/* dq_off is guarded by dqio_mutex */
 	if (!dquot->dq_off)
 		if ((ret = dq_insert_tree(dquot)) < 0) {
 			printk(KERN_ERR "VFS: Error %zd occurred while creating quota.\n", ret);
diff --git a/fs/ramfs/file-mmu.c b/fs/ramfs/file-mmu.c
index 2115383dcc8d..00a933eb820c 100644
--- a/fs/ramfs/file-mmu.c
+++ b/fs/ramfs/file-mmu.c
@@ -24,18 +24,7 @@
  * caches is sufficient.
  */
 
-#include <linux/module.h>
 #include <linux/fs.h>
-#include <linux/pagemap.h>
-#include <linux/highmem.h>
-#include <linux/init.h>
-#include <linux/string.h>
-#include <linux/smp_lock.h>
-#include <linux/backing-dev.h>
-#include <linux/ramfs.h>
-
-#include <asm/uaccess.h>
-#include "internal.h"
 
 struct address_space_operations ramfs_aops = {
 	.readpage	= simple_readpage,
@@ -43,7 +32,7 @@ struct address_space_operations ramfs_aops = {
 	.commit_write	= simple_commit_write
 };
 
-struct file_operations ramfs_file_operations = {
+const struct file_operations ramfs_file_operations = {
 	.read		= generic_file_read,
 	.write		= generic_file_write,
 	.mmap		= generic_file_mmap,
diff --git a/fs/ramfs/file-nommu.c b/fs/ramfs/file-nommu.c
index 3f810acd0bfa..f443a84b98a5 100644
--- a/fs/ramfs/file-nommu.c
+++ b/fs/ramfs/file-nommu.c
@@ -33,7 +33,7 @@ struct address_space_operations ramfs_aops = {
 	.commit_write		= simple_commit_write
 };
 
-struct file_operations ramfs_file_operations = {
+const struct file_operations ramfs_file_operations = {
 	.mmap			= ramfs_nommu_mmap,
 	.get_unmapped_area	= ramfs_nommu_get_unmapped_area,
 	.read			= generic_file_read,
@@ -87,8 +87,7 @@ static int ramfs_nommu_expand_for_mapping(struct inode *inode, size_t newsize)
 	xpages = 1UL << order;
 	npages = (newsize + PAGE_SIZE - 1) >> PAGE_SHIFT;
 
-	for (loop = 0; loop < npages; loop++)
-		set_page_count(pages + loop, 1);
+	split_page(pages, order);
 
 	/* trim off any pages we don't actually require */
 	for (loop = npages; loop < xpages; loop++)
diff --git a/fs/ramfs/internal.h b/fs/ramfs/internal.h
index 272c8a7120b0..313237631b49 100644
--- a/fs/ramfs/internal.h
+++ b/fs/ramfs/internal.h
@@ -11,5 +11,5 @@
 
 
 extern struct address_space_operations ramfs_aops;
-extern struct file_operations ramfs_file_operations;
+extern const struct file_operations ramfs_file_operations;
 extern struct inode_operations ramfs_file_inode_operations;
diff --git a/fs/read_write.c b/fs/read_write.c
index 3f7a1a62165f..6256ca81a718 100644
--- a/fs/read_write.c
+++ b/fs/read_write.c
@@ -19,7 +19,7 @@
 #include <asm/uaccess.h>
 #include <asm/unistd.h>
 
-struct file_operations generic_ro_fops = {
+const struct file_operations generic_ro_fops = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.mmap		= generic_file_readonly_mmap,
@@ -470,7 +470,7 @@ static ssize_t do_readv_writev(int type, struct file *file,
 	 * verify all the pointers
 	 */
 	ret = -EINVAL;
-	if ((nr_segs > UIO_MAXIOV) || (nr_segs <= 0))
+	if (nr_segs > UIO_MAXIOV)
 		goto out;
 	if (!file->f_op)
 		goto out;
diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c
index d71ac6579289..973c819f8033 100644
--- a/fs/reiserfs/dir.c
+++ b/fs/reiserfs/dir.c
@@ -18,7 +18,7 @@ static int reiserfs_readdir(struct file *, void *, filldir_t);
 static int reiserfs_dir_fsync(struct file *filp, struct dentry *dentry,
 			      int datasync);
 
-struct file_operations reiserfs_dir_operations = {
+const struct file_operations reiserfs_dir_operations = {
 	.read = generic_read_dir,
 	.readdir = reiserfs_readdir,
 	.fsync = reiserfs_dir_fsync,
diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c
index be12879bb179..cf6e1cf40351 100644
--- a/fs/reiserfs/file.c
+++ b/fs/reiserfs/file.c
@@ -1532,7 +1532,7 @@ static ssize_t reiserfs_file_write(struct file *file,	/* the file we are going t
 		buf += write_bytes;
 		*ppos = pos += write_bytes;
 		count -= write_bytes;
-		balance_dirty_pages_ratelimited(inode->i_mapping);
+		balance_dirty_pages_ratelimited_nr(inode->i_mapping, num_pages);
 	}
 
 	/* this is only true on error */
@@ -1546,10 +1546,10 @@ static ssize_t reiserfs_file_write(struct file *file,	/* the file we are going t
 		}
 	}
 
-	if ((file->f_flags & O_SYNC) || IS_SYNC(inode))
-		res =
-		    generic_osync_inode(inode, file->f_mapping,
-					OSYNC_METADATA | OSYNC_DATA);
+	if (likely(res >= 0) &&
+	    (unlikely((file->f_flags & O_SYNC) || IS_SYNC(inode))))
+		res = generic_osync_inode(inode, file->f_mapping,
+		                          OSYNC_METADATA | OSYNC_DATA);
 
 	mutex_unlock(&inode->i_mutex);
 	reiserfs_async_progress_wait(inode->i_sb);
@@ -1566,7 +1566,7 @@ static ssize_t reiserfs_aio_write(struct kiocb *iocb, const char __user * buf,
 	return generic_file_aio_write(iocb, buf, count, pos);
 }
 
-struct file_operations reiserfs_file_operations = {
+const struct file_operations reiserfs_file_operations = {
 	.read = generic_file_read,
 	.write = reiserfs_file_write,
 	.ioctl = reiserfs_ioctl,
@@ -1576,6 +1576,8 @@ struct file_operations reiserfs_file_operations = {
 	.sendfile = generic_file_sendfile,
 	.aio_read = generic_file_aio_read,
 	.aio_write = reiserfs_aio_write,
+	.splice_read = generic_file_splice_read,
+	.splice_write = generic_file_splice_write,
 };
 
 struct inode_operations reiserfs_file_inode_operations = {
diff --git a/fs/reiserfs/fix_node.c b/fs/reiserfs/fix_node.c
index aa22588019ec..5600d3d60cf7 100644
--- a/fs/reiserfs/fix_node.c
+++ b/fs/reiserfs/fix_node.c
@@ -191,9 +191,7 @@ static void create_virtual_node(struct tree_balance *tb, int h)
 					       "vs-8045: create_virtual_node: rdkey %k, affected item==%d (mode==%c) Must be %c",
 					       key, vn->vn_affected_item_num,
 					       vn->vn_mode, M_DELETE);
-			} else
-				/* we can delete directory item, that has only one directory entry in it */
-				;
+			}
 		}
 #endif
 
diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
index d60f6238c66a..9857e50f85e7 100644
--- a/fs/reiserfs/inode.c
+++ b/fs/reiserfs/inode.c
@@ -466,7 +466,6 @@ static int reiserfs_get_block_create_0(struct inode *inode, sector_t block,
    direct_IO request. */
 static int reiserfs_get_blocks_direct_io(struct inode *inode,
 					 sector_t iblock,
-					 unsigned long max_blocks,
 					 struct buffer_head *bh_result,
 					 int create)
 {
@@ -2793,7 +2792,7 @@ static int invalidatepage_can_drop(struct inode *inode, struct buffer_head *bh)
 }
 
 /* clm -- taken from fs/buffer.c:block_invalidate_page */
-static int reiserfs_invalidatepage(struct page *page, unsigned long offset)
+static void reiserfs_invalidatepage(struct page *page, unsigned long offset)
 {
 	struct buffer_head *head, *bh, *next;
 	struct inode *inode = page->mapping->host;
@@ -2832,10 +2831,12 @@ static int reiserfs_invalidatepage(struct page *page, unsigned long offset)
 	 * The get_block cached value has been unconditionally invalidated,
 	 * so real IO is not possible anymore.
 	 */
-	if (!offset && ret)
+	if (!offset && ret) {
 		ret = try_to_release_page(page, 0);
+		/* maybe should BUG_ON(!ret); - neilb */
+	}
       out:
-	return ret;
+	return;
 }
 
 static int reiserfs_set_page_dirty(struct page *page)
diff --git a/fs/reiserfs/item_ops.c b/fs/reiserfs/item_ops.c
index e237cd668e5b..7a88adbceef6 100644
--- a/fs/reiserfs/item_ops.c
+++ b/fs/reiserfs/item_ops.c
@@ -275,7 +275,7 @@ static void indirect_print_item(struct item_head *ih, char *item)
 	int j;
 	__le32 *unp;
 	__u32 prev = INT_MAX;
-	int num;
+	int num = 0;
 
 	unp = (__le32 *) item;
 
diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c
index 5a9d2722fa0a..1b73529b8099 100644
--- a/fs/reiserfs/journal.c
+++ b/fs/reiserfs/journal.c
@@ -2227,6 +2227,9 @@ static int journal_read_transaction(struct super_block *p_s_sb,
 	journal->j_start = cur_dblock - SB_ONDISK_JOURNAL_1st_BLOCK(p_s_sb);
 	journal->j_last_flush_trans_id = trans_id;
 	journal->j_trans_id = trans_id + 1;
+	/* check for trans_id overflow */
+	if (journal->j_trans_id == 0)
+		journal->j_trans_id = 10;
 	brelse(c_bh);
 	brelse(d_bh);
 	kfree(log_blocks);
@@ -2450,6 +2453,9 @@ static int journal_read(struct super_block *p_s_sb)
 		journal->j_start = le32_to_cpu(jh->j_first_unflushed_offset);
 		journal->j_trans_id =
 		    le32_to_cpu(jh->j_last_flush_trans_id) + 1;
+		/* check for trans_id overflow */
+		if (journal->j_trans_id == 0)
+			journal->j_trans_id = 10;
 		journal->j_last_flush_trans_id =
 		    le32_to_cpu(jh->j_last_flush_trans_id);
 		journal->j_mount_id = le32_to_cpu(jh->j_mount_id) + 1;
@@ -3873,8 +3879,8 @@ static int do_journal_end(struct reiserfs_transaction_handle *th,
 	int cur_write_start = 0;	/* start index of current log write */
 	int old_start;
 	int i;
-	int flush = flags & FLUSH_ALL;
-	int wait_on_commit = flags & WAIT;
+	int flush;
+	int wait_on_commit;
 	struct reiserfs_journal_list *jl, *temp_jl;
 	struct list_head *entry, *safe;
 	unsigned long jindex;
@@ -3884,6 +3890,13 @@ static int do_journal_end(struct reiserfs_transaction_handle *th,
 	BUG_ON(th->t_refcount > 1);
 	BUG_ON(!th->t_trans_id);
 
+	/* protect flush_older_commits from doing mistakes if the
+           transaction ID counter gets overflowed.  */
+	if (th->t_trans_id == ~0UL)
+		flags |= FLUSH_ALL | COMMIT_NOW | WAIT;
+	flush = flags & FLUSH_ALL;
+	wait_on_commit = flags & WAIT;
+
 	put_fs_excl();
 	current->journal_info = th->t_handle_save;
 	reiserfs_check_lock_depth(p_s_sb, "journal end");
@@ -4105,7 +4118,9 @@ static int do_journal_end(struct reiserfs_transaction_handle *th,
 	journal->j_first = NULL;
 	journal->j_len = 0;
 	journal->j_trans_start_time = 0;
-	journal->j_trans_id++;
+	/* check for trans_id overflow */
+	if (++journal->j_trans_id == 0)
+		journal->j_trans_id = 10;
 	journal->j_current_jl->j_trans_id = journal->j_trans_id;
 	journal->j_must_wait = 0;
 	journal->j_len_alloc = 0;
diff --git a/fs/reiserfs/prints.c b/fs/reiserfs/prints.c
index d55e164bd5c2..27bd3a1df2ad 100644
--- a/fs/reiserfs/prints.c
+++ b/fs/reiserfs/prints.c
@@ -143,7 +143,7 @@ static void sprintf_buffer_head(char *buf, struct buffer_head *bh)
 	char b[BDEVNAME_SIZE];
 
 	sprintf(buf,
-		"dev %s, size %d, blocknr %llu, count %d, state 0x%lx, page %p, (%s, %s, %s)",
+		"dev %s, size %zd, blocknr %llu, count %d, state 0x%lx, page %p, (%s, %s, %s)",
 		bdevname(bh->b_bdev, b), bh->b_size,
 		(unsigned long long)bh->b_blocknr, atomic_read(&(bh->b_count)),
 		bh->b_state, bh->b_page,
@@ -601,8 +601,7 @@ void store_print_tb(struct tree_balance *tb)
 		tb->tb_mode, PATH_LAST_POSITION(tb->tb_path),
 		tb->tb_path->pos_in_item);
 
-	for (h = 0; h < sizeof(tb->insert_size) / sizeof(tb->insert_size[0]);
-	     h++) {
+	for (h = 0; h < ARRAY_SIZE(tb->insert_size); h++) {
 		if (PATH_H_PATH_OFFSET(tb->tb_path, h) <=
 		    tb->tb_path->path_length
 		    && PATH_H_PATH_OFFSET(tb->tb_path,
@@ -658,15 +657,13 @@ void store_print_tb(struct tree_balance *tb)
 
 	/* print FEB list (list of buffers in form (bh (b_blocknr, b_count), that will be used for new nodes) */
 	h = 0;
-	for (i = 0; i < sizeof(tb->FEB) / sizeof(tb->FEB[0]); i++)
+	for (i = 0; i < ARRAY_SIZE(tb->FEB); i++)
 		sprintf(print_tb_buf + strlen(print_tb_buf),
 			"%p (%llu %d)%s", tb->FEB[i],
 			tb->FEB[i] ? (unsigned long long)tb->FEB[i]->
 			b_blocknr : 0ULL,
 			tb->FEB[i] ? atomic_read(&(tb->FEB[i]->b_count)) : 0,
-			(i ==
-			 sizeof(tb->FEB) / sizeof(tb->FEB[0]) -
-			 1) ? "\n" : ", ");
+			(i == ARRAY_SIZE(tb->FEB) - 1) ? "\n" : ", ");
 
 	sprintf(print_tb_buf + strlen(print_tb_buf),
 		"======================== the end ====================================\n");
diff --git a/fs/reiserfs/procfs.c b/fs/reiserfs/procfs.c
index ef6caed9336b..731688e1cfe3 100644
--- a/fs/reiserfs/procfs.c
+++ b/fs/reiserfs/procfs.c
@@ -470,7 +470,7 @@ static int r_open(struct inode *inode, struct file *file)
 	return ret;
 }
 
-static struct file_operations r_file_operations = {
+static const struct file_operations r_file_operations = {
 	.open = r_open,
 	.read = seq_read,
 	.llseek = seq_lseek,
diff --git a/fs/reiserfs/stree.c b/fs/reiserfs/stree.c
index e2d08d7bcffc..d2b25e1ba6e9 100644
--- a/fs/reiserfs/stree.c
+++ b/fs/reiserfs/stree.c
@@ -981,6 +981,8 @@ static inline int prepare_for_direntry_item(struct path *path,
 	return M_CUT;
 }
 
+#define JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD (2 * JOURNAL_PER_BALANCE_CNT + 1)
+
 /*  If the path points to a directory or direct item, calculate mode and the size cut, for balance.
     If the path points to an indirect item, remove some number of its unformatted nodes.
     In case of file truncate calculate whether this item must be deleted/truncated or last
@@ -1020,148 +1022,79 @@ static char prepare_for_delete_or_cut(struct reiserfs_transaction_handle *th, st
 
 	/* Case of an indirect item. */
 	{
-		int n_unfm_number,	/* Number of the item unformatted nodes. */
-		 n_counter, n_blk_size;
-		__le32 *p_n_unfm_pointer;	/* Pointer to the unformatted node number. */
-		__u32 tmp;
-		struct item_head s_ih;	/* Item header. */
-		char c_mode;	/* Returned mode of the balance. */
-		int need_research;
-
-		n_blk_size = p_s_sb->s_blocksize;
-
-		/* Search for the needed object indirect item until there are no unformatted nodes to be removed. */
-		do {
-			need_research = 0;
-			p_s_bh = PATH_PLAST_BUFFER(p_s_path);
-			/* Copy indirect item header to a temp variable. */
-			copy_item_head(&s_ih, PATH_PITEM_HEAD(p_s_path));
-			/* Calculate number of unformatted nodes in this item. */
-			n_unfm_number = I_UNFM_NUM(&s_ih);
-
-			RFALSE(!is_indirect_le_ih(&s_ih) || !n_unfm_number ||
-			       pos_in_item(p_s_path) + 1 != n_unfm_number,
-			       "PAP-5240: invalid item %h "
-			       "n_unfm_number = %d *p_n_pos_in_item = %d",
-			       &s_ih, n_unfm_number, pos_in_item(p_s_path));
-
-			/* Calculate balance mode and position in the item to remove unformatted nodes. */
-			if (n_new_file_length == max_reiserfs_offset(inode)) {	/* Case of delete. */
-				pos_in_item(p_s_path) = 0;
-				*p_n_cut_size = -(IH_SIZE + ih_item_len(&s_ih));
-				c_mode = M_DELETE;
-			} else {	/* Case of truncate. */
-				if (n_new_file_length < le_ih_k_offset(&s_ih)) {
-					pos_in_item(p_s_path) = 0;
-					*p_n_cut_size =
-					    -(IH_SIZE + ih_item_len(&s_ih));
-					c_mode = M_DELETE;	/* Delete this item. */
-				} else {
-					/* indirect item must be truncated starting from *p_n_pos_in_item-th position */
-					pos_in_item(p_s_path) =
-					    (n_new_file_length + n_blk_size -
-					     le_ih_k_offset(&s_ih)) >> p_s_sb->
-					    s_blocksize_bits;
-
-					RFALSE(pos_in_item(p_s_path) >
-					       n_unfm_number,
-					       "PAP-5250: invalid position in the item");
-
-					/* Either convert last unformatted node of indirect item to direct item or increase
-					   its free space.  */
-					if (pos_in_item(p_s_path) ==
-					    n_unfm_number) {
-						*p_n_cut_size = 0;	/* Nothing to cut. */
-						return M_CONVERT;	/* Maybe convert last unformatted node to the direct item. */
-					}
-					/* Calculate size to cut. */
-					*p_n_cut_size =
-					    -(ih_item_len(&s_ih) -
-					      pos_in_item(p_s_path) *
-					      UNFM_P_SIZE);
-
-					c_mode = M_CUT;	/* Cut from this indirect item. */
-				}
-			}
+	    int blk_size = p_s_sb->s_blocksize;
+	    struct item_head s_ih;
+	    int need_re_search;
+	    int delete = 0;
+	    int result = M_CUT;
+	    int pos = 0;
+
+	    if ( n_new_file_length == max_reiserfs_offset (inode) ) {
+		/* prepare_for_delete_or_cut() is called by
+		 * reiserfs_delete_item() */
+		n_new_file_length = 0;
+		delete = 1;
+	    }
+
+	    do {
+		need_re_search = 0;
+		*p_n_cut_size = 0;
+		p_s_bh = PATH_PLAST_BUFFER(p_s_path);
+		copy_item_head(&s_ih, PATH_PITEM_HEAD(p_s_path));
+		pos = I_UNFM_NUM(&s_ih);
 
-			RFALSE(n_unfm_number <= pos_in_item(p_s_path),
-			       "PAP-5260: invalid position in the indirect item");
-
-			/* pointers to be cut */
-			n_unfm_number -= pos_in_item(p_s_path);
-			/* Set pointer to the last unformatted node pointer that is to be cut. */
-			p_n_unfm_pointer =
-			    (__le32 *) B_I_PITEM(p_s_bh,
-						 &s_ih) + I_UNFM_NUM(&s_ih) -
-			    1 - *p_n_removed;
-
-			/* We go through the unformatted nodes pointers of the indirect
-			   item and look for the unformatted nodes in the cache. If we
-			   found some of them we free it, zero corresponding indirect item
-			   entry and log buffer containing that indirect item. For this we
-			   need to prepare last path element for logging. If some
-			   unformatted node has b_count > 1 we must not free this
-			   unformatted node since it is in use. */
-			reiserfs_prepare_for_journal(p_s_sb, p_s_bh, 1);
-			// note: path could be changed, first line in for loop takes care
-			// of it
+		while (le_ih_k_offset (&s_ih) + (pos - 1) * blk_size > n_new_file_length) {
+		    __u32 *unfm, block;
 
-			for (n_counter = *p_n_removed;
-			     n_counter < n_unfm_number;
-			     n_counter++, p_n_unfm_pointer--) {
+		    /* Each unformatted block deletion may involve one additional
+		     * bitmap block into the transaction, thereby the initial
+		     * journal space reservation might not be enough. */
+		    if (!delete && (*p_n_cut_size) != 0 &&
+			reiserfs_transaction_free_space(th) < JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
+			break;
+		    }
 
-				cond_resched();
-				if (item_moved(&s_ih, p_s_path)) {
-					need_research = 1;
-					break;
-				}
-				RFALSE(p_n_unfm_pointer <
-				       (__le32 *) B_I_PITEM(p_s_bh, &s_ih)
-				       || p_n_unfm_pointer >
-				       (__le32 *) B_I_PITEM(p_s_bh,
-							    &s_ih) +
-				       I_UNFM_NUM(&s_ih) - 1,
-				       "vs-5265: pointer out of range");
-
-				/* Hole, nothing to remove. */
-				if (!get_block_num(p_n_unfm_pointer, 0)) {
-					(*p_n_removed)++;
-					continue;
-				}
+		    unfm = (__u32 *)B_I_PITEM(p_s_bh, &s_ih) + pos - 1;
+		    block = get_block_num(unfm, 0);
 
-				(*p_n_removed)++;
+		    if (block != 0) {
+			reiserfs_prepare_for_journal(p_s_sb, p_s_bh, 1);
+			put_block_num(unfm, 0, 0);
+			journal_mark_dirty (th, p_s_sb, p_s_bh);
+			reiserfs_free_block(th, inode, block, 1);
+		    }
 
-				tmp = get_block_num(p_n_unfm_pointer, 0);
-				put_block_num(p_n_unfm_pointer, 0, 0);
-				journal_mark_dirty(th, p_s_sb, p_s_bh);
-				reiserfs_free_block(th, inode, tmp, 1);
-				if (item_moved(&s_ih, p_s_path)) {
-					need_research = 1;
-					break;
-				}
-			}
+		    cond_resched();
 
-			/* a trick.  If the buffer has been logged, this
-			 ** will do nothing.  If we've broken the loop without
-			 ** logging it, it will restore the buffer
-			 **
-			 */
-			reiserfs_restore_prepared_buffer(p_s_sb, p_s_bh);
-
-			/* This loop can be optimized. */
-		} while ((*p_n_removed < n_unfm_number || need_research) &&
-			 search_for_position_by_key(p_s_sb, p_s_item_key,
-						    p_s_path) ==
-			 POSITION_FOUND);
-
-		RFALSE(*p_n_removed < n_unfm_number,
-		       "PAP-5310: indirect item is not found");
-		RFALSE(item_moved(&s_ih, p_s_path),
-		       "after while, comp failed, retry");
-
-		if (c_mode == M_CUT)
-			pos_in_item(p_s_path) *= UNFM_P_SIZE;
-		return c_mode;
+		    if (item_moved (&s_ih, p_s_path))  {
+			need_re_search = 1;
+			break;
+		    }
+
+		    pos --;
+		    (*p_n_removed) ++;
+		    (*p_n_cut_size) -= UNFM_P_SIZE;
+
+		    if (pos == 0) {
+			(*p_n_cut_size) -= IH_SIZE;
+			result = M_DELETE;
+			break;
+		    }
+		}
+		/* a trick.  If the buffer has been logged, this will do nothing.  If
+		** we've broken the loop without logging it, it will restore the
+		** buffer */
+		reiserfs_restore_prepared_buffer(p_s_sb, p_s_bh);
+	    } while (need_re_search &&
+		     search_for_position_by_key(p_s_sb, p_s_item_key, p_s_path) == POSITION_FOUND);
+	    pos_in_item(p_s_path) = pos * UNFM_P_SIZE;
+
+	    if (*p_n_cut_size == 0) {
+		/* Nothing were cut. maybe convert last unformatted node to the
+		 * direct item? */
+		result = M_CONVERT;
+	    }
+	    return result;
 	}
 }
 
@@ -1948,7 +1881,8 @@ int reiserfs_do_truncate(struct reiserfs_transaction_handle *th, struct inode *p
 		 ** sure the file is consistent before ending the current trans
 		 ** and starting a new one
 		 */
-		if (journal_transaction_should_end(th, th->t_blocks_allocated)) {
+		if (journal_transaction_should_end(th, 0) ||
+		    reiserfs_transaction_free_space(th) <= JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
 			int orig_len_alloc = th->t_blocks_allocated;
 			decrement_counters_in_path(&s_search_path);
 
@@ -1962,7 +1896,7 @@ int reiserfs_do_truncate(struct reiserfs_transaction_handle *th, struct inode *p
 			if (err)
 				goto out;
 			err = journal_begin(th, p_s_inode->i_sb,
-					    JOURNAL_PER_BALANCE_CNT * 6);
+					    JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD + JOURNAL_PER_BALANCE_CNT * 4) ;
 			if (err)
 				goto out;
 			reiserfs_update_inode_transaction(p_s_inode);
diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c
index d63da756eb49..cae2abbc0c71 100644
--- a/fs/reiserfs/super.c
+++ b/fs/reiserfs/super.c
@@ -521,7 +521,8 @@ static int init_inodecache(void)
 	reiserfs_inode_cachep = kmem_cache_create("reiser_inode_cache",
 						  sizeof(struct
 							 reiserfs_inode_info),
-						  0, SLAB_RECLAIM_ACCOUNT,
+						  0, (SLAB_RECLAIM_ACCOUNT|
+							SLAB_MEM_SPREAD),
 						  init_once, NULL);
 	if (reiserfs_inode_cachep == NULL)
 		return -ENOMEM;
@@ -684,14 +685,14 @@ static const arg_desc_t logging_mode[] = {
 	 (1 << REISERFS_DATA_ORDERED | 1 << REISERFS_DATA_WRITEBACK)},
 	{"writeback", 1 << REISERFS_DATA_WRITEBACK,
 	 (1 << REISERFS_DATA_ORDERED | 1 << REISERFS_DATA_LOG)},
-	{NULL, 0}
+	{.value = NULL}
 };
 
 /* possible values for -o barrier= */
 static const arg_desc_t barrier_mode[] = {
 	{"none", 1 << REISERFS_BARRIER_NONE, 1 << REISERFS_BARRIER_FLUSH},
 	{"flush", 1 << REISERFS_BARRIER_FLUSH, 1 << REISERFS_BARRIER_NONE},
-	{NULL, 0}
+	{.value = NULL}
 };
 
 /* possible values for "-o block-allocator=" and bits which are to be set in
@@ -889,7 +890,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options,	/* strin
 		{"acl",.setmask = 1 << REISERFS_UNSUPPORTED_OPT},
 		{"noacl",.clrmask = 1 << REISERFS_UNSUPPORTED_OPT},
 #endif
-		{"nolog",},	/* This is unsupported */
+		{.option_name = "nolog"},
 		{"replayonly",.setmask = 1 << REPLAYONLY},
 		{"block-allocator",.arg_required = 'a',.values = balloc},
 		{"data",.arg_required = 'd',.values = logging_mode},
@@ -907,7 +908,7 @@ static int reiserfs_parse_options(struct super_block *s, char *options,	/* strin
 		{"grpjquota",.arg_required =
 		 'g' | (1 << REISERFS_OPT_ALLOWEMPTY),.values = NULL},
 		{"jqfmt",.arg_required = 'f',.values = NULL},
-		{NULL,}
+		{.option_name = NULL}
 	};
 
 	*blocks = 0;
diff --git a/fs/reiserfs/xattr_acl.c b/fs/reiserfs/xattr_acl.c
index ab8894c3b9e5..58c418fbca2c 100644
--- a/fs/reiserfs/xattr_acl.c
+++ b/fs/reiserfs/xattr_acl.c
@@ -182,7 +182,7 @@ struct posix_acl *reiserfs_get_acl(struct inode *inode, int type)
 {
 	char *name, *value;
 	struct posix_acl *acl, **p_acl;
-	size_t size;
+	int size;
 	int retval;
 	struct reiserfs_inode_info *reiserfs_i = REISERFS_I(inode);
 
@@ -206,7 +206,7 @@ struct posix_acl *reiserfs_get_acl(struct inode *inode, int type)
 		return posix_acl_dup(*p_acl);
 
 	size = reiserfs_xattr_get(inode, name, NULL, 0);
-	if ((int)size < 0) {
+	if (size < 0) {
 		if (size == -ENODATA || size == -ENOSYS) {
 			*p_acl = ERR_PTR(-ENODATA);
 			return NULL;
diff --git a/fs/relayfs/Makefile b/fs/relayfs/Makefile
deleted file mode 100644
index e76e182cdb38..000000000000
--- a/fs/relayfs/Makefile
+++ /dev/null
@@ -1,4 +0,0 @@
-obj-$(CONFIG_RELAYFS_FS) += relayfs.o
-
-relayfs-y := relay.o inode.o buffers.o
-
diff --git a/fs/relayfs/buffers.c b/fs/relayfs/buffers.c
deleted file mode 100644
index 10187812771e..000000000000
--- a/fs/relayfs/buffers.c
+++ /dev/null
@@ -1,190 +0,0 @@
-/*
- * RelayFS buffer management code.
- *
- * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
- * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
- *
- * This file is released under the GPL.
- */
-
-#include <linux/module.h>
-#include <linux/vmalloc.h>
-#include <linux/mm.h>
-#include <linux/relayfs_fs.h>
-#include "relay.h"
-#include "buffers.h"
-
-/*
- * close() vm_op implementation for relayfs file mapping.
- */
-static void relay_file_mmap_close(struct vm_area_struct *vma)
-{
-	struct rchan_buf *buf = vma->vm_private_data;
-	buf->chan->cb->buf_unmapped(buf, vma->vm_file);
-}
-
-/*
- * nopage() vm_op implementation for relayfs file mapping.
- */
-static struct page *relay_buf_nopage(struct vm_area_struct *vma,
-				     unsigned long address,
-				     int *type)
-{
-	struct page *page;
-	struct rchan_buf *buf = vma->vm_private_data;
-	unsigned long offset = address - vma->vm_start;
-
-	if (address > vma->vm_end)
-		return NOPAGE_SIGBUS; /* Disallow mremap */
-	if (!buf)
-		return NOPAGE_OOM;
-
-	page = vmalloc_to_page(buf->start + offset);
-	if (!page)
-		return NOPAGE_OOM;
-	get_page(page);
-
-	if (type)
-		*type = VM_FAULT_MINOR;
-
-	return page;
-}
-
-/*
- * vm_ops for relay file mappings.
- */
-static struct vm_operations_struct relay_file_mmap_ops = {
-	.nopage = relay_buf_nopage,
-	.close = relay_file_mmap_close,
-};
-
-/**
- *	relay_mmap_buf: - mmap channel buffer to process address space
- *	@buf: relay channel buffer
- *	@vma: vm_area_struct describing memory to be mapped
- *
- *	Returns 0 if ok, negative on error
- *
- *	Caller should already have grabbed mmap_sem.
- */
-int relay_mmap_buf(struct rchan_buf *buf, struct vm_area_struct *vma)
-{
-	unsigned long length = vma->vm_end - vma->vm_start;
-	struct file *filp = vma->vm_file;
-
-	if (!buf)
-		return -EBADF;
-
-	if (length != (unsigned long)buf->chan->alloc_size)
-		return -EINVAL;
-
-	vma->vm_ops = &relay_file_mmap_ops;
-	vma->vm_private_data = buf;
-	buf->chan->cb->buf_mapped(buf, filp);
-
-	return 0;
-}
-
-/**
- *	relay_alloc_buf - allocate a channel buffer
- *	@buf: the buffer struct
- *	@size: total size of the buffer
- *
- *	Returns a pointer to the resulting buffer, NULL if unsuccessful
- */
-static void *relay_alloc_buf(struct rchan_buf *buf, unsigned long size)
-{
-	void *mem;
-	unsigned int i, j, n_pages;
-
-	size = PAGE_ALIGN(size);
-	n_pages = size >> PAGE_SHIFT;
-
-	buf->page_array = kcalloc(n_pages, sizeof(struct page *), GFP_KERNEL);
-	if (!buf->page_array)
-		return NULL;
-
-	for (i = 0; i < n_pages; i++) {
-		buf->page_array[i] = alloc_page(GFP_KERNEL);
-		if (unlikely(!buf->page_array[i]))
-			goto depopulate;
-	}
-	mem = vmap(buf->page_array, n_pages, VM_MAP, PAGE_KERNEL);
-	if (!mem)
-		goto depopulate;
-
-	memset(mem, 0, size);
-	buf->page_count = n_pages;
-	return mem;
-
-depopulate:
-	for (j = 0; j < i; j++)
-		__free_page(buf->page_array[j]);
-	kfree(buf->page_array);
-	return NULL;
-}
-
-/**
- *	relay_create_buf - allocate and initialize a channel buffer
- *	@alloc_size: size of the buffer to allocate
- *	@n_subbufs: number of sub-buffers in the channel
- *
- *	Returns channel buffer if successful, NULL otherwise
- */
-struct rchan_buf *relay_create_buf(struct rchan *chan)
-{
-	struct rchan_buf *buf = kcalloc(1, sizeof(struct rchan_buf), GFP_KERNEL);
-	if (!buf)
-		return NULL;
-
-	buf->padding = kmalloc(chan->n_subbufs * sizeof(size_t *), GFP_KERNEL);
-	if (!buf->padding)
-		goto free_buf;
-
-	buf->start = relay_alloc_buf(buf, chan->alloc_size);
-	if (!buf->start)
-		goto free_buf;
-
-	buf->chan = chan;
-	kref_get(&buf->chan->kref);
-	return buf;
-
-free_buf:
-	kfree(buf->padding);
-	kfree(buf);
-	return NULL;
-}
-
-/**
- *	relay_destroy_buf - destroy an rchan_buf struct and associated buffer
- *	@buf: the buffer struct
- */
-void relay_destroy_buf(struct rchan_buf *buf)
-{
-	struct rchan *chan = buf->chan;
-	unsigned int i;
-
-	if (likely(buf->start)) {
-		vunmap(buf->start);
-		for (i = 0; i < buf->page_count; i++)
-			__free_page(buf->page_array[i]);
-		kfree(buf->page_array);
-	}
-	kfree(buf->padding);
-	kfree(buf);
-	kref_put(&chan->kref, relay_destroy_channel);
-}
-
-/**
- *	relay_remove_buf - remove a channel buffer
- *
- *	Removes the file from the relayfs fileystem, which also frees the
- *	rchan_buf_struct and the channel buffer.  Should only be called from
- *	kref_put().
- */
-void relay_remove_buf(struct kref *kref)
-{
-	struct rchan_buf *buf = container_of(kref, struct rchan_buf, kref);
-	buf->chan->cb->remove_buf_file(buf->dentry);
-	relay_destroy_buf(buf);
-}
diff --git a/fs/relayfs/buffers.h b/fs/relayfs/buffers.h
deleted file mode 100644
index 37a12493f641..000000000000
--- a/fs/relayfs/buffers.h
+++ /dev/null
@@ -1,12 +0,0 @@
-#ifndef _BUFFERS_H
-#define _BUFFERS_H
-
-/* This inspired by rtai/shmem */
-#define FIX_SIZE(x) (((x) - 1) & PAGE_MASK) + PAGE_SIZE
-
-extern int relay_mmap_buf(struct rchan_buf *buf, struct vm_area_struct *vma);
-extern struct rchan_buf *relay_create_buf(struct rchan *chan);
-extern void relay_destroy_buf(struct rchan_buf *buf);
-extern void relay_remove_buf(struct kref *kref);
-
-#endif/* _BUFFERS_H */
diff --git a/fs/relayfs/inode.c b/fs/relayfs/inode.c
deleted file mode 100644
index 383523011aad..000000000000
--- a/fs/relayfs/inode.c
+++ /dev/null
@@ -1,581 +0,0 @@
-/*
- * VFS-related code for RelayFS, a high-speed data relay filesystem.
- *
- * Copyright (C) 2003-2005 - Tom Zanussi <zanussi@us.ibm.com>, IBM Corp
- * Copyright (C) 2003-2005 - Karim Yaghmour <karim@opersys.com>
- *
- * Based on ramfs, Copyright (C) 2002 - Linus Torvalds
- *
- * This file is released under the GPL.
- */
-
-#include <linux/module.h>
-#include <linux/fs.h>
-#include <linux/mount.h>
-#include <linux/pagemap.h>
-#include <linux/init.h>
-#include <linux/string.h>
-#include <linux/backing-dev.h>
-#include <linux/namei.h>
-#include <linux/poll.h>
-#include <linux/relayfs_fs.h>
-#include "relay.h"
-#include "buffers.h"
-
-#define RELAYFS_MAGIC			0xF0B4A981
-
-static struct vfsmount *		relayfs_mount;
-static int				relayfs_mount_count;
-
-static struct backing_dev_info		relayfs_backing_dev_info = {
-	.ra_pages	= 0,	/* No readahead */
-	.capabilities	= BDI_CAP_NO_ACCT_DIRTY | BDI_CAP_NO_WRITEBACK,
-};
-
-static struct inode *relayfs_get_inode(struct super_block *sb,
-				       int mode,
- 				       struct file_operations *fops,
-				       void *data)
-{
-	struct inode *inode;
-
-	inode = new_inode(sb);
-	if (!inode)
-		return NULL;
-
-	inode->i_mode = mode;
-	inode->i_uid = 0;
-	inode->i_gid = 0;
-	inode->i_blksize = PAGE_CACHE_SIZE;
-	inode->i_blocks = 0;
-	inode->i_mapping->backing_dev_info = &relayfs_backing_dev_info;
-	inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-	switch (mode & S_IFMT) {
-	case S_IFREG:
-		inode->i_fop = fops;
-		if (data)
-			inode->u.generic_ip = data;
-		break;
-	case S_IFDIR:
-		inode->i_op = &simple_dir_inode_operations;
-		inode->i_fop = &simple_dir_operations;
-
-		/* directory inodes start off with i_nlink == 2 (for "." entry) */
-		inode->i_nlink++;
-		break;
-	default:
-		break;
-	}
-
-	return inode;
-}
-
-/**
- *	relayfs_create_entry - create a relayfs directory or file
- *	@name: the name of the file to create
- *	@parent: parent directory
- *	@mode: mode
- *	@fops: file operations to use for the file
- *	@data: user-associated data for this file
- *
- *	Returns the new dentry, NULL on failure
- *
- *	Creates a file or directory with the specifed permissions.
- */
-static struct dentry *relayfs_create_entry(const char *name,
-					   struct dentry *parent,
-					   int mode,
-					   struct file_operations *fops,
-					   void *data)
-{
-	struct dentry *d;
-	struct inode *inode;
-	int error = 0;
-
-	BUG_ON(!name || !(S_ISREG(mode) || S_ISDIR(mode)));
-
-	error = simple_pin_fs("relayfs", &relayfs_mount, &relayfs_mount_count);
-	if (error) {
-		printk(KERN_ERR "Couldn't mount relayfs: errcode %d\n", error);
-		return NULL;
-	}
-
-	if (!parent && relayfs_mount && relayfs_mount->mnt_sb)
-		parent = relayfs_mount->mnt_sb->s_root;
-
-	if (!parent) {
-		simple_release_fs(&relayfs_mount, &relayfs_mount_count);
-		return NULL;
-	}
-
-	parent = dget(parent);
-	mutex_lock(&parent->d_inode->i_mutex);
-	d = lookup_one_len(name, parent, strlen(name));
-	if (IS_ERR(d)) {
-		d = NULL;
-		goto release_mount;
-	}
-
-	if (d->d_inode) {
-		d = NULL;
-		goto release_mount;
-	}
-
-	inode = relayfs_get_inode(parent->d_inode->i_sb, mode, fops, data);
-	if (!inode) {
-		d = NULL;
-		goto release_mount;
-	}
-
-	d_instantiate(d, inode);
-	dget(d);	/* Extra count - pin the dentry in core */
-
-	if (S_ISDIR(mode))
-		parent->d_inode->i_nlink++;
-
-	goto exit;
-
-release_mount:
-	simple_release_fs(&relayfs_mount, &relayfs_mount_count);
-
-exit:
-	mutex_unlock(&parent->d_inode->i_mutex);
-	dput(parent);
-	return d;
-}
-
-/**
- *	relayfs_create_file - create a file in the relay filesystem
- *	@name: the name of the file to create
- *	@parent: parent directory
- *	@mode: mode, if not specied the default perms are used
- *	@fops: file operations to use for the file
- *	@data: user-associated data for this file
- *
- *	Returns file dentry if successful, NULL otherwise.
- *
- *	The file will be created user r on behalf of current user.
- */
-struct dentry *relayfs_create_file(const char *name,
-				   struct dentry *parent,
-				   int mode,
-				   struct file_operations *fops,
-				   void *data)
-{
-	BUG_ON(!fops);
-
-	if (!mode)
-		mode = S_IRUSR;
-	mode = (mode & S_IALLUGO) | S_IFREG;
-
-	return relayfs_create_entry(name, parent, mode, fops, data);
-}
-
-/**
- *	relayfs_create_dir - create a directory in the relay filesystem
- *	@name: the name of the directory to create
- *	@parent: parent directory, NULL if parent should be fs root
- *
- *	Returns directory dentry if successful, NULL otherwise.
- *
- *	The directory will be created world rwx on behalf of current user.
- */
-struct dentry *relayfs_create_dir(const char *name, struct dentry *parent)
-{
-	int mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
-	return relayfs_create_entry(name, parent, mode, NULL, NULL);
-}
-
-/**
- *	relayfs_remove - remove a file or directory in the relay filesystem
- *	@dentry: file or directory dentry
- *
- *	Returns 0 if successful, negative otherwise.
- */
-int relayfs_remove(struct dentry *dentry)
-{
-	struct dentry *parent;
-	int error = 0;
-
-	if (!dentry)
-		return -EINVAL;
-	parent = dentry->d_parent;
-	if (!parent)
-		return -EINVAL;
-
-	parent = dget(parent);
-	mutex_lock(&parent->d_inode->i_mutex);
-	if (dentry->d_inode) {
-		if (S_ISDIR(dentry->d_inode->i_mode))
-			error = simple_rmdir(parent->d_inode, dentry);
-		else
-			error = simple_unlink(parent->d_inode, dentry);
-		if (!error)
-			d_delete(dentry);
-	}
-	if (!error)
-		dput(dentry);
-	mutex_unlock(&parent->d_inode->i_mutex);
-	dput(parent);
-
-	if (!error)
-		simple_release_fs(&relayfs_mount, &relayfs_mount_count);
-
-	return error;
-}
-
-/**
- *	relayfs_remove_file - remove a file from relay filesystem
- *	@dentry: directory dentry
- *
- *	Returns 0 if successful, negative otherwise.
- */
-int relayfs_remove_file(struct dentry *dentry)
-{
-	return relayfs_remove(dentry);
-}
-
-/**
- *	relayfs_remove_dir - remove a directory in the relay filesystem
- *	@dentry: directory dentry
- *
- *	Returns 0 if successful, negative otherwise.
- */
-int relayfs_remove_dir(struct dentry *dentry)
-{
-	return relayfs_remove(dentry);
-}
-
-/**
- *	relay_file_open - open file op for relay files
- *	@inode: the inode
- *	@filp: the file
- *
- *	Increments the channel buffer refcount.
- */
-static int relay_file_open(struct inode *inode, struct file *filp)
-{
-	struct rchan_buf *buf = inode->u.generic_ip;
-	kref_get(&buf->kref);
-	filp->private_data = buf;
-
-	return 0;
-}
-
-/**
- *	relay_file_mmap - mmap file op for relay files
- *	@filp: the file
- *	@vma: the vma describing what to map
- *
- *	Calls upon relay_mmap_buf to map the file into user space.
- */
-static int relay_file_mmap(struct file *filp, struct vm_area_struct *vma)
-{
-	struct rchan_buf *buf = filp->private_data;
-	return relay_mmap_buf(buf, vma);
-}
-
-/**
- *	relay_file_poll - poll file op for relay files
- *	@filp: the file
- *	@wait: poll table
- *
- *	Poll implemention.
- */
-static unsigned int relay_file_poll(struct file *filp, poll_table *wait)
-{
-	unsigned int mask = 0;
-	struct rchan_buf *buf = filp->private_data;
-
-	if (buf->finalized)
-		return POLLERR;
-
-	if (filp->f_mode & FMODE_READ) {
-		poll_wait(filp, &buf->read_wait, wait);
-		if (!relay_buf_empty(buf))
-			mask |= POLLIN | POLLRDNORM;
-	}
-
-	return mask;
-}
-
-/**
- *	relay_file_release - release file op for relay files
- *	@inode: the inode
- *	@filp: the file
- *
- *	Decrements the channel refcount, as the filesystem is
- *	no longer using it.
- */
-static int relay_file_release(struct inode *inode, struct file *filp)
-{
-	struct rchan_buf *buf = filp->private_data;
-	kref_put(&buf->kref, relay_remove_buf);
-
-	return 0;
-}
-
-/**
- *	relay_file_read_consume - update the consumed count for the buffer
- */
-static void relay_file_read_consume(struct rchan_buf *buf,
-				    size_t read_pos,
-				    size_t bytes_consumed)
-{
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-	size_t read_subbuf;
-
-	if (buf->bytes_consumed + bytes_consumed > subbuf_size) {
-		relay_subbufs_consumed(buf->chan, buf->cpu, 1);
-		buf->bytes_consumed = 0;
-	}
-
-	buf->bytes_consumed += bytes_consumed;
-	read_subbuf = read_pos / buf->chan->subbuf_size;
-	if (buf->bytes_consumed + buf->padding[read_subbuf] == subbuf_size) {
-		if ((read_subbuf == buf->subbufs_produced % n_subbufs) &&
-		    (buf->offset == subbuf_size))
-			return;
-		relay_subbufs_consumed(buf->chan, buf->cpu, 1);
-		buf->bytes_consumed = 0;
-	}
-}
-
-/**
- *	relay_file_read_avail - boolean, are there unconsumed bytes available?
- */
-static int relay_file_read_avail(struct rchan_buf *buf, size_t read_pos)
-{
-	size_t bytes_produced, bytes_consumed, write_offset;
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-	size_t produced = buf->subbufs_produced % n_subbufs;
-	size_t consumed = buf->subbufs_consumed % n_subbufs;
-
-	write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-
-	if (consumed > produced) {
-		if ((produced > n_subbufs) &&
-		    (produced + n_subbufs - consumed <= n_subbufs))
-			produced += n_subbufs;
-	} else if (consumed == produced) {
-		if (buf->offset > subbuf_size) {
-			produced += n_subbufs;
-			if (buf->subbufs_produced == buf->subbufs_consumed)
-				consumed += n_subbufs;
-		}
-	}
-
-	if (buf->offset > subbuf_size)
-		bytes_produced = (produced - 1) * subbuf_size + write_offset;
-	else
-		bytes_produced = produced * subbuf_size + write_offset;
-	bytes_consumed = consumed * subbuf_size + buf->bytes_consumed;
-
-	if (bytes_produced == bytes_consumed)
-		return 0;
-
-	relay_file_read_consume(buf, read_pos, 0);
-
-	return 1;
-}
-
-/**
- *	relay_file_read_subbuf_avail - return bytes available in sub-buffer
- */
-static size_t relay_file_read_subbuf_avail(size_t read_pos,
-					   struct rchan_buf *buf)
-{
-	size_t padding, avail = 0;
-	size_t read_subbuf, read_offset, write_subbuf, write_offset;
-	size_t subbuf_size = buf->chan->subbuf_size;
-
-	write_subbuf = (buf->data - buf->start) / subbuf_size;
-	write_offset = buf->offset > subbuf_size ? subbuf_size : buf->offset;
-	read_subbuf = read_pos / subbuf_size;
-	read_offset = read_pos % subbuf_size;
-	padding = buf->padding[read_subbuf];
-
-	if (read_subbuf == write_subbuf) {
-		if (read_offset + padding < write_offset)
-			avail = write_offset - (read_offset + padding);
-	} else
-		avail = (subbuf_size - padding) - read_offset;
-
-	return avail;
-}
-
-/**
- *	relay_file_read_start_pos - find the first available byte to read
- *
- *	If the read_pos is in the middle of padding, return the
- *	position of the first actually available byte, otherwise
- *	return the original value.
- */
-static size_t relay_file_read_start_pos(size_t read_pos,
-					struct rchan_buf *buf)
-{
-	size_t read_subbuf, padding, padding_start, padding_end;
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-
-	read_subbuf = read_pos / subbuf_size;
-	padding = buf->padding[read_subbuf];
-	padding_start = (read_subbuf + 1) * subbuf_size - padding;
-	padding_end = (read_subbuf + 1) * subbuf_size;
-	if (read_pos >= padding_start && read_pos < padding_end) {
-		read_subbuf = (read_subbuf + 1) % n_subbufs;
-		read_pos = read_subbuf * subbuf_size;
-	}
-
-	return read_pos;
-}
-
-/**
- *	relay_file_read_end_pos - return the new read position
- */
-static size_t relay_file_read_end_pos(struct rchan_buf *buf,
-				      size_t read_pos,
-				      size_t count)
-{
-	size_t read_subbuf, padding, end_pos;
-	size_t subbuf_size = buf->chan->subbuf_size;
-	size_t n_subbufs = buf->chan->n_subbufs;
-
-	read_subbuf = read_pos / subbuf_size;
-	padding = buf->padding[read_subbuf];
-	if (read_pos % subbuf_size + count + padding == subbuf_size)
-		end_pos = (read_subbuf + 1) * subbuf_size;
-	else
-		end_pos = read_pos + count;
-	if (end_pos >= subbuf_size * n_subbufs)
-		end_pos = 0;
-
-	return end_pos;
-}
-
-/**
- *	relay_file_read - read file op for relay files
- *	@filp: the file
- *	@buffer: the userspace buffer
- *	@count: number of bytes to read
- *	@ppos: position to read from
- *
- *	Reads count bytes or the number of bytes available in the
- *	current sub-buffer being read, whichever is smaller.
- */
-static ssize_t relay_file_read(struct file *filp,
-			       char __user *buffer,
-			       size_t count,
-			       loff_t *ppos)
-{
-	struct rchan_buf *buf = filp->private_data;
-	struct inode *inode = filp->f_dentry->d_inode;
-	size_t read_start, avail;
-	ssize_t ret = 0;
-	void *from;
-
-	mutex_lock(&inode->i_mutex);
-	if(!relay_file_read_avail(buf, *ppos))
-		goto out;
-
-	read_start = relay_file_read_start_pos(*ppos, buf);
-	avail = relay_file_read_subbuf_avail(read_start, buf);
-	if (!avail)
-		goto out;
-
-	from = buf->start + read_start;
-	ret = count = min(count, avail);
-	if (copy_to_user(buffer, from, count)) {
-		ret = -EFAULT;
-		goto out;
-	}
-	relay_file_read_consume(buf, read_start, count);
-	*ppos = relay_file_read_end_pos(buf, read_start, count);
-out:
-	mutex_unlock(&inode->i_mutex);
-	return ret;
-}
-
-struct file_operations relay_file_operations = {
-	.open		= relay_file_open,
-	.poll		= relay_file_poll,
-	.mmap		= relay_file_mmap,
-	.read		= relay_file_read,
-	.llseek		= no_llseek,
-	.release	= relay_file_release,
-};
-
-static struct super_operations relayfs_ops = {
-	.statfs		= simple_statfs,
-	.drop_inode	= generic_delete_inode,
-};
-
-static int relayfs_fill_super(struct super_block * sb, void * data, int silent)
-{
-	struct inode *inode;
-	struct dentry *root;
-	int mode = S_IFDIR | S_IRWXU | S_IRUGO | S_IXUGO;
-
-	sb->s_blocksize = PAGE_CACHE_SIZE;
-	sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
-	sb->s_magic = RELAYFS_MAGIC;
-	sb->s_op = &relayfs_ops;
-	inode = relayfs_get_inode(sb, mode, NULL, NULL);
-
-	if (!inode)
-		return -ENOMEM;
-
-	root = d_alloc_root(inode);
-	if (!root) {
-		iput(inode);
-		return -ENOMEM;
-	}
-	sb->s_root = root;
-
-	return 0;
-}
-
-static struct super_block * relayfs_get_sb(struct file_system_type *fs_type,
-					   int flags, const char *dev_name,
-					   void *data)
-{
-	return get_sb_single(fs_type, flags, data, relayfs_fill_super);
-}
-
-static struct file_system_type relayfs_fs_type = {
-	.owner		= THIS_MODULE,
-	.name		= "relayfs",
-	.get_sb		= relayfs_get_sb,
-	.kill_sb	= kill_litter_super,
-};
-
-static int __init init_relayfs_fs(void)
-{
-	return register_filesystem(&relayfs_fs_type);
-}
-
-static void __exit exit_relayfs_fs(void)
-{
-
-
-
-
-
-	unregister_filesystem(&relayfs_fs_type);
-}
-
-module_init(init_relayfs_fs)
-module_exit(exit_relayfs_fs)
-
-EXPORT_SYMBOL_GPL(relay_file_operations);
-EXPORT_SYMBOL_GPL(relayfs_create_dir);
-EXPORT_SYMBOL_GPL(relayfs_remove_dir);
-EXPORT_SYMBOL_GPL(relayfs_create_file);
-EXPORT_SYMBOL_GPL(relayfs_remove_file);
-
-MODULE_AUTHOR("Tom Zanussi <zanussi@us.ibm.com> and Karim Yaghmour <karim@opersys.com>");
-MODULE_DESCRIPTION("Relay Filesystem");
-MODULE_LICENSE("GPL");
-
diff --git a/fs/relayfs/relay.c b/fs/relayfs/relay.c
deleted file mode 100644
index abf3ceaace49..000000000000
--- a/fs/relayfs/relay.c
+++ /dev/null
@@ -1,482 +0,0 @@
-/*
- * Public API and common code for RelayFS.
- *
- * See Documentation/filesystems/relayfs.txt for an overview of relayfs.
- *
- * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
- * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
- *
- * This file is released under the GPL.
- */
-
-#include <linux/errno.h>
-#include <linux/stddef.h>
-#include <linux/slab.h>
-#include <linux/module.h>
-#include <linux/string.h>
-#include <linux/relayfs_fs.h>
-#include "relay.h"
-#include "buffers.h"
-
-/**
- *	relay_buf_empty - boolean, is the channel buffer empty?
- *	@buf: channel buffer
- *
- *	Returns 1 if the buffer is empty, 0 otherwise.
- */
-int relay_buf_empty(struct rchan_buf *buf)
-{
-	return (buf->subbufs_produced - buf->subbufs_consumed) ? 0 : 1;
-}
-
-/**
- *	relay_buf_full - boolean, is the channel buffer full?
- *	@buf: channel buffer
- *
- *	Returns 1 if the buffer is full, 0 otherwise.
- */
-int relay_buf_full(struct rchan_buf *buf)
-{
-	size_t ready = buf->subbufs_produced - buf->subbufs_consumed;
-	return (ready >= buf->chan->n_subbufs) ? 1 : 0;
-}
-
-/*
- * High-level relayfs kernel API and associated functions.
- */
-
-/*
- * rchan_callback implementations defining default channel behavior.  Used
- * in place of corresponding NULL values in client callback struct.
- */
-
-/*
- * subbuf_start() default callback.  Does nothing.
- */
-static int subbuf_start_default_callback (struct rchan_buf *buf,
-					  void *subbuf,
-					  void *prev_subbuf,
-					  size_t prev_padding)
-{
-	if (relay_buf_full(buf))
-		return 0;
-
-	return 1;
-}
-
-/*
- * buf_mapped() default callback.  Does nothing.
- */
-static void buf_mapped_default_callback(struct rchan_buf *buf,
-					struct file *filp)
-{
-}
-
-/*
- * buf_unmapped() default callback.  Does nothing.
- */
-static void buf_unmapped_default_callback(struct rchan_buf *buf,
-					  struct file *filp)
-{
-}
-
-/*
- * create_buf_file_create() default callback.  Creates file to represent buf.
- */
-static struct dentry *create_buf_file_default_callback(const char *filename,
-						       struct dentry *parent,
-						       int mode,
-						       struct rchan_buf *buf,
-						       int *is_global)
-{
-	return relayfs_create_file(filename, parent, mode,
-				   &relay_file_operations, buf);
-}
-
-/*
- * remove_buf_file() default callback.  Removes file representing relay buffer.
- */
-static int remove_buf_file_default_callback(struct dentry *dentry)
-{
-	return relayfs_remove(dentry);
-}
-
-/* relay channel default callbacks */
-static struct rchan_callbacks default_channel_callbacks = {
-	.subbuf_start = subbuf_start_default_callback,
-	.buf_mapped = buf_mapped_default_callback,
-	.buf_unmapped = buf_unmapped_default_callback,
-	.create_buf_file = create_buf_file_default_callback,
-	.remove_buf_file = remove_buf_file_default_callback,
-};
-
-/**
- *	wakeup_readers - wake up readers waiting on a channel
- *	@private: the channel buffer
- *
- *	This is the work function used to defer reader waking.  The
- *	reason waking is deferred is that calling directly from write
- *	causes problems if you're writing from say the scheduler.
- */
-static void wakeup_readers(void *private)
-{
-	struct rchan_buf *buf = private;
-	wake_up_interruptible(&buf->read_wait);
-}
-
-/**
- *	__relay_reset - reset a channel buffer
- *	@buf: the channel buffer
- *	@init: 1 if this is a first-time initialization
- *
- *	See relay_reset for description of effect.
- */
-static inline void __relay_reset(struct rchan_buf *buf, unsigned int init)
-{
-	size_t i;
-
-	if (init) {
-		init_waitqueue_head(&buf->read_wait);
-		kref_init(&buf->kref);
-		INIT_WORK(&buf->wake_readers, NULL, NULL);
-	} else {
-		cancel_delayed_work(&buf->wake_readers);
-		flush_scheduled_work();
-	}
-
-	buf->subbufs_produced = 0;
-	buf->subbufs_consumed = 0;
-	buf->bytes_consumed = 0;
-	buf->finalized = 0;
-	buf->data = buf->start;
-	buf->offset = 0;
-
-	for (i = 0; i < buf->chan->n_subbufs; i++)
-		buf->padding[i] = 0;
-
-	buf->chan->cb->subbuf_start(buf, buf->data, NULL, 0);
-}
-
-/**
- *	relay_reset - reset the channel
- *	@chan: the channel
- *
- *	This has the effect of erasing all data from all channel buffers
- *	and restarting the channel in its initial state.  The buffers
- *	are not freed, so any mappings are still in effect.
- *
- *	NOTE: Care should be taken that the channel isn't actually
- *	being used by anything when this call is made.
- */
-void relay_reset(struct rchan *chan)
-{
-	unsigned int i;
-	struct rchan_buf *prev = NULL;
-
-	if (!chan)
-		return;
-
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		__relay_reset(chan->buf[i], 0);
-		prev = chan->buf[i];
-	}
-}
-
-/**
- *	relay_open_buf - create a new channel buffer in relayfs
- *
- *	Internal - used by relay_open().
- */
-static struct rchan_buf *relay_open_buf(struct rchan *chan,
-					const char *filename,
-					struct dentry *parent,
-					int *is_global)
-{
-	struct rchan_buf *buf;
-	struct dentry *dentry;
-
-	if (*is_global)
-		return chan->buf[0];
-
- 	buf = relay_create_buf(chan);
- 	if (!buf)
- 		return NULL;
-
-	/* Create file in fs */
- 	dentry = chan->cb->create_buf_file(filename, parent, S_IRUSR,
- 					   buf, is_global);
- 	if (!dentry) {
- 		relay_destroy_buf(buf);
-		return NULL;
- 	}
-
-	buf->dentry = dentry;
-	__relay_reset(buf, 1);
-
-	return buf;
-}
-
-/**
- *	relay_close_buf - close a channel buffer
- *	@buf: channel buffer
- *
- *	Marks the buffer finalized and restores the default callbacks.
- *	The channel buffer and channel buffer data structure are then freed
- *	automatically when the last reference is given up.
- */
-static inline void relay_close_buf(struct rchan_buf *buf)
-{
-	buf->finalized = 1;
-	buf->chan->cb = &default_channel_callbacks;
-	cancel_delayed_work(&buf->wake_readers);
-	flush_scheduled_work();
-	kref_put(&buf->kref, relay_remove_buf);
-}
-
-static inline void setup_callbacks(struct rchan *chan,
-				   struct rchan_callbacks *cb)
-{
-	if (!cb) {
-		chan->cb = &default_channel_callbacks;
-		return;
-	}
-
-	if (!cb->subbuf_start)
-		cb->subbuf_start = subbuf_start_default_callback;
-	if (!cb->buf_mapped)
-		cb->buf_mapped = buf_mapped_default_callback;
-	if (!cb->buf_unmapped)
-		cb->buf_unmapped = buf_unmapped_default_callback;
-	if (!cb->create_buf_file)
-		cb->create_buf_file = create_buf_file_default_callback;
-	if (!cb->remove_buf_file)
-		cb->remove_buf_file = remove_buf_file_default_callback;
-	chan->cb = cb;
-}
-
-/**
- *	relay_open - create a new relayfs channel
- *	@base_filename: base name of files to create
- *	@parent: dentry of parent directory, NULL for root directory
- *	@subbuf_size: size of sub-buffers
- *	@n_subbufs: number of sub-buffers
- *	@cb: client callback functions
- *
- *	Returns channel pointer if successful, NULL otherwise.
- *
- *	Creates a channel buffer for each cpu using the sizes and
- *	attributes specified.  The created channel buffer files
- *	will be named base_filename0...base_filenameN-1.  File
- *	permissions will be S_IRUSR.
- */
-struct rchan *relay_open(const char *base_filename,
-			 struct dentry *parent,
-			 size_t subbuf_size,
-			 size_t n_subbufs,
-			 struct rchan_callbacks *cb)
-{
-	unsigned int i;
-	struct rchan *chan;
-	char *tmpname;
-	int is_global = 0;
-
-	if (!base_filename)
-		return NULL;
-
-	if (!(subbuf_size && n_subbufs))
-		return NULL;
-
-	chan = kcalloc(1, sizeof(struct rchan), GFP_KERNEL);
-	if (!chan)
-		return NULL;
-
-	chan->version = RELAYFS_CHANNEL_VERSION;
-	chan->n_subbufs = n_subbufs;
-	chan->subbuf_size = subbuf_size;
-	chan->alloc_size = FIX_SIZE(subbuf_size * n_subbufs);
-	setup_callbacks(chan, cb);
-	kref_init(&chan->kref);
-
-	tmpname = kmalloc(NAME_MAX + 1, GFP_KERNEL);
-	if (!tmpname)
-		goto free_chan;
-
-	for_each_online_cpu(i) {
-		sprintf(tmpname, "%s%d", base_filename, i);
-		chan->buf[i] = relay_open_buf(chan, tmpname, parent,
-					      &is_global);
-		chan->buf[i]->cpu = i;
-		if (!chan->buf[i])
-			goto free_bufs;
-	}
-
-	kfree(tmpname);
-	return chan;
-
-free_bufs:
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i])
-			break;
-		relay_close_buf(chan->buf[i]);
-		if (is_global)
-			break;
-	}
-	kfree(tmpname);
-
-free_chan:
-	kref_put(&chan->kref, relay_destroy_channel);
-	return NULL;
-}
-
-/**
- *	relay_switch_subbuf - switch to a new sub-buffer
- *	@buf: channel buffer
- *	@length: size of current event
- *
- *	Returns either the length passed in or 0 if full.
-
- *	Performs sub-buffer-switch tasks such as invoking callbacks,
- *	updating padding counts, waking up readers, etc.
- */
-size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length)
-{
-	void *old, *new;
-	size_t old_subbuf, new_subbuf;
-
-	if (unlikely(length > buf->chan->subbuf_size))
-		goto toobig;
-
-	if (buf->offset != buf->chan->subbuf_size + 1) {
-		buf->prev_padding = buf->chan->subbuf_size - buf->offset;
-		old_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
-		buf->padding[old_subbuf] = buf->prev_padding;
-		buf->subbufs_produced++;
-		if (waitqueue_active(&buf->read_wait)) {
-			PREPARE_WORK(&buf->wake_readers, wakeup_readers, buf);
-			schedule_delayed_work(&buf->wake_readers, 1);
-		}
-	}
-
-	old = buf->data;
-	new_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
-	new = buf->start + new_subbuf * buf->chan->subbuf_size;
-	buf->offset = 0;
-	if (!buf->chan->cb->subbuf_start(buf, new, old, buf->prev_padding)) {
-		buf->offset = buf->chan->subbuf_size + 1;
-		return 0;
-	}
-	buf->data = new;
-	buf->padding[new_subbuf] = 0;
-
-	if (unlikely(length + buf->offset > buf->chan->subbuf_size))
-		goto toobig;
-
-	return length;
-
-toobig:
-	buf->chan->last_toobig = length;
-	return 0;
-}
-
-/**
- *	relay_subbufs_consumed - update the buffer's sub-buffers-consumed count
- *	@chan: the channel
- *	@cpu: the cpu associated with the channel buffer to update
- *	@subbufs_consumed: number of sub-buffers to add to current buf's count
- *
- *	Adds to the channel buffer's consumed sub-buffer count.
- *	subbufs_consumed should be the number of sub-buffers newly consumed,
- *	not the total consumed.
- *
- *	NOTE: kernel clients don't need to call this function if the channel
- *	mode is 'overwrite'.
- */
-void relay_subbufs_consumed(struct rchan *chan,
-			    unsigned int cpu,
-			    size_t subbufs_consumed)
-{
-	struct rchan_buf *buf;
-
-	if (!chan)
-		return;
-
-	if (cpu >= NR_CPUS || !chan->buf[cpu])
-		return;
-
-	buf = chan->buf[cpu];
-	buf->subbufs_consumed += subbufs_consumed;
-	if (buf->subbufs_consumed > buf->subbufs_produced)
-		buf->subbufs_consumed = buf->subbufs_produced;
-}
-
-/**
- *	relay_destroy_channel - free the channel struct
- *
- *	Should only be called from kref_put().
- */
-void relay_destroy_channel(struct kref *kref)
-{
-	struct rchan *chan = container_of(kref, struct rchan, kref);
-	kfree(chan);
-}
-
-/**
- *	relay_close - close the channel
- *	@chan: the channel
- *
- *	Closes all channel buffers and frees the channel.
- */
-void relay_close(struct rchan *chan)
-{
-	unsigned int i;
-	struct rchan_buf *prev = NULL;
-
-	if (!chan)
-		return;
-
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		relay_close_buf(chan->buf[i]);
-		prev = chan->buf[i];
-	}
-
-	if (chan->last_toobig)
-		printk(KERN_WARNING "relayfs: one or more items not logged "
-		       "[item size (%Zd) > sub-buffer size (%Zd)]\n",
-		       chan->last_toobig, chan->subbuf_size);
-
-	kref_put(&chan->kref, relay_destroy_channel);
-}
-
-/**
- *	relay_flush - close the channel
- *	@chan: the channel
- *
- *	Flushes all channel buffers i.e. forces buffer switch.
- */
-void relay_flush(struct rchan *chan)
-{
-	unsigned int i;
-	struct rchan_buf *prev = NULL;
-
-	if (!chan)
-		return;
-
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!chan->buf[i] || chan->buf[i] == prev)
-			break;
-		relay_switch_subbuf(chan->buf[i], 0);
-		prev = chan->buf[i];
-	}
-}
-
-EXPORT_SYMBOL_GPL(relay_open);
-EXPORT_SYMBOL_GPL(relay_close);
-EXPORT_SYMBOL_GPL(relay_flush);
-EXPORT_SYMBOL_GPL(relay_reset);
-EXPORT_SYMBOL_GPL(relay_subbufs_consumed);
-EXPORT_SYMBOL_GPL(relay_switch_subbuf);
-EXPORT_SYMBOL_GPL(relay_buf_full);
diff --git a/fs/relayfs/relay.h b/fs/relayfs/relay.h
deleted file mode 100644
index 0993d3e5753b..000000000000
--- a/fs/relayfs/relay.h
+++ /dev/null
@@ -1,8 +0,0 @@
-#ifndef _RELAY_H
-#define _RELAY_H
-
-extern int relayfs_remove(struct dentry *dentry);
-extern int relay_buf_empty(struct rchan_buf *buf);
-extern void relay_destroy_channel(struct kref *kref);
-
-#endif /* _RELAY_H */
diff --git a/fs/romfs/inode.c b/fs/romfs/inode.c
index 0a13859fd57b..9b9eda7b335c 100644
--- a/fs/romfs/inode.c
+++ b/fs/romfs/inode.c
@@ -463,7 +463,7 @@ static struct address_space_operations romfs_aops = {
 	.readpage = romfs_readpage
 };
 
-static struct file_operations romfs_dir_operations = {
+static const struct file_operations romfs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= romfs_readdir,
 };
@@ -579,7 +579,8 @@ static int init_inodecache(void)
 {
 	romfs_inode_cachep = kmem_cache_create("romfs_inode_cache",
 					     sizeof(struct romfs_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (romfs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/select.c b/fs/select.c
index 1815a57d2255..b3a3a1326af6 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -29,12 +29,6 @@
 #define ROUND_UP(x,y) (((x)+(y)-1)/(y))
 #define DEFAULT_POLLMASK (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM)
 
-struct poll_table_entry {
-	struct file * filp;
-	wait_queue_t wait;
-	wait_queue_head_t * wait_address;
-};
-
 struct poll_table_page {
 	struct poll_table_page * next;
 	struct poll_table_entry * entry;
@@ -64,13 +58,23 @@ void poll_initwait(struct poll_wqueues *pwq)
 	init_poll_funcptr(&pwq->pt, __pollwait);
 	pwq->error = 0;
 	pwq->table = NULL;
+	pwq->inline_index = 0;
 }
 
 EXPORT_SYMBOL(poll_initwait);
 
+static void free_poll_entry(struct poll_table_entry *entry)
+{
+	remove_wait_queue(entry->wait_address,&entry->wait);
+	fput(entry->filp);
+}
+
 void poll_freewait(struct poll_wqueues *pwq)
 {
 	struct poll_table_page * p = pwq->table;
+	int i;
+	for (i = 0; i < pwq->inline_index; i++)
+		free_poll_entry(pwq->inline_entries + i);
 	while (p) {
 		struct poll_table_entry * entry;
 		struct poll_table_page *old;
@@ -78,8 +82,7 @@ void poll_freewait(struct poll_wqueues *pwq)
 		entry = p->entry;
 		do {
 			entry--;
-			remove_wait_queue(entry->wait_address,&entry->wait);
-			fput(entry->filp);
+			free_poll_entry(entry);
 		} while (entry > p->entries);
 		old = p;
 		p = p->next;
@@ -89,12 +92,14 @@ void poll_freewait(struct poll_wqueues *pwq)
 
 EXPORT_SYMBOL(poll_freewait);
 
-static void __pollwait(struct file *filp, wait_queue_head_t *wait_address,
-		       poll_table *_p)
+static struct poll_table_entry *poll_get_entry(poll_table *_p)
 {
 	struct poll_wqueues *p = container_of(_p, struct poll_wqueues, pt);
 	struct poll_table_page *table = p->table;
 
+	if (p->inline_index < N_INLINE_POLL_ENTRIES)
+		return p->inline_entries + p->inline_index++;
+
 	if (!table || POLL_TABLE_FULL(table)) {
 		struct poll_table_page *new_table;
 
@@ -102,7 +107,7 @@ static void __pollwait(struct file *filp, wait_queue_head_t *wait_address,
 		if (!new_table) {
 			p->error = -ENOMEM;
 			__set_current_state(TASK_RUNNING);
-			return;
+			return NULL;
 		}
 		new_table->entry = new_table->entries;
 		new_table->next = table;
@@ -110,16 +115,21 @@ static void __pollwait(struct file *filp, wait_queue_head_t *wait_address,
 		table = new_table;
 	}
 
-	/* Add a new entry */
-	{
-		struct poll_table_entry * entry = table->entry;
-		table->entry = entry+1;
-	 	get_file(filp);
-	 	entry->filp = filp;
-		entry->wait_address = wait_address;
-		init_waitqueue_entry(&entry->wait, current);
-		add_wait_queue(wait_address,&entry->wait);
-	}
+	return table->entry++;
+}
+
+/* Add a new entry */
+static void __pollwait(struct file *filp, wait_queue_head_t *wait_address,
+				poll_table *p)
+{
+	struct poll_table_entry *entry = poll_get_entry(p);
+	if (!entry)
+		return;
+	get_file(filp);
+	entry->filp = filp;
+	entry->wait_address = wait_address;
+	init_waitqueue_entry(&entry->wait, current);
+	add_wait_queue(wait_address,&entry->wait);
 }
 
 #define FDS_IN(fds, n)		(fds->in + n)
@@ -210,7 +220,7 @@ int do_select(int n, fd_set_bits *fds, s64 *timeout)
 		for (i = 0; i < n; ++rinp, ++routp, ++rexp) {
 			unsigned long in, out, ex, all_bits, bit = 1, mask, j;
 			unsigned long res_in = 0, res_out = 0, res_ex = 0;
-			struct file_operations *f_op = NULL;
+			const struct file_operations *f_op = NULL;
 			struct file *file = NULL;
 
 			in = *inp++; out = *outp++; ex = *exp++;
@@ -221,17 +231,18 @@ int do_select(int n, fd_set_bits *fds, s64 *timeout)
 			}
 
 			for (j = 0; j < __NFDBITS; ++j, ++i, bit <<= 1) {
+				int fput_needed;
 				if (i >= n)
 					break;
 				if (!(bit & all_bits))
 					continue;
-				file = fget(i);
+				file = fget_light(i, &fput_needed);
 				if (file) {
 					f_op = file->f_op;
 					mask = DEFAULT_POLLMASK;
 					if (f_op && f_op->poll)
 						mask = (*f_op->poll)(file, retval ? NULL : wait);
-					fput(file);
+					fput_light(file, fput_needed);
 					if ((mask & POLLIN_SET) && (in & bit)) {
 						res_in |= bit;
 						retval++;
@@ -284,16 +295,6 @@ int do_select(int n, fd_set_bits *fds, s64 *timeout)
 	return retval;
 }
 
-static void *select_bits_alloc(int size)
-{
-	return kmalloc(6 * size, GFP_KERNEL);
-}
-
-static void select_bits_free(void *bits, int size)
-{
-	kfree(bits);
-}
-
 /*
  * We can actually return ERESTARTSYS instead of EINTR, but I'd
  * like to be certain this leads to no problems. So I return
@@ -312,6 +313,8 @@ static int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
 	char *bits;
 	int ret, size, max_fdset;
 	struct fdtable *fdt;
+	/* Allocate small arguments on the stack to save memory and be faster */
+	char stack_fds[SELECT_STACK_ALLOC];
 
 	ret = -EINVAL;
 	if (n < 0)
@@ -332,7 +335,10 @@ static int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
 	 */
 	ret = -ENOMEM;
 	size = FDS_BYTES(n);
-	bits = select_bits_alloc(size);
+	if (6*size < SELECT_STACK_ALLOC)
+		bits = stack_fds;
+	else
+		bits = kmalloc(6 * size, GFP_KERNEL);
 	if (!bits)
 		goto out_nofds;
 	fds.in      = (unsigned long *)  bits;
@@ -367,7 +373,8 @@ static int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
 		ret = -EFAULT;
 
 out:
-	select_bits_free(bits, size);
+	if (bits != stack_fds)
+		kfree(bits);
 out_nofds:
 	return ret;
 }
@@ -551,14 +558,15 @@ static void do_pollfd(unsigned int num, struct pollfd * fdpage,
 		fdp = fdpage+i;
 		fd = fdp->fd;
 		if (fd >= 0) {
-			struct file * file = fget(fd);
+			int fput_needed;
+			struct file * file = fget_light(fd, &fput_needed);
 			mask = POLLNVAL;
 			if (file != NULL) {
 				mask = DEFAULT_POLLMASK;
 				if (file->f_op && file->f_op->poll)
 					mask = file->f_op->poll(file, *pwait);
 				mask &= fdp->events | POLLERR | POLLHUP;
-				fput(file);
+				fput_light(file, fput_needed);
 			}
 			if (mask) {
 				*pwait = NULL;
@@ -619,6 +627,9 @@ static int do_poll(unsigned int nfds,  struct poll_list *list,
 	return count;
 }
 
+#define N_STACK_PPS ((sizeof(stack_pps) - sizeof(struct poll_list))  / \
+			sizeof(struct pollfd))
+
 int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds, s64 *timeout)
 {
 	struct poll_wqueues table;
@@ -628,6 +639,9 @@ int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds, s64 *timeout)
  	struct poll_list *walk;
 	struct fdtable *fdt;
 	int max_fdset;
+	/* Allocate small arguments on the stack to save memory and be faster */
+	char stack_pps[POLL_STACK_ALLOC];
+	struct poll_list *stack_pp = NULL;
 
 	/* Do a sanity check on nfds ... */
 	rcu_read_lock();
@@ -645,14 +659,23 @@ int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds, s64 *timeout)
 	err = -ENOMEM;
 	while(i!=0) {
 		struct poll_list *pp;
-		pp = kmalloc(sizeof(struct poll_list)+
-				sizeof(struct pollfd)*
-				(i>POLLFD_PER_PAGE?POLLFD_PER_PAGE:i),
-					GFP_KERNEL);
-		if(pp==NULL)
-			goto out_fds;
+		int num, size;
+		if (stack_pp == NULL)
+			num = N_STACK_PPS;
+		else
+			num = POLLFD_PER_PAGE;
+		if (num > i)
+			num = i;
+		size = sizeof(struct poll_list) + sizeof(struct pollfd)*num;
+		if (!stack_pp)
+			stack_pp = pp = (struct poll_list *)stack_pps;
+		else {
+			pp = kmalloc(size, GFP_KERNEL);
+			if (!pp)
+				goto out_fds;
+		}
 		pp->next=NULL;
-		pp->len = (i>POLLFD_PER_PAGE?POLLFD_PER_PAGE:i);
+		pp->len = num;
 		if (head == NULL)
 			head = pp;
 		else
@@ -660,7 +683,7 @@ int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds, s64 *timeout)
 
 		walk = pp;
 		if (copy_from_user(pp->entries, ufds + nfds-i, 
-				sizeof(struct pollfd)*pp->len)) {
+				sizeof(struct pollfd)*num)) {
 			err = -EFAULT;
 			goto out_fds;
 		}
@@ -689,7 +712,8 @@ out_fds:
 	walk = head;
 	while(walk!=NULL) {
 		struct poll_list *pp = walk->next;
-		kfree(walk);
+		if (walk != stack_pp)
+			kfree(walk);
 		walk = pp;
 	}
 	poll_freewait(&table);
diff --git a/fs/seq_file.c b/fs/seq_file.c
index 7c40570b71dc..555b9ac04c25 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -37,7 +37,7 @@ int seq_open(struct file *file, struct seq_operations *op)
 		file->private_data = p;
 	}
 	memset(p, 0, sizeof(*p));
-	sema_init(&p->sem, 1);
+	mutex_init(&p->lock);
 	p->op = op;
 
 	/*
@@ -71,7 +71,7 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
 	void *p;
 	int err = 0;
 
-	down(&m->sem);
+	mutex_lock(&m->lock);
 	/*
 	 * seq_file->op->..m_start/m_stop/m_next may do special actions
 	 * or optimisations based on the file->f_version, so we want to
@@ -164,7 +164,7 @@ Done:
 	else
 		*ppos += copied;
 	file->f_version = m->version;
-	up(&m->sem);
+	mutex_unlock(&m->lock);
 	return copied;
 Enomem:
 	err = -ENOMEM;
@@ -237,7 +237,7 @@ loff_t seq_lseek(struct file *file, loff_t offset, int origin)
 	struct seq_file *m = (struct seq_file *)file->private_data;
 	long long retval = -EINVAL;
 
-	down(&m->sem);
+	mutex_lock(&m->lock);
 	m->version = file->f_version;
 	switch (origin) {
 		case 1:
@@ -260,7 +260,7 @@ loff_t seq_lseek(struct file *file, loff_t offset, int origin)
 				}
 			}
 	}
-	up(&m->sem);
+	mutex_unlock(&m->lock);
 	file->f_version = m->version;
 	return retval;
 }
diff --git a/fs/smbfs/dir.c b/fs/smbfs/dir.c
index 0424d06b147e..34c7a11d91f0 100644
--- a/fs/smbfs/dir.c
+++ b/fs/smbfs/dir.c
@@ -34,7 +34,7 @@ static int smb_rename(struct inode *, struct dentry *,
 static int smb_make_node(struct inode *,struct dentry *,int,dev_t);
 static int smb_link(struct dentry *, struct inode *, struct dentry *);
 
-struct file_operations smb_dir_operations =
+const struct file_operations smb_dir_operations =
 {
 	.read		= generic_read_dir,
 	.readdir	= smb_readdir,
diff --git a/fs/smbfs/file.c b/fs/smbfs/file.c
index 7042e62726a4..c56bd99a9701 100644
--- a/fs/smbfs/file.c
+++ b/fs/smbfs/file.c
@@ -401,7 +401,7 @@ smb_file_permission(struct inode *inode, int mask, struct nameidata *nd)
 	return error;
 }
 
-struct file_operations smb_file_operations =
+const struct file_operations smb_file_operations =
 {
 	.llseek		= remote_llseek,
 	.read		= smb_file_read,
diff --git a/fs/smbfs/inode.c b/fs/smbfs/inode.c
index 02e3e82d465c..fdeabc0a34f7 100644
--- a/fs/smbfs/inode.c
+++ b/fs/smbfs/inode.c
@@ -80,7 +80,8 @@ static int init_inodecache(void)
 {
 	smb_inode_cachep = kmem_cache_create("smb_inode_cache",
 					     sizeof(struct smb_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (smb_inode_cachep == NULL)
 		return -ENOMEM;
@@ -216,7 +217,7 @@ smb_set_inode_attr(struct inode *inode, struct smb_fattr *fattr)
 	if (inode->i_mtime.tv_sec != last_time || inode->i_size != last_sz) {
 		VERBOSE("%ld changed, old=%ld, new=%ld, oz=%ld, nz=%ld\n",
 			inode->i_ino,
-			(long) last_time, (long) inode->i_mtime,
+			(long) last_time, (long) inode->i_mtime.tv_sec,
 			(long) last_sz, (long) inode->i_size);
 
 		if (!S_ISDIR(inode->i_mode))
diff --git a/fs/smbfs/proto.h b/fs/smbfs/proto.h
index e866ec8660d0..47664597e6b1 100644
--- a/fs/smbfs/proto.h
+++ b/fs/smbfs/proto.h
@@ -35,7 +35,7 @@ extern int smb_proc_symlink(struct smb_sb_info *server, struct dentry *d, const
 extern int smb_proc_link(struct smb_sb_info *server, struct dentry *dentry, struct dentry *new_dentry);
 extern void smb_install_null_ops(struct smb_ops *ops);
 /* dir.c */
-extern struct file_operations smb_dir_operations;
+extern const struct file_operations smb_dir_operations;
 extern struct inode_operations smb_dir_inode_operations;
 extern struct inode_operations smb_dir_inode_operations_unix;
 extern void smb_new_dentry(struct dentry *dentry);
@@ -64,7 +64,7 @@ extern int smb_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat
 extern int smb_notify_change(struct dentry *dentry, struct iattr *attr);
 /* file.c */
 extern struct address_space_operations smb_file_aops;
-extern struct file_operations smb_file_operations;
+extern const struct file_operations smb_file_operations;
 extern struct inode_operations smb_file_inode_operations;
 /* ioctl.c */
 extern int smb_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg);
diff --git a/fs/splice.c b/fs/splice.c
new file mode 100644
index 000000000000..4a026f95884f
--- /dev/null
+++ b/fs/splice.c
@@ -0,0 +1,659 @@
+/*
+ * "splice": joining two ropes together by interweaving their strands.
+ *
+ * This is the "extended pipe" functionality, where a pipe is used as
+ * an arbitrary in-memory buffer. Think of a pipe as a small kernel
+ * buffer that you can use to transfer data from one end to the other.
+ *
+ * The traditional unix read/write is extended with a "splice()" operation
+ * that transfers data buffers to or from a pipe buffer.
+ *
+ * Named by Larry McVoy, original implementation from Linus, extended by
+ * Jens to support splicing to files and fixing the initial implementation
+ * bugs.
+ *
+ * Copyright (C) 2005 Jens Axboe <axboe@suse.de>
+ * Copyright (C) 2005 Linus Torvalds <torvalds@osdl.org>
+ *
+ */
+#include <linux/fs.h>
+#include <linux/file.h>
+#include <linux/pagemap.h>
+#include <linux/pipe_fs_i.h>
+#include <linux/mm_inline.h>
+#include <linux/swap.h>
+
+/*
+ * Passed to the actors
+ */
+struct splice_desc {
+	unsigned int len, total_len;	/* current and remaining length */
+	unsigned int flags;		/* splice flags */
+	struct file *file;		/* file to read/write */
+	loff_t pos;			/* file position */
+};
+
+static int page_cache_pipe_buf_steal(struct pipe_inode_info *info,
+				     struct pipe_buffer *buf)
+{
+	struct page *page = buf->page;
+
+	WARN_ON(!PageLocked(page));
+	WARN_ON(!PageUptodate(page));
+
+	if (!remove_mapping(page_mapping(page), page))
+		return 1;
+
+	if (PageLRU(page)) {
+		struct zone *zone = page_zone(page);
+
+		spin_lock_irq(&zone->lru_lock);
+		BUG_ON(!PageLRU(page));
+		__ClearPageLRU(page);
+		del_page_from_lru(zone, page);
+		spin_unlock_irq(&zone->lru_lock);
+	}
+
+	buf->stolen = 1;
+	return 0;
+}
+
+static void page_cache_pipe_buf_release(struct pipe_inode_info *info,
+					struct pipe_buffer *buf)
+{
+	page_cache_release(buf->page);
+	buf->page = NULL;
+	buf->stolen = 0;
+}
+
+static void *page_cache_pipe_buf_map(struct file *file,
+				     struct pipe_inode_info *info,
+				     struct pipe_buffer *buf)
+{
+	struct page *page = buf->page;
+
+	lock_page(page);
+
+	if (!PageUptodate(page)) {
+		unlock_page(page);
+		return ERR_PTR(-EIO);
+	}
+
+	if (!page->mapping) {
+		unlock_page(page);
+		return ERR_PTR(-ENODATA);
+	}
+
+	return kmap(buf->page);
+}
+
+static void page_cache_pipe_buf_unmap(struct pipe_inode_info *info,
+				      struct pipe_buffer *buf)
+{
+	if (!buf->stolen)
+		unlock_page(buf->page);
+	kunmap(buf->page);
+}
+
+static struct pipe_buf_operations page_cache_pipe_buf_ops = {
+	.can_merge = 0,
+	.map = page_cache_pipe_buf_map,
+	.unmap = page_cache_pipe_buf_unmap,
+	.release = page_cache_pipe_buf_release,
+	.steal = page_cache_pipe_buf_steal,
+};
+
+static ssize_t move_to_pipe(struct inode *inode, struct page **pages,
+			    int nr_pages, unsigned long offset,
+			    unsigned long len)
+{
+	struct pipe_inode_info *info;
+	int ret, do_wakeup, i;
+
+	ret = 0;
+	do_wakeup = 0;
+	i = 0;
+
+	mutex_lock(PIPE_MUTEX(*inode));
+
+	info = inode->i_pipe;
+	for (;;) {
+		int bufs;
+
+		if (!PIPE_READERS(*inode)) {
+			send_sig(SIGPIPE, current, 0);
+			if (!ret)
+				ret = -EPIPE;
+			break;
+		}
+
+		bufs = info->nrbufs;
+		if (bufs < PIPE_BUFFERS) {
+			int newbuf = (info->curbuf + bufs) & (PIPE_BUFFERS - 1);
+			struct pipe_buffer *buf = info->bufs + newbuf;
+			struct page *page = pages[i++];
+			unsigned long this_len;
+
+			this_len = PAGE_CACHE_SIZE - offset;
+			if (this_len > len)
+				this_len = len;
+
+			buf->page = page;
+			buf->offset = offset;
+			buf->len = this_len;
+			buf->ops = &page_cache_pipe_buf_ops;
+			info->nrbufs = ++bufs;
+			do_wakeup = 1;
+
+			ret += this_len;
+			len -= this_len;
+			offset = 0;
+			if (!--nr_pages)
+				break;
+			if (!len)
+				break;
+			if (bufs < PIPE_BUFFERS)
+				continue;
+
+			break;
+		}
+
+		if (signal_pending(current)) {
+			if (!ret)
+				ret = -ERESTARTSYS;
+			break;
+		}
+
+		if (do_wakeup) {
+			wake_up_interruptible_sync(PIPE_WAIT(*inode));
+			kill_fasync(PIPE_FASYNC_READERS(*inode), SIGIO,
+				    POLL_IN);
+			do_wakeup = 0;
+		}
+
+		PIPE_WAITING_WRITERS(*inode)++;
+		pipe_wait(inode);
+		PIPE_WAITING_WRITERS(*inode)--;
+	}
+
+	mutex_unlock(PIPE_MUTEX(*inode));
+
+	if (do_wakeup) {
+		wake_up_interruptible(PIPE_WAIT(*inode));
+		kill_fasync(PIPE_FASYNC_READERS(*inode), SIGIO, POLL_IN);
+	}
+
+	while (i < nr_pages)
+		page_cache_release(pages[i++]);
+
+	return ret;
+}
+
+static int __generic_file_splice_read(struct file *in, struct inode *pipe,
+				      size_t len)
+{
+	struct address_space *mapping = in->f_mapping;
+	unsigned int offset, nr_pages;
+	struct page *pages[PIPE_BUFFERS], *shadow[PIPE_BUFFERS];
+	struct page *page;
+	pgoff_t index, pidx;
+	int i, j;
+
+	index = in->f_pos >> PAGE_CACHE_SHIFT;
+	offset = in->f_pos & ~PAGE_CACHE_MASK;
+	nr_pages = (len + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
+
+	if (nr_pages > PIPE_BUFFERS)
+		nr_pages = PIPE_BUFFERS;
+
+	/*
+	 * initiate read-ahead on this page range
+	 */
+	do_page_cache_readahead(mapping, in, index, nr_pages);
+
+	/*
+	 * Get as many pages from the page cache as possible..
+	 * Start IO on the page cache entries we create (we
+	 * can assume that any pre-existing ones we find have
+	 * already had IO started on them).
+	 */
+	i = find_get_pages(mapping, index, nr_pages, pages);
+
+	/*
+	 * common case - we found all pages and they are contiguous,
+	 * kick them off
+	 */
+	if (i && (pages[i - 1]->index == index + i - 1))
+		goto splice_them;
+
+	/*
+	 * fill shadow[] with pages at the right locations, so we only
+	 * have to fill holes
+	 */
+	memset(shadow, 0, i * sizeof(struct page *));
+	for (j = 0, pidx = index; j < i; pidx++, j++)
+		shadow[pages[j]->index - pidx] = pages[j];
+
+	/*
+	 * now fill in the holes
+	 */
+	for (i = 0, pidx = index; i < nr_pages; pidx++, i++) {
+		int error;
+
+		if (shadow[i])
+			continue;
+
+		/*
+		 * no page there, look one up / create it
+		 */
+		page = find_or_create_page(mapping, pidx,
+						   mapping_gfp_mask(mapping));
+		if (!page)
+			break;
+
+		if (PageUptodate(page))
+			unlock_page(page);
+		else {
+			error = mapping->a_ops->readpage(in, page);
+
+			if (unlikely(error)) {
+				page_cache_release(page);
+				break;
+			}
+		}
+		shadow[i] = page;
+	}
+
+	if (!i) {
+		for (i = 0; i < nr_pages; i++) {
+			 if (shadow[i])
+				page_cache_release(shadow[i]);
+		}
+		return 0;
+	}
+
+	memcpy(pages, shadow, i * sizeof(struct page *));
+
+	/*
+	 * Now we splice them into the pipe..
+	 */
+splice_them:
+	return move_to_pipe(pipe, pages, i, offset, len);
+}
+
+ssize_t generic_file_splice_read(struct file *in, struct inode *pipe,
+				 size_t len, unsigned int flags)
+{
+	ssize_t spliced;
+	int ret;
+
+	ret = 0;
+	spliced = 0;
+	while (len) {
+		ret = __generic_file_splice_read(in, pipe, len);
+
+		if (ret <= 0)
+			break;
+
+		in->f_pos += ret;
+		len -= ret;
+		spliced += ret;
+	}
+
+	if (spliced)
+		return spliced;
+
+	return ret;
+}
+
+/*
+ * Send 'len' bytes to socket from 'file' at position 'pos' using sendpage().
+ */
+static int pipe_to_sendpage(struct pipe_inode_info *info,
+			    struct pipe_buffer *buf, struct splice_desc *sd)
+{
+	struct file *file = sd->file;
+	loff_t pos = sd->pos;
+	unsigned int offset;
+	ssize_t ret;
+	void *ptr;
+
+	/*
+	 * sub-optimal, but we are limited by the pipe ->map. we don't
+	 * need a kmap'ed buffer here, we just want to make sure we
+	 * have the page pinned if the pipe page originates from the
+	 * page cache
+	 */
+	ptr = buf->ops->map(file, info, buf);
+	if (IS_ERR(ptr))
+		return PTR_ERR(ptr);
+
+	offset = pos & ~PAGE_CACHE_MASK;
+
+	ret = file->f_op->sendpage(file, buf->page, offset, sd->len, &pos,
+					sd->len < sd->total_len);
+
+	buf->ops->unmap(info, buf);
+	if (ret == sd->len)
+		return 0;
+
+	return -EIO;
+}
+
+/*
+ * This is a little more tricky than the file -> pipe splicing. There are
+ * basically three cases:
+ *
+ *	- Destination page already exists in the address space and there
+ *	  are users of it. For that case we have no other option that
+ *	  copying the data. Tough luck.
+ *	- Destination page already exists in the address space, but there
+ *	  are no users of it. Make sure it's uptodate, then drop it. Fall
+ *	  through to last case.
+ *	- Destination page does not exist, we can add the pipe page to
+ *	  the page cache and avoid the copy.
+ *
+ * For now we just do the slower thing and always copy pages over, it's
+ * easier than migrating pages from the pipe to the target file. For the
+ * case of doing file | file splicing, the migrate approach had some LRU
+ * nastiness...
+ */
+static int pipe_to_file(struct pipe_inode_info *info, struct pipe_buffer *buf,
+			struct splice_desc *sd)
+{
+	struct file *file = sd->file;
+	struct address_space *mapping = file->f_mapping;
+	unsigned int offset;
+	struct page *page;
+	pgoff_t index;
+	char *src;
+	int ret;
+
+	/*
+	 * after this, page will be locked and unmapped
+	 */
+	src = buf->ops->map(file, info, buf);
+	if (IS_ERR(src))
+		return PTR_ERR(src);
+
+	index = sd->pos >> PAGE_CACHE_SHIFT;
+	offset = sd->pos & ~PAGE_CACHE_MASK;
+
+	/*
+	 * reuse buf page, if SPLICE_F_MOVE is set
+	 */
+	if (sd->flags & SPLICE_F_MOVE) {
+		if (buf->ops->steal(info, buf))
+			goto find_page;
+
+		page = buf->page;
+		if (add_to_page_cache_lru(page, mapping, index,
+						mapping_gfp_mask(mapping)))
+			goto find_page;
+	} else {
+find_page:
+		ret = -ENOMEM;
+		page = find_or_create_page(mapping, index,
+						mapping_gfp_mask(mapping));
+		if (!page)
+			goto out;
+
+		/*
+		 * If the page is uptodate, it is also locked. If it isn't
+		 * uptodate, we can mark it uptodate if we are filling the
+		 * full page. Otherwise we need to read it in first...
+		 */
+		if (!PageUptodate(page)) {
+			if (sd->len < PAGE_CACHE_SIZE) {
+				ret = mapping->a_ops->readpage(file, page);
+				if (unlikely(ret))
+					goto out;
+
+				lock_page(page);
+
+				if (!PageUptodate(page)) {
+					/*
+					 * page got invalidated, repeat
+					 */
+					if (!page->mapping) {
+						unlock_page(page);
+						page_cache_release(page);
+						goto find_page;
+					}
+					ret = -EIO;
+					goto out;
+				}
+			} else {
+				WARN_ON(!PageLocked(page));
+				SetPageUptodate(page);
+			}
+		}
+	}
+
+	ret = mapping->a_ops->prepare_write(file, page, 0, sd->len);
+	if (ret)
+		goto out;
+
+	if (!buf->stolen) {
+		char *dst = kmap_atomic(page, KM_USER0);
+
+		memcpy(dst + offset, src + buf->offset, sd->len);
+		flush_dcache_page(page);
+		kunmap_atomic(dst, KM_USER0);
+	}
+
+	ret = mapping->a_ops->commit_write(file, page, 0, sd->len);
+	if (ret < 0)
+		goto out;
+
+	set_page_dirty(page);
+	ret = write_one_page(page, 0);
+out:
+	if (ret < 0)
+		unlock_page(page);
+	if (!buf->stolen)
+		page_cache_release(page);
+	buf->ops->unmap(info, buf);
+	return ret;
+}
+
+typedef int (splice_actor)(struct pipe_inode_info *, struct pipe_buffer *,
+			   struct splice_desc *);
+
+static ssize_t move_from_pipe(struct inode *inode, struct file *out,
+			      size_t len, unsigned int flags,
+			      splice_actor *actor)
+{
+	struct pipe_inode_info *info;
+	int ret, do_wakeup, err;
+	struct splice_desc sd;
+
+	ret = 0;
+	do_wakeup = 0;
+
+	sd.total_len = len;
+	sd.flags = flags;
+	sd.file = out;
+	sd.pos = out->f_pos;
+
+	mutex_lock(PIPE_MUTEX(*inode));
+
+	info = inode->i_pipe;
+	for (;;) {
+		int bufs = info->nrbufs;
+
+		if (bufs) {
+			int curbuf = info->curbuf;
+			struct pipe_buffer *buf = info->bufs + curbuf;
+			struct pipe_buf_operations *ops = buf->ops;
+
+			sd.len = buf->len;
+			if (sd.len > sd.total_len)
+				sd.len = sd.total_len;
+
+			err = actor(info, buf, &sd);
+			if (err) {
+				if (!ret && err != -ENODATA)
+					ret = err;
+
+				break;
+			}
+
+			ret += sd.len;
+			buf->offset += sd.len;
+			buf->len -= sd.len;
+			if (!buf->len) {
+				buf->ops = NULL;
+				ops->release(info, buf);
+				curbuf = (curbuf + 1) & (PIPE_BUFFERS - 1);
+				info->curbuf = curbuf;
+				info->nrbufs = --bufs;
+				do_wakeup = 1;
+			}
+
+			sd.pos += sd.len;
+			sd.total_len -= sd.len;
+			if (!sd.total_len)
+				break;
+		}
+
+		if (bufs)
+			continue;
+		if (!PIPE_WRITERS(*inode))
+			break;
+		if (!PIPE_WAITING_WRITERS(*inode)) {
+			if (ret)
+				break;
+		}
+
+		if (signal_pending(current)) {
+			if (!ret)
+				ret = -ERESTARTSYS;
+			break;
+		}
+
+		if (do_wakeup) {
+			wake_up_interruptible_sync(PIPE_WAIT(*inode));
+			kill_fasync(PIPE_FASYNC_WRITERS(*inode),SIGIO,POLL_OUT);
+			do_wakeup = 0;
+		}
+
+		pipe_wait(inode);
+	}
+
+	mutex_unlock(PIPE_MUTEX(*inode));
+
+	if (do_wakeup) {
+		wake_up_interruptible(PIPE_WAIT(*inode));
+		kill_fasync(PIPE_FASYNC_WRITERS(*inode), SIGIO, POLL_OUT);
+	}
+
+	mutex_lock(&out->f_mapping->host->i_mutex);
+	out->f_pos = sd.pos;
+	mutex_unlock(&out->f_mapping->host->i_mutex);
+	return ret;
+
+}
+
+ssize_t generic_file_splice_write(struct inode *inode, struct file *out,
+				  size_t len, unsigned int flags)
+{
+	return move_from_pipe(inode, out, len, flags, pipe_to_file);
+}
+
+ssize_t generic_splice_sendpage(struct inode *inode, struct file *out,
+				size_t len, unsigned int flags)
+{
+	return move_from_pipe(inode, out, len, flags, pipe_to_sendpage);
+}
+
+static long do_splice_from(struct inode *pipe, struct file *out, size_t len,
+			   unsigned int flags)
+{
+	loff_t pos;
+	int ret;
+
+	if (!out->f_op || !out->f_op->splice_write)
+		return -EINVAL;
+
+	if (!(out->f_mode & FMODE_WRITE))
+		return -EBADF;
+
+	pos = out->f_pos;
+	ret = rw_verify_area(WRITE, out, &pos, len);
+	if (unlikely(ret < 0))
+		return ret;
+
+	return out->f_op->splice_write(pipe, out, len, flags);
+}
+
+static long do_splice_to(struct file *in, struct inode *pipe, size_t len,
+			 unsigned int flags)
+{
+	loff_t pos, isize, left;
+	int ret;
+
+	if (!in->f_op || !in->f_op->splice_read)
+		return -EINVAL;
+
+	if (!(in->f_mode & FMODE_READ))
+		return -EBADF;
+
+	pos = in->f_pos;
+	ret = rw_verify_area(READ, in, &pos, len);
+	if (unlikely(ret < 0))
+		return ret;
+
+	isize = i_size_read(in->f_mapping->host);
+	if (unlikely(in->f_pos >= isize))
+		return 0;
+	
+	left = isize - in->f_pos;
+	if (left < len)
+		len = left;
+
+	return in->f_op->splice_read(in, pipe, len, flags);
+}
+
+static long do_splice(struct file *in, struct file *out, size_t len,
+		      unsigned int flags)
+{
+	struct inode *pipe;
+
+	pipe = in->f_dentry->d_inode;
+	if (pipe->i_pipe)
+		return do_splice_from(pipe, out, len, flags);
+
+	pipe = out->f_dentry->d_inode;
+	if (pipe->i_pipe)
+		return do_splice_to(in, pipe, len, flags);
+
+	return -EINVAL;
+}
+
+asmlinkage long sys_splice(int fdin, int fdout, size_t len, unsigned int flags)
+{
+	long error;
+	struct file *in, *out;
+	int fput_in, fput_out;
+
+	if (unlikely(!len))
+		return 0;
+
+	error = -EBADF;
+	in = fget_light(fdin, &fput_in);
+	if (in) {
+		if (in->f_mode & FMODE_READ) {
+			out = fget_light(fdout, &fput_out);
+			if (out) {
+				if (out->f_mode & FMODE_WRITE)
+					error = do_splice(in, out, len, flags);
+				fput_light(out, fput_out);
+			}
+		}
+
+		fput_light(in, fput_in);
+	}
+
+	return error;
+}
diff --git a/fs/super.c b/fs/super.c
index e20b5580afd5..a66f66bb8049 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -37,6 +37,7 @@
 #include <linux/writeback.h>		/* for the emergency remount stuff */
 #include <linux/idr.h>
 #include <linux/kobject.h>
+#include <linux/mutex.h>
 #include <asm/uaccess.h>
 
 
@@ -55,11 +56,10 @@ DEFINE_SPINLOCK(sb_lock);
  */
 static struct super_block *alloc_super(void)
 {
-	struct super_block *s = kmalloc(sizeof(struct super_block),  GFP_USER);
+	struct super_block *s = kzalloc(sizeof(struct super_block),  GFP_USER);
 	static struct super_operations default_op;
 
 	if (s) {
-		memset(s, 0, sizeof(struct super_block));
 		if (security_sb_alloc(s)) {
 			kfree(s);
 			s = NULL;
@@ -76,9 +76,9 @@ static struct super_block *alloc_super(void)
 		down_write(&s->s_umount);
 		s->s_count = S_BIAS;
 		atomic_set(&s->s_active, 1);
-		sema_init(&s->s_vfs_rename_sem,1);
-		sema_init(&s->s_dquot.dqio_sem, 1);
-		sema_init(&s->s_dquot.dqonoff_sem, 1);
+		mutex_init(&s->s_vfs_rename_mutex);
+		mutex_init(&s->s_dquot.dqio_mutex);
+		mutex_init(&s->s_dquot.dqonoff_mutex);
 		init_rwsem(&s->s_dquot.dqptr_sem);
 		init_waitqueue_head(&s->s_wait_unfrozen);
 		s->s_maxbytes = MAX_NON_LFS;
@@ -381,9 +381,9 @@ restart:
 void sync_filesystems(int wait)
 {
 	struct super_block *sb;
-	static DECLARE_MUTEX(mutex);
+	static DEFINE_MUTEX(mutex);
 
-	down(&mutex);		/* Could be down_interruptible */
+	mutex_lock(&mutex);		/* Could be down_interruptible */
 	spin_lock(&sb_lock);
 	list_for_each_entry(sb, &super_blocks, s_list) {
 		if (!sb->s_op->sync_fs)
@@ -412,7 +412,7 @@ restart:
 			goto restart;
 	}
 	spin_unlock(&sb_lock);
-	up(&mutex);
+	mutex_unlock(&mutex);
 }
 
 /**
@@ -693,9 +693,9 @@ struct super_block *get_sb_bdev(struct file_system_type *fs_type,
 	 * will protect the lockfs code from trying to start a snapshot
 	 * while we are mounting
 	 */
-	down(&bdev->bd_mount_sem);
+	mutex_lock(&bdev->bd_mount_mutex);
 	s = sget(fs_type, test_bdev_super, set_bdev_super, bdev);
-	up(&bdev->bd_mount_sem);
+	mutex_unlock(&bdev->bd_mount_mutex);
 	if (IS_ERR(s))
 		goto out;
 
@@ -712,7 +712,7 @@ struct super_block *get_sb_bdev(struct file_system_type *fs_type,
 		s->s_flags = flags;
 		strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id));
 		sb_set_blocksize(s, block_size(bdev));
-		error = fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
+		error = fill_super(s, data, flags & MS_SILENT ? 1 : 0);
 		if (error) {
 			up_write(&s->s_umount);
 			deactivate_super(s);
@@ -756,7 +756,7 @@ struct super_block *get_sb_nodev(struct file_system_type *fs_type,
 
 	s->s_flags = flags;
 
-	error = fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
+	error = fill_super(s, data, flags & MS_SILENT ? 1 : 0);
 	if (error) {
 		up_write(&s->s_umount);
 		deactivate_super(s);
@@ -785,7 +785,7 @@ struct super_block *get_sb_single(struct file_system_type *fs_type,
 		return s;
 	if (!s->s_root) {
 		s->s_flags = flags;
-		error = fill_super(s, data, flags & MS_VERBOSE ? 1 : 0);
+		error = fill_super(s, data, flags & MS_SILENT ? 1 : 0);
 		if (error) {
 			up_write(&s->s_umount);
 			deactivate_super(s);
diff --git a/fs/sysfs/bin.c b/fs/sysfs/bin.c
index 78899eeab974..c16a93c353c0 100644
--- a/fs/sysfs/bin.c
+++ b/fs/sysfs/bin.c
@@ -163,7 +163,7 @@ static int release(struct inode * inode, struct file * file)
 	return 0;
 }
 
-struct file_operations bin_fops = {
+const struct file_operations bin_fops = {
 	.read		= read,
 	.write		= write,
 	.mmap		= mmap,
diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c
index 49bd219275db..f26880a4785e 100644
--- a/fs/sysfs/dir.c
+++ b/fs/sysfs/dir.c
@@ -50,6 +50,32 @@ static struct sysfs_dirent * sysfs_new_dirent(struct sysfs_dirent * parent_sd,
 	return sd;
 }
 
+/**
+ *
+ * Return -EEXIST if there is already a sysfs element with the same name for
+ * the same parent.
+ *
+ * called with parent inode's i_mutex held
+ */
+int sysfs_dirent_exist(struct sysfs_dirent *parent_sd,
+			  const unsigned char *new)
+{
+	struct sysfs_dirent * sd;
+
+	list_for_each_entry(sd, &parent_sd->s_children, s_sibling) {
+		if (sd->s_element) {
+			const unsigned char *existing = sysfs_get_name(sd);
+			if (strcmp(existing, new))
+				continue;
+			else
+				return -EEXIST;
+		}
+	}
+
+	return 0;
+}
+
+
 int sysfs_make_dirent(struct sysfs_dirent * parent_sd, struct dentry * dentry,
 			void * element, umode_t mode, int type)
 {
@@ -102,7 +128,11 @@ static int create_dir(struct kobject * k, struct dentry * p,
 	mutex_lock(&p->d_inode->i_mutex);
 	*d = lookup_one_len(n, p, strlen(n));
 	if (!IS_ERR(*d)) {
-		error = sysfs_make_dirent(p->d_fsdata, *d, k, mode, SYSFS_DIR);
+ 		if (sysfs_dirent_exist(p->d_fsdata, n))
+  			error = -EEXIST;
+  		else
+			error = sysfs_make_dirent(p->d_fsdata, *d, k, mode,
+								SYSFS_DIR);
 		if (!error) {
 			error = sysfs_create(*d, mode, init_dir);
 			if (!error) {
@@ -302,6 +332,7 @@ void sysfs_remove_dir(struct kobject * kobj)
 	 * Drop reference from dget() on entrance.
 	 */
 	dput(dentry);
+	kobj->dentry = NULL;
 }
 
 int sysfs_rename_dir(struct kobject * kobj, const char *new_name)
@@ -472,14 +503,10 @@ static loff_t sysfs_dir_lseek(struct file * file, loff_t offset, int origin)
 	return offset;
 }
 
-struct file_operations sysfs_dir_operations = {
+const struct file_operations sysfs_dir_operations = {
 	.open		= sysfs_dir_open,
 	.release	= sysfs_dir_close,
 	.llseek		= sysfs_dir_lseek,
 	.read		= generic_read_dir,
 	.readdir	= sysfs_readdir,
 };
-
-EXPORT_SYMBOL_GPL(sysfs_create_dir);
-EXPORT_SYMBOL_GPL(sysfs_remove_dir);
-EXPORT_SYMBOL_GPL(sysfs_rename_dir);
diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c
index d0e3d8495165..830f76fa098c 100644
--- a/fs/sysfs/file.c
+++ b/fs/sysfs/file.c
@@ -301,9 +301,8 @@ static int check_perm(struct inode * inode, struct file * file)
 	/* No error? Great, allocate a buffer for the file, and store it
 	 * it in file->private_data for easy access.
 	 */
-	buffer = kmalloc(sizeof(struct sysfs_buffer),GFP_KERNEL);
+	buffer = kzalloc(sizeof(struct sysfs_buffer), GFP_KERNEL);
 	if (buffer) {
-		memset(buffer,0,sizeof(struct sysfs_buffer));
 		init_MUTEX(&buffer->sem);
 		buffer->needs_read_fill = 1;
 		buffer->ops = ops;
@@ -349,7 +348,7 @@ static int sysfs_release(struct inode * inode, struct file * filp)
 	return 0;
 }
 
-struct file_operations sysfs_file_operations = {
+const struct file_operations sysfs_file_operations = {
 	.read		= sysfs_read_file,
 	.write		= sysfs_write_file,
 	.llseek		= generic_file_llseek,
@@ -362,10 +361,12 @@ int sysfs_add_file(struct dentry * dir, const struct attribute * attr, int type)
 {
 	struct sysfs_dirent * parent_sd = dir->d_fsdata;
 	umode_t mode = (attr->mode & S_IALLUGO) | S_IFREG;
-	int error = 0;
+	int error = -EEXIST;
 
 	mutex_lock(&dir->d_inode->i_mutex);
-	error = sysfs_make_dirent(parent_sd, NULL, (void *) attr, mode, type);
+	if (!sysfs_dirent_exist(parent_sd, attr->name))
+		error = sysfs_make_dirent(parent_sd, NULL, (void *)attr,
+					  mode, type);
 	mutex_unlock(&dir->d_inode->i_mutex);
 
 	return error;
diff --git a/fs/sysfs/inode.c b/fs/sysfs/inode.c
index 689f7bcfaf30..4c29ac41ac3e 100644
--- a/fs/sysfs/inode.c
+++ b/fs/sysfs/inode.c
@@ -54,11 +54,10 @@ int sysfs_setattr(struct dentry * dentry, struct iattr * iattr)
 
 	if (!sd_iattr) {
 		/* setting attributes for the first time, allocate now */
-		sd_iattr = kmalloc(sizeof(struct iattr), GFP_KERNEL);
+		sd_iattr = kzalloc(sizeof(struct iattr), GFP_KERNEL);
 		if (!sd_iattr)
 			return -ENOMEM;
 		/* assign default attributes */
-		memset(sd_iattr, 0, sizeof(struct iattr));
 		sd_iattr->ia_mode = sd->s_mode;
 		sd_iattr->ia_uid = 0;
 		sd_iattr->ia_gid = 0;
@@ -227,12 +226,16 @@ void sysfs_drop_dentry(struct sysfs_dirent * sd, struct dentry * parent)
 void sysfs_hash_and_remove(struct dentry * dir, const char * name)
 {
 	struct sysfs_dirent * sd;
-	struct sysfs_dirent * parent_sd = dir->d_fsdata;
+	struct sysfs_dirent * parent_sd;
+
+	if (!dir)
+		return;
 
 	if (dir->d_inode == NULL)
 		/* no inode means this hasn't been made visible yet */
 		return;
 
+	parent_sd = dir->d_fsdata;
 	mutex_lock(&dir->d_inode->i_mutex);
 	list_for_each_entry(sd, &parent_sd->s_children, s_sibling) {
 		if (!sd->s_element)
diff --git a/fs/sysfs/symlink.c b/fs/sysfs/symlink.c
index e38d6338a20d..d2eac3ceed5f 100644
--- a/fs/sysfs/symlink.c
+++ b/fs/sysfs/symlink.c
@@ -66,6 +66,7 @@ static int sysfs_add_link(struct dentry * parent, const char * name, struct kobj
 	if (!error)
 		return 0;
 
+	kobject_put(target);
 	kfree(sl->link_name);
 exit2:
 	kfree(sl);
@@ -82,12 +83,13 @@ exit1:
 int sysfs_create_link(struct kobject * kobj, struct kobject * target, const char * name)
 {
 	struct dentry * dentry = kobj->dentry;
-	int error = 0;
+	int error = -EEXIST;
 
 	BUG_ON(!kobj || !kobj->dentry || !name);
 
 	mutex_lock(&dentry->d_inode->i_mutex);
-	error = sysfs_add_link(dentry, name, target);
+	if (!sysfs_dirent_exist(dentry->d_fsdata, name))
+		error = sysfs_add_link(dentry, name, target);
 	mutex_unlock(&dentry->d_inode->i_mutex);
 	return error;
 }
diff --git a/fs/sysfs/sysfs.h b/fs/sysfs/sysfs.h
index 3f8953e0e5d0..32958a7c50e9 100644
--- a/fs/sysfs/sysfs.h
+++ b/fs/sysfs/sysfs.h
@@ -5,6 +5,7 @@ extern kmem_cache_t *sysfs_dir_cachep;
 extern struct inode * sysfs_new_inode(mode_t mode, struct sysfs_dirent *);
 extern int sysfs_create(struct dentry *, int mode, int (*init)(struct inode *));
 
+extern int sysfs_dirent_exist(struct sysfs_dirent *, const unsigned char *);
 extern int sysfs_make_dirent(struct sysfs_dirent *, struct dentry *, void *,
 				umode_t, int);
 
@@ -20,9 +21,9 @@ extern int sysfs_setattr(struct dentry *dentry, struct iattr *iattr);
 
 extern struct rw_semaphore sysfs_rename_sem;
 extern struct super_block * sysfs_sb;
-extern struct file_operations sysfs_dir_operations;
-extern struct file_operations sysfs_file_operations;
-extern struct file_operations bin_fops;
+extern const struct file_operations sysfs_dir_operations;
+extern const struct file_operations sysfs_file_operations;
+extern const struct file_operations bin_fops;
 extern struct inode_operations sysfs_dir_inode_operations;
 extern struct inode_operations sysfs_symlink_inode_operations;
 
diff --git a/fs/sysv/dir.c b/fs/sysv/dir.c
index cce8b05cba5a..8c66e9270dd6 100644
--- a/fs/sysv/dir.c
+++ b/fs/sysv/dir.c
@@ -20,7 +20,7 @@
 
 static int sysv_readdir(struct file *, void *, filldir_t);
 
-struct file_operations sysv_dir_operations = {
+const struct file_operations sysv_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= sysv_readdir,
 	.fsync		= sysv_sync_file,
diff --git a/fs/sysv/file.c b/fs/sysv/file.c
index da69abc06240..a59e303135fa 100644
--- a/fs/sysv/file.c
+++ b/fs/sysv/file.c
@@ -19,7 +19,7 @@
  * We have mostly NULLs here: the current defaults are OK for
  * the coh filesystem.
  */
-struct file_operations sysv_file_operations = {
+const struct file_operations sysv_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/sysv/inode.c b/fs/sysv/inode.c
index fa33eceb0011..3ff89cc5833a 100644
--- a/fs/sysv/inode.c
+++ b/fs/sysv/inode.c
@@ -342,7 +342,7 @@ int __init sysv_init_icache(void)
 {
 	sysv_inode_cachep = kmem_cache_create("sysv_inode_cache",
 			sizeof(struct sysv_inode_info), 0,
-			SLAB_RECLAIM_ACCOUNT,
+			SLAB_RECLAIM_ACCOUNT|SLAB_MEM_SPREAD,
 			init_once, NULL);
 	if (!sysv_inode_cachep)
 		return -ENOMEM;
diff --git a/fs/sysv/namei.c b/fs/sysv/namei.c
index 7f0e4b53085e..b8a73f716fbe 100644
--- a/fs/sysv/namei.c
+++ b/fs/sysv/namei.c
@@ -16,18 +16,6 @@
 #include <linux/smp_lock.h>
 #include "sysv.h"
 
-static inline void inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static int add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = sysv_add_link(dentry, inode);
@@ -35,7 +23,7 @@ static int add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -124,7 +112,7 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	goto out;
 }
@@ -138,7 +126,7 @@ static int sysv_link(struct dentry * old_dentry, struct inode * dir,
 		return -EMLINK;
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
 	return add_nondir(dentry, inode);
@@ -151,7 +139,7 @@ static int sysv_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 
 	if (dir->i_nlink >= SYSV_SB(dir->i_sb)->s_link_max) 
 		goto out;
-	inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = sysv_new_inode(dir, S_IFDIR|mode);
 	err = PTR_ERR(inode);
@@ -160,7 +148,7 @@ static int sysv_mkdir(struct inode * dir, struct dentry *dentry, int mode)
 
 	sysv_set_inode(inode, 0);
 
-	inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = sysv_make_empty(inode, dir);
 	if (err)
@@ -175,11 +163,11 @@ out:
 	return err;
 
 out_fail:
-	dec_count(inode);
-	dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 out_dir:
-	dec_count(dir);
+	inode_dec_link_count(dir);
 	goto out;
 }
 
@@ -199,7 +187,7 @@ static int sysv_unlink(struct inode * dir, struct dentry * dentry)
 		goto out;
 
 	inode->i_ctime = dir->i_ctime;
-	dec_count(inode);
+	inode_dec_link_count(inode);
 out:
 	return err;
 }
@@ -213,8 +201,8 @@ static int sysv_rmdir(struct inode * dir, struct dentry * dentry)
 		err = sysv_unlink(dir, dentry);
 		if (!err) {
 			inode->i_size = 0;
-			dec_count(inode);
-			dec_count(dir);
+			inode_dec_link_count(inode);
+			inode_dec_link_count(dir);
 		}
 	}
 	return err;
@@ -258,34 +246,34 @@ static int sysv_rename(struct inode * old_dir, struct dentry * old_dentry,
 		new_de = sysv_find_entry(new_dentry, &new_page);
 		if (!new_de)
 			goto out_dir;
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		sysv_set_link(new_de, new_page, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= SYSV_SB(new_dir->i_sb)->s_link_max)
 				goto out_dir;
 		}
-		inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = sysv_add_link(new_dentry, old_inode);
 		if (err) {
-			dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	sysv_delete_entry(old_de, old_page);
-	dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		sysv_set_link(dir_de, dir_page, new_dir);
-		dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	return 0;
 
diff --git a/fs/sysv/super.c b/fs/sysv/super.c
index 59e76b51142b..e92b991e6dda 100644
--- a/fs/sysv/super.c
+++ b/fs/sysv/super.c
@@ -377,10 +377,10 @@ static int sysv_fill_super(struct super_block *sb, void *data, int silent)
 	sbi->s_sb = sb;
 	sbi->s_block_base = 0;
 	sb->s_fs_info = sbi;
-	
+
 	sb_set_blocksize(sb, BLOCK_SIZE);
 
-	for (i = 0; i < sizeof(flavours)/sizeof(flavours[0]) && !size; i++) {
+	for (i = 0; i < ARRAY_SIZE(flavours) && !size; i++) {
 		brelse(bh);
 		bh = sb_bread(sb, flavours[i].block);
 		if (!bh)
diff --git a/fs/sysv/sysv.h b/fs/sysv/sysv.h
index b7f9b4a42aab..393a480e4deb 100644
--- a/fs/sysv/sysv.h
+++ b/fs/sysv/sysv.h
@@ -159,8 +159,8 @@ extern ino_t sysv_inode_by_name(struct dentry *);
 extern struct inode_operations sysv_file_inode_operations;
 extern struct inode_operations sysv_dir_inode_operations;
 extern struct inode_operations sysv_fast_symlink_inode_operations;
-extern struct file_operations sysv_file_operations;
-extern struct file_operations sysv_dir_operations;
+extern const struct file_operations sysv_file_operations;
+extern const struct file_operations sysv_dir_operations;
 extern struct address_space_operations sysv_aops;
 extern struct super_operations sysv_sops;
 extern struct dentry_operations sysv_dentry_operations;
diff --git a/fs/udf/balloc.c b/fs/udf/balloc.c
index 201049ac8a96..ea521f846d97 100644
--- a/fs/udf/balloc.c
+++ b/fs/udf/balloc.c
@@ -152,7 +152,7 @@ static void udf_bitmap_free_blocks(struct super_block * sb,
 	int bitmap_nr;
 	unsigned long overflow;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (bloc.logicalBlockNum < 0 ||
 		(bloc.logicalBlockNum + count) > UDF_SB_PARTLEN(sb, bloc.partitionReferenceNum))
 	{
@@ -211,7 +211,7 @@ error_return:
 	sb->s_dirt = 1;
 	if (UDF_SB_LVIDBH(sb))
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return;
 }
 
@@ -226,7 +226,7 @@ static int udf_bitmap_prealloc_blocks(struct super_block * sb,
 	int nr_groups, bitmap_nr;
 	struct buffer_head *bh;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (first_block < 0 || first_block >= UDF_SB_PARTLEN(sb, partition))
 		goto out;
 
@@ -275,7 +275,7 @@ out:
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
 	}
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return alloc_count;
 }
 
@@ -291,7 +291,7 @@ static int udf_bitmap_new_block(struct super_block * sb,
 	int newblock = 0;
 
 	*err = -ENOSPC;
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 
 repeat:
 	if (goal < 0 || goal >= UDF_SB_PARTLEN(sb, partition))
@@ -364,7 +364,7 @@ repeat:
 	}
 	if (i >= (nr_groups*2))
 	{
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		return newblock;
 	}
 	if (bit < sb->s_blocksize << 3)
@@ -373,7 +373,7 @@ repeat:
 		bit = udf_find_next_one_bit(bh->b_data, sb->s_blocksize << 3, group_start << 3);
 	if (bit >= sb->s_blocksize << 3)
 	{
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		return 0;
 	}
 
@@ -387,7 +387,7 @@ got_block:
 	 */
 	if (inode && DQUOT_ALLOC_BLOCK(inode, 1))
 	{
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		*err = -EDQUOT;
 		return 0;
 	}
@@ -410,13 +410,13 @@ got_block:
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
 	}
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	*err = 0;
 	return newblock;
 
 error_return:
 	*err = -EIO;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return 0;
 }
 
@@ -433,7 +433,7 @@ static void udf_table_free_blocks(struct super_block * sb,
 	int8_t etype;
 	int i;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (bloc.logicalBlockNum < 0 ||
 		(bloc.logicalBlockNum + count) > UDF_SB_PARTLEN(sb, bloc.partitionReferenceNum))
 	{
@@ -666,7 +666,7 @@ static void udf_table_free_blocks(struct super_block * sb,
 
 error_return:
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return;
 }
 
@@ -692,7 +692,7 @@ static int udf_table_prealloc_blocks(struct super_block * sb,
 	else
 		return 0;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	extoffset = sizeof(struct unallocSpaceEntry);
 	bloc = UDF_I_LOCATION(table);
 
@@ -736,7 +736,7 @@ static int udf_table_prealloc_blocks(struct super_block * sb,
 		mark_buffer_dirty(UDF_SB_LVIDBH(sb));
 		sb->s_dirt = 1;
 	}
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	return alloc_count;
 }
 
@@ -761,7 +761,7 @@ static int udf_table_new_block(struct super_block * sb,
 	else
 		return newblock;
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (goal < 0 || goal >= UDF_SB_PARTLEN(sb, partition))
 		goal = 0;
 
@@ -811,7 +811,7 @@ static int udf_table_new_block(struct super_block * sb,
 	if (spread == 0xFFFFFFFF)
 	{
 		udf_release_data(goal_bh);
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		return 0;
 	}
 
@@ -827,7 +827,7 @@ static int udf_table_new_block(struct super_block * sb,
 	if (inode && DQUOT_ALLOC_BLOCK(inode, 1))
 	{
 		udf_release_data(goal_bh);
-		up(&sbi->s_alloc_sem);
+		mutex_unlock(&sbi->s_alloc_mutex);
 		*err = -EDQUOT;
 		return 0;
 	}
@@ -846,7 +846,7 @@ static int udf_table_new_block(struct super_block * sb,
 	}
 
 	sb->s_dirt = 1;
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 	*err = 0;
 	return newblock;
 }
diff --git a/fs/udf/dir.c b/fs/udf/dir.c
index f5222527fe39..8c28efa3b8ff 100644
--- a/fs/udf/dir.c
+++ b/fs/udf/dir.c
@@ -42,7 +42,7 @@ static int do_udf_readdir(struct inode *, struct file *, filldir_t, void *);
 
 /* readdir and lookup functions */
 
-struct file_operations udf_dir_operations = {
+const struct file_operations udf_dir_operations = {
 	.read			= generic_read_dir,
 	.readdir		= udf_readdir,
 	.ioctl			= udf_ioctl,
diff --git a/fs/udf/file.c b/fs/udf/file.c
index a6f2acc1f15c..e34b00e303f1 100644
--- a/fs/udf/file.c
+++ b/fs/udf/file.c
@@ -248,7 +248,7 @@ static int udf_release_file(struct inode * inode, struct file * filp)
 	return 0;
 }
 
-struct file_operations udf_file_operations = {
+const struct file_operations udf_file_operations = {
 	.read			= generic_file_read,
 	.ioctl			= udf_ioctl,
 	.open			= generic_file_open,
diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c
index c9b707b470ca..3873c672cb4c 100644
--- a/fs/udf/ialloc.c
+++ b/fs/udf/ialloc.c
@@ -42,7 +42,7 @@ void udf_free_inode(struct inode * inode)
 
 	clear_inode(inode);
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	if (sbi->s_lvidbh) {
 		if (S_ISDIR(inode->i_mode))
 			UDF_SB_LVIDIU(sb)->numDirs =
@@ -53,7 +53,7 @@ void udf_free_inode(struct inode * inode)
 		
 		mark_buffer_dirty(sbi->s_lvidbh);
 	}
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 
 	udf_free_blocks(sb, NULL, UDF_I_LOCATION(inode), 0, 1);
 }
@@ -83,7 +83,7 @@ struct inode * udf_new_inode (struct inode *dir, int mode, int * err)
 		return NULL;
 	}
 
-	down(&sbi->s_alloc_sem);
+	mutex_lock(&sbi->s_alloc_mutex);
 	UDF_I_UNIQUE(inode) = 0;
 	UDF_I_LENEXTENTS(inode) = 0;
 	UDF_I_NEXT_ALLOC_BLOCK(inode) = 0;
@@ -148,7 +148,7 @@ struct inode * udf_new_inode (struct inode *dir, int mode, int * err)
 		UDF_I_CRTIME(inode) = current_fs_time(inode->i_sb);
 	insert_inode_hash(inode);
 	mark_inode_dirty(inode);
-	up(&sbi->s_alloc_sem);
+	mutex_unlock(&sbi->s_alloc_mutex);
 
 	if (DQUOT_ALLOC_INODE(inode))
 	{
diff --git a/fs/udf/inode.c b/fs/udf/inode.c
index d04cff2273b6..81e0e8459af1 100644
--- a/fs/udf/inode.c
+++ b/fs/udf/inode.c
@@ -1341,13 +1341,11 @@ udf_update_inode(struct inode *inode, int do_sync)
 
 	if (UDF_QUERY_FLAG(inode->i_sb, UDF_FLAG_UID_FORGET))
 		fe->uid = cpu_to_le32(-1);
-	else if (inode->i_uid != UDF_SB(inode->i_sb)->s_uid)
-		fe->uid = cpu_to_le32(inode->i_uid);
+	else fe->uid = cpu_to_le32(inode->i_uid);
 
 	if (UDF_QUERY_FLAG(inode->i_sb, UDF_FLAG_GID_FORGET))
 		fe->gid = cpu_to_le32(-1);
-	else if (inode->i_gid != UDF_SB(inode->i_sb)->s_gid)
-		fe->gid = cpu_to_le32(inode->i_gid);
+	else fe->gid = cpu_to_le32(inode->i_gid);
 
 	udfperms =	((inode->i_mode & S_IRWXO)     ) |
 			((inode->i_mode & S_IRWXG) << 2) |
diff --git a/fs/udf/super.c b/fs/udf/super.c
index 368d8f81fe54..e45789fe38e8 100644
--- a/fs/udf/super.c
+++ b/fs/udf/super.c
@@ -140,7 +140,8 @@ static int init_inodecache(void)
 {
 	udf_inode_cachep = kmem_cache_create("udf_inode_cache",
 					     sizeof(struct udf_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (udf_inode_cachep == NULL)
 		return -ENOMEM;
@@ -660,8 +661,7 @@ udf_find_anchor(struct super_block *sb)
 		 *     lastblock
 		 *  however, if the disc isn't closed, it could be 512 */
 
-		for (i=0; (!lastblock && i<sizeof(last)/sizeof(int)); i++)
-		{
+		for (i = 0; !lastblock && i < ARRAY_SIZE(last); i++) {
 			if (last[i] < 0 || !(bh = sb_bread(sb, last[i])))
 			{
 				ident = location = 0;
@@ -672,7 +672,7 @@ udf_find_anchor(struct super_block *sb)
 				location = le32_to_cpu(((tag *)bh->b_data)->tagLocation);
 				udf_release_data(bh);
 			}
-	
+
 			if (ident == TAG_IDENT_AVDP)
 			{
 				if (location == last[i] - UDF_SB_SESSION(sb))
@@ -753,8 +753,7 @@ udf_find_anchor(struct super_block *sb)
 		}
 	}
 
-	for (i=0; i<sizeof(UDF_SB_ANCHOR(sb))/sizeof(int); i++)
-	{
+	for (i = 0; i < ARRAY_SIZE(UDF_SB_ANCHOR(sb)); i++) {
 		if (UDF_SB_ANCHOR(sb)[i])
 		{
 			if (!(bh = udf_read_tagged(sb,
@@ -1313,8 +1312,7 @@ udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset)
 	if (!sb)
 		return 1;
 
-	for (i=0; i<sizeof(UDF_SB_ANCHOR(sb))/sizeof(int); i++)
-	{
+	for (i = 0; i < ARRAY_SIZE(UDF_SB_ANCHOR(sb)); i++) {
 		if (UDF_SB_ANCHOR(sb)[i] && (bh = udf_read_tagged(sb,
 			UDF_SB_ANCHOR(sb)[i], UDF_SB_ANCHOR(sb)[i], &ident)))
 		{
@@ -1325,7 +1323,7 @@ udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset)
 			main_e = le32_to_cpu( anchor->mainVolDescSeqExt.extLength );
 			main_e = main_e >> sb->s_blocksize_bits;
 			main_e += main_s;
-	
+
 			/* Locate the reserve sequence */
 			reserve_s = le32_to_cpu(anchor->reserveVolDescSeqExt.extLocation);
 			reserve_e = le32_to_cpu(anchor->reserveVolDescSeqExt.extLength);
@@ -1344,12 +1342,10 @@ udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset)
 		}
 	}
 
-	if (i == sizeof(UDF_SB_ANCHOR(sb))/sizeof(int))
-	{
+	if (i == ARRAY_SIZE(UDF_SB_ANCHOR(sb))) {
 		udf_debug("No Anchor block found\n");
 		return 1;
-	}
-	else
+	} else
 		udf_debug("Using anchor in block %d\n", UDF_SB_ANCHOR(sb)[i]);
 
 	for (i=0; i<UDF_SB_NUMPARTS(sb); i++)
@@ -1515,7 +1511,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent)
 	sb->s_fs_info = sbi;
 	memset(UDF_SB(sb), 0x00, sizeof(struct udf_sb_info));
 
-	init_MUTEX(&sbi->s_alloc_sem);
+	mutex_init(&sbi->s_alloc_mutex);
 
 	if (!udf_parse_options((char *)options, &uopt))
 		goto error_out;
diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h
index 1d5800e0cbe7..023e19ba5a2e 100644
--- a/fs/udf/udfdecl.h
+++ b/fs/udf/udfdecl.h
@@ -44,9 +44,9 @@ struct buffer_head;
 struct super_block;
 
 extern struct inode_operations udf_dir_inode_operations;
-extern struct file_operations udf_dir_operations;
+extern const struct file_operations udf_dir_operations;
 extern struct inode_operations udf_file_inode_operations;
-extern struct file_operations udf_file_operations;
+extern const struct file_operations udf_file_operations;
 extern struct address_space_operations udf_aops;
 extern struct address_space_operations udf_adinicb_aops;
 extern struct address_space_operations udf_symlink_aops;
diff --git a/fs/ufs/dir.c b/fs/ufs/dir.c
index 7c10c68902ae..1a561202d3f4 100644
--- a/fs/ufs/dir.c
+++ b/fs/ufs/dir.c
@@ -620,7 +620,7 @@ int ufs_empty_dir (struct inode * inode)
 	return 1;
 }
 
-struct file_operations ufs_dir_operations = {
+const struct file_operations ufs_dir_operations = {
 	.read		= generic_read_dir,
 	.readdir	= ufs_readdir,
 	.fsync		= file_fsync,
diff --git a/fs/ufs/file.c b/fs/ufs/file.c
index ed69d7fe1b5d..312fd3f86313 100644
--- a/fs/ufs/file.c
+++ b/fs/ufs/file.c
@@ -23,25 +23,15 @@
  *  ext2 fs regular file handling primitives
  */
 
-#include <asm/uaccess.h>
-#include <asm/system.h>
-
-#include <linux/errno.h>
 #include <linux/fs.h>
 #include <linux/ufs_fs.h>
-#include <linux/fcntl.h>
-#include <linux/time.h>
-#include <linux/stat.h>
-#include <linux/mm.h>
-#include <linux/pagemap.h>
-#include <linux/smp_lock.h>
 
 /*
  * We have mostly NULL's here: the current defaults are ok for
  * the ufs filesystem.
  */
  
-struct file_operations ufs_file_operations = {
+const struct file_operations ufs_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= generic_file_read,
 	.write		= generic_file_write,
diff --git a/fs/ufs/namei.c b/fs/ufs/namei.c
index 2958cde7d3d6..8d5f98a01c74 100644
--- a/fs/ufs/namei.c
+++ b/fs/ufs/namei.c
@@ -43,18 +43,6 @@
 #define UFSD(x)
 #endif
 
-static inline void ufs_inc_count(struct inode *inode)
-{
-	inode->i_nlink++;
-	mark_inode_dirty(inode);
-}
-
-static inline void ufs_dec_count(struct inode *inode)
-{
-	inode->i_nlink--;
-	mark_inode_dirty(inode);
-}
-
 static inline int ufs_add_nondir(struct dentry *dentry, struct inode *inode)
 {
 	int err = ufs_add_link(dentry, inode);
@@ -62,7 +50,7 @@ static inline int ufs_add_nondir(struct dentry *dentry, struct inode *inode)
 		d_instantiate(dentry, inode);
 		return 0;
 	}
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	return err;
 }
@@ -173,7 +161,7 @@ out:
 	return err;
 
 out_fail:
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
 	iput(inode);
 	goto out;
 }
@@ -191,7 +179,7 @@ static int ufs_link (struct dentry * old_dentry, struct inode * dir,
 	}
 
 	inode->i_ctime = CURRENT_TIME_SEC;
-	ufs_inc_count(inode);
+	inode_inc_link_count(inode);
 	atomic_inc(&inode->i_count);
 
 	error = ufs_add_nondir(dentry, inode);
@@ -208,7 +196,7 @@ static int ufs_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 		goto out;
 
 	lock_kernel();
-	ufs_inc_count(dir);
+	inode_inc_link_count(dir);
 
 	inode = ufs_new_inode(dir, S_IFDIR|mode);
 	err = PTR_ERR(inode);
@@ -218,7 +206,7 @@ static int ufs_mkdir(struct inode * dir, struct dentry * dentry, int mode)
 	inode->i_op = &ufs_dir_inode_operations;
 	inode->i_fop = &ufs_dir_operations;
 
-	ufs_inc_count(inode);
+	inode_inc_link_count(inode);
 
 	err = ufs_make_empty(inode, dir);
 	if (err)
@@ -234,11 +222,11 @@ out:
 	return err;
 
 out_fail:
-	ufs_dec_count(inode);
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
+	inode_dec_link_count(inode);
 	iput (inode);
 out_dir:
-	ufs_dec_count(dir);
+	inode_dec_link_count(dir);
 	unlock_kernel();
 	goto out;
 }
@@ -260,7 +248,7 @@ static int ufs_unlink(struct inode * dir, struct dentry *dentry)
 		goto out;
 
 	inode->i_ctime = dir->i_ctime;
-	ufs_dec_count(inode);
+	inode_dec_link_count(inode);
 	err = 0;
 out:
 	unlock_kernel();
@@ -277,8 +265,8 @@ static int ufs_rmdir (struct inode * dir, struct dentry *dentry)
 		err = ufs_unlink(dir, dentry);
 		if (!err) {
 			inode->i_size = 0;
-			ufs_dec_count(inode);
-			ufs_dec_count(dir);
+			inode_dec_link_count(inode);
+			inode_dec_link_count(dir);
 		}
 	}
 	unlock_kernel();
@@ -319,35 +307,35 @@ static int ufs_rename (struct inode * old_dir, struct dentry * old_dentry,
 		new_de = ufs_find_entry (new_dentry, &new_bh);
 		if (!new_de)
 			goto out_dir;
-		ufs_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		ufs_set_link(new_dir, new_de, new_bh, old_inode);
 		new_inode->i_ctime = CURRENT_TIME_SEC;
 		if (dir_de)
 			new_inode->i_nlink--;
-		ufs_dec_count(new_inode);
+		inode_dec_link_count(new_inode);
 	} else {
 		if (dir_de) {
 			err = -EMLINK;
 			if (new_dir->i_nlink >= UFS_LINK_MAX)
 				goto out_dir;
 		}
-		ufs_inc_count(old_inode);
+		inode_inc_link_count(old_inode);
 		err = ufs_add_link(new_dentry, old_inode);
 		if (err) {
-			ufs_dec_count(old_inode);
+			inode_dec_link_count(old_inode);
 			goto out_dir;
 		}
 		if (dir_de)
-			ufs_inc_count(new_dir);
+			inode_inc_link_count(new_dir);
 	}
 
 	ufs_delete_entry (old_dir, old_de, old_bh);
 
-	ufs_dec_count(old_inode);
+	inode_dec_link_count(old_inode);
 
 	if (dir_de) {
 		ufs_set_link(old_inode, dir_de, dir_bh, new_dir);
-		ufs_dec_count(old_dir);
+		inode_dec_link_count(old_dir);
 	}
 	unlock_kernel();
 	return 0;
diff --git a/fs/ufs/super.c b/fs/ufs/super.c
index e9055ef7f5ac..db98a4c71e63 100644
--- a/fs/ufs/super.c
+++ b/fs/ufs/super.c
@@ -575,7 +575,7 @@ static int ufs_fill_super(struct super_block *sb, void *data, int silent)
 		if (!silent)
 			printk("You didn't specify the type of your ufs filesystem\n\n"
 			"mount -t ufs -o ufstype="
-			"sun|sunx86|44bsd|ufs2|5xbsd|old|hp|nextstep|netxstep-cd|openstep ...\n\n"
+			"sun|sunx86|44bsd|ufs2|5xbsd|old|hp|nextstep|nextstep-cd|openstep ...\n\n"
 			">>>WARNING<<< Wrong ufstype may corrupt your filesystem, "
 			"default is ufstype=old\n");
 		ufs_set_opt (sbi->s_mount_opt, UFSTYPE_OLD);
@@ -1184,7 +1184,8 @@ static int init_inodecache(void)
 {
 	ufs_inode_cachep = kmem_cache_create("ufs_inode_cache",
 					     sizeof(struct ufs_inode_info),
-					     0, SLAB_RECLAIM_ACCOUNT,
+					     0, (SLAB_RECLAIM_ACCOUNT|
+						SLAB_MEM_SPREAD),
 					     init_once, NULL);
 	if (ufs_inode_cachep == NULL)
 		return -ENOMEM;
diff --git a/fs/xattr.c b/fs/xattr.c
index 80eca7d3d69f..e416190f5e9c 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
@@ -17,6 +17,7 @@
 #include <linux/syscalls.h>
 #include <linux/module.h>
 #include <linux/fsnotify.h>
+#include <linux/audit.h>
 #include <asm/uaccess.h>
 
 
@@ -234,12 +235,15 @@ sys_fsetxattr(int fd, char __user *name, void __user *value,
 	      size_t size, int flags)
 {
 	struct file *f;
+	struct dentry *dentry;
 	int error = -EBADF;
 
 	f = fget(fd);
 	if (!f)
 		return error;
-	error = setxattr(f->f_dentry, name, value, size, flags);
+	dentry = f->f_dentry;
+	audit_inode(NULL, dentry->d_inode, 0);
+	error = setxattr(dentry, name, value, size, flags);
 	fput(f);
 	return error;
 }
@@ -458,12 +462,15 @@ asmlinkage long
 sys_fremovexattr(int fd, char __user *name)
 {
 	struct file *f;
+	struct dentry *dentry;
 	int error = -EBADF;
 
 	f = fget(fd);
 	if (!f)
 		return error;
-	error = removexattr(f->f_dentry, name);
+	dentry = f->f_dentry;
+	audit_inode(NULL, dentry->d_inode, 0);
+	error = removexattr(dentry, name);
 	fput(f);
 	return error;
 }
diff --git a/fs/xfs/Makefile-linux-2.6 b/fs/xfs/Makefile-linux-2.6
index 97bd4743b461..5d73eaa1971f 100644
--- a/fs/xfs/Makefile-linux-2.6
+++ b/fs/xfs/Makefile-linux-2.6
@@ -1,33 +1,19 @@
 #
-# Copyright (c) 2000-2005 Silicon Graphics, Inc.  All Rights Reserved.
+# Copyright (c) 2000-2005 Silicon Graphics, Inc.
+# All Rights Reserved.
 #
-# This program is free software; you can redistribute it and/or modify it
-# under the terms of version 2 of the GNU General Public License as
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License as
 # published by the Free Software Foundation.
 #
-# This program is distributed in the hope that it would be useful, but
-# WITHOUT ANY WARRANTY; without even the implied warranty of
-# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+# This program is distributed in the hope that it would be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
 #
-# Further, this software is distributed without any warranty that it is
-# free of the rightful claim of any third person regarding infringement
-# or the like.  Any license provided herein, whether implied or
-# otherwise, applies only to this software file.  Patent licenses, if
-# any, provided herein do not apply to combinations of this program with
-# other software, or any other product whatsoever.
-#
-# You should have received a copy of the GNU General Public License along
-# with this program; if not, write the Free Software Foundation, Inc., 59
-# Temple Place - Suite 330, Boston MA 02111-1307, USA.
-#
-# Contact information: Silicon Graphics, Inc., 1600 Amphitheatre Pkwy,
-# Mountain View, CA  94043, or:
-#
-# http://www.sgi.com
-#
-# For further information regarding this notice, see:
-#
-# http://oss.sgi.com/projects/GenInfo/SGIGPLNoticeExplan/
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write the Free Software Foundation,
+# Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #
 
 EXTRA_CFLAGS +=	 -Ifs/xfs -Ifs/xfs/linux-2.6 -funsigned-char
@@ -36,7 +22,7 @@ XFS_LINUX := linux-2.6
 
 ifeq ($(CONFIG_XFS_DEBUG),y)
 	EXTRA_CFLAGS += -g -DSTATIC="" -DDEBUG
-	EXTRA_CFLAGS += -DPAGEBUF_LOCK_TRACKING
+	EXTRA_CFLAGS += -DXFS_BUF_LOCK_TRACKING
 endif
 ifeq ($(CONFIG_XFS_TRACE),y)
 	EXTRA_CFLAGS += -DXFS_ALLOC_TRACE
@@ -50,7 +36,7 @@ ifeq ($(CONFIG_XFS_TRACE),y)
 	EXTRA_CFLAGS += -DXFS_ILOCK_TRACE
 	EXTRA_CFLAGS += -DXFS_LOG_TRACE
 	EXTRA_CFLAGS += -DXFS_RW_TRACE
-	EXTRA_CFLAGS += -DPAGEBUF_TRACE
+	EXTRA_CFLAGS += -DXFS_BUF_TRACE
 	EXTRA_CFLAGS += -DXFS_VNODE_TRACE
 endif
 
diff --git a/fs/xfs/linux-2.6/kmem.h b/fs/xfs/linux-2.6/kmem.h
index c64a29cdfff3..2cfd33d4d8aa 100644
--- a/fs/xfs/linux-2.6/kmem.h
+++ b/fs/xfs/linux-2.6/kmem.h
@@ -23,17 +23,8 @@
 #include <linux/mm.h>
 
 /*
- * memory management routines
+ * Process flags handling
  */
-#define KM_SLEEP	0x0001u
-#define KM_NOSLEEP	0x0002u
-#define KM_NOFS		0x0004u
-#define KM_MAYFAIL	0x0008u
-
-#define	kmem_zone	kmem_cache
-#define kmem_zone_t	struct kmem_cache
-
-typedef unsigned long xfs_pflags_t;
 
 #define PFLAGS_TEST_NOIO()              (current->flags & PF_NOIO)
 #define PFLAGS_TEST_FSTRANS()           (current->flags & PF_FSTRANS)
@@ -67,74 +58,102 @@ typedef unsigned long xfs_pflags_t;
 	*(NSTATEP) = *(OSTATEP);	\
 } while (0)
 
-static __inline gfp_t kmem_flags_convert(unsigned int __nocast flags)
+/*
+ * General memory allocation interfaces
+ */
+
+#define KM_SLEEP	0x0001u
+#define KM_NOSLEEP	0x0002u
+#define KM_NOFS		0x0004u
+#define KM_MAYFAIL	0x0008u
+
+/*
+ * We use a special process flag to avoid recursive callbacks into
+ * the filesystem during transactions.  We will also issue our own
+ * warnings, so we explicitly skip any generic ones (silly of us).
+ */
+static inline gfp_t
+kmem_flags_convert(unsigned int __nocast flags)
 {
-	gfp_t	lflags = __GFP_NOWARN;	/* we'll report problems, if need be */
+	gfp_t	lflags;
 
-#ifdef DEBUG
-	if (unlikely(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL))) {
-		printk(KERN_WARNING
-		    "XFS: memory allocation with wrong flags (%x)\n", flags);
-		BUG();
-	}
-#endif
+	BUG_ON(flags & ~(KM_SLEEP|KM_NOSLEEP|KM_NOFS|KM_MAYFAIL));
 
 	if (flags & KM_NOSLEEP) {
-		lflags |= GFP_ATOMIC;
+		lflags = GFP_ATOMIC | __GFP_NOWARN;
 	} else {
-		lflags |= GFP_KERNEL;
-
-		/* avoid recusive callbacks to filesystem during transactions */
+		lflags = GFP_KERNEL | __GFP_NOWARN;
 		if (PFLAGS_TEST_FSTRANS() || (flags & KM_NOFS))
 			lflags &= ~__GFP_FS;
 	}
-
-        return lflags;
+	return lflags;
 }
 
-static __inline kmem_zone_t *
+extern void *kmem_alloc(size_t, unsigned int __nocast);
+extern void *kmem_realloc(void *, size_t, size_t, unsigned int __nocast);
+extern void *kmem_zalloc(size_t, unsigned int __nocast);
+extern void  kmem_free(void *, size_t);
+
+/*
+ * Zone interfaces
+ */
+
+#define KM_ZONE_HWALIGN	SLAB_HWCACHE_ALIGN
+#define KM_ZONE_RECLAIM	SLAB_RECLAIM_ACCOUNT
+#define KM_ZONE_SPREAD	SLAB_MEM_SPREAD
+
+#define kmem_zone	kmem_cache
+#define kmem_zone_t	struct kmem_cache
+
+static inline kmem_zone_t *
 kmem_zone_init(int size, char *zone_name)
 {
 	return kmem_cache_create(zone_name, size, 0, 0, NULL, NULL);
 }
 
-static __inline void
+static inline kmem_zone_t *
+kmem_zone_init_flags(int size, char *zone_name, unsigned long flags,
+		     void (*construct)(void *, kmem_zone_t *, unsigned long))
+{
+	return kmem_cache_create(zone_name, size, 0, flags, construct, NULL);
+}
+
+static inline void
 kmem_zone_free(kmem_zone_t *zone, void *ptr)
 {
 	kmem_cache_free(zone, ptr);
 }
 
-static __inline void
+static inline void
 kmem_zone_destroy(kmem_zone_t *zone)
 {
 	if (zone && kmem_cache_destroy(zone))
 		BUG();
 }
 
-extern void *kmem_zone_zalloc(kmem_zone_t *, unsigned int __nocast);
 extern void *kmem_zone_alloc(kmem_zone_t *, unsigned int __nocast);
+extern void *kmem_zone_zalloc(kmem_zone_t *, unsigned int __nocast);
 
-extern void *kmem_alloc(size_t, unsigned int __nocast);
-extern void *kmem_realloc(void *, size_t, size_t, unsigned int __nocast);
-extern void *kmem_zalloc(size_t, unsigned int __nocast);
-extern void  kmem_free(void *, size_t);
+/*
+ * Low memory cache shrinkers
+ */
 
 typedef struct shrinker *kmem_shaker_t;
 typedef int (*kmem_shake_func_t)(int, gfp_t);
 
-static __inline kmem_shaker_t
+static inline kmem_shaker_t
 kmem_shake_register(kmem_shake_func_t sfunc)
 {
 	return set_shrinker(DEFAULT_SEEKS, sfunc);
 }
 
-static __inline void
+static inline void
 kmem_shake_deregister(kmem_shaker_t shrinker)
 {
 	remove_shrinker(shrinker);
 }
 
-static __inline int
+static inline int
 kmem_shake_allow(gfp_t gfp_mask)
 {
 	return (gfp_mask & __GFP_WAIT);
diff --git a/fs/xfs/linux-2.6/mrlock.h b/fs/xfs/linux-2.6/mrlock.h
index 16b44c3c2362..1b262b790d9c 100644
--- a/fs/xfs/linux-2.6/mrlock.h
+++ b/fs/xfs/linux-2.6/mrlock.h
@@ -79,7 +79,7 @@ static inline void mrdemote(mrlock_t *mrp)
  * Debug-only routine, without some platform-specific asm code, we can
  * now only answer requests regarding whether we hold the lock for write
  * (reader state is outside our visibility, we only track writer state).
- * Note: means !ismrlocked would give false positivies, so don't do that.
+ * Note: means !ismrlocked would give false positives, so don't do that.
  */
 static inline int ismrlocked(mrlock_t *mrp, int type)
 {
diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c
index 74d8be87f983..6cbbd165c60d 100644
--- a/fs/xfs/linux-2.6/xfs_aops.c
+++ b/fs/xfs/linux-2.6/xfs_aops.c
@@ -43,7 +43,29 @@
 #include <linux/pagevec.h>
 #include <linux/writeback.h>
 
-STATIC void xfs_count_page_state(struct page *, int *, int *, int *);
+STATIC void
+xfs_count_page_state(
+	struct page		*page,
+	int			*delalloc,
+	int			*unmapped,
+	int			*unwritten)
+{
+	struct buffer_head	*bh, *head;
+
+	*delalloc = *unmapped = *unwritten = 0;
+
+	bh = head = page_buffers(page);
+	do {
+		if (buffer_uptodate(bh) && !buffer_mapped(bh))
+			(*unmapped) = 1;
+		else if (buffer_unwritten(bh) && !buffer_delay(bh))
+			clear_buffer_unwritten(bh);
+		else if (buffer_unwritten(bh))
+			(*unwritten) = 1;
+		else if (buffer_delay(bh))
+			(*delalloc) = 1;
+	} while ((bh = bh->b_this_page) != head);
+}
 
 #if defined(XFS_RW_TRACE)
 void
@@ -54,7 +76,7 @@ xfs_page_trace(
 	int		mask)
 {
 	xfs_inode_t	*ip;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	loff_t		isize = i_size_read(inode);
 	loff_t		offset = page_offset(page);
 	int		delalloc = -1, unmapped = -1, unwritten = -1;
@@ -81,7 +103,7 @@ xfs_page_trace(
 		(void *)((unsigned long)delalloc),
 		(void *)((unsigned long)unmapped),
 		(void *)((unsigned long)unwritten),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL);
 }
 #else
@@ -192,7 +214,7 @@ xfs_alloc_ioend(
 	ioend->io_uptodate = 1; /* cleared if any I/O fails */
 	ioend->io_list = NULL;
 	ioend->io_type = type;
-	ioend->io_vnode = LINVFS_GET_VP(inode);
+	ioend->io_vnode = vn_from_inode(inode);
 	ioend->io_buffer_head = NULL;
 	ioend->io_buffer_tail = NULL;
 	atomic_inc(&ioend->io_vnode->v_iocount);
@@ -217,7 +239,7 @@ xfs_map_blocks(
 	xfs_iomap_t		*mapp,
 	int			flags)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error, nmaps = 1;
 
 	VOP_BMAP(vp, offset, count, flags, mapp, &nmaps, error);
@@ -350,7 +372,7 @@ static inline int bio_add_buffer(struct bio *bio, struct buffer_head *bh)
  * assumes that all buffers on the page are started at the same time.
  *
  * The fix is two passes across the ioend list - one to start writeback on the
- * bufferheads, and then the second one submit them for I/O.
+ * buffer_heads, and then submit them for I/O on the second pass.
  */
 STATIC void
 xfs_submit_ioend(
@@ -462,28 +484,37 @@ xfs_add_to_ioend(
 }
 
 STATIC void
+xfs_map_buffer(
+	struct buffer_head	*bh,
+	xfs_iomap_t		*mp,
+	xfs_off_t		offset,
+	uint			block_bits)
+{
+	sector_t		bn;
+
+	ASSERT(mp->iomap_bn != IOMAP_DADDR_NULL);
+
+	bn = (mp->iomap_bn >> (block_bits - BBSHIFT)) +
+	      ((offset - mp->iomap_offset) >> block_bits);
+
+	ASSERT(bn || (mp->iomap_flags & IOMAP_REALTIME));
+
+	bh->b_blocknr = bn;
+	set_buffer_mapped(bh);
+}
+
+STATIC void
 xfs_map_at_offset(
 	struct buffer_head	*bh,
 	loff_t			offset,
 	int			block_bits,
 	xfs_iomap_t		*iomapp)
 {
-	xfs_daddr_t		bn;
-	int			sector_shift;
-
 	ASSERT(!(iomapp->iomap_flags & IOMAP_HOLE));
 	ASSERT(!(iomapp->iomap_flags & IOMAP_DELAY));
-	ASSERT(iomapp->iomap_bn != IOMAP_DADDR_NULL);
-
-	sector_shift = block_bits - BBSHIFT;
-	bn = (iomapp->iomap_bn >> sector_shift) +
-	      ((offset - iomapp->iomap_offset) >> block_bits);
-
-	ASSERT(bn || (iomapp->iomap_flags & IOMAP_REALTIME));
-	ASSERT((bn << sector_shift) >= iomapp->iomap_bn);
 
 	lock_buffer(bh);
-	bh->b_blocknr = bn;
+	xfs_map_buffer(bh, iomapp, offset, block_bits);
 	bh->b_bdev = iomapp->iomap_target->bt_bdev;
 	set_buffer_mapped(bh);
 	clear_buffer_delay(bh);
@@ -616,7 +647,7 @@ xfs_is_delayed_page(
 				acceptable = (type == IOMAP_UNWRITTEN);
 			else if (buffer_delay(bh))
 				acceptable = (type == IOMAP_DELAY);
-			else if (buffer_mapped(bh))
+			else if (buffer_dirty(bh) && buffer_mapped(bh))
 				acceptable = (type == 0);
 			else
 				break;
@@ -668,7 +699,7 @@ xfs_convert_page(
 
 	/*
 	 * page_dirty is initially a count of buffers on the page before
-	 * EOF and is decrememted as we move each into a cleanable state.
+	 * EOF and is decremented as we move each into a cleanable state.
 	 *
 	 * Derivation:
 	 *
@@ -811,7 +842,7 @@ xfs_cluster_write(
  * page if possible.
  * The bh->b_state's cannot know if any of the blocks or which block for
  * that matter are dirty due to mmap writes, and therefore bh uptodate is
- * only vaild if the page itself isn't completely uptodate.  Some layers
+ * only valid if the page itself isn't completely uptodate.  Some layers
  * may clear the page dirty flag prior to calling write page, under the
  * assumption the entire page will be written out; by not writing out the
  * whole page the page can be reused before all valid dirty data is
@@ -861,7 +892,7 @@ xfs_page_state_convert(
 
 	/*
 	 * page_dirty is initially a count of buffers on the page before
-	 * EOF and is decrememted as we move each into a cleanable state.
+	 * EOF and is decremented as we move each into a cleanable state.
 	 *
 	 * Derivation:
 	 *
@@ -1040,54 +1071,191 @@ error:
 	return err;
 }
 
+/*
+ * writepage: Called from one of two places:
+ *
+ * 1. we are flushing a delalloc buffer head.
+ *
+ * 2. we are writing out a dirty page. Typically the page dirty
+ *    state is cleared before we get here. In this case is it
+ *    conceivable we have no buffer heads.
+ *
+ * For delalloc space on the page we need to allocate space and
+ * flush it. For unmapped buffer heads on the page we should
+ * allocate space if the page is uptodate. For any other dirty
+ * buffer heads on the page we should flush them.
+ *
+ * If we detect that a transaction would be required to flush
+ * the page, we have to check the process flags first, if we
+ * are already in a transaction or disk I/O during allocations
+ * is off, we need to fail the writepage and redirty the page.
+ */
+
 STATIC int
-__linvfs_get_block(
+xfs_vm_writepage(
+	struct page		*page,
+	struct writeback_control *wbc)
+{
+	int			error;
+	int			need_trans;
+	int			delalloc, unmapped, unwritten;
+	struct inode		*inode = page->mapping->host;
+
+	xfs_page_trace(XFS_WRITEPAGE_ENTER, inode, page, 0);
+
+	/*
+	 * We need a transaction if:
+	 *  1. There are delalloc buffers on the page
+	 *  2. The page is uptodate and we have unmapped buffers
+	 *  3. The page is uptodate and we have no buffers
+	 *  4. There are unwritten buffers on the page
+	 */
+
+	if (!page_has_buffers(page)) {
+		unmapped = 1;
+		need_trans = 1;
+	} else {
+		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
+		if (!PageUptodate(page))
+			unmapped = 0;
+		need_trans = delalloc + unmapped + unwritten;
+	}
+
+	/*
+	 * If we need a transaction and the process flags say
+	 * we are already in a transaction, or no IO is allowed
+	 * then mark the page dirty again and leave the page
+	 * as is.
+	 */
+	if (PFLAGS_TEST_FSTRANS() && need_trans)
+		goto out_fail;
+
+	/*
+	 * Delay hooking up buffer heads until we have
+	 * made our go/no-go decision.
+	 */
+	if (!page_has_buffers(page))
+		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
+
+	/*
+	 * Convert delayed allocate, unwritten or unmapped space
+	 * to real space and flush out to disk.
+	 */
+	error = xfs_page_state_convert(inode, page, wbc, 1, unmapped);
+	if (error == -EAGAIN)
+		goto out_fail;
+	if (unlikely(error < 0))
+		goto out_unlock;
+
+	return 0;
+
+out_fail:
+	redirty_page_for_writepage(wbc, page);
+	unlock_page(page);
+	return 0;
+out_unlock:
+	unlock_page(page);
+	return error;
+}
+
+/*
+ * Called to move a page into cleanable state - and from there
+ * to be released. Possibly the page is already clean. We always
+ * have buffer heads in this call.
+ *
+ * Returns 0 if the page is ok to release, 1 otherwise.
+ *
+ * Possible scenarios are:
+ *
+ * 1. We are being called to release a page which has been written
+ *    to via regular I/O. buffer heads will be dirty and possibly
+ *    delalloc. If no delalloc buffer heads in this case then we
+ *    can just return zero.
+ *
+ * 2. We are called to release a page which has been written via
+ *    mmap, all we need to do is ensure there is no delalloc
+ *    state in the buffer heads, if not we can let the caller
+ *    free them and we should come back later via writepage.
+ */
+STATIC int
+xfs_vm_releasepage(
+	struct page		*page,
+	gfp_t			gfp_mask)
+{
+	struct inode		*inode = page->mapping->host;
+	int			dirty, delalloc, unmapped, unwritten;
+	struct writeback_control wbc = {
+		.sync_mode = WB_SYNC_ALL,
+		.nr_to_write = 1,
+	};
+
+	xfs_page_trace(XFS_RELEASEPAGE_ENTER, inode, page, gfp_mask);
+
+	if (!page_has_buffers(page))
+		return 0;
+
+	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
+	if (!delalloc && !unwritten)
+		goto free_buffers;
+
+	if (!(gfp_mask & __GFP_FS))
+		return 0;
+
+	/* If we are already inside a transaction or the thread cannot
+	 * do I/O, we cannot release this page.
+	 */
+	if (PFLAGS_TEST_FSTRANS())
+		return 0;
+
+	/*
+	 * Convert delalloc space to real space, do not flush the
+	 * data out to disk, that will be done by the caller.
+	 * Never need to allocate space here - we will always
+	 * come back to writepage in that case.
+	 */
+	dirty = xfs_page_state_convert(inode, page, &wbc, 0, 0);
+	if (dirty == 0 && !unwritten)
+		goto free_buffers;
+	return 0;
+
+free_buffers:
+	return try_to_free_buffers(page);
+}
+
+STATIC int
+__xfs_get_blocks(
 	struct inode		*inode,
 	sector_t		iblock,
-	unsigned long		blocks,
 	struct buffer_head	*bh_result,
 	int			create,
 	int			direct,
 	bmapi_flags_t		flags)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	xfs_iomap_t		iomap;
 	xfs_off_t		offset;
 	ssize_t			size;
-	int			retpbbm = 1;
+	int			niomap = 1;
 	int			error;
 
 	offset = (xfs_off_t)iblock << inode->i_blkbits;
-	if (blocks)
-		size = (ssize_t) min_t(xfs_off_t, LONG_MAX,
-					(xfs_off_t)blocks << inode->i_blkbits);
-	else
-		size = 1 << inode->i_blkbits;
-
+	ASSERT(bh_result->b_size >= (1 << inode->i_blkbits));
+	size = bh_result->b_size;
 	VOP_BMAP(vp, offset, size,
-		create ? flags : BMAPI_READ, &iomap, &retpbbm, error);
+		create ? flags : BMAPI_READ, &iomap, &niomap, error);
 	if (error)
 		return -error;
-
-	if (retpbbm == 0)
+	if (niomap == 0)
 		return 0;
 
 	if (iomap.iomap_bn != IOMAP_DADDR_NULL) {
-		xfs_daddr_t	bn;
-		xfs_off_t	delta;
-
-		/* For unwritten extents do not report a disk address on
+		/*
+		 * For unwritten extents do not report a disk address on
 		 * the read case (treat as if we're reading into a hole).
 		 */
 		if (create || !(iomap.iomap_flags & IOMAP_UNWRITTEN)) {
-			delta = offset - iomap.iomap_offset;
-			delta >>= inode->i_blkbits;
-
-			bn = iomap.iomap_bn >> (inode->i_blkbits - BBSHIFT);
-			bn += delta;
-			BUG_ON(!bn && !(iomap.iomap_flags & IOMAP_REALTIME));
-			bh_result->b_blocknr = bn;
-			set_buffer_mapped(bh_result);
+			xfs_map_buffer(bh_result, &iomap, offset,
+				       inode->i_blkbits);
 		}
 		if (create && (iomap.iomap_flags & IOMAP_UNWRITTEN)) {
 			if (direct)
@@ -1097,12 +1265,16 @@ __linvfs_get_block(
 		}
 	}
 
-	/* If this is a realtime file, data might be on a new device */
+	/*
+	 * If this is a realtime file, data may be on a different device.
+	 * to that pointed to from the buffer_head b_bdev currently.
+	 */
 	bh_result->b_bdev = iomap.iomap_target->bt_bdev;
 
-	/* If we previously allocated a block out beyond eof and
-	 * we are now coming back to use it then we will need to
-	 * flag it as new even if it has a disk address.
+	/*
+	 * If we previously allocated a block out beyond eof and we are
+	 * now coming back to use it then we will need to flag it as new
+	 * even if it has a disk address.
 	 */
 	if (create &&
 	    ((!buffer_mapped(bh_result) && !buffer_uptodate(bh_result)) ||
@@ -1118,42 +1290,40 @@ __linvfs_get_block(
 		}
 	}
 
-	if (blocks) {
+	if (direct || size > (1 << inode->i_blkbits)) {
 		ASSERT(iomap.iomap_bsize - iomap.iomap_delta > 0);
 		offset = min_t(xfs_off_t,
-				iomap.iomap_bsize - iomap.iomap_delta,
-				(xfs_off_t)blocks << inode->i_blkbits);
-		bh_result->b_size = (u32) min_t(xfs_off_t, UINT_MAX, offset);
+				iomap.iomap_bsize - iomap.iomap_delta, size);
+		bh_result->b_size = (ssize_t)min_t(xfs_off_t, LONG_MAX, offset);
 	}
 
 	return 0;
 }
 
 int
-linvfs_get_block(
+xfs_get_blocks(
 	struct inode		*inode,
 	sector_t		iblock,
 	struct buffer_head	*bh_result,
 	int			create)
 {
-	return __linvfs_get_block(inode, iblock, 0, bh_result,
-					create, 0, BMAPI_WRITE);
+	return __xfs_get_blocks(inode, iblock,
+				bh_result, create, 0, BMAPI_WRITE);
 }
 
 STATIC int
-linvfs_get_blocks_direct(
+xfs_get_blocks_direct(
 	struct inode		*inode,
 	sector_t		iblock,
-	unsigned long		max_blocks,
 	struct buffer_head	*bh_result,
 	int			create)
 {
-	return __linvfs_get_block(inode, iblock, max_blocks, bh_result,
-					create, 1, BMAPI_WRITE|BMAPI_DIRECT);
+	return __xfs_get_blocks(inode, iblock,
+				bh_result, create, 1, BMAPI_WRITE|BMAPI_DIRECT);
 }
 
 STATIC void
-linvfs_end_io_direct(
+xfs_end_io_direct(
 	struct kiocb	*iocb,
 	loff_t		offset,
 	ssize_t		size,
@@ -1164,9 +1334,9 @@ linvfs_end_io_direct(
 	/*
 	 * Non-NULL private data means we need to issue a transaction to
 	 * convert a range from unwritten to written extents.  This needs
-	 * to happen from process contect but aio+dio I/O completion
+	 * to happen from process context but aio+dio I/O completion
 	 * happens from irq context so we need to defer it to a workqueue.
-	 * This is not nessecary for synchronous direct I/O, but we do
+	 * This is not necessary for synchronous direct I/O, but we do
 	 * it anyway to keep the code uniform and simpler.
 	 *
 	 * The core direct I/O code might be changed to always call the
@@ -1183,7 +1353,7 @@ linvfs_end_io_direct(
 	}
 
 	/*
-	 * blockdev_direct_IO can return an error even afer the I/O
+	 * blockdev_direct_IO can return an error even after the I/O
 	 * completion handler was called.  Thus we need to protect
 	 * against double-freeing.
 	 */
@@ -1191,7 +1361,7 @@ linvfs_end_io_direct(
 }
 
 STATIC ssize_t
-linvfs_direct_IO(
+xfs_vm_direct_IO(
 	int			rw,
 	struct kiocb		*iocb,
 	const struct iovec	*iov,
@@ -1200,7 +1370,7 @@ linvfs_direct_IO(
 {
 	struct file	*file = iocb->ki_filp;
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	xfs_iomap_t	iomap;
 	int		maps = 1;
 	int		error;
@@ -1215,253 +1385,79 @@ linvfs_direct_IO(
 	ret = blockdev_direct_IO_own_locking(rw, iocb, inode,
 		iomap.iomap_target->bt_bdev,
 		iov, offset, nr_segs,
-		linvfs_get_blocks_direct,
-		linvfs_end_io_direct);
+		xfs_get_blocks_direct,
+		xfs_end_io_direct);
 
 	if (unlikely(ret <= 0 && iocb->private))
 		xfs_destroy_ioend(iocb->private);
 	return ret;
 }
 
+STATIC int
+xfs_vm_prepare_write(
+	struct file		*file,
+	struct page		*page,
+	unsigned int		from,
+	unsigned int		to)
+{
+	return block_prepare_write(page, from, to, xfs_get_blocks);
+}
 
 STATIC sector_t
-linvfs_bmap(
+xfs_vm_bmap(
 	struct address_space	*mapping,
 	sector_t		block)
 {
 	struct inode		*inode = (struct inode *)mapping->host;
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error;
 
-	vn_trace_entry(vp, "linvfs_bmap", (inst_t *)__return_address);
+	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
 
 	VOP_RWLOCK(vp, VRWLOCK_READ);
 	VOP_FLUSH_PAGES(vp, (xfs_off_t)0, -1, 0, FI_REMAPF, error);
 	VOP_RWUNLOCK(vp, VRWLOCK_READ);
-	return generic_block_bmap(mapping, block, linvfs_get_block);
+	return generic_block_bmap(mapping, block, xfs_get_blocks);
 }
 
 STATIC int
-linvfs_readpage(
+xfs_vm_readpage(
 	struct file		*unused,
 	struct page		*page)
 {
-	return mpage_readpage(page, linvfs_get_block);
+	return mpage_readpage(page, xfs_get_blocks);
 }
 
 STATIC int
-linvfs_readpages(
+xfs_vm_readpages(
 	struct file		*unused,
 	struct address_space	*mapping,
 	struct list_head	*pages,
 	unsigned		nr_pages)
 {
-	return mpage_readpages(mapping, pages, nr_pages, linvfs_get_block);
+	return mpage_readpages(mapping, pages, nr_pages, xfs_get_blocks);
 }
 
 STATIC void
-xfs_count_page_state(
-	struct page		*page,
-	int			*delalloc,
-	int			*unmapped,
-	int			*unwritten)
-{
-	struct buffer_head	*bh, *head;
-
-	*delalloc = *unmapped = *unwritten = 0;
-
-	bh = head = page_buffers(page);
-	do {
-		if (buffer_uptodate(bh) && !buffer_mapped(bh))
-			(*unmapped) = 1;
-		else if (buffer_unwritten(bh) && !buffer_delay(bh))
-			clear_buffer_unwritten(bh);
-		else if (buffer_unwritten(bh))
-			(*unwritten) = 1;
-		else if (buffer_delay(bh))
-			(*delalloc) = 1;
-	} while ((bh = bh->b_this_page) != head);
-}
-
-
-/*
- * writepage: Called from one of two places:
- *
- * 1. we are flushing a delalloc buffer head.
- *
- * 2. we are writing out a dirty page. Typically the page dirty
- *    state is cleared before we get here. In this case is it
- *    conceivable we have no buffer heads.
- *
- * For delalloc space on the page we need to allocate space and
- * flush it. For unmapped buffer heads on the page we should
- * allocate space if the page is uptodate. For any other dirty
- * buffer heads on the page we should flush them.
- *
- * If we detect that a transaction would be required to flush
- * the page, we have to check the process flags first, if we
- * are already in a transaction or disk I/O during allocations
- * is off, we need to fail the writepage and redirty the page.
- */
-
-STATIC int
-linvfs_writepage(
-	struct page		*page,
-	struct writeback_control *wbc)
-{
-	int			error;
-	int			need_trans;
-	int			delalloc, unmapped, unwritten;
-	struct inode		*inode = page->mapping->host;
-
-	xfs_page_trace(XFS_WRITEPAGE_ENTER, inode, page, 0);
-
-	/*
-	 * We need a transaction if:
-	 *  1. There are delalloc buffers on the page
-	 *  2. The page is uptodate and we have unmapped buffers
-	 *  3. The page is uptodate and we have no buffers
-	 *  4. There are unwritten buffers on the page
-	 */
-
-	if (!page_has_buffers(page)) {
-		unmapped = 1;
-		need_trans = 1;
-	} else {
-		xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-		if (!PageUptodate(page))
-			unmapped = 0;
-		need_trans = delalloc + unmapped + unwritten;
-	}
-
-	/*
-	 * If we need a transaction and the process flags say
-	 * we are already in a transaction, or no IO is allowed
-	 * then mark the page dirty again and leave the page
-	 * as is.
-	 */
-	if (PFLAGS_TEST_FSTRANS() && need_trans)
-		goto out_fail;
-
-	/*
-	 * Delay hooking up buffer heads until we have
-	 * made our go/no-go decision.
-	 */
-	if (!page_has_buffers(page))
-		create_empty_buffers(page, 1 << inode->i_blkbits, 0);
-
-	/*
-	 * Convert delayed allocate, unwritten or unmapped space
-	 * to real space and flush out to disk.
-	 */
-	error = xfs_page_state_convert(inode, page, wbc, 1, unmapped);
-	if (error == -EAGAIN)
-		goto out_fail;
-	if (unlikely(error < 0))
-		goto out_unlock;
-
-	return 0;
-
-out_fail:
-	redirty_page_for_writepage(wbc, page);
-	unlock_page(page);
-	return 0;
-out_unlock:
-	unlock_page(page);
-	return error;
-}
-
-STATIC int
-linvfs_invalidate_page(
+xfs_vm_invalidatepage(
 	struct page		*page,
 	unsigned long		offset)
 {
 	xfs_page_trace(XFS_INVALIDPAGE_ENTER,
 			page->mapping->host, page, offset);
-	return block_invalidatepage(page, offset);
-}
-
-/*
- * Called to move a page into cleanable state - and from there
- * to be released. Possibly the page is already clean. We always
- * have buffer heads in this call.
- *
- * Returns 0 if the page is ok to release, 1 otherwise.
- *
- * Possible scenarios are:
- *
- * 1. We are being called to release a page which has been written
- *    to via regular I/O. buffer heads will be dirty and possibly
- *    delalloc. If no delalloc buffer heads in this case then we
- *    can just return zero.
- *
- * 2. We are called to release a page which has been written via
- *    mmap, all we need to do is ensure there is no delalloc
- *    state in the buffer heads, if not we can let the caller
- *    free them and we should come back later via writepage.
- */
-STATIC int
-linvfs_release_page(
-	struct page		*page,
-	gfp_t			gfp_mask)
-{
-	struct inode		*inode = page->mapping->host;
-	int			dirty, delalloc, unmapped, unwritten;
-	struct writeback_control wbc = {
-		.sync_mode = WB_SYNC_ALL,
-		.nr_to_write = 1,
-	};
-
-	xfs_page_trace(XFS_RELEASEPAGE_ENTER, inode, page, gfp_mask);
-
-	xfs_count_page_state(page, &delalloc, &unmapped, &unwritten);
-	if (!delalloc && !unwritten)
-		goto free_buffers;
-
-	if (!(gfp_mask & __GFP_FS))
-		return 0;
-
-	/* If we are already inside a transaction or the thread cannot
-	 * do I/O, we cannot release this page.
-	 */
-	if (PFLAGS_TEST_FSTRANS())
-		return 0;
-
-	/*
-	 * Convert delalloc space to real space, do not flush the
-	 * data out to disk, that will be done by the caller.
-	 * Never need to allocate space here - we will always
-	 * come back to writepage in that case.
-	 */
-	dirty = xfs_page_state_convert(inode, page, &wbc, 0, 0);
-	if (dirty == 0 && !unwritten)
-		goto free_buffers;
-	return 0;
-
-free_buffers:
-	return try_to_free_buffers(page);
-}
-
-STATIC int
-linvfs_prepare_write(
-	struct file		*file,
-	struct page		*page,
-	unsigned int		from,
-	unsigned int		to)
-{
-	return block_prepare_write(page, from, to, linvfs_get_block);
+	block_invalidatepage(page, offset);
 }
 
-struct address_space_operations linvfs_aops = {
-	.readpage		= linvfs_readpage,
-	.readpages		= linvfs_readpages,
-	.writepage		= linvfs_writepage,
+struct address_space_operations xfs_address_space_operations = {
+	.readpage		= xfs_vm_readpage,
+	.readpages		= xfs_vm_readpages,
+	.writepage		= xfs_vm_writepage,
 	.sync_page		= block_sync_page,
-	.releasepage		= linvfs_release_page,
-	.invalidatepage		= linvfs_invalidate_page,
-	.prepare_write		= linvfs_prepare_write,
+	.releasepage		= xfs_vm_releasepage,
+	.invalidatepage		= xfs_vm_invalidatepage,
+	.prepare_write		= xfs_vm_prepare_write,
 	.commit_write		= generic_commit_write,
-	.bmap			= linvfs_bmap,
-	.direct_IO		= linvfs_direct_IO,
+	.bmap			= xfs_vm_bmap,
+	.direct_IO		= xfs_vm_direct_IO,
 	.migratepage		= buffer_migrate_page,
 };
diff --git a/fs/xfs/linux-2.6/xfs_aops.h b/fs/xfs/linux-2.6/xfs_aops.h
index 55339dd5a30d..60716543c68b 100644
--- a/fs/xfs/linux-2.6/xfs_aops.h
+++ b/fs/xfs/linux-2.6/xfs_aops.h
@@ -40,7 +40,7 @@ typedef struct xfs_ioend {
 	struct work_struct	io_work;	/* xfsdatad work queue */
 } xfs_ioend_t;
 
-extern struct address_space_operations linvfs_aops;
-extern int linvfs_get_block(struct inode *, sector_t, struct buffer_head *, int);
+extern struct address_space_operations xfs_address_space_operations;
+extern int xfs_get_blocks(struct inode *, sector_t, struct buffer_head *, int);
 
 #endif /* __XFS_IOPS_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c
index bfb4f2917bb6..9fb0312665ca 100644
--- a/fs/xfs/linux-2.6/xfs_buf.c
+++ b/fs/xfs/linux-2.6/xfs_buf.c
@@ -29,6 +29,7 @@
 #include <linux/blkdev.h>
 #include <linux/hash.h>
 #include <linux/kthread.h>
+#include <linux/migrate.h>
 #include "xfs_linux.h"
 
 STATIC kmem_zone_t *xfs_buf_zone;
@@ -1805,13 +1806,12 @@ xfs_flush_buftarg(
 int __init
 xfs_buf_init(void)
 {
-	int		error = -ENOMEM;
-
 #ifdef XFS_BUF_TRACE
 	xfs_buf_trace_buf = ktrace_alloc(XFS_BUF_TRACE_SIZE, KM_SLEEP);
 #endif
 
-	xfs_buf_zone = kmem_zone_init(sizeof(xfs_buf_t), "xfs_buf");
+	xfs_buf_zone = kmem_zone_init_flags(sizeof(xfs_buf_t), "xfs_buf",
+						KM_ZONE_HWALIGN, NULL);
 	if (!xfs_buf_zone)
 		goto out_free_trace_buf;
 
@@ -1839,7 +1839,7 @@ xfs_buf_init(void)
 #ifdef XFS_BUF_TRACE
 	ktrace_free(xfs_buf_trace_buf);
 #endif
-	return error;
+	return -ENOMEM;
 }
 
 void
diff --git a/fs/xfs/linux-2.6/xfs_export.c b/fs/xfs/linux-2.6/xfs_export.c
index 80eb249f2fa0..b768ea910bbe 100644
--- a/fs/xfs/linux-2.6/xfs_export.c
+++ b/fs/xfs/linux-2.6/xfs_export.c
@@ -25,6 +25,8 @@
 #include "xfs_mount.h"
 #include "xfs_export.h"
 
+STATIC struct dentry dotdot = { .d_name.name = "..", .d_name.len = 2, };
+
 /*
  * XFS encodes and decodes the fileid portion of NFS filehandles
  * itself instead of letting the generic NFS code do it.  This
@@ -37,7 +39,7 @@
  */
 
 STATIC struct dentry *
-linvfs_decode_fh(
+xfs_fs_decode_fh(
 	struct super_block	*sb,
 	__u32			*fh,
 	int			fh_len,
@@ -78,12 +80,12 @@ linvfs_decode_fh(
 	}
 
 	fh = (__u32 *)&ifid;
-	return find_exported_dentry(sb, fh, parent, acceptable, context);
+	return sb->s_export_op->find_exported_dentry(sb, fh, parent, acceptable, context);
 }
 
 
 STATIC int
-linvfs_encode_fh(
+xfs_fs_encode_fh(
 	struct dentry		*dentry,
 	__u32			*fh,
 	int			*max_len,
@@ -95,7 +97,7 @@ linvfs_encode_fh(
 	int			len;
 	int			is64 = 0;
 #if XFS_BIG_INUMS
-	vfs_t			*vfs = LINVFS_GET_VFS(inode->i_sb);
+	vfs_t			*vfs = vfs_from_sb(inode->i_sb);
 
 	if (!(vfs->vfs_flag & VFS_32BITINODES)) {
 		/* filesystem may contain 64bit inode numbers */
@@ -130,21 +132,21 @@ linvfs_encode_fh(
 }
 
 STATIC struct dentry *
-linvfs_get_dentry(
+xfs_fs_get_dentry(
 	struct super_block	*sb,
 	void			*data)
 {
 	vnode_t			*vp;
 	struct inode		*inode;
 	struct dentry		*result;
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_VGET(vfsp, &vp, (fid_t *)data, error);
 	if (error || vp == NULL)
 		return ERR_PTR(-ESTALE) ;
 
-	inode = LINVFS_GET_IP(vp);
+	inode = vn_to_inode(vp);
 	result = d_alloc_anon(inode);
         if (!result) {
 		iput(inode);
@@ -154,25 +156,20 @@ linvfs_get_dentry(
 }
 
 STATIC struct dentry *
-linvfs_get_parent(
+xfs_fs_get_parent(
 	struct dentry		*child)
 {
 	int			error;
 	vnode_t			*vp, *cvp;
 	struct dentry		*parent;
-	struct dentry		dotdot;
-
-	dotdot.d_name.name = "..";
-	dotdot.d_name.len = 2;
-	dotdot.d_inode = NULL;
 
 	cvp = NULL;
-	vp = LINVFS_GET_VP(child->d_inode);
+	vp = vn_from_inode(child->d_inode);
 	VOP_LOOKUP(vp, &dotdot, &cvp, 0, NULL, NULL, error);
 	if (unlikely(error))
 		return ERR_PTR(-error);
 
-	parent = d_alloc_anon(LINVFS_GET_IP(cvp));
+	parent = d_alloc_anon(vn_to_inode(cvp));
 	if (unlikely(!parent)) {
 		VN_RELE(cvp);
 		return ERR_PTR(-ENOMEM);
@@ -180,9 +177,9 @@ linvfs_get_parent(
 	return parent;
 }
 
-struct export_operations linvfs_export_ops = {
-	.decode_fh		= linvfs_decode_fh,
-	.encode_fh		= linvfs_encode_fh,
-	.get_parent		= linvfs_get_parent,
-	.get_dentry		= linvfs_get_dentry,
+struct export_operations xfs_export_operations = {
+	.decode_fh		= xfs_fs_decode_fh,
+	.encode_fh		= xfs_fs_encode_fh,
+	.get_parent		= xfs_fs_get_parent,
+	.get_dentry		= xfs_fs_get_dentry,
 };
diff --git a/fs/xfs/linux-2.6/xfs_export.h b/fs/xfs/linux-2.6/xfs_export.h
index e5b0559700a4..e794ca4efc76 100644
--- a/fs/xfs/linux-2.6/xfs_export.h
+++ b/fs/xfs/linux-2.6/xfs_export.h
@@ -54,7 +54,7 @@
  * Note, the NFS filehandle also includes an fsid portion which
  * may have an inode number in it.  That number is hardcoded to
  * 32bits and there is no way for XFS to intercept it.  In
- * practice this means when exporting an XFS filesytem with 64bit
+ * practice this means when exporting an XFS filesystem with 64bit
  * inodes you should either export the mountpoint (rather than
  * a subdirectory) or use the "fsid" export option.
  */
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index ced4404339c7..85997b1205f5 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -43,13 +43,13 @@
 #include <linux/dcache.h>
 #include <linux/smp_lock.h>
 
-static struct vm_operations_struct linvfs_file_vm_ops;
+static struct vm_operations_struct xfs_file_vm_ops;
 #ifdef CONFIG_XFS_DMAPI
-static struct vm_operations_struct linvfs_dmapi_file_vm_ops;
+static struct vm_operations_struct xfs_dmapi_file_vm_ops;
 #endif
 
 STATIC inline ssize_t
-__linvfs_read(
+__xfs_file_read(
 	struct kiocb		*iocb,
 	char			__user *buf,
 	int			ioflags,
@@ -58,7 +58,7 @@ __linvfs_read(
 {
 	struct iovec		iov = {buf, count};
 	struct file		*file = iocb->ki_filp;
-	vnode_t			*vp = LINVFS_GET_VP(file->f_dentry->d_inode);
+	vnode_t			*vp = vn_from_inode(file->f_dentry->d_inode);
 	ssize_t			rval;
 
 	BUG_ON(iocb->ki_pos != pos);
@@ -71,28 +71,28 @@ __linvfs_read(
 
 
 STATIC ssize_t
-linvfs_aio_read(
+xfs_file_aio_read(
 	struct kiocb		*iocb,
 	char			__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_read(iocb, buf, IO_ISAIO, count, pos);
+	return __xfs_file_read(iocb, buf, IO_ISAIO, count, pos);
 }
 
 STATIC ssize_t
-linvfs_aio_read_invis(
+xfs_file_aio_read_invis(
 	struct kiocb		*iocb,
 	char			__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_read(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
+	return __xfs_file_read(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
 }
 
 
 STATIC inline ssize_t
-__linvfs_write(
+__xfs_file_write(
 	struct kiocb	*iocb,
 	const char	__user *buf,
 	int		ioflags,
@@ -102,7 +102,7 @@ __linvfs_write(
 	struct iovec	iov = {(void __user *)buf, count};
 	struct file	*file = iocb->ki_filp;
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	ssize_t		rval;
 
 	BUG_ON(iocb->ki_pos != pos);
@@ -115,28 +115,28 @@ __linvfs_write(
 
 
 STATIC ssize_t
-linvfs_aio_write(
+xfs_file_aio_write(
 	struct kiocb		*iocb,
 	const char		__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_write(iocb, buf, IO_ISAIO, count, pos);
+	return __xfs_file_write(iocb, buf, IO_ISAIO, count, pos);
 }
 
 STATIC ssize_t
-linvfs_aio_write_invis(
+xfs_file_aio_write_invis(
 	struct kiocb		*iocb,
 	const char		__user *buf,
 	size_t			count,
 	loff_t			pos)
 {
-	return __linvfs_write(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
+	return __xfs_file_write(iocb, buf, IO_ISAIO|IO_INVIS, count, pos);
 }
 
 
 STATIC inline ssize_t
-__linvfs_readv(
+__xfs_file_readv(
 	struct file		*file,
 	const struct iovec 	*iov,
 	int			ioflags,
@@ -144,8 +144,8 @@ __linvfs_readv(
 	loff_t			*ppos)
 {
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
-	struct		kiocb kiocb;
+	vnode_t		*vp = vn_from_inode(inode);
+	struct kiocb	kiocb;
 	ssize_t		rval;
 
 	init_sync_kiocb(&kiocb, file);
@@ -160,28 +160,28 @@ __linvfs_readv(
 }
 
 STATIC ssize_t
-linvfs_readv(
+xfs_file_readv(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_readv(file, iov, 0, nr_segs, ppos);
+	return __xfs_file_readv(file, iov, 0, nr_segs, ppos);
 }
 
 STATIC ssize_t
-linvfs_readv_invis(
+xfs_file_readv_invis(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_readv(file, iov, IO_INVIS, nr_segs, ppos);
+	return __xfs_file_readv(file, iov, IO_INVIS, nr_segs, ppos);
 }
 
 
 STATIC inline ssize_t
-__linvfs_writev(
+__xfs_file_writev(
 	struct file		*file,
 	const struct iovec 	*iov,
 	int			ioflags,
@@ -189,8 +189,8 @@ __linvfs_writev(
 	loff_t			*ppos)
 {
 	struct inode	*inode = file->f_mapping->host;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
-	struct		kiocb kiocb;
+	vnode_t		*vp = vn_from_inode(inode);
+	struct kiocb	kiocb;
 	ssize_t		rval;
 
 	init_sync_kiocb(&kiocb, file);
@@ -206,34 +206,34 @@ __linvfs_writev(
 
 
 STATIC ssize_t
-linvfs_writev(
+xfs_file_writev(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_writev(file, iov, 0, nr_segs, ppos);
+	return __xfs_file_writev(file, iov, 0, nr_segs, ppos);
 }
 
 STATIC ssize_t
-linvfs_writev_invis(
+xfs_file_writev_invis(
 	struct file		*file,
 	const struct iovec 	*iov,
 	unsigned long		nr_segs,
 	loff_t			*ppos)
 {
-	return __linvfs_writev(file, iov, IO_INVIS, nr_segs, ppos);
+	return __xfs_file_writev(file, iov, IO_INVIS, nr_segs, ppos);
 }
 
 STATIC ssize_t
-linvfs_sendfile(
+xfs_file_sendfile(
 	struct file		*filp,
 	loff_t			*ppos,
 	size_t			count,
 	read_actor_t		actor,
 	void			*target)
 {
-	vnode_t			*vp = LINVFS_GET_VP(filp->f_dentry->d_inode);
+	vnode_t			*vp = vn_from_inode(filp->f_dentry->d_inode);
 	ssize_t			rval;
 
 	VOP_SENDFILE(vp, filp, ppos, 0, count, actor, target, NULL, rval);
@@ -242,11 +242,11 @@ linvfs_sendfile(
 
 
 STATIC int
-linvfs_open(
+xfs_file_open(
 	struct inode	*inode,
 	struct file	*filp)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
 
 	if (!(filp->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
@@ -259,11 +259,11 @@ linvfs_open(
 
 
 STATIC int
-linvfs_release(
+xfs_file_release(
 	struct inode	*inode,
 	struct file	*filp)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error = 0;
 
 	if (vp)
@@ -273,13 +273,13 @@ linvfs_release(
 
 
 STATIC int
-linvfs_fsync(
+xfs_file_fsync(
 	struct file	*filp,
 	struct dentry	*dentry,
 	int		datasync)
 {
 	struct inode	*inode = dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
 	int		flags = FSYNC_WAIT;
 
@@ -292,7 +292,7 @@ linvfs_fsync(
 }
 
 /*
- * linvfs_readdir maps to VOP_READDIR().
+ * xfs_file_readdir maps to VOP_READDIR().
  * We need to build a uio, cred, ...
  */
 
@@ -301,13 +301,13 @@ linvfs_fsync(
 #ifdef CONFIG_XFS_DMAPI
 
 STATIC struct page *
-linvfs_filemap_nopage(
+xfs_vm_nopage(
 	struct vm_area_struct	*area,
 	unsigned long		address,
 	int			*type)
 {
 	struct inode	*inode = area->vm_file->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	xfs_mount_t	*mp = XFS_VFSTOM(vp->v_vfsp);
 	int		error;
 
@@ -324,7 +324,7 @@ linvfs_filemap_nopage(
 
 
 STATIC int
-linvfs_readdir(
+xfs_file_readdir(
 	struct file	*filp,
 	void		*dirent,
 	filldir_t	filldir)
@@ -340,7 +340,7 @@ linvfs_readdir(
 	xfs_off_t	start_offset, curr_offset;
 	xfs_dirent_t	*dbp = NULL;
 
-	vp = LINVFS_GET_VP(filp->f_dentry->d_inode);
+	vp = vn_from_inode(filp->f_dentry->d_inode);
 	ASSERT(vp);
 
 	/* Try fairly hard to get memory */
@@ -404,39 +404,40 @@ done:
 
 
 STATIC int
-linvfs_file_mmap(
+xfs_file_mmap(
 	struct file	*filp,
 	struct vm_area_struct *vma)
 {
 	struct inode	*ip = filp->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(ip);
-	vattr_t		va = { .va_mask = XFS_AT_UPDATIME };
+	vnode_t		*vp = vn_from_inode(ip);
+	vattr_t		vattr;
 	int		error;
 
-	vma->vm_ops = &linvfs_file_vm_ops;
+	vma->vm_ops = &xfs_file_vm_ops;
 
 #ifdef CONFIG_XFS_DMAPI
 	if (vp->v_vfsp->vfs_flag & VFS_DMI) {
-		vma->vm_ops = &linvfs_dmapi_file_vm_ops;
+		vma->vm_ops = &xfs_dmapi_file_vm_ops;
 	}
 #endif /* CONFIG_XFS_DMAPI */
 
-	VOP_SETATTR(vp, &va, XFS_AT_UPDATIME, NULL, error);
-	if (!error)
-		vn_revalidate(vp);	/* update Linux inode flags */
+	vattr.va_mask = XFS_AT_UPDATIME;
+	VOP_SETATTR(vp, &vattr, XFS_AT_UPDATIME, NULL, error);
+	if (likely(!error))
+		__vn_revalidate(vp, &vattr);	/* update flags */
 	return 0;
 }
 
 
 STATIC long
-linvfs_ioctl(
+xfs_file_ioctl(
 	struct file	*filp,
 	unsigned int	cmd,
 	unsigned long	arg)
 {
 	int		error;
-	struct inode *inode = filp->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	struct inode	*inode = filp->f_dentry->d_inode;
+	vnode_t		*vp = vn_from_inode(inode);
 
 	VOP_IOCTL(vp, inode, filp, 0, cmd, (void __user *)arg, error);
 	VMODIFY(vp);
@@ -451,14 +452,14 @@ linvfs_ioctl(
 }
 
 STATIC long
-linvfs_ioctl_invis(
+xfs_file_ioctl_invis(
 	struct file	*filp,
 	unsigned int	cmd,
 	unsigned long	arg)
 {
 	int		error;
-	struct inode *inode = filp->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	struct inode	*inode = filp->f_dentry->d_inode;
+	vnode_t		*vp = vn_from_inode(inode);
 
 	ASSERT(vp);
 	VOP_IOCTL(vp, inode, filp, IO_INVIS, cmd, (void __user *)arg, error);
@@ -476,11 +477,11 @@ linvfs_ioctl_invis(
 #ifdef CONFIG_XFS_DMAPI
 #ifdef HAVE_VMOP_MPROTECT
 STATIC int
-linvfs_mprotect(
+xfs_vm_mprotect(
 	struct vm_area_struct *vma,
 	unsigned int	newflags)
 {
-	vnode_t		*vp = LINVFS_GET_VP(vma->vm_file->f_dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(vma->vm_file->f_dentry->d_inode);
 	int		error = 0;
 
 	if (vp->v_vfsp->vfs_flag & VFS_DMI) {
@@ -503,10 +504,10 @@ linvfs_mprotect(
  * it back online.
  */
 STATIC int
-linvfs_open_exec(
+xfs_file_open_exec(
 	struct inode	*inode)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	xfs_mount_t	*mp = XFS_VFSTOM(vp->v_vfsp);
 	int		error = 0;
 	xfs_inode_t	*ip;
@@ -527,69 +528,69 @@ open_exec_out:
 }
 #endif /* HAVE_FOP_OPEN_EXEC */
 
-struct file_operations linvfs_file_operations = {
+const struct file_operations xfs_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
-	.readv		= linvfs_readv,
-	.writev		= linvfs_writev,
-	.aio_read	= linvfs_aio_read,
-	.aio_write	= linvfs_aio_write,
-	.sendfile	= linvfs_sendfile,
-	.unlocked_ioctl	= linvfs_ioctl,
+	.readv		= xfs_file_readv,
+	.writev		= xfs_file_writev,
+	.aio_read	= xfs_file_aio_read,
+	.aio_write	= xfs_file_aio_write,
+	.sendfile	= xfs_file_sendfile,
+	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
-	.compat_ioctl	= linvfs_compat_ioctl,
+	.compat_ioctl	= xfs_file_compat_ioctl,
 #endif
-	.mmap		= linvfs_file_mmap,
-	.open		= linvfs_open,
-	.release	= linvfs_release,
-	.fsync		= linvfs_fsync,
+	.mmap		= xfs_file_mmap,
+	.open		= xfs_file_open,
+	.release	= xfs_file_release,
+	.fsync		= xfs_file_fsync,
 #ifdef HAVE_FOP_OPEN_EXEC
-	.open_exec	= linvfs_open_exec,
+	.open_exec	= xfs_file_open_exec,
 #endif
 };
 
-struct file_operations linvfs_invis_file_operations = {
+const struct file_operations xfs_invis_file_operations = {
 	.llseek		= generic_file_llseek,
 	.read		= do_sync_read,
 	.write		= do_sync_write,
-	.readv		= linvfs_readv_invis,
-	.writev		= linvfs_writev_invis,
-	.aio_read	= linvfs_aio_read_invis,
-	.aio_write	= linvfs_aio_write_invis,
-	.sendfile	= linvfs_sendfile,
-	.unlocked_ioctl	= linvfs_ioctl_invis,
+	.readv		= xfs_file_readv_invis,
+	.writev		= xfs_file_writev_invis,
+	.aio_read	= xfs_file_aio_read_invis,
+	.aio_write	= xfs_file_aio_write_invis,
+	.sendfile	= xfs_file_sendfile,
+	.unlocked_ioctl	= xfs_file_ioctl_invis,
 #ifdef CONFIG_COMPAT
-	.compat_ioctl	= linvfs_compat_invis_ioctl,
+	.compat_ioctl	= xfs_file_compat_invis_ioctl,
 #endif
-	.mmap		= linvfs_file_mmap,
-	.open		= linvfs_open,
-	.release	= linvfs_release,
-	.fsync		= linvfs_fsync,
+	.mmap		= xfs_file_mmap,
+	.open		= xfs_file_open,
+	.release	= xfs_file_release,
+	.fsync		= xfs_file_fsync,
 };
 
 
-struct file_operations linvfs_dir_operations = {
+const struct file_operations xfs_dir_file_operations = {
 	.read		= generic_read_dir,
-	.readdir	= linvfs_readdir,
-	.unlocked_ioctl	= linvfs_ioctl,
+	.readdir	= xfs_file_readdir,
+	.unlocked_ioctl	= xfs_file_ioctl,
 #ifdef CONFIG_COMPAT
-	.compat_ioctl	= linvfs_compat_ioctl,
+	.compat_ioctl	= xfs_file_compat_ioctl,
 #endif
-	.fsync		= linvfs_fsync,
+	.fsync		= xfs_file_fsync,
 };
 
-static struct vm_operations_struct linvfs_file_vm_ops = {
+static struct vm_operations_struct xfs_file_vm_ops = {
 	.nopage		= filemap_nopage,
 	.populate	= filemap_populate,
 };
 
 #ifdef CONFIG_XFS_DMAPI
-static struct vm_operations_struct linvfs_dmapi_file_vm_ops = {
-	.nopage		= linvfs_filemap_nopage,
+static struct vm_operations_struct xfs_dmapi_file_vm_ops = {
+	.nopage		= xfs_vm_nopage,
 	.populate	= filemap_populate,
 #ifdef HAVE_VMOP_MPROTECT
-	.mprotect	= linvfs_mprotect,
+	.mprotect	= xfs_vm_mprotect,
 #endif
 };
 #endif /* CONFIG_XFS_DMAPI */
diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c
index 4fa4b1a5187e..575f2a790f31 100644
--- a/fs/xfs/linux-2.6/xfs_fs_subr.c
+++ b/fs/xfs/linux-2.6/xfs_fs_subr.c
@@ -57,7 +57,7 @@ fs_tosspages(
 	int		fiopt)
 {
 	vnode_t		*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 
 	if (VN_CACHED(vp))
 		truncate_inode_pages(ip->i_mapping, first);
@@ -76,7 +76,7 @@ fs_flushinval_pages(
 	int		fiopt)
 {
 	vnode_t		*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 
 	if (VN_CACHED(vp)) {
 		filemap_write_and_wait(ip->i_mapping);
@@ -98,7 +98,7 @@ fs_flush_pages(
 	int		fiopt)
 {
 	vnode_t		*vp = BHV_TO_VNODE(bdp);
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 
 	if (VN_CACHED(vp)) {
 		filemap_fdatawrite(ip->i_mapping);
diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c
index 4db47790415c..84478491609b 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl.c
@@ -138,7 +138,7 @@ xfs_find_handle(
 	}
 
 	/* we need the vnode */
-	vp = LINVFS_GET_VP(inode);
+	vp = vn_from_inode(inode);
 
 	/* now we can grab the fsid */
 	memcpy(&handle.ha_fsid, vp->v_vfsp->vfs_altfsid, sizeof(xfs_fsid_t));
@@ -256,7 +256,7 @@ xfs_vget_fsop_handlereq(
 	}
 
 	vpp = XFS_ITOV(ip);
-	inodep = LINVFS_GET_IP(vpp);
+	inodep = vn_to_inode(vpp);
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
 	*vp = vpp;
@@ -344,7 +344,7 @@ xfs_open_by_handle(
 		return -XFS_ERROR(-PTR_ERR(filp));
 	}
 	if (inode->i_mode & S_IFREG)
-		filp->f_op = &linvfs_invis_file_operations;
+		filp->f_op = &xfs_invis_file_operations;
 
 	fd_install(new_fd, filp);
 	return new_fd;
@@ -715,7 +715,7 @@ xfs_ioctl(
 	xfs_inode_t		*ip;
 	xfs_mount_t		*mp;
 
-	vp = LINVFS_GET_VP(inode);
+	vp = vn_from_inode(inode);
 
 	vn_trace_entry(vp, "xfs_ioctl", (inst_t *)__return_address);
 
@@ -1160,105 +1160,129 @@ xfs_ioc_xattr(
 	void			__user *arg)
 {
 	struct fsxattr		fa;
-	vattr_t			va;
-	int			error;
+	struct vattr		*vattr;
+	int			error = 0;
 	int			attr_flags;
 	unsigned int		flags;
 
+	vattr = kmalloc(sizeof(*vattr), GFP_KERNEL);
+	if (unlikely(!vattr))
+		return -ENOMEM;
+
 	switch (cmd) {
 	case XFS_IOC_FSGETXATTR: {
-		va.va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
-			     XFS_AT_NEXTENTS | XFS_AT_PROJID;
-		VOP_GETATTR(vp, &va, 0, NULL, error);
-		if (error)
-			return -error;
+		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
+				 XFS_AT_NEXTENTS | XFS_AT_PROJID;
+		VOP_GETATTR(vp, vattr, 0, NULL, error);
+		if (unlikely(error)) {
+			error = -error;
+			break;
+		}
 
-		fa.fsx_xflags	= va.va_xflags;
-		fa.fsx_extsize	= va.va_extsize;
-		fa.fsx_nextents = va.va_nextents;
-		fa.fsx_projid	= va.va_projid;
+		fa.fsx_xflags	= vattr->va_xflags;
+		fa.fsx_extsize	= vattr->va_extsize;
+		fa.fsx_nextents = vattr->va_nextents;
+		fa.fsx_projid	= vattr->va_projid;
 
-		if (copy_to_user(arg, &fa, sizeof(fa)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+		if (copy_to_user(arg, &fa, sizeof(fa))) {
+			error = -EFAULT;
+			break;
+		}
+		break;
 	}
 
 	case XFS_IOC_FSSETXATTR: {
-		if (copy_from_user(&fa, arg, sizeof(fa)))
-			return -XFS_ERROR(EFAULT);
+		if (copy_from_user(&fa, arg, sizeof(fa))) {
+			error = -EFAULT;
+			break;
+		}
 
 		attr_flags = 0;
 		if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
 			attr_flags |= ATTR_NONBLOCK;
 
-		va.va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID;
-		va.va_xflags  = fa.fsx_xflags;
-		va.va_extsize = fa.fsx_extsize;
-		va.va_projid  = fa.fsx_projid;
+		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID;
+		vattr->va_xflags  = fa.fsx_xflags;
+		vattr->va_extsize = fa.fsx_extsize;
+		vattr->va_projid  = fa.fsx_projid;
 
-		VOP_SETATTR(vp, &va, attr_flags, NULL, error);
-		if (!error)
-			vn_revalidate(vp);	/* update Linux inode flags */
-		return -error;
+		VOP_SETATTR(vp, vattr, attr_flags, NULL, error);
+		if (likely(!error))
+			__vn_revalidate(vp, vattr);	/* update flags */
+		error = -error;
+		break;
 	}
 
 	case XFS_IOC_FSGETXATTRA: {
-		va.va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
-			     XFS_AT_ANEXTENTS | XFS_AT_PROJID;
-		VOP_GETATTR(vp, &va, 0, NULL, error);
-		if (error)
-			return -error;
+		vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | \
+				 XFS_AT_ANEXTENTS | XFS_AT_PROJID;
+		VOP_GETATTR(vp, vattr, 0, NULL, error);
+		if (unlikely(error)) {
+			error = -error;
+			break;
+		}
 
-		fa.fsx_xflags	= va.va_xflags;
-		fa.fsx_extsize	= va.va_extsize;
-		fa.fsx_nextents = va.va_anextents;
-		fa.fsx_projid	= va.va_projid;
+		fa.fsx_xflags	= vattr->va_xflags;
+		fa.fsx_extsize	= vattr->va_extsize;
+		fa.fsx_nextents = vattr->va_anextents;
+		fa.fsx_projid	= vattr->va_projid;
 
-		if (copy_to_user(arg, &fa, sizeof(fa)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+		if (copy_to_user(arg, &fa, sizeof(fa))) {
+			error = -EFAULT;
+			break;
+		}
+		break;
 	}
 
 	case XFS_IOC_GETXFLAGS: {
 		flags = xfs_di2lxflags(ip->i_d.di_flags);
 		if (copy_to_user(arg, &flags, sizeof(flags)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+			error = -EFAULT;
+		break;
 	}
 
 	case XFS_IOC_SETXFLAGS: {
-		if (copy_from_user(&flags, arg, sizeof(flags)))
-			return -XFS_ERROR(EFAULT);
+		if (copy_from_user(&flags, arg, sizeof(flags))) {
+			error = -EFAULT;
+			break;
+		}
 
 		if (flags & ~(LINUX_XFLAG_IMMUTABLE | LINUX_XFLAG_APPEND | \
 			      LINUX_XFLAG_NOATIME | LINUX_XFLAG_NODUMP | \
-			      LINUX_XFLAG_SYNC))
-			return -XFS_ERROR(EOPNOTSUPP);
+			      LINUX_XFLAG_SYNC)) {
+			error = -EOPNOTSUPP;
+			break;
+		}
 
 		attr_flags = 0;
 		if (filp->f_flags & (O_NDELAY|O_NONBLOCK))
 			attr_flags |= ATTR_NONBLOCK;
 
-		va.va_mask = XFS_AT_XFLAGS;
-		va.va_xflags = xfs_merge_ioc_xflags(flags,
-				xfs_ip2xflags(ip));
+		vattr->va_mask = XFS_AT_XFLAGS;
+		vattr->va_xflags = xfs_merge_ioc_xflags(flags,
+							xfs_ip2xflags(ip));
 
-		VOP_SETATTR(vp, &va, attr_flags, NULL, error);
-		if (!error)
-			vn_revalidate(vp);	/* update Linux inode flags */
-		return -error;
+		VOP_SETATTR(vp, vattr, attr_flags, NULL, error);
+		if (likely(!error))
+			__vn_revalidate(vp, vattr);	/* update flags */
+		error = -error;
+		break;
 	}
 
 	case XFS_IOC_GETVERSION: {
-		flags = LINVFS_GET_IP(vp)->i_generation;
+		flags = vn_to_inode(vp)->i_generation;
 		if (copy_to_user(arg, &flags, sizeof(flags)))
-			return -XFS_ERROR(EFAULT);
-		return 0;
+			error = -EFAULT;
+		break;
 	}
 
 	default:
-		return -ENOTTY;
+		error = -ENOTTY;
+		break;
 	}
+
+	kfree(vattr);
+	return error;
 }
 
 STATIC int
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.c b/fs/xfs/linux-2.6/xfs_ioctl32.c
index a7c9ba1a9f7b..251bfe451a3f 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.c
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.c
@@ -72,7 +72,7 @@ xfs_ioctl32_flock(
 	    copy_in_user(&p->l_pid,	&p32->l_pid,	sizeof(u32)) ||
 	    copy_in_user(&p->l_pad,	&p32->l_pad,	4*sizeof(u32)))
 		return -EFAULT;
-	
+
 	return (unsigned long)p;
 }
 
@@ -107,11 +107,15 @@ xfs_ioctl32_bulkstat(
 #endif
 
 STATIC long
-__linvfs_compat_ioctl(int mode, struct file *f, unsigned cmd, unsigned long arg)
+xfs_compat_ioctl(
+	int		mode,
+	struct file	*file,
+	unsigned	cmd,
+	unsigned long	arg)
 {
+	struct inode	*inode = file->f_dentry->d_inode;
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
-	struct		inode *inode = f->f_dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
 
 	switch (cmd) {
 	case XFS_IOC_DIOINFO:
@@ -189,26 +193,26 @@ __linvfs_compat_ioctl(int mode, struct file *f, unsigned cmd, unsigned long arg)
 		return -ENOIOCTLCMD;
 	}
 
-	VOP_IOCTL(vp, inode, f, mode, cmd, (void __user *)arg, error);
+	VOP_IOCTL(vp, inode, file, mode, cmd, (void __user *)arg, error);
 	VMODIFY(vp);
 
 	return error;
 }
 
 long
-linvfs_compat_ioctl(
-	struct file		*f,
+xfs_file_compat_ioctl(
+	struct file		*file,
 	unsigned		cmd,
 	unsigned long		arg)
 {
-	return __linvfs_compat_ioctl(0, f, cmd, arg);
+	return xfs_compat_ioctl(0, file, cmd, arg);
 }
 
 long
-linvfs_compat_invis_ioctl(
-	struct file		*f,
+xfs_file_compat_invis_ioctl(
+	struct file		*file,
 	unsigned		cmd,
 	unsigned long		arg)
 {
-	return __linvfs_compat_ioctl(IO_INVIS, f, cmd, arg);
+	return xfs_compat_ioctl(IO_INVIS, file, cmd, arg);
 }
diff --git a/fs/xfs/linux-2.6/xfs_ioctl32.h b/fs/xfs/linux-2.6/xfs_ioctl32.h
index 011c273bec50..02de6e62ee37 100644
--- a/fs/xfs/linux-2.6/xfs_ioctl32.h
+++ b/fs/xfs/linux-2.6/xfs_ioctl32.h
@@ -18,7 +18,7 @@
 #ifndef __XFS_IOCTL32_H__
 #define __XFS_IOCTL32_H__
 
-extern long linvfs_compat_ioctl(struct file *, unsigned, unsigned long);
-extern long linvfs_compat_invis_ioctl(struct file *f, unsigned, unsigned long);
+extern long xfs_file_compat_ioctl(struct file *, unsigned, unsigned long);
+extern long xfs_file_compat_invis_ioctl(struct file *, unsigned, unsigned long);
 
 #endif /* __XFS_IOCTL32_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c
index d7f6f2d8ac8e..149237304fb6 100644
--- a/fs/xfs/linux-2.6/xfs_iops.c
+++ b/fs/xfs/linux-2.6/xfs_iops.c
@@ -106,7 +106,7 @@ xfs_ichgtime(
 	xfs_inode_t	*ip,
 	int		flags)
 {
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 	timespec_t	tv;
 
 	nanotime(&tv);
@@ -198,22 +198,22 @@ xfs_ichgtime_fast(
  * Pull the link count and size up from the xfs inode to the linux inode
  */
 STATIC void
-validate_fields(
-	struct inode	*ip)
+xfs_validate_fields(
+	struct inode	*ip,
+	struct vattr	*vattr)
 {
-	vnode_t		*vp = LINVFS_GET_VP(ip);
-	vattr_t		va;
+	vnode_t		*vp = vn_from_inode(ip);
 	int		error;
 
-	va.va_mask = XFS_AT_NLINK|XFS_AT_SIZE|XFS_AT_NBLOCKS;
-	VOP_GETATTR(vp, &va, ATTR_LAZY, NULL, error);
-	if (likely(!error)) {
-		ip->i_nlink = va.va_nlink;
-		ip->i_blocks = va.va_nblocks;
+	vattr->va_mask = XFS_AT_NLINK|XFS_AT_SIZE|XFS_AT_NBLOCKS;
+	VOP_GETATTR(vp, vattr, ATTR_LAZY, NULL, error);
+  	if (likely(!error)) {
+		ip->i_nlink = vattr->va_nlink;
+		ip->i_blocks = vattr->va_nblocks;
 
-		/* we're under i_mutex so i_size can't change under us */
-		if (i_size_read(ip) != va.va_size)
-			i_size_write(ip, va.va_size);
+		/* we're under i_sem so i_size can't change under us */
+		if (i_size_read(ip) != vattr->va_size)
+			i_size_write(ip, vattr->va_size);
 	}
 }
 
@@ -224,11 +224,11 @@ validate_fields(
  * inode, of course, such that log replay can't cause these to be lost).
  */
 STATIC int
-linvfs_init_security(
+xfs_init_security(
 	struct vnode	*vp,
 	struct inode	*dir)
 {
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 	size_t		length;
 	void		*value;
 	char		*name;
@@ -257,46 +257,46 @@ linvfs_init_security(
  * XXX(hch):  nfsd is broken, better fix it instead.
  */
 STATIC inline int
-has_fs_struct(struct task_struct *task)
+xfs_has_fs_struct(struct task_struct *task)
 {
 	return (task->fs != init_task.fs);
 }
 
 STATIC inline void
-cleanup_inode(
+xfs_cleanup_inode(
 	vnode_t		*dvp,
 	vnode_t		*vp,
-	struct dentry	*dentry,	
+	struct dentry	*dentry,
 	int		mode)
 {
 	struct dentry   teardown = {};
-	int             err2;
+	int             error;
 
 	/* Oh, the horror.
-	 * If we can't add the ACL or we fail in 
-	 * linvfs_init_security we must back out.
+	 * If we can't add the ACL or we fail in
+	 * xfs_init_security we must back out.
 	 * ENOSPC can hit here, among other things.
 	 */
-	teardown.d_inode = LINVFS_GET_IP(vp);
+	teardown.d_inode = vn_to_inode(vp);
 	teardown.d_name = dentry->d_name;
 
 	if (S_ISDIR(mode))
-	  	VOP_RMDIR(dvp, &teardown, NULL, err2);
+	  	VOP_RMDIR(dvp, &teardown, NULL, error);
 	else
-		VOP_REMOVE(dvp, &teardown, NULL, err2);
+		VOP_REMOVE(dvp, &teardown, NULL, error);
 	VN_RELE(vp);
 }
 
 STATIC int
-linvfs_mknod(
+xfs_vn_mknod(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	int		mode,
 	dev_t		rdev)
 {
 	struct inode	*ip;
-	vattr_t		va;
-	vnode_t		*vp = NULL, *dvp = LINVFS_GET_VP(dir);
+	vattr_t		vattr = { 0 };
+	vnode_t		*vp = NULL, *dvp = vn_from_inode(dir);
 	xfs_acl_t	*default_acl = NULL;
 	attrexists_t	test_default_acl = _ACL_DEFAULT_EXISTS;
 	int		error;
@@ -305,99 +305,98 @@ linvfs_mknod(
 	 * Irix uses Missed'em'V split, but doesn't want to see
 	 * the upper 5 bits of (14bit) major.
 	 */
-	if (!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff)
+	if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff))
 		return -EINVAL;
 
-	if (test_default_acl && test_default_acl(dvp)) {
-		if (!_ACL_ALLOC(default_acl))
+	if (unlikely(test_default_acl && test_default_acl(dvp))) {
+		if (!_ACL_ALLOC(default_acl)) {
 			return -ENOMEM;
+		}
 		if (!_ACL_GET_DEFAULT(dvp, default_acl)) {
 			_ACL_FREE(default_acl);
 			default_acl = NULL;
 		}
 	}
 
-	if (IS_POSIXACL(dir) && !default_acl && has_fs_struct(current))
+	if (IS_POSIXACL(dir) && !default_acl && xfs_has_fs_struct(current))
 		mode &= ~current->fs->umask;
 
-	memset(&va, 0, sizeof(va));
-	va.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
-	va.va_mode = mode;
+	vattr.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
+	vattr.va_mode = mode;
 
 	switch (mode & S_IFMT) {
 	case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK:
-		va.va_rdev = sysv_encode_dev(rdev);
-		va.va_mask |= XFS_AT_RDEV;
+		vattr.va_rdev = sysv_encode_dev(rdev);
+		vattr.va_mask |= XFS_AT_RDEV;
 		/*FALLTHROUGH*/
 	case S_IFREG:
-		VOP_CREATE(dvp, dentry, &va, &vp, NULL, error);
+		VOP_CREATE(dvp, dentry, &vattr, &vp, NULL, error);
 		break;
 	case S_IFDIR:
-		VOP_MKDIR(dvp, dentry, &va, &vp, NULL, error);
+		VOP_MKDIR(dvp, dentry, &vattr, &vp, NULL, error);
 		break;
 	default:
 		error = EINVAL;
 		break;
 	}
 
-	if (!error)
-	{
-		error = linvfs_init_security(vp, dir);
+	if (unlikely(!error)) {
+		error = xfs_init_security(vp, dir);
 		if (error)
-			cleanup_inode(dvp, vp, dentry, mode);
+			xfs_cleanup_inode(dvp, vp, dentry, mode);
 	}
 
-	if (default_acl) {
+	if (unlikely(default_acl)) {
 		if (!error) {
-			error = _ACL_INHERIT(vp, &va, default_acl);
-			if (!error) 
+			error = _ACL_INHERIT(vp, &vattr, default_acl);
+			if (!error)
 				VMODIFY(vp);
 			else
-				cleanup_inode(dvp, vp, dentry, mode);
+				xfs_cleanup_inode(dvp, vp, dentry, mode);
 		}
 		_ACL_FREE(default_acl);
 	}
 
-	if (!error) {
+	if (likely(!error)) {
 		ASSERT(vp);
-		ip = LINVFS_GET_IP(vp);
+		ip = vn_to_inode(vp);
 
 		if (S_ISCHR(mode) || S_ISBLK(mode))
 			ip->i_rdev = rdev;
 		else if (S_ISDIR(mode))
-			validate_fields(ip);
+			xfs_validate_fields(ip, &vattr);
 		d_instantiate(dentry, ip);
-		validate_fields(dir);
+		xfs_validate_fields(dir, &vattr);
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_create(
+xfs_vn_create(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	int		mode,
 	struct nameidata *nd)
 {
-	return linvfs_mknod(dir, dentry, mode, 0);
+	return xfs_vn_mknod(dir, dentry, mode, 0);
 }
 
 STATIC int
-linvfs_mkdir(
+xfs_vn_mkdir(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	int		mode)
 {
-	return linvfs_mknod(dir, dentry, mode|S_IFDIR, 0);
+	return xfs_vn_mknod(dir, dentry, mode|S_IFDIR, 0);
 }
 
 STATIC struct dentry *
-linvfs_lookup(
+xfs_vn_lookup(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	struct nameidata *nd)
 {
-	struct vnode	*vp = LINVFS_GET_VP(dir), *cvp;
+	struct vnode	*vp = vn_from_inode(dir), *cvp;
 	int		error;
 
 	if (dentry->d_name.len >= MAXNAMELEN)
@@ -411,11 +410,11 @@ linvfs_lookup(
 		return NULL;
 	}
 
-	return d_splice_alias(LINVFS_GET_IP(cvp), dentry);
+	return d_splice_alias(vn_to_inode(cvp), dentry);
 }
 
 STATIC int
-linvfs_link(
+xfs_vn_link(
 	struct dentry	*old_dentry,
 	struct inode	*dir,
 	struct dentry	*dentry)
@@ -423,99 +422,102 @@ linvfs_link(
 	struct inode	*ip;	/* inode of guy being linked to */
 	vnode_t		*tdvp;	/* target directory for new name/link */
 	vnode_t		*vp;	/* vp of name being linked */
+	vattr_t		vattr;
 	int		error;
 
 	ip = old_dentry->d_inode;	/* inode being linked to */
 	if (S_ISDIR(ip->i_mode))
 		return -EPERM;
 
-	tdvp = LINVFS_GET_VP(dir);
-	vp = LINVFS_GET_VP(ip);
+	tdvp = vn_from_inode(dir);
+	vp = vn_from_inode(ip);
 
 	VOP_LINK(tdvp, vp, dentry, NULL, error);
-	if (!error) {
+	if (likely(!error)) {
 		VMODIFY(tdvp);
 		VN_HOLD(vp);
-		validate_fields(ip);
+		xfs_validate_fields(ip, &vattr);
 		d_instantiate(dentry, ip);
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_unlink(
+xfs_vn_unlink(
 	struct inode	*dir,
 	struct dentry	*dentry)
 {
 	struct inode	*inode;
 	vnode_t		*dvp;	/* directory containing name to remove */
+	vattr_t		vattr;
 	int		error;
 
 	inode = dentry->d_inode;
-	dvp = LINVFS_GET_VP(dir);
+	dvp = vn_from_inode(dir);
 
 	VOP_REMOVE(dvp, dentry, NULL, error);
-	if (!error) {
-		validate_fields(dir);	/* For size only */
-		validate_fields(inode);
+	if (likely(!error)) {
+		xfs_validate_fields(dir, &vattr);	/* size needs update */
+		xfs_validate_fields(inode, &vattr);
 	}
-
 	return -error;
 }
 
 STATIC int
-linvfs_symlink(
+xfs_vn_symlink(
 	struct inode	*dir,
 	struct dentry	*dentry,
 	const char	*symname)
 {
 	struct inode	*ip;
-	vattr_t		va;
+	vattr_t		vattr = { 0 };
 	vnode_t		*dvp;	/* directory containing name of symlink */
 	vnode_t		*cvp;	/* used to lookup symlink to put in dentry */
 	int		error;
 
-	dvp = LINVFS_GET_VP(dir);
+	dvp = vn_from_inode(dir);
 	cvp = NULL;
 
-	memset(&va, 0, sizeof(va));
-	va.va_mode = S_IFLNK |
+	vattr.va_mode = S_IFLNK |
 		(irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO);
-	va.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
+	vattr.va_mask = XFS_AT_TYPE|XFS_AT_MODE;
 
 	error = 0;
-	VOP_SYMLINK(dvp, dentry, &va, (char *)symname, &cvp, NULL, error);
+	VOP_SYMLINK(dvp, dentry, &vattr, (char *)symname, &cvp, NULL, error);
 	if (likely(!error && cvp)) {
-		error = linvfs_init_security(cvp, dir);
+		error = xfs_init_security(cvp, dir);
 		if (likely(!error)) {
-			ip = LINVFS_GET_IP(cvp);
+			ip = vn_to_inode(cvp);
 			d_instantiate(dentry, ip);
-			validate_fields(dir);
-			validate_fields(ip);
+			xfs_validate_fields(dir, &vattr);
+			xfs_validate_fields(ip, &vattr);
+		} else {
+			xfs_cleanup_inode(dvp, cvp, dentry, 0);
 		}
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_rmdir(
+xfs_vn_rmdir(
 	struct inode	*dir,
 	struct dentry	*dentry)
 {
 	struct inode	*inode = dentry->d_inode;
-	vnode_t		*dvp = LINVFS_GET_VP(dir);
+	vnode_t		*dvp = vn_from_inode(dir);
+	vattr_t		vattr;
 	int		error;
 
 	VOP_RMDIR(dvp, dentry, NULL, error);
-	if (!error) {
-		validate_fields(inode);
-		validate_fields(dir);
+	if (likely(!error)) {
+		xfs_validate_fields(inode, &vattr);
+		xfs_validate_fields(dir, &vattr);
 	}
 	return -error;
 }
 
 STATIC int
-linvfs_rename(
+xfs_vn_rename(
 	struct inode	*odir,
 	struct dentry	*odentry,
 	struct inode	*ndir,
@@ -524,22 +526,21 @@ linvfs_rename(
 	struct inode	*new_inode = ndentry->d_inode;
 	vnode_t		*fvp;	/* from directory */
 	vnode_t		*tvp;	/* target directory */
+	vattr_t		vattr;
 	int		error;
 
-	fvp = LINVFS_GET_VP(odir);
-	tvp = LINVFS_GET_VP(ndir);
+	fvp = vn_from_inode(odir);
+	tvp = vn_from_inode(ndir);
 
 	VOP_RENAME(fvp, odentry, tvp, ndentry, NULL, error);
-	if (error)
-		return -error;
-
-	if (new_inode)
-		validate_fields(new_inode);
-
-	validate_fields(odir);
-	if (ndir != odir)
-		validate_fields(ndir);
-	return 0;
+	if (likely(!error)) {
+		if (new_inode)
+			xfs_validate_fields(new_inode, &vattr);
+		xfs_validate_fields(odir, &vattr);
+		if (ndir != odir)
+			xfs_validate_fields(ndir, &vattr);
+	}
+	return -error;
 }
 
 /*
@@ -548,7 +549,7 @@ linvfs_rename(
  * uio is kmalloced for this reason...
  */
 STATIC void *
-linvfs_follow_link(
+xfs_vn_follow_link(
 	struct dentry		*dentry,
 	struct nameidata	*nd)
 {
@@ -574,7 +575,7 @@ linvfs_follow_link(
 		return NULL;
 	}
 
-	vp = LINVFS_GET_VP(dentry->d_inode);
+	vp = vn_from_inode(dentry->d_inode);
 
 	iov.iov_base = link;
 	iov.iov_len = MAXPATHLEN;
@@ -599,7 +600,7 @@ linvfs_follow_link(
 }
 
 STATIC void
-linvfs_put_link(
+xfs_vn_put_link(
 	struct dentry	*dentry,
 	struct nameidata *nd,
 	void		*p)
@@ -612,12 +613,12 @@ linvfs_put_link(
 
 #ifdef CONFIG_XFS_POSIX_ACL
 STATIC int
-linvfs_permission(
+xfs_vn_permission(
 	struct inode	*inode,
 	int		mode,
 	struct nameidata *nd)
 {
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error;
 
 	mode <<= 6;		/* convert from linux to vnode access bits */
@@ -625,17 +626,17 @@ linvfs_permission(
 	return -error;
 }
 #else
-#define linvfs_permission NULL
+#define xfs_vn_permission NULL
 #endif
 
 STATIC int
-linvfs_getattr(
+xfs_vn_getattr(
 	struct vfsmount	*mnt,
 	struct dentry	*dentry,
 	struct kstat	*stat)
 {
 	struct inode	*inode = dentry->d_inode;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
+	vnode_t		*vp = vn_from_inode(inode);
 	int		error = 0;
 
 	if (unlikely(vp->v_flag & VMODIFIED))
@@ -646,18 +647,17 @@ linvfs_getattr(
 }
 
 STATIC int
-linvfs_setattr(
+xfs_vn_setattr(
 	struct dentry	*dentry,
 	struct iattr	*attr)
 {
 	struct inode	*inode = dentry->d_inode;
 	unsigned int	ia_valid = attr->ia_valid;
-	vnode_t		*vp = LINVFS_GET_VP(inode);
-	vattr_t		vattr;
+	vnode_t		*vp = vn_from_inode(inode);
+	vattr_t		vattr = { 0 };
 	int		flags = 0;
 	int		error;
 
-	memset(&vattr, 0, sizeof(vattr_t));
 	if (ia_valid & ATTR_UID) {
 		vattr.va_mask |= XFS_AT_UID;
 		vattr.va_uid = attr->ia_uid;
@@ -699,28 +699,27 @@ linvfs_setattr(
 #endif
 
 	VOP_SETATTR(vp, &vattr, flags, NULL, error);
-	if (error)
-		return -error;
-	vn_revalidate(vp);
-	return error;
+	if (likely(!error))
+		__vn_revalidate(vp, &vattr);
+	return -error;
 }
 
 STATIC void
-linvfs_truncate(
+xfs_vn_truncate(
 	struct inode	*inode)
 {
-	block_truncate_page(inode->i_mapping, inode->i_size, linvfs_get_block);
+	block_truncate_page(inode->i_mapping, inode->i_size, xfs_get_blocks);
 }
 
 STATIC int
-linvfs_setxattr(
+xfs_vn_setxattr(
 	struct dentry	*dentry,
 	const char	*name,
 	const void	*data,
 	size_t		size,
 	int		flags)
 {
-	vnode_t		*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(dentry->d_inode);
 	char		*attr = (char *)name;
 	attrnames_t	*namesp;
 	int		xflags = 0;
@@ -744,13 +743,13 @@ linvfs_setxattr(
 }
 
 STATIC ssize_t
-linvfs_getxattr(
+xfs_vn_getxattr(
 	struct dentry	*dentry,
 	const char	*name,
 	void		*data,
 	size_t		size)
 {
-	vnode_t		*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(dentry->d_inode);
 	char		*attr = (char *)name;
 	attrnames_t	*namesp;
 	int		xflags = 0;
@@ -774,12 +773,12 @@ linvfs_getxattr(
 }
 
 STATIC ssize_t
-linvfs_listxattr(
+xfs_vn_listxattr(
 	struct dentry		*dentry,
 	char			*data,
 	size_t			size)
 {
-	vnode_t			*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t			*vp = vn_from_inode(dentry->d_inode);
 	int			error, xflags = ATTR_KERNAMELS;
 	ssize_t			result;
 
@@ -794,11 +793,11 @@ linvfs_listxattr(
 }
 
 STATIC int
-linvfs_removexattr(
+xfs_vn_removexattr(
 	struct dentry	*dentry,
 	const char	*name)
 {
-	vnode_t		*vp = LINVFS_GET_VP(dentry->d_inode);
+	vnode_t		*vp = vn_from_inode(dentry->d_inode);
 	char		*attr = (char *)name;
 	attrnames_t	*namesp;
 	int		xflags = 0;
@@ -816,45 +815,45 @@ linvfs_removexattr(
 }
 
 
-struct inode_operations linvfs_file_inode_operations = {
-	.permission		= linvfs_permission,
-	.truncate		= linvfs_truncate,
-	.getattr		= linvfs_getattr,
-	.setattr		= linvfs_setattr,
-	.setxattr		= linvfs_setxattr,
-	.getxattr		= linvfs_getxattr,
-	.listxattr		= linvfs_listxattr,
-	.removexattr		= linvfs_removexattr,
+struct inode_operations xfs_inode_operations = {
+	.permission		= xfs_vn_permission,
+	.truncate		= xfs_vn_truncate,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= xfs_vn_setxattr,
+	.getxattr		= xfs_vn_getxattr,
+	.listxattr		= xfs_vn_listxattr,
+	.removexattr		= xfs_vn_removexattr,
 };
 
-struct inode_operations linvfs_dir_inode_operations = {
-	.create			= linvfs_create,
-	.lookup			= linvfs_lookup,
-	.link			= linvfs_link,
-	.unlink			= linvfs_unlink,
-	.symlink		= linvfs_symlink,
-	.mkdir			= linvfs_mkdir,
-	.rmdir			= linvfs_rmdir,
-	.mknod			= linvfs_mknod,
-	.rename			= linvfs_rename,
-	.permission		= linvfs_permission,
-	.getattr		= linvfs_getattr,
-	.setattr		= linvfs_setattr,
-	.setxattr		= linvfs_setxattr,
-	.getxattr		= linvfs_getxattr,
-	.listxattr		= linvfs_listxattr,
-	.removexattr		= linvfs_removexattr,
+struct inode_operations xfs_dir_inode_operations = {
+	.create			= xfs_vn_create,
+	.lookup			= xfs_vn_lookup,
+	.link			= xfs_vn_link,
+	.unlink			= xfs_vn_unlink,
+	.symlink		= xfs_vn_symlink,
+	.mkdir			= xfs_vn_mkdir,
+	.rmdir			= xfs_vn_rmdir,
+	.mknod			= xfs_vn_mknod,
+	.rename			= xfs_vn_rename,
+	.permission		= xfs_vn_permission,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= xfs_vn_setxattr,
+	.getxattr		= xfs_vn_getxattr,
+	.listxattr		= xfs_vn_listxattr,
+	.removexattr		= xfs_vn_removexattr,
 };
 
-struct inode_operations linvfs_symlink_inode_operations = {
+struct inode_operations xfs_symlink_inode_operations = {
 	.readlink		= generic_readlink,
-	.follow_link		= linvfs_follow_link,
-	.put_link		= linvfs_put_link,
-	.permission		= linvfs_permission,
-	.getattr		= linvfs_getattr,
-	.setattr		= linvfs_setattr,
-	.setxattr		= linvfs_setxattr,
-	.getxattr		= linvfs_getxattr,
-	.listxattr		= linvfs_listxattr,
-	.removexattr		= linvfs_removexattr,
+	.follow_link		= xfs_vn_follow_link,
+	.put_link		= xfs_vn_put_link,
+	.permission		= xfs_vn_permission,
+	.getattr		= xfs_vn_getattr,
+	.setattr		= xfs_vn_setattr,
+	.setxattr		= xfs_vn_setxattr,
+	.getxattr		= xfs_vn_getxattr,
+	.listxattr		= xfs_vn_listxattr,
+	.removexattr		= xfs_vn_removexattr,
 };
diff --git a/fs/xfs/linux-2.6/xfs_iops.h b/fs/xfs/linux-2.6/xfs_iops.h
index 6899a6b4a50a..ad6173da5678 100644
--- a/fs/xfs/linux-2.6/xfs_iops.h
+++ b/fs/xfs/linux-2.6/xfs_iops.h
@@ -18,13 +18,13 @@
 #ifndef __XFS_IOPS_H__
 #define __XFS_IOPS_H__
 
-extern struct inode_operations linvfs_file_inode_operations;
-extern struct inode_operations linvfs_dir_inode_operations;
-extern struct inode_operations linvfs_symlink_inode_operations;
+extern struct inode_operations xfs_inode_operations;
+extern struct inode_operations xfs_dir_inode_operations;
+extern struct inode_operations xfs_symlink_inode_operations;
 
-extern struct file_operations linvfs_file_operations;
-extern struct file_operations linvfs_invis_file_operations;
-extern struct file_operations linvfs_dir_operations;
+extern const struct file_operations xfs_file_operations;
+extern const struct file_operations xfs_dir_file_operations;
+extern const struct file_operations xfs_invis_file_operations;
 
 extern int xfs_ioctl(struct bhv_desc *, struct inode *, struct file *,
                         int, unsigned int, void __user *);
diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h
index 67389b745526..1fe09f2d6519 100644
--- a/fs/xfs/linux-2.6/xfs_linux.h
+++ b/fs/xfs/linux-2.6/xfs_linux.h
@@ -73,6 +73,9 @@
 #include <linux/list.h>
 #include <linux/proc_fs.h>
 #include <linux/sort.h>
+#include <linux/cpu.h>
+#include <linux/notifier.h>
+#include <linux/delay.h>
 
 #include <asm/page.h>
 #include <asm/div64.h>
@@ -100,6 +103,11 @@
  */
 #undef  HAVE_REFCACHE	/* reference cache not needed for NFS in 2.6 */
 #define HAVE_SENDFILE	/* sendfile(2) exists in 2.6, but not in 2.4 */
+#ifdef CONFIG_SMP
+#define HAVE_PERCPU_SB	/* per cpu superblock counters are a 2.6 feature */
+#else
+#undef  HAVE_PERCPU_SB	/* per cpu superblock counters are a 2.6 feature */
+#endif
 
 /*
  * State flag for unwritten extent buffers.
@@ -226,7 +234,7 @@ BUFFER_FNS(PrivateStart, unwritten);
 #define xfs_sort(a,n,s,fn)	sort(a,n,s,fn,NULL)
 #define xfs_stack_trace()	dump_stack()
 #define xfs_itruncate_data(ip, off)	\
-	(-vmtruncate(LINVFS_GET_IP(XFS_ITOV(ip)), (off)))
+	(-vmtruncate(vn_to_inode(XFS_ITOV(ip)), (off)))
 #define xfs_statvfs_fsid(statp, mp)	\
 	({ u64 id = huge_encode_dev((mp)->m_ddev_targp->bt_dev); \
 	   __kernel_fsid_t *fsid = &(statp)->f_fsid;	\
diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c
index e0ab45fbfebd..84ddf1893894 100644
--- a/fs/xfs/linux-2.6/xfs_lrw.c
+++ b/fs/xfs/linux-2.6/xfs_lrw.c
@@ -83,7 +83,7 @@ xfs_rw_enter_trace(
 		(void *)((unsigned long)ioflags),
 		(void *)((unsigned long)((io->io_new_size >> 32) & 0xffffffff)),
 		(void *)((unsigned long)(io->io_new_size & 0xffffffff)),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL,
 		(void *)NULL,
 		(void *)NULL,
@@ -113,7 +113,7 @@ xfs_inval_cached_trace(
 		(void *)((unsigned long)(first & 0xffffffff)),
 		(void *)((unsigned long)((last >> 32) & 0xffffffff)),
 		(void *)((unsigned long)(last & 0xffffffff)),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL,
 		(void *)NULL,
 		(void *)NULL,
@@ -249,9 +249,8 @@ xfs_read(
 	if (n < size)
 		size = n;
 
-	if (XFS_FORCED_SHUTDOWN(mp)) {
+	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
-	}
 
 	if (unlikely(ioflags & IO_ISDIRECT))
 		mutex_lock(&inode->i_mutex);
@@ -267,10 +266,14 @@ xfs_read(
 					dmflags, &locktype);
 		if (ret) {
 			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
-			goto unlock_isem;
+			goto unlock_mutex;
 		}
 	}
 
+	if (unlikely((ioflags & IO_ISDIRECT) && VN_CACHED(vp)))
+		VOP_FLUSHINVAL_PAGES(vp, ctooff(offtoct(*offset)),
+						-1, FI_REMAPF_LOCKED);
+
 	xfs_rw_enter_trace(XFS_READ_ENTER, &ip->i_iocore,
 				(void *)iovp, segs, *offset, ioflags);
 	ret = __generic_file_aio_read(iocb, iovp, segs, offset);
@@ -281,7 +284,7 @@ xfs_read(
 
 	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
 
-unlock_isem:
+unlock_mutex:
 	if (unlikely(ioflags & IO_ISDIRECT))
 		mutex_unlock(&inode->i_mutex);
 	return ret;
@@ -432,7 +435,7 @@ xfs_zero_eof(
 	xfs_fsize_t	isize,		/* current inode size */
 	xfs_fsize_t	end_size)	/* terminal inode size */
 {
-	struct inode	*ip = LINVFS_GET_IP(vp);
+	struct inode	*ip = vn_to_inode(vp);
 	xfs_fileoff_t	start_zero_fsb;
 	xfs_fileoff_t	end_zero_fsb;
 	xfs_fileoff_t	zero_count_fsb;
@@ -573,7 +576,7 @@ xfs_write(
 	vrwlock_t		locktype;
 	size_t			ocount = 0, count;
 	loff_t			pos;
-	int			need_isem = 1, need_flush = 0;
+	int			need_i_mutex = 1, need_flush = 0;
 
 	XFS_STATS_INC(xs_write_calls);
 
@@ -622,14 +625,14 @@ xfs_write(
 			return XFS_ERROR(-EINVAL);
 
 		if (!VN_CACHED(vp) && pos < i_size_read(inode))
-			need_isem = 0;
+			need_i_mutex = 0;
 
 		if (VN_CACHED(vp))
 			need_flush = 1;
 	}
 
 relock:
-	if (need_isem) {
+	if (need_i_mutex) {
 		iolock = XFS_IOLOCK_EXCL;
 		locktype = VRWLOCK_WRITE;
 
@@ -651,7 +654,7 @@ start:
 					S_ISBLK(inode->i_mode));
 	if (error) {
 		xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-		goto out_unlock_isem;
+		goto out_unlock_mutex;
 	}
 
 	new_size = pos + count;
@@ -663,7 +666,7 @@ start:
 		loff_t		savedsize = pos;
 		int		dmflags = FILP_DELAY_FLAG(file);
 
-		if (need_isem)
+		if (need_i_mutex)
 			dmflags |= DM_FLAGS_IMUX;
 
 		xfs_iunlock(xip, XFS_ILOCK_EXCL);
@@ -672,13 +675,13 @@ start:
 				      dmflags, &locktype);
 		if (error) {
 			xfs_iunlock(xip, iolock);
-			goto out_unlock_isem;
+			goto out_unlock_mutex;
 		}
 		xfs_ilock(xip, XFS_ILOCK_EXCL);
 		eventsent = 1;
 
 		/*
-		 * The iolock was dropped and reaquired in XFS_SEND_DATA
+		 * The iolock was dropped and reacquired in XFS_SEND_DATA
 		 * so we have to recheck the size when appending.
 		 * We will only "goto start;" once, since having sent the
 		 * event prevents another call to XFS_SEND_DATA, which is
@@ -710,7 +713,7 @@ start:
 					isize, pos + count);
 		if (error) {
 			xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock);
-			goto out_unlock_isem;
+			goto out_unlock_mutex;
 		}
 	}
 	xfs_iunlock(xip, XFS_ILOCK_EXCL);
@@ -731,7 +734,7 @@ start:
 			error = -remove_suid(file->f_dentry);
 		if (unlikely(error)) {
 			xfs_iunlock(xip, iolock);
-			goto out_unlock_isem;
+			goto out_unlock_mutex;
 		}
 	}
 
@@ -747,14 +750,14 @@ retry:
 					-1, FI_REMAPF_LOCKED);
 		}
 
-		if (need_isem) {
+		if (need_i_mutex) {
 			/* demote the lock now the cached pages are gone */
 			XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL);
 			mutex_unlock(&inode->i_mutex);
 
 			iolock = XFS_IOLOCK_SHARED;
 			locktype = VRWLOCK_WRITE_DIRECT;
-			need_isem = 0;
+			need_i_mutex = 0;
 		}
 
  		xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs,
@@ -772,7 +775,7 @@ retry:
 			pos += ret;
 			count -= ret;
 
-			need_isem = 1;
+			need_i_mutex = 1;
 			ioflags &= ~IO_ISDIRECT;
 			xfs_iunlock(xip, iolock);
 			goto relock;
@@ -794,14 +797,14 @@ retry:
 	    !(ioflags & IO_INVIS)) {
 
 		xfs_rwunlock(bdp, locktype);
-		if (need_isem)
+		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
 		error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp,
 				DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL,
 				0, 0, 0); /* Delay flag intentionally  unused */
 		if (error)
 			goto out_nounlocks;
-		if (need_isem)
+		if (need_i_mutex)
 			mutex_lock(&inode->i_mutex);
 		xfs_rwlock(bdp, locktype);
 		pos = xip->i_d.di_size;
@@ -905,9 +908,9 @@ retry:
 			if (error)
 				goto out_unlock_internal;
 		}
-	
+
 		xfs_rwunlock(bdp, locktype);
-		if (need_isem)
+		if (need_i_mutex)
 			mutex_unlock(&inode->i_mutex);
 
 		error = sync_page_range(inode, mapping, pos, ret);
@@ -918,8 +921,8 @@ retry:
 
  out_unlock_internal:
 	xfs_rwunlock(bdp, locktype);
- out_unlock_isem:
-	if (need_isem)
+ out_unlock_mutex:
+	if (need_i_mutex)
 		mutex_unlock(&inode->i_mutex);
  out_nounlocks:
 	return -error;
diff --git a/fs/xfs/linux-2.6/xfs_stats.c b/fs/xfs/linux-2.6/xfs_stats.c
index 8955720a2c6b..1f0589a05eca 100644
--- a/fs/xfs/linux-2.6/xfs_stats.c
+++ b/fs/xfs/linux-2.6/xfs_stats.c
@@ -56,24 +56,21 @@ xfs_read_xfsstats(
 	};
 
 	/* Loop over all stats groups */
-	for (i=j=len = 0; i < sizeof(xstats)/sizeof(struct xstats_entry); i++) {
+	for (i=j=len = 0; i < ARRAY_SIZE(xstats); i++) {
 		len += sprintf(buffer + len, xstats[i].desc);
 		/* inner loop does each group */
 		while (j < xstats[i].endpoint) {
 			val = 0;
 			/* sum over all cpus */
-			for (c = 0; c < NR_CPUS; c++) {
-				if (!cpu_possible(c)) continue;
+			for_each_cpu(c)
 				val += *(((__u32*)&per_cpu(xfsstats, c) + j));
-			}
 			len += sprintf(buffer + len, " %u", val);
 			j++;
 		}
 		buffer[len++] = '\n';
 	}
 	/* extra precision counters */
-	for (i = 0; i < NR_CPUS; i++) {
-		if (!cpu_possible(i)) continue;
+	for_each_cpu(i) {
 		xs_xstrat_bytes += per_cpu(xfsstats, i).xs_xstrat_bytes;
 		xs_write_bytes += per_cpu(xfsstats, i).xs_write_bytes;
 		xs_read_bytes += per_cpu(xfsstats, i).xs_read_bytes;
diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c
index f22e426d9e42..1884300417e3 100644
--- a/fs/xfs/linux-2.6/xfs_super.c
+++ b/fs/xfs/linux-2.6/xfs_super.c
@@ -59,8 +59,8 @@
 #include <linux/writeback.h>
 #include <linux/kthread.h>
 
-STATIC struct quotactl_ops linvfs_qops;
-STATIC struct super_operations linvfs_sops;
+STATIC struct quotactl_ops xfs_quotactl_operations;
+STATIC struct super_operations xfs_super_operations;
 STATIC kmem_zone_t *xfs_vnode_zone;
 STATIC kmem_zone_t *xfs_ioend_zone;
 mempool_t *xfs_ioend_pool;
@@ -76,8 +76,6 @@ xfs_args_allocate(
 	strncpy(args->fsname, sb->s_id, MAXNAMELEN);
 
 	/* Copy the already-parsed mount(2) flags we're interested in */
-	if (sb->s_flags & MS_NOATIME)
-		args->flags |= XFSMNT_NOATIME;
 	if (sb->s_flags & MS_DIRSYNC)
 		args->flags |= XFSMNT_DIRSYNC;
 	if (sb->s_flags & MS_SYNCHRONOUS)
@@ -129,21 +127,21 @@ xfs_set_inodeops(
 {
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFREG:
-		inode->i_op = &linvfs_file_inode_operations;
-		inode->i_fop = &linvfs_file_operations;
-		inode->i_mapping->a_ops = &linvfs_aops;
+		inode->i_op = &xfs_inode_operations;
+		inode->i_fop = &xfs_file_operations;
+		inode->i_mapping->a_ops = &xfs_address_space_operations;
 		break;
 	case S_IFDIR:
-		inode->i_op = &linvfs_dir_inode_operations;
-		inode->i_fop = &linvfs_dir_operations;
+		inode->i_op = &xfs_dir_inode_operations;
+		inode->i_fop = &xfs_dir_file_operations;
 		break;
 	case S_IFLNK:
-		inode->i_op = &linvfs_symlink_inode_operations;
+		inode->i_op = &xfs_symlink_inode_operations;
 		if (inode->i_blocks)
-			inode->i_mapping->a_ops = &linvfs_aops;
+			inode->i_mapping->a_ops = &xfs_address_space_operations;
 		break;
 	default:
-		inode->i_op = &linvfs_file_inode_operations;
+		inode->i_op = &xfs_inode_operations;
 		init_special_inode(inode, inode->i_mode, inode->i_rdev);
 		break;
 	}
@@ -155,7 +153,7 @@ xfs_revalidate_inode(
 	vnode_t			*vp,
 	xfs_inode_t		*ip)
 {
-	struct inode		*inode = LINVFS_GET_IP(vp);
+	struct inode		*inode = vn_to_inode(vp);
 
 	inode->i_mode	= ip->i_d.di_mode;
 	inode->i_nlink	= ip->i_d.di_nlink;
@@ -212,7 +210,7 @@ xfs_initialize_vnode(
 	int			unlock)
 {
 	xfs_inode_t		*ip = XFS_BHVTOI(inode_bhv);
-	struct inode		*inode = LINVFS_GET_IP(vp);
+	struct inode		*inode = vn_to_inode(vp);
 
 	if (!inode_bhv->bd_vobj) {
 		vp->v_vfsp = bhvtovfs(bdp);
@@ -230,7 +228,7 @@ xfs_initialize_vnode(
 	if (ip->i_d.di_mode != 0 && unlock && (inode->i_state & I_NEW)) {
 		xfs_revalidate_inode(XFS_BHVTOM(bdp), vp, ip);
 		xfs_set_inodeops(inode);
-	
+
 		ip->i_flags &= ~XFS_INEW;
 		barrier();
 
@@ -334,43 +332,42 @@ xfs_blkdev_issue_flush(
 }
 
 STATIC struct inode *
-linvfs_alloc_inode(
+xfs_fs_alloc_inode(
 	struct super_block	*sb)
 {
 	vnode_t			*vp;
 
-	vp = kmem_cache_alloc(xfs_vnode_zone, kmem_flags_convert(KM_SLEEP));
-	if (!vp)
+	vp = kmem_zone_alloc(xfs_vnode_zone, KM_SLEEP);
+	if (unlikely(!vp))
 		return NULL;
-	return LINVFS_GET_IP(vp);
+	return vn_to_inode(vp);
 }
 
 STATIC void
-linvfs_destroy_inode(
+xfs_fs_destroy_inode(
 	struct inode		*inode)
 {
-	kmem_zone_free(xfs_vnode_zone, LINVFS_GET_VP(inode));
+	kmem_zone_free(xfs_vnode_zone, vn_from_inode(inode));
 }
 
 STATIC void
-linvfs_inode_init_once(
-	void			*data,
-	kmem_cache_t		*cachep,
+xfs_fs_inode_init_once(
+	void			*vnode,
+	kmem_zone_t		*zonep,
 	unsigned long		flags)
 {
-	vnode_t			*vp = (vnode_t *)data;
-
 	if ((flags & (SLAB_CTOR_VERIFY|SLAB_CTOR_CONSTRUCTOR)) ==
-	    SLAB_CTOR_CONSTRUCTOR)
-		inode_init_once(LINVFS_GET_IP(vp));
+		      SLAB_CTOR_CONSTRUCTOR)
+		inode_init_once(vn_to_inode((vnode_t *)vnode));
 }
 
 STATIC int
-linvfs_init_zones(void)
+xfs_init_zones(void)
 {
-	xfs_vnode_zone = kmem_cache_create("xfs_vnode",
-				sizeof(vnode_t), 0, SLAB_RECLAIM_ACCOUNT,
-				linvfs_inode_init_once, NULL);
+	xfs_vnode_zone = kmem_zone_init_flags(sizeof(vnode_t), "xfs_vnode_t",
+					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
+					KM_ZONE_SPREAD,
+					xfs_fs_inode_init_once);
 	if (!xfs_vnode_zone)
 		goto out;
 
@@ -378,15 +375,12 @@ linvfs_init_zones(void)
 	if (!xfs_ioend_zone)
 		goto out_destroy_vnode_zone;
 
-	xfs_ioend_pool = mempool_create(4 * MAX_BUF_PER_PAGE,
-			mempool_alloc_slab, mempool_free_slab,
-			xfs_ioend_zone);
+	xfs_ioend_pool = mempool_create_slab_pool(4 * MAX_BUF_PER_PAGE,
+						  xfs_ioend_zone);
 	if (!xfs_ioend_pool)
 		goto out_free_ioend_zone;
-
 	return 0;
 
-
  out_free_ioend_zone:
 	kmem_zone_destroy(xfs_ioend_zone);
  out_destroy_vnode_zone:
@@ -396,7 +390,7 @@ linvfs_init_zones(void)
 }
 
 STATIC void
-linvfs_destroy_zones(void)
+xfs_destroy_zones(void)
 {
 	mempool_destroy(xfs_ioend_pool);
 	kmem_zone_destroy(xfs_vnode_zone);
@@ -407,14 +401,14 @@ linvfs_destroy_zones(void)
  * Attempt to flush the inode, this will actually fail
  * if the inode is pinned, but we dirty the inode again
  * at the point when it is unpinned after a log write,
- * since this is when the inode itself becomes flushable. 
+ * since this is when the inode itself becomes flushable.
  */
 STATIC int
-linvfs_write_inode(
+xfs_fs_write_inode(
 	struct inode		*inode,
 	int			sync)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error = 0, flags = FLUSH_INODE;
 
 	if (vp) {
@@ -434,13 +428,13 @@ linvfs_write_inode(
 }
 
 STATIC void
-linvfs_clear_inode(
+xfs_fs_clear_inode(
 	struct inode		*inode)
 {
-	vnode_t			*vp = LINVFS_GET_VP(inode);
+	vnode_t			*vp = vn_from_inode(inode);
 	int			error, cache;
 
-	vn_trace_entry(vp, "clear_inode", (inst_t *)__return_address);
+	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
 
 	XFS_STATS_INC(vn_rele);
 	XFS_STATS_INC(vn_remove);
@@ -516,7 +510,7 @@ void
 xfs_flush_inode(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 	struct vfs	*vfs = XFS_MTOVFS(ip->i_mount);
 
 	igrab(inode);
@@ -541,7 +535,7 @@ void
 xfs_flush_device(
 	xfs_inode_t	*ip)
 {
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 	struct vfs	*vfs = XFS_MTOVFS(ip->i_mount);
 
 	igrab(inode);
@@ -550,7 +544,7 @@ xfs_flush_device(
 	xfs_log_force(ip->i_mount, (xfs_lsn_t)0, XFS_LOG_FORCE|XFS_LOG_SYNC);
 }
 
-#define SYNCD_FLAGS	(SYNC_FSDATA|SYNC_BDFLUSH|SYNC_ATTR)
+#define SYNCD_FLAGS	(SYNC_FSDATA|SYNC_BDFLUSH|SYNC_ATTR|SYNC_REFCACHE)
 STATIC void
 vfs_sync_worker(
 	vfs_t		*vfsp,
@@ -613,7 +607,7 @@ xfssyncd(
 }
 
 STATIC int
-linvfs_start_syncd(
+xfs_fs_start_syncd(
 	vfs_t			*vfsp)
 {
 	vfsp->vfs_sync_work.w_syncer = vfs_sync_worker;
@@ -625,20 +619,20 @@ linvfs_start_syncd(
 }
 
 STATIC void
-linvfs_stop_syncd(
+xfs_fs_stop_syncd(
 	vfs_t			*vfsp)
 {
 	kthread_stop(vfsp->vfs_sync_task);
 }
 
 STATIC void
-linvfs_put_super(
+xfs_fs_put_super(
 	struct super_block	*sb)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
-	linvfs_stop_syncd(vfsp);
+	xfs_fs_stop_syncd(vfsp);
 	VFS_SYNC(vfsp, SYNC_ATTR|SYNC_DELWRI, NULL, error);
 	if (!error)
 		VFS_UNMOUNT(vfsp, 0, NULL, error);
@@ -652,10 +646,10 @@ linvfs_put_super(
 }
 
 STATIC void
-linvfs_write_super(
+xfs_fs_write_super(
 	struct super_block	*sb)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	if (sb->s_flags & MS_RDONLY) {
@@ -668,11 +662,11 @@ linvfs_write_super(
 }
 
 STATIC int
-linvfs_sync_super(
+xfs_fs_sync_super(
 	struct super_block	*sb,
 	int			wait)
 {
-	vfs_t		*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t		*vfsp = vfs_from_sb(sb);
 	int		error;
 	int		flags = SYNC_FSDATA;
 
@@ -707,11 +701,11 @@ linvfs_sync_super(
 }
 
 STATIC int
-linvfs_statfs(
+xfs_fs_statfs(
 	struct super_block	*sb,
 	struct kstatfs		*statp)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_STATVFS(vfsp, statp, NULL, error);
@@ -719,12 +713,12 @@ linvfs_statfs(
 }
 
 STATIC int
-linvfs_remount(
+xfs_fs_remount(
 	struct super_block	*sb,
 	int			*flags,
 	char			*options)
 {
-	vfs_t			*vfsp = LINVFS_GET_VFS(sb);
+	vfs_t			*vfsp = vfs_from_sb(sb);
 	struct xfs_mount_args	*args = xfs_args_allocate(sb);
 	int			error;
 
@@ -736,18 +730,18 @@ linvfs_remount(
 }
 
 STATIC void
-linvfs_freeze_fs(
+xfs_fs_lockfs(
 	struct super_block	*sb)
 {
-	VFS_FREEZE(LINVFS_GET_VFS(sb));
+	VFS_FREEZE(vfs_from_sb(sb));
 }
 
 STATIC int
-linvfs_show_options(
+xfs_fs_show_options(
 	struct seq_file		*m,
 	struct vfsmount		*mnt)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(mnt->mnt_sb);
+	struct vfs		*vfsp = vfs_from_sb(mnt->mnt_sb);
 	int			error;
 
 	VFS_SHOWARGS(vfsp, m, error);
@@ -755,11 +749,11 @@ linvfs_show_options(
 }
 
 STATIC int
-linvfs_quotasync(
+xfs_fs_quotasync(
 	struct super_block	*sb,
 	int			type)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_QUOTACTL(vfsp, Q_XQUOTASYNC, 0, (caddr_t)NULL, error);
@@ -767,11 +761,11 @@ linvfs_quotasync(
 }
 
 STATIC int
-linvfs_getxstate(
+xfs_fs_getxstate(
 	struct super_block	*sb,
 	struct fs_quota_stat	*fqs)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_QUOTACTL(vfsp, Q_XGETQSTAT, 0, (caddr_t)fqs, error);
@@ -779,12 +773,12 @@ linvfs_getxstate(
 }
 
 STATIC int
-linvfs_setxstate(
+xfs_fs_setxstate(
 	struct super_block	*sb,
 	unsigned int		flags,
 	int			op)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error;
 
 	VFS_QUOTACTL(vfsp, op, 0, (caddr_t)&flags, error);
@@ -792,13 +786,13 @@ linvfs_setxstate(
 }
 
 STATIC int
-linvfs_getxquota(
+xfs_fs_getxquota(
 	struct super_block	*sb,
 	int			type,
 	qid_t			id,
 	struct fs_disk_quota	*fdq)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error, getmode;
 
 	getmode = (type == USRQUOTA) ? Q_XGETQUOTA :
@@ -808,13 +802,13 @@ linvfs_getxquota(
 }
 
 STATIC int
-linvfs_setxquota(
+xfs_fs_setxquota(
 	struct super_block	*sb,
 	int			type,
 	qid_t			id,
 	struct fs_disk_quota	*fdq)
 {
-	struct vfs		*vfsp = LINVFS_GET_VFS(sb);
+	struct vfs		*vfsp = vfs_from_sb(sb);
 	int			error, setmode;
 
 	setmode = (type == USRQUOTA) ? Q_XSETQLIM :
@@ -824,21 +818,17 @@ linvfs_setxquota(
 }
 
 STATIC int
-linvfs_fill_super(
+xfs_fs_fill_super(
 	struct super_block	*sb,
 	void			*data,
 	int			silent)
 {
 	vnode_t			*rootvp;
-	struct vfs		*vfsp = vfs_allocate();
+	struct vfs		*vfsp = vfs_allocate(sb);
 	struct xfs_mount_args	*args = xfs_args_allocate(sb);
 	struct kstatfs		statvfs;
 	int			error, error2;
 
-	vfsp->vfs_super = sb;
-	LINVFS_SET_VFS(sb, vfsp);
-	if (sb->s_flags & MS_RDONLY)
-		vfsp->vfs_flag |= VFS_RDONLY;
 	bhv_insert_all_vfsops(vfsp);
 
 	VFS_PARSEARGS(vfsp, (char *)data, args, 0, error);
@@ -849,10 +839,10 @@ linvfs_fill_super(
 
 	sb_min_blocksize(sb, BBSIZE);
 #ifdef CONFIG_XFS_EXPORT
-	sb->s_export_op = &linvfs_export_ops;
+	sb->s_export_op = &xfs_export_operations;
 #endif
-	sb->s_qcop = &linvfs_qops;
-	sb->s_op = &linvfs_sops;
+	sb->s_qcop = &xfs_quotactl_operations;
+	sb->s_op = &xfs_super_operations;
 
 	VFS_MOUNT(vfsp, args, NULL, error);
 	if (error) {
@@ -876,7 +866,7 @@ linvfs_fill_super(
 	if (error)
 		goto fail_unmount;
 
-	sb->s_root = d_alloc_root(LINVFS_GET_IP(rootvp));
+	sb->s_root = d_alloc_root(vn_to_inode(rootvp));
 	if (!sb->s_root) {
 		error = ENOMEM;
 		goto fail_vnrele;
@@ -885,7 +875,7 @@ linvfs_fill_super(
 		error = EINVAL;
 		goto fail_vnrele;
 	}
-	if ((error = linvfs_start_syncd(vfsp)))
+	if ((error = xfs_fs_start_syncd(vfsp)))
 		goto fail_vnrele;
 	vn_trace_exit(rootvp, __FUNCTION__, (inst_t *)__return_address);
 
@@ -910,41 +900,41 @@ fail_vfsop:
 }
 
 STATIC struct super_block *
-linvfs_get_sb(
+xfs_fs_get_sb(
 	struct file_system_type	*fs_type,
 	int			flags,
 	const char		*dev_name,
 	void			*data)
 {
-	return get_sb_bdev(fs_type, flags, dev_name, data, linvfs_fill_super);
-}
-
-STATIC struct super_operations linvfs_sops = {
-	.alloc_inode		= linvfs_alloc_inode,
-	.destroy_inode		= linvfs_destroy_inode,
-	.write_inode		= linvfs_write_inode,
-	.clear_inode		= linvfs_clear_inode,
-	.put_super		= linvfs_put_super,
-	.write_super		= linvfs_write_super,
-	.sync_fs		= linvfs_sync_super,
-	.write_super_lockfs	= linvfs_freeze_fs,
-	.statfs			= linvfs_statfs,
-	.remount_fs		= linvfs_remount,
-	.show_options		= linvfs_show_options,
+	return get_sb_bdev(fs_type, flags, dev_name, data, xfs_fs_fill_super);
+}
+
+STATIC struct super_operations xfs_super_operations = {
+	.alloc_inode		= xfs_fs_alloc_inode,
+	.destroy_inode		= xfs_fs_destroy_inode,
+	.write_inode		= xfs_fs_write_inode,
+	.clear_inode		= xfs_fs_clear_inode,
+	.put_super		= xfs_fs_put_super,
+	.write_super		= xfs_fs_write_super,
+	.sync_fs		= xfs_fs_sync_super,
+	.write_super_lockfs	= xfs_fs_lockfs,
+	.statfs			= xfs_fs_statfs,
+	.remount_fs		= xfs_fs_remount,
+	.show_options		= xfs_fs_show_options,
 };
 
-STATIC struct quotactl_ops linvfs_qops = {
-	.quota_sync		= linvfs_quotasync,
-	.get_xstate		= linvfs_getxstate,
-	.set_xstate		= linvfs_setxstate,
-	.get_xquota		= linvfs_getxquota,
-	.set_xquota		= linvfs_setxquota,
+STATIC struct quotactl_ops xfs_quotactl_operations = {
+	.quota_sync		= xfs_fs_quotasync,
+	.get_xstate		= xfs_fs_getxstate,
+	.set_xstate		= xfs_fs_setxstate,
+	.get_xquota		= xfs_fs_getxquota,
+	.set_xquota		= xfs_fs_setxquota,
 };
 
 STATIC struct file_system_type xfs_fs_type = {
 	.owner			= THIS_MODULE,
 	.name			= "xfs",
-	.get_sb			= linvfs_get_sb,
+	.get_sb			= xfs_fs_get_sb,
 	.kill_sb		= kill_block_super,
 	.fs_flags		= FS_REQUIRES_DEV,
 };
@@ -965,7 +955,7 @@ init_xfs_fs( void )
 
 	ktrace_init(64);
 
-	error = linvfs_init_zones();
+	error = xfs_init_zones();
 	if (error < 0)
 		goto undo_zones;
 
@@ -981,14 +971,13 @@ init_xfs_fs( void )
 	error = register_filesystem(&xfs_fs_type);
 	if (error)
 		goto undo_register;
-	XFS_DM_INIT(&xfs_fs_type);
 	return 0;
 
 undo_register:
 	xfs_buf_terminate();
 
 undo_buffers:
-	linvfs_destroy_zones();
+	xfs_destroy_zones();
 
 undo_zones:
 	return error;
@@ -998,11 +987,10 @@ STATIC void __exit
 exit_xfs_fs( void )
 {
 	vfs_exitquota();
-	XFS_DM_EXIT(&xfs_fs_type);
 	unregister_filesystem(&xfs_fs_type);
 	xfs_cleanup();
 	xfs_buf_terminate();
-	linvfs_destroy_zones();
+	xfs_destroy_zones();
 	ktrace_uninit();
 }
 
diff --git a/fs/xfs/linux-2.6/xfs_super.h b/fs/xfs/linux-2.6/xfs_super.h
index df59408dca06..376b96cb513a 100644
--- a/fs/xfs/linux-2.6/xfs_super.h
+++ b/fs/xfs/linux-2.6/xfs_super.h
@@ -98,11 +98,6 @@ extern void xfs_qm_exit(void);
 				XFS_DMAPI_STRING \
 				XFS_DBG_STRING /* DBG must be last */
 
-#define LINVFS_GET_VFS(s) \
-	(vfs_t *)((s)->s_fs_info)
-#define LINVFS_SET_VFS(s, vfsp) \
-	((s)->s_fs_info = vfsp)
-
 struct xfs_inode;
 struct xfs_mount;
 struct xfs_buftarg;
@@ -120,6 +115,6 @@ extern int  xfs_blkdev_get(struct xfs_mount *, const char *,
 extern void xfs_blkdev_put(struct block_device *);
 extern void xfs_blkdev_issue_flush(struct xfs_buftarg *);
 
-extern struct export_operations linvfs_export_ops;
+extern struct export_operations xfs_export_operations;
 
 #endif	/* __XFS_SUPER_H__ */
diff --git a/fs/xfs/linux-2.6/xfs_sysctl.c b/fs/xfs/linux-2.6/xfs_sysctl.c
index a02564972420..7079cc837210 100644
--- a/fs/xfs/linux-2.6/xfs_sysctl.c
+++ b/fs/xfs/linux-2.6/xfs_sysctl.c
@@ -38,8 +38,7 @@ xfs_stats_clear_proc_handler(
 
 	if (!ret && write && *valp) {
 		printk("XFS Clearing xfsstats\n");
-		for (c = 0; c < NR_CPUS; c++) {
-			if (!cpu_possible(c)) continue;
+		for_each_cpu(c) {
 			preempt_disable();
 			/* save vn_active, it's a universal truth! */
 			vn_active = per_cpu(xfsstats, c).vn_active;
diff --git a/fs/xfs/linux-2.6/xfs_vfs.c b/fs/xfs/linux-2.6/xfs_vfs.c
index c855d62e5344..6f7c9f7a8624 100644
--- a/fs/xfs/linux-2.6/xfs_vfs.c
+++ b/fs/xfs/linux-2.6/xfs_vfs.c
@@ -227,7 +227,8 @@ vfs_freeze(
 }
 
 vfs_t *
-vfs_allocate( void )
+vfs_allocate(
+	struct super_block	*sb)
 {
 	struct vfs		*vfsp;
 
@@ -236,9 +237,23 @@ vfs_allocate( void )
 	INIT_LIST_HEAD(&vfsp->vfs_sync_list);
 	spin_lock_init(&vfsp->vfs_sync_lock);
 	init_waitqueue_head(&vfsp->vfs_wait_single_sync_task);
+
+	vfsp->vfs_super = sb;
+	sb->s_fs_info = vfsp;
+
+	if (sb->s_flags & MS_RDONLY)
+		vfsp->vfs_flag |= VFS_RDONLY;
+
 	return vfsp;
 }
 
+vfs_t *
+vfs_from_sb(
+	struct super_block	*sb)
+{
+	return (vfs_t *)sb->s_fs_info;
+}
+
 void
 vfs_deallocate(
 	struct vfs		*vfsp)
@@ -295,7 +310,7 @@ bhv_remove_all_vfsops(
 	bhv_remove_vfsops(vfsp, VFS_POSITION_DM);
 	if (!freebase)
 		return;
-	mp = XFS_BHVTOM(bhv_lookup(VFS_BHVHEAD(vfsp), &xfs_vfsops));
+	mp = XFS_VFSTOM(vfsp);
 	VFS_REMOVEBHV(vfsp, &mp->m_bhv);
 	xfs_mount_free(mp, 0);
 }
diff --git a/fs/xfs/linux-2.6/xfs_vfs.h b/fs/xfs/linux-2.6/xfs_vfs.h
index 57caf9eddee0..841200c03092 100644
--- a/fs/xfs/linux-2.6/xfs_vfs.h
+++ b/fs/xfs/linux-2.6/xfs_vfs.h
@@ -92,7 +92,7 @@ typedef enum {
 #define SYNC_FSDATA		0x0020	/* flush fs data (e.g. superblocks) */
 #define SYNC_REFCACHE		0x0040  /* prune some of the nfs ref cache */
 #define SYNC_REMOUNT		0x0080  /* remount readonly, no dummy LRs */
-#define SYNC_QUIESCE		0x0100  /* quiesce fileystem for a snapshot */
+#define SYNC_QUIESCE		0x0100  /* quiesce filesystem for a snapshot */
 
 typedef int	(*vfs_mount_t)(bhv_desc_t *,
 				struct xfs_mount_args *, struct cred *);
@@ -193,7 +193,8 @@ typedef struct bhv_vfsops {
 #define vfs_bhv_set_custom(b,o)	( (b)->bhv_custom = (void *)(o))
 #define vfs_bhv_clr_custom(b)	( (b)->bhv_custom = NULL )
 
-extern vfs_t *vfs_allocate(void);
+extern vfs_t *vfs_allocate(struct super_block *);
+extern vfs_t *vfs_from_sb(struct super_block *);
 extern void vfs_deallocate(vfs_t *);
 extern void vfs_insertops(vfs_t *, bhv_vfsops_t *);
 extern void vfs_insertbhv(vfs_t *, bhv_desc_t *, vfsops_t *, void *);
diff --git a/fs/xfs/linux-2.6/xfs_vnode.c b/fs/xfs/linux-2.6/xfs_vnode.c
index 260dd8415dd7..d27c25b27ccd 100644
--- a/fs/xfs/linux-2.6/xfs_vnode.c
+++ b/fs/xfs/linux-2.6/xfs_vnode.c
@@ -58,7 +58,7 @@ struct vnode *
 vn_initialize(
 	struct inode	*inode)
 {
-	struct vnode	*vp = LINVFS_GET_VP(inode);
+	struct vnode	*vp = vn_from_inode(inode);
 
 	XFS_STATS_INC(vn_active);
 	XFS_STATS_INC(vn_alloc);
@@ -83,7 +83,7 @@ vn_initialize(
 	vp->v_trace = ktrace_alloc(VNODE_TRACE_SIZE, KM_SLEEP);
 #endif	/* XFS_VNODE_TRACE */
 
-	vn_trace_exit(vp, "vn_initialize", (inst_t *)__return_address);
+	vn_trace_exit(vp, __FUNCTION__, (inst_t *)__return_address);
 	return vp;
 }
 
@@ -97,7 +97,7 @@ vn_revalidate_core(
 	struct vnode	*vp,
 	vattr_t		*vap)
 {
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	inode->i_mode	    = vap->va_mode;
 	inode->i_nlink	    = vap->va_nlink;
@@ -129,24 +129,31 @@ vn_revalidate_core(
  * Revalidate the Linux inode from the vnode.
  */
 int
-vn_revalidate(
-	struct vnode	*vp)
+__vn_revalidate(
+	struct vnode	*vp,
+	struct vattr	*vattr)
 {
-	vattr_t		va;
 	int		error;
 
-	vn_trace_entry(vp, "vn_revalidate", (inst_t *)__return_address);
-	ASSERT(vp->v_fbhv != NULL);
-
-	va.va_mask = XFS_AT_STAT|XFS_AT_XFLAGS;
-	VOP_GETATTR(vp, &va, 0, NULL, error);
-	if (!error) {
-		vn_revalidate_core(vp, &va);
+	vn_trace_entry(vp, __FUNCTION__, (inst_t *)__return_address);
+	vattr->va_mask = XFS_AT_STAT | XFS_AT_XFLAGS;
+	VOP_GETATTR(vp, vattr, 0, NULL, error);
+	if (likely(!error)) {
+		vn_revalidate_core(vp, vattr);
 		VUNMODIFY(vp);
 	}
 	return -error;
 }
 
+int
+vn_revalidate(
+	struct vnode	*vp)
+{
+	vattr_t		vattr;
+
+	return __vn_revalidate(vp, &vattr);
+}
+
 /*
  * Add a reference to a referenced vnode.
  */
@@ -159,7 +166,7 @@ vn_hold(
 	XFS_STATS_INC(vn_hold);
 
 	VN_LOCK(vp);
-	inode = igrab(LINVFS_GET_IP(vp));
+	inode = igrab(vn_to_inode(vp));
 	ASSERT(inode);
 	VN_UNLOCK(vp, 0);
 
diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h
index 0fe2419461d6..06f5845e9568 100644
--- a/fs/xfs/linux-2.6/xfs_vnode.h
+++ b/fs/xfs/linux-2.6/xfs_vnode.h
@@ -116,8 +116,14 @@ typedef enum {
 /*
  * Vnode to Linux inode mapping.
  */
-#define LINVFS_GET_VP(inode)	((vnode_t *)list_entry(inode, vnode_t, v_inode))
-#define LINVFS_GET_IP(vp)	(&(vp)->v_inode)
+static inline struct vnode *vn_from_inode(struct inode *inode)
+{
+	return (vnode_t *)list_entry(inode, vnode_t, v_inode);
+}
+static inline struct inode *vn_to_inode(struct vnode *vnode)
+{
+	return &vnode->v_inode;
+}
 
 /*
  * Vnode flags.
@@ -490,6 +496,7 @@ typedef struct vnode_map {
 			 (vmap).v_ino	 = (vp)->v_inode.i_ino; }
 
 extern int	vn_revalidate(struct vnode *);
+extern int	__vn_revalidate(struct vnode *, vattr_t *);
 extern void	vn_revalidate_core(struct vnode *, vattr_t *);
 
 extern void	vn_iowait(struct vnode *vp);
@@ -497,7 +504,7 @@ extern void	vn_iowake(struct vnode *vp);
 
 static inline int vn_count(struct vnode *vp)
 {
-	return atomic_read(&LINVFS_GET_IP(vp)->i_count);
+	return atomic_read(&vn_to_inode(vp)->i_count);
 }
 
 /*
@@ -511,16 +518,16 @@ extern vnode_t	*vn_hold(struct vnode *);
 	  vn_trace_hold(vp, __FILE__, __LINE__, (inst_t *)__return_address))
 #define VN_RELE(vp)		\
 	  (vn_trace_rele(vp, __FILE__, __LINE__, (inst_t *)__return_address), \
-	   iput(LINVFS_GET_IP(vp)))
+	   iput(vn_to_inode(vp)))
 #else
 #define VN_HOLD(vp)		((void)vn_hold(vp))
-#define VN_RELE(vp)		(iput(LINVFS_GET_IP(vp)))
+#define VN_RELE(vp)		(iput(vn_to_inode(vp)))
 #endif
 
 static inline struct vnode *vn_grab(struct vnode *vp)
 {
-	struct inode *inode = igrab(LINVFS_GET_IP(vp));
-	return inode ? LINVFS_GET_VP(inode) : NULL;
+	struct inode *inode = igrab(vn_to_inode(vp));
+	return inode ? vn_from_inode(inode) : NULL;
 }
 
 /*
@@ -528,7 +535,7 @@ static inline struct vnode *vn_grab(struct vnode *vp)
  */
 #define VNAME(dentry)		((char *) (dentry)->d_name.name)
 #define VNAMELEN(dentry)	((dentry)->d_name.len)
-#define VNAME_TO_VNODE(dentry)	(LINVFS_GET_VP((dentry)->d_inode))
+#define VNAME_TO_VNODE(dentry)	(vn_from_inode((dentry)->d_inode))
 
 /*
  * Vnode spinlock manipulation.
@@ -557,12 +564,12 @@ static __inline__ void vn_flagclr(struct vnode *vp, uint flag)
  */
 static inline void vn_mark_bad(struct vnode *vp)
 {
-	make_bad_inode(LINVFS_GET_IP(vp));
+	make_bad_inode(vn_to_inode(vp));
 }
 
 static inline int VN_BAD(struct vnode *vp)
 {
-	return is_bad_inode(LINVFS_GET_IP(vp));
+	return is_bad_inode(vn_to_inode(vp));
 }
 
 /*
@@ -587,9 +594,9 @@ static inline void vn_atime_to_time_t(struct vnode *vp, time_t *tt)
 /*
  * Some useful predicates.
  */
-#define VN_MAPPED(vp)	mapping_mapped(LINVFS_GET_IP(vp)->i_mapping)
-#define VN_CACHED(vp)	(LINVFS_GET_IP(vp)->i_mapping->nrpages)
-#define VN_DIRTY(vp)	mapping_tagged(LINVFS_GET_IP(vp)->i_mapping, \
+#define VN_MAPPED(vp)	mapping_mapped(vn_to_inode(vp)->i_mapping)
+#define VN_CACHED(vp)	(vn_to_inode(vp)->i_mapping->nrpages)
+#define VN_DIRTY(vp)	mapping_tagged(vn_to_inode(vp)->i_mapping, \
 					PAGECACHE_TAG_DIRTY)
 #define VMODIFY(vp)	VN_FLAGSET(vp, VMODIFIED)
 #define VUNMODIFY(vp)	VN_FLAGCLR(vp, VMODIFIED)
diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c
index 2ec6b441849c..546f48af882a 100644
--- a/fs/xfs/quota/xfs_dquot_item.c
+++ b/fs/xfs/quota/xfs_dquot_item.c
@@ -79,9 +79,11 @@ xfs_qm_dquot_logitem_format(
 
 	logvec->i_addr = (xfs_caddr_t)&logitem->qli_format;
 	logvec->i_len  = sizeof(xfs_dq_logformat_t);
+	XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_QFORMAT);
 	logvec++;
 	logvec->i_addr = (xfs_caddr_t)&logitem->qli_dquot->q_core;
 	logvec->i_len  = sizeof(xfs_disk_dquot_t);
+	XLOG_VEC_SET_TYPE(logvec, XLOG_REG_TYPE_DQUOT);
 
 	ASSERT(2 == logitem->qli_item.li_desc->lid_size);
 	logitem->qli_format.qlf_size = 2;
@@ -219,7 +221,7 @@ xfs_qm_dqunpin_wait(
  * as possible.
  *
  * We must not be holding the AIL_LOCK at this point. Calling incore() to
- * search the buffercache can be a time consuming thing, and AIL_LOCK is a
+ * search the buffer cache can be a time consuming thing, and AIL_LOCK is a
  * spinlock.
  */
 STATIC void
diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c
index 7c0e39dc6189..73c1e5e80c07 100644
--- a/fs/xfs/quota/xfs_qm.c
+++ b/fs/xfs/quota/xfs_qm.c
@@ -289,7 +289,7 @@ xfs_qm_rele_quotafs_ref(
 
 /*
  * This is called at mount time from xfs_mountfs to initialize the quotainfo
- * structure and start the global quotamanager (xfs_Gqm) if it hasn't done
+ * structure and start the global quota manager (xfs_Gqm) if it hasn't done
  * so already.	Note that the superblock has not been read in yet.
  */
 void
@@ -807,7 +807,7 @@ xfs_qm_dqattach_one(
  * Given a udquot and gdquot, attach a ptr to the group dquot in the
  * udquot as a hint for future lookups. The idea sounds simple, but the
  * execution isn't, because the udquot might have a group dquot attached
- * already and getting rid of that gets us into lock ordering contraints.
+ * already and getting rid of that gets us into lock ordering constraints.
  * The process is complicated more by the fact that the dquots may or may not
  * be locked on entry.
  */
@@ -1094,10 +1094,10 @@ xfs_qm_sync(
 			}
 			/*
 			 * If we can't grab the flush lock then if the caller
-			 * really wanted us to give this our best shot,
+			 * really wanted us to give this our best shot, so
 			 * see if we can give a push to the buffer before we wait
 			 * on the flush lock. At this point, we know that
-			 * eventhough the dquot is being flushed,
+			 * even though the dquot is being flushed,
 			 * it has (new) dirty data.
 			 */
 			xfs_qm_dqflock_pushbuf_wait(dqp);
@@ -1491,7 +1491,7 @@ xfs_qm_reset_dqcounts(
 		/*
 		 * Do a sanity check, and if needed, repair the dqblk. Don't
 		 * output any warnings because it's perfectly possible to
-		 * find unitialized dquot blks. See comment in xfs_qm_dqcheck.
+		 * find uninitialised dquot blks. See comment in xfs_qm_dqcheck.
 		 */
 		(void) xfs_qm_dqcheck(ddq, id+j, type, XFS_QMOPT_DQREPAIR,
 				      "xfs_quotacheck");
@@ -1580,7 +1580,7 @@ xfs_qm_dqiterate(
 
 	error = 0;
 	/*
-	 * This looks racey, but we can't keep an inode lock across a
+	 * This looks racy, but we can't keep an inode lock across a
 	 * trans_reserve. But, this gets called during quotacheck, and that
 	 * happens only at mount time which is single threaded.
 	 */
@@ -1704,9 +1704,9 @@ xfs_qm_get_rtblks(
 	xfs_qcnt_t	*O_rtblks)
 {
 	xfs_filblks_t	rtblks;			/* total rt blks */
+	xfs_extnum_t	idx;			/* extent record index */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
 	xfs_extnum_t	nextents;		/* number of extent entries */
-	xfs_bmbt_rec_t	*base;			/* base of extent array */
 	xfs_bmbt_rec_t	*ep;			/* pointer to an extent entry */
 	int		error;
 
@@ -1717,10 +1717,11 @@ xfs_qm_get_rtblks(
 			return error;
 	}
 	rtblks = 0;
-	nextents = ifp->if_bytes / sizeof(xfs_bmbt_rec_t);
-	base = &ifp->if_u1.if_extents[0];
-	for (ep = base; ep < &base[nextents]; ep++)
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	for (idx = 0; idx < nextents; idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		rtblks += xfs_bmbt_get_blockcount(ep);
+	}
 	*O_rtblks = (xfs_qcnt_t)rtblks;
 	return 0;
 }
@@ -1823,7 +1824,7 @@ xfs_qm_dqusage_adjust(
 	 * we have to start from the beginning anyway.
 	 * Once we're done, we'll log all the dquot bufs.
 	 *
-	 * The *QUOTA_ON checks below may look pretty racey, but quotachecks
+	 * The *QUOTA_ON checks below may look pretty racy, but quotachecks
 	 * and quotaoffs don't race. (Quotachecks happen at mount time only).
 	 */
 	if (XFS_IS_UQUOTA_ON(mp)) {
@@ -2788,9 +2789,7 @@ xfs_qm_freelist_destroy(xfs_frlist_t *ql)
 		xfs_qm_dqdestroy(dqp);
 		dqp = nextdqp;
 	}
-	/*
-	 * Don't bother about unlocking.
-	 */
+	mutex_unlock(&ql->qh_lock);
 	mutex_destroy(&ql->qh_lock);
 
 	ASSERT(ql->qh_nelems == 0);
diff --git a/fs/xfs/quota/xfs_qm_bhv.c b/fs/xfs/quota/xfs_qm_bhv.c
index 90402a1c3983..6838b36d95a9 100644
--- a/fs/xfs/quota/xfs_qm_bhv.c
+++ b/fs/xfs/quota/xfs_qm_bhv.c
@@ -374,7 +374,7 @@ xfs_qm_exit(void)
 	vfs_bhv_clr_custom(&xfs_qmops);
 	xfs_qm_cleanup_procfs();
 	if (qm_dqzone)
-		kmem_cache_destroy(qm_dqzone);
+		kmem_zone_destroy(qm_dqzone);
 	if (qm_dqtrxzone)
-		kmem_cache_destroy(qm_dqtrxzone);
+		kmem_zone_destroy(qm_dqtrxzone);
 }
diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c
index 676884394aae..c55db463bbf2 100644
--- a/fs/xfs/quota/xfs_qm_syscalls.c
+++ b/fs/xfs/quota/xfs_qm_syscalls.c
@@ -912,7 +912,7 @@ xfs_qm_export_dquot(
 
 	/*
 	 * Internally, we don't reset all the timers when quota enforcement
-	 * gets turned off. No need to confuse the userlevel code,
+	 * gets turned off. No need to confuse the user level code,
 	 * so return zeroes in that case.
 	 */
 	if (! XFS_IS_QUOTA_ENFORCED(mp)) {
diff --git a/fs/xfs/quota/xfs_trans_dquot.c b/fs/xfs/quota/xfs_trans_dquot.c
index 3290975d31f7..d8e131ec0aa8 100644
--- a/fs/xfs/quota/xfs_trans_dquot.c
+++ b/fs/xfs/quota/xfs_trans_dquot.c
@@ -804,7 +804,7 @@ xfs_trans_reserve_quota_bydquots(
 	}
 
 	/*
-	 * Didnt change anything critical, so, no need to log
+	 * Didn't change anything critical, so, no need to log
 	 */
 	return (0);
 }
diff --git a/fs/xfs/support/ktrace.c b/fs/xfs/support/ktrace.c
index 841aa4c15b8a..addf5a7ea06c 100644
--- a/fs/xfs/support/ktrace.c
+++ b/fs/xfs/support/ktrace.c
@@ -39,8 +39,8 @@ ktrace_init(int zentries)
 void
 ktrace_uninit(void)
 {
-	kmem_cache_destroy(ktrace_hdr_zone);
-	kmem_cache_destroy(ktrace_ent_zone);
+	kmem_zone_destroy(ktrace_hdr_zone);
+	kmem_zone_destroy(ktrace_ent_zone);
 }
 
 /*
diff --git a/fs/xfs/support/uuid.c b/fs/xfs/support/uuid.c
index a3d565a67734..e157015c70ff 100644
--- a/fs/xfs/support/uuid.c
+++ b/fs/xfs/support/uuid.c
@@ -21,13 +21,6 @@ static mutex_t	uuid_monitor;
 static int	uuid_table_size;
 static uuid_t	*uuid_table;
 
-void
-uuid_init(void)
-{
-	mutex_init(&uuid_monitor);
-}
-
-
 /* IRIX interpretation of an uuid_t */
 typedef struct {
 	__be32	uu_timelow;
@@ -50,7 +43,7 @@ uuid_getnodeuniq(uuid_t *uuid, int fsid [2])
 
 	fsid[0] = (be16_to_cpu(uup->uu_clockseq) << 16) |
 		   be16_to_cpu(uup->uu_timemid);
-	fsid[1] = be16_to_cpu(uup->uu_timelow);
+	fsid[1] = be32_to_cpu(uup->uu_timelow);
 }
 
 void
@@ -139,3 +132,9 @@ uuid_table_remove(uuid_t *uuid)
 	ASSERT(i < uuid_table_size);
 	mutex_unlock(&uuid_monitor);
 }
+
+void
+uuid_init(void)
+{
+	mutex_init(&uuid_monitor);
+}
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
index 4ff0f4e41c61..2539af34eb63 100644
--- a/fs/xfs/xfs_acl.c
+++ b/fs/xfs/xfs_acl.c
@@ -395,7 +395,7 @@ xfs_acl_allow_set(
  * The access control process to determine the access permission:
  *	if uid == file owner id, use the file owner bits.
  *	if gid == file owner group id, use the file group bits.
- *	scan ACL for a maching user or group, and use matched entry
+ *	scan ACL for a matching user or group, and use matched entry
  *	permission. Use total permissions of all matching group entries,
  *	until all acl entries are exhausted. The final permission produced
  *	by matching acl entry or entries needs to be & with group permission.
diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h
index f9315bc960cb..538d0d65b04c 100644
--- a/fs/xfs/xfs_acl.h
+++ b/fs/xfs/xfs_acl.h
@@ -55,8 +55,8 @@ struct xfs_inode;
 
 extern struct kmem_zone *xfs_acl_zone;
 #define xfs_acl_zone_init(zone, name)	\
-		(zone) = kmem_zone_init(sizeof(xfs_acl_t), name)
-#define xfs_acl_zone_destroy(zone)	kmem_cache_destroy(zone)
+		(zone) = kmem_zone_init(sizeof(xfs_acl_t), (name))
+#define xfs_acl_zone_destroy(zone)	kmem_zone_destroy(zone)
 
 extern int xfs_acl_inherit(struct vnode *, struct vattr *, xfs_acl_t *);
 extern int xfs_acl_iaccess(struct xfs_inode *, mode_t, cred_t *);
diff --git a/fs/xfs/xfs_ag.h b/fs/xfs/xfs_ag.h
index a96e2ffce0cc..dc2361dd740a 100644
--- a/fs/xfs/xfs_ag.h
+++ b/fs/xfs/xfs_ag.h
@@ -179,7 +179,7 @@ typedef struct xfs_perag
 {
 	char		pagf_init;	/* this agf's entry is initialized */
 	char		pagi_init;	/* this agi's entry is initialized */
-	char		pagf_metadata;	/* the agf is prefered to be metadata */
+	char		pagf_metadata;	/* the agf is preferred to be metadata */
 	char		pagi_inodeok;	/* The agi is ok for inodes */
 	__uint8_t	pagf_levels[XFS_BTNUM_AGF];
 					/* # of levels in bno & cnt btree */
diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c
index f4328e1e2a74..64ee07db0d5e 100644
--- a/fs/xfs/xfs_alloc.c
+++ b/fs/xfs/xfs_alloc.c
@@ -511,7 +511,7 @@ STATIC void
 xfs_alloc_trace_busy(
 	char		*name,		/* function tag string */
 	char		*str,		/* additional string */
-	xfs_mount_t	*mp,		/* file system mount poing */
+	xfs_mount_t	*mp,		/* file system mount point */
 	xfs_agnumber_t	agno,		/* allocation group number */
 	xfs_agblock_t	agbno,		/* a.g. relative block number */
 	xfs_extlen_t	len,		/* length of extent */
@@ -1843,7 +1843,7 @@ xfs_alloc_fix_freelist(
 	} else
 		agbp = NULL;
 
-	/* If this is a metadata prefered pag and we are user data
+	/* If this is a metadata preferred pag and we are user data
 	 * then try somewhere else if we are not being asked to
 	 * try harder at this point
 	 */
@@ -2458,7 +2458,7 @@ error0:
 /*
  * AG Busy list management
  * The busy list contains block ranges that have been freed but whose
- * transacations have not yet hit disk.  If any block listed in a busy
+ * transactions have not yet hit disk.  If any block listed in a busy
  * list is reused, the transaction that freed it must be forced to disk
  * before continuing to use the block.
  *
diff --git a/fs/xfs/xfs_alloc.h b/fs/xfs/xfs_alloc.h
index 3546dea27b7d..2d1f8928b267 100644
--- a/fs/xfs/xfs_alloc.h
+++ b/fs/xfs/xfs_alloc.h
@@ -68,7 +68,7 @@ typedef struct xfs_alloc_arg {
 	xfs_alloctype_t	otype;		/* original allocation type */
 	char		wasdel;		/* set if allocation was prev delayed */
 	char		wasfromfl;	/* set if allocation is from freelist */
-	char		isfl;		/* set if is freelist blocks - !actg */
+	char		isfl;		/* set if is freelist blocks - !acctg */
 	char		userdata;	/* set if this is user data */
 } xfs_alloc_arg_t;
 
diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c
index e5e91e9c7e89..b6e1e02bbb28 100644
--- a/fs/xfs/xfs_attr.c
+++ b/fs/xfs/xfs_attr.c
@@ -294,7 +294,7 @@ xfs_attr_set_int(xfs_inode_t *dp, const char *name, int namelen,
 	xfs_trans_ihold(args.trans, dp);
 
 	/*
-	 * If the attribute list is non-existant or a shortform list,
+	 * If the attribute list is non-existent or a shortform list,
 	 * upgrade it to a single-leaf-block attribute list.
 	 */
 	if ((dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) ||
@@ -1127,8 +1127,7 @@ xfs_attr_leaf_list(xfs_attr_list_context_t *context)
 		return(error);
 	ASSERT(bp != NULL);
 	leaf = bp->data;
-	if (unlikely(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						!= XFS_ATTR_LEAF_MAGIC)) {
+	if (unlikely(be16_to_cpu(leaf->hdr.info.magic) != XFS_ATTR_LEAF_MAGIC)) {
 		XFS_CORRUPTION_ERROR("xfs_attr_leaf_list", XFS_ERRLEVEL_LOW,
 				     context->dp->i_mount, leaf);
 		xfs_da_brelse(NULL, bp);
@@ -1541,8 +1540,8 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 						     XFS_ATTR_FORK);
 		if (error)
 			goto out;
-		ASSERT(INT_GET(((xfs_attr_leafblock_t *)
-				      bp->data)->hdr.info.magic, ARCH_CONVERT)
+		ASSERT(be16_to_cpu(((xfs_attr_leafblock_t *)
+				      bp->data)->hdr.info.magic)
 						       == XFS_ATTR_LEAF_MAGIC);
 
 		if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
@@ -1585,7 +1584,7 @@ out:
  * Fill in the disk block numbers in the state structure for the buffers
  * that are attached to the state structure.
  * This is done so that we can quickly reattach ourselves to those buffers
- * after some set of transaction commit's has released these buffers.
+ * after some set of transaction commits have released these buffers.
  */
 STATIC int
 xfs_attr_fillstate(xfs_da_state_t *state)
@@ -1632,7 +1631,7 @@ xfs_attr_fillstate(xfs_da_state_t *state)
 /*
  * Reattach the buffers to the state structure based on the disk block
  * numbers stored in the state structure.
- * This is done after some set of transaction commit's has released those
+ * This is done after some set of transaction commits have released those
  * buffers from our grip.
  */
 STATIC int
@@ -1763,7 +1762,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 			return(error);
 		if (bp) {
 			node = bp->data;
-			switch (INT_GET(node->hdr.info.magic, ARCH_CONVERT)) {
+			switch (be16_to_cpu(node->hdr.info.magic)) {
 			case XFS_DA_NODE_MAGIC:
 				xfs_attr_trace_l_cn("wrong blk", context, node);
 				xfs_da_brelse(NULL, bp);
@@ -1771,18 +1770,14 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 				break;
 			case XFS_ATTR_LEAF_MAGIC:
 				leaf = bp->data;
-				if (cursor->hashval >
-				    INT_GET(leaf->entries[
-					 INT_GET(leaf->hdr.count,
-						ARCH_CONVERT)-1].hashval,
-							ARCH_CONVERT)) {
+				if (cursor->hashval > be32_to_cpu(leaf->entries[
+				    be16_to_cpu(leaf->hdr.count)-1].hashval)) {
 					xfs_attr_trace_l_cl("wrong blk",
 							   context, leaf);
 					xfs_da_brelse(NULL, bp);
 					bp = NULL;
 				} else if (cursor->hashval <=
-					     INT_GET(leaf->entries[0].hashval,
-							ARCH_CONVERT)) {
+					     be32_to_cpu(leaf->entries[0].hashval)) {
 					xfs_attr_trace_l_cl("maybe wrong blk",
 							   context, leaf);
 					xfs_da_brelse(NULL, bp);
@@ -1817,10 +1812,10 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 				return(XFS_ERROR(EFSCORRUPTED));
 			}
 			node = bp->data;
-			if (INT_GET(node->hdr.info.magic, ARCH_CONVERT)
+			if (be16_to_cpu(node->hdr.info.magic)
 							== XFS_ATTR_LEAF_MAGIC)
 				break;
-			if (unlikely(INT_GET(node->hdr.info.magic, ARCH_CONVERT)
+			if (unlikely(be16_to_cpu(node->hdr.info.magic)
 							!= XFS_DA_NODE_MAGIC)) {
 				XFS_CORRUPTION_ERROR("xfs_attr_node_list(3)",
 						     XFS_ERRLEVEL_LOW,
@@ -1830,19 +1825,17 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 				return(XFS_ERROR(EFSCORRUPTED));
 			}
 			btree = node->btree;
-			for (i = 0;
-				i < INT_GET(node->hdr.count, ARCH_CONVERT);
+			for (i = 0; i < be16_to_cpu(node->hdr.count);
 								btree++, i++) {
 				if (cursor->hashval
-						<= INT_GET(btree->hashval,
-							    ARCH_CONVERT)) {
-					cursor->blkno = INT_GET(btree->before, ARCH_CONVERT);
+						<= be32_to_cpu(btree->hashval)) {
+					cursor->blkno = be32_to_cpu(btree->before);
 					xfs_attr_trace_l_cb("descending",
 							    context, btree);
 					break;
 				}
 			}
-			if (i == INT_GET(node->hdr.count, ARCH_CONVERT)) {
+			if (i == be16_to_cpu(node->hdr.count)) {
 				xfs_da_brelse(NULL, bp);
 				return(0);
 			}
@@ -1858,7 +1851,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 	 */
 	for (;;) {
 		leaf = bp->data;
-		if (unlikely(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
+		if (unlikely(be16_to_cpu(leaf->hdr.info.magic)
 						!= XFS_ATTR_LEAF_MAGIC)) {
 			XFS_CORRUPTION_ERROR("xfs_attr_node_list(4)",
 					     XFS_ERRLEVEL_LOW,
@@ -1869,7 +1862,7 @@ xfs_attr_node_list(xfs_attr_list_context_t *context)
 		error = xfs_attr_leaf_list_int(bp, context);
 		if (error || !leaf->hdr.info.forw)
 			break;	/* not really an error, buffer full or EOF */
-		cursor->blkno = INT_GET(leaf->hdr.info.forw, ARCH_CONVERT);
+		cursor->blkno = be32_to_cpu(leaf->hdr.info.forw);
 		xfs_da_brelse(NULL, bp);
 		error = xfs_da_read_buf(NULL, context->dp, cursor->blkno, -1,
 					      &bp, XFS_ATTR_FORK);
@@ -2232,9 +2225,10 @@ xfs_attr_trace_l_cn(char *where, struct xfs_attr_list_context *context,
 				: 0,
 		(__psunsigned_t)context->dupcnt,
 		(__psunsigned_t)context->flags,
-		(__psunsigned_t)INT_GET(node->hdr.count, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(node->btree[0].hashval, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(node->btree[INT_GET(node->hdr.count, ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
+		(__psunsigned_t)be16_to_cpu(node->hdr.count),
+		(__psunsigned_t)be32_to_cpu(node->btree[0].hashval),
+		(__psunsigned_t)be32_to_cpu(node->btree[
+				    be16_to_cpu(node->hdr.count)-1].hashval));
 }
 
 /*
@@ -2261,8 +2255,8 @@ xfs_attr_trace_l_cb(char *where, struct xfs_attr_list_context *context,
 				: 0,
 		(__psunsigned_t)context->dupcnt,
 		(__psunsigned_t)context->flags,
-		(__psunsigned_t)INT_GET(btree->hashval, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(btree->before, ARCH_CONVERT),
+		(__psunsigned_t)be32_to_cpu(btree->hashval),
+		(__psunsigned_t)be32_to_cpu(btree->before),
 		(__psunsigned_t)NULL);
 }
 
@@ -2290,9 +2284,10 @@ xfs_attr_trace_l_cl(char *where, struct xfs_attr_list_context *context,
 				: 0,
 		(__psunsigned_t)context->dupcnt,
 		(__psunsigned_t)context->flags,
-		(__psunsigned_t)INT_GET(leaf->hdr.count, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(leaf->entries[0].hashval, ARCH_CONVERT),
-		(__psunsigned_t)INT_GET(leaf->entries[INT_GET(leaf->hdr.count, ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
+		(__psunsigned_t)be16_to_cpu(leaf->hdr.count),
+		(__psunsigned_t)be32_to_cpu(leaf->entries[0].hashval),
+		(__psunsigned_t)be32_to_cpu(leaf->entries[
+				be16_to_cpu(leaf->hdr.count)-1].hashval));
 }
 
 /*
@@ -2522,7 +2517,7 @@ attr_user_capable(
 	struct vnode	*vp,
 	cred_t		*cred)
 {
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return -EPERM;
@@ -2540,7 +2535,7 @@ attr_trusted_capable(
 	struct vnode	*vp,
 	cred_t		*cred)
 {
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return -EPERM;
diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c
index fe91eac4e2a7..9462be86aa14 100644
--- a/fs/xfs/xfs_attr_leaf.c
+++ b/fs/xfs/xfs_attr_leaf.c
@@ -194,7 +194,7 @@ xfs_attr_shortform_create(xfs_da_args_t *args)
 	xfs_idata_realloc(dp, sizeof(*hdr), XFS_ATTR_FORK);
 	hdr = (xfs_attr_sf_hdr_t *)ifp->if_u1.if_data;
 	hdr->count = 0;
-	INT_SET(hdr->totsize, ARCH_CONVERT, sizeof(*hdr));
+	hdr->totsize = cpu_to_be16(sizeof(*hdr));
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_ADATA);
 }
 
@@ -224,8 +224,7 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT);
-				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
+	for (i = 0; i < sf->hdr.count; sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
 #ifdef DEBUG
 		if (sfe->namelen != args->namelen)
 			continue;
@@ -248,13 +247,13 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 	sfe = (xfs_attr_sf_entry_t *)((char *)sf + offset);
 
 	sfe->namelen = args->namelen;
-	INT_SET(sfe->valuelen, ARCH_CONVERT, args->valuelen);
+	sfe->valuelen = args->valuelen;
 	sfe->flags = (args->flags & ATTR_SECURE) ? XFS_ATTR_SECURE :
 			((args->flags & ATTR_ROOT) ? XFS_ATTR_ROOT : 0);
 	memcpy(sfe->nameval, args->name, args->namelen);
 	memcpy(&sfe->nameval[args->namelen], args->value, args->valuelen);
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, 1);
-	INT_MOD(sf->hdr.totsize, ARCH_CONVERT, size);
+	sf->hdr.count++;
+	be16_add(&sf->hdr.totsize, size);
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_ADATA);
 
 	xfs_sbversion_add_attr2(mp, args->trans);
@@ -277,7 +276,7 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	base = sizeof(xfs_attr_sf_hdr_t);
 	sf = (xfs_attr_shortform_t *)dp->i_afp->if_u1.if_data;
 	sfe = &sf->list[0];
-	end = INT_GET(sf->hdr.count, ARCH_CONVERT);
+	end = sf->hdr.count;
 	for (i = 0; i < end; sfe = XFS_ATTR_SF_NEXTENTRY(sfe),
 					base += size, i++) {
 		size = XFS_ATTR_SF_ENTSIZE(sfe);
@@ -300,11 +299,11 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	 * Fix up the attribute fork data, covering the hole
 	 */
 	end = base + size;
-	totsize = INT_GET(sf->hdr.totsize, ARCH_CONVERT);
+	totsize = be16_to_cpu(sf->hdr.totsize);
 	if (end != totsize)
 		memmove(&((char *)sf)[base], &((char *)sf)[end], totsize - end);
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, -1);
-	INT_MOD(sf->hdr.totsize, ARCH_CONVERT, -size);
+	sf->hdr.count--;
+	be16_add(&sf->hdr.totsize, -size);
 
 	/*
 	 * Fix up the start offset of the attribute fork
@@ -360,7 +359,7 @@ xfs_attr_shortform_lookup(xfs_da_args_t *args)
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT);
+	for (i = 0; i < sf->hdr.count;
 				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
 		if (sfe->namelen != args->namelen)
 			continue;
@@ -391,7 +390,7 @@ xfs_attr_shortform_getvalue(xfs_da_args_t *args)
 	ASSERT(args->dp->i_d.di_aformat == XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)args->dp->i_afp->if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT);
+	for (i = 0; i < sf->hdr.count;
 				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
 		if (sfe->namelen != args->namelen)
 			continue;
@@ -404,14 +403,14 @@ xfs_attr_shortform_getvalue(xfs_da_args_t *args)
 		    ((sfe->flags & XFS_ATTR_ROOT) != 0))
 			continue;
 		if (args->flags & ATTR_KERNOVAL) {
-			args->valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+			args->valuelen = sfe->valuelen;
 			return(XFS_ERROR(EEXIST));
 		}
-		if (args->valuelen < INT_GET(sfe->valuelen, ARCH_CONVERT)) {
-			args->valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+		if (args->valuelen < sfe->valuelen) {
+			args->valuelen = sfe->valuelen;
 			return(XFS_ERROR(ERANGE));
 		}
-		args->valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+		args->valuelen = sfe->valuelen;
 		memcpy(args->value, &sfe->nameval[args->namelen],
 						    args->valuelen);
 		return(XFS_ERROR(EEXIST));
@@ -438,7 +437,7 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	dp = args->dp;
 	ifp = dp->i_afp;
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
-	size = INT_GET(sf->hdr.totsize, ARCH_CONVERT);
+	size = be16_to_cpu(sf->hdr.totsize);
 	tmpbuffer = kmem_alloc(size, KM_SLEEP);
 	ASSERT(tmpbuffer != NULL);
 	memcpy(tmpbuffer, ifp->if_u1.if_data, size);
@@ -481,11 +480,11 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 	nargs.oknoent = 1;
 
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0; i < sf->hdr.count; i++) {
 		nargs.name = (char *)sfe->nameval;
 		nargs.namelen = sfe->namelen;
 		nargs.value = (char *)&sfe->nameval[nargs.namelen];
-		nargs.valuelen = INT_GET(sfe->valuelen, ARCH_CONVERT);
+		nargs.valuelen = sfe->valuelen;
 		nargs.hashval = xfs_da_hashname((char *)sfe->nameval,
 						sfe->namelen);
 		nargs.flags = (sfe->flags & XFS_ATTR_SECURE) ? ATTR_SECURE :
@@ -514,11 +513,9 @@ xfs_attr_shortform_compare(const void *a, const void *b)
 
 	sa = (xfs_attr_sf_sort_t *)a;
 	sb = (xfs_attr_sf_sort_t *)b;
-	if (INT_GET(sa->hash, ARCH_CONVERT)
-				< INT_GET(sb->hash, ARCH_CONVERT)) {
+	if (sa->hash < sb->hash) {
 		return(-1);
-	} else if (INT_GET(sa->hash, ARCH_CONVERT)
-				> INT_GET(sb->hash, ARCH_CONVERT)) {
+	} else if (sa->hash > sb->hash) {
 		return(1);
 	} else {
 		return(sa->entno - sb->entno);
@@ -527,7 +524,7 @@ xfs_attr_shortform_compare(const void *a, const void *b)
 
 /*
  * Copy out entries of shortform attribute lists for attr_list().
- * Shortform atrtribute lists are not stored in hashval sorted order.
+ * Shortform attribute lists are not stored in hashval sorted order.
  * If the output buffer is not large enough to hold them all, then we
  * we have to calculate each entries' hashvalue and sort them before
  * we can begin returning them to the user.
@@ -560,10 +557,8 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	 * If the buffer is large enough, do not bother with sorting.
 	 * Note the generous fudge factor of 16 overhead bytes per entry.
 	 */
-	if ((dp->i_afp->if_bytes + INT_GET(sf->hdr.count, ARCH_CONVERT) * 16)
-							< context->bufsize) {
-		for (i = 0, sfe = &sf->list[0];
-				i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	if ((dp->i_afp->if_bytes + sf->hdr.count * 16) < context->bufsize) {
+		for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
 			attrnames_t	*namesp;
 
 			if (((context->flags & ATTR_SECURE) != 0) !=
@@ -584,14 +579,13 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			if (context->flags & ATTR_KERNOVAL) {
 				ASSERT(context->flags & ATTR_KERNAMELS);
 				context->count += namesp->attr_namelen +
-					INT_GET(sfe->namelen, ARCH_CONVERT) + 1;
+					sfe->namelen + 1;
 			}
 			else {
 				if (xfs_attr_put_listent(context, namesp,
 						   (char *)sfe->nameval,
 						   (int)sfe->namelen,
-						   (int)INT_GET(sfe->valuelen,
-								ARCH_CONVERT)))
+						   (int)sfe->valuelen))
 					break;
 			}
 			sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
@@ -603,7 +597,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	/*
 	 * It didn't all fit, so we have to sort everything on hashval.
 	 */
-	sbsize = INT_GET(sf->hdr.count, ARCH_CONVERT) * sizeof(*sbuf);
+	sbsize = sf->hdr.count * sizeof(*sbuf);
 	sbp = sbuf = kmem_alloc(sbsize, KM_SLEEP);
 
 	/*
@@ -611,8 +605,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	 * the relevant info from only those that match into a buffer.
 	 */
 	nsbuf = 0;
-	for (i = 0, sfe = &sf->list[0];
-			i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
 		if (unlikely(
 		    ((char *)sfe < (char *)sf) ||
 		    ((char *)sfe >= ((char *)sf + dp->i_afp->if_bytes)))) {
@@ -636,8 +629,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			continue;
 		}
 		sbp->entno = i;
-		INT_SET(sbp->hash, ARCH_CONVERT,
-			xfs_da_hashname((char *)sfe->nameval, sfe->namelen));
+		sbp->hash = xfs_da_hashname((char *)sfe->nameval, sfe->namelen);
 		sbp->name = (char *)sfe->nameval;
 		sbp->namelen = sfe->namelen;
 		/* These are bytes, and both on-disk, don't endian-flip */
@@ -660,12 +652,12 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 	cursor->initted = 1;
 	cursor->blkno = 0;
 	for (sbp = sbuf, i = 0; i < nsbuf; i++, sbp++) {
-		if (INT_GET(sbp->hash, ARCH_CONVERT) == cursor->hashval) {
+		if (sbp->hash == cursor->hashval) {
 			if (cursor->offset == count) {
 				break;
 			}
 			count++;
-		} else if (INT_GET(sbp->hash, ARCH_CONVERT) > cursor->hashval) {
+		} else if (sbp->hash > cursor->hashval) {
 			break;
 		}
 	}
@@ -685,8 +677,8 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 			((sbp->flags & XFS_ATTR_ROOT) ? &attr_trusted :
 			  &attr_user);
 
-		if (cursor->hashval != INT_GET(sbp->hash, ARCH_CONVERT)) {
-			cursor->hashval = INT_GET(sbp->hash, ARCH_CONVERT);
+		if (cursor->hashval != sbp->hash) {
+			cursor->hashval = sbp->hash;
 			cursor->offset = 0;
 		}
 		if (context->flags & ATTR_KERNOVAL) {
@@ -696,7 +688,7 @@ xfs_attr_shortform_list(xfs_attr_list_context_t *context)
 		} else {
 			if (xfs_attr_put_listent(context, namesp,
 					sbp->name, sbp->namelen,
-					INT_GET(sbp->valuelen, ARCH_CONVERT)))
+					sbp->valuelen))
 				break;
 		}
 		cursor->offset++;
@@ -720,12 +712,11 @@ xfs_attr_shortform_allfit(xfs_dabuf_t *bp, xfs_inode_t *dp)
 	int bytes, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 
 	entry = &leaf->entries[0];
 	bytes = sizeof(struct xfs_attr_sf_hdr);
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
 			continue;		/* don't copy partial entries */
 		if (!(entry->flags & XFS_ATTR_LOCAL))
@@ -733,11 +724,11 @@ xfs_attr_shortform_allfit(xfs_dabuf_t *bp, xfs_inode_t *dp)
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, i);
 		if (name_loc->namelen >= XFS_ATTR_SF_ENTSIZE_MAX)
 			return(0);
-		if (INT_GET(name_loc->valuelen, ARCH_CONVERT) >= XFS_ATTR_SF_ENTSIZE_MAX)
+		if (be16_to_cpu(name_loc->valuelen) >= XFS_ATTR_SF_ENTSIZE_MAX)
 			return(0);
 		bytes += sizeof(struct xfs_attr_sf_entry)-1
 				+ name_loc->namelen
-				+ INT_GET(name_loc->valuelen, ARCH_CONVERT);
+				+ be16_to_cpu(name_loc->valuelen);
 	}
 	if ((dp->i_mount->m_flags & XFS_MOUNT_ATTR2) &&
 	    (bytes == sizeof(struct xfs_attr_sf_hdr)))
@@ -766,8 +757,7 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	ASSERT(bp != NULL);
 	memcpy(tmpbuffer, bp->data, XFS_LBSIZE(dp->i_mount));
 	leaf = (xfs_attr_leafblock_t *)tmpbuffer;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	memset(bp->data, 0, XFS_LBSIZE(dp->i_mount));
 
 	/*
@@ -810,7 +800,7 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 	nargs.trans = args->trans;
 	nargs.oknoent = 1;
 	entry = &leaf->entries[0];
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
 		if (entry->flags & XFS_ATTR_INCOMPLETE)
 			continue;	/* don't copy partial entries */
 		if (!entry->nameidx)
@@ -820,8 +810,8 @@ xfs_attr_leaf_to_shortform(xfs_dabuf_t *bp, xfs_da_args_t *args, int forkoff)
 		nargs.name = (char *)name_loc->nameval;
 		nargs.namelen = name_loc->namelen;
 		nargs.value = (char *)&name_loc->nameval[nargs.namelen];
-		nargs.valuelen = INT_GET(name_loc->valuelen, ARCH_CONVERT);
-		nargs.hashval = INT_GET(entry->hashval, ARCH_CONVERT);
+		nargs.valuelen = be16_to_cpu(name_loc->valuelen);
+		nargs.hashval = be32_to_cpu(entry->hashval);
 		nargs.flags = (entry->flags & XFS_ATTR_SECURE) ? ATTR_SECURE :
 			      ((entry->flags & XFS_ATTR_ROOT) ? ATTR_ROOT : 0);
 		xfs_attr_shortform_add(&nargs, forkoff);
@@ -875,13 +865,12 @@ xfs_attr_leaf_to_node(xfs_da_args_t *args)
 		goto out;
 	node = bp1->data;
 	leaf = bp2->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	/* both on-disk, don't endian-flip twice */
 	node->btree[0].hashval =
-		leaf->entries[INT_GET(leaf->hdr.count, ARCH_CONVERT)-1 ].hashval;
-	INT_SET(node->btree[0].before, ARCH_CONVERT, blkno);
-	INT_SET(node->hdr.count, ARCH_CONVERT, 1);
+		leaf->entries[be16_to_cpu(leaf->hdr.count)-1 ].hashval;
+	node->btree[0].before = cpu_to_be32(blkno);
+	node->hdr.count = cpu_to_be16(1);
 	xfs_da_log_buf(args->trans, bp1, 0, XFS_LBSIZE(dp->i_mount) - 1);
 	error = 0;
 out:
@@ -920,19 +909,16 @@ xfs_attr_leaf_create(xfs_da_args_t *args, xfs_dablk_t blkno, xfs_dabuf_t **bpp)
 	leaf = bp->data;
 	memset((char *)leaf, 0, XFS_LBSIZE(dp->i_mount));
 	hdr = &leaf->hdr;
-	INT_SET(hdr->info.magic, ARCH_CONVERT, XFS_ATTR_LEAF_MAGIC);
-	INT_SET(hdr->firstused, ARCH_CONVERT, XFS_LBSIZE(dp->i_mount));
+	hdr->info.magic = cpu_to_be16(XFS_ATTR_LEAF_MAGIC);
+	hdr->firstused = cpu_to_be16(XFS_LBSIZE(dp->i_mount));
 	if (!hdr->firstused) {
-		INT_SET(hdr->firstused, ARCH_CONVERT,
+		hdr->firstused = cpu_to_be16(
 			XFS_LBSIZE(dp->i_mount) - XFS_ATTR_LEAF_NAME_ALIGN);
 	}
 
-	INT_SET(hdr->freemap[0].base, ARCH_CONVERT,
-						sizeof(xfs_attr_leaf_hdr_t));
-	INT_SET(hdr->freemap[0].size, ARCH_CONVERT,
-					  INT_GET(hdr->firstused, ARCH_CONVERT)
-					- INT_GET(hdr->freemap[0].base,
-								ARCH_CONVERT));
+	hdr->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
+	hdr->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr->firstused) -
+					   sizeof(xfs_attr_leaf_hdr_t));
 
 	xfs_da_log_buf(args->trans, bp, 0, XFS_LBSIZE(dp->i_mount) - 1);
 
@@ -1004,10 +990,9 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	int tablesize, entsize, sum, tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	ASSERT((args->index >= 0)
-		&& (args->index <= INT_GET(leaf->hdr.count, ARCH_CONVERT)));
+		&& (args->index <= be16_to_cpu(leaf->hdr.count)));
 	hdr = &leaf->hdr;
 	entsize = xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
 			   args->trans->t_mountp->m_sb.sb_blocksize, NULL);
@@ -1016,26 +1001,25 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * Search through freemap for first-fit on new name length.
 	 * (may need to figure in size of entry struct too)
 	 */
-	tablesize = (INT_GET(hdr->count, ARCH_CONVERT) + 1)
+	tablesize = (be16_to_cpu(hdr->count) + 1)
 					* sizeof(xfs_attr_leaf_entry_t)
 					+ sizeof(xfs_attr_leaf_hdr_t);
 	map = &hdr->freemap[XFS_ATTR_LEAF_MAPSIZE-1];
 	for (sum = 0, i = XFS_ATTR_LEAF_MAPSIZE-1; i >= 0; map--, i--) {
-		if (tablesize > INT_GET(hdr->firstused, ARCH_CONVERT)) {
-			sum += INT_GET(map->size, ARCH_CONVERT);
+		if (tablesize > be16_to_cpu(hdr->firstused)) {
+			sum += be16_to_cpu(map->size);
 			continue;
 		}
 		if (!map->size)
 			continue;	/* no space in this map */
 		tmp = entsize;
-		if (INT_GET(map->base, ARCH_CONVERT)
-				< INT_GET(hdr->firstused, ARCH_CONVERT))
+		if (be16_to_cpu(map->base) < be16_to_cpu(hdr->firstused))
 			tmp += sizeof(xfs_attr_leaf_entry_t);
-		if (INT_GET(map->size, ARCH_CONVERT) >= tmp) {
+		if (be16_to_cpu(map->size) >= tmp) {
 			tmp = xfs_attr_leaf_add_work(bp, args, i);
 			return(tmp);
 		}
-		sum += INT_GET(map->size, ARCH_CONVERT);
+		sum += be16_to_cpu(map->size);
 	}
 
 	/*
@@ -1056,7 +1040,7 @@ xfs_attr_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * After compaction, the block is guaranteed to have only one
 	 * free region, in freemap[0].  If it is not big enough, give up.
 	 */
-	if (INT_GET(hdr->freemap[0].size, ARCH_CONVERT)
+	if (be16_to_cpu(hdr->freemap[0].size)
 				< (entsize + sizeof(xfs_attr_leaf_entry_t)))
 		return(XFS_ERROR(ENOSPC));
 
@@ -1079,45 +1063,42 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	int tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	ASSERT((mapindex >= 0) && (mapindex < XFS_ATTR_LEAF_MAPSIZE));
-	ASSERT((args->index >= 0)
-		&& (args->index <= INT_GET(hdr->count, ARCH_CONVERT)));
+	ASSERT((args->index >= 0) && (args->index <= be16_to_cpu(hdr->count)));
 
 	/*
 	 * Force open some space in the entry array and fill it in.
 	 */
 	entry = &leaf->entries[args->index];
-	if (args->index < INT_GET(hdr->count, ARCH_CONVERT)) {
-		tmp  = INT_GET(hdr->count, ARCH_CONVERT) - args->index;
+	if (args->index < be16_to_cpu(hdr->count)) {
+		tmp  = be16_to_cpu(hdr->count) - args->index;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
 		memmove((char *)(entry+1), (char *)entry, tmp);
 		xfs_da_log_buf(args->trans, bp,
 		    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(*entry)));
 	}
-	INT_MOD(hdr->count, ARCH_CONVERT, 1);
+	be16_add(&hdr->count, 1);
 
 	/*
 	 * Allocate space for the new string (at the end of the run).
 	 */
 	map = &hdr->freemap[mapindex];
 	mp = args->trans->t_mountp;
-	ASSERT(INT_GET(map->base, ARCH_CONVERT) < XFS_LBSIZE(mp));
-	ASSERT((INT_GET(map->base, ARCH_CONVERT) & 0x3) == 0);
-	ASSERT(INT_GET(map->size, ARCH_CONVERT) >=
+	ASSERT(be16_to_cpu(map->base) < XFS_LBSIZE(mp));
+	ASSERT((be16_to_cpu(map->base) & 0x3) == 0);
+	ASSERT(be16_to_cpu(map->size) >=
 		xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
 					 mp->m_sb.sb_blocksize, NULL));
-	ASSERT(INT_GET(map->size, ARCH_CONVERT) < XFS_LBSIZE(mp));
-	ASSERT((INT_GET(map->size, ARCH_CONVERT) & 0x3) == 0);
-	INT_MOD(map->size, ARCH_CONVERT,
+	ASSERT(be16_to_cpu(map->size) < XFS_LBSIZE(mp));
+	ASSERT((be16_to_cpu(map->size) & 0x3) == 0);
+	be16_add(&map->size,
 		-xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
 					  mp->m_sb.sb_blocksize, &tmp));
-	INT_SET(entry->nameidx, ARCH_CONVERT,
-					INT_GET(map->base, ARCH_CONVERT)
-				      + INT_GET(map->size, ARCH_CONVERT));
-	INT_SET(entry->hashval, ARCH_CONVERT, args->hashval);
+	entry->nameidx = cpu_to_be16(be16_to_cpu(map->base) +
+				     be16_to_cpu(map->size));
+	entry->hashval = cpu_to_be32(args->hashval);
 	entry->flags = tmp ? XFS_ATTR_LOCAL : 0;
 	entry->flags |= (args->flags & ATTR_SECURE) ? XFS_ATTR_SECURE :
 			((args->flags & ATTR_ROOT) ? XFS_ATTR_ROOT : 0);
@@ -1130,12 +1111,10 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	}
 	xfs_da_log_buf(args->trans, bp,
 			  XFS_DA_LOGRANGE(leaf, entry, sizeof(*entry)));
-	ASSERT((args->index == 0) || (INT_GET(entry->hashval, ARCH_CONVERT)
-						>= INT_GET((entry-1)->hashval,
-							    ARCH_CONVERT)));
-	ASSERT((args->index == INT_GET(hdr->count, ARCH_CONVERT)-1) ||
-	       (INT_GET(entry->hashval, ARCH_CONVERT)
-			    <= (INT_GET((entry+1)->hashval, ARCH_CONVERT))));
+	ASSERT((args->index == 0) ||
+	       (be32_to_cpu(entry->hashval) >= be32_to_cpu((entry-1)->hashval)));
+	ASSERT((args->index == be16_to_cpu(hdr->count)-1) ||
+	       (be32_to_cpu(entry->hashval) <= be32_to_cpu((entry+1)->hashval)));
 
 	/*
 	 * Copy the attribute name and value into the new space.
@@ -1149,10 +1128,10 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	if (entry->flags & XFS_ATTR_LOCAL) {
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
 		name_loc->namelen = args->namelen;
-		INT_SET(name_loc->valuelen, ARCH_CONVERT, args->valuelen);
+		name_loc->valuelen = cpu_to_be16(args->valuelen);
 		memcpy((char *)name_loc->nameval, args->name, args->namelen);
 		memcpy((char *)&name_loc->nameval[args->namelen], args->value,
-				   INT_GET(name_loc->valuelen, ARCH_CONVERT));
+				   be16_to_cpu(name_loc->valuelen));
 	} else {
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
 		name_rmt->namelen = args->namelen;
@@ -1171,28 +1150,23 @@ xfs_attr_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int mapindex)
 	/*
 	 * Update the control info for this leaf node
 	 */
-	if (INT_GET(entry->nameidx, ARCH_CONVERT)
-				< INT_GET(hdr->firstused, ARCH_CONVERT)) {
+	if (be16_to_cpu(entry->nameidx) < be16_to_cpu(hdr->firstused)) {
 		/* both on-disk, don't endian-flip twice */
 		hdr->firstused = entry->nameidx;
 	}
-	ASSERT(INT_GET(hdr->firstused, ARCH_CONVERT)
-				>= ((INT_GET(hdr->count, ARCH_CONVERT)
-					* sizeof(*entry))+sizeof(*hdr)));
-	tmp = (INT_GET(hdr->count, ARCH_CONVERT)-1)
-					* sizeof(xfs_attr_leaf_entry_t)
+	ASSERT(be16_to_cpu(hdr->firstused) >=
+	       ((be16_to_cpu(hdr->count) * sizeof(*entry)) + sizeof(*hdr)));
+	tmp = (be16_to_cpu(hdr->count)-1) * sizeof(xfs_attr_leaf_entry_t)
 					+ sizeof(xfs_attr_leaf_hdr_t);
 	map = &hdr->freemap[0];
 	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; map++, i++) {
-		if (INT_GET(map->base, ARCH_CONVERT) == tmp) {
-			INT_MOD(map->base, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_entry_t));
-			INT_MOD(map->size, ARCH_CONVERT,
-					-sizeof(xfs_attr_leaf_entry_t));
+		if (be16_to_cpu(map->base) == tmp) {
+			be16_add(&map->base, sizeof(xfs_attr_leaf_entry_t));
+			be16_add(&map->size,
+				 -((int)sizeof(xfs_attr_leaf_entry_t)));
 		}
 	}
-	INT_MOD(hdr->usedbytes, ARCH_CONVERT,
-				xfs_attr_leaf_entsize(leaf, args->index));
+	be16_add(&hdr->usedbytes, xfs_attr_leaf_entsize(leaf, args->index));
 	xfs_da_log_buf(args->trans, bp,
 		XFS_DA_LOGRANGE(leaf, hdr, sizeof(*hdr)));
 	return(0);
@@ -1223,28 +1197,25 @@ xfs_attr_leaf_compact(xfs_trans_t *trans, xfs_dabuf_t *bp)
 	hdr_s = &leaf_s->hdr;
 	hdr_d = &leaf_d->hdr;
 	hdr_d->info = hdr_s->info;	/* struct copy */
-	INT_SET(hdr_d->firstused, ARCH_CONVERT, XFS_LBSIZE(mp));
+	hdr_d->firstused = cpu_to_be16(XFS_LBSIZE(mp));
 	/* handle truncation gracefully */
 	if (!hdr_d->firstused) {
-		INT_SET(hdr_d->firstused, ARCH_CONVERT,
+		hdr_d->firstused = cpu_to_be16(
 				XFS_LBSIZE(mp) - XFS_ATTR_LEAF_NAME_ALIGN);
 	}
 	hdr_d->usedbytes = 0;
 	hdr_d->count = 0;
 	hdr_d->holes = 0;
-	INT_SET(hdr_d->freemap[0].base, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_hdr_t));
-	INT_SET(hdr_d->freemap[0].size, ARCH_CONVERT,
-				INT_GET(hdr_d->firstused, ARCH_CONVERT)
-			      - INT_GET(hdr_d->freemap[0].base, ARCH_CONVERT));
+	hdr_d->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
+	hdr_d->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr_d->firstused) -
+					     sizeof(xfs_attr_leaf_hdr_t));
 
 	/*
 	 * Copy all entry's in the same (sorted) order,
 	 * but allocate name/value pairs packed and in sequence.
 	 */
 	xfs_attr_leaf_moveents(leaf_s, 0, leaf_d, 0,
-				(int)INT_GET(hdr_s->count, ARCH_CONVERT), mp);
-
+				be16_to_cpu(hdr_s->count), mp);
 	xfs_da_log_buf(trans, bp, 0, XFS_LBSIZE(mp) - 1);
 
 	kmem_free(tmpbuffer, XFS_LBSIZE(mp));
@@ -1279,10 +1250,8 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	ASSERT(blk2->magic == XFS_ATTR_LEAF_MAGIC);
 	leaf1 = blk1->bp->data;
 	leaf2 = blk2->bp->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	args = state->args;
 
 	/*
@@ -1319,22 +1288,21 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	/*
 	 * Move any entries required from leaf to leaf:
 	 */
-	if (count < INT_GET(hdr1->count, ARCH_CONVERT)) {
+	if (count < be16_to_cpu(hdr1->count)) {
 		/*
 		 * Figure the total bytes to be added to the destination leaf.
 		 */
 		/* number entries being moved */
-		count = INT_GET(hdr1->count, ARCH_CONVERT) - count;
-		space  = INT_GET(hdr1->usedbytes, ARCH_CONVERT) - totallen;
+		count = be16_to_cpu(hdr1->count) - count;
+		space  = be16_to_cpu(hdr1->usedbytes) - totallen;
 		space += count * sizeof(xfs_attr_leaf_entry_t);
 
 		/*
 		 * leaf2 is the destination, compact it if it looks tight.
 		 */
-		max  = INT_GET(hdr2->firstused, ARCH_CONVERT)
+		max  = be16_to_cpu(hdr2->firstused)
 						- sizeof(xfs_attr_leaf_hdr_t);
-		max -= INT_GET(hdr2->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t);
+		max -= be16_to_cpu(hdr2->count) * sizeof(xfs_attr_leaf_entry_t);
 		if (space > max) {
 			xfs_attr_leaf_compact(args->trans, blk2->bp);
 		}
@@ -1342,13 +1310,12 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		/*
 		 * Move high entries from leaf1 to low end of leaf2.
 		 */
-		xfs_attr_leaf_moveents(leaf1,
-				INT_GET(hdr1->count, ARCH_CONVERT)-count,
+		xfs_attr_leaf_moveents(leaf1, be16_to_cpu(hdr1->count) - count,
 				leaf2, 0, count, state->mp);
 
 		xfs_da_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
 		xfs_da_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
-	} else if (count > INT_GET(hdr1->count, ARCH_CONVERT)) {
+	} else if (count > be16_to_cpu(hdr1->count)) {
 		/*
 		 * I assert that since all callers pass in an empty
 		 * second buffer, this code should never execute.
@@ -1358,17 +1325,16 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		 * Figure the total bytes to be added to the destination leaf.
 		 */
 		/* number entries being moved */
-		count -= INT_GET(hdr1->count, ARCH_CONVERT);
-		space  = totallen - INT_GET(hdr1->usedbytes, ARCH_CONVERT);
+		count -= be16_to_cpu(hdr1->count);
+		space  = totallen - be16_to_cpu(hdr1->usedbytes);
 		space += count * sizeof(xfs_attr_leaf_entry_t);
 
 		/*
 		 * leaf1 is the destination, compact it if it looks tight.
 		 */
-		max  = INT_GET(hdr1->firstused, ARCH_CONVERT)
+		max  = be16_to_cpu(hdr1->firstused)
 						- sizeof(xfs_attr_leaf_hdr_t);
-		max -= INT_GET(hdr1->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t);
+		max -= be16_to_cpu(hdr1->count) * sizeof(xfs_attr_leaf_entry_t);
 		if (space > max) {
 			xfs_attr_leaf_compact(args->trans, blk1->bp);
 		}
@@ -1377,8 +1343,7 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		 * Move low entries from leaf2 to high end of leaf1.
 		 */
 		xfs_attr_leaf_moveents(leaf2, 0, leaf1,
-				(int)INT_GET(hdr1->count, ARCH_CONVERT), count,
-				state->mp);
+				be16_to_cpu(hdr1->count), count, state->mp);
 
 		xfs_da_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
 		xfs_da_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
@@ -1387,12 +1352,10 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	/*
 	 * Copy out last hashval in each block for B-tree code.
 	 */
-	blk1->hashval =
-	    INT_GET(leaf1->entries[INT_GET(leaf1->hdr.count,
-				    ARCH_CONVERT)-1].hashval, ARCH_CONVERT);
-	blk2->hashval =
-	    INT_GET(leaf2->entries[INT_GET(leaf2->hdr.count,
-				    ARCH_CONVERT)-1].hashval, ARCH_CONVERT);
+	blk1->hashval = be32_to_cpu(
+		leaf1->entries[be16_to_cpu(leaf1->hdr.count)-1].hashval);
+	blk2->hashval = be32_to_cpu(
+		leaf2->entries[be16_to_cpu(leaf2->hdr.count)-1].hashval);
 
 	/*
 	 * Adjust the expected index for insertion.
@@ -1406,13 +1369,12 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * inserting.  The index/blkno fields refer to the "old" entry,
 	 * while the index2/blkno2 fields refer to the "new" entry.
 	 */
-	if (blk1->index > INT_GET(leaf1->hdr.count, ARCH_CONVERT)) {
+	if (blk1->index > be16_to_cpu(leaf1->hdr.count)) {
 		ASSERT(state->inleaf == 0);
-		blk2->index = blk1->index
-				- INT_GET(leaf1->hdr.count, ARCH_CONVERT);
+		blk2->index = blk1->index - be16_to_cpu(leaf1->hdr.count);
 		args->index = args->index2 = blk2->index;
 		args->blkno = args->blkno2 = blk2->blkno;
-	} else if (blk1->index == INT_GET(leaf1->hdr.count, ARCH_CONVERT)) {
+	} else if (blk1->index == be16_to_cpu(leaf1->hdr.count)) {
 		if (state->inleaf) {
 			args->index = blk1->index;
 			args->blkno = blk1->blkno;
@@ -1420,7 +1382,7 @@ xfs_attr_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 			args->blkno2 = blk2->blkno;
 		} else {
 			blk2->index = blk1->index
-				    - INT_GET(leaf1->hdr.count, ARCH_CONVERT);
+				    - be16_to_cpu(leaf1->hdr.count);
 			args->index = args->index2 = blk2->index;
 			args->blkno = args->blkno2 = blk2->blkno;
 		}
@@ -1464,15 +1426,14 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 	 * Examine entries until we reduce the absolute difference in
 	 * byte usage between the two blocks to a minimum.
 	 */
-	max = INT_GET(hdr1->count, ARCH_CONVERT)
-			+ INT_GET(hdr2->count, ARCH_CONVERT);
+	max = be16_to_cpu(hdr1->count) + be16_to_cpu(hdr2->count);
 	half  = (max+1) * sizeof(*entry);
-	half += INT_GET(hdr1->usedbytes, ARCH_CONVERT)
-				+ INT_GET(hdr2->usedbytes, ARCH_CONVERT)
-				+ xfs_attr_leaf_newentsize(
-						state->args->namelen,
-						state->args->valuelen,
-						state->blocksize, NULL);
+	half += be16_to_cpu(hdr1->usedbytes) +
+		be16_to_cpu(hdr2->usedbytes) +
+		xfs_attr_leaf_newentsize(
+				state->args->namelen,
+				state->args->valuelen,
+				state->blocksize, NULL);
 	half /= 2;
 	lastdelta = state->blocksize;
 	entry = &leaf1->entries[0];
@@ -1498,7 +1459,7 @@ xfs_attr_leaf_figure_balance(xfs_da_state_t *state,
 		/*
 		 * Wrap around into the second block if necessary.
 		 */
-		if (count == INT_GET(hdr1->count, ARCH_CONVERT)) {
+		if (count == be16_to_cpu(hdr1->count)) {
 			leaf1 = leaf2;
 			entry = &leaf1->entries[0];
 			index = 0;
@@ -1566,12 +1527,12 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC);
 	leaf = (xfs_attr_leafblock_t *)info;
-	count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+	count = be16_to_cpu(leaf->hdr.count);
 	bytes = sizeof(xfs_attr_leaf_hdr_t) +
 		count * sizeof(xfs_attr_leaf_entry_t) +
-		INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+		be16_to_cpu(leaf->hdr.usedbytes);
 	if (bytes > (state->blocksize >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);
@@ -1580,7 +1541,7 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	/*
 	 * Check for the degenerate case of the block being empty.
 	 * If the block is empty, we'll simply delete it, no need to
-	 * coalesce it with a sibling block.  We choose (aribtrarily)
+	 * coalesce it with a sibling block.  We choose (arbitrarily)
 	 * to merge with the forward block unless it is NULL.
 	 */
 	if (count == 0) {
@@ -1588,7 +1549,7 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward,
 						 0, &retval);
@@ -1610,13 +1571,12 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * to shrink an attribute list over time.
 	 */
 	/* start with smaller blk num */
-	forward = (INT_GET(info->forw, ARCH_CONVERT)
-					< INT_GET(info->back, ARCH_CONVERT));
+	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));
 	for (i = 0; i < 2; forward = !forward, i++) {
 		if (forward)
-			blkno = INT_GET(info->forw, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->forw);
 		else
-			blkno = INT_GET(info->back, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		error = xfs_da_read_buf(state->args->trans, state->args->dp,
@@ -1626,14 +1586,13 @@ xfs_attr_leaf_toosmall(xfs_da_state_t *state, int *action)
 		ASSERT(bp != NULL);
 
 		leaf = (xfs_attr_leafblock_t *)info;
-		count  = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		count  = be16_to_cpu(leaf->hdr.count);
 		bytes  = state->blocksize - (state->blocksize>>2);
-		bytes -= INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+		bytes -= be16_to_cpu(leaf->hdr.usedbytes);
 		leaf = bp->data;
-		ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-		count += INT_GET(leaf->hdr.count, ARCH_CONVERT);
-		bytes -= INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+		count += be16_to_cpu(leaf->hdr.count);
+		bytes -= be16_to_cpu(leaf->hdr.usedbytes);
 		bytes -= count * sizeof(xfs_attr_leaf_entry_t);
 		bytes -= sizeof(xfs_attr_leaf_hdr_t);
 		xfs_da_brelse(state->args->trans, bp);
@@ -1685,21 +1644,18 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	xfs_mount_t *mp;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	mp = args->trans->t_mountp;
-	ASSERT((INT_GET(hdr->count, ARCH_CONVERT) > 0)
-		&& (INT_GET(hdr->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8)));
+	ASSERT((be16_to_cpu(hdr->count) > 0)
+		&& (be16_to_cpu(hdr->count) < (XFS_LBSIZE(mp)/8)));
 	ASSERT((args->index >= 0)
-		&& (args->index < INT_GET(hdr->count, ARCH_CONVERT)));
-	ASSERT(INT_GET(hdr->firstused, ARCH_CONVERT)
-				>= ((INT_GET(hdr->count, ARCH_CONVERT)
-					* sizeof(*entry))+sizeof(*hdr)));
+		&& (args->index < be16_to_cpu(hdr->count)));
+	ASSERT(be16_to_cpu(hdr->firstused) >=
+	       ((be16_to_cpu(hdr->count) * sizeof(*entry)) + sizeof(*hdr)));
 	entry = &leaf->entries[args->index];
-	ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT)
-				>= INT_GET(hdr->firstused, ARCH_CONVERT));
-	ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT) < XFS_LBSIZE(mp));
+	ASSERT(be16_to_cpu(entry->nameidx) >= be16_to_cpu(hdr->firstused));
+	ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
 
 	/*
 	 * Scan through free region table:
@@ -1707,33 +1663,30 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 *    find smallest free region in case we need to replace it,
 	 *    adjust any map that borders the entry table,
 	 */
-	tablesize = INT_GET(hdr->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t)
+	tablesize = be16_to_cpu(hdr->count) * sizeof(xfs_attr_leaf_entry_t)
 					+ sizeof(xfs_attr_leaf_hdr_t);
 	map = &hdr->freemap[0];
-	tmp = INT_GET(map->size, ARCH_CONVERT);
+	tmp = be16_to_cpu(map->size);
 	before = after = -1;
 	smallest = XFS_ATTR_LEAF_MAPSIZE - 1;
 	entsize = xfs_attr_leaf_entsize(leaf, args->index);
 	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; map++, i++) {
-		ASSERT(INT_GET(map->base, ARCH_CONVERT) < XFS_LBSIZE(mp));
-		ASSERT(INT_GET(map->size, ARCH_CONVERT) < XFS_LBSIZE(mp));
-		if (INT_GET(map->base, ARCH_CONVERT) == tablesize) {
-			INT_MOD(map->base, ARCH_CONVERT,
-					-sizeof(xfs_attr_leaf_entry_t));
-			INT_MOD(map->size, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_entry_t));
+		ASSERT(be16_to_cpu(map->base) < XFS_LBSIZE(mp));
+		ASSERT(be16_to_cpu(map->size) < XFS_LBSIZE(mp));
+		if (be16_to_cpu(map->base) == tablesize) {
+			be16_add(&map->base,
+				 -((int)sizeof(xfs_attr_leaf_entry_t)));
+			be16_add(&map->size, sizeof(xfs_attr_leaf_entry_t));
 		}
 
-		if ((INT_GET(map->base, ARCH_CONVERT)
-					+ INT_GET(map->size, ARCH_CONVERT))
-				== INT_GET(entry->nameidx, ARCH_CONVERT)) {
+		if ((be16_to_cpu(map->base) + be16_to_cpu(map->size))
+				== be16_to_cpu(entry->nameidx)) {
 			before = i;
-		} else if (INT_GET(map->base, ARCH_CONVERT)
-			== (INT_GET(entry->nameidx, ARCH_CONVERT) + entsize)) {
+		} else if (be16_to_cpu(map->base)
+			== (be16_to_cpu(entry->nameidx) + entsize)) {
 			after = i;
-		} else if (INT_GET(map->size, ARCH_CONVERT) < tmp) {
-			tmp = INT_GET(map->size, ARCH_CONVERT);
+		} else if (be16_to_cpu(map->size) < tmp) {
+			tmp = be16_to_cpu(map->size);
 			smallest = i;
 		}
 	}
@@ -1745,38 +1698,35 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	if ((before >= 0) || (after >= 0)) {
 		if ((before >= 0) && (after >= 0)) {
 			map = &hdr->freemap[before];
-			INT_MOD(map->size, ARCH_CONVERT, entsize);
-			INT_MOD(map->size, ARCH_CONVERT,
-				INT_GET(hdr->freemap[after].size,
-							ARCH_CONVERT));
+			be16_add(&map->size, entsize);
+			be16_add(&map->size,
+				 be16_to_cpu(hdr->freemap[after].size));
 			hdr->freemap[after].base = 0;
 			hdr->freemap[after].size = 0;
 		} else if (before >= 0) {
 			map = &hdr->freemap[before];
-			INT_MOD(map->size, ARCH_CONVERT, entsize);
+			be16_add(&map->size, entsize);
 		} else {
 			map = &hdr->freemap[after];
 			/* both on-disk, don't endian flip twice */
 			map->base = entry->nameidx;
-			INT_MOD(map->size, ARCH_CONVERT, entsize);
+			be16_add(&map->size, entsize);
 		}
 	} else {
 		/*
 		 * Replace smallest region (if it is smaller than free'd entry)
 		 */
 		map = &hdr->freemap[smallest];
-		if (INT_GET(map->size, ARCH_CONVERT) < entsize) {
-			INT_SET(map->base, ARCH_CONVERT,
-					INT_GET(entry->nameidx, ARCH_CONVERT));
-			INT_SET(map->size, ARCH_CONVERT, entsize);
+		if (be16_to_cpu(map->size) < entsize) {
+			map->base = cpu_to_be16(be16_to_cpu(entry->nameidx));
+			map->size = cpu_to_be16(entsize);
 		}
 	}
 
 	/*
 	 * Did we remove the first entry?
 	 */
-	if (INT_GET(entry->nameidx, ARCH_CONVERT)
-				== INT_GET(hdr->firstused, ARCH_CONVERT))
+	if (be16_to_cpu(entry->nameidx) == be16_to_cpu(hdr->firstused))
 		smallest = 1;
 	else
 		smallest = 0;
@@ -1785,18 +1735,18 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * Compress the remaining entries and zero out the removed stuff.
 	 */
 	memset(XFS_ATTR_LEAF_NAME(leaf, args->index), 0, entsize);
-	INT_MOD(hdr->usedbytes, ARCH_CONVERT, -entsize);
+	be16_add(&hdr->usedbytes, -entsize);
 	xfs_da_log_buf(args->trans, bp,
 	     XFS_DA_LOGRANGE(leaf, XFS_ATTR_LEAF_NAME(leaf, args->index),
 				   entsize));
 
-	tmp = (INT_GET(hdr->count, ARCH_CONVERT) - args->index)
+	tmp = (be16_to_cpu(hdr->count) - args->index)
 					* sizeof(xfs_attr_leaf_entry_t);
 	memmove((char *)entry, (char *)(entry+1), tmp);
-	INT_MOD(hdr->count, ARCH_CONVERT, -1);
+	be16_add(&hdr->count, -1);
 	xfs_da_log_buf(args->trans, bp,
 	    XFS_DA_LOGRANGE(leaf, entry, tmp + sizeof(*entry)));
-	entry = &leaf->entries[INT_GET(hdr->count, ARCH_CONVERT)];
+	entry = &leaf->entries[be16_to_cpu(hdr->count)];
 	memset((char *)entry, 0, sizeof(xfs_attr_leaf_entry_t));
 
 	/*
@@ -1808,18 +1758,17 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	if (smallest) {
 		tmp = XFS_LBSIZE(mp);
 		entry = &leaf->entries[0];
-		for (i = INT_GET(hdr->count, ARCH_CONVERT)-1;
-						i >= 0; entry++, i--) {
-			ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT)
-				>= INT_GET(hdr->firstused, ARCH_CONVERT));
-			ASSERT(INT_GET(entry->nameidx, ARCH_CONVERT)
-							< XFS_LBSIZE(mp));
-			if (INT_GET(entry->nameidx, ARCH_CONVERT) < tmp)
-				tmp = INT_GET(entry->nameidx, ARCH_CONVERT);
+		for (i = be16_to_cpu(hdr->count)-1; i >= 0; entry++, i--) {
+			ASSERT(be16_to_cpu(entry->nameidx) >=
+			       be16_to_cpu(hdr->firstused));
+			ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
+
+			if (be16_to_cpu(entry->nameidx) < tmp)
+				tmp = be16_to_cpu(entry->nameidx);
 		}
-		INT_SET(hdr->firstused, ARCH_CONVERT, tmp);
+		hdr->firstused = cpu_to_be16(tmp);
 		if (!hdr->firstused) {
-			INT_SET(hdr->firstused, ARCH_CONVERT,
+			hdr->firstused = cpu_to_be16(
 					tmp - XFS_ATTR_LEAF_NAME_ALIGN);
 		}
 	} else {
@@ -1833,9 +1782,8 @@ xfs_attr_leaf_remove(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	 * "join" the leaf with a sibling if so.
 	 */
 	tmp  = sizeof(xfs_attr_leaf_hdr_t);
-	tmp += INT_GET(leaf->hdr.count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t);
-	tmp += INT_GET(leaf->hdr.usedbytes, ARCH_CONVERT);
+	tmp += be16_to_cpu(leaf->hdr.count) * sizeof(xfs_attr_leaf_entry_t);
+	tmp += be16_to_cpu(leaf->hdr.usedbytes);
 	return(tmp < mp->m_attr_magicpct); /* leaf is < 37% full */
 }
 
@@ -1859,20 +1807,16 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	ASSERT(save_blk->magic == XFS_ATTR_LEAF_MAGIC);
 	drop_leaf = drop_blk->bp->data;
 	save_leaf = save_blk->bp->data;
-	ASSERT(INT_GET(drop_leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(save_leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	drop_hdr = &drop_leaf->hdr;
 	save_hdr = &save_leaf->hdr;
 
 	/*
 	 * Save last hashval from dying block for later Btree fixup.
 	 */
-	drop_blk->hashval =
-		INT_GET(drop_leaf->entries[INT_GET(drop_leaf->hdr.count,
-						ARCH_CONVERT)-1].hashval,
-								ARCH_CONVERT);
+	drop_blk->hashval = be32_to_cpu(
+		drop_leaf->entries[be16_to_cpu(drop_leaf->hdr.count)-1].hashval);
 
 	/*
 	 * Check if we need a temp buffer, or can we do it in place.
@@ -1886,12 +1830,11 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 		 */
 		if (xfs_attr_leaf_order(save_blk->bp, drop_blk->bp)) {
 			xfs_attr_leaf_moveents(drop_leaf, 0, save_leaf, 0,
-			     (int)INT_GET(drop_hdr->count, ARCH_CONVERT), mp);
+			     be16_to_cpu(drop_hdr->count), mp);
 		} else {
 			xfs_attr_leaf_moveents(drop_leaf, 0, save_leaf,
-				  INT_GET(save_hdr->count, ARCH_CONVERT),
-				  (int)INT_GET(drop_hdr->count, ARCH_CONVERT),
-				  mp);
+				  be16_to_cpu(save_hdr->count),
+				  be16_to_cpu(drop_hdr->count), mp);
 		}
 	} else {
 		/*
@@ -1905,28 +1848,24 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 		tmp_hdr = &tmp_leaf->hdr;
 		tmp_hdr->info = save_hdr->info;	/* struct copy */
 		tmp_hdr->count = 0;
-		INT_SET(tmp_hdr->firstused, ARCH_CONVERT, state->blocksize);
+		tmp_hdr->firstused = cpu_to_be16(state->blocksize);
 		if (!tmp_hdr->firstused) {
-			INT_SET(tmp_hdr->firstused, ARCH_CONVERT,
+			tmp_hdr->firstused = cpu_to_be16(
 				state->blocksize - XFS_ATTR_LEAF_NAME_ALIGN);
 		}
 		tmp_hdr->usedbytes = 0;
 		if (xfs_attr_leaf_order(save_blk->bp, drop_blk->bp)) {
 			xfs_attr_leaf_moveents(drop_leaf, 0, tmp_leaf, 0,
-				(int)INT_GET(drop_hdr->count, ARCH_CONVERT),
-				mp);
+				be16_to_cpu(drop_hdr->count), mp);
 			xfs_attr_leaf_moveents(save_leaf, 0, tmp_leaf,
-				  INT_GET(tmp_leaf->hdr.count, ARCH_CONVERT),
-				 (int)INT_GET(save_hdr->count, ARCH_CONVERT),
-				 mp);
+				  be16_to_cpu(tmp_leaf->hdr.count),
+				  be16_to_cpu(save_hdr->count), mp);
 		} else {
 			xfs_attr_leaf_moveents(save_leaf, 0, tmp_leaf, 0,
-				(int)INT_GET(save_hdr->count, ARCH_CONVERT),
-				mp);
+				be16_to_cpu(save_hdr->count), mp);
 			xfs_attr_leaf_moveents(drop_leaf, 0, tmp_leaf,
-				INT_GET(tmp_leaf->hdr.count, ARCH_CONVERT),
-				(int)INT_GET(drop_hdr->count, ARCH_CONVERT),
-				mp);
+				be16_to_cpu(tmp_leaf->hdr.count),
+				be16_to_cpu(drop_hdr->count), mp);
 		}
 		memcpy((char *)save_leaf, (char *)tmp_leaf, state->blocksize);
 		kmem_free(tmpbuffer, state->blocksize);
@@ -1938,10 +1877,8 @@ xfs_attr_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	/*
 	 * Copy out last hashval in each block for B-tree code.
 	 */
-	save_blk->hashval =
-		INT_GET(save_leaf->entries[INT_GET(save_leaf->hdr.count,
-						ARCH_CONVERT)-1].hashval,
-								ARCH_CONVERT);
+	save_blk->hashval = be32_to_cpu(
+		save_leaf->entries[be16_to_cpu(save_leaf->hdr.count)-1].hashval);
 }
 
 /*========================================================================
@@ -1972,48 +1909,45 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	xfs_dahash_t hashval;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT)
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.count)
 					< (XFS_LBSIZE(args->dp->i_mount)/8));
 
 	/*
 	 * Binary search.  (note: small blocks will skip this loop)
 	 */
 	hashval = args->hashval;
-	probe = span = INT_GET(leaf->hdr.count, ARCH_CONVERT) / 2;
+	probe = span = be16_to_cpu(leaf->hdr.count) / 2;
 	for (entry = &leaf->entries[probe]; span > 4;
 		   entry = &leaf->entries[probe]) {
 		span /= 2;
-		if (INT_GET(entry->hashval, ARCH_CONVERT) < hashval)
+		if (be32_to_cpu(entry->hashval) < hashval)
 			probe += span;
-		else if (INT_GET(entry->hashval, ARCH_CONVERT) > hashval)
+		else if (be32_to_cpu(entry->hashval) > hashval)
 			probe -= span;
 		else
 			break;
 	}
 	ASSERT((probe >= 0) && 
 	       (!leaf->hdr.count
-	       || (probe < INT_GET(leaf->hdr.count, ARCH_CONVERT))));
-	ASSERT((span <= 4) || (INT_GET(entry->hashval, ARCH_CONVERT)
-							== hashval));
+	       || (probe < be16_to_cpu(leaf->hdr.count))));
+	ASSERT((span <= 4) || (be32_to_cpu(entry->hashval) == hashval));
 
 	/*
 	 * Since we may have duplicate hashval's, find the first matching
 	 * hashval in the leaf.
 	 */
-	while ((probe > 0) && (INT_GET(entry->hashval, ARCH_CONVERT)
-							>= hashval)) {
+	while ((probe > 0) && (be32_to_cpu(entry->hashval) >= hashval)) {
 		entry--;
 		probe--;
 	}
-	while ((probe < INT_GET(leaf->hdr.count, ARCH_CONVERT))
-		&& (INT_GET(entry->hashval, ARCH_CONVERT) < hashval)) {
+	while ((probe < be16_to_cpu(leaf->hdr.count)) &&
+	       (be32_to_cpu(entry->hashval) < hashval)) {
 		entry++;
 		probe++;
 	}
-	if ((probe == INT_GET(leaf->hdr.count, ARCH_CONVERT))
-		    || (INT_GET(entry->hashval, ARCH_CONVERT) != hashval)) {
+	if ((probe == be16_to_cpu(leaf->hdr.count)) ||
+	    (be32_to_cpu(entry->hashval) != hashval)) {
 		args->index = probe;
 		return(XFS_ERROR(ENOATTR));
 	}
@@ -2021,8 +1955,8 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	/*
 	 * Duplicate keys may be present, so search all of them for a match.
 	 */
-	for (  ; (probe < INT_GET(leaf->hdr.count, ARCH_CONVERT))
-			&& (INT_GET(entry->hashval, ARCH_CONVERT) == hashval);
+	for (  ; (probe < be16_to_cpu(leaf->hdr.count)) &&
+			(be32_to_cpu(entry->hashval) == hashval);
 			entry++, probe++) {
 /*
  * GROT: Add code to remove incomplete entries.
@@ -2064,11 +1998,9 @@ xfs_attr_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args)
 			    ((entry->flags & XFS_ATTR_ROOT) != 0))
 				continue;
 			args->index = probe;
-			args->rmtblkno
-				  = INT_GET(name_rmt->valueblk, ARCH_CONVERT);
+			args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
 			args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount,
-						   INT_GET(name_rmt->valuelen,
-								ARCH_CONVERT));
+						   be32_to_cpu(name_rmt->valuelen));
 			return(XFS_ERROR(EEXIST));
 		}
 	}
@@ -2090,18 +2022,17 @@ xfs_attr_leaf_getvalue(xfs_dabuf_t *bp, xfs_da_args_t *args)
 	xfs_attr_leaf_name_remote_t *name_rmt;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT)
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.count)
 					< (XFS_LBSIZE(args->dp->i_mount)/8));
-	ASSERT(args->index < ((int)INT_GET(leaf->hdr.count, ARCH_CONVERT)));
+	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
 
 	entry = &leaf->entries[args->index];
 	if (entry->flags & XFS_ATTR_LOCAL) {
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, args->index);
 		ASSERT(name_loc->namelen == args->namelen);
 		ASSERT(memcmp(args->name, name_loc->nameval, args->namelen) == 0);
-		valuelen = INT_GET(name_loc->valuelen, ARCH_CONVERT);
+		valuelen = be16_to_cpu(name_loc->valuelen);
 		if (args->flags & ATTR_KERNOVAL) {
 			args->valuelen = valuelen;
 			return(0);
@@ -2116,8 +2047,8 @@ xfs_attr_leaf_getvalue(xfs_dabuf_t *bp, xfs_da_args_t *args)
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
 		ASSERT(name_rmt->namelen == args->namelen);
 		ASSERT(memcmp(args->name, name_rmt->name, args->namelen) == 0);
-		valuelen = INT_GET(name_rmt->valuelen, ARCH_CONVERT);
-		args->rmtblkno = INT_GET(name_rmt->valueblk, ARCH_CONVERT);
+		valuelen = be32_to_cpu(name_rmt->valuelen);
+		args->rmtblkno = be32_to_cpu(name_rmt->valueblk);
 		args->rmtblkcnt = XFS_B_TO_FSB(args->dp->i_mount, valuelen);
 		if (args->flags & ATTR_KERNOVAL) {
 			args->valuelen = valuelen;
@@ -2159,32 +2090,29 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Set up environment.
 	 */
-	ASSERT(INT_GET(leaf_s->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf_d->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_s->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_d->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	hdr_s = &leaf_s->hdr;
 	hdr_d = &leaf_d->hdr;
-	ASSERT((INT_GET(hdr_s->count, ARCH_CONVERT) > 0)
-				&& (INT_GET(hdr_s->count, ARCH_CONVERT)
-						< (XFS_LBSIZE(mp)/8)));
-	ASSERT(INT_GET(hdr_s->firstused, ARCH_CONVERT) >=
-		((INT_GET(hdr_s->count, ARCH_CONVERT)
+	ASSERT((be16_to_cpu(hdr_s->count) > 0) &&
+	       (be16_to_cpu(hdr_s->count) < (XFS_LBSIZE(mp)/8)));
+	ASSERT(be16_to_cpu(hdr_s->firstused) >=
+		((be16_to_cpu(hdr_s->count)
 					* sizeof(*entry_s))+sizeof(*hdr_s)));
-	ASSERT(INT_GET(hdr_d->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8));
-	ASSERT(INT_GET(hdr_d->firstused, ARCH_CONVERT) >=
-		((INT_GET(hdr_d->count, ARCH_CONVERT)
+	ASSERT(be16_to_cpu(hdr_d->count) < (XFS_LBSIZE(mp)/8));
+	ASSERT(be16_to_cpu(hdr_d->firstused) >=
+		((be16_to_cpu(hdr_d->count)
 					* sizeof(*entry_d))+sizeof(*hdr_d)));
 
-	ASSERT(start_s < INT_GET(hdr_s->count, ARCH_CONVERT));
-	ASSERT(start_d <= INT_GET(hdr_d->count, ARCH_CONVERT));
-	ASSERT(count <= INT_GET(hdr_s->count, ARCH_CONVERT));
+	ASSERT(start_s < be16_to_cpu(hdr_s->count));
+	ASSERT(start_d <= be16_to_cpu(hdr_d->count));
+	ASSERT(count <= be16_to_cpu(hdr_s->count));
 
 	/*
 	 * Move the entries in the destination leaf up to make a hole?
 	 */
-	if (start_d < INT_GET(hdr_d->count, ARCH_CONVERT)) {
-		tmp  = INT_GET(hdr_d->count, ARCH_CONVERT) - start_d;
+	if (start_d < be16_to_cpu(hdr_d->count)) {
+		tmp  = be16_to_cpu(hdr_d->count) - start_d;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &leaf_d->entries[start_d];
 		entry_d = &leaf_d->entries[start_d + count];
@@ -2199,8 +2127,8 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	entry_d = &leaf_d->entries[start_d];
 	desti = start_d;
 	for (i = 0; i < count; entry_s++, entry_d++, desti++, i++) {
-		ASSERT(INT_GET(entry_s->nameidx, ARCH_CONVERT)
-				>= INT_GET(hdr_s->firstused, ARCH_CONVERT));
+		ASSERT(be16_to_cpu(entry_s->nameidx)
+				>= be16_to_cpu(hdr_s->firstused));
 		tmp = xfs_attr_leaf_entsize(leaf_s, start_s + i);
 #ifdef GROT
 		/*
@@ -2210,35 +2138,35 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 		 */
 		if (entry_s->flags & XFS_ATTR_INCOMPLETE) { /* skip partials? */
 			memset(XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), 0, tmp);
-			INT_MOD(hdr_s->usedbytes, ARCH_CONVERT, -tmp);
-			INT_MOD(hdr_s->count, ARCH_CONVERT, -1);
+			be16_add(&hdr_s->usedbytes, -tmp);
+			be16_add(&hdr_s->count, -1);
 			entry_d--;	/* to compensate for ++ in loop hdr */
 			desti--;
 			if ((start_s + i) < offset)
 				result++;	/* insertion index adjustment */
 		} else {
 #endif /* GROT */
-			INT_MOD(hdr_d->firstused, ARCH_CONVERT, -tmp);
+			be16_add(&hdr_d->firstused, -tmp);
 			/* both on-disk, don't endian flip twice */
 			entry_d->hashval = entry_s->hashval;
 			/* both on-disk, don't endian flip twice */
 			entry_d->nameidx = hdr_d->firstused;
 			entry_d->flags = entry_s->flags;
-			ASSERT(INT_GET(entry_d->nameidx, ARCH_CONVERT) + tmp
+			ASSERT(be16_to_cpu(entry_d->nameidx) + tmp
 							<= XFS_LBSIZE(mp));
 			memmove(XFS_ATTR_LEAF_NAME(leaf_d, desti),
 				XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), tmp);
-			ASSERT(INT_GET(entry_s->nameidx, ARCH_CONVERT) + tmp
+			ASSERT(be16_to_cpu(entry_s->nameidx) + tmp
 							<= XFS_LBSIZE(mp));
 			memset(XFS_ATTR_LEAF_NAME(leaf_s, start_s + i), 0, tmp);
-			INT_MOD(hdr_s->usedbytes, ARCH_CONVERT, -tmp);
-			INT_MOD(hdr_d->usedbytes, ARCH_CONVERT, tmp);
-			INT_MOD(hdr_s->count, ARCH_CONVERT, -1);
-			INT_MOD(hdr_d->count, ARCH_CONVERT, 1);
-			tmp = INT_GET(hdr_d->count, ARCH_CONVERT)
+			be16_add(&hdr_s->usedbytes, -tmp);
+			be16_add(&hdr_d->usedbytes, tmp);
+			be16_add(&hdr_s->count, -1);
+			be16_add(&hdr_d->count, 1);
+			tmp = be16_to_cpu(hdr_d->count)
 						* sizeof(xfs_attr_leaf_entry_t)
 						+ sizeof(xfs_attr_leaf_hdr_t);
-			ASSERT(INT_GET(hdr_d->firstused, ARCH_CONVERT) >= tmp);
+			ASSERT(be16_to_cpu(hdr_d->firstused) >= tmp);
 #ifdef GROT
 		}
 #endif /* GROT */
@@ -2247,7 +2175,7 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Zero out the entries we just copied.
 	 */
-	if (start_s == INT_GET(hdr_s->count, ARCH_CONVERT)) {
+	if (start_s == be16_to_cpu(hdr_s->count)) {
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &leaf_s->entries[start_s];
 		ASSERT(((char *)entry_s + tmp) <=
@@ -2258,15 +2186,14 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 		 * Move the remaining entries down to fill the hole,
 		 * then zero the entries at the top.
 		 */
-		tmp  = INT_GET(hdr_s->count, ARCH_CONVERT) - count;
+		tmp  = be16_to_cpu(hdr_s->count) - count;
 		tmp *= sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &leaf_s->entries[start_s + count];
 		entry_d = &leaf_s->entries[start_s];
 		memmove((char *)entry_d, (char *)entry_s, tmp);
 
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
-		entry_s = &leaf_s->entries[INT_GET(hdr_s->count,
-							ARCH_CONVERT)];
+		entry_s = &leaf_s->entries[be16_to_cpu(hdr_s->count)];
 		ASSERT(((char *)entry_s + tmp) <=
 		       ((char *)leaf_s + XFS_LBSIZE(mp)));
 		memset((char *)entry_s, 0, tmp);
@@ -2275,14 +2202,11 @@ xfs_attr_leaf_moveents(xfs_attr_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Fill in the freemap information
 	 */
-	INT_SET(hdr_d->freemap[0].base, ARCH_CONVERT,
-					sizeof(xfs_attr_leaf_hdr_t));
-	INT_MOD(hdr_d->freemap[0].base, ARCH_CONVERT,
-				INT_GET(hdr_d->count, ARCH_CONVERT)
-					* sizeof(xfs_attr_leaf_entry_t));
-	INT_SET(hdr_d->freemap[0].size, ARCH_CONVERT,
-				INT_GET(hdr_d->firstused, ARCH_CONVERT)
-			      - INT_GET(hdr_d->freemap[0].base, ARCH_CONVERT));
+	hdr_d->freemap[0].base = cpu_to_be16(sizeof(xfs_attr_leaf_hdr_t));
+	be16_add(&hdr_d->freemap[0].base, be16_to_cpu(hdr_d->count) *
+			sizeof(xfs_attr_leaf_entry_t));
+	hdr_d->freemap[0].size = cpu_to_be16(be16_to_cpu(hdr_d->firstused)
+			      - be16_to_cpu(hdr_d->freemap[0].base));
 	hdr_d->freemap[1].base = 0;
 	hdr_d->freemap[2].base = 0;
 	hdr_d->freemap[1].size = 0;
@@ -2301,18 +2225,16 @@ xfs_attr_leaf_order(xfs_dabuf_t *leaf1_bp, xfs_dabuf_t *leaf2_bp)
 
 	leaf1 = leaf1_bp->data;
 	leaf2 = leaf2_bp->data;
-	ASSERT((INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC) &&
-	       (INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC));
-	if (   (INT_GET(leaf1->hdr.count, ARCH_CONVERT) > 0)
-	    && (INT_GET(leaf2->hdr.count, ARCH_CONVERT) > 0)
-	    && (   (INT_GET(leaf2->entries[ 0 ].hashval, ARCH_CONVERT) <
-		      INT_GET(leaf1->entries[ 0 ].hashval, ARCH_CONVERT))
-		|| (INT_GET(leaf2->entries[INT_GET(leaf2->hdr.count,
-				ARCH_CONVERT)-1].hashval, ARCH_CONVERT) <
-		      INT_GET(leaf1->entries[INT_GET(leaf1->hdr.count,
-				ARCH_CONVERT)-1].hashval, ARCH_CONVERT))) ) {
+	ASSERT((be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC) &&
+	       (be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC));
+	if ((be16_to_cpu(leaf1->hdr.count) > 0) &&
+	    (be16_to_cpu(leaf2->hdr.count) > 0) &&
+	    ((be32_to_cpu(leaf2->entries[0].hashval) <
+	      be32_to_cpu(leaf1->entries[0].hashval)) ||
+	     (be32_to_cpu(leaf2->entries[
+			be16_to_cpu(leaf2->hdr.count)-1].hashval) <
+	      be32_to_cpu(leaf1->entries[
+			be16_to_cpu(leaf1->hdr.count)-1].hashval)))) {
 		return(1);
 	}
 	return(0);
@@ -2327,14 +2249,12 @@ xfs_attr_leaf_lasthash(xfs_dabuf_t *bp, int *count)
 	xfs_attr_leafblock_t *leaf;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	if (count)
-		*count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		*count = be16_to_cpu(leaf->hdr.count);
 	if (!leaf->hdr.count)
 		return(0);
-	return(INT_GET(leaf->entries[INT_GET(leaf->hdr.count,
-				ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
+	return be32_to_cpu(leaf->entries[be16_to_cpu(leaf->hdr.count)-1].hashval);
 }
 
 /*
@@ -2348,13 +2268,11 @@ xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index)
 	xfs_attr_leaf_name_remote_t *name_rmt;
 	int size;
 
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 	if (leaf->entries[index].flags & XFS_ATTR_LOCAL) {
 		name_loc = XFS_ATTR_LEAF_NAME_LOCAL(leaf, index);
 		size = XFS_ATTR_LEAF_ENTSIZE_LOCAL(name_loc->namelen,
-						   INT_GET(name_loc->valuelen,
-								ARCH_CONVERT));
+						   be16_to_cpu(name_loc->valuelen));
 	} else {
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, index);
 		size = XFS_ATTR_LEAF_ENTSIZE_REMOTE(name_rmt->namelen);
@@ -2412,22 +2330,20 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 	 */
 	if (context->resynch) {
 		entry = &leaf->entries[0];
-		for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT);
-							entry++, i++) {
-			if (INT_GET(entry->hashval, ARCH_CONVERT)
-							== cursor->hashval) {
+		for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+			if (be32_to_cpu(entry->hashval) == cursor->hashval) {
 				if (cursor->offset == context->dupcnt) {
 					context->dupcnt = 0;
 					break;
 				}
 				context->dupcnt++;
-			} else if (INT_GET(entry->hashval, ARCH_CONVERT)
-							> cursor->hashval) {
+			} else if (be32_to_cpu(entry->hashval) >
+					cursor->hashval) {
 				context->dupcnt = 0;
 				break;
 			}
 		}
-		if (i == INT_GET(leaf->hdr.count, ARCH_CONVERT)) {
+		if (i == be16_to_cpu(leaf->hdr.count)) {
 			xfs_attr_trace_l_c("not found", context);
 			return(0);
 		}
@@ -2441,12 +2357,12 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 	 * We have found our place, start copying out the new attributes.
 	 */
 	retval = 0;
-	for (  ; (i < INT_GET(leaf->hdr.count, ARCH_CONVERT))
+	for (  ; (i < be16_to_cpu(leaf->hdr.count))
 	     && (retval == 0); entry++, i++) {
 		attrnames_t	*namesp;
 
-		if (INT_GET(entry->hashval, ARCH_CONVERT) != cursor->hashval) {
-			cursor->hashval = INT_GET(entry->hashval, ARCH_CONVERT);
+		if (be32_to_cpu(entry->hashval) != cursor->hashval) {
+			cursor->hashval = be32_to_cpu(entry->hashval);
 			cursor->offset = 0;
 		}
 
@@ -2475,8 +2391,7 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 				retval = xfs_attr_put_listent(context, namesp,
 					(char *)name_loc->nameval,
 					(int)name_loc->namelen,
-					(int)INT_GET(name_loc->valuelen,
-								ARCH_CONVERT));
+					be16_to_cpu(name_loc->valuelen));
 			}
 		} else {
 			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
@@ -2488,8 +2403,7 @@ xfs_attr_leaf_list_int(xfs_dabuf_t *bp, xfs_attr_list_context_t *context)
 				retval = xfs_attr_put_listent(context, namesp,
 					(char *)name_rmt->name,
 					(int)name_rmt->namelen,
-					(int)INT_GET(name_rmt->valuelen,
-								ARCH_CONVERT));
+					be32_to_cpu(name_rmt->valuelen));
 			}
 		}
 		if (retval == 0) {
@@ -2596,9 +2510,8 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 	ASSERT(bp != NULL);
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < INT_GET(leaf->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
 	ASSERT(args->index >= 0);
 	entry = &leaf->entries[ args->index ];
 	ASSERT(entry->flags & XFS_ATTR_INCOMPLETE);
@@ -2613,7 +2526,7 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 		namelen = name_rmt->namelen;
 		name = (char *)name_rmt->name;
 	}
-	ASSERT(INT_GET(entry->hashval, ARCH_CONVERT) == args->hashval);
+	ASSERT(be32_to_cpu(entry->hashval) == args->hashval);
 	ASSERT(namelen == args->namelen);
 	ASSERT(memcmp(name, args->name, namelen) == 0);
 #endif /* DEBUG */
@@ -2625,8 +2538,8 @@ xfs_attr_leaf_clearflag(xfs_da_args_t *args)
 	if (args->rmtblkno) {
 		ASSERT((entry->flags & XFS_ATTR_LOCAL) == 0);
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, args->index);
-		INT_SET(name_rmt->valueblk, ARCH_CONVERT, args->rmtblkno);
-		INT_SET(name_rmt->valuelen, ARCH_CONVERT, args->valuelen);
+		name_rmt->valueblk = cpu_to_be32(args->rmtblkno);
+		name_rmt->valuelen = cpu_to_be32(args->valuelen);
 		xfs_da_log_buf(args->trans, bp,
 			 XFS_DA_LOGRANGE(leaf, name_rmt, sizeof(*name_rmt)));
 	}
@@ -2663,9 +2576,8 @@ xfs_attr_leaf_setflag(xfs_da_args_t *args)
 	ASSERT(bp != NULL);
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < INT_GET(leaf->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index < be16_to_cpu(leaf->hdr.count));
 	ASSERT(args->index >= 0);
 	entry = &leaf->entries[ args->index ];
 
@@ -2736,16 +2648,14 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 	}
 
 	leaf1 = bp1->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index < INT_GET(leaf1->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index < be16_to_cpu(leaf1->hdr.count));
 	ASSERT(args->index >= 0);
 	entry1 = &leaf1->entries[ args->index ];
 
 	leaf2 = bp2->data;
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
-	ASSERT(args->index2 < INT_GET(leaf2->hdr.count, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
+	ASSERT(args->index2 < be16_to_cpu(leaf2->hdr.count));
 	ASSERT(args->index2 >= 0);
 	entry2 = &leaf2->entries[ args->index2 ];
 
@@ -2768,7 +2678,7 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 		namelen2 = name_rmt->namelen;
 		name2 = (char *)name_rmt->name;
 	}
-	ASSERT(INT_GET(entry1->hashval, ARCH_CONVERT) == INT_GET(entry2->hashval, ARCH_CONVERT));
+	ASSERT(be32_to_cpu(entry1->hashval) == be32_to_cpu(entry2->hashval));
 	ASSERT(namelen1 == namelen2);
 	ASSERT(memcmp(name1, name2, namelen1) == 0);
 #endif /* DEBUG */
@@ -2782,8 +2692,8 @@ xfs_attr_leaf_flipflags(xfs_da_args_t *args)
 	if (args->rmtblkno) {
 		ASSERT((entry1->flags & XFS_ATTR_LOCAL) == 0);
 		name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf1, args->index);
-		INT_SET(name_rmt->valueblk, ARCH_CONVERT, args->rmtblkno);
-		INT_SET(name_rmt->valuelen, ARCH_CONVERT, args->valuelen);
+		name_rmt->valueblk = cpu_to_be32(args->rmtblkno);
+		name_rmt->valuelen = cpu_to_be32(args->valuelen);
 		xfs_da_log_buf(args->trans, bp1,
 			 XFS_DA_LOGRANGE(leaf1, name_rmt, sizeof(*name_rmt)));
 	}
@@ -2842,9 +2752,9 @@ xfs_attr_root_inactive(xfs_trans_t **trans, xfs_inode_t *dp)
 	 * This is a depth-first traversal!
 	 */
 	info = bp->data;
-	if (INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
+	if (be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC) {
 		error = xfs_attr_node_inactive(trans, dp, bp, 1);
-	} else if (INT_GET(info->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC) {
+	} else if (be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC) {
 		error = xfs_attr_leaf_inactive(trans, dp, bp);
 	} else {
 		error = XFS_ERROR(EIO);
@@ -2892,15 +2802,14 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 	}
 
 	node = bp->data;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT)
-						== XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	parent_blkno = xfs_da_blkno(bp);	/* save for re-read later */
-	count = INT_GET(node->hdr.count, ARCH_CONVERT);
+	count = be16_to_cpu(node->hdr.count);
 	if (!count) {
 		xfs_da_brelse(*trans, bp);
 		return(0);
 	}
-	child_fsb = INT_GET(node->btree[0].before, ARCH_CONVERT);
+	child_fsb = be32_to_cpu(node->btree[0].before);
 	xfs_da_brelse(*trans, bp);	/* no locks for later trans */
 
 	/*
@@ -2927,12 +2836,10 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 			 * Invalidate the subtree, however we have to.
 			 */
 			info = child_bp->data;
-			if (INT_GET(info->magic, ARCH_CONVERT)
-							== XFS_DA_NODE_MAGIC) {
+			if (be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC) {
 				error = xfs_attr_node_inactive(trans, dp,
 						child_bp, level+1);
-			} else if (INT_GET(info->magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC) {
+			} else if (be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC) {
 				error = xfs_attr_leaf_inactive(trans, dp,
 						child_bp);
 			} else {
@@ -2962,7 +2869,7 @@ xfs_attr_node_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp,
 				&bp, XFS_ATTR_FORK);
 			if (error)
 				return(error);
-			child_fsb = INT_GET(node->btree[i+1].before, ARCH_CONVERT);
+			child_fsb = be32_to_cpu(node->btree[i+1].before);
 			xfs_da_brelse(*trans, bp);
 		}
 		/*
@@ -2991,17 +2898,16 @@ xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
 	int error, count, size, tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT)
-						== XFS_ATTR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_ATTR_LEAF_MAGIC);
 
 	/*
 	 * Count the number of "remote" value extents.
 	 */
 	count = 0;
 	entry = &leaf->entries[0];
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
-		if (   INT_GET(entry->nameidx, ARCH_CONVERT)
-		    && ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+		if (be16_to_cpu(entry->nameidx) &&
+		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
 			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
 			if (name_rmt->valueblk)
 				count++;
@@ -3027,17 +2933,14 @@ xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
 	 */
 	lp = list;
 	entry = &leaf->entries[0];
-	for (i = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); entry++, i++) {
-		if (   INT_GET(entry->nameidx, ARCH_CONVERT)
-		    && ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
+	for (i = 0; i < be16_to_cpu(leaf->hdr.count); entry++, i++) {
+		if (be16_to_cpu(entry->nameidx) &&
+		    ((entry->flags & XFS_ATTR_LOCAL) == 0)) {
 			name_rmt = XFS_ATTR_LEAF_NAME_REMOTE(leaf, i);
 			if (name_rmt->valueblk) {
-				/* both on-disk, don't endian flip twice */
-				lp->valueblk = name_rmt->valueblk;
-				INT_SET(lp->valuelen, ARCH_CONVERT,
-						XFS_B_TO_FSB(dp->i_mount,
-						    INT_GET(name_rmt->valuelen,
-							      ARCH_CONVERT)));
+				lp->valueblk = be32_to_cpu(name_rmt->valueblk);
+				lp->valuelen = XFS_B_TO_FSB(dp->i_mount,
+						    be32_to_cpu(name_rmt->valuelen));
 				lp++;
 			}
 		}
@@ -3050,10 +2953,8 @@ xfs_attr_leaf_inactive(xfs_trans_t **trans, xfs_inode_t *dp, xfs_dabuf_t *bp)
 	error = 0;
 	for (lp = list, i = 0; i < count; i++, lp++) {
 		tmp = xfs_attr_leaf_freextent(trans, dp,
-						     INT_GET(lp->valueblk,
-								ARCH_CONVERT),
-						     INT_GET(lp->valuelen,
-								ARCH_CONVERT));
+				lp->valueblk, lp->valuelen);
+
 		if (error == 0)
 			error = tmp;	/* save only the 1st errno */
 	}
diff --git a/fs/xfs/xfs_attr_leaf.h b/fs/xfs/xfs_attr_leaf.h
index 541e34109bb9..51c3ee156b2f 100644
--- a/fs/xfs/xfs_attr_leaf.h
+++ b/fs/xfs/xfs_attr_leaf.h
@@ -73,39 +73,39 @@ struct xfs_trans;
 #define XFS_ATTR_LEAF_MAPSIZE	3	/* how many freespace slots */
 
 typedef struct xfs_attr_leaf_map {	/* RLE map of free bytes */
-	__uint16_t	base;	 	/* base of free region */
-	__uint16_t	size;	  	/* length of free region */
+	__be16	base;			  /* base of free region */
+	__be16	size;			  /* length of free region */
 } xfs_attr_leaf_map_t;
 
 typedef struct xfs_attr_leaf_hdr {	/* constant-structure header block */
 	xfs_da_blkinfo_t info;		/* block type, links, etc. */
-	__uint16_t	count;		/* count of active leaf_entry's */
-	__uint16_t	usedbytes;	/* num bytes of names/values stored */
-	__uint16_t	firstused;	/* first used byte in name area */
-	__uint8_t	holes;		/* != 0 if blk needs compaction */
-	__uint8_t	pad1;
+	__be16	count;			/* count of active leaf_entry's */
+	__be16	usedbytes;		/* num bytes of names/values stored */
+	__be16	firstused;		/* first used byte in name area */
+	__u8	holes;			/* != 0 if blk needs compaction */
+	__u8	pad1;
 	xfs_attr_leaf_map_t freemap[XFS_ATTR_LEAF_MAPSIZE];
 					/* N largest free regions */
 } xfs_attr_leaf_hdr_t;
 
 typedef struct xfs_attr_leaf_entry {	/* sorted on key, not name */
-	xfs_dahash_t	hashval;	/* hash value of name */
-	__uint16_t	nameidx;	/* index into buffer of name/value */
-	__uint8_t	flags;		/* LOCAL/ROOT/SECURE/INCOMPLETE flag */
-	__uint8_t	pad2;		/* unused pad byte */
+	__be32	hashval;		/* hash value of name */
+ 	__be16	nameidx;		/* index into buffer of name/value */
+	__u8	flags;			/* LOCAL/ROOT/SECURE/INCOMPLETE flag */
+	__u8	pad2;			/* unused pad byte */
 } xfs_attr_leaf_entry_t;
 
 typedef struct xfs_attr_leaf_name_local {
-	__uint16_t	valuelen;	/* number of bytes in value */
-	__uint8_t	namelen;	/* length of name bytes */
-	__uint8_t	nameval[1];	/* name/value bytes */
+	__be16	valuelen;		/* number of bytes in value */
+	__u8	namelen;		/* length of name bytes */
+	__u8	nameval[1];		/* name/value bytes */
 } xfs_attr_leaf_name_local_t;
 
 typedef struct xfs_attr_leaf_name_remote {
-	xfs_dablk_t	valueblk;	/* block number of value bytes */
-	__uint32_t	valuelen;	/* number of bytes in value */
-	__uint8_t	namelen;	/* length of name bytes */
-	__uint8_t	name[1];	/* name bytes */
+	__be32	valueblk;		/* block number of value bytes */
+	__be32	valuelen;		/* number of bytes in value */
+	__u8	namelen;		/* length of name bytes */
+	__u8	name[1];		/* name bytes */
 } xfs_attr_leaf_name_remote_t;
 
 typedef struct xfs_attr_leafblock {
@@ -143,8 +143,8 @@ typedef struct xfs_attr_leafblock {
 static inline xfs_attr_leaf_name_remote_t *
 xfs_attr_leaf_name_remote(xfs_attr_leafblock_t *leafp, int idx)
 {
-	return (xfs_attr_leaf_name_remote_t *) &((char *)
-		(leafp))[INT_GET((leafp)->entries[idx].nameidx, ARCH_CONVERT)];
+	return (xfs_attr_leaf_name_remote_t *)
+		&((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
 }
 
 #define XFS_ATTR_LEAF_NAME_LOCAL(leafp,idx)	\
@@ -152,16 +152,15 @@ xfs_attr_leaf_name_remote(xfs_attr_leafblock_t *leafp, int idx)
 static inline xfs_attr_leaf_name_local_t *
 xfs_attr_leaf_name_local(xfs_attr_leafblock_t *leafp, int idx)
 {
-	return (xfs_attr_leaf_name_local_t *) &((char *)
-		(leafp))[INT_GET((leafp)->entries[idx].nameidx, ARCH_CONVERT)];
+	return (xfs_attr_leaf_name_local_t *)
+		&((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
 }
 
 #define XFS_ATTR_LEAF_NAME(leafp,idx)		\
 	xfs_attr_leaf_name(leafp,idx)
 static inline char *xfs_attr_leaf_name(xfs_attr_leafblock_t *leafp, int idx)
 {
-	return (&((char *)
-		(leafp))[INT_GET((leafp)->entries[idx].nameidx, ARCH_CONVERT)]);
+	return &((char *)leafp)[be16_to_cpu(leafp->entries[idx].nameidx)];
 }
 
 /*
diff --git a/fs/xfs/xfs_attr_sf.h b/fs/xfs/xfs_attr_sf.h
index ffed6ca81a52..f67f917803b1 100644
--- a/fs/xfs/xfs_attr_sf.h
+++ b/fs/xfs/xfs_attr_sf.h
@@ -32,8 +32,8 @@ struct xfs_inode;
  */
 typedef struct xfs_attr_shortform {
 	struct xfs_attr_sf_hdr {	/* constant-structure header block */
-		__uint16_t totsize;	/* total bytes in shortform list */
-		__uint8_t count;	/* count of active entries */
+		__be16	totsize;	/* total bytes in shortform list */
+		__u8	count;	/* count of active entries */
 	} hdr;
 	struct xfs_attr_sf_entry {
 		__uint8_t namelen;	/* actual length of name (no NULL) */
@@ -66,8 +66,8 @@ typedef struct xfs_attr_sf_sort {
 #define XFS_ATTR_SF_NEXTENTRY(sfep)		/* next entry in struct */ \
 	((xfs_attr_sf_entry_t *)((char *)(sfep) + XFS_ATTR_SF_ENTSIZE(sfep)))
 #define XFS_ATTR_SF_TOTSIZE(dp)			/* total space in use */ \
-	(INT_GET(((xfs_attr_shortform_t *)	\
-		((dp)->i_afp->if_u1.if_data))->hdr.totsize, ARCH_CONVERT))
+	(be16_to_cpu(((xfs_attr_shortform_t *)	\
+		((dp)->i_afp->if_u1.if_data))->hdr.totsize))
 
 #if defined(XFS_ATTR_TRACE)
 /*
diff --git a/fs/xfs/xfs_behavior.c b/fs/xfs/xfs_behavior.c
index 9880adae3938..f4fe3715a803 100644
--- a/fs/xfs/xfs_behavior.c
+++ b/fs/xfs/xfs_behavior.c
@@ -31,7 +31,7 @@
  * The behavior chain is ordered based on the 'position' number which
  * lives in the first field of the ops vector (higher numbers first).
  *
- * Attemps to insert duplicate ops result in an EINVAL return code.
+ * Attempts to insert duplicate ops result in an EINVAL return code.
  * Otherwise, return 0 to indicate success.
  */
 int
@@ -84,7 +84,7 @@ bhv_insert(bhv_head_t *bhp, bhv_desc_t *bdp)
 
 /*
  * Remove a behavior descriptor from a position in a behavior chain;
- * the postition is guaranteed not to be the first position.
+ * the position is guaranteed not to be the first position.
  * Should only be called by the bhv_remove() macro.
  */
 void
diff --git a/fs/xfs/xfs_behavior.h b/fs/xfs/xfs_behavior.h
index 2cd89bb5ab10..1d8ff103201c 100644
--- a/fs/xfs/xfs_behavior.h
+++ b/fs/xfs/xfs_behavior.h
@@ -39,7 +39,7 @@
  * behaviors is synchronized with operations-in-progress (oip's) so that
  * the oip's always see a consistent view of the chain.
  *
- * The term "interpostion" is used to refer to the act of inserting
+ * The term "interposition" is used to refer to the act of inserting
  * a behavior such that it interposes on (i.e., is inserted in front
  * of) a particular other behavior.  A key example of this is when a
  * system implementing distributed single system image wishes to
@@ -51,7 +51,7 @@
  *
  * Behavior synchronization is logic which is necessary under certain
  * circumstances that there is no conflict between ongoing operations
- * traversing the behavior chain and those dunamically modifying the
+ * traversing the behavior chain and those dynamically modifying the
  * behavior chain.  Because behavior synchronization adds extra overhead
  * to virtual operation invocation, we want to restrict, as much as
  * we can, the requirement for this extra code, to those situations
diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c
index 70625e577c70..d384e489705f 100644
--- a/fs/xfs/xfs_bmap.c
+++ b/fs/xfs/xfs_bmap.c
@@ -89,7 +89,7 @@ xfs_bmap_add_attrfork_local(
 	int			*flags);	/* inode logging flags */
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after allocating space (or doing a delayed allocation).
  */
 STATIC int				/* error */
@@ -97,7 +97,7 @@ xfs_bmap_add_extent(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	int			*logflagsp, /* inode logging flags */
@@ -113,7 +113,7 @@ xfs_bmap_add_extent_delay_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
@@ -129,7 +129,7 @@ xfs_bmap_add_extent_hole_delay(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp,/* inode logging flags */
 	int			rsvd);	/* OK to allocate reserved blocks */
 
@@ -142,7 +142,7 @@ xfs_bmap_add_extent_hole_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp, /* inode logging flags */
 	int			whichfork); /* data or attr fork */
 
@@ -155,7 +155,7 @@ xfs_bmap_add_extent_unwritten_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp); /* inode logging flags */
 
 /*
@@ -169,7 +169,7 @@ xfs_bmap_alloc(
 /*
  * Transform a btree format file with only one leaf node, where the
  * extents list will fit in the inode, into an extents format file.
- * Since the extent list is already in-core, all we have to do is
+ * Since the file extents are already in-core, all we have to do is
  * give up the space for the btree root and pitch the leaf block.
  */
 STATIC int				/* error */
@@ -191,7 +191,7 @@ xfs_bmap_check_extents(
 #endif
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after removing space (or undoing a delayed allocation).
  */
 STATIC int				/* error */
@@ -201,7 +201,7 @@ xfs_bmap_del_extent(
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp,/* inode logging flags */
 	int			whichfork, /* data or attr fork */
 	int			rsvd);	 /* OK to allocate reserved blocks */
@@ -217,18 +217,6 @@ xfs_bmap_del_free(
 	xfs_bmap_free_item_t	*free);	/* list item to be freed */
 
 /*
- * Remove count entries from the extents array for inode "ip", starting
- * at index "idx".  Copies the remaining items down over the deleted ones,
- * and gives back the excess memory.
- */
-STATIC void
-xfs_bmap_delete_exlist(
-	xfs_inode_t	*ip,		/* incode inode pointer */
-	xfs_extnum_t	idx,		/* starting delete index */
-	xfs_extnum_t	count,		/* count of items to delete */
-	int		whichfork);	/* data or attr fork */
-
-/*
  * Convert an extents-format file into a btree-format file.
  * The new file will have a root block (in the inode) and a single child block.
  */
@@ -244,18 +232,6 @@ xfs_bmap_extents_to_btree(
 	int			whichfork);	/* data or attr fork */
 
 /*
- * Insert new item(s) in the extent list for inode "ip".
- * Count new items are inserted at offset idx.
- */
-STATIC void
-xfs_bmap_insert_exlist(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* starting index of new items */
-	xfs_extnum_t	count,		/* number of inserted items */
-	xfs_bmbt_irec_t	*new,		/* items to insert */
-	int		whichfork);	/* data or attr fork */
-
-/*
  * Convert a local file to an extents file.
  * This code is sort of bogus, since the file data needs to get
  * logged so it won't be lost.  The bmap-level manipulations are ok, though.
@@ -316,7 +292,7 @@ xfs_bmap_trace_addentry(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_delete_exlist.
+ * Add bmap trace entry prior to a call to xfs_iext_remove.
  */
 STATIC void
 xfs_bmap_trace_delete(
@@ -328,7 +304,7 @@ xfs_bmap_trace_delete(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_insert_exlist, or
+ * Add bmap trace entry prior to a call to xfs_iext_insert, or
  * reading in the extents list from the disk (in the btree).
  */
 STATIC void
@@ -343,7 +319,7 @@ xfs_bmap_trace_insert(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry after updating an extent list entry in place.
+ * Add bmap trace entry after updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_post_update(
@@ -354,7 +330,7 @@ xfs_bmap_trace_post_update(
 	int		whichfork);	/* data or attr fork */
 
 /*
- * Add bmap trace entry prior to updating an extent list entry in place.
+ * Add bmap trace entry prior to updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_pre_update(
@@ -413,19 +389,24 @@ STATIC int
 xfs_bmap_count_tree(
 	xfs_mount_t     *mp,
 	xfs_trans_t     *tp,
+	xfs_ifork_t	*ifp,
 	xfs_fsblock_t   blockno,
 	int             levelin,
 	int		*count);
 
 STATIC int
 xfs_bmap_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
 	int			numrecs,
 	int			*count);
 
 STATIC int
 xfs_bmap_disk_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_mount_t		*mp,
+	xfs_extnum_t		idx,
+	xfs_bmbt_block_t	*block,
 	int			numrecs,
 	int			*count);
 
@@ -537,7 +518,7 @@ xfs_bmap_add_attrfork_local(
 }
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after allocating space (or doing a delayed allocation).
  */
 STATIC int				/* error */
@@ -545,7 +526,7 @@ xfs_bmap_add_extent(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	int			*logflagsp, /* inode logging flags */
@@ -578,7 +559,7 @@ xfs_bmap_add_extent(
 	if (nextents == 0) {
 		xfs_bmap_trace_insert(fname, "insert empty", ip, 0, 1, new,
 			NULL, whichfork);
-		xfs_bmap_insert_exlist(ip, 0, 1, new, whichfork);
+		xfs_iext_insert(ifp, 0, 1, new);
 		ASSERT(cur == NULL);
 		ifp->if_lastex = 0;
 		if (!ISNULLSTARTBLOCK(new->br_startblock)) {
@@ -614,7 +595,7 @@ xfs_bmap_add_extent(
 		/*
 		 * Get the record referred to by idx.
 		 */
-		xfs_bmbt_get_all(&ifp->if_u1.if_extents[idx], &prev);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx), &prev);
 		/*
 		 * If it's a real allocation record, and the new allocation ends
 		 * after the start of the referred to record, then we're filling
@@ -714,14 +695,13 @@ xfs_bmap_add_extent_delay_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	xfs_filblks_t		*dnew,	/* new delayed-alloc indirect blocks */
 	xfs_fsblock_t		*first,	/* pointer to firstblock variable */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	int			*logflagsp, /* inode logging flags */
 	int			rsvd)	/* OK to use reserved data block allocation */
 {
-	xfs_bmbt_rec_t		*base;	/* base of extent entry list */
 	xfs_btree_cur_t		*cur;	/* btree cursor */
 	int			diff;	/* temp value */
 	xfs_bmbt_rec_t		*ep;	/* extent entry for idx */
@@ -730,6 +710,7 @@ xfs_bmap_add_extent_delay_real(
 	static char		fname[] = "xfs_bmap_add_extent_delay_real";
 #endif
 	int			i;	/* temp state */
+	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
 	xfs_bmbt_irec_t		r[3];	/* neighbor extent entries */
 					/* left is 0, right is 1, prev is 2 */
@@ -763,8 +744,8 @@ xfs_bmap_add_extent_delay_real(
 	 * Set up a bunch of variables to make the tests simpler.
 	 */
 	cur = *curp;
-	base = ip->i_df.if_u1.if_extents;
-	ep = &base[idx];
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ep = xfs_iext_get_ext(ifp, idx);
 	xfs_bmbt_get_all(ep, &PREV);
 	new_endoff = new->br_startoff + new->br_blockcount;
 	ASSERT(PREV.br_startoff <= new->br_startoff);
@@ -781,7 +762,7 @@ xfs_bmap_add_extent_delay_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &LEFT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(LEFT.br_startblock));
 	}
 	STATE_SET(LEFT_CONTIG,
@@ -798,7 +779,7 @@ xfs_bmap_add_extent_delay_real(
 	if (STATE_SET_TEST(RIGHT_VALID,
 			idx <
 			ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1)) {
-		xfs_bmbt_get_all(ep + 1, &RIGHT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
 		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(RIGHT.br_startblock));
 	}
 	STATE_SET(RIGHT_CONTIG,
@@ -825,14 +806,14 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount +
 			RIGHT.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
 		xfs_bmap_trace_delete(fname, "LF|RF|LC|RC", ip, idx, 2,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 2, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 2);
 		ip->i_df.if_lastex = idx - 1;
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
@@ -867,14 +848,14 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
 		ip->i_df.if_lastex = idx - 1;
 		xfs_bmap_trace_delete(fname, "LF|RF|LC", ip, idx, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -908,7 +889,7 @@ xfs_bmap_add_extent_delay_real(
 		ip->i_df.if_lastex = idx;
 		xfs_bmap_trace_delete(fname, "LF|RF|RC", ip, idx + 1, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx + 1, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx + 1, 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_DEXT;
 		else {
@@ -964,7 +945,7 @@ xfs_bmap_add_extent_delay_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + new->br_blockcount);
 		xfs_bmbt_set_startoff(ep,
 			PREV.br_startoff + new->br_blockcount);
@@ -1010,7 +991,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_bmbt_set_blockcount(ep, temp);
 		xfs_bmap_trace_insert(fname, "LF", ip, idx, 1, new, NULL,
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ip->i_df.if_lastex = idx;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1039,8 +1020,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
 			STARTBLOCKVAL(PREV.br_startblock) -
 			(cur ? cur->bc_private.b.allocated : 0));
-		base = ip->i_df.if_u1.if_extents;
-		ep = &base[idx + 1];
+		ep = xfs_iext_get_ext(ifp, idx + 1);
 		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
 		xfs_bmap_trace_post_update(fname, "LF", ip, idx + 1,
 			XFS_DATA_FORK);
@@ -1058,7 +1038,8 @@ xfs_bmap_add_extent_delay_real(
 		xfs_bmap_trace_pre_update(fname, "RF|RC", ip, idx + 1,
 			XFS_DATA_FORK);
 		xfs_bmbt_set_blockcount(ep, temp);
-		xfs_bmbt_set_allf(ep + 1, new->br_startoff, new->br_startblock,
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
+			new->br_startoff, new->br_startblock,
 			new->br_blockcount + RIGHT.br_blockcount,
 			RIGHT.br_state);
 		xfs_bmap_trace_post_update(fname, "RF|RC", ip, idx + 1,
@@ -1098,7 +1079,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_bmbt_set_blockcount(ep, temp);
 		xfs_bmap_trace_insert(fname, "RF", ip, idx + 1, 1,
 			new, NULL, XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 1, new);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1127,8 +1108,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(ip, temp),
 			STARTBLOCKVAL(PREV.br_startblock) -
 			(cur ? cur->bc_private.b.allocated : 0));
-		base = ip->i_df.if_u1.if_extents;
-		ep = &base[idx];
+		ep = xfs_iext_get_ext(ifp, idx);
 		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
 		xfs_bmap_trace_post_update(fname, "RF", ip, idx, XFS_DATA_FORK);
 		*dnew = temp;
@@ -1149,7 +1129,7 @@ xfs_bmap_add_extent_delay_real(
 		r[1].br_blockcount = temp2;
 		xfs_bmap_trace_insert(fname, "0", ip, idx + 1, 2, &r[0], &r[1],
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 2, &r[0], XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 2, &r[0]);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1204,13 +1184,13 @@ xfs_bmap_add_extent_delay_real(
 				}
 			}
 		}
-		base = ip->i_df.if_u1.if_extents;
-		ep = &base[idx];
+		ep = xfs_iext_get_ext(ifp, idx);
 		xfs_bmbt_set_startblock(ep, NULLSTARTBLOCK((int)temp));
 		xfs_bmap_trace_post_update(fname, "0", ip, idx, XFS_DATA_FORK);
 		xfs_bmap_trace_pre_update(fname, "0", ip, idx + 2,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_startblock(ep + 2, NULLSTARTBLOCK((int)temp2));
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx + 2),
+			NULLSTARTBLOCK((int)temp2));
 		xfs_bmap_trace_post_update(fname, "0", ip, idx + 2,
 			XFS_DATA_FORK);
 		*dnew = temp + temp2;
@@ -1254,10 +1234,9 @@ xfs_bmap_add_extent_unwritten_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		**curp,	/* if *curp is null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp) /* inode logging flags */
 {
-	xfs_bmbt_rec_t		*base;	/* base of extent entry list */
 	xfs_btree_cur_t		*cur;	/* btree cursor */
 	xfs_bmbt_rec_t		*ep;	/* extent entry for idx */
 	int			error;	/* error return value */
@@ -1265,6 +1244,7 @@ xfs_bmap_add_extent_unwritten_real(
 	static char		fname[] = "xfs_bmap_add_extent_unwritten_real";
 #endif
 	int			i;	/* temp state */
+	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
 	xfs_exntst_t		newext;	/* new extent state */
 	xfs_exntst_t		oldext;	/* old extent state */
@@ -1298,8 +1278,8 @@ xfs_bmap_add_extent_unwritten_real(
 	 */
 	error = 0;
 	cur = *curp;
-	base = ip->i_df.if_u1.if_extents;
-	ep = &base[idx];
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ep = xfs_iext_get_ext(ifp, idx);
 	xfs_bmbt_get_all(ep, &PREV);
 	newext = new->br_state;
 	oldext = (newext == XFS_EXT_UNWRITTEN) ?
@@ -1320,7 +1300,7 @@ xfs_bmap_add_extent_unwritten_real(
 	 * Don't set contiguous if the combined extent would be too large.
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &LEFT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &LEFT);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(LEFT.br_startblock));
 	}
 	STATE_SET(LEFT_CONTIG,
@@ -1337,7 +1317,7 @@ xfs_bmap_add_extent_unwritten_real(
 	if (STATE_SET_TEST(RIGHT_VALID,
 			idx <
 			ip->i_df.if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - 1)) {
-		xfs_bmbt_get_all(ep + 1, &RIGHT);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx + 1), &RIGHT);
 		STATE_SET(RIGHT_DELAY, ISNULLSTARTBLOCK(RIGHT.br_startblock));
 	}
 	STATE_SET(RIGHT_CONTIG,
@@ -1363,14 +1343,14 @@ xfs_bmap_add_extent_unwritten_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount +
 			RIGHT.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
 		xfs_bmap_trace_delete(fname, "LF|RF|LC|RC", ip, idx, 2,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 2, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 2);
 		ip->i_df.if_lastex = idx - 1;
 		ip->i_d.di_nextents -= 2;
 		if (cur == NULL)
@@ -1409,14 +1389,14 @@ xfs_bmap_add_extent_unwritten_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + PREV.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LF|RF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
 		ip->i_df.if_lastex = idx - 1;
 		xfs_bmap_trace_delete(fname, "LF|RF|LC", ip, idx, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 1);
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1456,7 +1436,7 @@ xfs_bmap_add_extent_unwritten_real(
 		ip->i_df.if_lastex = idx;
 		xfs_bmap_trace_delete(fname, "LF|RF|RC", ip, idx + 1, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx + 1, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx + 1, 1);
 		ip->i_d.di_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1516,7 +1496,7 @@ xfs_bmap_add_extent_unwritten_real(
 		 */
 		xfs_bmap_trace_pre_update(fname, "LF|LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			LEFT.br_blockcount + new->br_blockcount);
 		xfs_bmbt_set_startoff(ep,
 			PREV.br_startoff + new->br_blockcount);
@@ -1571,7 +1551,7 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_bmap_trace_post_update(fname, "LF", ip, idx, XFS_DATA_FORK);
 		xfs_bmap_trace_insert(fname, "LF", ip, idx, 1, new, NULL,
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ip->i_df.if_lastex = idx;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1609,7 +1589,8 @@ xfs_bmap_add_extent_unwritten_real(
 			PREV.br_blockcount - new->br_blockcount);
 		xfs_bmap_trace_post_update(fname, "RF|RC", ip, idx,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_allf(ep + 1, new->br_startoff, new->br_startblock,
+		xfs_bmbt_set_allf(xfs_iext_get_ext(ifp, idx + 1),
+			new->br_startoff, new->br_startblock,
 			new->br_blockcount + RIGHT.br_blockcount, newext);
 		xfs_bmap_trace_post_update(fname, "RF|RC", ip, idx + 1,
 			XFS_DATA_FORK);
@@ -1649,7 +1630,7 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_bmap_trace_post_update(fname, "RF", ip, idx, XFS_DATA_FORK);
 		xfs_bmap_trace_insert(fname, "RF", ip, idx + 1, 1,
 			new, NULL, XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 1, new);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents++;
 		if (cur == NULL)
@@ -1696,7 +1677,7 @@ xfs_bmap_add_extent_unwritten_real(
 		r[1].br_state = oldext;
 		xfs_bmap_trace_insert(fname, "0", ip, idx + 1, 2, &r[0], &r[1],
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx + 1, 2, &r[0], XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx + 1, 2, &r[0]);
 		ip->i_df.if_lastex = idx + 1;
 		ip->i_d.di_nextents += 2;
 		if (cur == NULL)
@@ -1770,15 +1751,15 @@ xfs_bmap_add_extent_hole_delay(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp, /* inode logging flags */
 	int			rsvd)		/* OK to allocate reserved blocks */
 {
-	xfs_bmbt_rec_t		*base;	/* base of extent entry list */
-	xfs_bmbt_rec_t		*ep;	/* extent list entry for idx */
+	xfs_bmbt_rec_t		*ep;	/* extent record for idx */
 #ifdef XFS_BMAP_TRACE
 	static char		fname[] = "xfs_bmap_add_extent_hole_delay";
 #endif
+	xfs_ifork_t		*ifp;	/* inode fork pointer */
 	xfs_bmbt_irec_t		left;	/* left neighbor extent entry */
 	xfs_filblks_t		newlen=0;	/* new indirect size */
 	xfs_filblks_t		oldlen=0;	/* old indirect size */
@@ -1799,15 +1780,15 @@ xfs_bmap_add_extent_hole_delay(
 				       ((state &= ~MASK(b)), 0))
 #define	SWITCH_STATE		(state & MASK2(LEFT_CONTIG, RIGHT_CONTIG))
 
-	base = ip->i_df.if_u1.if_extents;
-	ep = &base[idx];
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	ep = xfs_iext_get_ext(ifp, idx);
 	state = 0;
 	ASSERT(ISNULLSTARTBLOCK(new->br_startblock));
 	/*
 	 * Check and set flags if this segment has a left neighbor
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &left);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(left.br_startblock));
 	}
 	/*
@@ -1844,23 +1825,24 @@ xfs_bmap_add_extent_hole_delay(
 		/*
 		 * New allocation is contiguous with delayed allocations
 		 * on the left and on the right.
-		 * Merge all three into a single extent list entry.
+		 * Merge all three into a single extent record.
 		 */
 		temp = left.br_blockcount + new->br_blockcount +
 			right.br_blockcount;
 		xfs_bmap_trace_pre_update(fname, "LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1, temp);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
 		oldlen = STARTBLOCKVAL(left.br_startblock) +
 			STARTBLOCKVAL(new->br_startblock) +
 			STARTBLOCKVAL(right.br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_startblock(ep - 1, NULLSTARTBLOCK((int)newlen));
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
+			NULLSTARTBLOCK((int)newlen));
 		xfs_bmap_trace_post_update(fname, "LC|RC", ip, idx - 1,
 			XFS_DATA_FORK);
 		xfs_bmap_trace_delete(fname, "LC|RC", ip, idx, 1,
 			XFS_DATA_FORK);
-		xfs_bmap_delete_exlist(ip, idx, 1, XFS_DATA_FORK);
+		xfs_iext_remove(ifp, idx, 1);
 		ip->i_df.if_lastex = idx - 1;
 		break;
 
@@ -1873,11 +1855,12 @@ xfs_bmap_add_extent_hole_delay(
 		temp = left.br_blockcount + new->br_blockcount;
 		xfs_bmap_trace_pre_update(fname, "LC", ip, idx - 1,
 			XFS_DATA_FORK);
-		xfs_bmbt_set_blockcount(ep - 1, temp);
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1), temp);
 		oldlen = STARTBLOCKVAL(left.br_startblock) +
 			STARTBLOCKVAL(new->br_startblock);
 		newlen = xfs_bmap_worst_indlen(ip, temp);
-		xfs_bmbt_set_startblock(ep - 1, NULLSTARTBLOCK((int)newlen));
+		xfs_bmbt_set_startblock(xfs_iext_get_ext(ifp, idx - 1),
+			NULLSTARTBLOCK((int)newlen));
 		xfs_bmap_trace_post_update(fname, "LC", ip, idx - 1,
 			XFS_DATA_FORK);
 		ip->i_df.if_lastex = idx - 1;
@@ -1909,7 +1892,7 @@ xfs_bmap_add_extent_hole_delay(
 		oldlen = newlen = 0;
 		xfs_bmap_trace_insert(fname, "0", ip, idx, 1, new, NULL,
 			XFS_DATA_FORK);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, XFS_DATA_FORK);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ip->i_df.if_lastex = idx;
 		break;
 	}
@@ -1940,7 +1923,7 @@ xfs_bmap_add_extent_hole_real(
 	xfs_inode_t		*ip,	/* incore inode pointer */
 	xfs_extnum_t		idx,	/* extent number to update/insert */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*new,	/* new data to put in extent list */
+	xfs_bmbt_irec_t		*new,	/* new data to add to file extents */
 	int			*logflagsp, /* inode logging flags */
 	int			whichfork) /* data or attr fork */
 {
@@ -1970,13 +1953,13 @@ xfs_bmap_add_extent_hole_real(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(idx <= ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
-	ep = &ifp->if_u1.if_extents[idx];
+	ep = xfs_iext_get_ext(ifp, idx);
 	state = 0;
 	/*
 	 * Check and set flags if this segment has a left neighbor.
 	 */
 	if (STATE_SET_TEST(LEFT_VALID, idx > 0)) {
-		xfs_bmbt_get_all(ep - 1, &left);
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, idx - 1), &left);
 		STATE_SET(LEFT_DELAY, ISNULLSTARTBLOCK(left.br_startblock));
 	}
 	/*
@@ -2019,18 +2002,18 @@ xfs_bmap_add_extent_hole_real(
 		/*
 		 * New allocation is contiguous with real allocations on the
 		 * left and on the right.
-		 * Merge all three into a single extent list entry.
+		 * Merge all three into a single extent record.
 		 */
 		xfs_bmap_trace_pre_update(fname, "LC|RC", ip, idx - 1,
 			whichfork);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			left.br_blockcount + new->br_blockcount +
 			right.br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LC|RC", ip, idx - 1,
 			whichfork);
 		xfs_bmap_trace_delete(fname, "LC|RC", ip,
 			idx, 1, whichfork);
-		xfs_bmap_delete_exlist(ip, idx, 1, whichfork);
+		xfs_iext_remove(ifp, idx, 1);
 		ifp->if_lastex = idx - 1;
 		XFS_IFORK_NEXT_SET(ip, whichfork,
 			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
@@ -2062,7 +2045,7 @@ xfs_bmap_add_extent_hole_real(
 		 * Merge the new allocation with the left neighbor.
 		 */
 		xfs_bmap_trace_pre_update(fname, "LC", ip, idx - 1, whichfork);
-		xfs_bmbt_set_blockcount(ep - 1,
+		xfs_bmbt_set_blockcount(xfs_iext_get_ext(ifp, idx - 1),
 			left.br_blockcount + new->br_blockcount);
 		xfs_bmap_trace_post_update(fname, "LC", ip, idx - 1, whichfork);
 		ifp->if_lastex = idx - 1;
@@ -2116,7 +2099,7 @@ xfs_bmap_add_extent_hole_real(
 		 */
 		xfs_bmap_trace_insert(fname, "0", ip, idx, 1, new, NULL,
 			whichfork);
-		xfs_bmap_insert_exlist(ip, idx, 1, new, whichfork);
+		xfs_iext_insert(ifp, idx, 1, new);
 		ifp->if_lastex = idx;
 		XFS_IFORK_NEXT_SET(ip, whichfork,
 			XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
@@ -2311,25 +2294,15 @@ xfs_bmap_extsize_align(
 
 #define XFS_ALLOC_GAP_UNITS	4
 
-/*
- * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
- * It figures out where to ask the underlying allocator to put the new extent.
- */
 STATIC int
-xfs_bmap_alloc(
+xfs_bmap_adjacent(
 	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
 {
 	xfs_fsblock_t	adjust;		/* adjustment to block numbers */
-	xfs_alloctype_t	atype=0;	/* type for allocation routines */
-	int		error;		/* error return value */
 	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
 	xfs_mount_t	*mp;		/* mount point structure */
 	int		nullfb;		/* true if ap->firstblock isn't set */
 	int		rt;		/* true if inode is realtime */
-	xfs_extlen_t	prod = 0;	/* product factor for allocators */
-	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
-	xfs_extlen_t	align;		/* minimum allocation alignment */
-	xfs_rtblock_t	rtx;
 
 #define	ISVALID(x,y)	\
 	(rt ? \
@@ -2338,75 +2311,10 @@ xfs_bmap_alloc(
 		XFS_FSB_TO_AGNO(mp, x) < mp->m_sb.sb_agcount && \
 		XFS_FSB_TO_AGBNO(mp, x) < mp->m_sb.sb_agblocks)
 
-	/*
-	 * Set up variables.
-	 */
 	mp = ap->ip->i_mount;
 	nullfb = ap->firstblock == NULLFSBLOCK;
 	rt = XFS_IS_REALTIME_INODE(ap->ip) && ap->userdata;
 	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
-	if (rt) {
-		align = ap->ip->i_d.di_extsize ?
-			ap->ip->i_d.di_extsize : mp->m_sb.sb_rextsize;
-		/* Set prod to match the extent size */
-		prod = align / mp->m_sb.sb_rextsize;
-
-		error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
-						align, rt, ap->eof, 0,
-						ap->conv, &ap->off, &ap->alen);
-		if (error)
-			return error;
-		ASSERT(ap->alen);
-		ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0);
-
-		/*
-		 * If the offset & length are not perfectly aligned
-		 * then kill prod, it will just get us in trouble.
-		 */
-		if (do_mod(ap->off, align) || ap->alen % align)
-			prod = 1;
-		/*
-		 * Set ralen to be the actual requested length in rtextents.
-		 */
-		ralen = ap->alen / mp->m_sb.sb_rextsize;
-		/*
-		 * If the old value was close enough to MAXEXTLEN that
-		 * we rounded up to it, cut it back so it's valid again.
-		 * Note that if it's a really large request (bigger than
-		 * MAXEXTLEN), we don't hear about that number, and can't
-		 * adjust the starting point to match it.
-		 */
-		if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
-			ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
-		/*
-		 * If it's an allocation to an empty file at offset 0,
-		 * pick an extent that will space things out in the rt area.
-		 */
-		if (ap->eof && ap->off == 0) {
-			error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
-			if (error)
-				return error;
-			ap->rval = rtx * mp->m_sb.sb_rextsize;
-		} else
-			ap->rval = 0;
-	} else {
-		align = (ap->userdata && ap->ip->i_d.di_extsize &&
-			(ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)) ?
-			ap->ip->i_d.di_extsize : 0;
-		if (unlikely(align)) {
-			error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
-							align, rt,
-							ap->eof, 0, ap->conv,
-							&ap->off, &ap->alen);
-			ASSERT(!error);
-			ASSERT(ap->alen);
-		}
-		if (nullfb)
-			ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
-		else
-			ap->rval = ap->firstblock;
-	}
-
 	/*
 	 * If allocating at eof, and there's a previous real block,
 	 * try to use it's last block as our starting point.
@@ -2531,287 +2439,384 @@ xfs_bmap_alloc(
 		else if (gotbno != NULLFSBLOCK)
 			ap->rval = gotbno;
 	}
+#undef ISVALID
+	return 0;
+}
+
+STATIC int
+xfs_bmap_rtalloc(
+	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+{
+	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
+	int		error;		/* error return value */
+	xfs_mount_t	*mp;		/* mount point structure */
+	xfs_extlen_t	prod = 0;	/* product factor for allocators */
+	xfs_extlen_t	ralen = 0;	/* realtime allocation length */
+	xfs_extlen_t	align;		/* minimum allocation alignment */
+	xfs_rtblock_t	rtx;		/* realtime extent number */
+	xfs_rtblock_t	rtb;
+
+	mp = ap->ip->i_mount;
+	align = ap->ip->i_d.di_extsize ?
+		ap->ip->i_d.di_extsize : mp->m_sb.sb_rextsize;
+	prod = align / mp->m_sb.sb_rextsize;
+	error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
+					align, 1, ap->eof, 0,
+					ap->conv, &ap->off, &ap->alen);
+	if (error)
+		return error;
+	ASSERT(ap->alen);
+	ASSERT(ap->alen % mp->m_sb.sb_rextsize == 0);
+
+	/*
+	 * If the offset & length are not perfectly aligned
+	 * then kill prod, it will just get us in trouble.
+	 */
+	if (do_mod(ap->off, align) || ap->alen % align)
+		prod = 1;
+	/*
+	 * Set ralen to be the actual requested length in rtextents.
+	 */
+	ralen = ap->alen / mp->m_sb.sb_rextsize;
+	/*
+	 * If the old value was close enough to MAXEXTLEN that
+	 * we rounded up to it, cut it back so it's valid again.
+	 * Note that if it's a really large request (bigger than
+	 * MAXEXTLEN), we don't hear about that number, and can't
+	 * adjust the starting point to match it.
+	 */
+	if (ralen * mp->m_sb.sb_rextsize >= MAXEXTLEN)
+		ralen = MAXEXTLEN / mp->m_sb.sb_rextsize;
+	/*
+	 * If it's an allocation to an empty file at offset 0,
+	 * pick an extent that will space things out in the rt area.
+	 */
+	if (ap->eof && ap->off == 0) {
+		error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
+		if (error)
+			return error;
+		ap->rval = rtx * mp->m_sb.sb_rextsize;
+	} else {
+		ap->rval = 0;
+	}
+
+	xfs_bmap_adjacent(ap);
+
+	/*
+	 * Realtime allocation, done through xfs_rtallocate_extent.
+	 */
+	atype = ap->rval == 0 ?  XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
+	do_div(ap->rval, mp->m_sb.sb_rextsize);
+	rtb = ap->rval;
+	ap->alen = ralen;
+	if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen,
+				&ralen, atype, ap->wasdel, prod, &rtb)))
+		return error;
+	if (rtb == NULLFSBLOCK && prod > 1 &&
+	    (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1,
+					   ap->alen, &ralen, atype,
+					   ap->wasdel, 1, &rtb)))
+		return error;
+	ap->rval = rtb;
+	if (ap->rval != NULLFSBLOCK) {
+		ap->rval *= mp->m_sb.sb_rextsize;
+		ralen *= mp->m_sb.sb_rextsize;
+		ap->alen = ralen;
+		ap->ip->i_d.di_nblocks += ralen;
+		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
+		if (ap->wasdel)
+			ap->ip->i_delayed_blks -= ralen;
+		/*
+		 * Adjust the disk quota also. This was reserved
+		 * earlier.
+		 */
+		XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
+			ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
+					XFS_TRANS_DQ_RTBCOUNT, (long) ralen);
+	} else {
+		ap->alen = 0;
+	}
+	return 0;
+}
+
+STATIC int
+xfs_bmap_btalloc(
+	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+{
+	xfs_mount_t	*mp;		/* mount point structure */
+	xfs_alloctype_t	atype = 0;	/* type for allocation routines */
+	xfs_extlen_t	align;		/* minimum allocation alignment */
+	xfs_agnumber_t	ag;
+	xfs_agnumber_t	fb_agno;	/* ag number of ap->firstblock */
+	xfs_agnumber_t	startag;
+	xfs_alloc_arg_t	args;
+	xfs_extlen_t	blen;
+	xfs_extlen_t	delta;
+	xfs_extlen_t	longest;
+	xfs_extlen_t	need;
+	xfs_extlen_t	nextminlen = 0;
+	xfs_perag_t	*pag;
+	int		nullfb;		/* true if ap->firstblock isn't set */
+	int		isaligned;
+	int		notinit;
+	int		tryagain;
+	int		error;
+
+	mp = ap->ip->i_mount;
+	align = (ap->userdata && ap->ip->i_d.di_extsize &&
+		(ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE)) ?
+		ap->ip->i_d.di_extsize : 0;
+	if (unlikely(align)) {
+		error = xfs_bmap_extsize_align(mp, ap->gotp, ap->prevp,
+						align, 0, ap->eof, 0, ap->conv,
+						&ap->off, &ap->alen);
+		ASSERT(!error);
+		ASSERT(ap->alen);
+	}
+	nullfb = ap->firstblock == NULLFSBLOCK;
+	fb_agno = nullfb ? NULLAGNUMBER : XFS_FSB_TO_AGNO(mp, ap->firstblock);
+	if (nullfb)
+		ap->rval = XFS_INO_TO_FSB(mp, ap->ip->i_ino);
+	else
+		ap->rval = ap->firstblock;
+
+	xfs_bmap_adjacent(ap);
+
 	/*
 	 * If allowed, use ap->rval; otherwise must use firstblock since
 	 * it's in the right allocation group.
 	 */
-	if (nullfb || rt || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno)
+	if (nullfb || XFS_FSB_TO_AGNO(mp, ap->rval) == fb_agno)
 		;
 	else
 		ap->rval = ap->firstblock;
 	/*
-	 * Realtime allocation, done through xfs_rtallocate_extent.
+	 * Normal allocation, done through xfs_alloc_vextent.
 	 */
-	if (rt) {
-#ifndef __KERNEL__
-		ASSERT(0);
-#else
-		xfs_rtblock_t	rtb;
-
-		atype = ap->rval == 0 ?
-			XFS_ALLOCTYPE_ANY_AG : XFS_ALLOCTYPE_NEAR_BNO;
-		do_div(ap->rval, mp->m_sb.sb_rextsize);
-		rtb = ap->rval;
-		ap->alen = ralen;
-		if ((error = xfs_rtallocate_extent(ap->tp, ap->rval, 1, ap->alen,
-				&ralen, atype, ap->wasdel, prod, &rtb)))
-			return error;
-		if (rtb == NULLFSBLOCK && prod > 1 &&
-		    (error = xfs_rtallocate_extent(ap->tp, ap->rval, 1,
-						   ap->alen, &ralen, atype,
-						   ap->wasdel, 1, &rtb)))
-			return error;
-		ap->rval = rtb;
-		if (ap->rval != NULLFSBLOCK) {
-			ap->rval *= mp->m_sb.sb_rextsize;
-			ralen *= mp->m_sb.sb_rextsize;
-			ap->alen = ralen;
-			ap->ip->i_d.di_nblocks += ralen;
-			xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
-			if (ap->wasdel)
-				ap->ip->i_delayed_blks -= ralen;
+	tryagain = isaligned = 0;
+	args.tp = ap->tp;
+	args.mp = mp;
+	args.fsbno = ap->rval;
+	args.maxlen = MIN(ap->alen, mp->m_sb.sb_agblocks);
+	blen = 0;
+	if (nullfb) {
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		args.total = ap->total;
+		/*
+		 * Find the longest available space.
+		 * We're going to try for the whole allocation at once.
+		 */
+		startag = ag = XFS_FSB_TO_AGNO(mp, args.fsbno);
+		notinit = 0;
+		down_read(&mp->m_peraglock);
+		while (blen < ap->alen) {
+			pag = &mp->m_perag[ag];
+			if (!pag->pagf_init &&
+			    (error = xfs_alloc_pagf_init(mp, args.tp,
+				    ag, XFS_ALLOC_FLAG_TRYLOCK))) {
+				up_read(&mp->m_peraglock);
+				return error;
+			}
 			/*
-			 * Adjust the disk quota also. This was reserved
-			 * earlier.
+			 * See xfs_alloc_fix_freelist...
 			 */
-			XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
-				ap->wasdel ? XFS_TRANS_DQ_DELRTBCOUNT :
-						XFS_TRANS_DQ_RTBCOUNT,
-				(long) ralen);
-		} else
-			ap->alen = 0;
-#endif	/* __KERNEL__ */
+			if (pag->pagf_init) {
+				need = XFS_MIN_FREELIST_PAG(pag, mp);
+				delta = need > pag->pagf_flcount ?
+					need - pag->pagf_flcount : 0;
+				longest = (pag->pagf_longest > delta) ?
+					(pag->pagf_longest - delta) :
+					(pag->pagf_flcount > 0 ||
+					 pag->pagf_longest > 0);
+				if (blen < longest)
+					blen = longest;
+			} else
+				notinit = 1;
+			if (++ag == mp->m_sb.sb_agcount)
+				ag = 0;
+			if (ag == startag)
+				break;
+		}
+		up_read(&mp->m_peraglock);
+		/*
+		 * Since the above loop did a BUF_TRYLOCK, it is
+		 * possible that there is space for this request.
+		 */
+		if (notinit || blen < ap->minlen)
+			args.minlen = ap->minlen;
+		/*
+		 * If the best seen length is less than the request
+		 * length, use the best as the minimum.
+		 */
+		else if (blen < ap->alen)
+			args.minlen = blen;
+		/*
+		 * Otherwise we've seen an extent as big as alen,
+		 * use that as the minimum.
+		 */
+		else
+			args.minlen = ap->alen;
+	} else if (ap->low) {
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.total = args.minlen = ap->minlen;
+	} else {
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+		args.total = ap->total;
+		args.minlen = ap->minlen;
+	}
+	if (unlikely(ap->userdata && ap->ip->i_d.di_extsize &&
+		    (ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE))) {
+		args.prod = ap->ip->i_d.di_extsize;
+		if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod)))
+			args.mod = (xfs_extlen_t)(args.prod - args.mod);
+	} else if (unlikely(mp->m_sb.sb_blocksize >= NBPP)) {
+		args.prod = 1;
+		args.mod = 0;
+	} else {
+		args.prod = NBPP >> mp->m_sb.sb_blocklog;
+		if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod))))
+			args.mod = (xfs_extlen_t)(args.prod - args.mod);
 	}
 	/*
-	 * Normal allocation, done through xfs_alloc_vextent.
+	 * If we are not low on available data blocks, and the
+	 * underlying logical volume manager is a stripe, and
+	 * the file offset is zero then try to allocate data
+	 * blocks on stripe unit boundary.
+	 * NOTE: ap->aeof is only set if the allocation length
+	 * is >= the stripe unit and the allocation offset is
+	 * at the end of file.
 	 */
-	else {
-		xfs_agnumber_t	ag;
-		xfs_alloc_arg_t	args;
-		xfs_extlen_t	blen;
-		xfs_extlen_t	delta;
-		int		isaligned;
-		xfs_extlen_t	longest;
-		xfs_extlen_t	need;
-		xfs_extlen_t	nextminlen=0;
-		int		notinit;
-		xfs_perag_t	*pag;
-		xfs_agnumber_t	startag;
-		int		tryagain;
-
-		tryagain = isaligned = 0;
-		args.tp = ap->tp;
-		args.mp = mp;
-		args.fsbno = ap->rval;
-		args.maxlen = MIN(ap->alen, mp->m_sb.sb_agblocks);
-		blen = 0;
-		if (nullfb) {
-			args.type = XFS_ALLOCTYPE_START_BNO;
-			args.total = ap->total;
-			/*
-			 * Find the longest available space.
-			 * We're going to try for the whole allocation at once.
-			 */
-			startag = ag = XFS_FSB_TO_AGNO(mp, args.fsbno);
-			notinit = 0;
-			down_read(&mp->m_peraglock);
-			while (blen < ap->alen) {
-				pag = &mp->m_perag[ag];
-				if (!pag->pagf_init &&
-				    (error = xfs_alloc_pagf_init(mp, args.tp,
-					    ag, XFS_ALLOC_FLAG_TRYLOCK))) {
-					up_read(&mp->m_peraglock);
-					return error;
-				}
-				/*
-				 * See xfs_alloc_fix_freelist...
-				 */
-				if (pag->pagf_init) {
-					need = XFS_MIN_FREELIST_PAG(pag, mp);
-					delta = need > pag->pagf_flcount ?
-						need - pag->pagf_flcount : 0;
-					longest = (pag->pagf_longest > delta) ?
-						(pag->pagf_longest - delta) :
-						(pag->pagf_flcount > 0 ||
-						 pag->pagf_longest > 0);
-					if (blen < longest)
-						blen = longest;
-				} else
-					notinit = 1;
-				if (++ag == mp->m_sb.sb_agcount)
-					ag = 0;
-				if (ag == startag)
-					break;
-			}
-			up_read(&mp->m_peraglock);
+	if (!ap->low && ap->aeof) {
+		if (!ap->off) {
+			args.alignment = mp->m_dalign;
+			atype = args.type;
+			isaligned = 1;
 			/*
-			 * Since the above loop did a BUF_TRYLOCK, it is
-			 * possible that there is space for this request.
+			 * Adjust for alignment
 			 */
-			if (notinit || blen < ap->minlen)
-				args.minlen = ap->minlen;
+			if (blen > args.alignment && blen <= ap->alen)
+				args.minlen = blen - args.alignment;
+			args.minalignslop = 0;
+		} else {
 			/*
-			 * If the best seen length is less than the request
-			 * length, use the best as the minimum.
+			 * First try an exact bno allocation.
+			 * If it fails then do a near or start bno
+			 * allocation with alignment turned on.
 			 */
-			else if (blen < ap->alen)
-				args.minlen = blen;
+			atype = args.type;
+			tryagain = 1;
+			args.type = XFS_ALLOCTYPE_THIS_BNO;
+			args.alignment = 1;
 			/*
-			 * Otherwise we've seen an extent as big as alen,
-			 * use that as the minimum.
+			 * Compute the minlen+alignment for the
+			 * next case.  Set slop so that the value
+			 * of minlen+alignment+slop doesn't go up
+			 * between the calls.
 			 */
+			if (blen > mp->m_dalign && blen <= ap->alen)
+				nextminlen = blen - mp->m_dalign;
 			else
-				args.minlen = ap->alen;
-		} else if (ap->low) {
-			args.type = XFS_ALLOCTYPE_FIRST_AG;
-			args.total = args.minlen = ap->minlen;
-		} else {
-			args.type = XFS_ALLOCTYPE_NEAR_BNO;
-			args.total = ap->total;
-			args.minlen = ap->minlen;
-		}
-		if (unlikely(ap->userdata && ap->ip->i_d.di_extsize &&
-			    (ap->ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE))) {
-			args.prod = ap->ip->i_d.di_extsize;
-			if ((args.mod = (xfs_extlen_t)do_mod(ap->off, args.prod)))
-				args.mod = (xfs_extlen_t)(args.prod - args.mod);
-		} else if (unlikely(mp->m_sb.sb_blocksize >= NBPP)) {
-			args.prod = 1;
-			args.mod = 0;
-		} else {
-			args.prod = NBPP >> mp->m_sb.sb_blocklog;
-			if ((args.mod = (xfs_extlen_t)(do_mod(ap->off, args.prod))))
-				args.mod = (xfs_extlen_t)(args.prod - args.mod);
+				nextminlen = args.minlen;
+			if (nextminlen + mp->m_dalign > args.minlen + 1)
+				args.minalignslop =
+					nextminlen + mp->m_dalign -
+					args.minlen - 1;
+			else
+				args.minalignslop = 0;
 		}
+	} else {
+		args.alignment = 1;
+		args.minalignslop = 0;
+	}
+	args.minleft = ap->minleft;
+	args.wasdel = ap->wasdel;
+	args.isfl = 0;
+	args.userdata = ap->userdata;
+	if ((error = xfs_alloc_vextent(&args)))
+		return error;
+	if (tryagain && args.fsbno == NULLFSBLOCK) {
 		/*
-		 * If we are not low on available data blocks, and the
-		 * underlying logical volume manager is a stripe, and
-		 * the file offset is zero then try to allocate data
-		 * blocks on stripe unit boundary.
-		 * NOTE: ap->aeof is only set if the allocation length
-		 * is >= the stripe unit and the allocation offset is
-		 * at the end of file.
+		 * Exact allocation failed. Now try with alignment
+		 * turned on.
 		 */
-		if (!ap->low && ap->aeof) {
-			if (!ap->off) {
-				args.alignment = mp->m_dalign;
-				atype = args.type;
-				isaligned = 1;
-				/*
-				 * Adjust for alignment
-				 */
-				if (blen > args.alignment && blen <= ap->alen)
-					args.minlen = blen - args.alignment;
-				args.minalignslop = 0;
-			} else {
-				/*
-				 * First try an exact bno allocation.
-				 * If it fails then do a near or start bno
-				 * allocation with alignment turned on.
-				 */
-				atype = args.type;
-				tryagain = 1;
-				args.type = XFS_ALLOCTYPE_THIS_BNO;
-				args.alignment = 1;
-				/*
-				 * Compute the minlen+alignment for the
-				 * next case.  Set slop so that the value
-				 * of minlen+alignment+slop doesn't go up
-				 * between the calls.
-				 */
-				if (blen > mp->m_dalign && blen <= ap->alen)
-					nextminlen = blen - mp->m_dalign;
-				else
-					nextminlen = args.minlen;
-				if (nextminlen + mp->m_dalign > args.minlen + 1)
-					args.minalignslop =
-						nextminlen + mp->m_dalign -
-						args.minlen - 1;
-				else
-					args.minalignslop = 0;
-			}
-		} else {
-			args.alignment = 1;
-			args.minalignslop = 0;
-		}
-		args.minleft = ap->minleft;
-		args.wasdel = ap->wasdel;
-		args.isfl = 0;
-		args.userdata = ap->userdata;
+		args.type = atype;
+		args.fsbno = ap->rval;
+		args.alignment = mp->m_dalign;
+		args.minlen = nextminlen;
+		args.minalignslop = 0;
+		isaligned = 1;
 		if ((error = xfs_alloc_vextent(&args)))
 			return error;
-		if (tryagain && args.fsbno == NULLFSBLOCK) {
-			/*
-			 * Exact allocation failed. Now try with alignment
-			 * turned on.
-			 */
-			args.type = atype;
-			args.fsbno = ap->rval;
-			args.alignment = mp->m_dalign;
-			args.minlen = nextminlen;
-			args.minalignslop = 0;
-			isaligned = 1;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-		}
-		if (isaligned && args.fsbno == NULLFSBLOCK) {
-			/*
-			 * allocation failed, so turn off alignment and
-			 * try again.
-			 */
-			args.type = atype;
-			args.fsbno = ap->rval;
-			args.alignment = 0;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-		}
-		if (args.fsbno == NULLFSBLOCK && nullfb &&
-		    args.minlen > ap->minlen) {
-			args.minlen = ap->minlen;
-			args.type = XFS_ALLOCTYPE_START_BNO;
-			args.fsbno = ap->rval;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-		}
-		if (args.fsbno == NULLFSBLOCK && nullfb) {
-			args.fsbno = 0;
-			args.type = XFS_ALLOCTYPE_FIRST_AG;
-			args.total = ap->minlen;
-			args.minleft = 0;
-			if ((error = xfs_alloc_vextent(&args)))
-				return error;
-			ap->low = 1;
-		}
-		if (args.fsbno != NULLFSBLOCK) {
-			ap->firstblock = ap->rval = args.fsbno;
-			ASSERT(nullfb || fb_agno == args.agno ||
-			       (ap->low && fb_agno < args.agno));
-			ap->alen = args.len;
-			ap->ip->i_d.di_nblocks += args.len;
-			xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
-			if (ap->wasdel)
-				ap->ip->i_delayed_blks -= args.len;
-			/*
-			 * Adjust the disk quota also. This was reserved
-			 * earlier.
-			 */
-			XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
-				ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
-						XFS_TRANS_DQ_BCOUNT,
-				(long) args.len);
-		} else {
-			ap->rval = NULLFSBLOCK;
-			ap->alen = 0;
-		}
+	}
+	if (isaligned && args.fsbno == NULLFSBLOCK) {
+		/*
+		 * allocation failed, so turn off alignment and
+		 * try again.
+		 */
+		args.type = atype;
+		args.fsbno = ap->rval;
+		args.alignment = 0;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+	}
+	if (args.fsbno == NULLFSBLOCK && nullfb &&
+	    args.minlen > ap->minlen) {
+		args.minlen = ap->minlen;
+		args.type = XFS_ALLOCTYPE_START_BNO;
+		args.fsbno = ap->rval;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+	}
+	if (args.fsbno == NULLFSBLOCK && nullfb) {
+		args.fsbno = 0;
+		args.type = XFS_ALLOCTYPE_FIRST_AG;
+		args.total = ap->minlen;
+		args.minleft = 0;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+		ap->low = 1;
+	}
+	if (args.fsbno != NULLFSBLOCK) {
+		ap->firstblock = ap->rval = args.fsbno;
+		ASSERT(nullfb || fb_agno == args.agno ||
+		       (ap->low && fb_agno < args.agno));
+		ap->alen = args.len;
+		ap->ip->i_d.di_nblocks += args.len;
+		xfs_trans_log_inode(ap->tp, ap->ip, XFS_ILOG_CORE);
+		if (ap->wasdel)
+			ap->ip->i_delayed_blks -= args.len;
+		/*
+		 * Adjust the disk quota also. This was reserved
+		 * earlier.
+		 */
+		XFS_TRANS_MOD_DQUOT_BYINO(mp, ap->tp, ap->ip,
+			ap->wasdel ? XFS_TRANS_DQ_DELBCOUNT :
+					XFS_TRANS_DQ_BCOUNT,
+			(long) args.len);
+	} else {
+		ap->rval = NULLFSBLOCK;
+		ap->alen = 0;
 	}
 	return 0;
-#undef	ISVALID
+}
+
+/*
+ * xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
+ * It figures out where to ask the underlying allocator to put the new extent.
+ */
+STATIC int
+xfs_bmap_alloc(
+	xfs_bmalloca_t	*ap)		/* bmap alloc argument struct */
+{
+	if ((ap->ip->i_d.di_flags & XFS_DIFLAG_REALTIME) && ap->userdata)
+		return xfs_bmap_rtalloc(ap);
+	return xfs_bmap_btalloc(ap);
 }
 
 /*
  * Transform a btree format file with only one leaf node, where the
  * extents list will fit in the inode, into an extents format file.
- * Since the extent list is already in-core, all we have to do is
+ * Since the file extents are already in-core, all we have to do is
  * give up the space for the btree root and pitch the leaf block.
  */
 STATIC int				/* error */
@@ -2868,7 +2873,7 @@ xfs_bmap_btree_to_extents(
 }
 
 /*
- * Called by xfs_bmapi to update extent list structure and the btree
+ * Called by xfs_bmapi to update file extent records and the btree
  * after removing space (or undoing a delayed allocation).
  */
 STATIC int				/* error */
@@ -2878,7 +2883,7 @@ xfs_bmap_del_extent(
 	xfs_extnum_t		idx,	/* extent number to update/delete */
 	xfs_bmap_free_t		*flist,	/* list of extents to be freed */
 	xfs_btree_cur_t		*cur,	/* if null, not a btree */
-	xfs_bmbt_irec_t		*del,	/* data to remove from extent list */
+	xfs_bmbt_irec_t		*del,	/* data to remove from extents */
 	int			*logflagsp, /* inode logging flags */
 	int			whichfork, /* data or attr fork */
 	int			rsvd)	/* OK to allocate reserved blocks */
@@ -2903,7 +2908,6 @@ xfs_bmap_del_extent(
 	xfs_filblks_t		nblks;	/* quota/sb block count */
 	xfs_bmbt_irec_t		new;	/* new record to be inserted */
 	/* REFERENCED */
-	xfs_extnum_t		nextents;	/* number of extents in list */
 	uint			qfield;	/* quota field to update */
 	xfs_filblks_t		temp;	/* for indirect length calculations */
 	xfs_filblks_t		temp2;	/* for indirect length calculations */
@@ -2911,10 +2915,10 @@ xfs_bmap_del_extent(
 	XFS_STATS_INC(xs_del_exlist);
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	ASSERT(idx >= 0 && idx < nextents);
+	ASSERT((idx >= 0) && (idx < ifp->if_bytes /
+		(uint)sizeof(xfs_bmbt_rec_t)));
 	ASSERT(del->br_blockcount > 0);
-	ep = &ifp->if_u1.if_extents[idx];
+	ep = xfs_iext_get_ext(ifp, idx);
 	xfs_bmbt_get_all(ep, &got);
 	ASSERT(got.br_startoff <= del->br_startoff);
 	del_endoff = del->br_startoff + del->br_blockcount;
@@ -2990,7 +2994,7 @@ xfs_bmap_del_extent(
 		 * Matches the whole extent.  Delete the entry.
 		 */
 		xfs_bmap_trace_delete(fname, "3", ip, idx, 1, whichfork);
-		xfs_bmap_delete_exlist(ip, idx, 1, whichfork);
+		xfs_iext_remove(ifp, idx, 1);
 		ifp->if_lastex = idx;
 		if (delay)
 			break;
@@ -3160,7 +3164,7 @@ xfs_bmap_del_extent(
 		xfs_bmap_trace_post_update(fname, "0", ip, idx, whichfork);
 		xfs_bmap_trace_insert(fname, "0", ip, idx + 1, 1, &new, NULL,
 			whichfork);
-		xfs_bmap_insert_exlist(ip, idx + 1, 1, &new, whichfork);
+		xfs_iext_insert(ifp, idx + 1, 1, &new);
 		ifp->if_lastex = idx + 1;
 		break;
 	}
@@ -3213,31 +3217,6 @@ xfs_bmap_del_free(
 }
 
 /*
- * Remove count entries from the extents array for inode "ip", starting
- * at index "idx".  Copies the remaining items down over the deleted ones,
- * and gives back the excess memory.
- */
-STATIC void
-xfs_bmap_delete_exlist(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* starting delete index */
-	xfs_extnum_t	count,		/* count of items to delete */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_t	*base;		/* base of extent list */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_extnum_t	nextents;	/* number of extents in list after */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	base = ifp->if_u1.if_extents;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t) - count;
-	memmove(&base[idx], &base[idx + count],
-		(nextents - idx) * sizeof(*base));
-	xfs_iext_realloc(ip, -count, whichfork);
-}
-
-/*
  * Convert an extents-format file into a btree-format file.
  * The new file will have a root block (in the inode) and a single child block.
  */
@@ -3258,13 +3237,13 @@ xfs_bmap_extents_to_btree(
 	xfs_bmbt_rec_t		*arp;		/* child record pointer */
 	xfs_bmbt_block_t	*block;		/* btree root block */
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
-	xfs_bmbt_rec_t		*ep;		/* extent list pointer */
+	xfs_bmbt_rec_t		*ep;		/* extent record pointer */
 	int			error;		/* error return value */
-	xfs_extnum_t		i, cnt;		/* extent list index */
+	xfs_extnum_t		i, cnt;		/* extent record index */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
 	xfs_bmbt_key_t		*kp;		/* root block key pointer */
 	xfs_mount_t		*mp;		/* mount structure */
-	xfs_extnum_t		nextents;	/* extent list size */
+	xfs_extnum_t		nextents;	/* number of file extents */
 	xfs_bmbt_ptr_t		*pp;		/* root block address pointer */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
@@ -3343,7 +3322,8 @@ xfs_bmap_extents_to_btree(
 	ablock->bb_rightsib = cpu_to_be64(NULLDFSBNO);
 	arp = XFS_BMAP_REC_IADDR(ablock, 1, cur);
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (ep = ifp->if_u1.if_extents, cnt = i = 0; i < nextents; i++, ep++) {
+	for (cnt = i = 0; i < nextents; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
 		if (!ISNULLSTARTBLOCK(xfs_bmbt_get_startblock(ep))) {
 			arp->l0 = INT_GET(ep->l0, ARCH_CONVERT);
 			arp->l1 = INT_GET(ep->l1, ARCH_CONVERT);
@@ -3373,34 +3353,6 @@ xfs_bmap_extents_to_btree(
 }
 
 /*
- * Insert new item(s) in the extent list for inode "ip".
- * Count new items are inserted at offset idx.
- */
-STATIC void
-xfs_bmap_insert_exlist(
-	xfs_inode_t	*ip,		/* incore inode pointer */
-	xfs_extnum_t	idx,		/* starting index of new items */
-	xfs_extnum_t	count,		/* number of inserted items */
-	xfs_bmbt_irec_t	*new,		/* items to insert */
-	int		whichfork)	/* data or attr fork */
-{
-	xfs_bmbt_rec_t	*base;		/* extent list base */
-	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_extnum_t	nextents;	/* extent list size */
-	xfs_extnum_t	to;		/* extent list index */
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	xfs_iext_realloc(ip, count, whichfork);
-	base = ifp->if_u1.if_extents;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	memmove(&base[idx + count], &base[idx],
-		(nextents - (idx + count)) * sizeof(*base));
-	for (to = idx; to < idx + count; to++, new++)
-		xfs_bmbt_set_all(&base[to], new);
-}
-
-/*
  * Helper routine to reset inode di_forkoff field when switching
  * attribute fork from local to extent format - we reset it where
  * possible to make space available for inline data fork extents.
@@ -3457,12 +3409,13 @@ xfs_bmap_local_to_extents(
 	error = 0;
 	if (ifp->if_bytes) {
 		xfs_alloc_arg_t	args;	/* allocation arguments */
-		xfs_buf_t	*bp;	/* buffer for extent list block */
-		xfs_bmbt_rec_t	*ep;	/* extent list pointer */
+		xfs_buf_t	*bp;	/* buffer for extent block */
+		xfs_bmbt_rec_t	*ep;	/* extent record pointer */
 
 		args.tp = tp;
 		args.mp = ip->i_mount;
-		ASSERT(ifp->if_flags & XFS_IFINLINE);
+		ASSERT((ifp->if_flags &
+			(XFS_IFINLINE|XFS_IFEXTENTS|XFS_IFEXTIREC)) == XFS_IFINLINE);
 		/*
 		 * Allocate a block.  We know we need only one, since the
 		 * file currently fits in an inode.
@@ -3492,8 +3445,8 @@ xfs_bmap_local_to_extents(
 		xfs_trans_log_buf(tp, bp, 0, ifp->if_bytes - 1);
 		xfs_bmap_forkoff_reset(args.mp, ip, whichfork);
 		xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
-		xfs_iext_realloc(ip, 1, whichfork);
-		ep = ifp->if_u1.if_extents;
+		xfs_iext_add(ifp, 0, 1);
+		ep = xfs_iext_get_ext(ifp, 0);
 		xfs_bmbt_set_allf(ep, 0, args.fsbno, 1, XFS_EXT_NORM);
 		xfs_bmap_trace_post_update(fname, "new", ip, 0, whichfork);
 		XFS_IFORK_NEXT_SET(ip, whichfork, 1);
@@ -3514,110 +3467,54 @@ done:
 	return error;
 }
 
+/*
+ * Search the extent records for the entry containing block bno.
+ * If bno lies in a hole, point to the next entry.  If bno lies
+ * past eof, *eofp will be set, and *prevp will contain the last
+ * entry (null if none).  Else, *lastxp will be set to the index
+ * of the found entry; *gotp will contain the entry.
+ */
 xfs_bmbt_rec_t *			/* pointer to found extent entry */
-xfs_bmap_do_search_extents(
-	xfs_bmbt_rec_t	*base,		/* base of extent list */
-	xfs_extnum_t	lastx,		/* last extent index used */
-	xfs_extnum_t	nextents,	/* extent list size */
+xfs_bmap_search_multi_extents(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fileoff_t	bno,		/* block number searched for */
 	int		*eofp,		/* out: end of file found */
 	xfs_extnum_t	*lastxp,	/* out: last extent index */
 	xfs_bmbt_irec_t	*gotp,		/* out: extent entry found */
 	xfs_bmbt_irec_t	*prevp)		/* out: previous extent entry found */
 {
-	xfs_bmbt_rec_t	*ep;		/* extent list entry pointer */
-	xfs_bmbt_irec_t	got;		/* extent list entry, decoded */
-	int		high;		/* high index of binary search */
-	int		low;		/* low index of binary search */
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_extnum_t	lastx;		/* last extent index */
 
 	/*
 	 * Initialize the extent entry structure to catch access to
 	 * uninitialized br_startblock field.
 	 */
-	got.br_startoff = 0xffa5a5a5a5a5a5a5LL;
-	got.br_blockcount = 0xa55a5a5a5a5a5a5aLL;
-	got.br_state = XFS_EXT_INVALID;
-
+	gotp->br_startoff = 0xffa5a5a5a5a5a5a5LL;
+	gotp->br_blockcount = 0xa55a5a5a5a5a5a5aLL;
+	gotp->br_state = XFS_EXT_INVALID;
 #if XFS_BIG_BLKNOS
-	got.br_startblock = 0xffffa5a5a5a5a5a5LL;
+	gotp->br_startblock = 0xffffa5a5a5a5a5a5LL;
 #else
-	got.br_startblock = 0xffffa5a5;
+	gotp->br_startblock = 0xffffa5a5;
 #endif
-
-	if (lastx != NULLEXTNUM && lastx < nextents)
-		ep = base + lastx;
-	else
-		ep = NULL;
 	prevp->br_startoff = NULLFILEOFF;
-	if (ep && bno >= (got.br_startoff = xfs_bmbt_get_startoff(ep)) &&
-	    bno < got.br_startoff +
-		  (got.br_blockcount = xfs_bmbt_get_blockcount(ep)))
-		*eofp = 0;
-	else if (ep && lastx < nextents - 1 &&
-		 bno >= (got.br_startoff = xfs_bmbt_get_startoff(ep + 1)) &&
-		 bno < got.br_startoff +
-		       (got.br_blockcount = xfs_bmbt_get_blockcount(ep + 1))) {
-		lastx++;
-		ep++;
-		*eofp = 0;
-	} else if (nextents == 0)
-		*eofp = 1;
-	else if (bno == 0 &&
-		 (got.br_startoff = xfs_bmbt_get_startoff(base)) == 0) {
-		ep = base;
-		lastx = 0;
-		got.br_blockcount = xfs_bmbt_get_blockcount(ep);
+
+	ep = xfs_iext_bno_to_ext(ifp, bno, &lastx);
+	if (lastx > 0) {
+		xfs_bmbt_get_all(xfs_iext_get_ext(ifp, lastx - 1), prevp);
+	}
+	if (lastx < (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t))) {
+		xfs_bmbt_get_all(ep, gotp);
 		*eofp = 0;
 	} else {
-		/* binary search the extents array */
-		low = 0;
-		high = nextents - 1;
-		while (low <= high) {
-			XFS_STATS_INC(xs_cmp_exlist);
-			lastx = (low + high) >> 1;
-			ep = base + lastx;
-			got.br_startoff = xfs_bmbt_get_startoff(ep);
-			got.br_blockcount = xfs_bmbt_get_blockcount(ep);
-			if (bno < got.br_startoff)
-				high = lastx - 1;
-			else if (bno >= got.br_startoff + got.br_blockcount)
-				low = lastx + 1;
-			else {
-				got.br_startblock = xfs_bmbt_get_startblock(ep);
-				got.br_state = xfs_bmbt_get_state(ep);
-				*eofp = 0;
-				*lastxp = lastx;
-				*gotp = got;
-				return ep;
-			}
-		}
-		if (bno >= got.br_startoff + got.br_blockcount) {
-			lastx++;
-			if (lastx == nextents) {
-				*eofp = 1;
-				got.br_startblock = xfs_bmbt_get_startblock(ep);
-				got.br_state = xfs_bmbt_get_state(ep);
-				*prevp = got;
-				ep = NULL;
-			} else {
-				*eofp = 0;
-				xfs_bmbt_get_all(ep, prevp);
-				ep++;
-				got.br_startoff = xfs_bmbt_get_startoff(ep);
-				got.br_blockcount = xfs_bmbt_get_blockcount(ep);
-			}
-		} else {
-			*eofp = 0;
-			if (ep > base)
-				xfs_bmbt_get_all(ep - 1, prevp);
+		if (lastx > 0) {
+			*gotp = *prevp;
 		}
-	}
-	if (ep) {
-		got.br_startblock = xfs_bmbt_get_startblock(ep);
-		got.br_state = xfs_bmbt_get_state(ep);
+		*eofp = 1;
+		ep = NULL;
 	}
 	*lastxp = lastx;
-	*gotp = got;
 	return ep;
 }
 
@@ -3639,20 +3536,14 @@ xfs_bmap_search_extents(
 	xfs_bmbt_irec_t *prevp)         /* out: previous extent entry found */
 {
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t  *base;          /* base of extent list */
-	xfs_extnum_t    lastx;          /* last extent index used */
-	xfs_extnum_t    nextents;       /* extent list size */
-	xfs_bmbt_rec_t  *ep;            /* extent list entry pointer */
+	xfs_bmbt_rec_t  *ep;            /* extent record pointer */
 	int		rt;		/* realtime flag    */
 
 	XFS_STATS_INC(xs_look_exlist);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	lastx = ifp->if_lastex;
-	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	base = &ifp->if_u1.if_extents[0];
 
-	ep = xfs_bmap_do_search_extents(base, lastx, nextents, bno, eofp,
-					  lastxp, gotp, prevp);
+	ep = xfs_bmap_search_multi_extents(ifp, bno, eofp, lastxp, gotp, prevp);
+
 	rt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip);
 	if (unlikely(!rt && !gotp->br_startblock && (*lastxp != NULLEXTNUM))) {
                 cmn_err(CE_PANIC,"Access to block zero: fs: <%s> inode: %lld "
@@ -3732,7 +3623,7 @@ xfs_bmap_trace_addentry(
 }
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_delete_exlist.
+ * Add bmap trace entry prior to a call to xfs_iext_remove.
  */
 STATIC void
 xfs_bmap_trace_delete(
@@ -3747,13 +3638,13 @@ xfs_bmap_trace_delete(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_DELETE, fname, desc, ip, idx,
-		cnt, &ifp->if_u1.if_extents[idx],
-		cnt == 2 ? &ifp->if_u1.if_extents[idx + 1] : NULL,
+		cnt, xfs_iext_get_ext(ifp, idx),
+		cnt == 2 ? xfs_iext_get_ext(ifp, idx + 1) : NULL,
 		whichfork);
 }
 
 /*
- * Add bmap trace entry prior to a call to xfs_bmap_insert_exlist, or
+ * Add bmap trace entry prior to a call to xfs_iext_insert, or
  * reading in the extents list from the disk (in the btree).
  */
 STATIC void
@@ -3783,7 +3674,7 @@ xfs_bmap_trace_insert(
 }
 
 /*
- * Add bmap trace entry after updating an extent list entry in place.
+ * Add bmap trace entry after updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_post_update(
@@ -3797,11 +3688,11 @@ xfs_bmap_trace_post_update(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_POST_UP, fname, desc, ip, idx,
-		1, &ifp->if_u1.if_extents[idx], NULL, whichfork);
+		1, xfs_iext_get_ext(ifp, idx), NULL, whichfork);
 }
 
 /*
- * Add bmap trace entry prior to updating an extent list entry in place.
+ * Add bmap trace entry prior to updating an extent record in place.
  */
 STATIC void
 xfs_bmap_trace_pre_update(
@@ -3815,7 +3706,7 @@ xfs_bmap_trace_pre_update(
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_bmap_trace_addentry(XFS_BMAP_KTRACE_PRE_UP, fname, desc, ip, idx, 1,
-		&ifp->if_u1.if_extents[idx], NULL, whichfork);
+		xfs_iext_get_ext(ifp, idx), NULL, whichfork);
 }
 #endif	/* XFS_BMAP_TRACE */
 
@@ -3892,7 +3783,7 @@ xfs_bmap_add_attrfork(
 	int			rsvd)		/* xact may use reserved blks */
 {
 	xfs_fsblock_t		firstblock;	/* 1st block/ag allocated */
-	xfs_bmap_free_t		flist;		/* freed extent list */
+	xfs_bmap_free_t		flist;		/* freed extent records */
 	xfs_mount_t		*mp;		/* mount structure */
 	xfs_trans_t		*tp;		/* transaction pointer */
 	unsigned long		s;		/* spinlock spl value */
@@ -4146,7 +4037,7 @@ xfs_bmap_finish(
 	xfs_efd_log_item_t	*efd;		/* extent free data */
 	xfs_efi_log_item_t	*efi;		/* extent free intention */
 	int			error;		/* error return value */
-	xfs_bmap_free_item_t	*free;		/* free extent list item */
+	xfs_bmap_free_item_t	*free;		/* free extent item */
 	unsigned int		logres;		/* new log reservation */
 	unsigned int		logcount;	/* new log count */
 	xfs_mount_t		*mp;		/* filesystem mount structure */
@@ -4242,9 +4133,9 @@ xfs_bmap_first_unused(
 	xfs_fileoff_t	*first_unused,		/* unused block */
 	int		whichfork)		/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*base;			/* base of extent array */
 	xfs_bmbt_rec_t	*ep;			/* pointer to an extent entry */
 	int		error;			/* error return value */
+	int		idx;			/* extent record index */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
 	xfs_fileoff_t	lastaddr;		/* last block number seen */
 	xfs_fileoff_t	lowest;			/* lowest useful block */
@@ -4265,10 +4156,8 @@ xfs_bmap_first_unused(
 		return error;
 	lowest = *first_unused;
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	base = &ifp->if_u1.if_extents[0];
-	for (lastaddr = 0, max = lowest, ep = base;
-	     ep < &base[nextents];
-	     ep++) {
+	for (idx = 0, lastaddr = 0, max = lowest; idx < nextents; idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		off = xfs_bmbt_get_startoff(ep);
 		/*
 		 * See if the hole before this extent will work.
@@ -4287,8 +4176,8 @@ xfs_bmap_first_unused(
 /*
  * Returns the file-relative block number of the last block + 1 before
  * last_block (input value) in the file.
- * This is not based on i_size, it is based on the extent list.
- * Returns 0 for local files, as they do not have an extent list.
+ * This is not based on i_size, it is based on the extent records.
+ * Returns 0 for local files, as they do not have extent records.
  */
 int						/* error */
 xfs_bmap_last_before(
@@ -4335,8 +4224,8 @@ xfs_bmap_last_before(
 
 /*
  * Returns the file-relative block number of the first block past eof in
- * the file.  This is not based on i_size, it is based on the extent list.
- * Returns 0 for local files, as they do not have an extent list.
+ * the file.  This is not based on i_size, it is based on the extent records.
+ * Returns 0 for local files, as they do not have extent records.
  */
 int						/* error */
 xfs_bmap_last_offset(
@@ -4345,7 +4234,6 @@ xfs_bmap_last_offset(
 	xfs_fileoff_t	*last_block,		/* last block */
 	int		whichfork)		/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*base;			/* base of extent array */
 	xfs_bmbt_rec_t	*ep;			/* pointer to last extent */
 	int		error;			/* error return value */
 	xfs_ifork_t	*ifp;			/* inode fork pointer */
@@ -4368,9 +4256,7 @@ xfs_bmap_last_offset(
 		*last_block = 0;
 		return 0;
 	}
-	base = &ifp->if_u1.if_extents[0];
-	ASSERT(base != NULL);
-	ep = &base[nextents - 1];
+	ep = xfs_iext_get_ext(ifp, nextents - 1);
 	*last_block = xfs_bmbt_get_startoff(ep) + xfs_bmbt_get_blockcount(ep);
 	return 0;
 }
@@ -4400,7 +4286,7 @@ xfs_bmap_one_block(
 		return 0;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	ep = ifp->if_u1.if_extents;
+	ep = xfs_iext_get_ext(ifp, 0);
 	xfs_bmbt_get_all(ep, &s);
 	rval = s.br_startoff == 0 && s.br_blockcount == 1;
 	if (rval && whichfork == XFS_DATA_FORK)
@@ -4435,7 +4321,6 @@ xfs_bmap_read_extents(
 	xfs_bmbt_ptr_t		*pp;	/* pointer to block address */
 	/* REFERENCED */
 	xfs_extnum_t		room;	/* number of entries there's room for */
-	xfs_bmbt_rec_t		*trp;	/* target record pointer */
 
 	bno = NULLFSBLOCK;
 	mp = ip->i_mount;
@@ -4478,16 +4363,16 @@ xfs_bmap_read_extents(
 	/*
 	 * Here with bp and block set to the leftmost leaf node in the tree.
 	 */
-	room = ifp->if_bytes / (uint)sizeof(*trp);
-	trp = ifp->if_u1.if_extents;
+	room = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 	i = 0;
 	/*
-	 * Loop over all leaf nodes.  Copy information to the extent list.
+	 * Loop over all leaf nodes.  Copy information to the extent records.
 	 */
 	for (;;) {
-		xfs_bmbt_rec_t	*frp, *temp;
+		xfs_bmbt_rec_t	*frp, *trp;
 		xfs_fsblock_t	nextbno;
 		xfs_extnum_t	num_recs;
+		xfs_extnum_t	start;
 
 
 		num_recs = be16_to_cpu(block->bb_numrecs);
@@ -4511,12 +4396,13 @@ xfs_bmap_read_extents(
 		if (nextbno != NULLFSBLOCK)
 			xfs_btree_reada_bufl(mp, nextbno, 1);
 		/*
-		 * Copy records into the extent list.
+		 * Copy records into the extent records.
 		 */
 		frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
 			block, 1, mp->m_bmap_dmxr[0]);
-		temp = trp;
-		for (j = 0; j < num_recs; j++, frp++, trp++) {
+		start = i;
+		for (j = 0; j < num_recs; j++, i++, frp++) {
+			trp = xfs_iext_get_ext(ifp, i);
 			trp->l0 = INT_GET(frp->l0, ARCH_CONVERT);
 			trp->l1 = INT_GET(frp->l1, ARCH_CONVERT);
 		}
@@ -4526,14 +4412,14 @@ xfs_bmap_read_extents(
 			 * any "older" data bmap btree records for a
 			 * set bit in the "extent flag" position.
 			 */
-			if (unlikely(xfs_check_nostate_extents(temp, num_recs))) {
+			if (unlikely(xfs_check_nostate_extents(ifp,
+					start, num_recs))) {
 				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
 						 XFS_ERRLEVEL_LOW,
 						 ip->i_mount);
 				goto error0;
 			}
 		}
-		i += num_recs;
 		xfs_trans_brelse(tp, bp);
 		bno = nextbno;
 		/*
@@ -4546,7 +4432,7 @@ xfs_bmap_read_extents(
 			return error;
 		block = XFS_BUF_TO_BMBT_BLOCK(bp);
 	}
-	ASSERT(i == ifp->if_bytes / (uint)sizeof(*trp));
+	ASSERT(i == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
 	ASSERT(i == XFS_IFORK_NEXTENTS(ip, whichfork));
 	xfs_bmap_trace_exlist(fname, ip, i, whichfork);
 	return 0;
@@ -4557,7 +4443,7 @@ error0:
 
 #ifdef XFS_BMAP_TRACE
 /*
- * Add bmap trace insert entries for all the contents of the extent list.
+ * Add bmap trace insert entries for all the contents of the extent records.
  */
 void
 xfs_bmap_trace_exlist(
@@ -4566,16 +4452,15 @@ xfs_bmap_trace_exlist(
 	xfs_extnum_t	cnt,		/* count of entries in the list */
 	int		whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t	*base;		/* base of extent list */
-	xfs_bmbt_rec_t	*ep;		/* current entry in extent list */
-	xfs_extnum_t	idx;		/* extent list entry number */
+	xfs_bmbt_rec_t	*ep;		/* current extent record */
+	xfs_extnum_t	idx;		/* extent record index */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_irec_t	s;		/* extent list record */
+	xfs_bmbt_irec_t	s;		/* file extent record */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(cnt == ifp->if_bytes / (uint)sizeof(*base));
-	base = ifp->if_u1.if_extents;
-	for (idx = 0, ep = base; idx < cnt; idx++, ep++) {
+	ASSERT(cnt == (ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)));
+	for (idx = 0; idx < cnt; idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		xfs_bmbt_get_all(ep, &s);
 		xfs_bmap_trace_insert(fname, "exlist", ip, idx, 1, &s, NULL,
 			whichfork);
@@ -4661,14 +4546,10 @@ xfs_bmapi(
 	xfs_bmalloca_t	bma;		/* args for xfs_bmap_alloc */
 	xfs_btree_cur_t	*cur;		/* bmap btree cursor */
 	xfs_fileoff_t	end;		/* end of mapped file region */
-	int		eof;		/* we've hit the end of extent list */
-	char		contig;		/* allocation must be one extent */
-	char		delay;		/* this request is for delayed alloc */
-	char		exact;		/* don't do all of wasdelayed extent */
-	char		convert;	/* unwritten extent I/O completion */
-	xfs_bmbt_rec_t	*ep;		/* extent list entry pointer */
+	int		eof;		/* we've hit the end of extents */
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
 	int		error;		/* error return */
-	xfs_bmbt_irec_t	got;		/* current extent list record */
+	xfs_bmbt_irec_t	got;		/* current file extent record */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
 	xfs_extlen_t	indlen;		/* indirect blocks length */
 	xfs_extnum_t	lastx;		/* last useful extent number */
@@ -4680,17 +4561,13 @@ xfs_bmapi(
 	int		nallocs;	/* number of extents alloc\'d */
 	xfs_extnum_t	nextents;	/* number of extents in file */
 	xfs_fileoff_t	obno;		/* old block number (offset) */
-	xfs_bmbt_irec_t	prev;		/* previous extent list record */
+	xfs_bmbt_irec_t	prev;		/* previous file extent record */
 	int		tmp_logflags;	/* temp flags holder */
 	int		whichfork;	/* data or attr fork */
 	char		inhole;		/* current location is hole in file */
-	char		stateless;	/* ignore state flag set */
-	char		trim;		/* output trimmed to match range */
-	char		userdata;	/* allocating non-metadata */
 	char		wasdelay;	/* old extent was delayed */
 	char		wr;		/* this is a write request */
 	char		rt;		/* this is a realtime file */
-	char		rsvd;		/* OK to allocate reserved blocks */
 #ifdef DEBUG
 	xfs_fileoff_t	orig_bno;	/* original block number value */
 	int		orig_flags;	/* original flags arg value */
@@ -4727,15 +4604,8 @@ xfs_bmapi(
 		XFS_STATS_INC(xs_blk_mapw);
 	else
 		XFS_STATS_INC(xs_blk_mapr);
-	delay = (flags & XFS_BMAPI_DELAY) != 0;
-	trim = (flags & XFS_BMAPI_ENTIRE) == 0;
-	userdata = (flags & XFS_BMAPI_METADATA) == 0;
-	convert = (flags & XFS_BMAPI_CONVERT) != 0;
-	exact = (flags & XFS_BMAPI_EXACT) != 0;
-	rsvd = (flags & XFS_BMAPI_RSVBLOCKS) != 0;
-	contig = (flags & XFS_BMAPI_CONTIG) != 0;
 	/*
-	 * stateless is used to combine extents which
+	 * IGSTATE flag is used to combine extents which
 	 * differ only due to the state of the extents.
 	 * This technique is used from xfs_getbmap()
 	 * when the caller does not wish to see the
@@ -4751,10 +4621,9 @@ xfs_bmapi(
 	 * xfs_strat_comp(), where the xfs_bmapi() call
 	 * is transactioned, and the extents combined.
 	 */
-	stateless = (flags & XFS_BMAPI_IGSTATE) != 0;
-	if (stateless && wr)	/* if writing unwritten space, no */
-		wr = 0;		/* allocations are allowed */
-	ASSERT(wr || !delay);
+	if ((flags & XFS_BMAPI_IGSTATE) && wr)	/* if writing unwritten space */
+		wr = 0;				/* no allocations are allowed */
+	ASSERT(wr || !(flags & XFS_BMAPI_DELAY));
 	logflags = 0;
 	nallocs = 0;
 	cur = NULL;
@@ -4789,7 +4658,7 @@ xfs_bmapi(
 		if (eof && !wr)
 			got.br_startoff = end;
 		inhole = eof || got.br_startoff > bno;
-		wasdelay = wr && !inhole && !delay &&
+		wasdelay = wr && !inhole && !(flags & XFS_BMAPI_DELAY) &&
 			ISNULLSTARTBLOCK(got.br_startblock);
 		/*
 		 * First, deal with the hole before the allocated space
@@ -4801,11 +4670,11 @@ xfs_bmapi(
 			 * allocate the stuff asked for in this bmap call
 			 * but that wouldn't be as good.
 			 */
-			if (wasdelay && !exact) {
+			if (wasdelay && !(flags & XFS_BMAPI_EXACT)) {
 				alen = (xfs_extlen_t)got.br_blockcount;
 				aoff = got.br_startoff;
 				if (lastx != NULLEXTNUM && lastx) {
-					ep = &ifp->if_u1.if_extents[lastx - 1];
+					ep = xfs_iext_get_ext(ifp, lastx - 1);
 					xfs_bmbt_get_all(ep, &prev);
 				}
 			} else if (wasdelay) {
@@ -4823,8 +4692,8 @@ xfs_bmapi(
 							got.br_startoff - bno);
 				aoff = bno;
 			}
-			minlen = contig ? alen : 1;
-			if (delay) {
+			minlen = (flags & XFS_BMAPI_CONTIG) ? alen : 1;
+			if (flags & XFS_BMAPI_DELAY) {
 				xfs_extlen_t	extsz;
 
 				/* Figure out the extent size, adjust alen */
@@ -4837,7 +4706,9 @@ xfs_bmapi(
 				if (extsz) {
 					error = xfs_bmap_extsize_align(mp,
 							&got, &prev, extsz,
-							rt, eof, delay, convert,
+							rt, eof,
+							flags&XFS_BMAPI_DELAY,
+							flags&XFS_BMAPI_CONVERT,
 							&aoff, &alen);
 					ASSERT(!error);
 				}
@@ -4875,24 +4746,29 @@ xfs_bmapi(
 				if (rt) {
 					error = xfs_mod_incore_sb(mp,
 							XFS_SBS_FREXTENTS,
-							-(extsz), rsvd);
+							-(extsz), (flags &
+							XFS_BMAPI_RSVBLOCKS));
 				} else {
 					error = xfs_mod_incore_sb(mp,
 							XFS_SBS_FDBLOCKS,
-							-(alen), rsvd);
+							-(alen), (flags &
+							XFS_BMAPI_RSVBLOCKS));
 				}
 				if (!error) {
 					error = xfs_mod_incore_sb(mp,
 							XFS_SBS_FDBLOCKS,
-							-(indlen), rsvd);
+							-(indlen), (flags &
+							XFS_BMAPI_RSVBLOCKS));
 					if (error && rt)
 						xfs_mod_incore_sb(mp,
 							XFS_SBS_FREXTENTS,
-							extsz, rsvd);
+							extsz, (flags &
+							XFS_BMAPI_RSVBLOCKS));
 					else if (error)
 						xfs_mod_incore_sb(mp,
 							XFS_SBS_FDBLOCKS,
-							alen, rsvd);
+							alen, (flags &
+							XFS_BMAPI_RSVBLOCKS));
 				}
 
 				if (error) {
@@ -4925,7 +4801,7 @@ xfs_bmapi(
 				/* Indicate if this is the first user data
 				 * in the file, or just any user data.
 				 */
-				if (userdata) {
+				if (!(flags & XFS_BMAPI_METADATA)) {
 					bma.userdata = (aoff == 0) ?
 						XFS_ALLOC_INITIAL_USER_DATA :
 						XFS_ALLOC_USERDATA;
@@ -4937,7 +4813,7 @@ xfs_bmapi(
 				bma.firstblock = *firstblock;
 				bma.alen = alen;
 				bma.off = aoff;
-				bma.conv = convert;
+				bma.conv = (flags & XFS_BMAPI_CONVERT);
 				bma.wasdel = wasdelay;
 				bma.minlen = minlen;
 				bma.low = flist->xbf_low;
@@ -4948,7 +4824,8 @@ xfs_bmapi(
 				 * is larger than a stripe unit.
 				 */
 				if (mp->m_dalign && alen >= mp->m_dalign &&
-				    userdata && whichfork == XFS_DATA_FORK) {
+				    (!(flags & XFS_BMAPI_METADATA)) &&
+				    (whichfork == XFS_DATA_FORK)) {
 					if ((error = xfs_bmap_isaeof(ip, aoff,
 							whichfork, &bma.aeof)))
 						goto error0;
@@ -5011,19 +4888,19 @@ xfs_bmapi(
 			}
 			error = xfs_bmap_add_extent(ip, lastx, &cur, &got,
 				firstblock, flist, &tmp_logflags, whichfork,
-				rsvd);
+				(flags & XFS_BMAPI_RSVBLOCKS));
 			logflags |= tmp_logflags;
 			if (error)
 				goto error0;
 			lastx = ifp->if_lastex;
-			ep = &ifp->if_u1.if_extents[lastx];
+			ep = xfs_iext_get_ext(ifp, lastx);
 			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 			xfs_bmbt_get_all(ep, &got);
 			ASSERT(got.br_startoff <= aoff);
 			ASSERT(got.br_startoff + got.br_blockcount >=
 				aoff + alen);
 #ifdef DEBUG
-			if (delay) {
+			if (flags & XFS_BMAPI_DELAY) {
 				ASSERT(ISNULLSTARTBLOCK(got.br_startblock));
 				ASSERT(STARTBLOCKVAL(got.br_startblock) > 0);
 			}
@@ -5052,14 +4929,15 @@ xfs_bmapi(
 		 * Then deal with the allocated space we found.
 		 */
 		ASSERT(ep != NULL);
-		if (trim && (got.br_startoff + got.br_blockcount > obno)) {
+		if (!(flags & XFS_BMAPI_ENTIRE) &&
+		    (got.br_startoff + got.br_blockcount > obno)) {
 			if (obno > bno)
 				bno = obno;
 			ASSERT((bno >= obno) || (n == 0));
 			ASSERT(bno < end);
 			mval->br_startoff = bno;
 			if (ISNULLSTARTBLOCK(got.br_startblock)) {
-				ASSERT(!wr || delay);
+				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
 				mval->br_startblock = DELAYSTARTBLOCK;
 			} else
 				mval->br_startblock =
@@ -5081,7 +4959,7 @@ xfs_bmapi(
 		} else {
 			*mval = got;
 			if (ISNULLSTARTBLOCK(mval->br_startblock)) {
-				ASSERT(!wr || delay);
+				ASSERT(!wr || (flags & XFS_BMAPI_DELAY));
 				mval->br_startblock = DELAYSTARTBLOCK;
 			}
 		}
@@ -5107,12 +4985,12 @@ xfs_bmapi(
 			mval->br_state = XFS_EXT_NORM;
 			error = xfs_bmap_add_extent(ip, lastx, &cur, mval,
 				firstblock, flist, &tmp_logflags, whichfork,
-				rsvd);
+				(flags & XFS_BMAPI_RSVBLOCKS));
 			logflags |= tmp_logflags;
 			if (error)
 				goto error0;
 			lastx = ifp->if_lastex;
-			ep = &ifp->if_u1.if_extents[lastx];
+			ep = xfs_iext_get_ext(ifp, lastx);
 			nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
 			xfs_bmbt_get_all(ep, &got);
 			/*
@@ -5124,9 +5002,10 @@ xfs_bmapi(
 				continue;
 		}
 
-		ASSERT(!trim ||
+		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
 		       ((mval->br_startoff + mval->br_blockcount) <= end));
-		ASSERT(!trim || (mval->br_blockcount <= len) ||
+		ASSERT((flags & XFS_BMAPI_ENTIRE) ||
+		       (mval->br_blockcount <= len) ||
 		       (mval->br_startoff < obno));
 		bno = mval->br_startoff + mval->br_blockcount;
 		len = end - bno;
@@ -5141,7 +5020,8 @@ xfs_bmapi(
 			   mval[-1].br_startblock != HOLESTARTBLOCK &&
 			   mval->br_startblock ==
 			   mval[-1].br_startblock + mval[-1].br_blockcount &&
-			   (stateless || mval[-1].br_state == mval->br_state)) {
+			   ((flags & XFS_BMAPI_IGSTATE) ||
+				mval[-1].br_state == mval->br_state)) {
 			ASSERT(mval->br_startoff ==
 			       mval[-1].br_startoff + mval[-1].br_blockcount);
 			mval[-1].br_blockcount += mval->br_blockcount;
@@ -5168,8 +5048,7 @@ xfs_bmapi(
 		/*
 		 * Else go on to the next record.
 		 */
-		ep++;
-		lastx++;
+		ep = xfs_iext_get_ext(ifp, ++lastx);
 		if (lastx >= nextents) {
 			eof = 1;
 			prev = got;
@@ -5199,7 +5078,7 @@ xfs_bmapi(
 error0:
 	/*
 	 * Log everything.  Do this after conversion, there's no point in
-	 * logging the extent list if we've converted to btree format.
+	 * logging the extent records if we've converted to btree format.
 	 */
 	if ((logflags & XFS_ILOG_FEXT(whichfork)) &&
 	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
@@ -5252,12 +5131,12 @@ xfs_bmapi_single(
 	xfs_fsblock_t	*fsb,		/* output: mapped block */
 	xfs_fileoff_t	bno)		/* starting file offs. mapped */
 {
-	int		eof;		/* we've hit the end of extent list */
+	int		eof;		/* we've hit the end of extents */
 	int		error;		/* error return */
-	xfs_bmbt_irec_t	got;		/* current extent list record */
+	xfs_bmbt_irec_t	got;		/* current file extent record */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
 	xfs_extnum_t	lastx;		/* last useful extent number */
-	xfs_bmbt_irec_t	prev;		/* previous extent list record */
+	xfs_bmbt_irec_t	prev;		/* previous file extent record */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if (unlikely(
@@ -5312,18 +5191,18 @@ xfs_bunmapi(
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
 	xfs_bmbt_irec_t		del;		/* extent being deleted */
 	int			eof;		/* is deleting at eof */
-	xfs_bmbt_rec_t		*ep;		/* extent list entry pointer */
+	xfs_bmbt_rec_t		*ep;		/* extent record pointer */
 	int			error;		/* error return value */
 	xfs_extnum_t		extno;		/* extent number in list */
-	xfs_bmbt_irec_t		got;		/* current extent list entry */
+	xfs_bmbt_irec_t		got;		/* current extent record */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
 	int			isrt;		/* freeing in rt area */
 	xfs_extnum_t		lastx;		/* last extent index used */
 	int			logflags;	/* transaction logging flags */
 	xfs_extlen_t		mod;		/* rt extent offset */
 	xfs_mount_t		*mp;		/* mount structure */
-	xfs_extnum_t		nextents;	/* size of extent list */
-	xfs_bmbt_irec_t		prev;		/* previous extent list entry */
+	xfs_extnum_t		nextents;	/* number of file extents */
+	xfs_bmbt_irec_t		prev;		/* previous extent record */
 	xfs_fileoff_t		start;		/* first file offset deleted */
 	int			tmp_logflags;	/* partial logging flags */
 	int			wasdel;		/* was a delayed alloc extent */
@@ -5369,7 +5248,7 @@ xfs_bunmapi(
 	 * file, back up to the last block if so...
 	 */
 	if (eof) {
-		ep = &ifp->if_u1.if_extents[--lastx];
+		ep = xfs_iext_get_ext(ifp, --lastx);
 		xfs_bmbt_get_all(ep, &got);
 		bno = got.br_startoff + got.br_blockcount - 1;
 	}
@@ -5393,7 +5272,7 @@ xfs_bunmapi(
 		if (got.br_startoff > bno) {
 			if (--lastx < 0)
 				break;
-			ep--;
+			ep = xfs_iext_get_ext(ifp, lastx);
 			xfs_bmbt_get_all(ep, &got);
 		}
 		/*
@@ -5440,7 +5319,8 @@ xfs_bunmapi(
 					del.br_blockcount : mod;
 				if (bno < got.br_startoff) {
 					if (--lastx >= 0)
-						xfs_bmbt_get_all(--ep, &got);
+						xfs_bmbt_get_all(xfs_iext_get_ext(
+							ifp, lastx), &got);
 				}
 				continue;
 			}
@@ -5500,7 +5380,8 @@ xfs_bunmapi(
 				 * try again.
 				 */
 				ASSERT(lastx > 0);
-				xfs_bmbt_get_all(ep - 1, &prev);
+				xfs_bmbt_get_all(xfs_iext_get_ext(ifp,
+						lastx - 1), &prev);
 				ASSERT(prev.br_state == XFS_EXT_NORM);
 				ASSERT(!ISNULLSTARTBLOCK(prev.br_startblock));
 				ASSERT(del.br_startblock ==
@@ -5587,12 +5468,12 @@ nodelete:
 		 * If not done go on to the next (previous) record.
 		 * Reset ep in case the extents array was re-alloced.
 		 */
-		ep = &ifp->if_u1.if_extents[lastx];
+		ep = xfs_iext_get_ext(ifp, lastx);
 		if (bno != (xfs_fileoff_t)-1 && bno >= start) {
 			if (lastx >= XFS_IFORK_NEXTENTS(ip, whichfork) ||
 			    xfs_bmbt_get_startoff(ep) > bno) {
-				lastx--;
-				ep--;
+				if (--lastx >= 0)
+					ep = xfs_iext_get_ext(ifp, lastx);
 			}
 			if (lastx >= 0)
 				xfs_bmbt_get_all(ep, &got);
@@ -5636,7 +5517,7 @@ nodelete:
 error0:
 	/*
 	 * Log everything.  Do this after conversion, there's no point in
-	 * logging the extent list if we've converted to btree format.
+	 * logging the extent records if we've converted to btree format.
 	 */
 	if ((logflags & XFS_ILOG_FEXT(whichfork)) &&
 	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
@@ -5892,9 +5773,9 @@ xfs_bmap_isaeof(
 {
 	int		error;		/* error return value */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t	*lastrec;	/* extent list entry pointer */
-	xfs_extnum_t	nextents;	/* size of extent list */
-	xfs_bmbt_irec_t	s;		/* expanded extent list entry */
+	xfs_bmbt_rec_t	*lastrec;	/* extent record pointer */
+	xfs_extnum_t	nextents;	/* number of file extents */
+	xfs_bmbt_irec_t	s;		/* expanded extent record */
 
 	ASSERT(whichfork == XFS_DATA_FORK);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
@@ -5909,7 +5790,7 @@ xfs_bmap_isaeof(
 	/*
 	 * Go to the last extent
 	 */
-	lastrec = &ifp->if_u1.if_extents[nextents - 1];
+	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
 	xfs_bmbt_get_all(lastrec, &s);
 	/*
 	 * Check we are allocating in the last extent (for delayed allocations)
@@ -5936,8 +5817,8 @@ xfs_bmap_eof(
 	xfs_fsblock_t	blockcount;	/* extent block count */
 	int		error;		/* error return value */
 	xfs_ifork_t	*ifp;		/* inode fork pointer */
-	xfs_bmbt_rec_t	*lastrec;	/* extent list entry pointer */
-	xfs_extnum_t	nextents;	/* size of extent list */
+	xfs_bmbt_rec_t	*lastrec;	/* extent record pointer */
+	xfs_extnum_t	nextents;	/* number of file extents */
 	xfs_fileoff_t	startoff;	/* extent starting file offset */
 
 	ASSERT(whichfork == XFS_DATA_FORK);
@@ -5953,7 +5834,7 @@ xfs_bmap_eof(
 	/*
 	 * Go to the last extent
 	 */
-	lastrec = &ifp->if_u1.if_extents[nextents - 1];
+	lastrec = xfs_iext_get_ext(ifp, nextents - 1);
 	startoff = xfs_bmbt_get_startoff(lastrec);
 	blockcount = xfs_bmbt_get_blockcount(lastrec);
 	*eof = endoff >= startoff + blockcount;
@@ -5969,18 +5850,21 @@ xfs_bmap_check_extents(
 	xfs_inode_t		*ip,		/* incore inode pointer */
 	int			whichfork)	/* data or attr fork */
 {
-	xfs_bmbt_rec_t		*base;		/* base of extents list */
 	xfs_bmbt_rec_t		*ep;		/* current extent entry */
+	xfs_extnum_t		idx;		/* extent record index */
 	xfs_ifork_t		*ifp;		/* inode fork pointer */
 	xfs_extnum_t		nextents;	/* number of extents in list */
+	xfs_bmbt_rec_t		*nextp;		/* next extent entry */
 
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	base = ifp->if_u1.if_extents;
 	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
-	for (ep = base; ep < &base[nextents - 1]; ep++) {
+	ep = xfs_iext_get_ext(ifp, 0);
+	for (idx = 0; idx < nextents - 1; idx++) {
+		nextp = xfs_iext_get_ext(ifp, idx + 1);
 		xfs_btree_check_rec(XFS_BTNUM_BMAP, (void *)ep,
-			(void *)(ep + 1));
+			(void *)(nextp));
+		ep = nextp;
 	}
 }
 
@@ -6119,12 +6003,14 @@ xfs_bmap_check_leaf_extents(
 	xfs_fsblock_t		bno;	/* block # of "block" */
 	xfs_buf_t		*bp;	/* buffer for "block" */
 	int			error;	/* error return value */
-	xfs_extnum_t		i=0;	/* index into the extents list */
+	xfs_extnum_t		i=0, j;	/* index into the extents list */
 	xfs_ifork_t		*ifp;	/* fork structure */
 	int			level;	/* btree level, for checking */
 	xfs_mount_t		*mp;	/* file system mount structure */
 	xfs_bmbt_ptr_t		*pp;	/* pointer to block address */
-	xfs_bmbt_rec_t		*ep, *lastp;	/* extent pointers in block entry */
+	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
+	xfs_bmbt_rec_t		*lastp; /* pointer to previous extent */
+	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
 	int			bp_release = 0;
 
 	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) {
@@ -6194,7 +6080,6 @@ xfs_bmap_check_leaf_extents(
 	 */
 	lastp = NULL;
 	for (;;) {
-		xfs_bmbt_rec_t	*frp;
 		xfs_fsblock_t	nextbno;
 		xfs_extnum_t	num_recs;
 
@@ -6213,18 +6098,20 @@ xfs_bmap_check_leaf_extents(
 		 * conform with the first entry in this one.
 		 */
 
-		frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
+		ep = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
 			block, 1, mp->m_bmap_dmxr[0]);
-
-		for (ep = frp;ep < frp + (num_recs - 1); ep++) {
+		for (j = 1; j < num_recs; j++) {
+			nextp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize, xfs_bmbt,
+				block, j + 1, mp->m_bmap_dmxr[0]);
 			if (lastp) {
 				xfs_btree_check_rec(XFS_BTNUM_BMAP,
 					(void *)lastp, (void *)ep);
 			}
 			xfs_btree_check_rec(XFS_BTNUM_BMAP, (void *)ep,
-				(void *)(ep + 1));
+				(void *)(nextp));
+			lastp = ep;
+			ep = nextp;
 		}
-		lastp = frp + num_recs - 1; /* For the next iteration */
 
 		i += num_recs;
 		if (bp_release) {
@@ -6288,7 +6175,7 @@ xfs_bmap_count_blocks(
 	mp = ip->i_mount;
 	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if ( XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS ) {
-		if (unlikely(xfs_bmap_count_leaves(ifp->if_u1.if_extents,
+		if (unlikely(xfs_bmap_count_leaves(ifp, 0,
 			ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t),
 			count) < 0)) {
 			XFS_ERROR_REPORT("xfs_bmap_count_blocks(1)",
@@ -6310,7 +6197,7 @@ xfs_bmap_count_blocks(
 	ASSERT(XFS_FSB_TO_AGBNO(mp, INT_GET(*pp, ARCH_CONVERT)) < mp->m_sb.sb_agblocks);
 	bno = INT_GET(*pp, ARCH_CONVERT);
 
-	if (unlikely(xfs_bmap_count_tree(mp, tp, bno, level, count) < 0)) {
+	if (unlikely(xfs_bmap_count_tree(mp, tp, ifp, bno, level, count) < 0)) {
 		XFS_ERROR_REPORT("xfs_bmap_count_blocks(2)", XFS_ERRLEVEL_LOW,
 				 mp);
 		return XFS_ERROR(EFSCORRUPTED);
@@ -6327,6 +6214,7 @@ int                                     /* error */
 xfs_bmap_count_tree(
 	xfs_mount_t     *mp,            /* file system mount point */
 	xfs_trans_t     *tp,            /* transaction pointer */
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
 	xfs_fsblock_t   blockno,	/* file system block number */
 	int             levelin,	/* level in btree */
 	int		*count)		/* Count of blocks */
@@ -6339,7 +6227,6 @@ xfs_bmap_count_tree(
 	xfs_fsblock_t		nextbno;
 	xfs_bmbt_block_t        *block, *nextblock;
 	int			numrecs;
-	xfs_bmbt_rec_t		*frp;
 
 	if ((error = xfs_btree_read_bufl(mp, tp, bno, 0, &bp, XFS_BMAP_BTREE_REF)))
 		return error;
@@ -6364,7 +6251,7 @@ xfs_bmap_count_tree(
 			xfs_bmbt, block, 1, mp->m_bmap_dmxr[1]);
 		bno = INT_GET(*pp, ARCH_CONVERT);
 		if (unlikely((error =
-		     xfs_bmap_count_tree(mp, tp, bno, level, count)) < 0)) {
+		     xfs_bmap_count_tree(mp, tp, ifp, bno, level, count)) < 0)) {
 			xfs_trans_brelse(tp, bp);
 			XFS_ERROR_REPORT("xfs_bmap_count_tree(1)",
 					 XFS_ERRLEVEL_LOW, mp);
@@ -6376,9 +6263,8 @@ xfs_bmap_count_tree(
 		for (;;) {
 			nextbno = be64_to_cpu(block->bb_rightsib);
 			numrecs = be16_to_cpu(block->bb_numrecs);
-			frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize,
-				xfs_bmbt, block, 1, mp->m_bmap_dmxr[0]);
-			if (unlikely(xfs_bmap_disk_count_leaves(frp, numrecs, count) < 0)) {
+			if (unlikely(xfs_bmap_disk_count_leaves(ifp, mp,
+					0, block, numrecs, count) < 0)) {
 				xfs_trans_brelse(tp, bp);
 				XFS_ERROR_REPORT("xfs_bmap_count_tree(2)",
 						 XFS_ERRLEVEL_LOW, mp);
@@ -6399,33 +6285,45 @@ xfs_bmap_count_tree(
 }
 
 /*
- * Count leaf blocks given a pointer to an extent list.
+ * Count leaf blocks given a range of extent records.
  */
 int
 xfs_bmap_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
 	int			numrecs,
 	int			*count)
 {
 	int		b;
+	xfs_bmbt_rec_t	*frp;
 
-	for ( b = 1; b <= numrecs; b++, frp++)
+	for (b = 0; b < numrecs; b++) {
+		frp = xfs_iext_get_ext(ifp, idx + b);
 		*count += xfs_bmbt_get_blockcount(frp);
+	}
 	return 0;
 }
 
 /*
- * Count leaf blocks given a pointer to an extent list originally in btree format.
+ * Count leaf blocks given a range of extent records originally
+ * in btree format.
  */
 int
 xfs_bmap_disk_count_leaves(
-	xfs_bmbt_rec_t		*frp,
+	xfs_ifork_t		*ifp,
+	xfs_mount_t		*mp,
+	xfs_extnum_t		idx,
+	xfs_bmbt_block_t	*block,
 	int			numrecs,
 	int			*count)
 {
 	int		b;
+	xfs_bmbt_rec_t	*frp;
 
-	for ( b = 1; b <= numrecs; b++, frp++)
+	for (b = 1; b <= numrecs; b++) {
+		frp = XFS_BTREE_REC_ADDR(mp->m_sb.sb_blocksize,
+			xfs_bmbt, block, idx + b, mp->m_bmap_dmxr[0]);
 		*count += xfs_bmbt_disk_get_blockcount(frp);
+	}
 	return 0;
 }
diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h
index 12cc63dfc2c4..f83399c89ce3 100644
--- a/fs/xfs/xfs_bmap.h
+++ b/fs/xfs/xfs_bmap.h
@@ -20,6 +20,7 @@
 
 struct getbmap;
 struct xfs_bmbt_irec;
+struct xfs_ifork;
 struct xfs_inode;
 struct xfs_mount;
 struct xfs_trans;
@@ -347,9 +348,20 @@ xfs_bmap_count_blocks(
  */
 int
 xfs_check_nostate_extents(
-	xfs_bmbt_rec_t		*ep,
+	struct xfs_ifork	*ifp,
+	xfs_extnum_t		idx,
 	xfs_extnum_t		num);
 
+/*
+ * Call xfs_bmap_do_search_extents() to search for the extent
+ * record containing block bno. If in multi-level in-core extent
+ * allocation mode, find and extract the target extent buffer,
+ * otherwise just use the direct extent list.
+ */
+xfs_bmbt_rec_t *
+xfs_bmap_search_multi_extents(struct xfs_ifork *, xfs_fileoff_t, int *,
+			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
+
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_H__ */
diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c
index 3f1383d160e8..bea44709afbe 100644
--- a/fs/xfs/xfs_bmap_btree.c
+++ b/fs/xfs/xfs_bmap_btree.c
@@ -2754,7 +2754,7 @@ xfs_bmbt_update(
 }
 
 /*
- * Check an extent list, which has just been read, for
+ * Check extent records, which have just been read, for
  * any bit in the extent flag field. ASSERT on debug
  * kernels, as this condition should not occur.
  * Return an error condition (1) if any flags found,
@@ -2763,10 +2763,14 @@ xfs_bmbt_update(
 
 int
 xfs_check_nostate_extents(
-	xfs_bmbt_rec_t		*ep,
+	xfs_ifork_t		*ifp,
+	xfs_extnum_t		idx,
 	xfs_extnum_t		num)
 {
-	for (; num > 0; num--, ep++) {
+	xfs_bmbt_rec_t		*ep;
+
+	for (; num > 0; num--, idx++) {
+		ep = xfs_iext_get_ext(ifp, idx);
 		if ((ep->l0 >>
 		     (64 - BMBT_EXNTFLAG_BITLEN)) != 0) {
 			ASSERT(0);
diff --git a/fs/xfs/xfs_bmap_btree.h b/fs/xfs/xfs_bmap_btree.h
index e095a2d344ae..6478cfa0e539 100644
--- a/fs/xfs/xfs_bmap_btree.h
+++ b/fs/xfs/xfs_bmap_btree.h
@@ -372,14 +372,6 @@ extern int xfs_bmbt_get_rec(struct xfs_btree_cur *, xfs_fileoff_t *,
 				xfs_exntst_t *, int *);
 #endif
 
-/*
- * Search an extent list for the extent which includes block
- * bno.
- */
-xfs_bmbt_rec_t *xfs_bmap_do_search_extents(xfs_bmbt_rec_t *,
-			xfs_extnum_t, xfs_extnum_t, xfs_fileoff_t, int *,
-			xfs_extnum_t *, xfs_bmbt_irec_t *, xfs_bmbt_irec_t *);
-
 #endif	/* __KERNEL__ */
 
 #endif	/* __XFS_BMAP_BTREE_H__ */
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 07e2324152b1..5fed15682dda 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -98,12 +98,12 @@ xfs_buf_item_flush_log_debug(
 }
 
 /*
- * This function is called to verify that our caller's have logged
+ * This function is called to verify that our callers have logged
  * all the bytes that they changed.
  *
  * It does this by comparing the original copy of the buffer stored in
  * the buf log item's bli_orig array to the current copy of the buffer
- * and ensuring that all bytes which miscompare are set in the bli_logged
+ * and ensuring that all bytes which mismatch are set in the bli_logged
  * array of the buf log item.
  */
 STATIC void
diff --git a/fs/xfs/xfs_cap.h b/fs/xfs/xfs_cap.h
index 433ec537f9bd..d0035c6e9514 100644
--- a/fs/xfs/xfs_cap.h
+++ b/fs/xfs/xfs_cap.h
@@ -38,7 +38,7 @@ typedef struct xfs_cap_set {
 /*
  * For Linux, we take the bitfields directly from capability.h
  * and no longer attempt to keep this attribute ondisk compatible
- * with IRIX.  Since this attribute is only set on exectuables,
+ * with IRIX.  Since this attribute is only set on executables,
  * it just doesn't make much sense to try.  We do use a different
  * named attribute though, to avoid confusion.
  */
diff --git a/fs/xfs/xfs_clnt.h b/fs/xfs/xfs_clnt.h
index f57cc9ac875e..022fff62085b 100644
--- a/fs/xfs/xfs_clnt.h
+++ b/fs/xfs/xfs_clnt.h
@@ -68,8 +68,6 @@ struct xfs_mount_args {
 						 * enforcement */
 #define XFSMNT_PQUOTAENF	0x00000040	/* IRIX project quota limit
 						 * enforcement */
-#define XFSMNT_NOATIME		0x00000100	/* don't modify access
-						 * times on reads */
 #define XFSMNT_NOALIGN		0x00000200	/* don't allocate at
 						 * stripe boundaries*/
 #define XFSMNT_RETERR		0x00000400	/* return error to user */
diff --git a/fs/xfs/xfs_da_btree.c b/fs/xfs/xfs_da_btree.c
index 473671fa5c13..8988b9051175 100644
--- a/fs/xfs/xfs_da_btree.c
+++ b/fs/xfs/xfs_da_btree.c
@@ -126,10 +126,10 @@ xfs_da_node_create(xfs_da_args_t *args, xfs_dablk_t blkno, int level,
 	node = bp->data;
 	node->hdr.info.forw = 0;
 	node->hdr.info.back = 0;
-	INT_SET(node->hdr.info.magic, ARCH_CONVERT, XFS_DA_NODE_MAGIC);
+	node->hdr.info.magic = cpu_to_be16(XFS_DA_NODE_MAGIC);
 	node->hdr.info.pad = 0;
 	node->hdr.count = 0;
-	INT_SET(node->hdr.level, ARCH_CONVERT, level);
+	node->hdr.level = cpu_to_be16(level);
 
 	xfs_da_log_buf(tp, bp,
 		XFS_DA_LOGRANGE(node, &node->hdr, sizeof(node->hdr)));
@@ -290,28 +290,28 @@ xfs_da_split(xfs_da_state_t *state)
 
 	node = oldblk->bp->data;
 	if (node->hdr.info.forw) {
-		if (INT_GET(node->hdr.info.forw, ARCH_CONVERT) == addblk->blkno) {
+		if (be32_to_cpu(node->hdr.info.forw) == addblk->blkno) {
 			bp = addblk->bp;
 		} else {
 			ASSERT(state->extravalid);
 			bp = state->extrablk.bp;
 		}
 		node = bp->data;
-		INT_SET(node->hdr.info.back, ARCH_CONVERT, oldblk->blkno);
+		node->hdr.info.back = cpu_to_be32(oldblk->blkno);
 		xfs_da_log_buf(state->args->trans, bp,
 		    XFS_DA_LOGRANGE(node, &node->hdr.info,
 		    sizeof(node->hdr.info)));
 	}
 	node = oldblk->bp->data;
-	if (INT_GET(node->hdr.info.back, ARCH_CONVERT)) {
-		if (INT_GET(node->hdr.info.back, ARCH_CONVERT) == addblk->blkno) {
+	if (node->hdr.info.back) {
+		if (be32_to_cpu(node->hdr.info.back) == addblk->blkno) {
 			bp = addblk->bp;
 		} else {
 			ASSERT(state->extravalid);
 			bp = state->extrablk.bp;
 		}
 		node = bp->data;
-		INT_SET(node->hdr.info.forw, ARCH_CONVERT, oldblk->blkno);
+		node->hdr.info.forw = cpu_to_be32(oldblk->blkno);
 		xfs_da_log_buf(state->args->trans, bp,
 		    XFS_DA_LOGRANGE(node, &node->hdr.info,
 		    sizeof(node->hdr.info)));
@@ -359,14 +359,14 @@ xfs_da_root_split(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	ASSERT(bp != NULL);
 	node = bp->data;
 	oldroot = blk1->bp->data;
-	if (INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
-		size = (int)((char *)&oldroot->btree[INT_GET(oldroot->hdr.count, ARCH_CONVERT)] -
+	if (be16_to_cpu(oldroot->hdr.info.magic) == XFS_DA_NODE_MAGIC) {
+		size = (int)((char *)&oldroot->btree[be16_to_cpu(oldroot->hdr.count)] -
 			     (char *)oldroot);
 	} else {
 		ASSERT(XFS_DIR_IS_V2(mp));
-		ASSERT(INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+		ASSERT(be16_to_cpu(oldroot->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 		leaf = (xfs_dir2_leaf_t *)oldroot;
-		size = (int)((char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] -
+		size = (int)((char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] -
 			     (char *)leaf);
 	}
 	memcpy(node, oldroot, size);
@@ -381,18 +381,18 @@ xfs_da_root_split(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	error = xfs_da_node_create(args,
 		args->whichfork == XFS_DATA_FORK &&
 		XFS_DIR_IS_V2(mp) ? mp->m_dirleafblk : 0,
-		INT_GET(node->hdr.level, ARCH_CONVERT) + 1, &bp, args->whichfork);
+		be16_to_cpu(node->hdr.level) + 1, &bp, args->whichfork);
 	if (error)
 		return(error);
 	node = bp->data;
-	INT_SET(node->btree[0].hashval, ARCH_CONVERT, blk1->hashval);
-	INT_SET(node->btree[0].before, ARCH_CONVERT, blk1->blkno);
-	INT_SET(node->btree[1].hashval, ARCH_CONVERT, blk2->hashval);
-	INT_SET(node->btree[1].before, ARCH_CONVERT, blk2->blkno);
-	INT_SET(node->hdr.count, ARCH_CONVERT, 2);
+	node->btree[0].hashval = cpu_to_be32(blk1->hashval);
+	node->btree[0].before = cpu_to_be32(blk1->blkno);
+	node->btree[1].hashval = cpu_to_be32(blk2->hashval);
+	node->btree[1].before = cpu_to_be32(blk2->blkno);
+	node->hdr.count = cpu_to_be16(2);
 
 #ifdef DEBUG
-	if (INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC) {
+	if (be16_to_cpu(oldroot->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC) {
 		ASSERT(blk1->blkno >= mp->m_dirleafblk &&
 		       blk1->blkno < mp->m_dirfreeblk);
 		ASSERT(blk2->blkno >= mp->m_dirleafblk &&
@@ -424,7 +424,7 @@ xfs_da_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	int useextra;
 
 	node = oldblk->bp->data;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 
 	/*
 	 * With V2 the extra block is data or freespace.
@@ -435,7 +435,7 @@ xfs_da_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	/*
 	 * Do we have to split the node?
 	 */
-	if ((INT_GET(node->hdr.count, ARCH_CONVERT) + newcount) > state->node_ents) {
+	if ((be16_to_cpu(node->hdr.count) + newcount) > state->node_ents) {
 		/*
 		 * Allocate a new node, add to the doubly linked chain of
 		 * nodes, then move some of our excess entries into it.
@@ -472,7 +472,7 @@ xfs_da_node_split(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	 * If we had double-split op below us, then add the extra block too.
 	 */
 	node = oldblk->bp->data;
-	if (oldblk->index <= INT_GET(node->hdr.count, ARCH_CONVERT)) {
+	if (oldblk->index <= be16_to_cpu(node->hdr.count)) {
 		oldblk->index++;
 		xfs_da_node_add(state, oldblk, addblk);
 		if (useextra) {
@@ -516,17 +516,17 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 * Figure out how many entries need to move, and in which direction.
 	 * Swap the nodes around if that makes it simpler.
 	 */
-	if ((INT_GET(node1->hdr.count, ARCH_CONVERT) > 0) && (INT_GET(node2->hdr.count, ARCH_CONVERT) > 0) &&
-	    ((INT_GET(node2->btree[ 0 ].hashval, ARCH_CONVERT) < INT_GET(node1->btree[ 0 ].hashval, ARCH_CONVERT)) ||
-	     (INT_GET(node2->btree[ INT_GET(node2->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT) <
-	      INT_GET(node1->btree[ INT_GET(node1->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT)))) {
+	if ((be16_to_cpu(node1->hdr.count) > 0) && (be16_to_cpu(node2->hdr.count) > 0) &&
+	    ((be32_to_cpu(node2->btree[0].hashval) < be32_to_cpu(node1->btree[0].hashval)) ||
+	     (be32_to_cpu(node2->btree[be16_to_cpu(node2->hdr.count)-1].hashval) <
+	      be32_to_cpu(node1->btree[be16_to_cpu(node1->hdr.count)-1].hashval)))) {
 		tmpnode = node1;
 		node1 = node2;
 		node2 = tmpnode;
 	}
-	ASSERT(INT_GET(node1->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	ASSERT(INT_GET(node2->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	count = (INT_GET(node1->hdr.count, ARCH_CONVERT) - INT_GET(node2->hdr.count, ARCH_CONVERT)) / 2;
+	ASSERT(be16_to_cpu(node1->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node2->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	count = (be16_to_cpu(node1->hdr.count) - be16_to_cpu(node2->hdr.count)) / 2;
 	if (count == 0)
 		return;
 	tp = state->args->trans;
@@ -537,7 +537,7 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		/*
 		 * Move elements in node2 up to make a hole.
 		 */
-		if ((tmp = INT_GET(node2->hdr.count, ARCH_CONVERT)) > 0) {
+		if ((tmp = be16_to_cpu(node2->hdr.count)) > 0) {
 			tmp *= (uint)sizeof(xfs_da_node_entry_t);
 			btree_s = &node2->btree[0];
 			btree_d = &node2->btree[count];
@@ -548,13 +548,12 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		 * Move the req'd B-tree elements from high in node1 to
 		 * low in node2.
 		 */
-		INT_MOD(node2->hdr.count, ARCH_CONVERT, count);
+		be16_add(&node2->hdr.count, count);
 		tmp = count * (uint)sizeof(xfs_da_node_entry_t);
-		btree_s = &node1->btree[INT_GET(node1->hdr.count, ARCH_CONVERT) - count];
+		btree_s = &node1->btree[be16_to_cpu(node1->hdr.count) - count];
 		btree_d = &node2->btree[0];
 		memcpy(btree_d, btree_s, tmp);
-		INT_MOD(node1->hdr.count, ARCH_CONVERT, -(count));
-
+		be16_add(&node1->hdr.count, -count);
 	} else {
 		/*
 		 * Move the req'd B-tree elements from low in node2 to
@@ -563,21 +562,21 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 		count = -count;
 		tmp = count * (uint)sizeof(xfs_da_node_entry_t);
 		btree_s = &node2->btree[0];
-		btree_d = &node1->btree[INT_GET(node1->hdr.count, ARCH_CONVERT)];
+		btree_d = &node1->btree[be16_to_cpu(node1->hdr.count)];
 		memcpy(btree_d, btree_s, tmp);
-		INT_MOD(node1->hdr.count, ARCH_CONVERT, count);
+		be16_add(&node1->hdr.count, count);
 		xfs_da_log_buf(tp, blk1->bp,
 			XFS_DA_LOGRANGE(node1, btree_d, tmp));
 
 		/*
 		 * Move elements in node2 down to fill the hole.
 		 */
-		tmp  = INT_GET(node2->hdr.count, ARCH_CONVERT) - count;
+		tmp  = be16_to_cpu(node2->hdr.count) - count;
 		tmp *= (uint)sizeof(xfs_da_node_entry_t);
 		btree_s = &node2->btree[count];
 		btree_d = &node2->btree[0];
 		memmove(btree_d, btree_s, tmp);
-		INT_MOD(node2->hdr.count, ARCH_CONVERT, -(count));
+		be16_add(&node2->hdr.count, -count);
 	}
 
 	/*
@@ -588,7 +587,7 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	xfs_da_log_buf(tp, blk2->bp,
 		XFS_DA_LOGRANGE(node2, &node2->hdr,
 			sizeof(node2->hdr) +
-			sizeof(node2->btree[0]) * INT_GET(node2->hdr.count, ARCH_CONVERT)));
+			sizeof(node2->btree[0]) * be16_to_cpu(node2->hdr.count)));
 
 	/*
 	 * Record the last hashval from each block for upward propagation.
@@ -596,15 +595,15 @@ xfs_da_node_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	 */
 	node1 = blk1->bp->data;
 	node2 = blk2->bp->data;
-	blk1->hashval = INT_GET(node1->btree[ INT_GET(node1->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
-	blk2->hashval = INT_GET(node2->btree[ INT_GET(node2->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+	blk1->hashval = be32_to_cpu(node1->btree[be16_to_cpu(node1->hdr.count)-1].hashval);
+	blk2->hashval = be32_to_cpu(node2->btree[be16_to_cpu(node2->hdr.count)-1].hashval);
 
 	/*
 	 * Adjust the expected index for insertion.
 	 */
-	if (blk1->index >= INT_GET(node1->hdr.count, ARCH_CONVERT)) {
-		blk2->index = blk1->index - INT_GET(node1->hdr.count, ARCH_CONVERT);
-		blk1->index = INT_GET(node1->hdr.count, ARCH_CONVERT) + 1;	/* make it invalid */
+	if (blk1->index >= be16_to_cpu(node1->hdr.count)) {
+		blk2->index = blk1->index - be16_to_cpu(node1->hdr.count);
+		blk1->index = be16_to_cpu(node1->hdr.count) + 1;	/* make it invalid */
 	}
 }
 
@@ -622,8 +621,8 @@ xfs_da_node_add(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 
 	node = oldblk->bp->data;
 	mp = state->mp;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	ASSERT((oldblk->index >= 0) && (oldblk->index <= INT_GET(node->hdr.count, ARCH_CONVERT)));
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	ASSERT((oldblk->index >= 0) && (oldblk->index <= be16_to_cpu(node->hdr.count)));
 	ASSERT(newblk->blkno != 0);
 	if (state->args->whichfork == XFS_DATA_FORK && XFS_DIR_IS_V2(mp))
 		ASSERT(newblk->blkno >= mp->m_dirleafblk &&
@@ -634,22 +633,22 @@ xfs_da_node_add(xfs_da_state_t *state, xfs_da_state_blk_t *oldblk,
 	 */
 	tmp = 0;
 	btree = &node->btree[ oldblk->index ];
-	if (oldblk->index < INT_GET(node->hdr.count, ARCH_CONVERT)) {
-		tmp = (INT_GET(node->hdr.count, ARCH_CONVERT) - oldblk->index) * (uint)sizeof(*btree);
+	if (oldblk->index < be16_to_cpu(node->hdr.count)) {
+		tmp = (be16_to_cpu(node->hdr.count) - oldblk->index) * (uint)sizeof(*btree);
 		memmove(btree + 1, btree, tmp);
 	}
-	INT_SET(btree->hashval, ARCH_CONVERT, newblk->hashval);
-	INT_SET(btree->before, ARCH_CONVERT, newblk->blkno);
+	btree->hashval = cpu_to_be32(newblk->hashval);
+	btree->before = cpu_to_be32(newblk->blkno);
 	xfs_da_log_buf(state->args->trans, oldblk->bp,
 		XFS_DA_LOGRANGE(node, btree, tmp + sizeof(*btree)));
-	INT_MOD(node->hdr.count, ARCH_CONVERT, +1);
+	be16_add(&node->hdr.count, 1);
 	xfs_da_log_buf(state->args->trans, oldblk->bp,
 		XFS_DA_LOGRANGE(node, &node->hdr, sizeof(node->hdr)));
 
 	/*
 	 * Copy the last hash value from the oldblk to propagate upwards.
 	 */
-	oldblk->hashval = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+	oldblk->hashval = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1 ].hashval);
 }
 
 /*========================================================================
@@ -768,21 +767,21 @@ xfs_da_root_join(xfs_da_state_t *state, xfs_da_state_blk_t *root_blk)
 	ASSERT(args != NULL);
 	ASSERT(root_blk->magic == XFS_DA_NODE_MAGIC);
 	oldroot = root_blk->bp->data;
-	ASSERT(INT_GET(oldroot->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(oldroot->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	ASSERT(!oldroot->hdr.info.forw);
 	ASSERT(!oldroot->hdr.info.back);
 
 	/*
 	 * If the root has more than one child, then don't do anything.
 	 */
-	if (INT_GET(oldroot->hdr.count, ARCH_CONVERT) > 1)
+	if (be16_to_cpu(oldroot->hdr.count) > 1)
 		return(0);
 
 	/*
 	 * Read in the (only) child block, then copy those bytes into
 	 * the root block's buffer and free the original child block.
 	 */
-	child = INT_GET(oldroot->btree[ 0 ].before, ARCH_CONVERT);
+	child = be32_to_cpu(oldroot->btree[0].before);
 	ASSERT(child != 0);
 	error = xfs_da_read_buf(args->trans, args->dp, child, -1, &bp,
 					     args->whichfork);
@@ -790,11 +789,11 @@ xfs_da_root_join(xfs_da_state_t *state, xfs_da_state_blk_t *root_blk)
 		return(error);
 	ASSERT(bp != NULL);
 	blkinfo = bp->data;
-	if (INT_GET(oldroot->hdr.level, ARCH_CONVERT) == 1) {
-		ASSERT(INT_GET(blkinfo->magic, ARCH_CONVERT) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
-		       INT_GET(blkinfo->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
+	if (be16_to_cpu(oldroot->hdr.level) == 1) {
+		ASSERT(be16_to_cpu(blkinfo->magic) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
+		       be16_to_cpu(blkinfo->magic) == XFS_ATTR_LEAF_MAGIC);
 	} else {
-		ASSERT(INT_GET(blkinfo->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+		ASSERT(be16_to_cpu(blkinfo->magic) == XFS_DA_NODE_MAGIC);
 	}
 	ASSERT(!blkinfo->forw);
 	ASSERT(!blkinfo->back);
@@ -830,9 +829,9 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC);
 	node = (xfs_da_intnode_t *)info;
-	count = INT_GET(node->hdr.count, ARCH_CONVERT);
+	count = be16_to_cpu(node->hdr.count);
 	if (count > (state->node_ents >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);	/* blk over 50%, don't try to join */
@@ -841,7 +840,7 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 	/*
 	 * Check for the degenerate case of the block being empty.
 	 * If the block is empty, we'll simply delete it, no need to
-	 * coalesce it with a sibling block.  We choose (aribtrarily)
+	 * coalesce it with a sibling block.  We choose (arbitrarily)
 	 * to merge with the forward block unless it is NULL.
 	 */
 	if (count == 0) {
@@ -849,7 +848,7 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward,
 						 0, &retval);
@@ -871,13 +870,12 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 	 * to shrink a directory over time.
 	 */
 	/* start with smaller blk num */
-	forward = (INT_GET(info->forw, ARCH_CONVERT)
-				< INT_GET(info->back, ARCH_CONVERT));
+	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));
 	for (i = 0; i < 2; forward = !forward, i++) {
 		if (forward)
-			blkno = INT_GET(info->forw, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->forw);
 		else
-			blkno = INT_GET(info->back, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		error = xfs_da_read_buf(state->args->trans, state->args->dp,
@@ -889,10 +887,10 @@ xfs_da_node_toosmall(xfs_da_state_t *state, int *action)
 		node = (xfs_da_intnode_t *)info;
 		count  = state->node_ents;
 		count -= state->node_ents >> 2;
-		count -= INT_GET(node->hdr.count, ARCH_CONVERT);
+		count -= be16_to_cpu(node->hdr.count);
 		node = bp->data;
-		ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-		count -= INT_GET(node->hdr.count, ARCH_CONVERT);
+		ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+		count -= be16_to_cpu(node->hdr.count);
 		xfs_da_brelse(state->args->trans, bp);
 		if (count >= 0)
 			break;	/* fits with at least 25% to spare */
@@ -973,16 +971,16 @@ xfs_da_fixhashpath(xfs_da_state_t *state, xfs_da_state_path_t *path)
 	}
 	for (blk--, level--; level >= 0; blk--, level--) {
 		node = blk->bp->data;
-		ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+		ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 		btree = &node->btree[ blk->index ];
-		if (INT_GET(btree->hashval, ARCH_CONVERT) == lasthash)
+		if (be32_to_cpu(btree->hashval) == lasthash)
 			break;
 		blk->hashval = lasthash;
-		INT_SET(btree->hashval, ARCH_CONVERT, lasthash);
+		btree->hashval = cpu_to_be32(lasthash);
 		xfs_da_log_buf(state->args->trans, blk->bp,
 				  XFS_DA_LOGRANGE(node, btree, sizeof(*btree)));
 
-		lasthash = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+		lasthash = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 	}
 }
 
@@ -997,25 +995,25 @@ xfs_da_node_remove(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk)
 	int tmp;
 
 	node = drop_blk->bp->data;
-	ASSERT(drop_blk->index < INT_GET(node->hdr.count, ARCH_CONVERT));
+	ASSERT(drop_blk->index < be16_to_cpu(node->hdr.count));
 	ASSERT(drop_blk->index >= 0);
 
 	/*
 	 * Copy over the offending entry, or just zero it out.
 	 */
 	btree = &node->btree[drop_blk->index];
-	if (drop_blk->index < (INT_GET(node->hdr.count, ARCH_CONVERT)-1)) {
-		tmp  = INT_GET(node->hdr.count, ARCH_CONVERT) - drop_blk->index - 1;
+	if (drop_blk->index < (be16_to_cpu(node->hdr.count)-1)) {
+		tmp  = be16_to_cpu(node->hdr.count) - drop_blk->index - 1;
 		tmp *= (uint)sizeof(xfs_da_node_entry_t);
 		memmove(btree, btree + 1, tmp);
 		xfs_da_log_buf(state->args->trans, drop_blk->bp,
 		    XFS_DA_LOGRANGE(node, btree, tmp));
-		btree = &node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ];
+		btree = &node->btree[be16_to_cpu(node->hdr.count)-1];
 	}
 	memset((char *)btree, 0, sizeof(xfs_da_node_entry_t));
 	xfs_da_log_buf(state->args->trans, drop_blk->bp,
 	    XFS_DA_LOGRANGE(node, btree, sizeof(*btree)));
-	INT_MOD(node->hdr.count, ARCH_CONVERT, -1);
+	be16_add(&node->hdr.count, -1);
 	xfs_da_log_buf(state->args->trans, drop_blk->bp,
 	    XFS_DA_LOGRANGE(node, &node->hdr, sizeof(node->hdr)));
 
@@ -1023,7 +1021,7 @@ xfs_da_node_remove(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk)
 	 * Copy the last hash value from the block to propagate upwards.
 	 */
 	btree--;
-	drop_blk->hashval = INT_GET(btree->hashval, ARCH_CONVERT);
+	drop_blk->hashval = be32_to_cpu(btree->hashval);
 }
 
 /*
@@ -1041,40 +1039,40 @@ xfs_da_node_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 
 	drop_node = drop_blk->bp->data;
 	save_node = save_blk->bp->data;
-	ASSERT(INT_GET(drop_node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-	ASSERT(INT_GET(save_node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(drop_node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(save_node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	tp = state->args->trans;
 
 	/*
 	 * If the dying block has lower hashvals, then move all the
 	 * elements in the remaining block up to make a hole.
 	 */
-	if ((INT_GET(drop_node->btree[ 0 ].hashval, ARCH_CONVERT) < INT_GET(save_node->btree[ 0 ].hashval, ARCH_CONVERT)) ||
-	    (INT_GET(drop_node->btree[ INT_GET(drop_node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT) <
-	     INT_GET(save_node->btree[ INT_GET(save_node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT)))
+	if ((be32_to_cpu(drop_node->btree[0].hashval) < be32_to_cpu(save_node->btree[ 0 ].hashval)) ||
+	    (be32_to_cpu(drop_node->btree[be16_to_cpu(drop_node->hdr.count)-1].hashval) <
+	     be32_to_cpu(save_node->btree[be16_to_cpu(save_node->hdr.count)-1].hashval)))
 	{
-		btree = &save_node->btree[ INT_GET(drop_node->hdr.count, ARCH_CONVERT) ];
-		tmp = INT_GET(save_node->hdr.count, ARCH_CONVERT) * (uint)sizeof(xfs_da_node_entry_t);
+		btree = &save_node->btree[be16_to_cpu(drop_node->hdr.count)];
+		tmp = be16_to_cpu(save_node->hdr.count) * (uint)sizeof(xfs_da_node_entry_t);
 		memmove(btree, &save_node->btree[0], tmp);
 		btree = &save_node->btree[0];
 		xfs_da_log_buf(tp, save_blk->bp,
 			XFS_DA_LOGRANGE(save_node, btree,
-				(INT_GET(save_node->hdr.count, ARCH_CONVERT) + INT_GET(drop_node->hdr.count, ARCH_CONVERT)) *
+				(be16_to_cpu(save_node->hdr.count) + be16_to_cpu(drop_node->hdr.count)) *
 				sizeof(xfs_da_node_entry_t)));
 	} else {
-		btree = &save_node->btree[ INT_GET(save_node->hdr.count, ARCH_CONVERT) ];
+		btree = &save_node->btree[be16_to_cpu(save_node->hdr.count)];
 		xfs_da_log_buf(tp, save_blk->bp,
 			XFS_DA_LOGRANGE(save_node, btree,
-				INT_GET(drop_node->hdr.count, ARCH_CONVERT) *
+				be16_to_cpu(drop_node->hdr.count) *
 				sizeof(xfs_da_node_entry_t)));
 	}
 
 	/*
 	 * Move all the B-tree elements from drop_blk to save_blk.
 	 */
-	tmp = INT_GET(drop_node->hdr.count, ARCH_CONVERT) * (uint)sizeof(xfs_da_node_entry_t);
+	tmp = be16_to_cpu(drop_node->hdr.count) * (uint)sizeof(xfs_da_node_entry_t);
 	memcpy(btree, &drop_node->btree[0], tmp);
-	INT_MOD(save_node->hdr.count, ARCH_CONVERT, INT_GET(drop_node->hdr.count, ARCH_CONVERT));
+	be16_add(&save_node->hdr.count, be16_to_cpu(drop_node->hdr.count));
 
 	xfs_da_log_buf(tp, save_blk->bp,
 		XFS_DA_LOGRANGE(save_node, &save_node->hdr,
@@ -1083,7 +1081,7 @@ xfs_da_node_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	/*
 	 * Save the last hashval in the remaining block for upward propagation.
 	 */
-	save_blk->hashval = INT_GET(save_node->btree[ INT_GET(save_node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+	save_blk->hashval = be32_to_cpu(save_node->btree[be16_to_cpu(save_node->hdr.count)-1].hashval);
 }
 
 /*========================================================================
@@ -1138,46 +1136,46 @@ xfs_da_node_lookup_int(xfs_da_state_t *state, int *result)
 			return(error);
 		}
 		curr = blk->bp->data;
-		ASSERT(INT_GET(curr->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC ||
-		       INT_GET(curr->magic, ARCH_CONVERT) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
-		       INT_GET(curr->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
+		ASSERT(be16_to_cpu(curr->magic) == XFS_DA_NODE_MAGIC ||
+		       be16_to_cpu(curr->magic) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
+		       be16_to_cpu(curr->magic) == XFS_ATTR_LEAF_MAGIC);
 
 		/*
 		 * Search an intermediate node for a match.
 		 */
-		blk->magic = INT_GET(curr->magic, ARCH_CONVERT);
-		if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
+		blk->magic = be16_to_cpu(curr->magic);
+		if (blk->magic == XFS_DA_NODE_MAGIC) {
 			node = blk->bp->data;
-			blk->hashval = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+			blk->hashval = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 
 			/*
 			 * Binary search.  (note: small blocks will skip loop)
 			 */
-			max = INT_GET(node->hdr.count, ARCH_CONVERT);
+			max = be16_to_cpu(node->hdr.count);
 			probe = span = max / 2;
 			hashval = args->hashval;
 			for (btree = &node->btree[probe]; span > 4;
 				   btree = &node->btree[probe]) {
 				span /= 2;
-				if (INT_GET(btree->hashval, ARCH_CONVERT) < hashval)
+				if (be32_to_cpu(btree->hashval) < hashval)
 					probe += span;
-				else if (INT_GET(btree->hashval, ARCH_CONVERT) > hashval)
+				else if (be32_to_cpu(btree->hashval) > hashval)
 					probe -= span;
 				else
 					break;
 			}
 			ASSERT((probe >= 0) && (probe < max));
-			ASSERT((span <= 4) || (INT_GET(btree->hashval, ARCH_CONVERT) == hashval));
+			ASSERT((span <= 4) || (be32_to_cpu(btree->hashval) == hashval));
 
 			/*
 			 * Since we may have duplicate hashval's, find the first
 			 * matching hashval in the node.
 			 */
-			while ((probe > 0) && (INT_GET(btree->hashval, ARCH_CONVERT) >= hashval)) {
+			while ((probe > 0) && (be32_to_cpu(btree->hashval) >= hashval)) {
 				btree--;
 				probe--;
 			}
-			while ((probe < max) && (INT_GET(btree->hashval, ARCH_CONVERT) < hashval)) {
+			while ((probe < max) && (be32_to_cpu(btree->hashval) < hashval)) {
 				btree++;
 				probe++;
 			}
@@ -1187,21 +1185,21 @@ xfs_da_node_lookup_int(xfs_da_state_t *state, int *result)
 			 */
 			if (probe == max) {
 				blk->index = max-1;
-				blkno = INT_GET(node->btree[ max-1 ].before, ARCH_CONVERT);
+				blkno = be32_to_cpu(node->btree[max-1].before);
 			} else {
 				blk->index = probe;
-				blkno = INT_GET(btree->before, ARCH_CONVERT);
+				blkno = be32_to_cpu(btree->before);
 			}
 		}
-		else if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC) {
+		else if (be16_to_cpu(curr->magic) == XFS_ATTR_LEAF_MAGIC) {
 			blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL);
 			break;
 		}
-		else if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC) {
+		else if (be16_to_cpu(curr->magic) == XFS_DIR_LEAF_MAGIC) {
 			blk->hashval = xfs_dir_leaf_lasthash(blk->bp, NULL);
 			break;
 		}
-		else if (INT_GET(curr->magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC) {
+		else if (be16_to_cpu(curr->magic) == XFS_DIR2_LEAFN_MAGIC) {
 			blk->hashval = xfs_dir2_leafn_lasthash(blk->bp, NULL);
 			break;
 		}
@@ -1274,8 +1272,8 @@ xfs_da_blk_link(xfs_da_state_t *state, xfs_da_state_blk_t *old_blk,
 	ASSERT(old_blk->magic == XFS_DA_NODE_MAGIC ||
 	       old_blk->magic == XFS_DIRX_LEAF_MAGIC(state->mp) ||
 	       old_blk->magic == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(old_blk->magic == INT_GET(old_info->magic, ARCH_CONVERT));
-	ASSERT(new_blk->magic == INT_GET(new_info->magic, ARCH_CONVERT));
+	ASSERT(old_blk->magic == be16_to_cpu(old_info->magic));
+	ASSERT(new_blk->magic == be16_to_cpu(new_info->magic));
 	ASSERT(old_blk->magic == new_blk->magic);
 
 	switch (old_blk->magic) {
@@ -1302,47 +1300,44 @@ xfs_da_blk_link(xfs_da_state_t *state, xfs_da_state_blk_t *old_blk,
 		/*
 		 * Link new block in before existing block.
 		 */
-		INT_SET(new_info->forw, ARCH_CONVERT, old_blk->blkno);
-		new_info->back = old_info->back; /* INT_: direct copy */
-		if (INT_GET(old_info->back, ARCH_CONVERT)) {
+		new_info->forw = cpu_to_be32(old_blk->blkno);
+		new_info->back = old_info->back;
+		if (old_info->back) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(old_info->back,
-							ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(old_info->back),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT) == INT_GET(old_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->forw, ARCH_CONVERT) == old_blk->blkno);
-			INT_SET(tmp_info->forw, ARCH_CONVERT, new_blk->blkno);
+			ASSERT(be16_to_cpu(tmp_info->magic) == be16_to_cpu(old_info->magic));
+			ASSERT(be32_to_cpu(tmp_info->forw) == old_blk->blkno);
+			tmp_info->forw = cpu_to_be32(new_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0, sizeof(*tmp_info)-1);
 			xfs_da_buf_done(bp);
 		}
-		INT_SET(old_info->back, ARCH_CONVERT, new_blk->blkno);
+		old_info->back = cpu_to_be32(new_blk->blkno);
 	} else {
 		/*
 		 * Link new block in after existing block.
 		 */
-		new_info->forw = old_info->forw; /* INT_: direct copy */
-		INT_SET(new_info->back, ARCH_CONVERT, old_blk->blkno);
-		if (INT_GET(old_info->forw, ARCH_CONVERT)) {
+		new_info->forw = old_info->forw;
+		new_info->back = cpu_to_be32(old_blk->blkno);
+		if (old_info->forw) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(old_info->forw, ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(old_info->forw),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT)
-				    == INT_GET(old_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->back, ARCH_CONVERT)
-				    == old_blk->blkno);
-			INT_SET(tmp_info->back, ARCH_CONVERT, new_blk->blkno);
+			ASSERT(tmp_info->magic == old_info->magic);
+			ASSERT(be32_to_cpu(tmp_info->back) == old_blk->blkno);
+			tmp_info->back = cpu_to_be32(new_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0, sizeof(*tmp_info)-1);
 			xfs_da_buf_done(bp);
 		}
-		INT_SET(old_info->forw, ARCH_CONVERT, new_blk->blkno);
+		old_info->forw = cpu_to_be32(new_blk->blkno);
 	}
 
 	xfs_da_log_buf(args->trans, old_blk->bp, 0, sizeof(*tmp_info) - 1);
@@ -1360,13 +1355,13 @@ xfs_da_node_order(xfs_dabuf_t *node1_bp, xfs_dabuf_t *node2_bp)
 
 	node1 = node1_bp->data;
 	node2 = node2_bp->data;
-	ASSERT((INT_GET(node1->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) &&
-	       (INT_GET(node2->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC));
-	if ((INT_GET(node1->hdr.count, ARCH_CONVERT) > 0) && (INT_GET(node2->hdr.count, ARCH_CONVERT) > 0) &&
-	    ((INT_GET(node2->btree[ 0 ].hashval, ARCH_CONVERT) <
-	      INT_GET(node1->btree[ 0 ].hashval, ARCH_CONVERT)) ||
-	     (INT_GET(node2->btree[ INT_GET(node2->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT) <
-	      INT_GET(node1->btree[ INT_GET(node1->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT)))) {
+	ASSERT((be16_to_cpu(node1->hdr.info.magic) == XFS_DA_NODE_MAGIC) &&
+	       (be16_to_cpu(node2->hdr.info.magic) == XFS_DA_NODE_MAGIC));
+	if ((be16_to_cpu(node1->hdr.count) > 0) && (be16_to_cpu(node2->hdr.count) > 0) &&
+	    ((be32_to_cpu(node2->btree[0].hashval) <
+	      be32_to_cpu(node1->btree[0].hashval)) ||
+	     (be32_to_cpu(node2->btree[be16_to_cpu(node2->hdr.count)-1].hashval) <
+	      be32_to_cpu(node1->btree[be16_to_cpu(node1->hdr.count)-1].hashval)))) {
 		return(1);
 	}
 	return(0);
@@ -1381,12 +1376,12 @@ xfs_da_node_lasthash(xfs_dabuf_t *bp, int *count)
 	xfs_da_intnode_t *node;
 
 	node = bp->data;
-	ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+	ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
 	if (count)
-		*count = INT_GET(node->hdr.count, ARCH_CONVERT);
+		*count = be16_to_cpu(node->hdr.count);
 	if (!node->hdr.count)
 		return(0);
-	return(INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT));
+	return be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 }
 
 /*
@@ -1411,50 +1406,47 @@ xfs_da_blk_unlink(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	ASSERT(save_blk->magic == XFS_DA_NODE_MAGIC ||
 	       save_blk->magic == XFS_DIRX_LEAF_MAGIC(state->mp) ||
 	       save_blk->magic == XFS_ATTR_LEAF_MAGIC);
-	ASSERT(save_blk->magic == INT_GET(save_info->magic, ARCH_CONVERT));
-	ASSERT(drop_blk->magic == INT_GET(drop_info->magic, ARCH_CONVERT));
+	ASSERT(save_blk->magic == be16_to_cpu(save_info->magic));
+	ASSERT(drop_blk->magic == be16_to_cpu(drop_info->magic));
 	ASSERT(save_blk->magic == drop_blk->magic);
-	ASSERT((INT_GET(save_info->forw, ARCH_CONVERT) == drop_blk->blkno) ||
-	       (INT_GET(save_info->back, ARCH_CONVERT) == drop_blk->blkno));
-	ASSERT((INT_GET(drop_info->forw, ARCH_CONVERT) == save_blk->blkno) ||
-	       (INT_GET(drop_info->back, ARCH_CONVERT) == save_blk->blkno));
+	ASSERT((be32_to_cpu(save_info->forw) == drop_blk->blkno) ||
+	       (be32_to_cpu(save_info->back) == drop_blk->blkno));
+	ASSERT((be32_to_cpu(drop_info->forw) == save_blk->blkno) ||
+	       (be32_to_cpu(drop_info->back) == save_blk->blkno));
 
 	/*
 	 * Unlink the leaf block from the doubly linked chain of leaves.
 	 */
-	if (INT_GET(save_info->back, ARCH_CONVERT) == drop_blk->blkno) {
-		save_info->back = drop_info->back; /* INT_: direct copy */
-		if (INT_GET(drop_info->back, ARCH_CONVERT)) {
+	if (be32_to_cpu(save_info->back) == drop_blk->blkno) {
+		save_info->back = drop_info->back;
+		if (drop_info->back) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(drop_info->back,
-							ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(drop_info->back),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT) == INT_GET(save_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->forw, ARCH_CONVERT) == drop_blk->blkno);
-			INT_SET(tmp_info->forw, ARCH_CONVERT, save_blk->blkno);
+			ASSERT(tmp_info->magic == save_info->magic);
+			ASSERT(be32_to_cpu(tmp_info->forw) == drop_blk->blkno);
+			tmp_info->forw = cpu_to_be32(save_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0,
 						    sizeof(*tmp_info) - 1);
 			xfs_da_buf_done(bp);
 		}
 	} else {
-		save_info->forw = drop_info->forw; /* INT_: direct copy */
-		if (INT_GET(drop_info->forw, ARCH_CONVERT)) {
+		save_info->forw = drop_info->forw;
+		if (drop_info->forw) {
 			error = xfs_da_read_buf(args->trans, args->dp,
-						INT_GET(drop_info->forw, ARCH_CONVERT), -1, &bp,
-						args->whichfork);
+						be32_to_cpu(drop_info->forw),
+						-1, &bp, args->whichfork);
 			if (error)
 				return(error);
 			ASSERT(bp != NULL);
 			tmp_info = bp->data;
-			ASSERT(INT_GET(tmp_info->magic, ARCH_CONVERT)
-				    == INT_GET(save_info->magic, ARCH_CONVERT));
-			ASSERT(INT_GET(tmp_info->back, ARCH_CONVERT)
-				    == drop_blk->blkno);
-			INT_SET(tmp_info->back, ARCH_CONVERT, save_blk->blkno);
+			ASSERT(tmp_info->magic == save_info->magic);
+			ASSERT(be32_to_cpu(tmp_info->back) == drop_blk->blkno);
+			tmp_info->back = cpu_to_be32(save_blk->blkno);
 			xfs_da_log_buf(args->trans, bp, 0,
 						    sizeof(*tmp_info) - 1);
 			xfs_da_buf_done(bp);
@@ -1497,14 +1489,14 @@ xfs_da_path_shift(xfs_da_state_t *state, xfs_da_state_path_t *path,
 	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
 		ASSERT(blk->bp != NULL);
 		node = blk->bp->data;
-		ASSERT(INT_GET(node->hdr.info.magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
-		if (forward && (blk->index < INT_GET(node->hdr.count, ARCH_CONVERT)-1)) {
+		ASSERT(be16_to_cpu(node->hdr.info.magic) == XFS_DA_NODE_MAGIC);
+		if (forward && (blk->index < be16_to_cpu(node->hdr.count)-1)) {
 			blk->index++;
-			blkno = INT_GET(node->btree[ blk->index ].before, ARCH_CONVERT);
+			blkno = be32_to_cpu(node->btree[blk->index].before);
 			break;
 		} else if (!forward && (blk->index > 0)) {
 			blk->index--;
-			blkno = INT_GET(node->btree[ blk->index ].before, ARCH_CONVERT);
+			blkno = be32_to_cpu(node->btree[blk->index].before);
 			break;
 		}
 	}
@@ -1536,18 +1528,18 @@ xfs_da_path_shift(xfs_da_state_t *state, xfs_da_state_path_t *path,
 			return(error);
 		ASSERT(blk->bp != NULL);
 		info = blk->bp->data;
-		ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC ||
-		       INT_GET(info->magic, ARCH_CONVERT) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
-		       INT_GET(info->magic, ARCH_CONVERT) == XFS_ATTR_LEAF_MAGIC);
-		blk->magic = INT_GET(info->magic, ARCH_CONVERT);
-		if (INT_GET(info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC) {
+		ASSERT(be16_to_cpu(info->magic) == XFS_DA_NODE_MAGIC ||
+		       be16_to_cpu(info->magic) == XFS_DIRX_LEAF_MAGIC(state->mp) ||
+		       be16_to_cpu(info->magic) == XFS_ATTR_LEAF_MAGIC);
+		blk->magic = be16_to_cpu(info->magic);
+		if (blk->magic == XFS_DA_NODE_MAGIC) {
 			node = (xfs_da_intnode_t *)info;
-			blk->hashval = INT_GET(node->btree[ INT_GET(node->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT);
+			blk->hashval = be32_to_cpu(node->btree[be16_to_cpu(node->hdr.count)-1].hashval);
 			if (forward)
 				blk->index = 0;
 			else
-				blk->index = INT_GET(node->hdr.count, ARCH_CONVERT)-1;
-			blkno = INT_GET(node->btree[ blk->index ].before, ARCH_CONVERT);
+				blk->index = be16_to_cpu(node->hdr.count)-1;
+			blkno = be32_to_cpu(node->btree[blk->index].before);
 		} else {
 			ASSERT(level == path->active-1);
 			blk->index = 0;
@@ -1788,40 +1780,40 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	/*
 	 * Get values from the moved block.
 	 */
-	if (INT_GET(dead_info->magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC) {
+	if (be16_to_cpu(dead_info->magic) == XFS_DIR_LEAF_MAGIC) {
 		ASSERT(XFS_DIR_IS_V1(mp));
 		dead_leaf = (xfs_dir_leafblock_t *)dead_info;
 		dead_level = 0;
 		dead_hash =
 			INT_GET(dead_leaf->entries[INT_GET(dead_leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
-	} else if (INT_GET(dead_info->magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC) {
+	} else if (be16_to_cpu(dead_info->magic) == XFS_DIR2_LEAFN_MAGIC) {
 		ASSERT(XFS_DIR_IS_V2(mp));
 		dead_leaf2 = (xfs_dir2_leaf_t *)dead_info;
 		dead_level = 0;
-		dead_hash = INT_GET(dead_leaf2->ents[INT_GET(dead_leaf2->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+		dead_hash = be32_to_cpu(dead_leaf2->ents[be16_to_cpu(dead_leaf2->hdr.count) - 1].hashval);
 	} else {
-		ASSERT(INT_GET(dead_info->magic, ARCH_CONVERT) == XFS_DA_NODE_MAGIC);
+		ASSERT(be16_to_cpu(dead_info->magic) == XFS_DA_NODE_MAGIC);
 		dead_node = (xfs_da_intnode_t *)dead_info;
-		dead_level = INT_GET(dead_node->hdr.level, ARCH_CONVERT);
-		dead_hash = INT_GET(dead_node->btree[INT_GET(dead_node->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+		dead_level = be16_to_cpu(dead_node->hdr.level);
+		dead_hash = be32_to_cpu(dead_node->btree[be16_to_cpu(dead_node->hdr.count) - 1].hashval);
 	}
 	sib_buf = par_buf = NULL;
 	/*
 	 * If the moved block has a left sibling, fix up the pointers.
 	 */
-	if ((sib_blkno = INT_GET(dead_info->back, ARCH_CONVERT))) {
+	if ((sib_blkno = be32_to_cpu(dead_info->back))) {
 		if ((error = xfs_da_read_buf(tp, ip, sib_blkno, -1, &sib_buf, w)))
 			goto done;
 		sib_info = sib_buf->data;
 		if (unlikely(
-		    INT_GET(sib_info->forw, ARCH_CONVERT) != last_blkno ||
-		    INT_GET(sib_info->magic, ARCH_CONVERT) != INT_GET(dead_info->magic, ARCH_CONVERT))) {
+		    be32_to_cpu(sib_info->forw) != last_blkno ||
+		    sib_info->magic != dead_info->magic)) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(2)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		INT_SET(sib_info->forw, ARCH_CONVERT, dead_blkno);
+		sib_info->forw = cpu_to_be32(dead_blkno);
 		xfs_da_log_buf(tp, sib_buf,
 			XFS_DA_LOGRANGE(sib_info, &sib_info->forw,
 					sizeof(sib_info->forw)));
@@ -1831,20 +1823,19 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	/*
 	 * If the moved block has a right sibling, fix up the pointers.
 	 */
-	if ((sib_blkno = INT_GET(dead_info->forw, ARCH_CONVERT))) {
+	if ((sib_blkno = be32_to_cpu(dead_info->forw))) {
 		if ((error = xfs_da_read_buf(tp, ip, sib_blkno, -1, &sib_buf, w)))
 			goto done;
 		sib_info = sib_buf->data;
 		if (unlikely(
-		       INT_GET(sib_info->back, ARCH_CONVERT) != last_blkno
-		    || INT_GET(sib_info->magic, ARCH_CONVERT)
-				!= INT_GET(dead_info->magic, ARCH_CONVERT))) {
+		       be32_to_cpu(sib_info->back) != last_blkno ||
+		       sib_info->magic != dead_info->magic)) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(3)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		INT_SET(sib_info->back, ARCH_CONVERT, dead_blkno);
+		sib_info->back = cpu_to_be32(dead_blkno);
 		xfs_da_log_buf(tp, sib_buf,
 			XFS_DA_LOGRANGE(sib_info, &sib_info->back,
 					sizeof(sib_info->back)));
@@ -1861,26 +1852,26 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 			goto done;
 		par_node = par_buf->data;
 		if (unlikely(
-		    INT_GET(par_node->hdr.info.magic, ARCH_CONVERT) != XFS_DA_NODE_MAGIC ||
-		    (level >= 0 && level != INT_GET(par_node->hdr.level, ARCH_CONVERT) + 1))) {
+		    be16_to_cpu(par_node->hdr.info.magic) != XFS_DA_NODE_MAGIC ||
+		    (level >= 0 && level != be16_to_cpu(par_node->hdr.level) + 1))) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(4)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		level = INT_GET(par_node->hdr.level, ARCH_CONVERT);
+		level = be16_to_cpu(par_node->hdr.level);
 		for (entno = 0;
-		     entno < INT_GET(par_node->hdr.count, ARCH_CONVERT) &&
-		     INT_GET(par_node->btree[entno].hashval, ARCH_CONVERT) < dead_hash;
+		     entno < be16_to_cpu(par_node->hdr.count) &&
+		     be32_to_cpu(par_node->btree[entno].hashval) < dead_hash;
 		     entno++)
 			continue;
-		if (unlikely(entno == INT_GET(par_node->hdr.count, ARCH_CONVERT))) {
+		if (unlikely(entno == be16_to_cpu(par_node->hdr.count))) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(5)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
 			goto done;
 		}
-		par_blkno = INT_GET(par_node->btree[entno].before, ARCH_CONVERT);
+		par_blkno = be32_to_cpu(par_node->btree[entno].before);
 		if (level == dead_level + 1)
 			break;
 		xfs_da_brelse(tp, par_buf);
@@ -1892,13 +1883,13 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	 */
 	for (;;) {
 		for (;
-		     entno < INT_GET(par_node->hdr.count, ARCH_CONVERT) &&
-		     INT_GET(par_node->btree[entno].before, ARCH_CONVERT) != last_blkno;
+		     entno < be16_to_cpu(par_node->hdr.count) &&
+		     be32_to_cpu(par_node->btree[entno].before) != last_blkno;
 		     entno++)
 			continue;
-		if (entno < INT_GET(par_node->hdr.count, ARCH_CONVERT))
+		if (entno < be16_to_cpu(par_node->hdr.count))
 			break;
-		par_blkno = INT_GET(par_node->hdr.info.forw, ARCH_CONVERT);
+		par_blkno = be32_to_cpu(par_node->hdr.info.forw);
 		xfs_da_brelse(tp, par_buf);
 		par_buf = NULL;
 		if (unlikely(par_blkno == 0)) {
@@ -1911,8 +1902,8 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 			goto done;
 		par_node = par_buf->data;
 		if (unlikely(
-		    INT_GET(par_node->hdr.level, ARCH_CONVERT) != level ||
-		    INT_GET(par_node->hdr.info.magic, ARCH_CONVERT) != XFS_DA_NODE_MAGIC)) {
+		    be16_to_cpu(par_node->hdr.level) != level ||
+		    be16_to_cpu(par_node->hdr.info.magic) != XFS_DA_NODE_MAGIC)) {
 			XFS_ERROR_REPORT("xfs_da_swap_lastblock(7)",
 					 XFS_ERRLEVEL_LOW, mp);
 			error = XFS_ERROR(EFSCORRUPTED);
@@ -1923,7 +1914,7 @@ xfs_da_swap_lastblock(xfs_da_args_t *args, xfs_dablk_t *dead_blknop,
 	/*
 	 * Update the parent entry pointing to the moved block.
 	 */
-	INT_SET(par_node->btree[entno].before, ARCH_CONVERT, dead_blkno);
+	par_node->btree[entno].before = cpu_to_be32(dead_blkno);
 	xfs_da_log_buf(tp, par_buf,
 		XFS_DA_LOGRANGE(par_node, &par_node->btree[entno].before,
 				sizeof(par_node->btree[entno].before)));
@@ -2203,8 +2194,8 @@ xfs_da_do_buf(
 		info = rbp->data;
 		data = rbp->data;
 		free = rbp->data;
-		magic = INT_GET(info->magic, ARCH_CONVERT);
-		magic1 = INT_GET(data->hdr.magic, ARCH_CONVERT);
+		magic = be16_to_cpu(info->magic);
+		magic1 = be32_to_cpu(data->hdr.magic);
 		if (unlikely(
 		    XFS_TEST_ERROR((magic != XFS_DA_NODE_MAGIC) &&
 				   (magic != XFS_DIR_LEAF_MAGIC) &&
@@ -2213,7 +2204,7 @@ xfs_da_do_buf(
 				   (magic != XFS_DIR2_LEAFN_MAGIC) &&
 				   (magic1 != XFS_DIR2_BLOCK_MAGIC) &&
 				   (magic1 != XFS_DIR2_DATA_MAGIC) &&
-				   (INT_GET(free->hdr.magic, ARCH_CONVERT) != XFS_DIR2_FREE_MAGIC),
+				   (be32_to_cpu(free->hdr.magic) != XFS_DIR2_FREE_MAGIC),
 				mp, XFS_ERRTAG_DA_READ_BUF,
 				XFS_RANDOM_DA_READ_BUF))) {
 			xfs_buftrace("DA READ ERROR", rbp->bps[0]);
diff --git a/fs/xfs/xfs_da_btree.h b/fs/xfs/xfs_da_btree.h
index 41352113721a..243a730d5ec8 100644
--- a/fs/xfs/xfs_da_btree.h
+++ b/fs/xfs/xfs_da_btree.h
@@ -45,10 +45,10 @@ struct zone;
 	(XFS_DIR_IS_V1(mp) ? XFS_DIR_LEAF_MAGIC : XFS_DIR2_LEAFN_MAGIC)
 
 typedef struct xfs_da_blkinfo {
-	xfs_dablk_t forw;			/* previous block in list */
-	xfs_dablk_t back;			/* following block in list */
-	__uint16_t magic;			/* validity check on block */
-	__uint16_t pad;				/* unused */
+	__be32		forw;			/* previous block in list */
+	__be32		back;			/* following block in list */
+	__be16		magic;			/* validity check on block */
+	__be16		pad;			/* unused */
 } xfs_da_blkinfo_t;
 
 /*
@@ -65,12 +65,12 @@ typedef struct xfs_da_blkinfo {
 typedef struct xfs_da_intnode {
 	struct xfs_da_node_hdr {	/* constant-structure header block */
 		xfs_da_blkinfo_t info;	/* block type, links, etc. */
-		__uint16_t count;	/* count of active entries */
-		__uint16_t level;	/* level above leaves (leaf == 0) */
+		__be16	count;		/* count of active entries */
+		__be16	level;		/* level above leaves (leaf == 0) */
 	} hdr;
 	struct xfs_da_node_entry {
-		xfs_dahash_t hashval;	/* hash value for this descendant */
-		xfs_dablk_t before;	/* Btree block before this key */
+		__be32	hashval;	/* hash value for this descendant */
+		__be32	before;		/* Btree block before this key */
 	} btree[1];			/* variable sized array of keys */
 } xfs_da_intnode_t;
 typedef struct xfs_da_node_hdr xfs_da_node_hdr_t;
diff --git a/fs/xfs/xfs_dfrag.c b/fs/xfs/xfs_dfrag.c
index c6191d00ad27..4968a6358e61 100644
--- a/fs/xfs/xfs_dfrag.c
+++ b/fs/xfs/xfs_dfrag.c
@@ -83,7 +83,7 @@ xfs_swapext(
 
 	/* Pull information for the target fd */
 	if (((fp = fget((int)sxp->sx_fdtarget)) == NULL) ||
-	    ((vp = LINVFS_GET_VP(fp->f_dentry->d_inode)) == NULL))  {
+	    ((vp = vn_from_inode(fp->f_dentry->d_inode)) == NULL))  {
 		error = XFS_ERROR(EINVAL);
 		goto error0;
 	}
@@ -95,7 +95,7 @@ xfs_swapext(
 	}
 
 	if (((tfp = fget((int)sxp->sx_fdtmp)) == NULL) ||
-	    ((tvp = LINVFS_GET_VP(tfp->f_dentry->d_inode)) == NULL)) {
+	    ((tvp = vn_from_inode(tfp->f_dentry->d_inode)) == NULL)) {
 		error = XFS_ERROR(EINVAL);
 		goto error0;
 	}
diff --git a/fs/xfs/xfs_dir.c b/fs/xfs/xfs_dir.c
index bb87d2a700a9..9cc702a839a3 100644
--- a/fs/xfs/xfs_dir.c
+++ b/fs/xfs/xfs_dir.c
@@ -634,7 +634,7 @@ xfs_dir_leaf_removename(xfs_da_args_t *args, int *count, int *totallen)
 		return(retval);
 	ASSERT(bp != NULL);
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	retval = xfs_dir_leaf_lookup_int(bp, args, &index);
 	if (retval == EEXIST) {
 		(void)xfs_dir_leaf_remove(args->trans, bp, index);
@@ -912,7 +912,7 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 			return(error);
 		if (bp)
 			leaf = bp->data;
-		if (bp && INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) != XFS_DIR_LEAF_MAGIC) {
+		if (bp && be16_to_cpu(leaf->hdr.info.magic) != XFS_DIR_LEAF_MAGIC) {
 			xfs_dir_trace_g_dub("node: block not a leaf",
 						   dp, uio, bno);
 			xfs_da_brelse(trans, bp);
@@ -949,17 +949,17 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 			if (bp == NULL)
 				return(XFS_ERROR(EFSCORRUPTED));
 			node = bp->data;
-			if (INT_GET(node->hdr.info.magic, ARCH_CONVERT) != XFS_DA_NODE_MAGIC)
+			if (be16_to_cpu(node->hdr.info.magic) != XFS_DA_NODE_MAGIC)
 				break;
 			btree = &node->btree[0];
 			xfs_dir_trace_g_dun("node: node detail", dp, uio, node);
-			for (i = 0; i < INT_GET(node->hdr.count, ARCH_CONVERT); btree++, i++) {
-				if (INT_GET(btree->hashval, ARCH_CONVERT) >= cookhash) {
-					bno = INT_GET(btree->before, ARCH_CONVERT);
+			for (i = 0; i < be16_to_cpu(node->hdr.count); btree++, i++) {
+				if (be32_to_cpu(btree->hashval) >= cookhash) {
+					bno = be32_to_cpu(btree->before);
 					break;
 				}
 			}
-			if (i == INT_GET(node->hdr.count, ARCH_CONVERT)) {
+			if (i == be16_to_cpu(node->hdr.count)) {
 				xfs_da_brelse(trans, bp);
 				xfs_dir_trace_g_du("node: hash beyond EOF",
 							  dp, uio);
@@ -982,7 +982,7 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 	 */
 	for (;;) {
 		leaf = bp->data;
-		if (unlikely(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) != XFS_DIR_LEAF_MAGIC)) {
+		if (unlikely(be16_to_cpu(leaf->hdr.info.magic) != XFS_DIR_LEAF_MAGIC)) {
 			xfs_dir_trace_g_dul("node: not a leaf", dp, uio, leaf);
 			xfs_da_brelse(trans, bp);
 			XFS_CORRUPTION_ERROR("xfs_dir_node_getdents(1)",
@@ -990,7 +990,7 @@ xfs_dir_node_getdents(xfs_trans_t *trans, xfs_inode_t *dp, uio_t *uio,
 			return XFS_ERROR(EFSCORRUPTED);
 		}
 		xfs_dir_trace_g_dul("node: leaf detail", dp, uio, leaf);
-		if ((nextbno = INT_GET(leaf->hdr.info.forw, ARCH_CONVERT))) {
+		if ((nextbno = be32_to_cpu(leaf->hdr.info.forw))) {
 			nextda = xfs_da_reada_buf(trans, dp, nextbno,
 						  XFS_DATA_FORK);
 		} else
@@ -1118,21 +1118,20 @@ void
 xfs_dir_trace_g_dun(char *where, xfs_inode_t *dp, uio_t *uio,
 			xfs_da_intnode_t *node)
 {
-	int	last = INT_GET(node->hdr.count, ARCH_CONVERT) - 1;
+	int	last = be16_to_cpu(node->hdr.count) - 1;
 
 	xfs_dir_trace_enter(XFS_DIR_KTRACE_G_DUN, where,
 		     (void *)dp, (void *)dp->i_mount,
 		     (void *)((unsigned long)(uio->uio_offset >> 32)),
 		     (void *)((unsigned long)(uio->uio_offset & 0xFFFFFFFF)),
 		     (void *)(unsigned long)uio->uio_resid,
+		     (void *)(unsigned long)be32_to_cpu(node->hdr.info.forw),
 		     (void *)(unsigned long)
-			INT_GET(node->hdr.info.forw, ARCH_CONVERT),
+			be16_to_cpu(node->hdr.count),
 		     (void *)(unsigned long)
-			INT_GET(node->hdr.count, ARCH_CONVERT),
+			be32_to_cpu(node->btree[0].hashval),
 		     (void *)(unsigned long)
-			INT_GET(node->btree[0].hashval, ARCH_CONVERT),
-		     (void *)(unsigned long)
-			INT_GET(node->btree[last].hashval, ARCH_CONVERT),
+			be32_to_cpu(node->btree[last].hashval),
 		     NULL, NULL, NULL);
 }
 
@@ -1150,8 +1149,7 @@ xfs_dir_trace_g_dul(char *where, xfs_inode_t *dp, uio_t *uio,
 		     (void *)((unsigned long)(uio->uio_offset >> 32)),
 		     (void *)((unsigned long)(uio->uio_offset & 0xFFFFFFFF)),
 		     (void *)(unsigned long)uio->uio_resid,
-		     (void *)(unsigned long)
-			INT_GET(leaf->hdr.info.forw, ARCH_CONVERT),
+		     (void *)(unsigned long)be32_to_cpu(leaf->hdr.info.forw),
 		     (void *)(unsigned long)
 			INT_GET(leaf->hdr.count, ARCH_CONVERT),
 		     (void *)(unsigned long)
diff --git a/fs/xfs/xfs_dir2.h b/fs/xfs/xfs_dir2.h
index 3158f5dc431f..7dd364b1e038 100644
--- a/fs/xfs/xfs_dir2.h
+++ b/fs/xfs/xfs_dir2.h
@@ -55,16 +55,16 @@ typedef	__uint32_t	xfs_dir2_db_t;
 /*
  * Byte offset in a directory.
  */
-typedef	xfs_off_t		xfs_dir2_off_t;
+typedef	xfs_off_t	xfs_dir2_off_t;
 
 /*
  * For getdents, argument struct for put routines.
  */
 typedef int (*xfs_dir2_put_t)(struct xfs_dir2_put_args *pa);
 typedef struct xfs_dir2_put_args {
-	xfs_off_t		cook;		/* cookie of (next) entry */
+	xfs_off_t	cook;		/* cookie of (next) entry */
 	xfs_intino_t	ino;		/* inode number */
-	struct xfs_dirent	*dbp;		/* buffer pointer */
+	xfs_dirent_t	*dbp;		/* buffer pointer */
 	char		*name;		/* directory entry name */
 	int		namelen;	/* length of name */
 	int		done;		/* output: set if value was stored */
@@ -75,18 +75,13 @@ typedef struct xfs_dir2_put_args {
 /*
  * Other interfaces used by the rest of the dir v2 code.
  */
-extern int
-	xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
-			    xfs_dir2_db_t *dbp);
-
-extern int
-	xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *vp);
-
-extern int
-	xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *vp);
-
-extern int
-	xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
-			      struct xfs_dabuf *bp);
+extern int xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
+				xfs_dir2_db_t *dbp);
+extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp,
+				int *vp);
+extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp,
+				int *vp);
+extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
+				struct xfs_dabuf *bp);
 
 #endif	/* __XFS_DIR2_H__ */
diff --git a/fs/xfs/xfs_dir2_block.c b/fs/xfs/xfs_dir2_block.c
index 31bc99faa704..972ded595476 100644
--- a/fs/xfs/xfs_dir2_block.c
+++ b/fs/xfs/xfs_dir2_block.c
@@ -81,7 +81,7 @@ xfs_dir2_block_addname(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			needlog;	/* need to log header */
 	int			needscan;	/* need to rescan freespace */
-	xfs_dir2_data_off_t	*tagp;		/* pointer to tag value */
+	__be16			*tagp;		/* pointer to tag value */
 	xfs_trans_t		*tp;		/* transaction structure */
 
 	xfs_dir2_trace_args("block_addname", args);
@@ -100,8 +100,7 @@ xfs_dir2_block_addname(
 	/*
 	 * Check the magic number, corrupted if wrong.
 	 */
-	if (unlikely(INT_GET(block->hdr.magic, ARCH_CONVERT)
-						!= XFS_DIR2_BLOCK_MAGIC)) {
+	if (unlikely(be32_to_cpu(block->hdr.magic) != XFS_DIR2_BLOCK_MAGIC)) {
 		XFS_CORRUPTION_ERROR("xfs_dir2_block_addname",
 				     XFS_ERRLEVEL_LOW, mp, block);
 		xfs_da_brelse(tp, bp);
@@ -121,38 +120,38 @@ xfs_dir2_block_addname(
 		/*
 		 * Tag just before the first leaf entry.
 		 */
-		tagp = (xfs_dir2_data_off_t *)blp - 1;
+		tagp = (__be16 *)blp - 1;
 		/*
 		 * Data object just before the first leaf entry.
 		 */
-		enddup = (xfs_dir2_data_unused_t *)((char *)block + INT_GET(*tagp, ARCH_CONVERT));
+		enddup = (xfs_dir2_data_unused_t *)((char *)block + be16_to_cpu(*tagp));
 		/*
 		 * If it's not free then can't do this add without cleaning up:
 		 * the space before the first leaf entry needs to be free so it
 		 * can be expanded to hold the pointer to the new entry.
 		 */
-		if (INT_GET(enddup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG)
+		if (be16_to_cpu(enddup->freetag) != XFS_DIR2_DATA_FREE_TAG)
 			dup = enddup = NULL;
 		/*
 		 * Check out the biggest freespace and see if it's the same one.
 		 */
 		else {
 			dup = (xfs_dir2_data_unused_t *)
-			      ((char *)block + INT_GET(bf[0].offset, ARCH_CONVERT));
+			      ((char *)block + be16_to_cpu(bf[0].offset));
 			if (dup == enddup) {
 				/*
 				 * It is the biggest freespace, is it too small
 				 * to hold the new leaf too?
 				 */
-				if (INT_GET(dup->length, ARCH_CONVERT) < len + (uint)sizeof(*blp)) {
+				if (be16_to_cpu(dup->length) < len + (uint)sizeof(*blp)) {
 					/*
 					 * Yes, we use the second-largest
 					 * entry instead if it works.
 					 */
-					if (INT_GET(bf[1].length, ARCH_CONVERT) >= len)
+					if (be16_to_cpu(bf[1].length) >= len)
 						dup = (xfs_dir2_data_unused_t *)
 						      ((char *)block +
-						       INT_GET(bf[1].offset, ARCH_CONVERT));
+						       be16_to_cpu(bf[1].offset));
 					else
 						dup = NULL;
 				}
@@ -161,7 +160,7 @@ xfs_dir2_block_addname(
 				 * Not the same free entry,
 				 * just check its length.
 				 */
-				if (INT_GET(dup->length, ARCH_CONVERT) < len) {
+				if (be16_to_cpu(dup->length) < len) {
 					dup = NULL;
 				}
 			}
@@ -172,9 +171,9 @@ xfs_dir2_block_addname(
 	 * If there are stale entries we'll use one for the leaf.
 	 * Is the biggest entry enough to avoid compaction?
 	 */
-	else if (INT_GET(bf[0].length, ARCH_CONVERT) >= len) {
+	else if (be16_to_cpu(bf[0].length) >= len) {
 		dup = (xfs_dir2_data_unused_t *)
-		      ((char *)block + INT_GET(bf[0].offset, ARCH_CONVERT));
+		      ((char *)block + be16_to_cpu(bf[0].offset));
 		compact = 0;
 	}
 	/*
@@ -184,20 +183,20 @@ xfs_dir2_block_addname(
 		/*
 		 * Tag just before the first leaf entry.
 		 */
-		tagp = (xfs_dir2_data_off_t *)blp - 1;
+		tagp = (__be16 *)blp - 1;
 		/*
 		 * Data object just before the first leaf entry.
 		 */
-		dup = (xfs_dir2_data_unused_t *)((char *)block + INT_GET(*tagp, ARCH_CONVERT));
+		dup = (xfs_dir2_data_unused_t *)((char *)block + be16_to_cpu(*tagp));
 		/*
 		 * If it's not free then the data will go where the
 		 * leaf data starts now, if it works at all.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
-			if (INT_GET(dup->length, ARCH_CONVERT) + (INT_GET(btp->stale, ARCH_CONVERT) - 1) *
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			if (be16_to_cpu(dup->length) + (be32_to_cpu(btp->stale) - 1) *
 			    (uint)sizeof(*blp) < len)
 				dup = NULL;
-		} else if ((INT_GET(btp->stale, ARCH_CONVERT) - 1) * (uint)sizeof(*blp) < len)
+		} else if ((be32_to_cpu(btp->stale) - 1) * (uint)sizeof(*blp) < len)
 			dup = NULL;
 		else
 			dup = (xfs_dir2_data_unused_t *)blp;
@@ -243,11 +242,11 @@ xfs_dir2_block_addname(
 		int	fromidx;		/* source leaf index */
 		int	toidx;			/* target leaf index */
 
-		for (fromidx = toidx = INT_GET(btp->count, ARCH_CONVERT) - 1,
+		for (fromidx = toidx = be32_to_cpu(btp->count) - 1,
 			highstale = lfloghigh = -1;
 		     fromidx >= 0;
 		     fromidx--) {
-			if (INT_GET(blp[fromidx].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR) {
+			if (be32_to_cpu(blp[fromidx].address) == XFS_DIR2_NULL_DATAPTR) {
 				if (highstale == -1)
 					highstale = toidx;
 				else {
@@ -260,15 +259,15 @@ xfs_dir2_block_addname(
 				blp[toidx] = blp[fromidx];
 			toidx--;
 		}
-		lfloglow = toidx + 1 - (INT_GET(btp->stale, ARCH_CONVERT) - 1);
-		lfloghigh -= INT_GET(btp->stale, ARCH_CONVERT) - 1;
-		INT_MOD(btp->count, ARCH_CONVERT, -(INT_GET(btp->stale, ARCH_CONVERT) - 1));
+		lfloglow = toidx + 1 - (be32_to_cpu(btp->stale) - 1);
+		lfloghigh -= be32_to_cpu(btp->stale) - 1;
+		be32_add(&btp->count, -(be32_to_cpu(btp->stale) - 1));
 		xfs_dir2_data_make_free(tp, bp,
 			(xfs_dir2_data_aoff_t)((char *)blp - (char *)block),
-			(xfs_dir2_data_aoff_t)((INT_GET(btp->stale, ARCH_CONVERT) - 1) * sizeof(*blp)),
+			(xfs_dir2_data_aoff_t)((be32_to_cpu(btp->stale) - 1) * sizeof(*blp)),
 			&needlog, &needscan);
-		blp += INT_GET(btp->stale, ARCH_CONVERT) - 1;
-		INT_SET(btp->stale, ARCH_CONVERT, 1);
+		blp += be32_to_cpu(btp->stale) - 1;
+		btp->stale = cpu_to_be32(1);
 		/*
 		 * If we now need to rebuild the bestfree map, do so.
 		 * This needs to happen before the next call to use_free.
@@ -283,23 +282,23 @@ xfs_dir2_block_addname(
 	 * Set leaf logging boundaries to impossible state.
 	 * For the no-stale case they're set explicitly.
 	 */
-	else if (INT_GET(btp->stale, ARCH_CONVERT)) {
-		lfloglow = INT_GET(btp->count, ARCH_CONVERT);
+	else if (btp->stale) {
+		lfloglow = be32_to_cpu(btp->count);
 		lfloghigh = -1;
 	}
 	/*
 	 * Find the slot that's first lower than our hash value, -1 if none.
 	 */
-	for (low = 0, high = INT_GET(btp->count, ARCH_CONVERT) - 1; low <= high; ) {
+	for (low = 0, high = be32_to_cpu(btp->count) - 1; low <= high; ) {
 		mid = (low + high) >> 1;
-		if ((hash = INT_GET(blp[mid].hashval, ARCH_CONVERT)) == args->hashval)
+		if ((hash = be32_to_cpu(blp[mid].hashval)) == args->hashval)
 			break;
 		if (hash < args->hashval)
 			low = mid + 1;
 		else
 			high = mid - 1;
 	}
-	while (mid >= 0 && INT_GET(blp[mid].hashval, ARCH_CONVERT) >= args->hashval) {
+	while (mid >= 0 && be32_to_cpu(blp[mid].hashval) >= args->hashval) {
 		mid--;
 	}
 	/*
@@ -311,14 +310,14 @@ xfs_dir2_block_addname(
 		 */
 		xfs_dir2_data_use_free(tp, bp, enddup,
 			(xfs_dir2_data_aoff_t)
-			((char *)enddup - (char *)block + INT_GET(enddup->length, ARCH_CONVERT) -
+			((char *)enddup - (char *)block + be16_to_cpu(enddup->length) -
 			 sizeof(*blp)),
 			(xfs_dir2_data_aoff_t)sizeof(*blp),
 			&needlog, &needscan);
 		/*
 		 * Update the tail (entry count).
 		 */
-		INT_MOD(btp->count, ARCH_CONVERT, +1);
+		be32_add(&btp->count, 1);
 		/*
 		 * If we now need to rebuild the bestfree map, do so.
 		 * This needs to happen before the next call to use_free.
@@ -346,12 +345,12 @@ xfs_dir2_block_addname(
 	else {
 		for (lowstale = mid;
 		     lowstale >= 0 &&
-			INT_GET(blp[lowstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR;
+			be32_to_cpu(blp[lowstale].address) != XFS_DIR2_NULL_DATAPTR;
 		     lowstale--)
 			continue;
 		for (highstale = mid + 1;
-		     highstale < INT_GET(btp->count, ARCH_CONVERT) &&
-			INT_GET(blp[highstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR &&
+		     highstale < be32_to_cpu(btp->count) &&
+			be32_to_cpu(blp[highstale].address) != XFS_DIR2_NULL_DATAPTR &&
 			(lowstale < 0 || mid - lowstale > highstale - mid);
 		     highstale++)
 			continue;
@@ -359,7 +358,7 @@ xfs_dir2_block_addname(
 		 * Move entries toward the low-numbered stale entry.
 		 */
 		if (lowstale >= 0 &&
-		    (highstale == INT_GET(btp->count, ARCH_CONVERT) ||
+		    (highstale == be32_to_cpu(btp->count) ||
 		     mid - lowstale <= highstale - mid)) {
 			if (mid - lowstale)
 				memmove(&blp[lowstale], &blp[lowstale + 1],
@@ -371,7 +370,7 @@ xfs_dir2_block_addname(
 		 * Move entries toward the high-numbered stale entry.
 		 */
 		else {
-			ASSERT(highstale < INT_GET(btp->count, ARCH_CONVERT));
+			ASSERT(highstale < be32_to_cpu(btp->count));
 			mid++;
 			if (highstale - mid)
 				memmove(&blp[mid + 1], &blp[mid],
@@ -379,7 +378,7 @@ xfs_dir2_block_addname(
 			lfloglow = MIN(mid, lfloglow);
 			lfloghigh = MAX(highstale, lfloghigh);
 		}
-		INT_MOD(btp->stale, ARCH_CONVERT, -1);
+		be32_add(&btp->stale, -1);
 	}
 	/*
 	 * Point to the new data entry.
@@ -388,8 +387,9 @@ xfs_dir2_block_addname(
 	/*
 	 * Fill in the leaf entry.
 	 */
-	INT_SET(blp[mid].hashval, ARCH_CONVERT, args->hashval);
-	INT_SET(blp[mid].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp, (char *)dep - (char *)block));
+	blp[mid].hashval = cpu_to_be32(args->hashval);
+	blp[mid].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
+				(char *)dep - (char *)block));
 	xfs_dir2_block_log_leaf(tp, bp, lfloglow, lfloghigh);
 	/*
 	 * Mark space for the data entry used.
@@ -404,7 +404,7 @@ xfs_dir2_block_addname(
 	dep->namelen = args->namelen;
 	memcpy(dep->name, args->name, args->namelen);
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+	*tagp = cpu_to_be16((char *)dep - (char *)block);
 	/*
 	 * Clean up the bestfree array and log the header, tail, and entry.
 	 */
@@ -485,8 +485,8 @@ xfs_dir2_block_getdents(
 		/*
 		 * Unused, skip it.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
-			ptr += INT_GET(dup->length, ARCH_CONVERT);
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			ptr += be16_to_cpu(dup->length);
 			continue;
 		}
 
@@ -533,7 +533,7 @@ xfs_dir2_block_getdents(
 
 	/*
 	 * Reached the end of the block.
-	 * Set the offset to a nonexistent block 1 and return.
+	 * Set the offset to a non-existent block 1 and return.
 	 */
 	*eofp = 1;
 
@@ -622,7 +622,7 @@ xfs_dir2_block_lookup(
 	 * Get the offset from the leaf entry, to point to the data.
 	 */
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(blp[ent].address, ARCH_CONVERT)));
+	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(blp[ent].address)));
 	/*
 	 * Fill in inode number, release the block.
 	 */
@@ -674,10 +674,10 @@ xfs_dir2_block_lookup_int(
 	 * Loop doing a binary search for our hash value.
 	 * Find our entry, ENOENT if it's not there.
 	 */
-	for (low = 0, high = INT_GET(btp->count, ARCH_CONVERT) - 1; ; ) {
+	for (low = 0, high = be32_to_cpu(btp->count) - 1; ; ) {
 		ASSERT(low <= high);
 		mid = (low + high) >> 1;
-		if ((hash = INT_GET(blp[mid].hashval, ARCH_CONVERT)) == args->hashval)
+		if ((hash = be32_to_cpu(blp[mid].hashval)) == args->hashval)
 			break;
 		if (hash < args->hashval)
 			low = mid + 1;
@@ -692,7 +692,7 @@ xfs_dir2_block_lookup_int(
 	/*
 	 * Back up to the first one with the right hash value.
 	 */
-	while (mid > 0 && INT_GET(blp[mid - 1].hashval, ARCH_CONVERT) == args->hashval) {
+	while (mid > 0 && be32_to_cpu(blp[mid - 1].hashval) == args->hashval) {
 		mid--;
 	}
 	/*
@@ -700,7 +700,7 @@ xfs_dir2_block_lookup_int(
 	 * right hash value looking for our name.
 	 */
 	do {
-		if ((addr = INT_GET(blp[mid].address, ARCH_CONVERT)) == XFS_DIR2_NULL_DATAPTR)
+		if ((addr = be32_to_cpu(blp[mid].address)) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Get pointer to the entry from the leaf.
@@ -717,7 +717,7 @@ xfs_dir2_block_lookup_int(
 			*entno = mid;
 			return 0;
 		}
-	} while (++mid < INT_GET(btp->count, ARCH_CONVERT) && INT_GET(blp[mid].hashval, ARCH_CONVERT) == hash);
+	} while (++mid < be32_to_cpu(btp->count) && be32_to_cpu(blp[mid].hashval) == hash);
 	/*
 	 * No match, release the buffer and return ENOENT.
 	 */
@@ -767,7 +767,7 @@ xfs_dir2_block_removename(
 	 * Point to the data entry using the leaf entry.
 	 */
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(blp[ent].address, ARCH_CONVERT)));
+	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(blp[ent].address)));
 	/*
 	 * Mark the data entry's space free.
 	 */
@@ -778,12 +778,12 @@ xfs_dir2_block_removename(
 	/*
 	 * Fix up the block tail.
 	 */
-	INT_MOD(btp->stale, ARCH_CONVERT, +1);
+	be32_add(&btp->stale, 1);
 	xfs_dir2_block_log_tail(tp, bp);
 	/*
 	 * Remove the leaf entry by marking it stale.
 	 */
-	INT_SET(blp[ent].address, ARCH_CONVERT, XFS_DIR2_NULL_DATAPTR);
+	blp[ent].address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
 	xfs_dir2_block_log_leaf(tp, bp, ent, ent);
 	/*
 	 * Fix up bestfree, log the header if necessary.
@@ -843,7 +843,7 @@ xfs_dir2_block_replace(
 	 * Point to the data entry we need to change.
 	 */
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(blp[ent].address, ARCH_CONVERT)));
+	      ((char *)block + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(blp[ent].address)));
 	ASSERT(INT_GET(dep->inumber, ARCH_CONVERT) != args->inumber);
 	/*
 	 * Change the inode number to the new value.
@@ -868,8 +868,8 @@ xfs_dir2_block_sort(
 
 	la = a;
 	lb = b;
-	return INT_GET(la->hashval, ARCH_CONVERT) < INT_GET(lb->hashval, ARCH_CONVERT) ? -1 :
-		(INT_GET(la->hashval, ARCH_CONVERT) > INT_GET(lb->hashval, ARCH_CONVERT) ? 1 : 0);
+	return be32_to_cpu(la->hashval) < be32_to_cpu(lb->hashval) ? -1 :
+		(be32_to_cpu(la->hashval) > be32_to_cpu(lb->hashval) ? 1 : 0);
 }
 
 /*
@@ -881,7 +881,7 @@ xfs_dir2_leaf_to_block(
 	xfs_dabuf_t		*lbp,		/* leaf buffer */
 	xfs_dabuf_t		*dbp)		/* data buffer */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf bests table */
+	__be16			*bestsp;	/* leaf bests table */
 	xfs_dir2_block_t	*block;		/* block structure */
 	xfs_dir2_block_tail_t	*btp;		/* block tail */
 	xfs_inode_t		*dp;		/* incore directory inode */
@@ -896,7 +896,7 @@ xfs_dir2_leaf_to_block(
 	int			needscan;	/* need to scan for bestfree */
 	xfs_dir2_sf_hdr_t	sfh;		/* shortform header */
 	int			size;		/* bytes used */
-	xfs_dir2_data_off_t	*tagp;		/* end of entry (tag) */
+	__be16			*tagp;		/* end of entry (tag) */
 	int			to;		/* block/leaf to index */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
@@ -905,7 +905,7 @@ xfs_dir2_leaf_to_block(
 	tp = args->trans;
 	mp = dp->i_mount;
 	leaf = lbp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	/*
 	 * If there are data blocks other than the first one, take this
@@ -915,11 +915,11 @@ xfs_dir2_leaf_to_block(
 	 */
 	while (dp->i_d.di_size > mp->m_dirblksize) {
 		bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-		if (INT_GET(bestsp[INT_GET(ltp->bestcount, ARCH_CONVERT) - 1], ARCH_CONVERT) ==
+		if (be16_to_cpu(bestsp[be32_to_cpu(ltp->bestcount) - 1]) ==
 		    mp->m_dirblksize - (uint)sizeof(block->hdr)) {
 			if ((error =
 			    xfs_dir2_leaf_trim_data(args, lbp,
-				    (xfs_dir2_db_t)(INT_GET(ltp->bestcount, ARCH_CONVERT) - 1))))
+				    (xfs_dir2_db_t)(be32_to_cpu(ltp->bestcount) - 1))))
 				goto out;
 		} else {
 			error = 0;
@@ -935,28 +935,29 @@ xfs_dir2_leaf_to_block(
 		goto out;
 	}
 	block = dbp->data;
-	ASSERT(INT_GET(block->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC);
+	ASSERT(be32_to_cpu(block->hdr.magic) == XFS_DIR2_DATA_MAGIC);
 	/*
 	 * Size of the "leaf" area in the block.
 	 */
 	size = (uint)sizeof(block->tail) +
-	       (uint)sizeof(*lep) * (INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT));
+	       (uint)sizeof(*lep) * (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale));
 	/*
 	 * Look at the last data entry.
 	 */
-	tagp = (xfs_dir2_data_off_t *)((char *)block + mp->m_dirblksize) - 1;
-	dup = (xfs_dir2_data_unused_t *)((char *)block + INT_GET(*tagp, ARCH_CONVERT));
+	tagp = (__be16 *)((char *)block + mp->m_dirblksize) - 1;
+	dup = (xfs_dir2_data_unused_t *)((char *)block + be16_to_cpu(*tagp));
 	/*
 	 * If it's not free or is too short we can't do it.
 	 */
-	if (INT_GET(dup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG || INT_GET(dup->length, ARCH_CONVERT) < size) {
+	if (be16_to_cpu(dup->freetag) != XFS_DIR2_DATA_FREE_TAG ||
+	    be16_to_cpu(dup->length) < size) {
 		error = 0;
 		goto out;
 	}
 	/*
 	 * Start converting it to block form.
 	 */
-	INT_SET(block->hdr.magic, ARCH_CONVERT, XFS_DIR2_BLOCK_MAGIC);
+	block->hdr.magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
 	needlog = 1;
 	needscan = 0;
 	/*
@@ -968,20 +969,20 @@ xfs_dir2_leaf_to_block(
 	 * Initialize the block tail.
 	 */
 	btp = XFS_DIR2_BLOCK_TAIL_P(mp, block);
-	INT_SET(btp->count, ARCH_CONVERT, INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT));
+	btp->count = cpu_to_be32(be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale));
 	btp->stale = 0;
 	xfs_dir2_block_log_tail(tp, dbp);
 	/*
 	 * Initialize the block leaf area.  We compact out stale entries.
 	 */
 	lep = XFS_DIR2_BLOCK_LEAF_P(btp);
-	for (from = to = 0; from < INT_GET(leaf->hdr.count, ARCH_CONVERT); from++) {
-		if (INT_GET(leaf->ents[from].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	for (from = to = 0; from < be16_to_cpu(leaf->hdr.count); from++) {
+		if (be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		lep[to++] = leaf->ents[from];
 	}
-	ASSERT(to == INT_GET(btp->count, ARCH_CONVERT));
-	xfs_dir2_block_log_leaf(tp, dbp, 0, INT_GET(btp->count, ARCH_CONVERT) - 1);
+	ASSERT(to == be32_to_cpu(btp->count));
+	xfs_dir2_block_log_leaf(tp, dbp, 0, be32_to_cpu(btp->count) - 1);
 	/*
 	 * Scan the bestfree if we need it and log the data block header.
 	 */
@@ -1043,7 +1044,7 @@ xfs_dir2_sf_to_block(
 	int			offset;		/* target block offset */
 	xfs_dir2_sf_entry_t	*sfep;		/* sf entry pointer */
 	xfs_dir2_sf_t		*sfp;		/* shortform structure */
-	xfs_dir2_data_off_t	*tagp;		/* end of data entry */
+	__be16			*tagp;		/* end of data entry */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
 	xfs_dir2_trace_args("sf_to_block", args);
@@ -1095,12 +1096,12 @@ xfs_dir2_sf_to_block(
 		return error;
 	}
 	block = bp->data;
-	INT_SET(block->hdr.magic, ARCH_CONVERT, XFS_DIR2_BLOCK_MAGIC);
+	block->hdr.magic = cpu_to_be32(XFS_DIR2_BLOCK_MAGIC);
 	/*
 	 * Compute size of block "tail" area.
 	 */
 	i = (uint)sizeof(*btp) +
-	    (INT_GET(sfp->hdr.count, ARCH_CONVERT) + 2) * (uint)sizeof(xfs_dir2_leaf_entry_t);
+	    (sfp->hdr.count + 2) * (uint)sizeof(xfs_dir2_leaf_entry_t);
 	/*
 	 * The whole thing is initialized to free by the init routine.
 	 * Say we're using the leaf and tail area.
@@ -1114,7 +1115,7 @@ xfs_dir2_sf_to_block(
 	 * Fill in the tail.
 	 */
 	btp = XFS_DIR2_BLOCK_TAIL_P(mp, block);
-	INT_SET(btp->count, ARCH_CONVERT, INT_GET(sfp->hdr.count, ARCH_CONVERT) + 2);	/* ., .. */
+	btp->count = cpu_to_be32(sfp->hdr.count + 2);	/* ., .. */
 	btp->stale = 0;
 	blp = XFS_DIR2_BLOCK_LEAF_P(btp);
 	endoffset = (uint)((char *)blp - (char *)block);
@@ -1123,7 +1124,7 @@ xfs_dir2_sf_to_block(
 	 */
 	xfs_dir2_data_use_free(tp, bp, dup,
 		(xfs_dir2_data_aoff_t)((char *)dup - (char *)block),
-		INT_GET(dup->length, ARCH_CONVERT), &needlog, &needscan);
+		be16_to_cpu(dup->length), &needlog, &needscan);
 	/*
 	 * Create entry for .
 	 */
@@ -1133,10 +1134,11 @@ xfs_dir2_sf_to_block(
 	dep->namelen = 1;
 	dep->name[0] = '.';
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+	*tagp = cpu_to_be16((char *)dep - (char *)block);
 	xfs_dir2_data_log_entry(tp, bp, dep);
-	INT_SET(blp[0].hashval, ARCH_CONVERT, xfs_dir_hash_dot);
-	INT_SET(blp[0].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp, (char *)dep - (char *)block));
+	blp[0].hashval = cpu_to_be32(xfs_dir_hash_dot);
+	blp[0].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
+				(char *)dep - (char *)block));
 	/*
 	 * Create entry for ..
 	 */
@@ -1146,15 +1148,16 @@ xfs_dir2_sf_to_block(
 	dep->namelen = 2;
 	dep->name[0] = dep->name[1] = '.';
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+	*tagp = cpu_to_be16((char *)dep - (char *)block);
 	xfs_dir2_data_log_entry(tp, bp, dep);
-	INT_SET(blp[1].hashval, ARCH_CONVERT, xfs_dir_hash_dotdot);
-	INT_SET(blp[1].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp, (char *)dep - (char *)block));
+	blp[1].hashval = cpu_to_be32(xfs_dir_hash_dotdot);
+	blp[1].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
+				(char *)dep - (char *)block));
 	offset = XFS_DIR2_DATA_FIRST_OFFSET;
 	/*
 	 * Loop over existing entries, stuff them in.
 	 */
-	if ((i = 0) == INT_GET(sfp->hdr.count, ARCH_CONVERT))
+	if ((i = 0) == sfp->hdr.count)
 		sfep = NULL;
 	else
 		sfep = XFS_DIR2_SF_FIRSTENTRY(sfp);
@@ -1176,15 +1179,14 @@ xfs_dir2_sf_to_block(
 		if (offset < newoffset) {
 			dup = (xfs_dir2_data_unused_t *)
 			      ((char *)block + offset);
-			INT_SET(dup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-			INT_SET(dup->length, ARCH_CONVERT, newoffset - offset);
-			INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT,
-				(xfs_dir2_data_off_t)
+			dup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+			dup->length = cpu_to_be16(newoffset - offset);
+			*XFS_DIR2_DATA_UNUSED_TAG_P(dup) = cpu_to_be16(
 				((char *)dup - (char *)block));
 			xfs_dir2_data_log_unused(tp, bp, dup);
 			(void)xfs_dir2_data_freeinsert((xfs_dir2_data_t *)block,
 				dup, &dummy);
-			offset += INT_GET(dup->length, ARCH_CONVERT);
+			offset += be16_to_cpu(dup->length);
 			continue;
 		}
 		/*
@@ -1196,13 +1198,14 @@ xfs_dir2_sf_to_block(
 		dep->namelen = sfep->namelen;
 		memcpy(dep->name, sfep->name, dep->namelen);
 		tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-		INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)block));
+		*tagp = cpu_to_be16((char *)dep - (char *)block);
 		xfs_dir2_data_log_entry(tp, bp, dep);
-		INT_SET(blp[2 + i].hashval, ARCH_CONVERT, xfs_da_hashname((char *)sfep->name, sfep->namelen));
-		INT_SET(blp[2 + i].address, ARCH_CONVERT, XFS_DIR2_BYTE_TO_DATAPTR(mp,
+		blp[2 + i].hashval = cpu_to_be32(xfs_da_hashname(
+					(char *)sfep->name, sfep->namelen));
+		blp[2 + i].address = cpu_to_be32(XFS_DIR2_BYTE_TO_DATAPTR(mp,
 						 (char *)dep - (char *)block));
 		offset = (int)((char *)(tagp + 1) - (char *)block);
-		if (++i == INT_GET(sfp->hdr.count, ARCH_CONVERT))
+		if (++i == sfp->hdr.count)
 			sfep = NULL;
 		else
 			sfep = XFS_DIR2_SF_NEXTENTRY(sfp, sfep);
@@ -1212,13 +1215,13 @@ xfs_dir2_sf_to_block(
 	/*
 	 * Sort the leaf entries by hash value.
 	 */
-	xfs_sort(blp, INT_GET(btp->count, ARCH_CONVERT), sizeof(*blp), xfs_dir2_block_sort);
+	xfs_sort(blp, be32_to_cpu(btp->count), sizeof(*blp), xfs_dir2_block_sort);
 	/*
 	 * Log the leaf entry area and tail.
 	 * Already logged the header in data_init, ignore needlog.
 	 */
 	ASSERT(needscan == 0);
-	xfs_dir2_block_log_leaf(tp, bp, 0, INT_GET(btp->count, ARCH_CONVERT) - 1);
+	xfs_dir2_block_log_leaf(tp, bp, 0, be32_to_cpu(btp->count) - 1);
 	xfs_dir2_block_log_tail(tp, bp);
 	xfs_dir2_data_check(dp, bp);
 	xfs_da_buf_done(bp);
diff --git a/fs/xfs/xfs_dir2_block.h b/fs/xfs/xfs_dir2_block.h
index a2e5cb98a838..6722effd0b20 100644
--- a/fs/xfs/xfs_dir2_block.h
+++ b/fs/xfs/xfs_dir2_block.h
@@ -43,8 +43,8 @@ struct xfs_trans;
 #define	XFS_DIR2_BLOCK_MAGIC	0x58443242	/* XD2B: for one block dirs */
 
 typedef struct xfs_dir2_block_tail {
-	__uint32_t	count;			/* count of leaf entries */
-	__uint32_t	stale;			/* count of stale lf entries */
+	__be32		count;			/* count of leaf entries */
+	__be32		stale;			/* count of stale lf entries */
 } xfs_dir2_block_tail_t;
 
 /*
@@ -75,8 +75,7 @@ xfs_dir2_block_tail_p(struct xfs_mount *mp, xfs_dir2_block_t *block)
 static inline struct xfs_dir2_leaf_entry *
 xfs_dir2_block_leaf_p(xfs_dir2_block_tail_t *btp)
 {
-	return (((struct xfs_dir2_leaf_entry *)
-		(btp)) - INT_GET((btp)->count, ARCH_CONVERT));
+	return ((struct xfs_dir2_leaf_entry *)btp) - be32_to_cpu(btp->count);
 }
 
 /*
diff --git a/fs/xfs/xfs_dir2_data.c b/fs/xfs/xfs_dir2_data.c
index 5b7c47e2f14a..bb3d03ff002b 100644
--- a/fs/xfs/xfs_dir2_data.c
+++ b/fs/xfs/xfs_dir2_data.c
@@ -70,11 +70,11 @@ xfs_dir2_data_check(
 
 	mp = dp->i_mount;
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	bf = d->hdr.bestfree;
 	p = (char *)d->u;
-	if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
+	if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
 		btp = XFS_DIR2_BLOCK_TAIL_P(mp, (xfs_dir2_block_t *)d);
 		lep = XFS_DIR2_BLOCK_LEAF_P(btp);
 		endp = (char *)lep;
@@ -96,8 +96,8 @@ xfs_dir2_data_check(
 		ASSERT(!bf[2].offset);
 		freeseen |= 1 << 2;
 	}
-	ASSERT(INT_GET(bf[0].length, ARCH_CONVERT) >= INT_GET(bf[1].length, ARCH_CONVERT));
-	ASSERT(INT_GET(bf[1].length, ARCH_CONVERT) >= INT_GET(bf[2].length, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(bf[0].length) >= be16_to_cpu(bf[1].length));
+	ASSERT(be16_to_cpu(bf[1].length) >= be16_to_cpu(bf[2].length));
 	/*
 	 * Loop over the data/unused entries.
 	 */
@@ -108,18 +108,20 @@ xfs_dir2_data_check(
 		 * If we find it, account for that, else make sure it
 		 * doesn't need to be there.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
 			ASSERT(lastfree == 0);
-			ASSERT(INT_GET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT) ==
+			ASSERT(be16_to_cpu(*XFS_DIR2_DATA_UNUSED_TAG_P(dup)) ==
 			       (char *)dup - (char *)d);
 			dfp = xfs_dir2_data_freefind(d, dup);
 			if (dfp) {
 				i = (int)(dfp - bf);
 				ASSERT((freeseen & (1 << i)) == 0);
 				freeseen |= 1 << i;
-			} else
-				ASSERT(INT_GET(dup->length, ARCH_CONVERT) <= INT_GET(bf[2].length, ARCH_CONVERT));
-			p += INT_GET(dup->length, ARCH_CONVERT);
+			} else {
+				ASSERT(be16_to_cpu(dup->length) <=
+				       be16_to_cpu(bf[2].length));
+			}
+			p += be16_to_cpu(dup->length);
 			lastfree = 1;
 			continue;
 		}
@@ -132,21 +134,21 @@ xfs_dir2_data_check(
 		dep = (xfs_dir2_data_entry_t *)p;
 		ASSERT(dep->namelen != 0);
 		ASSERT(xfs_dir_ino_validate(mp, INT_GET(dep->inumber, ARCH_CONVERT)) == 0);
-		ASSERT(INT_GET(*XFS_DIR2_DATA_ENTRY_TAG_P(dep), ARCH_CONVERT) ==
+		ASSERT(be16_to_cpu(*XFS_DIR2_DATA_ENTRY_TAG_P(dep)) ==
 		       (char *)dep - (char *)d);
 		count++;
 		lastfree = 0;
-		if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
+		if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
 			addr = XFS_DIR2_DB_OFF_TO_DATAPTR(mp, mp->m_dirdatablk,
 				(xfs_dir2_data_aoff_t)
 				((char *)dep - (char *)d));
 			hash = xfs_da_hashname((char *)dep->name, dep->namelen);
-			for (i = 0; i < INT_GET(btp->count, ARCH_CONVERT); i++) {
-				if (INT_GET(lep[i].address, ARCH_CONVERT) == addr &&
-				    INT_GET(lep[i].hashval, ARCH_CONVERT) == hash)
+			for (i = 0; i < be32_to_cpu(btp->count); i++) {
+				if (be32_to_cpu(lep[i].address) == addr &&
+				    be32_to_cpu(lep[i].hashval) == hash)
 					break;
 			}
-			ASSERT(i < INT_GET(btp->count, ARCH_CONVERT));
+			ASSERT(i < be32_to_cpu(btp->count));
 		}
 		p += XFS_DIR2_DATA_ENTSIZE(dep->namelen);
 	}
@@ -154,15 +156,15 @@ xfs_dir2_data_check(
 	 * Need to have seen all the entries and all the bestfree slots.
 	 */
 	ASSERT(freeseen == 7);
-	if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
-		for (i = stale = 0; i < INT_GET(btp->count, ARCH_CONVERT); i++) {
-			if (INT_GET(lep[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
+		for (i = stale = 0; i < be32_to_cpu(btp->count); i++) {
+			if (be32_to_cpu(lep[i].address) == XFS_DIR2_NULL_DATAPTR)
 				stale++;
 			if (i > 0)
-				ASSERT(INT_GET(lep[i].hashval, ARCH_CONVERT) >= INT_GET(lep[i - 1].hashval, ARCH_CONVERT));
+				ASSERT(be32_to_cpu(lep[i].hashval) >= be32_to_cpu(lep[i - 1].hashval));
 		}
-		ASSERT(count == INT_GET(btp->count, ARCH_CONVERT) - INT_GET(btp->stale, ARCH_CONVERT));
-		ASSERT(stale == INT_GET(btp->stale, ARCH_CONVERT));
+		ASSERT(count == be32_to_cpu(btp->count) - be32_to_cpu(btp->stale));
+		ASSERT(stale == be32_to_cpu(btp->stale));
 	}
 }
 #endif
@@ -190,8 +192,8 @@ xfs_dir2_data_freefind(
 	 * Check order, non-overlapping entries, and if we find the
 	 * one we're looking for it has to be exact.
 	 */
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	for (dfp = &d->hdr.bestfree[0], seenzero = matched = 0;
 	     dfp < &d->hdr.bestfree[XFS_DIR2_DATA_FD_COUNT];
 	     dfp++) {
@@ -201,23 +203,24 @@ xfs_dir2_data_freefind(
 			continue;
 		}
 		ASSERT(seenzero == 0);
-		if (INT_GET(dfp->offset, ARCH_CONVERT) == off) {
+		if (be16_to_cpu(dfp->offset) == off) {
 			matched = 1;
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(dup->length, ARCH_CONVERT));
-		} else if (off < INT_GET(dfp->offset, ARCH_CONVERT))
-			ASSERT(off + INT_GET(dup->length, ARCH_CONVERT) <= INT_GET(dfp->offset, ARCH_CONVERT));
+			ASSERT(dfp->length == dup->length);
+		} else if (off < be16_to_cpu(dfp->offset))
+			ASSERT(off + be16_to_cpu(dup->length) <= be16_to_cpu(dfp->offset));
 		else
-			ASSERT(INT_GET(dfp->offset, ARCH_CONVERT) + INT_GET(dfp->length, ARCH_CONVERT) <= off);
-		ASSERT(matched || INT_GET(dfp->length, ARCH_CONVERT) >= INT_GET(dup->length, ARCH_CONVERT));
+			ASSERT(be16_to_cpu(dfp->offset) + be16_to_cpu(dfp->length) <= off);
+		ASSERT(matched || be16_to_cpu(dfp->length) >= be16_to_cpu(dup->length));
 		if (dfp > &d->hdr.bestfree[0])
-			ASSERT(INT_GET(dfp[-1].length, ARCH_CONVERT) >= INT_GET(dfp[0].length, ARCH_CONVERT));
+			ASSERT(be16_to_cpu(dfp[-1].length) >= be16_to_cpu(dfp[0].length));
 	}
 #endif
 	/*
 	 * If this is smaller than the smallest bestfree entry,
 	 * it can't be there since they're sorted.
 	 */
-	if (INT_GET(dup->length, ARCH_CONVERT) < INT_GET(d->hdr.bestfree[XFS_DIR2_DATA_FD_COUNT - 1].length, ARCH_CONVERT))
+	if (be16_to_cpu(dup->length) <
+	    be16_to_cpu(d->hdr.bestfree[XFS_DIR2_DATA_FD_COUNT - 1].length))
 		return NULL;
 	/*
 	 * Look at the three bestfree entries for our guy.
@@ -227,7 +230,7 @@ xfs_dir2_data_freefind(
 	     dfp++) {
 		if (!dfp->offset)
 			return NULL;
-		if (INT_GET(dfp->offset, ARCH_CONVERT) == off)
+		if (be16_to_cpu(dfp->offset) == off)
 			return dfp;
 	}
 	/*
@@ -249,29 +252,29 @@ xfs_dir2_data_freeinsert(
 	xfs_dir2_data_free_t	new;		/* new bestfree entry */
 
 #ifdef __KERNEL__
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 #endif
 	dfp = d->hdr.bestfree;
-	INT_COPY(new.length, dup->length, ARCH_CONVERT);
-	INT_SET(new.offset, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dup - (char *)d));
+	new.length = dup->length;
+	new.offset = cpu_to_be16((char *)dup - (char *)d);
 	/*
 	 * Insert at position 0, 1, or 2; or not at all.
 	 */
-	if (INT_GET(new.length, ARCH_CONVERT) > INT_GET(dfp[0].length, ARCH_CONVERT)) {
+	if (be16_to_cpu(new.length) > be16_to_cpu(dfp[0].length)) {
 		dfp[2] = dfp[1];
 		dfp[1] = dfp[0];
 		dfp[0] = new;
 		*loghead = 1;
 		return &dfp[0];
 	}
-	if (INT_GET(new.length, ARCH_CONVERT) > INT_GET(dfp[1].length, ARCH_CONVERT)) {
+	if (be16_to_cpu(new.length) > be16_to_cpu(dfp[1].length)) {
 		dfp[2] = dfp[1];
 		dfp[1] = new;
 		*loghead = 1;
 		return &dfp[1];
 	}
-	if (INT_GET(new.length, ARCH_CONVERT) > INT_GET(dfp[2].length, ARCH_CONVERT)) {
+	if (be16_to_cpu(new.length) > be16_to_cpu(dfp[2].length)) {
 		dfp[2] = new;
 		*loghead = 1;
 		return &dfp[2];
@@ -289,8 +292,8 @@ xfs_dir2_data_freeremove(
 	int			*loghead)	/* out: log data header */
 {
 #ifdef __KERNEL__
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 #endif
 	/*
 	 * It's the first entry, slide the next 2 up.
@@ -334,8 +337,8 @@ xfs_dir2_data_freescan(
 	char			*p;		/* current entry pointer */
 
 #ifdef __KERNEL__
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 #endif
 	/*
 	 * Start by clearing the table.
@@ -348,7 +351,7 @@ xfs_dir2_data_freescan(
 	p = (char *)d->u;
 	if (aendp)
 		endp = aendp;
-	else if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC) {
+	else if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC) {
 		btp = XFS_DIR2_BLOCK_TAIL_P(mp, (xfs_dir2_block_t *)d);
 		endp = (char *)XFS_DIR2_BLOCK_LEAF_P(btp);
 	} else
@@ -361,11 +364,11 @@ xfs_dir2_data_freescan(
 		/*
 		 * If it's a free entry, insert it.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
 			ASSERT((char *)dup - (char *)d ==
-			       INT_GET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT));
+			       be16_to_cpu(*XFS_DIR2_DATA_UNUSED_TAG_P(dup)));
 			xfs_dir2_data_freeinsert(d, dup, loghead);
-			p += INT_GET(dup->length, ARCH_CONVERT);
+			p += be16_to_cpu(dup->length);
 		}
 		/*
 		 * For active entries, check their tags and skip them.
@@ -373,7 +376,7 @@ xfs_dir2_data_freescan(
 		else {
 			dep = (xfs_dir2_data_entry_t *)p;
 			ASSERT((char *)dep - (char *)d ==
-			       INT_GET(*XFS_DIR2_DATA_ENTRY_TAG_P(dep), ARCH_CONVERT));
+			       be16_to_cpu(*XFS_DIR2_DATA_ENTRY_TAG_P(dep)));
 			p += XFS_DIR2_DATA_ENTSIZE(dep->namelen);
 		}
 	}
@@ -415,8 +418,8 @@ xfs_dir2_data_init(
 	 * Initialize the header.
 	 */
 	d = bp->data;
-	INT_SET(d->hdr.magic, ARCH_CONVERT, XFS_DIR2_DATA_MAGIC);
-	INT_SET(d->hdr.bestfree[0].offset, ARCH_CONVERT, (xfs_dir2_data_off_t)sizeof(d->hdr));
+	d->hdr.magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
+	d->hdr.bestfree[0].offset = cpu_to_be16(sizeof(d->hdr));
 	for (i = 1; i < XFS_DIR2_DATA_FD_COUNT; i++) {
 		d->hdr.bestfree[i].length = 0;
 		d->hdr.bestfree[i].offset = 0;
@@ -425,13 +428,12 @@ xfs_dir2_data_init(
 	 * Set up an unused entry for the block's body.
 	 */
 	dup = &d->u[0].unused;
-	INT_SET(dup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
+	dup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
 
 	t=mp->m_dirblksize - (uint)sizeof(d->hdr);
-	INT_SET(d->hdr.bestfree[0].length, ARCH_CONVERT, t);
-	INT_SET(dup->length, ARCH_CONVERT, t);
-	INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT,
-		(xfs_dir2_data_off_t)((char *)dup - (char *)d));
+	d->hdr.bestfree[0].length = cpu_to_be16(t);
+	dup->length = cpu_to_be16(t);
+	*XFS_DIR2_DATA_UNUSED_TAG_P(dup) = cpu_to_be16((char *)dup - (char *)d);
 	/*
 	 * Log it and return it.
 	 */
@@ -453,8 +455,8 @@ xfs_dir2_data_log_entry(
 	xfs_dir2_data_t		*d;		/* data block pointer */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)dep - (char *)d),
 		(uint)((char *)(XFS_DIR2_DATA_ENTRY_TAG_P(dep) + 1) -
 		       (char *)d - 1));
@@ -471,8 +473,8 @@ xfs_dir2_data_log_header(
 	xfs_dir2_data_t		*d;		/* data block pointer */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)&d->hdr - (char *)d),
 		(uint)(sizeof(d->hdr) - 1));
 }
@@ -489,8 +491,8 @@ xfs_dir2_data_log_unused(
 	xfs_dir2_data_t		*d;		/* data block pointer */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 	/*
 	 * Log the first part of the unused entry.
 	 */
@@ -533,12 +535,12 @@ xfs_dir2_data_make_free(
 	/*
 	 * Figure out where the end of the data area is.
 	 */
-	if (INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC)
+	if (be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC)
 		endptr = (char *)d + mp->m_dirblksize;
 	else {
 		xfs_dir2_block_tail_t	*btp;	/* block tail */
 
-		ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
+		ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
 		btp = XFS_DIR2_BLOCK_TAIL_P(mp, (xfs_dir2_block_t *)d);
 		endptr = (char *)XFS_DIR2_BLOCK_LEAF_P(btp);
 	}
@@ -547,11 +549,11 @@ xfs_dir2_data_make_free(
 	 * the previous entry and see if it's free.
 	 */
 	if (offset > sizeof(d->hdr)) {
-		xfs_dir2_data_off_t	*tagp;	/* tag just before us */
+		__be16			*tagp;	/* tag just before us */
 
-		tagp = (xfs_dir2_data_off_t *)((char *)d + offset) - 1;
-		prevdup = (xfs_dir2_data_unused_t *)((char *)d + INT_GET(*tagp, ARCH_CONVERT));
-		if (INT_GET(prevdup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG)
+		tagp = (__be16 *)((char *)d + offset) - 1;
+		prevdup = (xfs_dir2_data_unused_t *)((char *)d + be16_to_cpu(*tagp));
+		if (be16_to_cpu(prevdup->freetag) != XFS_DIR2_DATA_FREE_TAG)
 			prevdup = NULL;
 	} else
 		prevdup = NULL;
@@ -562,7 +564,7 @@ xfs_dir2_data_make_free(
 	if ((char *)d + offset + len < endptr) {
 		postdup =
 			(xfs_dir2_data_unused_t *)((char *)d + offset + len);
-		if (INT_GET(postdup->freetag, ARCH_CONVERT) != XFS_DIR2_DATA_FREE_TAG)
+		if (be16_to_cpu(postdup->freetag) != XFS_DIR2_DATA_FREE_TAG)
 			postdup = NULL;
 	} else
 		postdup = NULL;
@@ -586,13 +588,13 @@ xfs_dir2_data_make_free(
 		 * since the third bestfree is there, there might be more
 		 * entries.
 		 */
-		needscan = d->hdr.bestfree[2].length;
+		needscan = (d->hdr.bestfree[2].length != 0);
 		/*
 		 * Fix up the new big freespace.
 		 */
-		INT_MOD(prevdup->length, ARCH_CONVERT, len + INT_GET(postdup->length, ARCH_CONVERT));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)prevdup - (char *)d));
+		be16_add(&prevdup->length, len + be16_to_cpu(postdup->length));
+		*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup) =
+			cpu_to_be16((char *)prevdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, prevdup);
 		if (!needscan) {
 			/*
@@ -614,7 +616,7 @@ xfs_dir2_data_make_free(
 			 */
 			dfp = xfs_dir2_data_freeinsert(d, prevdup, needlogp);
 			ASSERT(dfp == &d->hdr.bestfree[0]);
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(prevdup->length, ARCH_CONVERT));
+			ASSERT(dfp->length == prevdup->length);
 			ASSERT(!dfp[1].length);
 			ASSERT(!dfp[2].length);
 		}
@@ -624,9 +626,9 @@ xfs_dir2_data_make_free(
 	 */
 	else if (prevdup) {
 		dfp = xfs_dir2_data_freefind(d, prevdup);
-		INT_MOD(prevdup->length, ARCH_CONVERT, len);
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)prevdup - (char *)d));
+		be16_add(&prevdup->length, len);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(prevdup) =
+			cpu_to_be16((char *)prevdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, prevdup);
 		/*
 		 * If the previous entry was in the table, the new entry
@@ -640,8 +642,10 @@ xfs_dir2_data_make_free(
 		/*
 		 * Otherwise we need a scan if the new entry is big enough.
 		 */
-		else
-			needscan = INT_GET(prevdup->length, ARCH_CONVERT) > INT_GET(d->hdr.bestfree[2].length, ARCH_CONVERT);
+		else {
+			needscan = be16_to_cpu(prevdup->length) >
+				   be16_to_cpu(d->hdr.bestfree[2].length);
+		}
 	}
 	/*
 	 * The following entry is free, merge with it.
@@ -649,10 +653,10 @@ xfs_dir2_data_make_free(
 	else if (postdup) {
 		dfp = xfs_dir2_data_freefind(d, postdup);
 		newdup = (xfs_dir2_data_unused_t *)((char *)d + offset);
-		INT_SET(newdup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup->length, ARCH_CONVERT, len + INT_GET(postdup->length, ARCH_CONVERT));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup->length = cpu_to_be16(len + be16_to_cpu(postdup->length));
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		/*
 		 * If the following entry was in the table, the new entry
@@ -666,18 +670,20 @@ xfs_dir2_data_make_free(
 		/*
 		 * Otherwise we need a scan if the new entry is big enough.
 		 */
-		else
-			needscan = INT_GET(newdup->length, ARCH_CONVERT) > INT_GET(d->hdr.bestfree[2].length, ARCH_CONVERT);
+		else {
+			needscan = be16_to_cpu(newdup->length) >
+				   be16_to_cpu(d->hdr.bestfree[2].length);
+		}
 	}
 	/*
 	 * Neither neighbor is free.  Make a new entry.
 	 */
 	else {
 		newdup = (xfs_dir2_data_unused_t *)((char *)d + offset);
-		INT_SET(newdup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup->length, ARCH_CONVERT, len);
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup->length = cpu_to_be16(len);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		(void)xfs_dir2_data_freeinsert(d, newdup, needlogp);
 	}
@@ -707,18 +713,18 @@ xfs_dir2_data_use_free(
 	int			oldlen;		/* old unused entry's length */
 
 	d = bp->data;
-	ASSERT(INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC ||
-	       INT_GET(d->hdr.magic, ARCH_CONVERT) == XFS_DIR2_BLOCK_MAGIC);
-	ASSERT(INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG);
+	ASSERT(be32_to_cpu(d->hdr.magic) == XFS_DIR2_DATA_MAGIC ||
+	       be32_to_cpu(d->hdr.magic) == XFS_DIR2_BLOCK_MAGIC);
+	ASSERT(be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG);
 	ASSERT(offset >= (char *)dup - (char *)d);
-	ASSERT(offset + len <= (char *)dup + INT_GET(dup->length, ARCH_CONVERT) - (char *)d);
-	ASSERT((char *)dup - (char *)d == INT_GET(*XFS_DIR2_DATA_UNUSED_TAG_P(dup), ARCH_CONVERT));
+	ASSERT(offset + len <= (char *)dup + be16_to_cpu(dup->length) - (char *)d);
+	ASSERT((char *)dup - (char *)d == be16_to_cpu(*XFS_DIR2_DATA_UNUSED_TAG_P(dup)));
 	/*
 	 * Look up the entry in the bestfree table.
 	 */
 	dfp = xfs_dir2_data_freefind(d, dup);
-	oldlen = INT_GET(dup->length, ARCH_CONVERT);
-	ASSERT(dfp || oldlen <= INT_GET(d->hdr.bestfree[2].length, ARCH_CONVERT));
+	oldlen = be16_to_cpu(dup->length);
+	ASSERT(dfp || oldlen <= be16_to_cpu(d->hdr.bestfree[2].length));
 	/*
 	 * Check for alignment with front and back of the entry.
 	 */
@@ -732,7 +738,7 @@ xfs_dir2_data_use_free(
 	 */
 	if (matchfront && matchback) {
 		if (dfp) {
-			needscan = d->hdr.bestfree[2].offset;
+			needscan = (d->hdr.bestfree[2].offset != 0);
 			if (!needscan)
 				xfs_dir2_data_freeremove(d, dfp, needlogp);
 		}
@@ -743,10 +749,10 @@ xfs_dir2_data_use_free(
 	 */
 	else if (matchfront) {
 		newdup = (xfs_dir2_data_unused_t *)((char *)d + offset + len);
-		INT_SET(newdup->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup->length, ARCH_CONVERT, oldlen - len);
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup->length = cpu_to_be16(oldlen - len);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		/*
 		 * If it was in the table, remove it and add the new one.
@@ -755,8 +761,8 @@ xfs_dir2_data_use_free(
 			xfs_dir2_data_freeremove(d, dfp, needlogp);
 			dfp = xfs_dir2_data_freeinsert(d, newdup, needlogp);
 			ASSERT(dfp != NULL);
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(newdup->length, ARCH_CONVERT));
-			ASSERT(INT_GET(dfp->offset, ARCH_CONVERT) == (char *)newdup - (char *)d);
+			ASSERT(dfp->length == newdup->length);
+			ASSERT(be16_to_cpu(dfp->offset) == (char *)newdup - (char *)d);
 			/*
 			 * If we got inserted at the last slot,
 			 * that means we don't know if there was a better
@@ -771,10 +777,9 @@ xfs_dir2_data_use_free(
 	 */
 	else if (matchback) {
 		newdup = dup;
-		INT_SET(newdup->length, ARCH_CONVERT, (xfs_dir2_data_off_t)
-			(((char *)d + offset) - (char *)newdup));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->length = cpu_to_be16(((char *)d + offset) - (char *)newdup);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		/*
 		 * If it was in the table, remove it and add the new one.
@@ -783,8 +788,8 @@ xfs_dir2_data_use_free(
 			xfs_dir2_data_freeremove(d, dfp, needlogp);
 			dfp = xfs_dir2_data_freeinsert(d, newdup, needlogp);
 			ASSERT(dfp != NULL);
-			ASSERT(INT_GET(dfp->length, ARCH_CONVERT) == INT_GET(newdup->length, ARCH_CONVERT));
-			ASSERT(INT_GET(dfp->offset, ARCH_CONVERT) == (char *)newdup - (char *)d);
+			ASSERT(dfp->length == newdup->length);
+			ASSERT(be16_to_cpu(dfp->offset) == (char *)newdup - (char *)d);
 			/*
 			 * If we got inserted at the last slot,
 			 * that means we don't know if there was a better
@@ -799,16 +804,15 @@ xfs_dir2_data_use_free(
 	 */
 	else {
 		newdup = dup;
-		INT_SET(newdup->length, ARCH_CONVERT, (xfs_dir2_data_off_t)
-			(((char *)d + offset) - (char *)newdup));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup - (char *)d));
+		newdup->length = cpu_to_be16(((char *)d + offset) - (char *)newdup);
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup) =
+			cpu_to_be16((char *)newdup - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup);
 		newdup2 = (xfs_dir2_data_unused_t *)((char *)d + offset + len);
-		INT_SET(newdup2->freetag, ARCH_CONVERT, XFS_DIR2_DATA_FREE_TAG);
-		INT_SET(newdup2->length, ARCH_CONVERT, oldlen - len - INT_GET(newdup->length, ARCH_CONVERT));
-		INT_SET(*XFS_DIR2_DATA_UNUSED_TAG_P(newdup2), ARCH_CONVERT,
-			(xfs_dir2_data_off_t)((char *)newdup2 - (char *)d));
+		newdup2->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
+		newdup2->length = cpu_to_be16(oldlen - len - be16_to_cpu(newdup->length));
+		*XFS_DIR2_DATA_UNUSED_TAG_P(newdup2) =
+			cpu_to_be16((char *)newdup2 - (char *)d);
 		xfs_dir2_data_log_unused(tp, bp, newdup2);
 		/*
 		 * If the old entry was in the table, we need to scan
@@ -819,7 +823,7 @@ xfs_dir2_data_use_free(
 		 * the 2 new will work.
 		 */
 		if (dfp) {
-			needscan = d->hdr.bestfree[2].length;
+			needscan = (d->hdr.bestfree[2].length != 0);
 			if (!needscan) {
 				xfs_dir2_data_freeremove(d, dfp, needlogp);
 				(void)xfs_dir2_data_freeinsert(d, newdup,
diff --git a/fs/xfs/xfs_dir2_data.h b/fs/xfs/xfs_dir2_data.h
index 5e3a7f9ec735..0847cbb53e17 100644
--- a/fs/xfs/xfs_dir2_data.h
+++ b/fs/xfs/xfs_dir2_data.h
@@ -65,8 +65,8 @@ struct xfs_trans;
  * The freespace will be formatted as a xfs_dir2_data_unused_t.
  */
 typedef struct xfs_dir2_data_free {
-	xfs_dir2_data_off_t	offset;		/* start of freespace */
-	xfs_dir2_data_off_t	length;		/* length of freespace */
+	__be16			offset;		/* start of freespace */
+	__be16			length;		/* length of freespace */
 } xfs_dir2_data_free_t;
 
 /*
@@ -75,7 +75,7 @@ typedef struct xfs_dir2_data_free {
  * The code knows that XFS_DIR2_DATA_FD_COUNT is 3.
  */
 typedef struct xfs_dir2_data_hdr {
-	__uint32_t		magic;		/* XFS_DIR2_DATA_MAGIC */
+	__be32			magic;		/* XFS_DIR2_DATA_MAGIC */
 						/* or XFS_DIR2_BLOCK_MAGIC */
 	xfs_dir2_data_free_t	bestfree[XFS_DIR2_DATA_FD_COUNT];
 } xfs_dir2_data_hdr_t;
@@ -97,10 +97,10 @@ typedef struct xfs_dir2_data_entry {
  * Tag appears as the last 2 bytes.
  */
 typedef struct xfs_dir2_data_unused {
-	__uint16_t		freetag;	/* XFS_DIR2_DATA_FREE_TAG */
-	xfs_dir2_data_off_t	length;		/* total free length */
+	__be16			freetag;	/* XFS_DIR2_DATA_FREE_TAG */
+	__be16			length;		/* total free length */
 						/* variable offset */
-	xfs_dir2_data_off_t	tag;		/* starting offset of us */
+	__be16			tag;		/* starting offset of us */
 } xfs_dir2_data_unused_t;
 
 typedef union {
@@ -134,12 +134,11 @@ static inline int xfs_dir2_data_entsize(int n)
  * Pointer to an entry's tag word.
  */
 #define	XFS_DIR2_DATA_ENTRY_TAG_P(dep)	xfs_dir2_data_entry_tag_p(dep)
-static inline xfs_dir2_data_off_t *
+static inline __be16 *
 xfs_dir2_data_entry_tag_p(xfs_dir2_data_entry_t *dep)
 {
-	return (xfs_dir2_data_off_t *) \
-		 ((char *)(dep) + XFS_DIR2_DATA_ENTSIZE((dep)->namelen) - \
-		  (uint)sizeof(xfs_dir2_data_off_t));
+	return (__be16 *)((char *)dep +
+		XFS_DIR2_DATA_ENTSIZE(dep->namelen) - sizeof(__be16));
 }
 
 /*
@@ -147,12 +146,11 @@ xfs_dir2_data_entry_tag_p(xfs_dir2_data_entry_t *dep)
  */
 #define	XFS_DIR2_DATA_UNUSED_TAG_P(dup) \
 	xfs_dir2_data_unused_tag_p(dup)
-static inline xfs_dir2_data_off_t *
+static inline __be16 *
 xfs_dir2_data_unused_tag_p(xfs_dir2_data_unused_t *dup)
 {
-	return (xfs_dir2_data_off_t *) \
-		 ((char *)(dup) + INT_GET((dup)->length, ARCH_CONVERT) \
-				- (uint)sizeof(xfs_dir2_data_off_t));
+	return (__be16 *)((char *)dup +
+			be16_to_cpu(dup->length) - sizeof(__be16));
 }
 
 /*
diff --git a/fs/xfs/xfs_dir2_leaf.c b/fs/xfs/xfs_dir2_leaf.c
index d342b6b55239..0f5e2f2ce6ec 100644
--- a/fs/xfs/xfs_dir2_leaf.c
+++ b/fs/xfs/xfs_dir2_leaf.c
@@ -66,7 +66,7 @@ xfs_dir2_block_to_leaf(
 	xfs_da_args_t		*args,		/* operation arguments */
 	xfs_dabuf_t		*dbp)		/* input block's buffer */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf's bestsp entries */
+	__be16			*bestsp;	/* leaf's bestsp entries */
 	xfs_dablk_t		blkno;		/* leaf block's bno */
 	xfs_dir2_block_t	*block;		/* block structure */
 	xfs_dir2_leaf_entry_t	*blp;		/* block's leaf entries */
@@ -111,14 +111,14 @@ xfs_dir2_block_to_leaf(
 	/*
 	 * Set the counts in the leaf header.
 	 */
-	INT_COPY(leaf->hdr.count, btp->count, ARCH_CONVERT); /* INT_: type change */
-	INT_COPY(leaf->hdr.stale, btp->stale, ARCH_CONVERT); /* INT_: type change */
+	leaf->hdr.count = cpu_to_be16(be32_to_cpu(btp->count));
+	leaf->hdr.stale = cpu_to_be16(be32_to_cpu(btp->stale));
 	/*
 	 * Could compact these but I think we always do the conversion
 	 * after squeezing out stale entries.
 	 */
-	memcpy(leaf->ents, blp, INT_GET(btp->count, ARCH_CONVERT) * sizeof(xfs_dir2_leaf_entry_t));
-	xfs_dir2_leaf_log_ents(tp, lbp, 0, INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1);
+	memcpy(leaf->ents, blp, be32_to_cpu(btp->count) * sizeof(xfs_dir2_leaf_entry_t));
+	xfs_dir2_leaf_log_ents(tp, lbp, 0, be16_to_cpu(leaf->hdr.count) - 1);
 	needscan = 0;
 	needlog = 1;
 	/*
@@ -133,7 +133,7 @@ xfs_dir2_block_to_leaf(
 	/*
 	 * Fix up the block header, make it a data block.
 	 */
-	INT_SET(block->hdr.magic, ARCH_CONVERT, XFS_DIR2_DATA_MAGIC);
+	block->hdr.magic = cpu_to_be32(XFS_DIR2_DATA_MAGIC);
 	if (needscan)
 		xfs_dir2_data_freescan(mp, (xfs_dir2_data_t *)block, &needlog,
 			NULL);
@@ -141,9 +141,9 @@ xfs_dir2_block_to_leaf(
 	 * Set up leaf tail and bests table.
 	 */
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
-	INT_SET(ltp->bestcount, ARCH_CONVERT, 1);
+	ltp->bestcount = cpu_to_be32(1);
 	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-	INT_COPY(bestsp[0], block->hdr.bestfree[0].length, ARCH_CONVERT);
+	bestsp[0] =  block->hdr.bestfree[0].length;
 	/*
 	 * Log the data header and leaf bests table.
 	 */
@@ -163,7 +163,7 @@ int						/* error */
 xfs_dir2_leaf_addname(
 	xfs_da_args_t		*args)		/* operation arguments */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* freespace table in leaf */
+	__be16			*bestsp;	/* freespace table in leaf */
 	int			compact;	/* need to compact leaves */
 	xfs_dir2_data_t		*data;		/* data block structure */
 	xfs_dabuf_t		*dbp;		/* data block buffer */
@@ -187,7 +187,7 @@ xfs_dir2_leaf_addname(
 	int			needbytes;	/* leaf block bytes needed */
 	int			needlog;	/* need to log data header */
 	int			needscan;	/* need to rescan data free */
-	xfs_dir2_data_off_t	*tagp;		/* end of data entry */
+	__be16			*tagp;		/* end of data entry */
 	xfs_trans_t		*tp;		/* transaction pointer */
 	xfs_dir2_db_t		use_block;	/* data block number */
 
@@ -222,14 +222,14 @@ xfs_dir2_leaf_addname(
 	 * in a data block, improving the lookup of those entries.
 	 */
 	for (use_block = -1, lep = &leaf->ents[index];
-	     index < INT_GET(leaf->hdr.count, ARCH_CONVERT) && INT_GET(lep->hashval, ARCH_CONVERT) == args->hashval;
+	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
 	     index++, lep++) {
-		if (INT_GET(lep->address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
-		i = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
-		ASSERT(i < INT_GET(ltp->bestcount, ARCH_CONVERT));
-		ASSERT(INT_GET(bestsp[i], ARCH_CONVERT) != NULLDATAOFF);
-		if (INT_GET(bestsp[i], ARCH_CONVERT) >= length) {
+		i = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
+		ASSERT(i < be32_to_cpu(ltp->bestcount));
+		ASSERT(be16_to_cpu(bestsp[i]) != NULLDATAOFF);
+		if (be16_to_cpu(bestsp[i]) >= length) {
 			use_block = i;
 			break;
 		}
@@ -238,13 +238,13 @@ xfs_dir2_leaf_addname(
 	 * Didn't find a block yet, linear search all the data blocks.
 	 */
 	if (use_block == -1) {
-		for (i = 0; i < INT_GET(ltp->bestcount, ARCH_CONVERT); i++) {
+		for (i = 0; i < be32_to_cpu(ltp->bestcount); i++) {
 			/*
 			 * Remember a block we see that's missing.
 			 */
-			if (INT_GET(bestsp[i], ARCH_CONVERT) == NULLDATAOFF && use_block == -1)
+			if (be16_to_cpu(bestsp[i]) == NULLDATAOFF && use_block == -1)
 				use_block = i;
-			else if (INT_GET(bestsp[i], ARCH_CONVERT) >= length) {
+			else if (be16_to_cpu(bestsp[i]) >= length) {
 				use_block = i;
 				break;
 			}
@@ -260,21 +260,21 @@ xfs_dir2_leaf_addname(
 	 * Now kill use_block if it refers to a missing block, so we
 	 * can use it as an indication of allocation needed.
 	 */
-	if (use_block != -1 && INT_GET(bestsp[use_block], ARCH_CONVERT) == NULLDATAOFF)
+	if (use_block != -1 && be16_to_cpu(bestsp[use_block]) == NULLDATAOFF)
 		use_block = -1;
 	/*
 	 * If we don't have enough free bytes but we can make enough
 	 * by compacting out stale entries, we'll do that.
 	 */
-	if ((char *)bestsp - (char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] < needbytes &&
-	    INT_GET(leaf->hdr.stale, ARCH_CONVERT) > 1) {
+	if ((char *)bestsp - (char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] < needbytes &&
+	    be16_to_cpu(leaf->hdr.stale) > 1) {
 		compact = 1;
 	}
 	/*
 	 * Otherwise if we don't have enough free bytes we need to
 	 * convert to node form.
 	 */
-	else if ((char *)bestsp - (char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] <
+	else if ((char *)bestsp - (char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] <
 		 needbytes) {
 		/*
 		 * Just checking or no space reservation, give up.
@@ -330,8 +330,8 @@ xfs_dir2_leaf_addname(
 	 * There are stale entries, so we'll need log-low and log-high
 	 * impossibly bad values later.
 	 */
-	else if (INT_GET(leaf->hdr.stale, ARCH_CONVERT)) {
-		lfloglow = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+	else if (be16_to_cpu(leaf->hdr.stale)) {
+		lfloglow = be16_to_cpu(leaf->hdr.count);
 		lfloghigh = -1;
 	}
 	/*
@@ -358,13 +358,13 @@ xfs_dir2_leaf_addname(
 		 * If we're adding a new data block on the end we need to
 		 * extend the bests table.  Copy it up one entry.
 		 */
-		if (use_block >= INT_GET(ltp->bestcount, ARCH_CONVERT)) {
+		if (use_block >= be32_to_cpu(ltp->bestcount)) {
 			bestsp--;
 			memmove(&bestsp[0], &bestsp[1],
-				INT_GET(ltp->bestcount, ARCH_CONVERT) * sizeof(bestsp[0]));
-			INT_MOD(ltp->bestcount, ARCH_CONVERT, +1);
+				be32_to_cpu(ltp->bestcount) * sizeof(bestsp[0]));
+			be32_add(&ltp->bestcount, 1);
 			xfs_dir2_leaf_log_tail(tp, lbp);
-			xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+			xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 		}
 		/*
 		 * If we're filling in a previously empty block just log it.
@@ -372,7 +372,7 @@ xfs_dir2_leaf_addname(
 		else
 			xfs_dir2_leaf_log_bests(tp, lbp, use_block, use_block);
 		data = dbp->data;
-		INT_COPY(bestsp[use_block], data->hdr.bestfree[0].length, ARCH_CONVERT);
+		bestsp[use_block] = data->hdr.bestfree[0].length;
 		grown = 1;
 	}
 	/*
@@ -394,8 +394,8 @@ xfs_dir2_leaf_addname(
 	 * Point to the biggest freespace in our data block.
 	 */
 	dup = (xfs_dir2_data_unused_t *)
-	      ((char *)data + INT_GET(data->hdr.bestfree[0].offset, ARCH_CONVERT));
-	ASSERT(INT_GET(dup->length, ARCH_CONVERT) >= length);
+	      ((char *)data + be16_to_cpu(data->hdr.bestfree[0].offset));
+	ASSERT(be16_to_cpu(dup->length) >= length);
 	needscan = needlog = 0;
 	/*
 	 * Mark the initial part of our freespace in use for the new entry.
@@ -411,7 +411,7 @@ xfs_dir2_leaf_addname(
 	dep->namelen = args->namelen;
 	memcpy(dep->name, args->name, dep->namelen);
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)data));
+	*tagp = cpu_to_be16((char *)dep - (char *)data);
 	/*
 	 * Need to scan fix up the bestfree table.
 	 */
@@ -427,8 +427,8 @@ xfs_dir2_leaf_addname(
 	 * If the bests table needs to be changed, do it.
 	 * Log the change unless we've already done that.
 	 */
-	if (INT_GET(bestsp[use_block], ARCH_CONVERT) != INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT)) {
-		INT_COPY(bestsp[use_block], data->hdr.bestfree[0].length, ARCH_CONVERT);
+	if (be16_to_cpu(bestsp[use_block]) != be16_to_cpu(data->hdr.bestfree[0].length)) {
+		bestsp[use_block] = data->hdr.bestfree[0].length;
 		if (!grown)
 			xfs_dir2_leaf_log_bests(tp, lbp, use_block, use_block);
 	}
@@ -440,15 +440,15 @@ xfs_dir2_leaf_addname(
 		/*
 		 * lep is still good as the index leaf entry.
 		 */
-		if (index < INT_GET(leaf->hdr.count, ARCH_CONVERT))
+		if (index < be16_to_cpu(leaf->hdr.count))
 			memmove(lep + 1, lep,
-				(INT_GET(leaf->hdr.count, ARCH_CONVERT) - index) * sizeof(*lep));
+				(be16_to_cpu(leaf->hdr.count) - index) * sizeof(*lep));
 		/*
 		 * Record low and high logging indices for the leaf.
 		 */
 		lfloglow = index;
-		lfloghigh = INT_GET(leaf->hdr.count, ARCH_CONVERT);
-		INT_MOD(leaf->hdr.count, ARCH_CONVERT, +1);
+		lfloghigh = be16_to_cpu(leaf->hdr.count);
+		be16_add(&leaf->hdr.count, 1);
 	}
 	/*
 	 * There are stale entries.
@@ -468,7 +468,7 @@ xfs_dir2_leaf_addname(
 			 */
 			for (lowstale = index - 1;
 			     lowstale >= 0 &&
-				INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) !=
+				be32_to_cpu(leaf->ents[lowstale].address) !=
 				XFS_DIR2_NULL_DATAPTR;
 			     lowstale--)
 				continue;
@@ -478,8 +478,8 @@ xfs_dir2_leaf_addname(
 			 * lowstale entry would be better.
 			 */
 			for (highstale = index;
-			     highstale < INT_GET(leaf->hdr.count, ARCH_CONVERT) &&
-				INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) !=
+			     highstale < be16_to_cpu(leaf->hdr.count) &&
+				be32_to_cpu(leaf->ents[highstale].address) !=
 				XFS_DIR2_NULL_DATAPTR &&
 				(lowstale < 0 ||
 				 index - lowstale - 1 >= highstale - index);
@@ -490,10 +490,10 @@ xfs_dir2_leaf_addname(
 		 * If the low one is better, use it.
 		 */
 		if (lowstale >= 0 &&
-		    (highstale == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
+		    (highstale == be16_to_cpu(leaf->hdr.count) ||
 		     index - lowstale - 1 < highstale - index)) {
 			ASSERT(index - lowstale - 1 >= 0);
-			ASSERT(INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[lowstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			/*
 			 * Copy entries up to cover the stale entry
@@ -512,10 +512,10 @@ xfs_dir2_leaf_addname(
 		 */
 		else {
 			ASSERT(highstale - index >= 0);
-			ASSERT(INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[highstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			/*
-			 * Copy entries down to copver the stale entry
+			 * Copy entries down to cover the stale entry
 			 * and make room for the new entry.
 			 */
 			if (highstale - index > 0)
@@ -526,13 +526,14 @@ xfs_dir2_leaf_addname(
 			lfloglow = MIN(index, lfloglow);
 			lfloghigh = MAX(highstale, lfloghigh);
 		}
-		INT_MOD(leaf->hdr.stale, ARCH_CONVERT, -1);
+		be16_add(&leaf->hdr.stale, -1);
 	}
 	/*
 	 * Fill in the new leaf entry.
 	 */
-	INT_SET(lep->hashval, ARCH_CONVERT, args->hashval);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_DB_OFF_TO_DATAPTR(mp, use_block, INT_GET(*tagp, ARCH_CONVERT)));
+	lep->hashval = cpu_to_be32(args->hashval);
+	lep->address = cpu_to_be32(XFS_DIR2_DB_OFF_TO_DATAPTR(mp, use_block,
+				be16_to_cpu(*tagp)));
 	/*
 	 * Log the leaf fields and give up the buffers.
 	 */
@@ -563,30 +564,30 @@ xfs_dir2_leaf_check(
 
 	leaf = bp->data;
 	mp = dp->i_mount;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	/*
 	 * This value is not restrictive enough.
 	 * Should factor in the size of the bests table as well.
 	 * We can deduce a value for that from di_size.
 	 */
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
+	ASSERT(be16_to_cpu(leaf->hdr.count) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	/*
 	 * Leaves and bests don't overlap.
 	 */
-	ASSERT((char *)&leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT)] <=
+	ASSERT((char *)&leaf->ents[be16_to_cpu(leaf->hdr.count)] <=
 	       (char *)XFS_DIR2_LEAF_BESTS_P(ltp));
 	/*
 	 * Check hash value order, count stale entries.
 	 */
-	for (i = stale = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); i++) {
-		if (i + 1 < INT_GET(leaf->hdr.count, ARCH_CONVERT))
-			ASSERT(INT_GET(leaf->ents[i].hashval, ARCH_CONVERT) <=
-			       INT_GET(leaf->ents[i + 1].hashval, ARCH_CONVERT));
-		if (INT_GET(leaf->ents[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	for (i = stale = 0; i < be16_to_cpu(leaf->hdr.count); i++) {
+		if (i + 1 < be16_to_cpu(leaf->hdr.count))
+			ASSERT(be32_to_cpu(leaf->ents[i].hashval) <=
+			       be32_to_cpu(leaf->ents[i + 1].hashval));
+		if (be32_to_cpu(leaf->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
 			stale++;
 	}
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) == stale);
+	ASSERT(be16_to_cpu(leaf->hdr.stale) == stale);
 }
 #endif	/* DEBUG */
 
@@ -611,8 +612,8 @@ xfs_dir2_leaf_compact(
 	/*
 	 * Compress out the stale entries in place.
 	 */
-	for (from = to = 0, loglow = -1; from < INT_GET(leaf->hdr.count, ARCH_CONVERT); from++) {
-		if (INT_GET(leaf->ents[from].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+	for (from = to = 0, loglow = -1; from < be16_to_cpu(leaf->hdr.count); from++) {
+		if (be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Only actually copy the entries that are different.
@@ -627,8 +628,8 @@ xfs_dir2_leaf_compact(
 	/*
 	 * Update and log the header, log the leaf entries.
 	 */
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) == from - to);
-	INT_MOD(leaf->hdr.count, ARCH_CONVERT, -(INT_GET(leaf->hdr.stale, ARCH_CONVERT)));
+	ASSERT(be16_to_cpu(leaf->hdr.stale) == from - to);
+	be16_add(&leaf->hdr.count, -(be16_to_cpu(leaf->hdr.stale)));
 	leaf->hdr.stale = 0;
 	xfs_dir2_leaf_log_header(args->trans, bp);
 	if (loglow != -1)
@@ -662,14 +663,14 @@ xfs_dir2_leaf_compact_x1(
 	int		to;		/* destination copy index */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) > 1);
+	ASSERT(be16_to_cpu(leaf->hdr.stale) > 1);
 	index = *indexp;
 	/*
 	 * Find the first stale entry before our index, if any.
 	 */
 	for (lowstale = index - 1;
 	     lowstale >= 0 &&
-		INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR;
+		be32_to_cpu(leaf->ents[lowstale].address) != XFS_DIR2_NULL_DATAPTR;
 	     lowstale--)
 		continue;
 	/*
@@ -677,8 +678,8 @@ xfs_dir2_leaf_compact_x1(
 	 * Stop if the answer would be worse than lowstale.
 	 */
 	for (highstale = index;
-	     highstale < INT_GET(leaf->hdr.count, ARCH_CONVERT) &&
-		INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) != XFS_DIR2_NULL_DATAPTR &&
+	     highstale < be16_to_cpu(leaf->hdr.count) &&
+		be32_to_cpu(leaf->ents[highstale].address) != XFS_DIR2_NULL_DATAPTR &&
 		(lowstale < 0 || index - lowstale > highstale - index);
 	     highstale++)
 		continue;
@@ -686,7 +687,7 @@ xfs_dir2_leaf_compact_x1(
 	 * Pick the better of lowstale and highstale.
 	 */
 	if (lowstale >= 0 &&
-	    (highstale == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
+	    (highstale == be16_to_cpu(leaf->hdr.count) ||
 	     index - lowstale <= highstale - index))
 		keepstale = lowstale;
 	else
@@ -695,14 +696,14 @@ xfs_dir2_leaf_compact_x1(
 	 * Copy the entries in place, removing all the stale entries
 	 * except keepstale.
 	 */
-	for (from = to = 0; from < INT_GET(leaf->hdr.count, ARCH_CONVERT); from++) {
+	for (from = to = 0; from < be16_to_cpu(leaf->hdr.count); from++) {
 		/*
 		 * Notice the new value of index.
 		 */
 		if (index == from)
 			newindex = to;
 		if (from != keepstale &&
-		    INT_GET(leaf->ents[from].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR) {
+		    be32_to_cpu(leaf->ents[from].address) == XFS_DIR2_NULL_DATAPTR) {
 			if (from == to)
 				*lowlogp = to;
 			continue;
@@ -730,8 +731,8 @@ xfs_dir2_leaf_compact_x1(
 	/*
 	 * Adjust the leaf header values.
 	 */
-	INT_MOD(leaf->hdr.count, ARCH_CONVERT, -(from - to));
-	INT_SET(leaf->hdr.stale, ARCH_CONVERT, 1);
+	be16_add(&leaf->hdr.count, -(from - to));
+	leaf->hdr.stale = cpu_to_be16(1);
 	/*
 	 * Remember the low/high stale value only in the "right"
 	 * direction.
@@ -739,8 +740,8 @@ xfs_dir2_leaf_compact_x1(
 	if (lowstale >= newindex)
 		lowstale = -1;
 	else
-		highstale = INT_GET(leaf->hdr.count, ARCH_CONVERT);
-	*highlogp = INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1;
+		highstale = be16_to_cpu(leaf->hdr.count);
+	*highlogp = be16_to_cpu(leaf->hdr.count) - 1;
 	*lowstalep = lowstale;
 	*highstalep = highstale;
 }
@@ -766,7 +767,7 @@ xfs_dir2_leaf_getdents(
 	xfs_dir2_data_entry_t	*dep;		/* data entry */
 	xfs_dir2_data_unused_t	*dup;		/* unused entry */
 	int			eof;		/* reached end of directory */
-	int			error=0;		/* error return value */
+	int			error = 0;	/* error return value */
 	int			i;		/* temporary loop index */
 	int			j;		/* temporary loop index */
 	int			length;		/* temporary length value */
@@ -778,8 +779,8 @@ xfs_dir2_leaf_getdents(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	xfs_dir2_off_t		newoff;		/* new curoff after new blk */
 	int			nmap;		/* mappings to ask xfs_bmapi */
-	xfs_dir2_put_args_t	p;		/* formatting arg bundle */
-	char			*ptr=NULL;		/* pointer to current data */
+	xfs_dir2_put_args_t	*p;		/* formatting arg bundle */
+	char			*ptr = NULL;	/* pointer to current data */
 	int			ra_current;	/* number of read-ahead blks */
 	int			ra_index;	/* *map index for read-ahead */
 	int			ra_offset;	/* map entry offset for ra */
@@ -797,9 +798,10 @@ xfs_dir2_leaf_getdents(
 	/*
 	 * Setup formatting arguments.
 	 */
-	p.dbp = dbp;
-	p.put = put;
-	p.uio = uio;
+	p = kmem_alloc(sizeof(*p), KM_SLEEP);
+	p->dbp = dbp;
+	p->put = put;
+	p->uio = uio;
 	/*
 	 * Set up to bmap a number of blocks based on the caller's
 	 * buffer size, the directory block size, and the filesystem
@@ -1046,11 +1048,10 @@ xfs_dir2_leaf_getdents(
 				while ((char *)ptr - (char *)data < byteoff) {
 					dup = (xfs_dir2_data_unused_t *)ptr;
 
-					if (INT_GET(dup->freetag, ARCH_CONVERT)
+					if (be16_to_cpu(dup->freetag)
 						  == XFS_DIR2_DATA_FREE_TAG) {
 
-						length = INT_GET(dup->length,
-								 ARCH_CONVERT);
+						length = be16_to_cpu(dup->length);
 						ptr += length;
 						continue;
 					}
@@ -1079,9 +1080,8 @@ xfs_dir2_leaf_getdents(
 		/*
 		 * No, it's unused, skip over it.
 		 */
-		if (INT_GET(dup->freetag, ARCH_CONVERT)
-						== XFS_DIR2_DATA_FREE_TAG) {
-			length = INT_GET(dup->length, ARCH_CONVERT);
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			length = be16_to_cpu(dup->length);
 			ptr += length;
 			curoff += length;
 			continue;
@@ -1092,24 +1092,24 @@ xfs_dir2_leaf_getdents(
 		 */
 		dep = (xfs_dir2_data_entry_t *)ptr;
 
-		p.namelen = dep->namelen;
+		p->namelen = dep->namelen;
 
-		length = XFS_DIR2_DATA_ENTSIZE(p.namelen);
+		length = XFS_DIR2_DATA_ENTSIZE(p->namelen);
 
-		p.cook = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff + length);
+		p->cook = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff + length);
 
-		p.ino = INT_GET(dep->inumber, ARCH_CONVERT);
+		p->ino = INT_GET(dep->inumber, ARCH_CONVERT);
 #if XFS_BIG_INUMS
-		p.ino += mp->m_inoadd;
+		p->ino += mp->m_inoadd;
 #endif
-		p.name = (char *)dep->name;
+		p->name = (char *)dep->name;
 
-		error = p.put(&p);
+		error = p->put(p);
 
 		/*
 		 * Won't fit.  Return to caller.
 		 */
-		if (!p.done) {
+		if (!p->done) {
 			eof = 0;
 			break;
 		}
@@ -1129,6 +1129,7 @@ xfs_dir2_leaf_getdents(
 	else
 		uio->uio_offset = XFS_DIR2_BYTE_TO_DATAPTR(mp, curoff);
 	kmem_free(map, map_size * sizeof(*map));
+	kmem_free(p, sizeof(*p));
 	if (bp)
 		xfs_da_brelse(tp, bp);
 	return error;
@@ -1171,7 +1172,7 @@ xfs_dir2_leaf_init(
 	/*
 	 * Initialize the header.
 	 */
-	INT_SET(leaf->hdr.info.magic, ARCH_CONVERT, magic);
+	leaf->hdr.info.magic = cpu_to_be16(magic);
 	leaf->hdr.info.forw = 0;
 	leaf->hdr.info.back = 0;
 	leaf->hdr.count = 0;
@@ -1201,13 +1202,13 @@ xfs_dir2_leaf_log_bests(
 	int			first,		/* first entry to log */
 	int			last)		/* last entry to log */
 {
-	xfs_dir2_data_off_t	*firstb;	/* pointer to first entry */
-	xfs_dir2_data_off_t	*lastb;		/* pointer to last entry */
+	__be16			*firstb;	/* pointer to first entry */
+	__be16			*lastb;		/* pointer to last entry */
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	ltp = XFS_DIR2_LEAF_TAIL_P(tp->t_mountp, leaf);
 	firstb = XFS_DIR2_LEAF_BESTS_P(ltp) + first;
 	lastb = XFS_DIR2_LEAF_BESTS_P(ltp) + last;
@@ -1230,8 +1231,8 @@ xfs_dir2_leaf_log_ents(
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC ||
-	       INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC ||
+	       be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	firstlep = &leaf->ents[first];
 	lastlep = &leaf->ents[last];
 	xfs_da_log_buf(tp, bp, (uint)((char *)firstlep - (char *)leaf),
@@ -1249,8 +1250,8 @@ xfs_dir2_leaf_log_header(
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC ||
-	       INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC ||
+	       be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)&leaf->hdr - (char *)leaf),
 		(uint)(sizeof(leaf->hdr) - 1));
 }
@@ -1269,7 +1270,7 @@ xfs_dir2_leaf_log_tail(
 
 	mp = tp->t_mountp;
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAF1_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAF1_MAGIC);
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	xfs_da_log_buf(tp, bp, (uint)((char *)ltp - (char *)leaf),
 		(uint)(mp->m_dirblksize - 1));
@@ -1314,7 +1315,7 @@ xfs_dir2_leaf_lookup(
 	 */
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->data +
-	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, INT_GET(lep->address, ARCH_CONVERT)));
+	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, be32_to_cpu(lep->address)));
 	/*
 	 * Return the found inode number.
 	 */
@@ -1373,17 +1374,17 @@ xfs_dir2_leaf_lookup_int(
 	 * looking to match the name.
 	 */
 	for (lep = &leaf->ents[index], dbp = NULL, curdb = -1;
-	     index < INT_GET(leaf->hdr.count, ARCH_CONVERT) && INT_GET(lep->hashval, ARCH_CONVERT) == args->hashval;
+	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
 	     lep++, index++) {
 		/*
 		 * Skip over stale leaf entries.
 		 */
-		if (INT_GET(lep->address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Get the new data block number.
 		 */
-		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 		/*
 		 * If it's not the same as the old data block number,
 		 * need to pitch the old one and read the new one.
@@ -1406,7 +1407,7 @@ xfs_dir2_leaf_lookup_int(
 		 */
 		dep = (xfs_dir2_data_entry_t *)
 		      ((char *)dbp->data +
-		       XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT)));
+		       XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
 		/*
 		 * If it matches then return it.
 		 */
@@ -1435,7 +1436,7 @@ int						/* error */
 xfs_dir2_leaf_removename(
 	xfs_da_args_t		*args)		/* operation arguments */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf block best freespace */
+	__be16			*bestsp;	/* leaf block best freespace */
 	xfs_dir2_data_t		*data;		/* data block structure */
 	xfs_dir2_db_t		db;		/* data block number */
 	xfs_dabuf_t		*dbp;		/* data block buffer */
@@ -1471,14 +1472,14 @@ xfs_dir2_leaf_removename(
 	 * Point to the leaf entry, use that to point to the data entry.
 	 */
 	lep = &leaf->ents[index];
-	db = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+	db = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)data + XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT)));
+	      ((char *)data + XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
 	needscan = needlog = 0;
-	oldbest = INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT);
+	oldbest = be16_to_cpu(data->hdr.bestfree[0].length);
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
 	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-	ASSERT(INT_GET(bestsp[db], ARCH_CONVERT) == oldbest);
+	ASSERT(be16_to_cpu(bestsp[db]) == oldbest);
 	/*
 	 * Mark the former data entry unused.
 	 */
@@ -1488,9 +1489,9 @@ xfs_dir2_leaf_removename(
 	/*
 	 * We just mark the leaf entry stale by putting a null in it.
 	 */
-	INT_MOD(leaf->hdr.stale, ARCH_CONVERT, +1);
+	be16_add(&leaf->hdr.stale, 1);
 	xfs_dir2_leaf_log_header(tp, lbp);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_NULL_DATAPTR);
+	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
 	xfs_dir2_leaf_log_ents(tp, lbp, index, index);
 	/*
 	 * Scan the freespace in the data block again if necessary,
@@ -1504,15 +1505,15 @@ xfs_dir2_leaf_removename(
 	 * If the longest freespace in the data block has changed,
 	 * put the new value in the bests table and log that.
 	 */
-	if (INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) != oldbest) {
-		INT_COPY(bestsp[db], data->hdr.bestfree[0].length, ARCH_CONVERT);
+	if (be16_to_cpu(data->hdr.bestfree[0].length) != oldbest) {
+		bestsp[db] = data->hdr.bestfree[0].length;
 		xfs_dir2_leaf_log_bests(tp, lbp, db, db);
 	}
 	xfs_dir2_data_check(dp, dbp);
 	/*
 	 * If the data block is now empty then get rid of the data block.
 	 */
-	if (INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) ==
+	if (be16_to_cpu(data->hdr.bestfree[0].length) ==
 	    mp->m_dirblksize - (uint)sizeof(data->hdr)) {
 		ASSERT(db != mp->m_dirdatablk);
 		if ((error = xfs_dir2_shrink_inode(args, db, dbp))) {
@@ -1535,12 +1536,12 @@ xfs_dir2_leaf_removename(
 		 * If this is the last data block then compact the
 		 * bests table by getting rid of entries.
 		 */
-		if (db == INT_GET(ltp->bestcount, ARCH_CONVERT) - 1) {
+		if (db == be32_to_cpu(ltp->bestcount) - 1) {
 			/*
 			 * Look for the last active entry (i).
 			 */
 			for (i = db - 1; i > 0; i--) {
-				if (INT_GET(bestsp[i], ARCH_CONVERT) != NULLDATAOFF)
+				if (be16_to_cpu(bestsp[i]) != NULLDATAOFF)
 					break;
 			}
 			/*
@@ -1548,12 +1549,12 @@ xfs_dir2_leaf_removename(
 			 * end are removed.
 			 */
 			memmove(&bestsp[db - i], bestsp,
-				(INT_GET(ltp->bestcount, ARCH_CONVERT) - (db - i)) * sizeof(*bestsp));
-			INT_MOD(ltp->bestcount, ARCH_CONVERT, -(db - i));
+				(be32_to_cpu(ltp->bestcount) - (db - i)) * sizeof(*bestsp));
+			be32_add(&ltp->bestcount, -(db - i));
 			xfs_dir2_leaf_log_tail(tp, lbp);
-			xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+			xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 		} else
-			INT_SET(bestsp[db], ARCH_CONVERT, NULLDATAOFF);
+			bestsp[db] = cpu_to_be16(NULLDATAOFF);
 	}
 	/*
 	 * If the data block was not the first one, drop it.
@@ -1604,7 +1605,7 @@ xfs_dir2_leaf_replace(
 	 */
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->data +
-	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, INT_GET(lep->address, ARCH_CONVERT)));
+	       XFS_DIR2_DATAPTR_TO_OFF(dp->i_mount, be32_to_cpu(lep->address)));
 	ASSERT(args->inumber != INT_GET(dep->inumber, ARCH_CONVERT));
 	/*
 	 * Put the new inode number in, log it.
@@ -1645,11 +1646,11 @@ xfs_dir2_leaf_search_hash(
 	 * Note, the table cannot be empty, so we have to go through the loop.
 	 * Binary search the leaf entries looking for our hash value.
 	 */
-	for (lep = leaf->ents, low = 0, high = INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1,
+	for (lep = leaf->ents, low = 0, high = be16_to_cpu(leaf->hdr.count) - 1,
 		hashwant = args->hashval;
 	     low <= high; ) {
 		mid = (low + high) >> 1;
-		if ((hash = INT_GET(lep[mid].hashval, ARCH_CONVERT)) == hashwant)
+		if ((hash = be32_to_cpu(lep[mid].hashval)) == hashwant)
 			break;
 		if (hash < hashwant)
 			low = mid + 1;
@@ -1660,7 +1661,7 @@ xfs_dir2_leaf_search_hash(
 	 * Found one, back up through all the equal hash values.
 	 */
 	if (hash == hashwant) {
-		while (mid > 0 && INT_GET(lep[mid - 1].hashval, ARCH_CONVERT) == hashwant) {
+		while (mid > 0 && be32_to_cpu(lep[mid - 1].hashval) == hashwant) {
 			mid--;
 		}
 	}
@@ -1682,7 +1683,7 @@ xfs_dir2_leaf_trim_data(
 	xfs_dabuf_t		*lbp,		/* leaf buffer */
 	xfs_dir2_db_t		db)		/* data block number */
 {
-	xfs_dir2_data_off_t	*bestsp;	/* leaf bests table */
+	__be16			*bestsp;	/* leaf bests table */
 #ifdef DEBUG
 	xfs_dir2_data_t		*data;		/* data block structure */
 #endif
@@ -1706,7 +1707,7 @@ xfs_dir2_leaf_trim_data(
 	}
 #ifdef DEBUG
 	data = dbp->data;
-	ASSERT(INT_GET(data->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC);
+	ASSERT(be32_to_cpu(data->hdr.magic) == XFS_DIR2_DATA_MAGIC);
 #endif
 	/* this seems to be an error
 	 * data is only valid if DEBUG is defined?
@@ -1715,9 +1716,9 @@ xfs_dir2_leaf_trim_data(
 
 	leaf = lbp->data;
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
-	ASSERT(INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) ==
+	ASSERT(be16_to_cpu(data->hdr.bestfree[0].length) ==
 	       mp->m_dirblksize - (uint)sizeof(data->hdr));
-	ASSERT(db == INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+	ASSERT(db == be32_to_cpu(ltp->bestcount) - 1);
 	/*
 	 * Get rid of the data block.
 	 */
@@ -1730,10 +1731,10 @@ xfs_dir2_leaf_trim_data(
 	 * Eliminate the last bests entry from the table.
 	 */
 	bestsp = XFS_DIR2_LEAF_BESTS_P(ltp);
-	INT_MOD(ltp->bestcount, ARCH_CONVERT, -1);
-	memmove(&bestsp[1], &bestsp[0], INT_GET(ltp->bestcount, ARCH_CONVERT) * sizeof(*bestsp));
+	be32_add(&ltp->bestcount, -1);
+	memmove(&bestsp[1], &bestsp[0], be32_to_cpu(ltp->bestcount) * sizeof(*bestsp));
 	xfs_dir2_leaf_log_tail(tp, lbp);
-	xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+	xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 	return 0;
 }
 
@@ -1805,7 +1806,7 @@ xfs_dir2_node_to_leaf(
 		return 0;
 	lbp = state->path.blk[0].bp;
 	leaf = lbp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * Read the freespace block.
 	 */
@@ -1814,15 +1815,15 @@ xfs_dir2_node_to_leaf(
 		return error;
 	}
 	free = fbp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	ASSERT(!free->hdr.firstdb);
 	/*
 	 * Now see if the leafn and free data will fit in a leaf1.
 	 * If not, release the buffer and give up.
 	 */
 	if ((uint)sizeof(leaf->hdr) +
-	    (INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT)) * (uint)sizeof(leaf->ents[0]) +
-	    INT_GET(free->hdr.nvalid, ARCH_CONVERT) * (uint)sizeof(leaf->bests[0]) +
+	    (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale)) * (uint)sizeof(leaf->ents[0]) +
+	    be32_to_cpu(free->hdr.nvalid) * (uint)sizeof(leaf->bests[0]) +
 	    (uint)sizeof(leaf->tail) >
 	    mp->m_dirblksize) {
 		xfs_da_brelse(tp, fbp);
@@ -1832,22 +1833,22 @@ xfs_dir2_node_to_leaf(
 	 * If the leaf has any stale entries in it, compress them out.
 	 * The compact routine will log the header.
 	 */
-	if (INT_GET(leaf->hdr.stale, ARCH_CONVERT))
+	if (be16_to_cpu(leaf->hdr.stale))
 		xfs_dir2_leaf_compact(args, lbp);
 	else
 		xfs_dir2_leaf_log_header(tp, lbp);
-	INT_SET(leaf->hdr.info.magic, ARCH_CONVERT, XFS_DIR2_LEAF1_MAGIC);
+	leaf->hdr.info.magic = cpu_to_be16(XFS_DIR2_LEAF1_MAGIC);
 	/*
 	 * Set up the leaf tail from the freespace block.
 	 */
 	ltp = XFS_DIR2_LEAF_TAIL_P(mp, leaf);
-	INT_COPY(ltp->bestcount, free->hdr.nvalid, ARCH_CONVERT);
+	ltp->bestcount = free->hdr.nvalid;
 	/*
 	 * Set up the leaf bests table.
 	 */
 	memcpy(XFS_DIR2_LEAF_BESTS_P(ltp), free->bests,
-		INT_GET(ltp->bestcount, ARCH_CONVERT) * sizeof(leaf->bests[0]));
-	xfs_dir2_leaf_log_bests(tp, lbp, 0, INT_GET(ltp->bestcount, ARCH_CONVERT) - 1);
+		be32_to_cpu(ltp->bestcount) * sizeof(leaf->bests[0]));
+	xfs_dir2_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 	xfs_dir2_leaf_log_tail(tp, lbp);
 	xfs_dir2_leaf_check(dp, lbp);
 	/*
diff --git a/fs/xfs/xfs_dir2_leaf.h b/fs/xfs/xfs_dir2_leaf.h
index 1393993d61e9..f57ca1162412 100644
--- a/fs/xfs/xfs_dir2_leaf.h
+++ b/fs/xfs/xfs_dir2_leaf.h
@@ -46,23 +46,23 @@ typedef	__uint32_t	xfs_dir2_dataptr_t;
  */
 typedef struct xfs_dir2_leaf_hdr {
 	xfs_da_blkinfo_t	info;		/* header for da routines */
-	__uint16_t		count;		/* count of entries */
-	__uint16_t		stale;		/* count of stale entries */
+	__be16			count;		/* count of entries */
+	__be16			stale;		/* count of stale entries */
 } xfs_dir2_leaf_hdr_t;
 
 /*
  * Leaf block entry.
  */
 typedef struct xfs_dir2_leaf_entry {
-	xfs_dahash_t		hashval;	/* hash value of name */
-	xfs_dir2_dataptr_t	address;	/* address of data entry */
+	__be32			hashval;	/* hash value of name */
+	__be32			address;	/* address of data entry */
 } xfs_dir2_leaf_entry_t;
 
 /*
  * Leaf block tail.
  */
 typedef struct xfs_dir2_leaf_tail {
-	__uint32_t		bestcount;
+	__be32			bestcount;
 } xfs_dir2_leaf_tail_t;
 
 /*
@@ -105,11 +105,10 @@ xfs_dir2_leaf_tail_p(struct xfs_mount *mp, xfs_dir2_leaf_t *lp)
  * Get address of the bests array in the single-leaf block.
  */
 #define	XFS_DIR2_LEAF_BESTS_P(ltp)	xfs_dir2_leaf_bests_p(ltp)
-static inline xfs_dir2_data_off_t *
+static inline __be16 *
 xfs_dir2_leaf_bests_p(xfs_dir2_leaf_tail_t *ltp)
 {
-	return (xfs_dir2_data_off_t *)
-		(ltp) - INT_GET((ltp)->bestcount, ARCH_CONVERT);
+	return (__be16 *)ltp - be32_to_cpu(ltp->bestcount);
 }
 
 /*
diff --git a/fs/xfs/xfs_dir2_node.c b/fs/xfs/xfs_dir2_node.c
index 641f8633d254..ac511ab9c52d 100644
--- a/fs/xfs/xfs_dir2_node.c
+++ b/fs/xfs/xfs_dir2_node.c
@@ -76,7 +76,7 @@ xfs_dir2_free_log_bests(
 	xfs_dir2_free_t		*free;		/* freespace structure */
 
 	free = bp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	xfs_da_log_buf(tp, bp,
 		(uint)((char *)&free->bests[first] - (char *)free),
 		(uint)((char *)&free->bests[last] - (char *)free +
@@ -94,7 +94,7 @@ xfs_dir2_free_log_header(
 	xfs_dir2_free_t		*free;		/* freespace structure */
 
 	free = bp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	xfs_da_log_buf(tp, bp, (uint)((char *)&free->hdr - (char *)free),
 		(uint)(sizeof(xfs_dir2_free_hdr_t) - 1));
 }
@@ -114,14 +114,14 @@ xfs_dir2_leaf_to_node(
 	xfs_dabuf_t		*fbp;		/* freespace buffer */
 	xfs_dir2_db_t		fdb;		/* freespace block number */
 	xfs_dir2_free_t		*free;		/* freespace structure */
-	xfs_dir2_data_off_t	*from;		/* pointer to freespace entry */
+	__be16			*from;		/* pointer to freespace entry */
 	int			i;		/* leaf freespace index */
 	xfs_dir2_leaf_t		*leaf;		/* leaf structure */
 	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			n;		/* count of live freespc ents */
 	xfs_dir2_data_off_t	off;		/* freespace entry value */
-	xfs_dir2_data_off_t	*to;		/* pointer to freespace entry */
+	__be16			*to;		/* pointer to freespace entry */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
 	xfs_dir2_trace_args_b("leaf_to_node", args, lbp);
@@ -149,28 +149,28 @@ xfs_dir2_leaf_to_node(
 	/*
 	 * Initialize the freespace block header.
 	 */
-	INT_SET(free->hdr.magic, ARCH_CONVERT, XFS_DIR2_FREE_MAGIC);
+	free->hdr.magic = cpu_to_be32(XFS_DIR2_FREE_MAGIC);
 	free->hdr.firstdb = 0;
-	ASSERT(INT_GET(ltp->bestcount, ARCH_CONVERT) <= (uint)dp->i_d.di_size / mp->m_dirblksize);
-	INT_COPY(free->hdr.nvalid, ltp->bestcount, ARCH_CONVERT);
+	ASSERT(be32_to_cpu(ltp->bestcount) <= (uint)dp->i_d.di_size / mp->m_dirblksize);
+	free->hdr.nvalid = ltp->bestcount;
 	/*
 	 * Copy freespace entries from the leaf block to the new block.
 	 * Count active entries.
 	 */
 	for (i = n = 0, from = XFS_DIR2_LEAF_BESTS_P(ltp), to = free->bests;
-	     i < INT_GET(ltp->bestcount, ARCH_CONVERT); i++, from++, to++) {
-		if ((off = INT_GET(*from, ARCH_CONVERT)) != NULLDATAOFF)
+	     i < be32_to_cpu(ltp->bestcount); i++, from++, to++) {
+		if ((off = be16_to_cpu(*from)) != NULLDATAOFF)
 			n++;
-		INT_SET(*to, ARCH_CONVERT, off);
+		*to = cpu_to_be16(off);
 	}
-	INT_SET(free->hdr.nused, ARCH_CONVERT, n);
-	INT_SET(leaf->hdr.info.magic, ARCH_CONVERT, XFS_DIR2_LEAFN_MAGIC);
+	free->hdr.nused = cpu_to_be32(n);
+	leaf->hdr.info.magic = cpu_to_be16(XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * Log everything.
 	 */
 	xfs_dir2_leaf_log_header(tp, lbp);
 	xfs_dir2_free_log_header(tp, fbp);
-	xfs_dir2_free_log_bests(tp, fbp, 0, INT_GET(free->hdr.nvalid, ARCH_CONVERT) - 1);
+	xfs_dir2_free_log_bests(tp, fbp, 0, be32_to_cpu(free->hdr.nvalid) - 1);
 	xfs_da_buf_done(fbp);
 	xfs_dir2_leafn_check(dp, lbp);
 	return 0;
@@ -217,15 +217,15 @@ xfs_dir2_leafn_add(
 	 * a compact.
 	 */
 
-	if (INT_GET(leaf->hdr.count, ARCH_CONVERT) == XFS_DIR2_MAX_LEAF_ENTS(mp)) {
+	if (be16_to_cpu(leaf->hdr.count) == XFS_DIR2_MAX_LEAF_ENTS(mp)) {
 		if (!leaf->hdr.stale)
 			return XFS_ERROR(ENOSPC);
-		compact = INT_GET(leaf->hdr.stale, ARCH_CONVERT) > 1;
+		compact = be16_to_cpu(leaf->hdr.stale) > 1;
 	} else
 		compact = 0;
-	ASSERT(index == 0 || INT_GET(leaf->ents[index - 1].hashval, ARCH_CONVERT) <= args->hashval);
-	ASSERT(index == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
-	       INT_GET(leaf->ents[index].hashval, ARCH_CONVERT) >= args->hashval);
+	ASSERT(index == 0 || be32_to_cpu(leaf->ents[index - 1].hashval) <= args->hashval);
+	ASSERT(index == be16_to_cpu(leaf->hdr.count) ||
+	       be32_to_cpu(leaf->ents[index].hashval) >= args->hashval);
 
 	if (args->justcheck)
 		return 0;
@@ -242,7 +242,7 @@ xfs_dir2_leafn_add(
 	 * Set impossible logging indices for this case.
 	 */
 	else if (leaf->hdr.stale) {
-		lfloglow = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		lfloglow = be16_to_cpu(leaf->hdr.count);
 		lfloghigh = -1;
 	}
 	/*
@@ -250,12 +250,12 @@ xfs_dir2_leafn_add(
 	 */
 	if (!leaf->hdr.stale) {
 		lep = &leaf->ents[index];
-		if (index < INT_GET(leaf->hdr.count, ARCH_CONVERT))
+		if (index < be16_to_cpu(leaf->hdr.count))
 			memmove(lep + 1, lep,
-				(INT_GET(leaf->hdr.count, ARCH_CONVERT) - index) * sizeof(*lep));
+				(be16_to_cpu(leaf->hdr.count) - index) * sizeof(*lep));
 		lfloglow = index;
-		lfloghigh = INT_GET(leaf->hdr.count, ARCH_CONVERT);
-		INT_MOD(leaf->hdr.count, ARCH_CONVERT, +1);
+		lfloghigh = be16_to_cpu(leaf->hdr.count);
+		be16_add(&leaf->hdr.count, 1);
 	}
 	/*
 	 * There are stale entries.  We'll use one for the new entry.
@@ -271,7 +271,7 @@ xfs_dir2_leafn_add(
 			 */
 			for (lowstale = index - 1;
 			     lowstale >= 0 &&
-				INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) !=
+				be32_to_cpu(leaf->ents[lowstale].address) !=
 				XFS_DIR2_NULL_DATAPTR;
 			     lowstale--)
 				continue;
@@ -281,8 +281,8 @@ xfs_dir2_leafn_add(
 			 * lowstale already found.
 			 */
 			for (highstale = index;
-			     highstale < INT_GET(leaf->hdr.count, ARCH_CONVERT) &&
-				INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) !=
+			     highstale < be16_to_cpu(leaf->hdr.count) &&
+				be32_to_cpu(leaf->ents[highstale].address) !=
 				XFS_DIR2_NULL_DATAPTR &&
 				(lowstale < 0 ||
 				 index - lowstale - 1 >= highstale - index);
@@ -294,9 +294,9 @@ xfs_dir2_leafn_add(
 		 * Shift entries up toward the stale slot.
 		 */
 		if (lowstale >= 0 &&
-		    (highstale == INT_GET(leaf->hdr.count, ARCH_CONVERT) ||
+		    (highstale == be16_to_cpu(leaf->hdr.count) ||
 		     index - lowstale - 1 < highstale - index)) {
-			ASSERT(INT_GET(leaf->ents[lowstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[lowstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			ASSERT(index - lowstale - 1 >= 0);
 			if (index - lowstale - 1 > 0)
@@ -312,7 +312,7 @@ xfs_dir2_leafn_add(
 		 * Shift entries down toward the stale slot.
 		 */
 		else {
-			ASSERT(INT_GET(leaf->ents[highstale].address, ARCH_CONVERT) ==
+			ASSERT(be32_to_cpu(leaf->ents[highstale].address) ==
 			       XFS_DIR2_NULL_DATAPTR);
 			ASSERT(highstale - index >= 0);
 			if (highstale - index > 0)
@@ -323,13 +323,14 @@ xfs_dir2_leafn_add(
 			lfloglow = MIN(index, lfloglow);
 			lfloghigh = MAX(highstale, lfloghigh);
 		}
-		INT_MOD(leaf->hdr.stale, ARCH_CONVERT, -1);
+		be16_add(&leaf->hdr.stale, -1);
 	}
 	/*
 	 * Insert the new entry, log everything.
 	 */
-	INT_SET(lep->hashval, ARCH_CONVERT, args->hashval);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_DB_OFF_TO_DATAPTR(mp, args->blkno, args->index));
+	lep->hashval = cpu_to_be32(args->hashval);
+	lep->address = cpu_to_be32(XFS_DIR2_DB_OFF_TO_DATAPTR(mp,
+				args->blkno, args->index));
 	xfs_dir2_leaf_log_header(tp, bp);
 	xfs_dir2_leaf_log_ents(tp, bp, lfloglow, lfloghigh);
 	xfs_dir2_leafn_check(dp, bp);
@@ -352,17 +353,17 @@ xfs_dir2_leafn_check(
 
 	leaf = bp->data;
 	mp = dp->i_mount;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
-	for (i = stale = 0; i < INT_GET(leaf->hdr.count, ARCH_CONVERT); i++) {
-		if (i + 1 < INT_GET(leaf->hdr.count, ARCH_CONVERT)) {
-			ASSERT(INT_GET(leaf->ents[i].hashval, ARCH_CONVERT) <=
-			       INT_GET(leaf->ents[i + 1].hashval, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.count) <= XFS_DIR2_MAX_LEAF_ENTS(mp));
+	for (i = stale = 0; i < be16_to_cpu(leaf->hdr.count); i++) {
+		if (i + 1 < be16_to_cpu(leaf->hdr.count)) {
+			ASSERT(be32_to_cpu(leaf->ents[i].hashval) <=
+			       be32_to_cpu(leaf->ents[i + 1].hashval));
 		}
-		if (INT_GET(leaf->ents[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(leaf->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
 			stale++;
 	}
-	ASSERT(INT_GET(leaf->hdr.stale, ARCH_CONVERT) == stale);
+	ASSERT(be16_to_cpu(leaf->hdr.stale) == stale);
 }
 #endif	/* DEBUG */
 
@@ -378,12 +379,12 @@ xfs_dir2_leafn_lasthash(
 	xfs_dir2_leaf_t	*leaf;			/* leaf structure */
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	if (count)
-		*count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
+		*count = be16_to_cpu(leaf->hdr.count);
 	if (!leaf->hdr.count)
 		return 0;
-	return INT_GET(leaf->ents[INT_GET(leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+	return be32_to_cpu(leaf->ents[be16_to_cpu(leaf->hdr.count) - 1].hashval);
 }
 
 /*
@@ -419,9 +420,9 @@ xfs_dir2_leafn_lookup_int(
 	tp = args->trans;
 	mp = dp->i_mount;
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 #ifdef __KERNEL__
-	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) > 0);
+	ASSERT(be16_to_cpu(leaf->hdr.count) > 0);
 #endif
 	xfs_dir2_leafn_check(dp, bp);
 	/*
@@ -443,7 +444,7 @@ xfs_dir2_leafn_lookup_int(
 		curdb = -1;
 		length = XFS_DIR2_DATA_ENTSIZE(args->namelen);
 		if ((free = (curbp ? curbp->data : NULL)))
-			ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+			ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	}
 	/*
 	 * For others, it's a data block buffer, get the block number.
@@ -456,17 +457,17 @@ xfs_dir2_leafn_lookup_int(
 	 * Loop over leaf entries with the right hash value.
 	 */
 	for (lep = &leaf->ents[index];
-	     index < INT_GET(leaf->hdr.count, ARCH_CONVERT) && INT_GET(lep->hashval, ARCH_CONVERT) == args->hashval;
+	     index < be16_to_cpu(leaf->hdr.count) && be32_to_cpu(lep->hashval) == args->hashval;
 	     lep++, index++) {
 		/*
 		 * Skip stale leaf entries.
 		 */
-		if (INT_GET(lep->address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Pull the data block number from the entry.
 		 */
-		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+		newdb = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 		/*
 		 * For addname, we're looking for a place to put the new entry.
 		 * We want to use a data block with an entry of equal
@@ -506,15 +507,15 @@ xfs_dir2_leafn_lookup_int(
 					}
 					curfdb = newfdb;
 					free = curbp->data;
-					ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) ==
+					ASSERT(be32_to_cpu(free->hdr.magic) ==
 					       XFS_DIR2_FREE_MAGIC);
-					ASSERT((INT_GET(free->hdr.firstdb, ARCH_CONVERT) %
+					ASSERT((be32_to_cpu(free->hdr.firstdb) %
 						XFS_DIR2_MAX_FREE_BESTS(mp)) ==
 					       0);
-					ASSERT(INT_GET(free->hdr.firstdb, ARCH_CONVERT) <= curdb);
+					ASSERT(be32_to_cpu(free->hdr.firstdb) <= curdb);
 					ASSERT(curdb <
-					       INT_GET(free->hdr.firstdb, ARCH_CONVERT) +
-					       INT_GET(free->hdr.nvalid, ARCH_CONVERT));
+					       be32_to_cpu(free->hdr.firstdb) +
+					       be32_to_cpu(free->hdr.nvalid));
 				}
 				/*
 				 * Get the index for our entry.
@@ -523,12 +524,12 @@ xfs_dir2_leafn_lookup_int(
 				/*
 				 * If it has room, return it.
 				 */
-				if (unlikely(INT_GET(free->bests[fi], ARCH_CONVERT) == NULLDATAOFF)) {
+				if (unlikely(be16_to_cpu(free->bests[fi]) == NULLDATAOFF)) {
 					XFS_ERROR_REPORT("xfs_dir2_leafn_lookup_int",
 							 XFS_ERRLEVEL_LOW, mp);
 					return XFS_ERROR(EFSCORRUPTED);
 				}
-				if (INT_GET(free->bests[fi], ARCH_CONVERT) >= length) {
+				if (be16_to_cpu(free->bests[fi]) >= length) {
 					*indexp = index;
 					state->extravalid = 1;
 					state->extrablk.bp = curbp;
@@ -572,7 +573,7 @@ xfs_dir2_leafn_lookup_int(
 			 */
 			dep = (xfs_dir2_data_entry_t *)
 			      ((char *)curbp->data +
-			       XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT)));
+			       XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address)));
 			/*
 			 * Compare the entry, return it if it matches.
 			 */
@@ -619,7 +620,7 @@ xfs_dir2_leafn_lookup_int(
 	 * Return the final index, that will be the insertion point.
 	 */
 	*indexp = index;
-	ASSERT(index == INT_GET(leaf->hdr.count, ARCH_CONVERT) || args->oknoent);
+	ASSERT(index == be16_to_cpu(leaf->hdr.count) || args->oknoent);
 	return XFS_ERROR(ENOENT);
 }
 
@@ -657,12 +658,12 @@ xfs_dir2_leafn_moveents(
 	 * destination leaf entries, open up a hole in the destination
 	 * to hold the new entries.
 	 */
-	if (start_d < INT_GET(leaf_d->hdr.count, ARCH_CONVERT)) {
+	if (start_d < be16_to_cpu(leaf_d->hdr.count)) {
 		memmove(&leaf_d->ents[start_d + count], &leaf_d->ents[start_d],
-			(INT_GET(leaf_d->hdr.count, ARCH_CONVERT) - start_d) *
+			(be16_to_cpu(leaf_d->hdr.count) - start_d) *
 			sizeof(xfs_dir2_leaf_entry_t));
 		xfs_dir2_leaf_log_ents(tp, bp_d, start_d + count,
-			count + INT_GET(leaf_d->hdr.count, ARCH_CONVERT) - 1);
+			count + be16_to_cpu(leaf_d->hdr.count) - 1);
 	}
 	/*
 	 * If the source has stale leaves, count the ones in the copy range
@@ -672,7 +673,7 @@ xfs_dir2_leafn_moveents(
 		int	i;			/* temp leaf index */
 
 		for (i = start_s, stale = 0; i < start_s + count; i++) {
-			if (INT_GET(leaf_s->ents[i].address, ARCH_CONVERT) == XFS_DIR2_NULL_DATAPTR)
+			if (be32_to_cpu(leaf_s->ents[i].address) == XFS_DIR2_NULL_DATAPTR)
 				stale++;
 		}
 	} else
@@ -687,7 +688,7 @@ xfs_dir2_leafn_moveents(
 	 * If there are source entries after the ones we copied,
 	 * delete the ones we copied by sliding the next ones down.
 	 */
-	if (start_s + count < INT_GET(leaf_s->hdr.count, ARCH_CONVERT)) {
+	if (start_s + count < be16_to_cpu(leaf_s->hdr.count)) {
 		memmove(&leaf_s->ents[start_s], &leaf_s->ents[start_s + count],
 			count * sizeof(xfs_dir2_leaf_entry_t));
 		xfs_dir2_leaf_log_ents(tp, bp_s, start_s, start_s + count - 1);
@@ -695,10 +696,10 @@ xfs_dir2_leafn_moveents(
 	/*
 	 * Update the headers and log them.
 	 */
-	INT_MOD(leaf_s->hdr.count, ARCH_CONVERT, -(count));
-	INT_MOD(leaf_s->hdr.stale, ARCH_CONVERT, -(stale));
-	INT_MOD(leaf_d->hdr.count, ARCH_CONVERT, count);
-	INT_MOD(leaf_d->hdr.stale, ARCH_CONVERT, stale);
+	be16_add(&leaf_s->hdr.count, -(count));
+	be16_add(&leaf_s->hdr.stale, -(stale));
+	be16_add(&leaf_d->hdr.count, count);
+	be16_add(&leaf_d->hdr.stale, stale);
 	xfs_dir2_leaf_log_header(tp, bp_s);
 	xfs_dir2_leaf_log_header(tp, bp_d);
 	xfs_dir2_leafn_check(args->dp, bp_s);
@@ -719,13 +720,13 @@ xfs_dir2_leafn_order(
 
 	leaf1 = leaf1_bp->data;
 	leaf2 = leaf2_bp->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	if (INT_GET(leaf1->hdr.count, ARCH_CONVERT) > 0 &&
-	    INT_GET(leaf2->hdr.count, ARCH_CONVERT) > 0 &&
-	    (INT_GET(leaf2->ents[0].hashval, ARCH_CONVERT) < INT_GET(leaf1->ents[0].hashval, ARCH_CONVERT) ||
-	     INT_GET(leaf2->ents[INT_GET(leaf2->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT) <
-	     INT_GET(leaf1->ents[INT_GET(leaf1->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT)))
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	if (be16_to_cpu(leaf1->hdr.count) > 0 &&
+	    be16_to_cpu(leaf2->hdr.count) > 0 &&
+	    (be32_to_cpu(leaf2->ents[0].hashval) < be32_to_cpu(leaf1->ents[0].hashval) ||
+	     be32_to_cpu(leaf2->ents[be16_to_cpu(leaf2->hdr.count) - 1].hashval) <
+	     be32_to_cpu(leaf1->ents[be16_to_cpu(leaf1->hdr.count) - 1].hashval)))
 		return 1;
 	return 0;
 }
@@ -768,9 +769,9 @@ xfs_dir2_leafn_rebalance(
 	}
 	leaf1 = blk1->bp->data;
 	leaf2 = blk2->bp->data;
-	oldsum = INT_GET(leaf1->hdr.count, ARCH_CONVERT) + INT_GET(leaf2->hdr.count, ARCH_CONVERT);
+	oldsum = be16_to_cpu(leaf1->hdr.count) + be16_to_cpu(leaf2->hdr.count);
 #ifdef DEBUG
-	oldstale = INT_GET(leaf1->hdr.stale, ARCH_CONVERT) + INT_GET(leaf2->hdr.stale, ARCH_CONVERT);
+	oldstale = be16_to_cpu(leaf1->hdr.stale) + be16_to_cpu(leaf2->hdr.stale);
 #endif
 	mid = oldsum >> 1;
 	/*
@@ -780,10 +781,10 @@ xfs_dir2_leafn_rebalance(
 	if (oldsum & 1) {
 		xfs_dahash_t	midhash;	/* middle entry hash value */
 
-		if (mid >= INT_GET(leaf1->hdr.count, ARCH_CONVERT))
-			midhash = INT_GET(leaf2->ents[mid - INT_GET(leaf1->hdr.count, ARCH_CONVERT)].hashval, ARCH_CONVERT);
+		if (mid >= be16_to_cpu(leaf1->hdr.count))
+			midhash = be32_to_cpu(leaf2->ents[mid - be16_to_cpu(leaf1->hdr.count)].hashval);
 		else
-			midhash = INT_GET(leaf1->ents[mid].hashval, ARCH_CONVERT);
+			midhash = be32_to_cpu(leaf1->ents[mid].hashval);
 		isleft = args->hashval <= midhash;
 	}
 	/*
@@ -797,30 +798,30 @@ xfs_dir2_leafn_rebalance(
 	 * Calculate moved entry count.  Positive means left-to-right,
 	 * negative means right-to-left.  Then move the entries.
 	 */
-	count = INT_GET(leaf1->hdr.count, ARCH_CONVERT) - mid + (isleft == 0);
+	count = be16_to_cpu(leaf1->hdr.count) - mid + (isleft == 0);
 	if (count > 0)
 		xfs_dir2_leafn_moveents(args, blk1->bp,
-			INT_GET(leaf1->hdr.count, ARCH_CONVERT) - count, blk2->bp, 0, count);
+			be16_to_cpu(leaf1->hdr.count) - count, blk2->bp, 0, count);
 	else if (count < 0)
 		xfs_dir2_leafn_moveents(args, blk2->bp, 0, blk1->bp,
-			INT_GET(leaf1->hdr.count, ARCH_CONVERT), count);
-	ASSERT(INT_GET(leaf1->hdr.count, ARCH_CONVERT) + INT_GET(leaf2->hdr.count, ARCH_CONVERT) == oldsum);
-	ASSERT(INT_GET(leaf1->hdr.stale, ARCH_CONVERT) + INT_GET(leaf2->hdr.stale, ARCH_CONVERT) == oldstale);
+			be16_to_cpu(leaf1->hdr.count), count);
+	ASSERT(be16_to_cpu(leaf1->hdr.count) + be16_to_cpu(leaf2->hdr.count) == oldsum);
+	ASSERT(be16_to_cpu(leaf1->hdr.stale) + be16_to_cpu(leaf2->hdr.stale) == oldstale);
 	/*
 	 * Mark whether we're inserting into the old or new leaf.
 	 */
-	if (INT_GET(leaf1->hdr.count, ARCH_CONVERT) < INT_GET(leaf2->hdr.count, ARCH_CONVERT))
+	if (be16_to_cpu(leaf1->hdr.count) < be16_to_cpu(leaf2->hdr.count))
 		state->inleaf = swap;
-	else if (INT_GET(leaf1->hdr.count, ARCH_CONVERT) > INT_GET(leaf2->hdr.count, ARCH_CONVERT))
+	else if (be16_to_cpu(leaf1->hdr.count) > be16_to_cpu(leaf2->hdr.count))
 		state->inleaf = !swap;
 	else
 		state->inleaf =
-			swap ^ (blk1->index <= INT_GET(leaf1->hdr.count, ARCH_CONVERT));
+			swap ^ (blk1->index <= be16_to_cpu(leaf1->hdr.count));
 	/*
 	 * Adjust the expected index for insertion.
 	 */
 	if (!state->inleaf)
-		blk2->index = blk1->index - INT_GET(leaf1->hdr.count, ARCH_CONVERT);
+		blk2->index = blk1->index - be16_to_cpu(leaf1->hdr.count);
 	
 	/* 
 	 * Finally sanity check just to make sure we are not returning a negative index 
@@ -829,7 +830,7 @@ xfs_dir2_leafn_rebalance(
 		state->inleaf = 1;
 		blk2->index = 0;
 		cmn_err(CE_ALERT,
-			"xfs_dir2_leafn_rebalance: picked the wrong leaf? reverting orignal leaf: "
+			"xfs_dir2_leafn_rebalance: picked the wrong leaf? reverting original leaf: "
 			"blk1->index %d\n",
 			blk1->index);
 	}
@@ -867,7 +868,7 @@ xfs_dir2_leafn_remove(
 	tp = args->trans;
 	mp = dp->i_mount;
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * Point to the entry we're removing.
 	 */
@@ -875,17 +876,17 @@ xfs_dir2_leafn_remove(
 	/*
 	 * Extract the data block and offset from the entry.
 	 */
-	db = XFS_DIR2_DATAPTR_TO_DB(mp, INT_GET(lep->address, ARCH_CONVERT));
+	db = XFS_DIR2_DATAPTR_TO_DB(mp, be32_to_cpu(lep->address));
 	ASSERT(dblk->blkno == db);
-	off = XFS_DIR2_DATAPTR_TO_OFF(mp, INT_GET(lep->address, ARCH_CONVERT));
+	off = XFS_DIR2_DATAPTR_TO_OFF(mp, be32_to_cpu(lep->address));
 	ASSERT(dblk->index == off);
 	/*
 	 * Kill the leaf entry by marking it stale.
 	 * Log the leaf block changes.
 	 */
-	INT_MOD(leaf->hdr.stale, ARCH_CONVERT, +1);
+	be16_add(&leaf->hdr.stale, 1);
 	xfs_dir2_leaf_log_header(tp, bp);
-	INT_SET(lep->address, ARCH_CONVERT, XFS_DIR2_NULL_DATAPTR);
+	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
 	xfs_dir2_leaf_log_ents(tp, bp, index, index);
 	/*
 	 * Make the data entry free.  Keep track of the longest freespace
@@ -894,7 +895,7 @@ xfs_dir2_leafn_remove(
 	dbp = dblk->bp;
 	data = dbp->data;
 	dep = (xfs_dir2_data_entry_t *)((char *)data + off);
-	longest = INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT);
+	longest = be16_to_cpu(data->hdr.bestfree[0].length);
 	needlog = needscan = 0;
 	xfs_dir2_data_make_free(tp, dbp, off,
 		XFS_DIR2_DATA_ENTSIZE(dep->namelen), &needlog, &needscan);
@@ -911,7 +912,7 @@ xfs_dir2_leafn_remove(
 	 * If the longest data block freespace changes, need to update
 	 * the corresponding freeblock entry.
 	 */
-	if (longest < INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT)) {
+	if (longest < be16_to_cpu(data->hdr.bestfree[0].length)) {
 		int		error;		/* error return value */
 		xfs_dabuf_t	*fbp;		/* freeblock buffer */
 		xfs_dir2_db_t	fdb;		/* freeblock block number */
@@ -929,15 +930,15 @@ xfs_dir2_leafn_remove(
 			return error;
 		}
 		free = fbp->data;
-		ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
-		ASSERT(INT_GET(free->hdr.firstdb, ARCH_CONVERT) ==
+		ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
+		ASSERT(be32_to_cpu(free->hdr.firstdb) ==
 		       XFS_DIR2_MAX_FREE_BESTS(mp) *
 		       (fdb - XFS_DIR2_FREE_FIRSTDB(mp)));
 		/*
 		 * Calculate which entry we need to fix.
 		 */
 		findex = XFS_DIR2_DB_TO_FDINDEX(mp, db);
-		longest = INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT);
+		longest = be16_to_cpu(data->hdr.bestfree[0].length);
 		/*
 		 * If the data block is now empty we can get rid of it
 		 * (usually).
@@ -969,7 +970,7 @@ xfs_dir2_leafn_remove(
 			/*
 			 * One less used entry in the free table.
 			 */
-			INT_MOD(free->hdr.nused, ARCH_CONVERT, -1);
+			free->hdr.nused = cpu_to_be32(-1);
 			xfs_dir2_free_log_header(tp, fbp);
 			/*
 			 * If this was the last entry in the table, we can
@@ -977,21 +978,21 @@ xfs_dir2_leafn_remove(
 			 * entries at the end referring to non-existent
 			 * data blocks, get those too.
 			 */
-			if (findex == INT_GET(free->hdr.nvalid, ARCH_CONVERT) - 1) {
+			if (findex == be32_to_cpu(free->hdr.nvalid) - 1) {
 				int	i;		/* free entry index */
 
 				for (i = findex - 1;
-				     i >= 0 && INT_GET(free->bests[i], ARCH_CONVERT) == NULLDATAOFF;
+				     i >= 0 && be16_to_cpu(free->bests[i]) == NULLDATAOFF;
 				     i--)
 					continue;
-				INT_SET(free->hdr.nvalid, ARCH_CONVERT, i + 1);
+				free->hdr.nvalid = cpu_to_be32(i + 1);
 				logfree = 0;
 			}
 			/*
 			 * Not the last entry, just punch it out.
 			 */
 			else {
-				INT_SET(free->bests[findex], ARCH_CONVERT, NULLDATAOFF);
+				free->bests[findex] = cpu_to_be16(NULLDATAOFF);
 				logfree = 1;
 			}
 			/*
@@ -1017,7 +1018,7 @@ xfs_dir2_leafn_remove(
 		 * the new value.
 		 */
 		else {
-			INT_SET(free->bests[findex], ARCH_CONVERT, longest);
+			free->bests[findex] = cpu_to_be16(longest);
 			logfree = 1;
 		}
 		/*
@@ -1039,7 +1040,7 @@ xfs_dir2_leafn_remove(
 	*rval =
 		((uint)sizeof(leaf->hdr) +
 		 (uint)sizeof(leaf->ents[0]) *
-		 (INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT))) <
+		 (be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale))) <
 		mp->m_dir_magicpct;
 	return 0;
 }
@@ -1138,9 +1139,9 @@ xfs_dir2_leafn_toosmall(
 	 */
 	blk = &state->path.blk[state->path.active - 1];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_DIR2_LEAFN_MAGIC);
 	leaf = (xfs_dir2_leaf_t *)info;
-	count = INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT);
+	count = be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale);
 	bytes = (uint)sizeof(leaf->hdr) + count * (uint)sizeof(leaf->ents[0]);
 	if (bytes > (state->blocksize >> 1)) {
 		/*
@@ -1160,7 +1161,7 @@ xfs_dir2_leafn_toosmall(
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward, 0,
 			&rval);
@@ -1176,9 +1177,9 @@ xfs_dir2_leafn_toosmall(
 	 * We prefer coalescing with the lower numbered sibling so as
 	 * to shrink a directory over time.
 	 */
-	forward = INT_GET(info->forw, ARCH_CONVERT) < INT_GET(info->back, ARCH_CONVERT);
+	forward = be32_to_cpu(info->forw) < be32_to_cpu(info->back);
 	for (i = 0, bp = NULL; i < 2; forward = !forward, i++) {
-		blkno = forward ?INT_GET( info->forw, ARCH_CONVERT) : INT_GET(info->back, ARCH_CONVERT);
+		blkno = forward ? be32_to_cpu(info->forw) : be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		/*
@@ -1194,11 +1195,11 @@ xfs_dir2_leafn_toosmall(
 		 * Count bytes in the two blocks combined.
 		 */
 		leaf = (xfs_dir2_leaf_t *)info;
-		count = INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT);
+		count = be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale);
 		bytes = state->blocksize - (state->blocksize >> 2);
 		leaf = bp->data;
-		ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-		count += INT_GET(leaf->hdr.count, ARCH_CONVERT) - INT_GET(leaf->hdr.stale, ARCH_CONVERT);
+		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+		count += be16_to_cpu(leaf->hdr.count) - be16_to_cpu(leaf->hdr.stale);
 		bytes -= count * (uint)sizeof(leaf->ents[0]);
 		/*
 		 * Fits with at least 25% to spare.
@@ -1256,27 +1257,27 @@ xfs_dir2_leafn_unbalance(
 	ASSERT(save_blk->magic == XFS_DIR2_LEAFN_MAGIC);
 	drop_leaf = drop_blk->bp->data;
 	save_leaf = save_blk->bp->data;
-	ASSERT(INT_GET(drop_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(INT_GET(save_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
+	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_DIR2_LEAFN_MAGIC);
 	/*
 	 * If there are any stale leaf entries, take this opportunity
 	 * to purge them.
 	 */
-	if (INT_GET(drop_leaf->hdr.stale, ARCH_CONVERT))
+	if (drop_leaf->hdr.stale)
 		xfs_dir2_leaf_compact(args, drop_blk->bp);
-	if (INT_GET(save_leaf->hdr.stale, ARCH_CONVERT))
+	if (save_leaf->hdr.stale)
 		xfs_dir2_leaf_compact(args, save_blk->bp);
 	/*
 	 * Move the entries from drop to the appropriate end of save.
 	 */
-	drop_blk->hashval = INT_GET(drop_leaf->ents[INT_GET(drop_leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+	drop_blk->hashval = be32_to_cpu(drop_leaf->ents[be16_to_cpu(drop_leaf->hdr.count) - 1].hashval);
 	if (xfs_dir2_leafn_order(save_blk->bp, drop_blk->bp))
 		xfs_dir2_leafn_moveents(args, drop_blk->bp, 0, save_blk->bp, 0,
-			INT_GET(drop_leaf->hdr.count, ARCH_CONVERT));
+			be16_to_cpu(drop_leaf->hdr.count));
 	else
 		xfs_dir2_leafn_moveents(args, drop_blk->bp, 0, save_blk->bp,
-			INT_GET(save_leaf->hdr.count, ARCH_CONVERT), INT_GET(drop_leaf->hdr.count, ARCH_CONVERT));
-	save_blk->hashval = INT_GET(save_leaf->ents[INT_GET(save_leaf->hdr.count, ARCH_CONVERT) - 1].hashval, ARCH_CONVERT);
+			be16_to_cpu(save_leaf->hdr.count), be16_to_cpu(drop_leaf->hdr.count));
+	save_blk->hashval = be32_to_cpu(save_leaf->ents[be16_to_cpu(save_leaf->hdr.count) - 1].hashval);
 	xfs_dir2_leafn_check(args->dp, save_blk->bp);
 }
 
@@ -1378,7 +1379,7 @@ xfs_dir2_node_addname_int(
 	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			needlog;	/* need to log data header */
 	int			needscan;	/* need to rescan data frees */
-	xfs_dir2_data_off_t	*tagp;		/* data entry tag pointer */
+	__be16			*tagp;		/* data entry tag pointer */
 	xfs_trans_t		*tp;		/* transaction pointer */
 
 	dp = args->dp;
@@ -1397,7 +1398,7 @@ xfs_dir2_node_addname_int(
 		 */
 		ifbno = fblk->blkno;
 		free = fbp->data;
-		ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+		ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 		findex = fblk->index;
 		/*
 		 * This means the free entry showed that the data block had
@@ -1405,10 +1406,10 @@ xfs_dir2_node_addname_int(
 		 * Use that data block.
 		 */
 		if (findex >= 0) {
-			ASSERT(findex < INT_GET(free->hdr.nvalid, ARCH_CONVERT));
-			ASSERT(INT_GET(free->bests[findex], ARCH_CONVERT) != NULLDATAOFF);
-			ASSERT(INT_GET(free->bests[findex], ARCH_CONVERT) >= length);
-			dbno = INT_GET(free->hdr.firstdb, ARCH_CONVERT) + findex;
+			ASSERT(findex < be32_to_cpu(free->hdr.nvalid));
+			ASSERT(be16_to_cpu(free->bests[findex]) != NULLDATAOFF);
+			ASSERT(be16_to_cpu(free->bests[findex]) >= length);
+			dbno = be32_to_cpu(free->hdr.firstdb) + findex;
 		}
 		/*
 		 * The data block looked at didn't have enough room.
@@ -1481,20 +1482,20 @@ xfs_dir2_node_addname_int(
 				continue;
 			}
 			free = fbp->data;
-			ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+			ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 			findex = 0;
 		}
 		/*
 		 * Look at the current free entry.  Is it good enough?
 		 */
-		if (INT_GET(free->bests[findex], ARCH_CONVERT) != NULLDATAOFF &&
-		    INT_GET(free->bests[findex], ARCH_CONVERT) >= length)
-			dbno = INT_GET(free->hdr.firstdb, ARCH_CONVERT) + findex;
+		if (be16_to_cpu(free->bests[findex]) != NULLDATAOFF &&
+		    be16_to_cpu(free->bests[findex]) >= length)
+			dbno = be32_to_cpu(free->hdr.firstdb) + findex;
 		else {
 			/*
 			 * Are we done with the freeblock?
 			 */
-			if (++findex == INT_GET(free->hdr.nvalid, ARCH_CONVERT)) {
+			if (++findex == be32_to_cpu(free->hdr.nvalid)) {
 				/*
 				 * Drop the block.
 				 */
@@ -1608,15 +1609,15 @@ xfs_dir2_node_addname_int(
 			 * its first slot as our empty slot.
 			 */
 			free = fbp->data;
-			INT_SET(free->hdr.magic, ARCH_CONVERT, XFS_DIR2_FREE_MAGIC);
-			INT_SET(free->hdr.firstdb, ARCH_CONVERT,
+			free->hdr.magic = cpu_to_be32(XFS_DIR2_FREE_MAGIC);
+			free->hdr.firstdb = cpu_to_be32(
 				(fbno - XFS_DIR2_FREE_FIRSTDB(mp)) *
 				XFS_DIR2_MAX_FREE_BESTS(mp));
 			free->hdr.nvalid = 0;
 			free->hdr.nused = 0;
 		} else {
 			free = fbp->data;
-			ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+			ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 		}
 
 		/*
@@ -1627,20 +1628,20 @@ xfs_dir2_node_addname_int(
 		 * If it's after the end of the current entries in the
 		 * freespace block, extend that table.
 		 */
-		if (findex >= INT_GET(free->hdr.nvalid, ARCH_CONVERT)) {
+		if (findex >= be32_to_cpu(free->hdr.nvalid)) {
 			ASSERT(findex < XFS_DIR2_MAX_FREE_BESTS(mp));
-			INT_SET(free->hdr.nvalid, ARCH_CONVERT, findex + 1);
+			free->hdr.nvalid = cpu_to_be32(findex + 1);
 			/*
 			 * Tag new entry so nused will go up.
 			 */
-			INT_SET(free->bests[findex], ARCH_CONVERT, NULLDATAOFF);
+			free->bests[findex] = cpu_to_be16(NULLDATAOFF);
 		}
 		/*
 		 * If this entry was for an empty data block
 		 * (this should always be true) then update the header.
 		 */
-		if (INT_GET(free->bests[findex], ARCH_CONVERT) == NULLDATAOFF) {
-			INT_MOD(free->hdr.nused, ARCH_CONVERT, +1);
+		if (be16_to_cpu(free->bests[findex]) == NULLDATAOFF) {
+			be32_add(&free->hdr.nused, 1);
 			xfs_dir2_free_log_header(tp, fbp);
 		}
 		/*
@@ -1649,7 +1650,7 @@ xfs_dir2_node_addname_int(
 		 * change again.
 		 */
 		data = dbp->data;
-		INT_COPY(free->bests[findex], data->hdr.bestfree[0].length, ARCH_CONVERT);
+		free->bests[findex] = data->hdr.bestfree[0].length;
 		logfree = 1;
 	}
 	/*
@@ -1677,12 +1678,12 @@ xfs_dir2_node_addname_int(
 		data = dbp->data;
 		logfree = 0;
 	}
-	ASSERT(INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT) >= length);
+	ASSERT(be16_to_cpu(data->hdr.bestfree[0].length) >= length);
 	/*
 	 * Point to the existing unused space.
 	 */
 	dup = (xfs_dir2_data_unused_t *)
-	      ((char *)data + INT_GET(data->hdr.bestfree[0].offset, ARCH_CONVERT));
+	      ((char *)data + be16_to_cpu(data->hdr.bestfree[0].offset));
 	needscan = needlog = 0;
 	/*
 	 * Mark the first part of the unused space, inuse for us.
@@ -1698,7 +1699,7 @@ xfs_dir2_node_addname_int(
 	dep->namelen = args->namelen;
 	memcpy(dep->name, args->name, dep->namelen);
 	tagp = XFS_DIR2_DATA_ENTRY_TAG_P(dep);
-	INT_SET(*tagp, ARCH_CONVERT, (xfs_dir2_data_off_t)((char *)dep - (char *)data));
+	*tagp = cpu_to_be16((char *)dep - (char *)data);
 	xfs_dir2_data_log_entry(tp, dbp, dep);
 	/*
 	 * Rescan the block for bestfree if needed.
@@ -1713,8 +1714,8 @@ xfs_dir2_node_addname_int(
 	/*
 	 * If the freespace entry is now wrong, update it.
 	 */
-	if (INT_GET(free->bests[findex], ARCH_CONVERT) != INT_GET(data->hdr.bestfree[0].length, ARCH_CONVERT)) {
-		INT_COPY(free->bests[findex], data->hdr.bestfree[0].length, ARCH_CONVERT);
+	if (be16_to_cpu(free->bests[findex]) != be16_to_cpu(data->hdr.bestfree[0].length)) {
+		free->bests[findex] = data->hdr.bestfree[0].length;
 		logfree = 1;
 	}
 	/*
@@ -1731,7 +1732,7 @@ xfs_dir2_node_addname_int(
 	 * Return the data block and offset in args, then drop the data block.
 	 */
 	args->blkno = (xfs_dablk_t)dbno;
-	args->index = INT_GET(*tagp, ARCH_CONVERT);
+	args->index = be16_to_cpu(*tagp);
 	xfs_da_buf_done(dbp);
 	return 0;
 }
@@ -1900,10 +1901,10 @@ xfs_dir2_node_replace(
 		 * Point to the data entry.
 		 */
 		data = state->extrablk.bp->data;
-		ASSERT(INT_GET(data->hdr.magic, ARCH_CONVERT) == XFS_DIR2_DATA_MAGIC);
+		ASSERT(be32_to_cpu(data->hdr.magic) == XFS_DIR2_DATA_MAGIC);
 		dep = (xfs_dir2_data_entry_t *)
 		      ((char *)data +
-		       XFS_DIR2_DATAPTR_TO_OFF(state->mp, INT_GET(lep->address, ARCH_CONVERT)));
+		       XFS_DIR2_DATAPTR_TO_OFF(state->mp, be32_to_cpu(lep->address)));
 		ASSERT(inum != INT_GET(dep->inumber, ARCH_CONVERT));
 		/*
 		 * Fill in the new inode number and log the entry.
@@ -1966,11 +1967,11 @@ xfs_dir2_node_trim_free(
 		return 0;
 	}
 	free = bp->data;
-	ASSERT(INT_GET(free->hdr.magic, ARCH_CONVERT) == XFS_DIR2_FREE_MAGIC);
+	ASSERT(be32_to_cpu(free->hdr.magic) == XFS_DIR2_FREE_MAGIC);
 	/*
 	 * If there are used entries, there's nothing to do.
 	 */
-	if (INT_GET(free->hdr.nused, ARCH_CONVERT) > 0) {
+	if (be32_to_cpu(free->hdr.nused) > 0) {
 		xfs_da_brelse(tp, bp);
 		*rvalp = 0;
 		return 0;
diff --git a/fs/xfs/xfs_dir2_node.h b/fs/xfs/xfs_dir2_node.h
index 0ab8fbd59512..c7c870ee7857 100644
--- a/fs/xfs/xfs_dir2_node.h
+++ b/fs/xfs/xfs_dir2_node.h
@@ -41,15 +41,15 @@ struct xfs_trans;
 #define	XFS_DIR2_FREE_MAGIC	0x58443246	/* XD2F */
 
 typedef	struct xfs_dir2_free_hdr {
-	__uint32_t		magic;		/* XFS_DIR2_FREE_MAGIC */
-	__int32_t		firstdb;	/* db of first entry */
-	__int32_t		nvalid;		/* count of valid entries */
-	__int32_t		nused;		/* count of used entries */
+	__be32			magic;		/* XFS_DIR2_FREE_MAGIC */
+	__be32			firstdb;	/* db of first entry */
+	__be32			nvalid;		/* count of valid entries */
+	__be32			nused;		/* count of used entries */
 } xfs_dir2_free_hdr_t;
 
 typedef struct xfs_dir2_free {
 	xfs_dir2_free_hdr_t	hdr;		/* block header */
-	xfs_dir2_data_off_t	bests[1];	/* best free counts */
+	__be16			bests[1];	/* best free counts */
 						/* unused entries are -1 */
 } xfs_dir2_free_t;
 
diff --git a/fs/xfs/xfs_dir2_sf.c b/fs/xfs/xfs_dir2_sf.c
index ec8e7476c8b7..d98a41d1fe63 100644
--- a/fs/xfs/xfs_dir2_sf.c
+++ b/fs/xfs/xfs_dir2_sf.c
@@ -98,8 +98,8 @@ xfs_dir2_block_sfsize(
 	/*
 	 * Iterate over the block's data entries by using the leaf pointers.
 	 */
-	for (i = 0; i < INT_GET(btp->count, ARCH_CONVERT); i++) {
-		if ((addr = INT_GET(blp[i].address, ARCH_CONVERT)) == XFS_DIR2_NULL_DATAPTR)
+	for (i = 0; i < be32_to_cpu(btp->count); i++) {
+		if ((addr = be32_to_cpu(blp[i].address)) == XFS_DIR2_NULL_DATAPTR)
 			continue;
 		/*
 		 * Calculate the pointer to the entry at hand.
@@ -220,8 +220,8 @@ xfs_dir2_block_to_sf(
 		 * If it's unused, just skip over it.
 		 */
 		dup = (xfs_dir2_data_unused_t *)ptr;
-		if (INT_GET(dup->freetag, ARCH_CONVERT) == XFS_DIR2_DATA_FREE_TAG) {
-			ptr += INT_GET(dup->length, ARCH_CONVERT);
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			ptr += be16_to_cpu(dup->length);
 			continue;
 		}
 		dep = (xfs_dir2_data_entry_t *)ptr;
diff --git a/fs/xfs/xfs_dir_leaf.c b/fs/xfs/xfs_dir_leaf.c
index e83074016abb..6d711869262f 100644
--- a/fs/xfs/xfs_dir_leaf.c
+++ b/fs/xfs/xfs_dir_leaf.c
@@ -176,7 +176,7 @@ xfs_dir_shortform_addname(xfs_da_args_t *args)
 	ASSERT(dp->i_df.if_u1.if_data != NULL);
 	sf = (xfs_dir_shortform_t *)dp->i_df.if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		if (sfe->namelen == args->namelen &&
 		    args->name[0] == sfe->name[0] &&
 		    memcmp(args->name, sfe->name, args->namelen) == 0)
@@ -193,7 +193,7 @@ xfs_dir_shortform_addname(xfs_da_args_t *args)
 	XFS_DIR_SF_PUT_DIRINO(&args->inumber, &sfe->inumber);
 	sfe->namelen = args->namelen;
 	memcpy(sfe->name, args->name, sfe->namelen);
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, +1);
+	sf->hdr.count++;
 
 	dp->i_d.di_size += size;
 	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_DDATA);
@@ -227,7 +227,7 @@ xfs_dir_shortform_removename(xfs_da_args_t *args)
 	base = sizeof(xfs_dir_sf_hdr_t);
 	sf = (xfs_dir_shortform_t *)dp->i_df.if_u1.if_data;
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		size = XFS_DIR_SF_ENTSIZE_BYENTRY(sfe);
 		if (sfe->namelen == args->namelen &&
 		    sfe->name[0] == args->name[0] &&
@@ -245,7 +245,7 @@ xfs_dir_shortform_removename(xfs_da_args_t *args)
 		memmove(&((char *)sf)[base], &((char *)sf)[base+size],
 					      dp->i_d.di_size - (base+size));
 	}
-	INT_MOD(sf->hdr.count, ARCH_CONVERT, -1);
+	sf->hdr.count--;
 
 	xfs_idata_realloc(dp, -size, XFS_DATA_FORK);
 	dp->i_d.di_size -= size;
@@ -288,7 +288,7 @@ xfs_dir_shortform_lookup(xfs_da_args_t *args)
 		return(XFS_ERROR(EEXIST));
 	}
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		if (sfe->namelen == args->namelen &&
 		    sfe->name[0] == args->name[0] &&
 		    memcmp(args->name, sfe->name, args->namelen) == 0) {
@@ -375,7 +375,7 @@ xfs_dir_shortform_to_leaf(xfs_da_args_t *iargs)
 		goto out;
 
 	sfe = &sf->list[0];
-	for (i = 0; i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0; i < sf->hdr.count; i++) {
 		args.name = (char *)(sfe->name);
 		args.namelen = sfe->namelen;
 		args.hashval = xfs_da_hashname((char *)(sfe->name),
@@ -428,7 +428,7 @@ xfs_dir_shortform_getdents(xfs_inode_t *dp, uio_t *uio, int *eofp,
 	sf = (xfs_dir_shortform_t *)dp->i_df.if_u1.if_data;
 	cookhash = XFS_DA_COOKIE_HASH(mp, uio->uio_offset);
 	want_entno = XFS_DA_COOKIE_ENTRY(mp, uio->uio_offset);
-	nsbuf = INT_GET(sf->hdr.count, ARCH_CONVERT) + 2;
+	nsbuf = sf->hdr.count + 2;
 	sbsize = (nsbuf + 1) * sizeof(*sbuf);
 	sbp = sbuf = kmem_alloc(sbsize, KM_SLEEP);
 
@@ -460,8 +460,7 @@ xfs_dir_shortform_getdents(xfs_inode_t *dp, uio_t *uio, int *eofp,
 	/*
 	 * Scan the directory data for the rest of the entries.
 	 */
-	for (i = 0, sfe = &sf->list[0];
-			i < INT_GET(sf->hdr.count, ARCH_CONVERT); i++) {
+	for (i = 0, sfe = &sf->list[0]; i < sf->hdr.count; i++) {
 
 		if (unlikely(
 		    ((char *)sfe < (char *)sf) ||
@@ -600,7 +599,7 @@ xfs_dir_shortform_replace(xfs_da_args_t *args)
 	}
 	ASSERT(args->namelen != 1 || args->name[0] != '.');
 	sfe = &sf->list[0];
-	for (i = INT_GET(sf->hdr.count, ARCH_CONVERT)-1; i >= 0; i--) {
+	for (i = sf->hdr.count-1; i >= 0; i--) {
 		if (sfe->namelen == args->namelen &&
 		    sfe->name[0] == args->name[0] &&
 		    memcmp(args->name, sfe->name, args->namelen) == 0) {
@@ -644,7 +643,7 @@ xfs_dir_leaf_to_shortform(xfs_da_args_t *iargs)
 	ASSERT(bp != NULL);
 	memcpy(tmpbuffer, bp->data, XFS_LBSIZE(dp->i_mount));
 	leaf = (xfs_dir_leafblock_t *)tmpbuffer;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	memset(bp->data, 0, XFS_LBSIZE(dp->i_mount));
 
 	/*
@@ -742,11 +741,13 @@ xfs_dir_leaf_to_node(xfs_da_args_t *args)
 	}
 	node = bp1->data;
 	leaf = bp2->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	INT_SET(node->btree[0].hashval, ARCH_CONVERT, INT_GET(leaf->entries[ INT_GET(leaf->hdr.count, ARCH_CONVERT)-1 ].hashval, ARCH_CONVERT));
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	node->btree[0].hashval = cpu_to_be32(
+		INT_GET(leaf->entries[
+			INT_GET(leaf->hdr.count, ARCH_CONVERT)-1].hashval, ARCH_CONVERT));
 	xfs_da_buf_done(bp2);
-	INT_SET(node->btree[0].before, ARCH_CONVERT, blkno);
-	INT_SET(node->hdr.count, ARCH_CONVERT, 1);
+	node->btree[0].before = cpu_to_be32(blkno);
+	node->hdr.count = cpu_to_be16(1);
 	xfs_da_log_buf(args->trans, bp1,
 		XFS_DA_LOGRANGE(node, &node->btree[0], sizeof(node->btree[0])));
 	xfs_da_buf_done(bp1);
@@ -781,7 +782,7 @@ xfs_dir_leaf_create(xfs_da_args_t *args, xfs_dablk_t blkno, xfs_dabuf_t **bpp)
 	leaf = bp->data;
 	memset((char *)leaf, 0, XFS_LBSIZE(dp->i_mount));
 	hdr = &leaf->hdr;
-	INT_SET(hdr->info.magic, ARCH_CONVERT, XFS_DIR_LEAF_MAGIC);
+	hdr->info.magic = cpu_to_be16(XFS_DIR_LEAF_MAGIC);
 	INT_SET(hdr->firstused, ARCH_CONVERT, XFS_LBSIZE(dp->i_mount));
 	if (!hdr->firstused)
 		INT_SET(hdr->firstused, ARCH_CONVERT, XFS_LBSIZE(dp->i_mount) - 1);
@@ -860,7 +861,7 @@ xfs_dir_leaf_add(xfs_dabuf_t *bp, xfs_da_args_t *args, int index)
 	int tablesize, entsize, sum, i, tmp, error;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	ASSERT((index >= 0) && (index <= INT_GET(leaf->hdr.count, ARCH_CONVERT)));
 	hdr = &leaf->hdr;
 	entsize = XFS_DIR_LEAF_ENTSIZE_BYNAME(args->namelen);
@@ -940,7 +941,7 @@ xfs_dir_leaf_add_work(xfs_dabuf_t *bp, xfs_da_args_t *args, int index,
 	int tmp, i;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	ASSERT((mapindex >= 0) && (mapindex < XFS_DIR_LEAF_MAPSIZE));
 	ASSERT((index >= 0) && (index <= INT_GET(hdr->count, ARCH_CONVERT)));
@@ -1097,8 +1098,8 @@ xfs_dir_leaf_rebalance(xfs_da_state_t *state, xfs_da_state_blk_t *blk1,
 	ASSERT(blk2->magic == XFS_DIR_LEAF_MAGIC);
 	leaf1 = blk1->bp->data;
 	leaf2 = blk2->bp->data;
-	ASSERT(INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf1->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf2->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 
 	/*
 	 * Check ordering of blocks, reverse if it makes things simpler.
@@ -1325,7 +1326,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 */
 	blk = &state->path.blk[ state->path.active-1 ];
 	info = blk->bp->data;
-	ASSERT(INT_GET(info->magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(info->magic) == XFS_DIR_LEAF_MAGIC);
 	leaf = (xfs_dir_leafblock_t *)info;
 	count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
 	bytes = (uint)sizeof(xfs_dir_leaf_hdr_t) +
@@ -1340,7 +1341,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 	/*
 	 * Check for the degenerate case of the block being empty.
 	 * If the block is empty, we'll simply delete it, no need to
-	 * coalesce it with a sibling block.  We choose (aribtrarily)
+	 * coalesce it with a sibling block.  We choose (arbitrarily)
 	 * to merge with the forward block unless it is NULL.
 	 */
 	if (count == 0) {
@@ -1348,7 +1349,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 		 * Make altpath point to the block we want to keep and
 		 * path point to the block we want to drop (this one).
 		 */
-		forward = info->forw;
+		forward = (info->forw != 0);
 		memcpy(&state->altpath, &state->path, sizeof(state->path));
 		error = xfs_da_path_shift(state, &state->altpath, forward,
 						 0, &retval);
@@ -1369,12 +1370,12 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 	 * We prefer coalescing with the lower numbered sibling so as
 	 * to shrink a directory over time.
 	 */
-	forward = (INT_GET(info->forw, ARCH_CONVERT) < INT_GET(info->back, ARCH_CONVERT));	/* start with smaller blk num */
+	forward = (be32_to_cpu(info->forw) < be32_to_cpu(info->back));	/* start with smaller blk num */
 	for (i = 0; i < 2; forward = !forward, i++) {
 		if (forward)
-			blkno = INT_GET(info->forw, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->forw);
 		else
-			blkno = INT_GET(info->back, ARCH_CONVERT);
+			blkno = be32_to_cpu(info->back);
 		if (blkno == 0)
 			continue;
 		error = xfs_da_read_buf(state->args->trans, state->args->dp,
@@ -1389,7 +1390,7 @@ xfs_dir_leaf_toosmall(xfs_da_state_t *state, int *action)
 		bytes  = state->blocksize - (state->blocksize>>2);
 		bytes -= INT_GET(leaf->hdr.namebytes, ARCH_CONVERT);
 		leaf = bp->data;
-		ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+		ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 		count += INT_GET(leaf->hdr.count, ARCH_CONVERT);
 		bytes -= INT_GET(leaf->hdr.namebytes, ARCH_CONVERT);
 		bytes -= count * ((uint)sizeof(xfs_dir_leaf_name_t) - 1);
@@ -1447,7 +1448,7 @@ xfs_dir_leaf_remove(xfs_trans_t *trans, xfs_dabuf_t *bp, int index)
 	xfs_mount_t *mp;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	hdr = &leaf->hdr;
 	mp = trans->t_mountp;
 	ASSERT((INT_GET(hdr->count, ARCH_CONVERT) > 0) && (INT_GET(hdr->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8)));
@@ -1599,8 +1600,8 @@ xfs_dir_leaf_unbalance(xfs_da_state_t *state, xfs_da_state_blk_t *drop_blk,
 	ASSERT(save_blk->magic == XFS_DIR_LEAF_MAGIC);
 	drop_leaf = drop_blk->bp->data;
 	save_leaf = save_blk->bp->data;
-	ASSERT(INT_GET(drop_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	ASSERT(INT_GET(save_leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(drop_leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(save_leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	drop_hdr = &drop_leaf->hdr;
 	save_hdr = &save_leaf->hdr;
 
@@ -1695,7 +1696,7 @@ xfs_dir_leaf_lookup_int(xfs_dabuf_t *bp, xfs_da_args_t *args, int *index)
 	xfs_dahash_t hashval;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	ASSERT(INT_GET(leaf->hdr.count, ARCH_CONVERT) < (XFS_LBSIZE(args->dp->i_mount)/8));
 
 	/*
@@ -1782,8 +1783,8 @@ xfs_dir_leaf_moveents(xfs_dir_leafblock_t *leaf_s, int start_s,
 	/*
 	 * Set up environment.
 	 */
-	ASSERT(INT_GET(leaf_s->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
-	ASSERT(INT_GET(leaf_d->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_s->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf_d->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	hdr_s = &leaf_s->hdr;
 	hdr_d = &leaf_d->hdr;
 	ASSERT((INT_GET(hdr_s->count, ARCH_CONVERT) > 0) && (INT_GET(hdr_s->count, ARCH_CONVERT) < (XFS_LBSIZE(mp)/8)));
@@ -1883,8 +1884,8 @@ xfs_dir_leaf_order(xfs_dabuf_t *leaf1_bp, xfs_dabuf_t *leaf2_bp)
 
 	leaf1 = leaf1_bp->data;
 	leaf2 = leaf2_bp->data;
-	ASSERT((INT_GET(leaf1->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC) &&
-	       (INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC));
+	ASSERT((be16_to_cpu(leaf1->hdr.info.magic) == XFS_DIR_LEAF_MAGIC) &&
+	       (be16_to_cpu(leaf2->hdr.info.magic) == XFS_DIR_LEAF_MAGIC));
 	if ((INT_GET(leaf1->hdr.count, ARCH_CONVERT) > 0) && (INT_GET(leaf2->hdr.count, ARCH_CONVERT) > 0) &&
 	    ((INT_GET(leaf2->entries[ 0 ].hashval, ARCH_CONVERT) <
 	      INT_GET(leaf1->entries[ 0 ].hashval, ARCH_CONVERT)) ||
@@ -1904,7 +1905,7 @@ xfs_dir_leaf_lasthash(xfs_dabuf_t *bp, int *count)
 	xfs_dir_leafblock_t *leaf;
 
 	leaf = bp->data;
-	ASSERT(INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) == XFS_DIR_LEAF_MAGIC);
+	ASSERT(be16_to_cpu(leaf->hdr.info.magic) == XFS_DIR_LEAF_MAGIC);
 	if (count)
 		*count = INT_GET(leaf->hdr.count, ARCH_CONVERT);
 	if (!leaf->hdr.count)
@@ -1940,7 +1941,7 @@ xfs_dir_leaf_getdents_int(
 
 	mp = dp->i_mount;
 	leaf = bp->data;
-	if (INT_GET(leaf->hdr.info.magic, ARCH_CONVERT) != XFS_DIR_LEAF_MAGIC) {
+	if (be16_to_cpu(leaf->hdr.info.magic) != XFS_DIR_LEAF_MAGIC) {
 		*eobp = 1;
 		return XFS_ERROR(ENOENT);	/* XXX wrong code */
 	}
@@ -1992,7 +1993,7 @@ xfs_dir_leaf_getdents_int(
 
 	if (i == INT_GET(leaf->hdr.count, ARCH_CONVERT)) {
 		xfs_dir_trace_g_du("leaf: hash not found", dp, uio);
-		if (!INT_GET(leaf->hdr.info.forw, ARCH_CONVERT))
+		if (!leaf->hdr.info.forw)
 			uio->uio_offset =
 				XFS_DA_MAKE_COOKIE(mp, 0, 0, XFS_DA_MAXHASH);
 		/*
@@ -2047,8 +2048,7 @@ xfs_dir_leaf_getdents_int(
 			xfs_dir_trace_g_duc("leaf: middle cookie  ",
 						   dp, uio, p.cook.o);
 
-		} else if ((thishash = INT_GET(leaf->hdr.info.forw,
-							ARCH_CONVERT))) {
+		} else if ((thishash = be32_to_cpu(leaf->hdr.info.forw))) {
 			xfs_dabuf_t *bp2;
 			xfs_dir_leafblock_t *leaf2;
 
@@ -2064,9 +2064,9 @@ xfs_dir_leaf_getdents_int(
 			leaf2 = bp2->data;
 
 			if (unlikely(
-			       (INT_GET(leaf2->hdr.info.magic, ARCH_CONVERT)
+			       (be16_to_cpu(leaf2->hdr.info.magic)
 						!= XFS_DIR_LEAF_MAGIC)
-			    || (INT_GET(leaf2->hdr.info.back, ARCH_CONVERT)
+			    || (be32_to_cpu(leaf2->hdr.info.back)
 						!= bno))) {	/* GROT */
 				XFS_CORRUPTION_ERROR("xfs_dir_leaf_getdents_int(3)",
 						     XFS_ERRLEVEL_LOW, mp,
diff --git a/fs/xfs/xfs_dir_sf.h b/fs/xfs/xfs_dir_sf.h
index fe44c6f4d560..5b20b4d3f57d 100644
--- a/fs/xfs/xfs_dir_sf.h
+++ b/fs/xfs/xfs_dir_sf.h
@@ -35,19 +35,21 @@ typedef struct { __uint8_t i[sizeof(xfs_ino_t)]; } xfs_dir_ino_t;
  * and the elements much be memcpy'd out into a work area to get correct
  * alignment for the inode number fields.
  */
+typedef struct xfs_dir_sf_hdr {		/* constant-structure header block */
+	xfs_dir_ino_t	parent;		/* parent dir inode number */
+	__uint8_t	count;		/* count of active entries */
+} xfs_dir_sf_hdr_t;
+
+typedef struct xfs_dir_sf_entry {
+	xfs_dir_ino_t	inumber;	/* referenced inode number */
+	__uint8_t	namelen;	/* actual length of name (no NULL) */
+	__uint8_t	name[1];	/* name */
+} xfs_dir_sf_entry_t;
+
 typedef struct xfs_dir_shortform {
-	struct xfs_dir_sf_hdr {		/* constant-structure header block */
-		xfs_dir_ino_t parent;	/* parent dir inode number */
-		__uint8_t count;	/* count of active entries */
-	} hdr;
-	struct xfs_dir_sf_entry {
-		xfs_dir_ino_t inumber;	/* referenced inode number */
-		__uint8_t namelen;	/* actual length of name (no NULL) */
-		__uint8_t name[1];	/* name */
-	} list[1];			/* variable sized array */
+	xfs_dir_sf_hdr_t	hdr;
+	xfs_dir_sf_entry_t	list[1];	/* variable sized array */
 } xfs_dir_shortform_t;
-typedef struct xfs_dir_sf_hdr xfs_dir_sf_hdr_t;
-typedef struct xfs_dir_sf_entry xfs_dir_sf_entry_t;
 
 /*
  * We generate this then sort it, so that readdirs are returned in
diff --git a/fs/xfs/xfs_dmapi.h b/fs/xfs/xfs_dmapi.h
index b4c7f2bc55a0..00b1540f8108 100644
--- a/fs/xfs/xfs_dmapi.h
+++ b/fs/xfs/xfs_dmapi.h
@@ -191,14 +191,4 @@ typedef enum {
 
 extern struct bhv_vfsops xfs_dmops;
 
-#ifdef CONFIG_XFS_DMAPI
-void xfs_dm_init(struct file_system_type *);
-void xfs_dm_exit(struct file_system_type *);
-#define XFS_DM_INIT(fstype)	xfs_dm_init(fstype)
-#define XFS_DM_EXIT(fstype)	xfs_dm_exit(fstype)
-#else
-#define XFS_DM_INIT(fstype)
-#define XFS_DM_EXIT(fstype)
-#endif
-
 #endif  /* __XFS_DMAPI_H__ */
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index b4d971b01588..dfa3527b20a7 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -462,6 +462,7 @@ xfs_fs_counts(
 {
 	unsigned long	s;
 
+	xfs_icsb_sync_counters_lazy(mp);
 	s = XFS_SB_LOCK(mp);
 	cnt->freedata = mp->m_sb.sb_fdblocks;
 	cnt->freertx = mp->m_sb.sb_frextents;
@@ -476,7 +477,7 @@ xfs_fs_counts(
  *
  * xfs_reserve_blocks is called to set m_resblks
  * in the in-core mount table. The number of unused reserved blocks
- * is kept in m_resbls_avail.
+ * is kept in m_resblks_avail.
  *
  * Reserve the requested number of blocks if available. Otherwise return
  * as many as possible to satisfy the request. The actual number
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 8f3fae1aa98a..4eeb856183b1 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -136,10 +136,8 @@ xfs_ialloc_ag_alloc(
 	int		ninodes;	/* num inodes per buf */
 	xfs_agino_t	thisino;	/* current inode number, for loop */
 	int		version;	/* inode version number to use */
-	int		isaligned;	/* inode allocation at stripe unit */
+	int		isaligned = 0;	/* inode allocation at stripe unit */
 					/* boundary */
-	xfs_dinode_core_t dic;          /* a dinode_core to copy to new */
-					/* inodes */
 
 	args.tp = tp;
 	args.mp = tp->t_mountp;
@@ -154,47 +152,75 @@ xfs_ialloc_ag_alloc(
 		return XFS_ERROR(ENOSPC);
 	args.minlen = args.maxlen = XFS_IALLOC_BLOCKS(args.mp);
 	/*
-	 * Set the alignment for the allocation.
-	 * If stripe alignment is turned on then align at stripe unit
-	 * boundary.
-	 * If the cluster size is smaller than a filesystem block
-	 * then we're doing I/O for inodes in filesystem block size pieces,
-	 * so don't need alignment anyway.
-	 */
-	isaligned = 0;
-	if (args.mp->m_sinoalign) {
-		ASSERT(!(args.mp->m_flags & XFS_MOUNT_NOALIGN));
-		args.alignment = args.mp->m_dalign;
-		isaligned = 1;
-	} else if (XFS_SB_VERSION_HASALIGN(&args.mp->m_sb) &&
-	    args.mp->m_sb.sb_inoalignmt >=
-	    XFS_B_TO_FSBT(args.mp, XFS_INODE_CLUSTER_SIZE(args.mp)))
-		args.alignment = args.mp->m_sb.sb_inoalignmt;
-	else
-		args.alignment = 1;
+	 * First try to allocate inodes contiguous with the last-allocated
+	 * chunk of inodes.  If the filesystem is striped, this will fill
+	 * an entire stripe unit with inodes.
+ 	 */
 	agi = XFS_BUF_TO_AGI(agbp);
-	/*
-	 * Need to figure out where to allocate the inode blocks.
-	 * Ideally they should be spaced out through the a.g.
-	 * For now, just allocate blocks up front.
-	 */
-	args.agbno = be32_to_cpu(agi->agi_root);
-	args.fsbno = XFS_AGB_TO_FSB(args.mp, be32_to_cpu(agi->agi_seqno),
-				    args.agbno);
-	/*
-	 * Allocate a fixed-size extent of inodes.
-	 */
-	args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	args.mod = args.total = args.wasdel = args.isfl = args.userdata =
-		args.minalignslop = 0;
-	args.prod = 1;
-	/*
-	 * Allow space for the inode btree to split.
-	 */
-	args.minleft = XFS_IN_MAXLEVELS(args.mp) - 1;
-	if ((error = xfs_alloc_vextent(&args)))
-		return error;
+	newino = be32_to_cpu(agi->agi_newino);
+	if(likely(newino != NULLAGINO)) {
+		args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
+				XFS_IALLOC_BLOCKS(args.mp);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp,
+				be32_to_cpu(agi->agi_seqno), args.agbno);
+		args.type = XFS_ALLOCTYPE_THIS_BNO;
+		args.mod = args.total = args.wasdel = args.isfl =
+			args.userdata = args.minalignslop = 0;
+		args.prod = 1;
+		args.alignment = 1;
+		/*
+		 * Allow space for the inode btree to split.
+		 */
+		args.minleft = XFS_IN_MAXLEVELS(args.mp) - 1;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+	} else
+		args.fsbno = NULLFSBLOCK;
 
+	if (unlikely(args.fsbno == NULLFSBLOCK)) {
+		/*
+		 * Set the alignment for the allocation.
+		 * If stripe alignment is turned on then align at stripe unit
+		 * boundary.
+		 * If the cluster size is smaller than a filesystem block 
+		 * then we're doing I/O for inodes in filesystem block size 
+		 * pieces, so don't need alignment anyway.
+		 */
+		isaligned = 0;
+		if (args.mp->m_sinoalign) {
+			ASSERT(!(args.mp->m_flags & XFS_MOUNT_NOALIGN));
+			args.alignment = args.mp->m_dalign;
+			isaligned = 1;
+		} else if (XFS_SB_VERSION_HASALIGN(&args.mp->m_sb) &&
+			   args.mp->m_sb.sb_inoalignmt >= 
+			   XFS_B_TO_FSBT(args.mp,
+			  	XFS_INODE_CLUSTER_SIZE(args.mp)))
+				args.alignment = args.mp->m_sb.sb_inoalignmt;
+		else
+			args.alignment = 1;
+		/*
+		 * Need to figure out where to allocate the inode blocks.
+		 * Ideally they should be spaced out through the a.g.
+		 * For now, just allocate blocks up front.
+		 */
+		args.agbno = be32_to_cpu(agi->agi_root);
+		args.fsbno = XFS_AGB_TO_FSB(args.mp,
+				be32_to_cpu(agi->agi_seqno), args.agbno);
+		/*
+		 * Allocate a fixed-size extent of inodes.
+		 */
+		args.type = XFS_ALLOCTYPE_NEAR_BNO;
+		args.mod = args.total = args.wasdel = args.isfl =
+			args.userdata = args.minalignslop = 0;
+		args.prod = 1;
+		/*
+		 * Allow space for the inode btree to split.
+		 */
+		args.minleft = XFS_IN_MAXLEVELS(args.mp) - 1;
+		if ((error = xfs_alloc_vextent(&args)))
+			return error;
+	}
+ 
 	/*
 	 * If stripe alignment is turned on, then try again with cluster
 	 * alignment.
@@ -250,10 +276,6 @@ xfs_ialloc_ag_alloc(
 	else
 		version = XFS_DINODE_VERSION_1;
 
-	memset(&dic, 0, sizeof(xfs_dinode_core_t));
-	INT_SET(dic.di_magic, ARCH_CONVERT, XFS_DINODE_MAGIC);
-	INT_SET(dic.di_version, ARCH_CONVERT, version);
-
 	for (j = 0; j < nbufs; j++) {
 		/*
 		 * Get the block.
@@ -266,12 +288,13 @@ xfs_ialloc_ag_alloc(
 		ASSERT(fbuf);
 		ASSERT(!XFS_BUF_GETERROR(fbuf));
 		/*
-		 * Loop over the inodes in this buffer.
+		 * Set initial values for the inodes in this buffer.
 		 */
-
+		xfs_biozero(fbuf, 0, ninodes << args.mp->m_sb.sb_inodelog);
 		for (i = 0; i < ninodes; i++) {
 			free = XFS_MAKE_IPTR(args.mp, fbuf, i);
-			memcpy(&(free->di_core), &dic, sizeof(xfs_dinode_core_t));
+			INT_SET(free->di_core.di_magic, ARCH_CONVERT, XFS_DINODE_MAGIC);
+			INT_SET(free->di_core.di_version, ARCH_CONVERT, version);
 			INT_SET(free->di_next_unlinked, ARCH_CONVERT, NULLAGINO);
 			xfs_ialloc_log_di(tp, fbuf, i,
 				XFS_DI_CORE_BITS | XFS_DI_NEXT_UNLINKED);
@@ -1028,7 +1051,7 @@ xfs_difree(
 	rec.ir_freecount++;
 
 	/*
-	 * When an inode cluster is free, it becomes elgible for removal
+	 * When an inode cluster is free, it becomes eligible for removal
 	 */
 	if ((mp->m_flags & XFS_MOUNT_IDELETE) &&
 	    (rec.ir_freecount == XFS_IALLOC_INODES(mp))) {
diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c
index 8e380a1fb79b..bb33113eef9f 100644
--- a/fs/xfs/xfs_iget.c
+++ b/fs/xfs/xfs_iget.c
@@ -258,7 +258,7 @@ again:
 				goto finish_inode;
 
 			} else if (vp != inode_vp) {
-				struct inode *inode = LINVFS_GET_IP(inode_vp);
+				struct inode *inode = vn_to_inode(inode_vp);
 
 				/* The inode is being torn down, pause and
 				 * try again.
@@ -495,7 +495,7 @@ retry:
 	if ((inode = iget_locked(XFS_MTOVFS(mp)->vfs_super, ino))) {
 		xfs_inode_t	*ip;
 
-		vp = LINVFS_GET_VP(inode);
+		vp = vn_from_inode(inode);
 		if (inode->i_state & I_NEW) {
 			vn_initialize(inode);
 			error = xfs_iget_core(vp, mp, tp, ino, flags,
@@ -509,7 +509,7 @@ retry:
 		} else {
 			/*
 			 * If the inode is not fully constructed due to
-			 * filehandle mistmatches wait for the inode to go
+			 * filehandle mismatches wait for the inode to go
 			 * away and try again.
 			 *
 			 * iget_locked will call __wait_on_freeing_inode
@@ -617,7 +617,7 @@ xfs_iput_new(xfs_inode_t	*ip,
 	     uint		lock_flags)
 {
 	vnode_t		*vp = XFS_ITOV(ip);
-	struct inode	*inode = LINVFS_GET_IP(vp);
+	struct inode	*inode = vn_to_inode(vp);
 
 	vn_trace_entry(vp, "xfs_iput_new", (inst_t *)__return_address);
 
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index 1d7f5a7e063e..48146bdc6bdd 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -76,16 +76,18 @@ STATIC int xfs_iformat_btree(xfs_inode_t *, xfs_dinode_t *, int);
  */
 STATIC void
 xfs_validate_extents(
-	xfs_bmbt_rec_t		*ep,
+	xfs_ifork_t		*ifp,
 	int			nrecs,
 	int			disk,
 	xfs_exntfmt_t		fmt)
 {
+	xfs_bmbt_rec_t		*ep;
 	xfs_bmbt_irec_t		irec;
 	xfs_bmbt_rec_t		rec;
 	int			i;
 
 	for (i = 0; i < nrecs; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
 		rec.l0 = get_unaligned((__uint64_t*)&ep->l0);
 		rec.l1 = get_unaligned((__uint64_t*)&ep->l1);
 		if (disk)
@@ -94,11 +96,10 @@ xfs_validate_extents(
 			xfs_bmbt_get_all(&rec, &irec);
 		if (fmt == XFS_EXTFMT_NOSTATE)
 			ASSERT(irec.br_state == XFS_EXT_NORM);
-		ep++;
 	}
 }
 #else /* DEBUG */
-#define xfs_validate_extents(ep, nrecs, disk, fmt)
+#define xfs_validate_extents(ifp, nrecs, disk, fmt)
 #endif /* DEBUG */
 
 /*
@@ -159,7 +160,7 @@ xfs_inotobp(
 	xfs_dinode_t	*dip;
 
 	/*
-	 * Call the space managment code to find the location of the
+	 * Call the space management code to find the location of the
 	 * inode on disk.
 	 */
 	imap.im_blkno = 0;
@@ -252,7 +253,8 @@ xfs_itobp(
 	xfs_inode_t	*ip,
 	xfs_dinode_t	**dipp,
 	xfs_buf_t	**bpp,
-	xfs_daddr_t	bno)
+	xfs_daddr_t	bno,
+	uint		imap_flags)
 {
 	xfs_buf_t	*bp;
 	int		error;
@@ -268,10 +270,9 @@ xfs_itobp(
 		 * inode on disk.
 		 */
 		imap.im_blkno = bno;
-		error = xfs_imap(mp, tp, ip->i_ino, &imap, XFS_IMAP_LOOKUP);
-		if (error != 0) {
+		if ((error = xfs_imap(mp, tp, ip->i_ino, &imap,
+					XFS_IMAP_LOOKUP | imap_flags)))
 			return error;
-		}
 
 		/*
 		 * If the inode number maps to a block outside the bounds
@@ -335,9 +336,10 @@ xfs_itobp(
 	 * (if DEBUG kernel) or the first inode in the buffer, otherwise.
 	 */
 #ifdef DEBUG
-	ni = BBTOB(imap.im_len) >> mp->m_sb.sb_inodelog;
+	ni = (imap_flags & XFS_IMAP_BULKSTAT) ? 0 :
+		(BBTOB(imap.im_len) >> mp->m_sb.sb_inodelog);
 #else
-	ni = 1;
+	ni = (imap_flags & XFS_IMAP_BULKSTAT) ? 0 : 1;
 #endif
 	for (i = 0; i < ni; i++) {
 		int		di_ok;
@@ -504,7 +506,7 @@ xfs_iformat(
 	switch (INT_GET(dip->di_core.di_aformat, ARCH_CONVERT)) {
 	case XFS_DINODE_FMT_LOCAL:
 		atp = (xfs_attr_shortform_t *)XFS_DFORK_APTR(dip);
-		size = (int)INT_GET(atp->hdr.totsize, ARCH_CONVERT);
+		size = be16_to_cpu(atp->hdr.totsize);
 		error = xfs_iformat_local(ip, dip, XFS_ATTR_FORK, size);
 		break;
 	case XFS_DINODE_FMT_EXTENTS:
@@ -597,7 +599,6 @@ xfs_iformat_extents(
 	xfs_bmbt_rec_t	*ep, *dp;
 	xfs_ifork_t	*ifp;
 	int		nex;
-	int		real_size;
 	int		size;
 	int		i;
 
@@ -619,23 +620,20 @@ xfs_iformat_extents(
 		return XFS_ERROR(EFSCORRUPTED);
 	}
 
-	real_size = 0;
+	ifp->if_real_bytes = 0;
 	if (nex == 0)
 		ifp->if_u1.if_extents = NULL;
 	else if (nex <= XFS_INLINE_EXTS)
 		ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
-	else {
-		ifp->if_u1.if_extents = kmem_alloc(size, KM_SLEEP);
-		ASSERT(ifp->if_u1.if_extents != NULL);
-		real_size = size;
-	}
+	else
+		xfs_iext_add(ifp, 0, nex);
+
 	ifp->if_bytes = size;
-	ifp->if_real_bytes = real_size;
 	if (size) {
 		dp = (xfs_bmbt_rec_t *) XFS_DFORK_PTR(dip, whichfork);
-		xfs_validate_extents(dp, nex, 1, XFS_EXTFMT_INODE(ip));
-		ep = ifp->if_u1.if_extents;
-		for (i = 0; i < nex; i++, ep++, dp++) {
+		xfs_validate_extents(ifp, nex, 1, XFS_EXTFMT_INODE(ip));
+		for (i = 0; i < nex; i++, dp++) {
+			ep = xfs_iext_get_ext(ifp, i);
 			ep->l0 = INT_GET(get_unaligned((__uint64_t*)&dp->l0),
 								ARCH_CONVERT);
 			ep->l1 = INT_GET(get_unaligned((__uint64_t*)&dp->l1),
@@ -646,7 +644,7 @@ xfs_iformat_extents(
 		if (whichfork != XFS_DATA_FORK ||
 			XFS_EXTFMT_INODE(ip) == XFS_EXTFMT_NOSTATE)
 				if (unlikely(xfs_check_nostate_extents(
-				    ifp->if_u1.if_extents, nex))) {
+				    ifp, 0, nex))) {
 					XFS_ERROR_REPORT("xfs_iformat_extents(2)",
 							 XFS_ERRLEVEL_LOW,
 							 ip->i_mount);
@@ -839,7 +837,7 @@ xfs_dic2xflags(
 
 /*
  * Given a mount structure and an inode number, return a pointer
- * to a newly allocated in-core inode coresponding to the given
+ * to a newly allocated in-core inode corresponding to the given
  * inode number.
  *
  * Initialize the inode's attributes and extent pointers if it
@@ -871,9 +869,8 @@ xfs_iread(
 	 * return NULL as well.  Set i_blkno to 0 so that xfs_itobp() will
 	 * know that this is a new incore inode.
 	 */
-	error = xfs_itobp(mp, tp, ip, &dip, &bp, bno);
-
-	if (error != 0) {
+	error = xfs_itobp(mp, tp, ip, &dip, &bp, bno, 0);
+	if (error) {
 		kmem_zone_free(xfs_inode_zone, ip);
 		return error;
 	}
@@ -1015,6 +1012,7 @@ xfs_iread_extents(
 {
 	int		error;
 	xfs_ifork_t	*ifp;
+	xfs_extnum_t	nextents;
 	size_t		size;
 
 	if (unlikely(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)) {
@@ -1022,26 +1020,24 @@ xfs_iread_extents(
 				 ip->i_mount);
 		return XFS_ERROR(EFSCORRUPTED);
 	}
-	size = XFS_IFORK_NEXTENTS(ip, whichfork) * (uint)sizeof(xfs_bmbt_rec_t);
+	nextents = XFS_IFORK_NEXTENTS(ip, whichfork);
+	size = nextents * sizeof(xfs_bmbt_rec_t);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
+
 	/*
 	 * We know that the size is valid (it's checked in iformat_btree)
 	 */
-	ifp->if_u1.if_extents = kmem_alloc(size, KM_SLEEP);
-	ASSERT(ifp->if_u1.if_extents != NULL);
 	ifp->if_lastex = NULLEXTNUM;
-	ifp->if_bytes = ifp->if_real_bytes = (int)size;
+	ifp->if_bytes = ifp->if_real_bytes = 0;
 	ifp->if_flags |= XFS_IFEXTENTS;
+	xfs_iext_add(ifp, 0, nextents);
 	error = xfs_bmap_read_extents(tp, ip, whichfork);
 	if (error) {
-		kmem_free(ifp->if_u1.if_extents, size);
-		ifp->if_u1.if_extents = NULL;
-		ifp->if_bytes = ifp->if_real_bytes = 0;
+		xfs_iext_destroy(ifp);
 		ifp->if_flags &= ~XFS_IFEXTENTS;
 		return error;
 	}
-	xfs_validate_extents((xfs_bmbt_rec_t *)ifp->if_u1.if_extents,
-		XFS_IFORK_NEXTENTS(ip, whichfork), 0, XFS_EXTFMT_INODE(ip));
+	xfs_validate_extents(ifp, nextents, 0, XFS_EXTFMT_INODE(ip));
 	return 0;
 }
 
@@ -1376,10 +1372,10 @@ xfs_itrunc_trace(
 		     (void*)(unsigned long)((toss_finish >> 32) & 0xffffffff),
 		     (void*)(unsigned long)(toss_finish & 0xffffffff),
 		     (void*)(unsigned long)current_cpu(),
-		     (void*)0,
-		     (void*)0,
-		     (void*)0,
-		     (void*)0);
+		     (void*)(unsigned long)current_pid(),
+		     (void*)NULL,
+		     (void*)NULL,
+		     (void*)NULL);
 }
 #else
 #define	xfs_itrunc_trace(tag, ip, flag, new_size, toss_start, toss_finish)
@@ -1397,6 +1393,16 @@ xfs_itrunc_trace(
  * calling into the buffer/page cache code and we can't hold the
  * inode lock when we do so.
  *
+ * We need to wait for any direct I/Os in flight to complete before we
+ * proceed with the truncate. This is needed to prevent the extents
+ * being read or written by the direct I/Os from being removed while the
+ * I/O is in flight as there is no other method of synchronising
+ * direct I/O with the truncate operation.  Also, because we hold
+ * the IOLOCK in exclusive mode, we prevent new direct I/Os from being
+ * started until the truncate completes and drops the lock. Essentially,
+ * the vn_iowait() call forms an I/O barrier that provides strict ordering
+ * between direct I/Os and the truncate operation.
+ *
  * The flags parameter can have either the value XFS_ITRUNC_DEFINITE
  * or XFS_ITRUNC_MAYBE.  The XFS_ITRUNC_MAYBE value should be used
  * in the case that the caller is locking things out of order and
@@ -1424,6 +1430,9 @@ xfs_itruncate_start(
 
 	mp = ip->i_mount;
 	vp = XFS_ITOV(ip);
+
+	vn_iowait(vp);  /* wait for the completion of any pending DIOs */
+	
 	/*
 	 * Call VOP_TOSS_PAGES() or VOP_FLUSHINVAL_PAGES() to get rid of pages and buffers
 	 * overlapping the region being removed.  We have to use
@@ -1899,7 +1908,7 @@ xfs_iunlink(
 		 * Here we put the head pointer into our next pointer,
 		 * and then we fall through to point the head at us.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
 		if (error) {
 			return error;
 		}
@@ -2008,7 +2017,7 @@ xfs_iunlink_remove(
 		 * of dealing with the buffer when there is no need to
 		 * change it.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
 		if (error) {
 			cmn_err(CE_WARN,
 				"xfs_iunlink_remove: xfs_itobp()  returned an error %d on %s.  Returning error.",
@@ -2070,7 +2079,7 @@ xfs_iunlink_remove(
 		 * Now last_ibp points to the buffer previous to us on
 		 * the unlinked list.  Pull us from the list.
 		 */
-		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0);
+		error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0);
 		if (error) {
 			cmn_err(CE_WARN,
 				"xfs_iunlink_remove: xfs_itobp()  returned an error %d on %s.  Returning error.",
@@ -2476,92 +2485,6 @@ xfs_iroot_realloc(
 
 
 /*
- * This is called when the amount of space needed for if_extents
- * is increased or decreased.  The change in size is indicated by
- * the number of extents that need to be added or deleted in the
- * ext_diff parameter.
- *
- * If the amount of space needed has decreased below the size of the
- * inline buffer, then switch to using the inline buffer.  Otherwise,
- * use kmem_realloc() or kmem_alloc() to adjust the size of the buffer
- * to what is needed.
- *
- * ip -- the inode whose if_extents area is changing
- * ext_diff -- the change in the number of extents, positive or negative,
- *	 requested for the if_extents array.
- */
-void
-xfs_iext_realloc(
-	xfs_inode_t	*ip,
-	int		ext_diff,
-	int		whichfork)
-{
-	int		byte_diff;
-	xfs_ifork_t	*ifp;
-	int		new_size;
-	uint		rnew_size;
-
-	if (ext_diff == 0) {
-		return;
-	}
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	byte_diff = ext_diff * (uint)sizeof(xfs_bmbt_rec_t);
-	new_size = (int)ifp->if_bytes + byte_diff;
-	ASSERT(new_size >= 0);
-
-	if (new_size == 0) {
-		if (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext) {
-			ASSERT(ifp->if_real_bytes != 0);
-			kmem_free(ifp->if_u1.if_extents, ifp->if_real_bytes);
-		}
-		ifp->if_u1.if_extents = NULL;
-		rnew_size = 0;
-	} else if (new_size <= sizeof(ifp->if_u2.if_inline_ext)) {
-		/*
-		 * If the valid extents can fit in if_inline_ext,
-		 * copy them from the malloc'd vector and free it.
-		 */
-		if (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext) {
-			/*
-			 * For now, empty files are format EXTENTS,
-			 * so the if_extents pointer is null.
-			 */
-			if (ifp->if_u1.if_extents) {
-				memcpy(ifp->if_u2.if_inline_ext,
-					ifp->if_u1.if_extents, new_size);
-				kmem_free(ifp->if_u1.if_extents,
-					  ifp->if_real_bytes);
-			}
-			ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
-		}
-		rnew_size = 0;
-	} else {
-		rnew_size = new_size;
-		if ((rnew_size & (rnew_size - 1)) != 0)
-			rnew_size = xfs_iroundup(rnew_size);
-		/*
-		 * Stuck with malloc/realloc.
-		 */
-		if (ifp->if_u1.if_extents == ifp->if_u2.if_inline_ext) {
-			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
-				kmem_alloc(rnew_size, KM_SLEEP);
-			memcpy(ifp->if_u1.if_extents, ifp->if_u2.if_inline_ext,
-			      sizeof(ifp->if_u2.if_inline_ext));
-		} else if (rnew_size != ifp->if_real_bytes) {
-			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
-			  kmem_realloc(ifp->if_u1.if_extents,
-					rnew_size,
-					ifp->if_real_bytes,
-					KM_NOFS);
-		}
-	}
-	ifp->if_real_bytes = rnew_size;
-	ifp->if_bytes = new_size;
-}
-
-
-/*
  * This is called when the amount of space needed for if_data
  * is increased or decreased.  The change in size is indicated by
  * the number of bytes that need to be added or deleted in the
@@ -2720,12 +2643,11 @@ xfs_idestroy_fork(
 			ifp->if_real_bytes = 0;
 		}
 	} else if ((ifp->if_flags & XFS_IFEXTENTS) &&
-		   (ifp->if_u1.if_extents != NULL) &&
-		   (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext)) {
+		   ((ifp->if_flags & XFS_IFEXTIREC) ||
+		    ((ifp->if_u1.if_extents != NULL) &&
+		     (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext)))) {
 		ASSERT(ifp->if_real_bytes != 0);
-		kmem_free(ifp->if_u1.if_extents, ifp->if_real_bytes);
-		ifp->if_u1.if_extents = NULL;
-		ifp->if_real_bytes = 0;
+		xfs_iext_destroy(ifp);
 	}
 	ASSERT(ifp->if_u1.if_extents == NULL ||
 	       ifp->if_u1.if_extents == ifp->if_u2.if_inline_ext);
@@ -2801,7 +2723,7 @@ xfs_ipin(
 /*
  * Decrement the pin count of the given inode, and wake up
  * anyone in xfs_iwait_unpin() if the count goes to 0.  The
- * inode must have been previoulsy pinned with a call to xfs_ipin().
+ * inode must have been previously pinned with a call to xfs_ipin().
  */
 void
 xfs_iunpin(
@@ -2814,7 +2736,7 @@ xfs_iunpin(
 
 		/* make sync come back and flush this inode */
 		if (vp) {
-			struct inode	*inode = LINVFS_GET_IP(vp);
+			struct inode	*inode = vn_to_inode(vp);
 
 			if (!(inode->i_state & I_NEW))
 				mark_inode_dirty_sync(inode);
@@ -2902,16 +2824,15 @@ xfs_iextents_copy(
 	 * the delayed ones.  There must be at least one
 	 * non-delayed extent.
 	 */
-	ep = ifp->if_u1.if_extents;
 	dest_ep = buffer;
 	copied = 0;
 	for (i = 0; i < nrecs; i++) {
+		ep = xfs_iext_get_ext(ifp, i);
 		start_block = xfs_bmbt_get_startblock(ep);
 		if (ISNULLSTARTBLOCK(start_block)) {
 			/*
 			 * It's a delayed allocation extent, so skip it.
 			 */
-			ep++;
 			continue;
 		}
 
@@ -2921,11 +2842,10 @@ xfs_iextents_copy(
 		put_unaligned(INT_GET(ep->l1, ARCH_CONVERT),
 			      (__uint64_t*)&dest_ep->l1);
 		dest_ep++;
-		ep++;
 		copied++;
 	}
 	ASSERT(copied != 0);
-	xfs_validate_extents(buffer, copied, 1, XFS_EXTFMT_INODE(ip));
+	xfs_validate_extents(ifp, copied, 1, XFS_EXTFMT_INODE(ip));
 
 	return (copied * (uint)sizeof(xfs_bmbt_rec_t));
 }
@@ -2995,8 +2915,10 @@ xfs_iflush_fork(
 	case XFS_DINODE_FMT_EXTENTS:
 		ASSERT((ifp->if_flags & XFS_IFEXTENTS) ||
 		       !(iip->ili_format.ilf_fields & extflag[whichfork]));
-		ASSERT((ifp->if_u1.if_extents != NULL) || (ifp->if_bytes == 0));
-		ASSERT((ifp->if_u1.if_extents == NULL) || (ifp->if_bytes > 0));
+		ASSERT((xfs_iext_get_ext(ifp, 0) != NULL) ||
+			(ifp->if_bytes == 0));
+		ASSERT((xfs_iext_get_ext(ifp, 0) == NULL) ||
+			(ifp->if_bytes > 0));
 		if ((iip->ili_format.ilf_fields & extflag[whichfork]) &&
 		    (ifp->if_bytes > 0)) {
 			ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) > 0);
@@ -3114,8 +3036,8 @@ xfs_iflush(
 	/*
 	 * Get the buffer containing the on-disk inode.
 	 */
-	error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0);
-	if (error != 0) {
+	error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0, 0);
+	if (error) {
 		xfs_ifunlock(ip);
 		return error;
 	}
@@ -3610,7 +3532,7 @@ xfs_iaccess(
 {
 	int		error;
 	mode_t		orgmode = mode;
-	struct inode	*inode = LINVFS_GET_IP(XFS_ITOV(ip));
+	struct inode	*inode = vn_to_inode(XFS_ITOV(ip));
 
 	if (mode & S_IWUSR) {
 		umode_t		imode = inode->i_mode;
@@ -3704,3 +3626,1100 @@ xfs_ilock_trace(xfs_inode_t *ip, int lock, unsigned int lockflags, inst_t *ra)
 		     NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL,NULL);
 }
 #endif
+
+/*
+ * Return a pointer to the extent record at file index idx.
+ */
+xfs_bmbt_rec_t *
+xfs_iext_get_ext(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx)		/* index of target extent */
+{
+	ASSERT(idx >= 0);
+	if ((ifp->if_flags & XFS_IFEXTIREC) && (idx == 0)) {
+		return ifp->if_u1.if_ext_irec->er_extbuf;
+	} else if (ifp->if_flags & XFS_IFEXTIREC) {
+		xfs_ext_irec_t	*erp;		/* irec pointer */
+		int		erp_idx = 0;	/* irec index */
+		xfs_extnum_t	page_idx = idx;	/* ext index in target list */
+
+		erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 0);
+		return &erp->er_extbuf[page_idx];
+	} else if (ifp->if_bytes) {
+		return &ifp->if_u1.if_extents[idx];
+	} else {
+		return NULL;
+	}
+}
+
+/*
+ * Insert new item(s) into the extent records for incore inode
+ * fork 'ifp'.  'count' new items are inserted at index 'idx'.
+ */
+void
+xfs_iext_insert(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* starting index of new items */
+	xfs_extnum_t	count,		/* number of inserted items */
+	xfs_bmbt_irec_t	*new)		/* items to insert */
+{
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_extnum_t	i;		/* extent record index */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
+	xfs_iext_add(ifp, idx, count);
+	for (i = idx; i < idx + count; i++, new++) {
+		ep = xfs_iext_get_ext(ifp, i);
+		xfs_bmbt_set_all(ep, new);
+	}
+}
+
+/*
+ * This is called when the amount of space required for incore file
+ * extents needs to be increased. The ext_diff parameter stores the
+ * number of new extents being added and the idx parameter contains
+ * the extent index where the new extents will be added. If the new
+ * extents are being appended, then we just need to (re)allocate and
+ * initialize the space. Otherwise, if the new extents are being
+ * inserted into the middle of the existing entries, a bit more work
+ * is required to make room for the new extents to be inserted. The
+ * caller is responsible for filling in the new extent entries upon
+ * return.
+ */
+void
+xfs_iext_add(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin adding exts */
+	int		ext_diff)	/* number of extents to add */
+{
+	int		byte_diff;	/* new bytes being added */
+	int		new_size;	/* size of extents after adding */
+	xfs_extnum_t	nextents;	/* number of extents in file */
+
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT((idx >= 0) && (idx <= nextents));
+	byte_diff = ext_diff * sizeof(xfs_bmbt_rec_t);
+	new_size = ifp->if_bytes + byte_diff;
+	/*
+	 * If the new number of extents (nextents + ext_diff)
+	 * fits inside the inode, then continue to use the inline
+	 * extent buffer.
+	 */
+	if (nextents + ext_diff <= XFS_INLINE_EXTS) {
+		if (idx < nextents) {
+			memmove(&ifp->if_u2.if_inline_ext[idx + ext_diff],
+				&ifp->if_u2.if_inline_ext[idx],
+				(nextents - idx) * sizeof(xfs_bmbt_rec_t));
+			memset(&ifp->if_u2.if_inline_ext[idx], 0, byte_diff);
+		}
+		ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
+		ifp->if_real_bytes = 0;
+		ifp->if_lastex = nextents + ext_diff;
+	}
+	/*
+	 * Otherwise use a linear (direct) extent list.
+	 * If the extents are currently inside the inode,
+	 * xfs_iext_realloc_direct will switch us from
+	 * inline to direct extent allocation mode.
+	 */
+	else if (nextents + ext_diff <= XFS_LINEAR_EXTS) {
+		xfs_iext_realloc_direct(ifp, new_size);
+		if (idx < nextents) {
+			memmove(&ifp->if_u1.if_extents[idx + ext_diff],
+				&ifp->if_u1.if_extents[idx],
+				(nextents - idx) * sizeof(xfs_bmbt_rec_t));
+			memset(&ifp->if_u1.if_extents[idx], 0, byte_diff);
+		}
+	}
+	/* Indirection array */
+	else {
+		xfs_ext_irec_t	*erp;
+		int		erp_idx = 0;
+		int		page_idx = idx;
+
+		ASSERT(nextents + ext_diff > XFS_LINEAR_EXTS);
+		if (ifp->if_flags & XFS_IFEXTIREC) {
+			erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 1);
+		} else {
+			xfs_iext_irec_init(ifp);
+			ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+			erp = ifp->if_u1.if_ext_irec;
+		}
+		/* Extents fit in target extent page */
+		if (erp && erp->er_extcount + ext_diff <= XFS_LINEAR_EXTS) {
+			if (page_idx < erp->er_extcount) {
+				memmove(&erp->er_extbuf[page_idx + ext_diff],
+					&erp->er_extbuf[page_idx],
+					(erp->er_extcount - page_idx) *
+					sizeof(xfs_bmbt_rec_t));
+				memset(&erp->er_extbuf[page_idx], 0, byte_diff);
+			}
+			erp->er_extcount += ext_diff;
+			xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff);
+		}
+		/* Insert a new extent page */
+		else if (erp) {
+			xfs_iext_add_indirect_multi(ifp,
+				erp_idx, page_idx, ext_diff);
+		}
+		/*
+		 * If extent(s) are being appended to the last page in
+		 * the indirection array and the new extent(s) don't fit
+		 * in the page, then erp is NULL and erp_idx is set to
+		 * the next index needed in the indirection array.
+		 */
+		else {
+			int	count = ext_diff;
+
+			while (count) {
+				erp = xfs_iext_irec_new(ifp, erp_idx);
+				erp->er_extcount = count;
+				count -= MIN(count, (int)XFS_LINEAR_EXTS);
+				if (count) {
+					erp_idx++;
+				}
+			}
+		}
+	}
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * This is called when incore extents are being added to the indirection
+ * array and the new extents do not fit in the target extent list. The
+ * erp_idx parameter contains the irec index for the target extent list
+ * in the indirection array, and the idx parameter contains the extent
+ * index within the list. The number of extents being added is stored
+ * in the count parameter.
+ *
+ *    |-------|   |-------|
+ *    |       |   |       |    idx - number of extents before idx
+ *    |  idx  |   | count |
+ *    |       |   |       |    count - number of extents being inserted at idx
+ *    |-------|   |-------|
+ *    | count |   | nex2  |    nex2 - number of extents after idx + count
+ *    |-------|   |-------|
+ */
+void
+xfs_iext_add_indirect_multi(
+	xfs_ifork_t	*ifp,			/* inode fork pointer */
+	int		erp_idx,		/* target extent irec index */
+	xfs_extnum_t	idx,			/* index within target list */
+	int		count)			/* new extents being added */
+{
+	int		byte_diff;		/* new bytes being added */
+	xfs_ext_irec_t	*erp;			/* pointer to irec entry */
+	xfs_extnum_t	ext_diff;		/* number of extents to add */
+	xfs_extnum_t	ext_cnt;		/* new extents still needed */
+	xfs_extnum_t	nex2;			/* extents after idx + count */
+	xfs_bmbt_rec_t	*nex2_ep = NULL;	/* temp list for nex2 extents */
+	int		nlists;			/* number of irec's (lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	erp = &ifp->if_u1.if_ext_irec[erp_idx];
+	nex2 = erp->er_extcount - idx;
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+
+	/*
+	 * Save second part of target extent list
+	 * (all extents past */
+	if (nex2) {
+		byte_diff = nex2 * sizeof(xfs_bmbt_rec_t);
+		nex2_ep = (xfs_bmbt_rec_t *) kmem_alloc(byte_diff, KM_SLEEP);
+		memmove(nex2_ep, &erp->er_extbuf[idx], byte_diff);
+		erp->er_extcount -= nex2;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -nex2);
+		memset(&erp->er_extbuf[idx], 0, byte_diff);
+	}
+
+	/*
+	 * Add the new extents to the end of the target
+	 * list, then allocate new irec record(s) and
+	 * extent buffer(s) as needed to store the rest
+	 * of the new extents.
+	 */
+	ext_cnt = count;
+	ext_diff = MIN(ext_cnt, (int)XFS_LINEAR_EXTS - erp->er_extcount);
+	if (ext_diff) {
+		erp->er_extcount += ext_diff;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff);
+		ext_cnt -= ext_diff;
+	}
+	while (ext_cnt) {
+		erp_idx++;
+		erp = xfs_iext_irec_new(ifp, erp_idx);
+		ext_diff = MIN(ext_cnt, (int)XFS_LINEAR_EXTS);
+		erp->er_extcount = ext_diff;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff);
+		ext_cnt -= ext_diff;
+	}
+
+	/* Add nex2 extents back to indirection array */
+	if (nex2) {
+		xfs_extnum_t	ext_avail;
+		int		i;
+
+		byte_diff = nex2 * sizeof(xfs_bmbt_rec_t);
+		ext_avail = XFS_LINEAR_EXTS - erp->er_extcount;
+		i = 0;
+		/*
+		 * If nex2 extents fit in the current page, append
+		 * nex2_ep after the new extents.
+		 */
+		if (nex2 <= ext_avail) {
+			i = erp->er_extcount;
+		}
+		/*
+		 * Otherwise, check if space is available in the
+		 * next page.
+		 */
+		else if ((erp_idx < nlists - 1) &&
+			 (nex2 <= (ext_avail = XFS_LINEAR_EXTS -
+			  ifp->if_u1.if_ext_irec[erp_idx+1].er_extcount))) {
+			erp_idx++;
+			erp++;
+			/* Create a hole for nex2 extents */
+			memmove(&erp->er_extbuf[nex2], erp->er_extbuf,
+				erp->er_extcount * sizeof(xfs_bmbt_rec_t));
+		}
+		/*
+		 * Final choice, create a new extent page for
+		 * nex2 extents.
+		 */
+		else {
+			erp_idx++;
+			erp = xfs_iext_irec_new(ifp, erp_idx);
+		}
+		memmove(&erp->er_extbuf[i], nex2_ep, byte_diff);
+		kmem_free(nex2_ep, byte_diff);
+		erp->er_extcount += nex2;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, nex2);
+	}
+}
+
+/*
+ * This is called when the amount of space required for incore file
+ * extents needs to be decreased. The ext_diff parameter stores the
+ * number of extents to be removed and the idx parameter contains
+ * the extent index where the extents will be removed from.
+ *
+ * If the amount of space needed has decreased below the linear
+ * limit, XFS_IEXT_BUFSZ, then switch to using the contiguous
+ * extent array.  Otherwise, use kmem_realloc() to adjust the
+ * size to what is needed.
+ */
+void
+xfs_iext_remove(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing exts */
+	int		ext_diff)	/* number of extents to remove */
+{
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		new_size;	/* size of extents after removal */
+
+	ASSERT(ext_diff > 0);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	new_size = (nextents - ext_diff) * sizeof(xfs_bmbt_rec_t);
+
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+	} else if (ifp->if_flags & XFS_IFEXTIREC) {
+		xfs_iext_remove_indirect(ifp, idx, ext_diff);
+	} else if (ifp->if_real_bytes) {
+		xfs_iext_remove_direct(ifp, idx, ext_diff);
+	} else {
+		xfs_iext_remove_inline(ifp, idx, ext_diff);
+	}
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * This removes ext_diff extents from the inline buffer, beginning
+ * at extent index idx.
+ */
+void
+xfs_iext_remove_inline(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing exts */
+	int		ext_diff)	/* number of extents to remove */
+{
+	int		nextents;	/* number of extents in file */
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
+	ASSERT(idx < XFS_INLINE_EXTS);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT(((nextents - ext_diff) > 0) &&
+		(nextents - ext_diff) < XFS_INLINE_EXTS);
+
+	if (idx + ext_diff < nextents) {
+		memmove(&ifp->if_u2.if_inline_ext[idx],
+			&ifp->if_u2.if_inline_ext[idx + ext_diff],
+			(nextents - (idx + ext_diff)) *
+			 sizeof(xfs_bmbt_rec_t));
+		memset(&ifp->if_u2.if_inline_ext[nextents - ext_diff],
+			0, ext_diff * sizeof(xfs_bmbt_rec_t));
+	} else {
+		memset(&ifp->if_u2.if_inline_ext[idx], 0,
+			ext_diff * sizeof(xfs_bmbt_rec_t));
+	}
+}
+
+/*
+ * This removes ext_diff extents from a linear (direct) extent list,
+ * beginning at extent index idx. If the extents are being removed
+ * from the end of the list (ie. truncate) then we just need to re-
+ * allocate the list to remove the extra space. Otherwise, if the
+ * extents are being removed from the middle of the existing extent
+ * entries, then we first need to move the extent records beginning
+ * at idx + ext_diff up in the list to overwrite the records being
+ * removed, then remove the extra space via kmem_realloc.
+ */
+void
+xfs_iext_remove_direct(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing exts */
+	int		ext_diff)	/* number of extents to remove */
+{
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		new_size;	/* size of extents after removal */
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
+	new_size = ifp->if_bytes -
+		(ext_diff * sizeof(xfs_bmbt_rec_t));
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+		return;
+	}
+	/* Move extents up in the list (if needed) */
+	if (idx + ext_diff < nextents) {
+		memmove(&ifp->if_u1.if_extents[idx],
+			&ifp->if_u1.if_extents[idx + ext_diff],
+			(nextents - (idx + ext_diff)) *
+			 sizeof(xfs_bmbt_rec_t));
+	}
+	memset(&ifp->if_u1.if_extents[nextents - ext_diff],
+		0, ext_diff * sizeof(xfs_bmbt_rec_t));
+	/*
+	 * Reallocate the direct extent list. If the extents
+	 * will fit inside the inode then xfs_iext_realloc_direct
+	 * will switch from direct to inline extent allocation
+	 * mode for us.
+	 */
+	xfs_iext_realloc_direct(ifp, new_size);
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * This is called when incore extents are being removed from the
+ * indirection array and the extents being removed span multiple extent
+ * buffers. The idx parameter contains the file extent index where we
+ * want to begin removing extents, and the count parameter contains
+ * how many extents need to be removed.
+ *
+ *    |-------|   |-------|
+ *    | nex1  |   |       |    nex1 - number of extents before idx
+ *    |-------|   | count |
+ *    |       |   |       |    count - number of extents being removed at idx
+ *    | count |   |-------|
+ *    |       |   | nex2  |    nex2 - number of extents after idx + count
+ *    |-------|   |-------|
+ */
+void
+xfs_iext_remove_indirect(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	idx,		/* index to begin removing extents */
+	int		count)		/* number of extents to remove */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	int		erp_idx = 0;	/* indirection array index */
+	xfs_extnum_t	ext_cnt;	/* extents left to remove */
+	xfs_extnum_t	ext_diff;	/* extents to remove in current list */
+	xfs_extnum_t	nex1;		/* number of extents before idx */
+	xfs_extnum_t	nex2;		/* extents after idx + count */
+	int		nlists;		/* entries in indirection array */
+	int		page_idx = idx;	/* index in target extent list */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	erp = xfs_iext_idx_to_irec(ifp,  &page_idx, &erp_idx, 0);
+	ASSERT(erp != NULL);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	nex1 = page_idx;
+	ext_cnt = count;
+	while (ext_cnt) {
+		nex2 = MAX((erp->er_extcount - (nex1 + ext_cnt)), 0);
+		ext_diff = MIN(ext_cnt, (erp->er_extcount - nex1));
+		/*
+		 * Check for deletion of entire list;
+		 * xfs_iext_irec_remove() updates extent offsets.
+		 */
+		if (ext_diff == erp->er_extcount) {
+			xfs_iext_irec_remove(ifp, erp_idx);
+			ext_cnt -= ext_diff;
+			nex1 = 0;
+			if (ext_cnt) {
+				ASSERT(erp_idx < ifp->if_real_bytes /
+					XFS_IEXT_BUFSZ);
+				erp = &ifp->if_u1.if_ext_irec[erp_idx];
+				nex1 = 0;
+				continue;
+			} else {
+				break;
+			}
+		}
+		/* Move extents up (if needed) */
+		if (nex2) {
+			memmove(&erp->er_extbuf[nex1],
+				&erp->er_extbuf[nex1 + ext_diff],
+				nex2 * sizeof(xfs_bmbt_rec_t));
+		}
+		/* Zero out rest of page */
+		memset(&erp->er_extbuf[nex1 + nex2], 0, (XFS_IEXT_BUFSZ -
+			((nex1 + nex2) * sizeof(xfs_bmbt_rec_t))));
+		/* Update remaining counters */
+		erp->er_extcount -= ext_diff;
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -ext_diff);
+		ext_cnt -= ext_diff;
+		nex1 = 0;
+		erp_idx++;
+		erp++;
+	}
+	ifp->if_bytes -= count * sizeof(xfs_bmbt_rec_t);
+	xfs_iext_irec_compact(ifp);
+}
+
+/*
+ * Create, destroy, or resize a linear (direct) block of extents.
+ */
+void
+xfs_iext_realloc_direct(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		new_size)	/* new size of extents */
+{
+	int		rnew_size;	/* real new size of extents */
+
+	rnew_size = new_size;
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC) ||
+		((new_size >= 0) && (new_size <= XFS_IEXT_BUFSZ) &&
+		 (new_size != ifp->if_real_bytes)));
+
+	/* Free extent records */
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+	}
+	/* Resize direct extent list and zero any new bytes */
+	else if (ifp->if_real_bytes) {
+		/* Check if extents will fit inside the inode */
+		if (new_size <= XFS_INLINE_EXTS * sizeof(xfs_bmbt_rec_t)) {
+			xfs_iext_direct_to_inline(ifp, new_size /
+				(uint)sizeof(xfs_bmbt_rec_t));
+			ifp->if_bytes = new_size;
+			return;
+		}
+		if ((new_size & (new_size - 1)) != 0) {
+			rnew_size = xfs_iroundup(new_size);
+		}
+		if (rnew_size != ifp->if_real_bytes) {
+			ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+				kmem_realloc(ifp->if_u1.if_extents,
+						rnew_size,
+						ifp->if_real_bytes,
+						KM_SLEEP);
+		}
+		if (rnew_size > ifp->if_real_bytes) {
+			memset(&ifp->if_u1.if_extents[ifp->if_bytes /
+				(uint)sizeof(xfs_bmbt_rec_t)], 0,
+				rnew_size - ifp->if_real_bytes);
+		}
+	}
+	/*
+	 * Switch from the inline extent buffer to a direct
+	 * extent list. Be sure to include the inline extent
+	 * bytes in new_size.
+	 */
+	else {
+		new_size += ifp->if_bytes;
+		if ((new_size & (new_size - 1)) != 0) {
+			rnew_size = xfs_iroundup(new_size);
+		}
+		xfs_iext_inline_to_direct(ifp, rnew_size);
+	}
+	ifp->if_real_bytes = rnew_size;
+	ifp->if_bytes = new_size;
+}
+
+/*
+ * Switch from linear (direct) extent records to inline buffer.
+ */
+void
+xfs_iext_direct_to_inline(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	nextents)	/* number of extents in file */
+{
+	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
+	ASSERT(nextents <= XFS_INLINE_EXTS);
+	/*
+	 * The inline buffer was zeroed when we switched
+	 * from inline to direct extent allocation mode,
+	 * so we don't need to clear it here.
+	 */
+	memcpy(ifp->if_u2.if_inline_ext, ifp->if_u1.if_extents,
+		nextents * sizeof(xfs_bmbt_rec_t));
+	kmem_free(ifp->if_u1.if_extents, KM_SLEEP);
+	ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext;
+	ifp->if_real_bytes = 0;
+}
+
+/*
+ * Switch from inline buffer to linear (direct) extent records.
+ * new_size should already be rounded up to the next power of 2
+ * by the caller (when appropriate), so use new_size as it is.
+ * However, since new_size may be rounded up, we can't update
+ * if_bytes here. It is the caller's responsibility to update
+ * if_bytes upon return.
+ */
+void
+xfs_iext_inline_to_direct(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		new_size)	/* number of extents in file */
+{
+	ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+		kmem_alloc(new_size, KM_SLEEP);
+	memset(ifp->if_u1.if_extents, 0, new_size);
+	if (ifp->if_bytes) {
+		memcpy(ifp->if_u1.if_extents, ifp->if_u2.if_inline_ext,
+			ifp->if_bytes);
+		memset(ifp->if_u2.if_inline_ext, 0, XFS_INLINE_EXTS *
+			sizeof(xfs_bmbt_rec_t));
+	}
+	ifp->if_real_bytes = new_size;
+}
+
+/*
+ * Resize an extent indirection array to new_size bytes.
+ */
+void
+xfs_iext_realloc_indirect(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		new_size)	/* new indirection array size */
+{
+	int		nlists;		/* number of irec's (ex lists) */
+	int		size;		/* current indirection array size */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	size = nlists * sizeof(xfs_ext_irec_t);
+	ASSERT(ifp->if_real_bytes);
+	ASSERT((new_size >= 0) && (new_size != size));
+	if (new_size == 0) {
+		xfs_iext_destroy(ifp);
+	} else {
+		ifp->if_u1.if_ext_irec = (xfs_ext_irec_t *)
+			kmem_realloc(ifp->if_u1.if_ext_irec,
+				new_size, size, KM_SLEEP);
+	}
+}
+
+/*
+ * Switch from indirection array to linear (direct) extent allocations.
+ */
+void
+xfs_iext_indirect_to_direct(
+	 xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_bmbt_rec_t	*ep;		/* extent record pointer */
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		size;		/* size of file extents */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT(nextents <= XFS_LINEAR_EXTS);
+	size = nextents * sizeof(xfs_bmbt_rec_t);
+
+	xfs_iext_irec_compact_full(ifp);
+	ASSERT(ifp->if_real_bytes == XFS_IEXT_BUFSZ);
+
+	ep = ifp->if_u1.if_ext_irec->er_extbuf;
+	kmem_free(ifp->if_u1.if_ext_irec, sizeof(xfs_ext_irec_t));
+	ifp->if_flags &= ~XFS_IFEXTIREC;
+	ifp->if_u1.if_extents = ep;
+	ifp->if_bytes = size;
+	if (nextents < XFS_LINEAR_EXTS) {
+		xfs_iext_realloc_direct(ifp, size);
+	}
+}
+
+/*
+ * Free incore file extents.
+ */
+void
+xfs_iext_destroy(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	if (ifp->if_flags & XFS_IFEXTIREC) {
+		int	erp_idx;
+		int	nlists;
+
+		nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+		for (erp_idx = nlists - 1; erp_idx >= 0 ; erp_idx--) {
+			xfs_iext_irec_remove(ifp, erp_idx);
+		}
+		ifp->if_flags &= ~XFS_IFEXTIREC;
+	} else if (ifp->if_real_bytes) {
+		kmem_free(ifp->if_u1.if_extents, ifp->if_real_bytes);
+	} else if (ifp->if_bytes) {
+		memset(ifp->if_u2.if_inline_ext, 0, XFS_INLINE_EXTS *
+			sizeof(xfs_bmbt_rec_t));
+	}
+	ifp->if_u1.if_extents = NULL;
+	ifp->if_real_bytes = 0;
+	ifp->if_bytes = 0;
+}
+
+/*
+ * Return a pointer to the extent record for file system block bno.
+ */
+xfs_bmbt_rec_t *			/* pointer to found extent record */
+xfs_iext_bno_to_ext(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fileoff_t	bno,		/* block number to search for */
+	xfs_extnum_t	*idxp)		/* index of target extent */
+{
+	xfs_bmbt_rec_t	*base;		/* pointer to first extent */
+	xfs_filblks_t	blockcount = 0;	/* number of blocks in extent */
+	xfs_bmbt_rec_t	*ep = NULL;	/* pointer to target extent */
+	xfs_ext_irec_t	*erp = NULL;	/* indirection array pointer */
+	int		high;		/* upper boundary in search */
+	xfs_extnum_t	idx = 0;	/* index of target extent */
+	int		low;		/* lower boundary in search */
+	xfs_extnum_t	nextents;	/* number of file extents */
+	xfs_fileoff_t	startoff = 0;	/* start offset of extent */
+
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	if (nextents == 0) {
+		*idxp = 0;
+		return NULL;
+	}
+	low = 0;
+	if (ifp->if_flags & XFS_IFEXTIREC) {
+		/* Find target extent list */
+		int	erp_idx = 0;
+		erp = xfs_iext_bno_to_irec(ifp, bno, &erp_idx);
+		base = erp->er_extbuf;
+		high = erp->er_extcount - 1;
+	} else {
+		base = ifp->if_u1.if_extents;
+		high = nextents - 1;
+	}
+	/* Binary search extent records */
+	while (low <= high) {
+		idx = (low + high) >> 1;
+		ep = base + idx;
+		startoff = xfs_bmbt_get_startoff(ep);
+		blockcount = xfs_bmbt_get_blockcount(ep);
+		if (bno < startoff) {
+			high = idx - 1;
+		} else if (bno >= startoff + blockcount) {
+			low = idx + 1;
+		} else {
+			/* Convert back to file-based extent index */
+			if (ifp->if_flags & XFS_IFEXTIREC) {
+				idx += erp->er_extoff;
+			}
+			*idxp = idx;
+			return ep;
+		}
+	}
+	/* Convert back to file-based extent index */
+	if (ifp->if_flags & XFS_IFEXTIREC) {
+		idx += erp->er_extoff;
+	}
+	if (bno >= startoff + blockcount) {
+		if (++idx == nextents) {
+			ep = NULL;
+		} else {
+			ep = xfs_iext_get_ext(ifp, idx);
+		}
+	}
+	*idxp = idx;
+	return ep;
+}
+
+/*
+ * Return a pointer to the indirection array entry containing the
+ * extent record for filesystem block bno. Store the index of the
+ * target irec in *erp_idxp.
+ */
+xfs_ext_irec_t *			/* pointer to found extent record */
+xfs_iext_bno_to_irec(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_fileoff_t	bno,		/* block number to search for */
+	int		*erp_idxp)	/* irec index of target ext list */
+{
+	xfs_ext_irec_t	*erp = NULL;	/* indirection array pointer */
+	xfs_ext_irec_t	*erp_next;	/* next indirection array entry */
+	int		erp_idx;	/* indirection array index */
+	int		nlists;		/* number of extent irec's (lists) */
+	int		high;		/* binary search upper limit */
+	int		low;		/* binary search lower limit */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp_idx = 0;
+	low = 0;
+	high = nlists - 1;
+	while (low <= high) {
+		erp_idx = (low + high) >> 1;
+		erp = &ifp->if_u1.if_ext_irec[erp_idx];
+		erp_next = erp_idx < nlists - 1 ? erp + 1 : NULL;
+		if (bno < xfs_bmbt_get_startoff(erp->er_extbuf)) {
+			high = erp_idx - 1;
+		} else if (erp_next && bno >=
+			   xfs_bmbt_get_startoff(erp_next->er_extbuf)) {
+			low = erp_idx + 1;
+		} else {
+			break;
+		}
+	}
+	*erp_idxp = erp_idx;
+	return erp;
+}
+
+/*
+ * Return a pointer to the indirection array entry containing the
+ * extent record at file extent index *idxp. Store the index of the
+ * target irec in *erp_idxp and store the page index of the target
+ * extent record in *idxp.
+ */
+xfs_ext_irec_t *
+xfs_iext_idx_to_irec(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	xfs_extnum_t	*idxp,		/* extent index (file -> page) */
+	int		*erp_idxp,	/* pointer to target irec */
+	int		realloc)	/* new bytes were just added */
+{
+	xfs_ext_irec_t	*prev;		/* pointer to previous irec */
+	xfs_ext_irec_t	*erp = NULL;	/* pointer to current irec */
+	int		erp_idx;	/* indirection array index */
+	int		nlists;		/* number of irec's (ex lists) */
+	int		high;		/* binary search upper limit */
+	int		low;		/* binary search lower limit */
+	xfs_extnum_t	page_idx = *idxp; /* extent index in target list */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	ASSERT(page_idx >= 0 && page_idx <=
+		ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp_idx = 0;
+	low = 0;
+	high = nlists - 1;
+
+	/* Binary search extent irec's */
+	while (low <= high) {
+		erp_idx = (low + high) >> 1;
+		erp = &ifp->if_u1.if_ext_irec[erp_idx];
+		prev = erp_idx > 0 ? erp - 1 : NULL;
+		if (page_idx < erp->er_extoff || (page_idx == erp->er_extoff &&
+		     realloc && prev && prev->er_extcount < XFS_LINEAR_EXTS)) {
+			high = erp_idx - 1;
+		} else if (page_idx > erp->er_extoff + erp->er_extcount ||
+			   (page_idx == erp->er_extoff + erp->er_extcount &&
+			    !realloc)) {
+			low = erp_idx + 1;
+		} else if (page_idx == erp->er_extoff + erp->er_extcount &&
+			   erp->er_extcount == XFS_LINEAR_EXTS) {
+			ASSERT(realloc);
+			page_idx = 0;
+			erp_idx++;
+			erp = erp_idx < nlists ? erp + 1 : NULL;
+			break;
+		} else {
+			page_idx -= erp->er_extoff;
+			break;
+		}
+	}
+	*idxp = page_idx;
+	*erp_idxp = erp_idx;
+	return(erp);
+}
+
+/*
+ * Allocate and initialize an indirection array once the space needed
+ * for incore extents increases above XFS_IEXT_BUFSZ.
+ */
+void
+xfs_iext_irec_init(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	xfs_extnum_t	nextents;	/* number of extents in file */
+
+	ASSERT(!(ifp->if_flags & XFS_IFEXTIREC));
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+	ASSERT(nextents <= XFS_LINEAR_EXTS);
+
+	erp = (xfs_ext_irec_t *)
+		kmem_alloc(sizeof(xfs_ext_irec_t), KM_SLEEP);
+
+	if (nextents == 0) {
+		ifp->if_u1.if_extents = (xfs_bmbt_rec_t *)
+			kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+	} else if (!ifp->if_real_bytes) {
+		xfs_iext_inline_to_direct(ifp, XFS_IEXT_BUFSZ);
+	} else if (ifp->if_real_bytes < XFS_IEXT_BUFSZ) {
+		xfs_iext_realloc_direct(ifp, XFS_IEXT_BUFSZ);
+	}
+	erp->er_extbuf = ifp->if_u1.if_extents;
+	erp->er_extcount = nextents;
+	erp->er_extoff = 0;
+
+	ifp->if_flags |= XFS_IFEXTIREC;
+	ifp->if_real_bytes = XFS_IEXT_BUFSZ;
+	ifp->if_bytes = nextents * sizeof(xfs_bmbt_rec_t);
+	ifp->if_u1.if_ext_irec = erp;
+
+	return;
+}
+
+/*
+ * Allocate and initialize a new entry in the indirection array.
+ */
+xfs_ext_irec_t *
+xfs_iext_irec_new(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		erp_idx)	/* index for new irec */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	int		i;		/* loop counter */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+
+	/* Resize indirection array */
+	xfs_iext_realloc_indirect(ifp, ++nlists *
+				  sizeof(xfs_ext_irec_t));
+	/*
+	 * Move records down in the array so the
+	 * new page can use erp_idx.
+	 */
+	erp = ifp->if_u1.if_ext_irec;
+	for (i = nlists - 1; i > erp_idx; i--) {
+		memmove(&erp[i], &erp[i-1], sizeof(xfs_ext_irec_t));
+	}
+	ASSERT(i == erp_idx);
+
+	/* Initialize new extent record */
+	erp = ifp->if_u1.if_ext_irec;
+	erp[erp_idx].er_extbuf = (xfs_bmbt_rec_t *)
+		kmem_alloc(XFS_IEXT_BUFSZ, KM_SLEEP);
+	ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ;
+	memset(erp[erp_idx].er_extbuf, 0, XFS_IEXT_BUFSZ);
+	erp[erp_idx].er_extcount = 0;
+	erp[erp_idx].er_extoff = erp_idx > 0 ?
+		erp[erp_idx-1].er_extoff + erp[erp_idx-1].er_extcount : 0;
+	return (&erp[erp_idx]);
+}
+
+/*
+ * Remove a record from the indirection array.
+ */
+void
+xfs_iext_irec_remove(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		erp_idx)	/* irec index to remove */
+{
+	xfs_ext_irec_t	*erp;		/* indirection array pointer */
+	int		i;		/* loop counter */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp = &ifp->if_u1.if_ext_irec[erp_idx];
+	if (erp->er_extbuf) {
+		xfs_iext_irec_update_extoffs(ifp, erp_idx + 1,
+			-erp->er_extcount);
+		kmem_free(erp->er_extbuf, XFS_IEXT_BUFSZ);
+	}
+	/* Compact extent records */
+	erp = ifp->if_u1.if_ext_irec;
+	for (i = erp_idx; i < nlists - 1; i++) {
+		memmove(&erp[i], &erp[i+1], sizeof(xfs_ext_irec_t));
+	}
+	/*
+	 * Manually free the last extent record from the indirection
+	 * array.  A call to xfs_iext_realloc_indirect() with a size
+	 * of zero would result in a call to xfs_iext_destroy() which
+	 * would in turn call this function again, creating a nasty
+	 * infinite loop.
+	 */
+	if (--nlists) {
+		xfs_iext_realloc_indirect(ifp,
+			nlists * sizeof(xfs_ext_irec_t));
+	} else {
+		kmem_free(ifp->if_u1.if_ext_irec,
+			sizeof(xfs_ext_irec_t));
+	}
+	ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ;
+}
+
+/*
+ * This is called to clean up large amounts of unused memory allocated
+ * by the indirection array.  Before compacting anything though, verify
+ * that the indirection array is still needed and switch back to the
+ * linear extent list (or even the inline buffer) if possible.  The
+ * compaction policy is as follows:
+ *
+ *    Full Compaction: Extents fit into a single page (or inline buffer)
+ *    Full Compaction: Extents occupy less than 10% of allocated space
+ * Partial Compaction: Extents occupy > 10% and < 50% of allocated space
+ *      No Compaction: Extents occupy at least 50% of allocated space
+ */
+void
+xfs_iext_irec_compact(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_extnum_t	nextents;	/* number of extents in file */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t);
+
+	if (nextents == 0) {
+		xfs_iext_destroy(ifp);
+	} else if (nextents <= XFS_INLINE_EXTS) {
+		xfs_iext_indirect_to_direct(ifp);
+		xfs_iext_direct_to_inline(ifp, nextents);
+	} else if (nextents <= XFS_LINEAR_EXTS) {
+		xfs_iext_indirect_to_direct(ifp);
+	} else if (nextents < (nlists * XFS_LINEAR_EXTS) >> 3) {
+		xfs_iext_irec_compact_full(ifp);
+	} else if (nextents < (nlists * XFS_LINEAR_EXTS) >> 1) {
+		xfs_iext_irec_compact_pages(ifp);
+	}
+}
+
+/*
+ * Combine extents from neighboring extent pages.
+ */
+void
+xfs_iext_irec_compact_pages(
+	xfs_ifork_t	*ifp)		/* inode fork pointer */
+{
+	xfs_ext_irec_t	*erp, *erp_next;/* pointers to irec entries */
+	int		erp_idx = 0;	/* indirection array index */
+	int		nlists;		/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	while (erp_idx < nlists - 1) {
+		erp = &ifp->if_u1.if_ext_irec[erp_idx];
+		erp_next = erp + 1;
+		if (erp_next->er_extcount <=
+		    (XFS_LINEAR_EXTS - erp->er_extcount)) {
+			memmove(&erp->er_extbuf[erp->er_extcount],
+				erp_next->er_extbuf, erp_next->er_extcount *
+				sizeof(xfs_bmbt_rec_t));
+			erp->er_extcount += erp_next->er_extcount;
+			/*
+			 * Free page before removing extent record
+			 * so er_extoffs don't get modified in
+			 * xfs_iext_irec_remove.
+			 */
+			kmem_free(erp_next->er_extbuf, XFS_IEXT_BUFSZ);
+			erp_next->er_extbuf = NULL;
+			xfs_iext_irec_remove(ifp, erp_idx + 1);
+			nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+		} else {
+			erp_idx++;
+		}
+	}
+}
+
+/*
+ * Fully compact the extent records managed by the indirection array.
+ */
+void
+xfs_iext_irec_compact_full(
+	xfs_ifork_t	*ifp)			/* inode fork pointer */
+{
+	xfs_bmbt_rec_t	*ep, *ep_next;		/* extent record pointers */
+	xfs_ext_irec_t	*erp, *erp_next;	/* extent irec pointers */
+	int		erp_idx = 0;		/* extent irec index */
+	int		ext_avail;		/* empty entries in ex list */
+	int		ext_diff;		/* number of exts to add */
+	int		nlists;			/* number of irec's (ex lists) */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	erp = ifp->if_u1.if_ext_irec;
+	ep = &erp->er_extbuf[erp->er_extcount];
+	erp_next = erp + 1;
+	ep_next = erp_next->er_extbuf;
+	while (erp_idx < nlists - 1) {
+		ext_avail = XFS_LINEAR_EXTS - erp->er_extcount;
+		ext_diff = MIN(ext_avail, erp_next->er_extcount);
+		memcpy(ep, ep_next, ext_diff * sizeof(xfs_bmbt_rec_t));
+		erp->er_extcount += ext_diff;
+		erp_next->er_extcount -= ext_diff;
+		/* Remove next page */
+		if (erp_next->er_extcount == 0) {
+			/*
+			 * Free page before removing extent record
+			 * so er_extoffs don't get modified in
+			 * xfs_iext_irec_remove.
+			 */
+			kmem_free(erp_next->er_extbuf,
+				erp_next->er_extcount * sizeof(xfs_bmbt_rec_t));
+			erp_next->er_extbuf = NULL;
+			xfs_iext_irec_remove(ifp, erp_idx + 1);
+			erp = &ifp->if_u1.if_ext_irec[erp_idx];
+			nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+		/* Update next page */
+		} else {
+			/* Move rest of page up to become next new page */
+			memmove(erp_next->er_extbuf, ep_next,
+				erp_next->er_extcount * sizeof(xfs_bmbt_rec_t));
+			ep_next = erp_next->er_extbuf;
+			memset(&ep_next[erp_next->er_extcount], 0,
+				(XFS_LINEAR_EXTS - erp_next->er_extcount) *
+				sizeof(xfs_bmbt_rec_t));
+		}
+		if (erp->er_extcount == XFS_LINEAR_EXTS) {
+			erp_idx++;
+			if (erp_idx < nlists)
+				erp = &ifp->if_u1.if_ext_irec[erp_idx];
+			else
+				break;
+		}
+		ep = &erp->er_extbuf[erp->er_extcount];
+		erp_next = erp + 1;
+		ep_next = erp_next->er_extbuf;
+	}
+}
+
+/*
+ * This is called to update the er_extoff field in the indirection
+ * array when extents have been added or removed from one of the
+ * extent lists. erp_idx contains the irec index to begin updating
+ * at and ext_diff contains the number of extents that were added
+ * or removed.
+ */
+void
+xfs_iext_irec_update_extoffs(
+	xfs_ifork_t	*ifp,		/* inode fork pointer */
+	int		erp_idx,	/* irec index to update */
+	int		ext_diff)	/* number of new extents */
+{
+	int		i;		/* loop counter */
+	int		nlists;		/* number of irec's (ex lists */
+
+	ASSERT(ifp->if_flags & XFS_IFEXTIREC);
+	nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ;
+	for (i = erp_idx; i < nlists; i++) {
+		ifp->if_u1.if_ext_irec[i].er_extoff += ext_diff;
+	}
+}
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 1cfbcf18ce86..39ef9c36ea55 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -25,10 +25,37 @@
 #define	XFS_ATTR_FORK	1
 
 /*
+ * The following xfs_ext_irec_t struct introduces a second (top) level
+ * to the in-core extent allocation scheme. These structs are allocated
+ * in a contiguous block, creating an indirection array where each entry
+ * (irec) contains a pointer to a buffer of in-core extent records which
+ * it manages. Each extent buffer is 4k in size, since 4k is the system
+ * page size on Linux i386 and systems with larger page sizes don't seem
+ * to gain much, if anything, by using their native page size as the
+ * extent buffer size. Also, using 4k extent buffers everywhere provides
+ * a consistent interface for CXFS across different platforms.
+ *
+ * There is currently no limit on the number of irec's (extent lists)
+ * allowed, so heavily fragmented files may require an indirection array
+ * which spans multiple system pages of memory. The number of extents
+ * which would require this amount of contiguous memory is very large
+ * and should not cause problems in the foreseeable future. However,
+ * if the memory needed for the contiguous array ever becomes a problem,
+ * it is possible that a third level of indirection may be required.
+ */
+typedef struct xfs_ext_irec {
+	xfs_bmbt_rec_t	*er_extbuf;	/* block of extent records */
+	xfs_extnum_t	er_extoff;	/* extent offset in file */
+	xfs_extnum_t	er_extcount;	/* number of extents in page/block */
+} xfs_ext_irec_t;
+
+/*
  * File incore extent information, present for each of data & attr forks.
  */
-#define	XFS_INLINE_EXTS	2
-#define	XFS_INLINE_DATA	32
+#define	XFS_IEXT_BUFSZ		4096
+#define	XFS_LINEAR_EXTS		(XFS_IEXT_BUFSZ / (uint)sizeof(xfs_bmbt_rec_t))
+#define	XFS_INLINE_EXTS		2
+#define	XFS_INLINE_DATA		32
 typedef struct xfs_ifork {
 	int			if_bytes;	/* bytes in if_u1 */
 	int			if_real_bytes;	/* bytes allocated in if_u1 */
@@ -39,6 +66,7 @@ typedef struct xfs_ifork {
 	xfs_extnum_t		if_lastex;	/* last if_extents used */
 	union {
 		xfs_bmbt_rec_t	*if_extents;	/* linear map file exts */
+		xfs_ext_irec_t	*if_ext_irec;	/* irec map file exts */
 		char		*if_data;	/* inline file data */
 	} if_u1;
 	union {
@@ -61,20 +89,16 @@ typedef struct xfs_ifork {
 /*
  * Per-fork incore inode flags.
  */
-#define	XFS_IFINLINE	0x0001	/* Inline data is read in */
-#define	XFS_IFEXTENTS	0x0002	/* All extent pointers are read in */
-#define	XFS_IFBROOT	0x0004	/* i_broot points to the bmap b-tree root */
+#define	XFS_IFINLINE	0x01	/* Inline data is read in */
+#define	XFS_IFEXTENTS	0x02	/* All extent pointers are read in */
+#define	XFS_IFBROOT	0x04	/* i_broot points to the bmap b-tree root */
+#define	XFS_IFEXTIREC	0x08	/* Indirection array of extent blocks */
 
 /*
- * Flags for xfs_imap() and xfs_dilocate().
+ * Flags for xfs_itobp(), xfs_imap() and xfs_dilocate().
  */
-#define	XFS_IMAP_LOOKUP		0x1
-
-/*
- * Maximum number of extent pointers in if_u1.if_extents.
- */
-#define	XFS_MAX_INCORE_EXTENTS	32768
-
+#define XFS_IMAP_LOOKUP		0x1
+#define XFS_IMAP_BULKSTAT	0x2
 
 #ifdef __KERNEL__
 struct bhv_desc;
@@ -398,7 +422,7 @@ int		xfs_finish_reclaim_all(struct xfs_mount *, int);
  */
 int		xfs_itobp(struct xfs_mount *, struct xfs_trans *,
 			  xfs_inode_t *, xfs_dinode_t **, struct xfs_buf **,
-			  xfs_daddr_t);
+			  xfs_daddr_t, uint);
 int		xfs_iread(struct xfs_mount *, struct xfs_trans *, xfs_ino_t,
 			  xfs_inode_t **, xfs_daddr_t);
 int		xfs_iread_extents(struct xfs_trans *, xfs_inode_t *, int);
@@ -440,6 +464,32 @@ xfs_inode_t	*xfs_vtoi(struct vnode *vp);
 
 void		xfs_synchronize_atime(xfs_inode_t *);
 
+xfs_bmbt_rec_t	*xfs_iext_get_ext(xfs_ifork_t *, xfs_extnum_t);
+void		xfs_iext_insert(xfs_ifork_t *, xfs_extnum_t, xfs_extnum_t,
+				xfs_bmbt_irec_t *);
+void		xfs_iext_add(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_add_indirect_multi(xfs_ifork_t *, int, xfs_extnum_t, int);
+void		xfs_iext_remove(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_remove_inline(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_remove_direct(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_remove_indirect(xfs_ifork_t *, xfs_extnum_t, int);
+void		xfs_iext_realloc_direct(xfs_ifork_t *, int);
+void		xfs_iext_realloc_indirect(xfs_ifork_t *, int);
+void		xfs_iext_indirect_to_direct(xfs_ifork_t *);
+void		xfs_iext_direct_to_inline(xfs_ifork_t *, xfs_extnum_t);
+void		xfs_iext_inline_to_direct(xfs_ifork_t *, int);
+void		xfs_iext_destroy(xfs_ifork_t *);
+xfs_bmbt_rec_t	*xfs_iext_bno_to_ext(xfs_ifork_t *, xfs_fileoff_t, int *);
+xfs_ext_irec_t	*xfs_iext_bno_to_irec(xfs_ifork_t *, xfs_fileoff_t, int *);
+xfs_ext_irec_t	*xfs_iext_idx_to_irec(xfs_ifork_t *, xfs_extnum_t *, int *, int);
+void		xfs_iext_irec_init(xfs_ifork_t *);
+xfs_ext_irec_t *xfs_iext_irec_new(xfs_ifork_t *, int);
+void		xfs_iext_irec_remove(xfs_ifork_t *, int);
+void		xfs_iext_irec_compact(xfs_ifork_t *);
+void		xfs_iext_irec_compact_pages(xfs_ifork_t *);
+void		xfs_iext_irec_compact_full(xfs_ifork_t *);
+void		xfs_iext_irec_update_extoffs(xfs_ifork_t *, int, int);
+
 #define xfs_ipincount(ip)	((unsigned int) atomic_read(&ip->i_pincount))
 
 #ifdef DEBUG
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 36aa1fcb90a5..7497a481b2f5 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -580,7 +580,7 @@ xfs_inode_item_unpin_remove(
  * been or is in the process of being flushed, then (ideally) we'd like to
  * see if the inode's buffer is still incore, and if so give it a nudge.
  * We delay doing so until the pushbuf routine, though, to avoid holding
- * the AIL lock across a call to the blackhole which is the buffercache.
+ * the AIL lock across a call to the blackhole which is the buffer cache.
  * Also we don't want to sleep in any device strategy routines, which can happen
  * if we do the subsequent bawrite in here.
  */
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index 788917f355c4..d5dfedcb8922 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -76,7 +76,7 @@ xfs_iomap_enter_trace(
 		(void *)((unsigned long)count),
 		(void *)((unsigned long)((io->io_new_size >> 32) & 0xffffffff)),
 		(void *)((unsigned long)(io->io_new_size & 0xffffffff)),
-		(void *)NULL,
+		(void *)((unsigned long)current_pid()),
 		(void *)NULL,
 		(void *)NULL,
 		(void *)NULL,
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index c59450e1be40..94068d014f27 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -272,7 +272,7 @@ xfs_bulkstat(
 	size_t			statstruct_size, /* sizeof struct filling */
 	char			__user *ubuffer, /* buffer with inode stats */
 	int			flags,	/* defined in xfs_itable.h */
-	int			*done)	/* 1 if there're more stats to get */
+	int			*done)	/* 1 if there are more stats to get */
 {
 	xfs_agblock_t		agbno=0;/* allocation group block number */
 	xfs_buf_t		*agbp;	/* agi header buffer */
@@ -562,7 +562,8 @@ xfs_bulkstat(
 						if (bp)
 							xfs_buf_relse(bp);
 						error = xfs_itobp(mp, NULL, ip,
-								  &dip, &bp, bno);
+								&dip, &bp, bno,
+								XFS_IMAP_BULKSTAT);
 						if (!error)
 							clustidx = ip->i_boffset / mp->m_sb.sb_inodesize;
 						kmem_zone_free(xfs_inode_zone, ip);
@@ -570,6 +571,8 @@ xfs_bulkstat(
 								   mp, XFS_ERRTAG_BULKSTAT_READ_CHUNK,
 								   XFS_RANDOM_BULKSTAT_READ_CHUNK)) {
 							bp = NULL;
+							ubleft = 0;
+							rval = error;
 							break;
 						}
 					}
@@ -673,7 +676,7 @@ xfs_bulkstat_single(
 	xfs_mount_t		*mp,	/* mount point for filesystem */
 	xfs_ino_t		*lastinop, /* inode to return */
 	char			__user *buffer, /* buffer with inode stats */
-	int			*done)	/* 1 if there're more stats to get */
+	int			*done)	/* 1 if there are more stats to get */
 {
 	int			count;	/* count value for bulkstat call */
 	int			error;	/* return value */
diff --git a/fs/xfs/xfs_itable.h b/fs/xfs/xfs_itable.h
index 047d834ed210..11eb4e1b18c4 100644
--- a/fs/xfs/xfs_itable.h
+++ b/fs/xfs/xfs_itable.h
@@ -60,7 +60,7 @@ xfs_bulkstat(
 	size_t		statstruct_size,/* sizeof struct that we're filling */
 	char		__user *ubuffer,/* buffer with inode stats */
 	int		flags,		/* flag to control access method */
-	int		*done);		/* 1 if there're more stats to get */
+	int		*done);		/* 1 if there are more stats to get */
 
 int
 xfs_bulkstat_single(
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index 9176995160ed..32e841d2f26d 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -59,7 +59,7 @@ STATIC xlog_t *  xlog_alloc_log(xfs_mount_t	*mp,
 				int		num_bblks);
 STATIC int	 xlog_space_left(xlog_t *log, int cycle, int bytes);
 STATIC int	 xlog_sync(xlog_t *log, xlog_in_core_t *iclog);
-STATIC void	 xlog_unalloc_log(xlog_t *log);
+STATIC void	 xlog_dealloc_log(xlog_t *log);
 STATIC int	 xlog_write(xfs_mount_t *mp, xfs_log_iovec_t region[],
 			    int nentries, xfs_log_ticket_t tic,
 			    xfs_lsn_t *start_lsn,
@@ -304,7 +304,7 @@ xfs_log_done(xfs_mount_t	*mp,
 	if ((ticket->t_flags & XLOG_TIC_PERM_RESERV) == 0 ||
 	    (flags & XFS_LOG_REL_PERM_RESERV)) {
 		/*
-		 * Release ticket if not permanent reservation or a specifc
+		 * Release ticket if not permanent reservation or a specific
 		 * request has been made to release a permanent reservation.
 		 */
 		xlog_trace_loggrant(log, ticket, "xfs_log_done: (non-permanent)");
@@ -511,7 +511,7 @@ xfs_log_mount(xfs_mount_t	*mp,
 			vfsp->vfs_flag |= VFS_RDONLY;
 		if (error) {
 			cmn_err(CE_WARN, "XFS: log mount/recovery failed: error %d", error);
-			xlog_unalloc_log(mp->m_log);
+			xlog_dealloc_log(mp->m_log);
 			return error;
 		}
 	}
@@ -667,7 +667,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 		 *
 		 * Go through the motions of sync'ing and releasing
 		 * the iclog, even though no I/O will actually happen,
-		 * we need to wait for other log I/O's that may already
+		 * we need to wait for other log I/Os that may already
 		 * be in progress.  Do this as a separate section of
 		 * code so we'll know if we ever get stuck here that
 		 * we're in this odd situation of trying to unmount
@@ -704,7 +704,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 void
 xfs_log_unmount_dealloc(xfs_mount_t *mp)
 {
-	xlog_unalloc_log(mp->m_log);
+	xlog_dealloc_log(mp->m_log);
 }
 
 /*
@@ -1492,7 +1492,7 @@ xlog_sync(xlog_t		*log,
 		ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);
 		ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);
 
-		/* account for internal log which does't start at block #0 */
+		/* account for internal log which doesn't start at block #0 */
 		XFS_BUF_SET_ADDR(bp, XFS_BUF_ADDR(bp) + log->l_logBBstart);
 		XFS_BUF_WRITE(bp);
 		if ((error = XFS_bwrite(bp))) {
@@ -1506,10 +1506,10 @@ xlog_sync(xlog_t		*log,
 
 
 /*
- * Unallocate a log structure
+ * Deallocate a log structure
  */
 void
-xlog_unalloc_log(xlog_t *log)
+xlog_dealloc_log(xlog_t *log)
 {
 	xlog_in_core_t	*iclog, *next_iclog;
 	xlog_ticket_t	*tic, *next_tic;
@@ -1539,7 +1539,7 @@ xlog_unalloc_log(xlog_t *log)
 	if ((log->l_ticket_cnt != log->l_ticket_tcnt)  &&
 	    !XLOG_FORCED_SHUTDOWN(log)) {
 		xfs_fs_cmn_err(CE_WARN, log->l_mp,
-			"xlog_unalloc_log: (cnt: %d, total: %d)",
+			"xlog_dealloc_log: (cnt: %d, total: %d)",
 			log->l_ticket_cnt, log->l_ticket_tcnt);
 		/* ASSERT(log->l_ticket_cnt == log->l_ticket_tcnt); */
 
@@ -1562,7 +1562,7 @@ xlog_unalloc_log(xlog_t *log)
 #endif
 	log->l_mp->m_log = NULL;
 	kmem_free(log, sizeof(xlog_t));
-}	/* xlog_unalloc_log */
+}	/* xlog_dealloc_log */
 
 /*
  * Update counters atomically now that memcpy is done.
@@ -2829,7 +2829,7 @@ xlog_state_release_iclog(xlog_t		*log,
 
 	/*
 	 * We let the log lock go, so it's possible that we hit a log I/O
-	 * error or someother SHUTDOWN condition that marks the iclog
+	 * error or some other SHUTDOWN condition that marks the iclog
 	 * as XLOG_STATE_IOERROR before the bwrite. However, we know that
 	 * this iclog has consistent data, so we ignore IOERROR
 	 * flags after this point.
diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h
index 4b2ac88dbb83..eacb3d4987f2 100644
--- a/fs/xfs/xfs_log.h
+++ b/fs/xfs/xfs_log.h
@@ -27,7 +27,7 @@
 
 #ifdef __KERNEL__
 /*
- * By comparing each compnent, we don't have to worry about extra
+ * By comparing each component, we don't have to worry about extra
  * endian issues in treating two 32 bit numbers as one 64 bit number
  */
 static inline xfs_lsn_t	_lsn_cmp(xfs_lsn_t lsn1, xfs_lsn_t lsn2)
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 7d46cbd6a07a..1f0016b0b4ec 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -583,7 +583,7 @@ xlog_find_head(
 		 *        x | x ... | x - 1 | x
 		 * Another case that fits this picture would be
 		 *        x | x + 1 | x ... | x
-		 * In this case the head really is somwhere at the end of the
+		 * In this case the head really is somewhere at the end of the
 		 * log, as one of the latest writes at the beginning was
 		 * incomplete.
 		 * One more case is
@@ -2799,7 +2799,7 @@ xlog_recover_do_trans(
 		 * we don't need to worry about the block number being
 		 * truncated in > 1 TB buffers because in user-land,
 		 * we're now n32 or 64-bit so xfs_daddr_t is 64-bits so
-		 * the blkno's will get through the user-mode buffer
+		 * the blknos will get through the user-mode buffer
 		 * cache properly.  The only bad case is o32 kernels
 		 * where xfs_daddr_t is 32-bits but mount will warn us
 		 * off a > 1 TB filesystem before we get here.
@@ -3249,7 +3249,7 @@ xlog_recover_process_iunlinks(
 					 * next inode in the bucket.
 					 */
 					error = xfs_itobp(mp, NULL, ip, &dip,
-							&ibp, 0);
+							&ibp, 0, 0);
 					ASSERT(error || (dip != NULL));
 				}
 
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 62188ea392c7..72e7e78bfff8 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -51,11 +51,32 @@ STATIC int	xfs_uuid_mount(xfs_mount_t *);
 STATIC void	xfs_uuid_unmount(xfs_mount_t *mp);
 STATIC void	xfs_unmountfs_wait(xfs_mount_t *);
 
+
+#ifdef HAVE_PERCPU_SB
+STATIC void	xfs_icsb_destroy_counters(xfs_mount_t *);
+STATIC void	xfs_icsb_balance_counter(xfs_mount_t *, xfs_sb_field_t, int);
+STATIC void	xfs_icsb_sync_counters(xfs_mount_t *);
+STATIC int	xfs_icsb_modify_counters(xfs_mount_t *, xfs_sb_field_t,
+						int, int);
+STATIC int	xfs_icsb_modify_counters_locked(xfs_mount_t *, xfs_sb_field_t,
+						int, int);
+STATIC int	xfs_icsb_disable_counter(xfs_mount_t *, xfs_sb_field_t);
+
+#else
+
+#define xfs_icsb_destroy_counters(mp)			do { } while (0)
+#define xfs_icsb_balance_counter(mp, a, b)		do { } while (0)
+#define xfs_icsb_sync_counters(mp)			do { } while (0)
+#define xfs_icsb_modify_counters(mp, a, b, c)		do { } while (0)
+#define xfs_icsb_modify_counters_locked(mp, a, b, c)	do { } while (0)
+
+#endif
+
 static const struct {
-    short offset;
-    short type;     /* 0 = integer
-		* 1 = binary / string (no translation)
-		*/
+	short offset;
+	short type;	/* 0 = integer
+			 * 1 = binary / string (no translation)
+			 */
 } xfs_sb_info[] = {
     { offsetof(xfs_sb_t, sb_magicnum),   0 },
     { offsetof(xfs_sb_t, sb_blocksize),  0 },
@@ -113,7 +134,11 @@ xfs_mount_init(void)
 {
 	xfs_mount_t *mp;
 
-	mp = kmem_zalloc(sizeof(*mp), KM_SLEEP);
+	mp = kmem_zalloc(sizeof(xfs_mount_t), KM_SLEEP);
+
+	if (xfs_icsb_init_counters(mp)) {
+		mp->m_flags |= XFS_MOUNT_NO_PERCPU_SB;
+	}
 
 	AIL_LOCKINIT(&mp->m_ail_lock, "xfs_ail");
 	spinlock_init(&mp->m_sb_lock, "xfs_sb");
@@ -136,8 +161,8 @@ xfs_mount_init(void)
  */
 void
 xfs_mount_free(
-	xfs_mount_t *mp,
-	int	    remove_bhv)
+	xfs_mount_t	*mp,
+	int		remove_bhv)
 {
 	if (mp->m_ihash)
 		xfs_ihash_free(mp);
@@ -177,6 +202,7 @@ xfs_mount_free(
 		VFS_REMOVEBHV(vfsp, &mp->m_bhv);
 	}
 
+	xfs_icsb_destroy_counters(mp);
 	kmem_free(mp, sizeof(xfs_mount_t));
 }
 
@@ -242,9 +268,12 @@ xfs_mount_validate_sb(
 	    sbp->sb_blocklog > XFS_MAX_BLOCKSIZE_LOG			||
 	    sbp->sb_inodesize < XFS_DINODE_MIN_SIZE			||
 	    sbp->sb_inodesize > XFS_DINODE_MAX_SIZE			||
+	    sbp->sb_inodelog < XFS_DINODE_MIN_LOG			||
+	    sbp->sb_inodelog > XFS_DINODE_MAX_LOG			||
+	    (sbp->sb_blocklog - sbp->sb_inodelog != sbp->sb_inopblog)	||
 	    (sbp->sb_rextsize * sbp->sb_blocksize > XFS_MAX_RTEXTSIZE)	||
 	    (sbp->sb_rextsize * sbp->sb_blocksize < XFS_MIN_RTEXTSIZE)	||
-	    sbp->sb_imax_pct > 100)) {
+	    (sbp->sb_imax_pct > 100 || sbp->sb_imax_pct < 1))) {
 		cmn_err(CE_WARN, "XFS: SB sanity check 1 failed");
 		XFS_CORRUPTION_ERROR("xfs_mount_validate_sb(3)",
 				     XFS_ERRLEVEL_LOW, mp, sbp);
@@ -364,7 +393,7 @@ xfs_initialize_perag(
 				break;
 			}
 
-			/* This ag is prefered for inodes */
+			/* This ag is preferred for inodes */
 			pag = &mp->m_perag[index];
 			pag->pagi_inodeok = 1;
 			if (index < max_metadata)
@@ -527,6 +556,10 @@ xfs_readsb(xfs_mount_t *mp)
 		ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
 	}
 
+	xfs_icsb_balance_counter(mp, XFS_SBS_ICOUNT, 0);
+	xfs_icsb_balance_counter(mp, XFS_SBS_IFREE, 0);
+	xfs_icsb_balance_counter(mp, XFS_SBS_FDBLOCKS, 0);
+
 	mp->m_sb_bp = bp;
 	xfs_buf_relse(bp);
 	ASSERT(XFS_BUF_VALUSEMA(bp) > 0);
@@ -1154,6 +1187,9 @@ xfs_unmountfs_writesb(xfs_mount_t *mp)
 	sbp = xfs_getsb(mp, 0);
 	if (!(XFS_MTOVFS(mp)->vfs_flag & VFS_RDONLY ||
 		XFS_FORCED_SHUTDOWN(mp))) {
+
+		xfs_icsb_sync_counters(mp);
+
 		/*
 		 * mark shared-readonly if desired
 		 */
@@ -1227,7 +1263,6 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
 
 	xfs_trans_log_buf(tp, bp, first, last);
 }
-
 /*
  * xfs_mod_incore_sb_unlocked() is a utility routine common used to apply
  * a delta to a specified field in the in-core superblock.  Simply
@@ -1237,7 +1272,7 @@ xfs_mod_sb(xfs_trans_t *tp, __int64_t fields)
  *
  * The SB_LOCK must be held when this routine is called.
  */
-STATIC int
+int
 xfs_mod_incore_sb_unlocked(xfs_mount_t *mp, xfs_sb_field_t field,
 			int delta, int rsvd)
 {
@@ -1406,9 +1441,26 @@ xfs_mod_incore_sb(xfs_mount_t *mp, xfs_sb_field_t field, int delta, int rsvd)
 	unsigned long	s;
 	int	status;
 
-	s = XFS_SB_LOCK(mp);
-	status = xfs_mod_incore_sb_unlocked(mp, field, delta, rsvd);
-	XFS_SB_UNLOCK(mp, s);
+	/* check for per-cpu counters */
+	switch (field) {
+#ifdef HAVE_PERCPU_SB
+	case XFS_SBS_ICOUNT:
+	case XFS_SBS_IFREE:
+	case XFS_SBS_FDBLOCKS:
+		if (!(mp->m_flags & XFS_MOUNT_NO_PERCPU_SB)) {
+			status = xfs_icsb_modify_counters(mp, field,
+							delta, rsvd);
+			break;
+		}
+		/* FALLTHROUGH */
+#endif
+	default:
+		s = XFS_SB_LOCK(mp);
+		status = xfs_mod_incore_sb_unlocked(mp, field, delta, rsvd);
+		XFS_SB_UNLOCK(mp, s);
+		break;
+	}
+
 	return status;
 }
 
@@ -1445,8 +1497,26 @@ xfs_mod_incore_sb_batch(xfs_mount_t *mp, xfs_mod_sb_t *msb, uint nmsb, int rsvd)
 		 * from the loop so we'll fall into the undo loop
 		 * below.
 		 */
-		status = xfs_mod_incore_sb_unlocked(mp, msbp->msb_field,
-						    msbp->msb_delta, rsvd);
+		switch (msbp->msb_field) {
+#ifdef HAVE_PERCPU_SB
+		case XFS_SBS_ICOUNT:
+		case XFS_SBS_IFREE:
+		case XFS_SBS_FDBLOCKS:
+			if (!(mp->m_flags & XFS_MOUNT_NO_PERCPU_SB)) {
+				status = xfs_icsb_modify_counters_locked(mp,
+							msbp->msb_field,
+							msbp->msb_delta, rsvd);
+				break;
+			}
+			/* FALLTHROUGH */
+#endif
+		default:
+			status = xfs_mod_incore_sb_unlocked(mp,
+						msbp->msb_field,
+						msbp->msb_delta, rsvd);
+			break;
+		}
+
 		if (status != 0) {
 			break;
 		}
@@ -1463,8 +1533,28 @@ xfs_mod_incore_sb_batch(xfs_mount_t *mp, xfs_mod_sb_t *msb, uint nmsb, int rsvd)
 	if (status != 0) {
 		msbp--;
 		while (msbp >= msb) {
-			status = xfs_mod_incore_sb_unlocked(mp,
-				    msbp->msb_field, -(msbp->msb_delta), rsvd);
+			switch (msbp->msb_field) {
+#ifdef HAVE_PERCPU_SB
+			case XFS_SBS_ICOUNT:
+			case XFS_SBS_IFREE:
+			case XFS_SBS_FDBLOCKS:
+				if (!(mp->m_flags & XFS_MOUNT_NO_PERCPU_SB)) {
+					status =
+					    xfs_icsb_modify_counters_locked(mp,
+							msbp->msb_field,
+							-(msbp->msb_delta),
+							rsvd);
+					break;
+				}
+				/* FALLTHROUGH */
+#endif
+			default:
+				status = xfs_mod_incore_sb_unlocked(mp,
+							msbp->msb_field,
+							-(msbp->msb_delta),
+							rsvd);
+				break;
+			}
 			ASSERT(status == 0);
 			msbp--;
 		}
@@ -1577,3 +1667,525 @@ xfs_mount_log_sbunit(
 	xfs_mod_sb(tp, fields);
 	xfs_trans_commit(tp, 0, NULL);
 }
+
+
+#ifdef HAVE_PERCPU_SB
+/*
+ * Per-cpu incore superblock counters
+ *
+ * Simple concept, difficult implementation
+ *
+ * Basically, replace the incore superblock counters with a distributed per cpu
+ * counter for contended fields (e.g.  free block count).
+ *
+ * Difficulties arise in that the incore sb is used for ENOSPC checking, and
+ * hence needs to be accurately read when we are running low on space. Hence
+ * there is a method to enable and disable the per-cpu counters based on how
+ * much "stuff" is available in them.
+ *
+ * Basically, a counter is enabled if there is enough free resource to justify
+ * running a per-cpu fast-path. If the per-cpu counter runs out (i.e. a local
+ * ENOSPC), then we disable the counters to synchronise all callers and
+ * re-distribute the available resources.
+ *
+ * If, once we redistributed the available resources, we still get a failure,
+ * we disable the per-cpu counter and go through the slow path.
+ *
+ * The slow path is the current xfs_mod_incore_sb() function.  This means that
+ * when we disable a per-cpu counter, we need to drain it's resources back to
+ * the global superblock. We do this after disabling the counter to prevent
+ * more threads from queueing up on the counter.
+ *
+ * Essentially, this means that we still need a lock in the fast path to enable
+ * synchronisation between the global counters and the per-cpu counters. This
+ * is not a problem because the lock will be local to a CPU almost all the time
+ * and have little contention except when we get to ENOSPC conditions.
+ *
+ * Basically, this lock becomes a barrier that enables us to lock out the fast
+ * path while we do things like enabling and disabling counters and
+ * synchronising the counters.
+ *
+ * Locking rules:
+ *
+ * 	1. XFS_SB_LOCK() before picking up per-cpu locks
+ * 	2. per-cpu locks always picked up via for_each_online_cpu() order
+ * 	3. accurate counter sync requires XFS_SB_LOCK + per cpu locks
+ * 	4. modifying per-cpu counters requires holding per-cpu lock
+ * 	5. modifying global counters requires holding XFS_SB_LOCK
+ *	6. enabling or disabling a counter requires holding the XFS_SB_LOCK
+ *	   and _none_ of the per-cpu locks.
+ *
+ * Disabled counters are only ever re-enabled by a balance operation
+ * that results in more free resources per CPU than a given threshold.
+ * To ensure counters don't remain disabled, they are rebalanced when
+ * the global resource goes above a higher threshold (i.e. some hysteresis
+ * is present to prevent thrashing).
+ */
+
+/*
+ * hot-plug CPU notifier support.
+ *
+ * We cannot use the hotcpu_register() function because it does
+ * not allow notifier instances. We need a notifier per filesystem
+ * as we need to be able to identify the filesystem to balance
+ * the counters out. This is achieved by having a notifier block
+ * embedded in the xfs_mount_t and doing pointer magic to get the
+ * mount pointer from the notifier block address.
+ */
+STATIC int
+xfs_icsb_cpu_notify(
+	struct notifier_block *nfb,
+	unsigned long action,
+	void *hcpu)
+{
+	xfs_icsb_cnts_t *cntp;
+	xfs_mount_t	*mp;
+	int		s;
+
+	mp = (xfs_mount_t *)container_of(nfb, xfs_mount_t, m_icsb_notifier);
+	cntp = (xfs_icsb_cnts_t *)
+			per_cpu_ptr(mp->m_sb_cnts, (unsigned long)hcpu);
+	switch (action) {
+	case CPU_UP_PREPARE:
+		/* Easy Case - initialize the area and locks, and
+		 * then rebalance when online does everything else for us. */
+		memset(cntp, 0, sizeof(xfs_icsb_cnts_t));
+		break;
+	case CPU_ONLINE:
+		xfs_icsb_balance_counter(mp, XFS_SBS_ICOUNT, 0);
+		xfs_icsb_balance_counter(mp, XFS_SBS_IFREE, 0);
+		xfs_icsb_balance_counter(mp, XFS_SBS_FDBLOCKS, 0);
+		break;
+	case CPU_DEAD:
+		/* Disable all the counters, then fold the dead cpu's
+		 * count into the total on the global superblock and
+		 * re-enable the counters. */
+		s = XFS_SB_LOCK(mp);
+		xfs_icsb_disable_counter(mp, XFS_SBS_ICOUNT);
+		xfs_icsb_disable_counter(mp, XFS_SBS_IFREE);
+		xfs_icsb_disable_counter(mp, XFS_SBS_FDBLOCKS);
+
+		mp->m_sb.sb_icount += cntp->icsb_icount;
+		mp->m_sb.sb_ifree += cntp->icsb_ifree;
+		mp->m_sb.sb_fdblocks += cntp->icsb_fdblocks;
+
+		memset(cntp, 0, sizeof(xfs_icsb_cnts_t));
+
+		xfs_icsb_balance_counter(mp, XFS_SBS_ICOUNT, XFS_ICSB_SB_LOCKED);
+		xfs_icsb_balance_counter(mp, XFS_SBS_IFREE, XFS_ICSB_SB_LOCKED);
+		xfs_icsb_balance_counter(mp, XFS_SBS_FDBLOCKS, XFS_ICSB_SB_LOCKED);
+		XFS_SB_UNLOCK(mp, s);
+		break;
+	}
+
+	return NOTIFY_OK;
+}
+
+int
+xfs_icsb_init_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	mp->m_sb_cnts = alloc_percpu(xfs_icsb_cnts_t);
+	if (mp->m_sb_cnts == NULL)
+		return -ENOMEM;
+
+	mp->m_icsb_notifier.notifier_call = xfs_icsb_cpu_notify;
+	mp->m_icsb_notifier.priority = 0;
+	register_cpu_notifier(&mp->m_icsb_notifier);
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		memset(cntp, 0, sizeof(xfs_icsb_cnts_t));
+	}
+	/*
+	 * start with all counters disabled so that the
+	 * initial balance kicks us off correctly
+	 */
+	mp->m_icsb_counters = -1;
+	return 0;
+}
+
+STATIC void
+xfs_icsb_destroy_counters(
+	xfs_mount_t	*mp)
+{
+	if (mp->m_sb_cnts) {
+		unregister_cpu_notifier(&mp->m_icsb_notifier);
+		free_percpu(mp->m_sb_cnts);
+	}
+}
+
+STATIC inline void
+xfs_icsb_lock_cntr(
+	xfs_icsb_cnts_t	*icsbp)
+{
+	while (test_and_set_bit(XFS_ICSB_FLAG_LOCK, &icsbp->icsb_flags)) {
+		ndelay(1000);
+	}
+}
+
+STATIC inline void
+xfs_icsb_unlock_cntr(
+	xfs_icsb_cnts_t	*icsbp)
+{
+	clear_bit(XFS_ICSB_FLAG_LOCK, &icsbp->icsb_flags);
+}
+
+
+STATIC inline void
+xfs_icsb_lock_all_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		xfs_icsb_lock_cntr(cntp);
+	}
+}
+
+STATIC inline void
+xfs_icsb_unlock_all_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		xfs_icsb_unlock_cntr(cntp);
+	}
+}
+
+STATIC void
+xfs_icsb_count(
+	xfs_mount_t	*mp,
+	xfs_icsb_cnts_t	*cnt,
+	int		flags)
+{
+	xfs_icsb_cnts_t *cntp;
+	int		i;
+
+	memset(cnt, 0, sizeof(xfs_icsb_cnts_t));
+
+	if (!(flags & XFS_ICSB_LAZY_COUNT))
+		xfs_icsb_lock_all_counters(mp);
+
+	for_each_online_cpu(i) {
+		cntp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, i);
+		cnt->icsb_icount += cntp->icsb_icount;
+		cnt->icsb_ifree += cntp->icsb_ifree;
+		cnt->icsb_fdblocks += cntp->icsb_fdblocks;
+	}
+
+	if (!(flags & XFS_ICSB_LAZY_COUNT))
+		xfs_icsb_unlock_all_counters(mp);
+}
+
+STATIC int
+xfs_icsb_counter_disabled(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field)
+{
+	ASSERT((field >= XFS_SBS_ICOUNT) && (field <= XFS_SBS_FDBLOCKS));
+	return test_bit(field, &mp->m_icsb_counters);
+}
+
+STATIC int
+xfs_icsb_disable_counter(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field)
+{
+	xfs_icsb_cnts_t	cnt;
+
+	ASSERT((field >= XFS_SBS_ICOUNT) && (field <= XFS_SBS_FDBLOCKS));
+
+	xfs_icsb_lock_all_counters(mp);
+	if (!test_and_set_bit(field, &mp->m_icsb_counters)) {
+		/* drain back to superblock */
+
+		xfs_icsb_count(mp, &cnt, XFS_ICSB_SB_LOCKED|XFS_ICSB_LAZY_COUNT);
+		switch(field) {
+		case XFS_SBS_ICOUNT:
+			mp->m_sb.sb_icount = cnt.icsb_icount;
+			break;
+		case XFS_SBS_IFREE:
+			mp->m_sb.sb_ifree = cnt.icsb_ifree;
+			break;
+		case XFS_SBS_FDBLOCKS:
+			mp->m_sb.sb_fdblocks = cnt.icsb_fdblocks;
+			break;
+		default:
+			BUG();
+		}
+	}
+
+	xfs_icsb_unlock_all_counters(mp);
+
+	return 0;
+}
+
+STATIC void
+xfs_icsb_enable_counter(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	uint64_t	count,
+	uint64_t	resid)
+{
+	xfs_icsb_cnts_t	*cntp;
+	int		i;
+
+	ASSERT((field >= XFS_SBS_ICOUNT) && (field <= XFS_SBS_FDBLOCKS));
+
+	xfs_icsb_lock_all_counters(mp);
+	for_each_online_cpu(i) {
+		cntp = per_cpu_ptr(mp->m_sb_cnts, i);
+		switch (field) {
+		case XFS_SBS_ICOUNT:
+			cntp->icsb_icount = count + resid;
+			break;
+		case XFS_SBS_IFREE:
+			cntp->icsb_ifree = count + resid;
+			break;
+		case XFS_SBS_FDBLOCKS:
+			cntp->icsb_fdblocks = count + resid;
+			break;
+		default:
+			BUG();
+			break;
+		}
+		resid = 0;
+	}
+	clear_bit(field, &mp->m_icsb_counters);
+	xfs_icsb_unlock_all_counters(mp);
+}
+
+STATIC void
+xfs_icsb_sync_counters_int(
+	xfs_mount_t	*mp,
+	int		flags)
+{
+	xfs_icsb_cnts_t	cnt;
+	int		s;
+
+	/* Pass 1: lock all counters */
+	if ((flags & XFS_ICSB_SB_LOCKED) == 0)
+		s = XFS_SB_LOCK(mp);
+
+	xfs_icsb_count(mp, &cnt, flags);
+
+	/* Step 3: update mp->m_sb fields */
+	if (!xfs_icsb_counter_disabled(mp, XFS_SBS_ICOUNT))
+		mp->m_sb.sb_icount = cnt.icsb_icount;
+	if (!xfs_icsb_counter_disabled(mp, XFS_SBS_IFREE))
+		mp->m_sb.sb_ifree = cnt.icsb_ifree;
+	if (!xfs_icsb_counter_disabled(mp, XFS_SBS_FDBLOCKS))
+		mp->m_sb.sb_fdblocks = cnt.icsb_fdblocks;
+
+	if ((flags & XFS_ICSB_SB_LOCKED) == 0)
+		XFS_SB_UNLOCK(mp, s);
+}
+
+/*
+ * Accurate update of per-cpu counters to incore superblock
+ */
+STATIC void
+xfs_icsb_sync_counters(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_sync_counters_int(mp, 0);
+}
+
+/*
+ * lazy addition used for things like df, background sb syncs, etc
+ */
+void
+xfs_icsb_sync_counters_lazy(
+	xfs_mount_t	*mp)
+{
+	xfs_icsb_sync_counters_int(mp, XFS_ICSB_LAZY_COUNT);
+}
+
+/*
+ * Balance and enable/disable counters as necessary.
+ *
+ * Thresholds for re-enabling counters are somewhat magic.
+ * inode counts are chosen to be the same number as single
+ * on disk allocation chunk per CPU, and free blocks is
+ * something far enough zero that we aren't going thrash
+ * when we get near ENOSPC.
+ */
+#define XFS_ICSB_INO_CNTR_REENABLE	64
+#define XFS_ICSB_FDBLK_CNTR_REENABLE	512
+STATIC void
+xfs_icsb_balance_counter(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t  field,
+	int		flags)
+{
+	uint64_t	count, resid = 0;
+	int		weight = num_online_cpus();
+	int		s;
+
+	if (!(flags & XFS_ICSB_SB_LOCKED))
+		s = XFS_SB_LOCK(mp);
+
+	/* disable counter and sync counter */
+	xfs_icsb_disable_counter(mp, field);
+
+	/* update counters  - first CPU gets residual*/
+	switch (field) {
+	case XFS_SBS_ICOUNT:
+		count = mp->m_sb.sb_icount;
+		resid = do_div(count, weight);
+		if (count < XFS_ICSB_INO_CNTR_REENABLE)
+			goto out;
+		break;
+	case XFS_SBS_IFREE:
+		count = mp->m_sb.sb_ifree;
+		resid = do_div(count, weight);
+		if (count < XFS_ICSB_INO_CNTR_REENABLE)
+			goto out;
+		break;
+	case XFS_SBS_FDBLOCKS:
+		count = mp->m_sb.sb_fdblocks;
+		resid = do_div(count, weight);
+		if (count < XFS_ICSB_FDBLK_CNTR_REENABLE)
+			goto out;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	xfs_icsb_enable_counter(mp, field, count, resid);
+out:
+	if (!(flags & XFS_ICSB_SB_LOCKED))
+		XFS_SB_UNLOCK(mp, s);
+}
+
+STATIC int
+xfs_icsb_modify_counters_int(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	int		delta,
+	int		rsvd,
+	int		flags)
+{
+	xfs_icsb_cnts_t	*icsbp;
+	long long	lcounter;	/* long counter for 64 bit fields */
+	int		cpu, s, locked = 0;
+	int		ret = 0, balance_done = 0;
+
+again:
+	cpu = get_cpu();
+	icsbp = (xfs_icsb_cnts_t *)per_cpu_ptr(mp->m_sb_cnts, cpu),
+	xfs_icsb_lock_cntr(icsbp);
+	if (unlikely(xfs_icsb_counter_disabled(mp, field)))
+		goto slow_path;
+
+	switch (field) {
+	case XFS_SBS_ICOUNT:
+		lcounter = icsbp->icsb_icount;
+		lcounter += delta;
+		if (unlikely(lcounter < 0))
+			goto slow_path;
+		icsbp->icsb_icount = lcounter;
+		break;
+
+	case XFS_SBS_IFREE:
+		lcounter = icsbp->icsb_ifree;
+		lcounter += delta;
+		if (unlikely(lcounter < 0))
+			goto slow_path;
+		icsbp->icsb_ifree = lcounter;
+		break;
+
+	case XFS_SBS_FDBLOCKS:
+		BUG_ON((mp->m_resblks - mp->m_resblks_avail) != 0);
+
+		lcounter = icsbp->icsb_fdblocks;
+		lcounter += delta;
+		if (unlikely(lcounter < 0))
+			goto slow_path;
+		icsbp->icsb_fdblocks = lcounter;
+		break;
+	default:
+		BUG();
+		break;
+	}
+	xfs_icsb_unlock_cntr(icsbp);
+	put_cpu();
+	if (locked)
+		XFS_SB_UNLOCK(mp, s);
+	return 0;
+
+	/*
+	 * The slow path needs to be run with the SBLOCK
+	 * held so that we prevent other threads from
+	 * attempting to run this path at the same time.
+	 * this provides exclusion for the balancing code,
+	 * and exclusive fallback if the balance does not
+	 * provide enough resources to continue in an unlocked
+	 * manner.
+	 */
+slow_path:
+	xfs_icsb_unlock_cntr(icsbp);
+	put_cpu();
+
+	/* need to hold superblock incase we need
+	 * to disable a counter */
+	if (!(flags & XFS_ICSB_SB_LOCKED)) {
+		s = XFS_SB_LOCK(mp);
+		locked = 1;
+		flags |= XFS_ICSB_SB_LOCKED;
+	}
+	if (!balance_done) {
+		xfs_icsb_balance_counter(mp, field, flags);
+		balance_done = 1;
+		goto again;
+	} else {
+		/*
+		 * we might not have enough on this local
+		 * cpu to allocate for a bulk request.
+		 * We need to drain this field from all CPUs
+		 * and disable the counter fastpath
+		 */
+		xfs_icsb_disable_counter(mp, field);
+	}
+
+	ret = xfs_mod_incore_sb_unlocked(mp, field, delta, rsvd);
+
+	if (locked)
+		XFS_SB_UNLOCK(mp, s);
+	return ret;
+}
+
+STATIC int
+xfs_icsb_modify_counters(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	int		delta,
+	int		rsvd)
+{
+	return xfs_icsb_modify_counters_int(mp, field, delta, rsvd, 0);
+}
+
+/*
+ * Called when superblock is already locked
+ */
+STATIC int
+xfs_icsb_modify_counters_locked(
+	xfs_mount_t	*mp,
+	xfs_sb_field_t	field,
+	int		delta,
+	int		rsvd)
+{
+	return xfs_icsb_modify_counters_int(mp, field, delta,
+						rsvd, XFS_ICSB_SB_LOCKED);
+}
+#endif
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index cd3cf9613a00..66cbee79864e 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -267,6 +267,34 @@ typedef struct xfs_ioops {
 #define XFS_IODONE(vfsp) \
 	(*(mp)->m_io_ops.xfs_iodone)(vfsp)
 
+#ifdef HAVE_PERCPU_SB
+
+/*
+ * Valid per-cpu incore superblock counters. Note that if you add new counters,
+ * you may need to define new counter disabled bit field descriptors as there
+ * are more possible fields in the superblock that can fit in a bitfield on a
+ * 32 bit platform. The XFS_SBS_* values for the current current counters just
+ * fit.
+ */
+typedef struct xfs_icsb_cnts {
+	uint64_t	icsb_fdblocks;
+	uint64_t	icsb_ifree;
+	uint64_t	icsb_icount;
+	unsigned long	icsb_flags;
+} xfs_icsb_cnts_t;
+
+#define XFS_ICSB_FLAG_LOCK	(1 << 0)	/* counter lock bit */
+
+#define XFS_ICSB_SB_LOCKED	(1 << 0)	/* sb already locked */
+#define XFS_ICSB_LAZY_COUNT	(1 << 1)	/* accuracy not needed */
+
+extern int	xfs_icsb_init_counters(struct xfs_mount *);
+extern void	xfs_icsb_sync_counters_lazy(struct xfs_mount *);
+
+#else
+#define xfs_icsb_init_counters(mp)	(0)
+#define xfs_icsb_sync_counters_lazy(mp)	do { } while (0)
+#endif
 
 typedef struct xfs_mount {
 	bhv_desc_t		m_bhv;		/* vfs xfs behavior */
@@ -351,7 +379,7 @@ typedef struct xfs_mount {
 #endif
 	int			m_dalign;	/* stripe unit */
 	int			m_swidth;	/* stripe width */
-	int			m_sinoalign;	/* stripe unit inode alignmnt */
+	int			m_sinoalign;	/* stripe unit inode alignment */
 	int			m_attr_magicpct;/* 37% of the blocksize */
 	int			m_dir_magicpct;	/* 37% of the dir blocksize */
 	__uint8_t		m_mk_sharedro;	/* mark shared ro on unmount */
@@ -372,6 +400,11 @@ typedef struct xfs_mount {
 	struct xfs_qmops	m_qm_ops;	/* vector of XQM ops */
 	struct xfs_ioops	m_io_ops;	/* vector of I/O ops */
 	atomic_t		m_active_trans;	/* number trans frozen */
+#ifdef HAVE_PERCPU_SB
+	xfs_icsb_cnts_t		*m_sb_cnts;	/* per-cpu superblock counters */
+	unsigned long		m_icsb_counters; /* disabled per-cpu counters */
+	struct notifier_block	m_icsb_notifier; /* hotplug cpu notifier */
+#endif
 } xfs_mount_t;
 
 /*
@@ -386,8 +419,6 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_FS_SHUTDOWN	(1ULL << 4)	/* atomic stop of all filesystem
 						   operations, typically for
 						   disk errors in metadata */
-#define XFS_MOUNT_NOATIME	(1ULL << 5)	/* don't modify inode access
-						   times on reads */
 #define XFS_MOUNT_RETERR	(1ULL << 6)     /* return alignment errors to
 						   user */
 #define XFS_MOUNT_NOALIGN	(1ULL << 7)	/* turn off stripe alignment
@@ -411,6 +442,8 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_DIRSYNC	(1ULL << 21)	/* synchronous directory ops */
 #define XFS_MOUNT_COMPAT_IOSIZE	(1ULL << 22)	/* don't report large preferred
 						 * I/O size in stat() */
+#define XFS_MOUNT_NO_PERCPU_SB	(1ULL << 23)	/* don't use per-cpu superblock
+						   counters */
 
 
 /*
@@ -473,11 +506,6 @@ xfs_preferred_iosize(xfs_mount_t *mp)
 #define XFS_SHUTDOWN_REMOTE_REQ 0x10	/* Shutdown came from remote cell */
 
 /*
- * xflags for xfs_syncsub
- */
-#define XFS_XSYNC_RELOC		0x01
-
-/*
  * Flags for xfs_mountfs
  */
 #define XFS_MFSI_SECOND		0x01	/* Secondary mount -- skip stuff */
@@ -548,6 +576,8 @@ extern void	xfs_unmountfs_close(xfs_mount_t *, struct cred *);
 extern int	xfs_unmountfs_writesb(xfs_mount_t *);
 extern int	xfs_unmount_flush(xfs_mount_t *, int);
 extern int	xfs_mod_incore_sb(xfs_mount_t *, xfs_sb_field_t, int, int);
+extern int	xfs_mod_incore_sb_unlocked(xfs_mount_t *, xfs_sb_field_t,
+			int, int);
 extern int	xfs_mod_incore_sb_batch(xfs_mount_t *, xfs_mod_sb_t *,
 			uint, int);
 extern struct xfs_buf *xfs_getsb(xfs_mount_t *, int);
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index 82a08baf437b..4f6a034de7f7 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -31,7 +31,7 @@
 typedef __uint32_t	xfs_dqid_t;
 
 /*
- * Eventhough users may not have quota limits occupying all 64-bits,
+ * Even though users may not have quota limits occupying all 64-bits,
  * they may need 64-bit accounting. Hence, 64-bit quota-counters,
  * and quota-limits. This is a waste in the common case, but hey ...
  */
@@ -246,7 +246,7 @@ typedef struct xfs_qoff_logformat {
 #ifdef __KERNEL__
 /*
  * This check is done typically without holding the inode lock;
- * that may seem racey, but it is harmless in the context that it is used.
+ * that may seem racy, but it is harmless in the context that it is used.
  * The inode cannot go inactive as long a reference is kept, and
  * therefore if dquot(s) were attached, they'll stay consistent.
  * If, for example, the ownership of the inode changes while
diff --git a/fs/xfs/xfs_rw.h b/fs/xfs/xfs_rw.h
index de85eefb7966..e63795644478 100644
--- a/fs/xfs/xfs_rw.h
+++ b/fs/xfs/xfs_rw.h
@@ -89,6 +89,7 @@ extern void xfs_ioerror_alert(char *func, struct xfs_mount *mp,
  */
 extern int xfs_rwlock(bhv_desc_t *bdp, vrwlock_t write_lock);
 extern void xfs_rwunlock(bhv_desc_t *bdp, vrwlock_t write_lock);
+extern int xfs_setattr(bhv_desc_t *bdp, vattr_t *vap, int flags, cred_t *credp);
 extern int xfs_change_file_space(bhv_desc_t *bdp, int cmd, xfs_flock64_t *bf,
 				 xfs_off_t offset, cred_t *credp, int flags);
 extern int xfs_set_dmattrs(bhv_desc_t *bdp, u_int evmask, u_int16_t state,
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index d3d714e6b32a..8d056cef5d1f 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -55,10 +55,141 @@ STATIC void	xfs_trans_committed(xfs_trans_t *, int);
 STATIC void	xfs_trans_chunk_committed(xfs_log_item_chunk_t *, xfs_lsn_t, int);
 STATIC void	xfs_trans_free(xfs_trans_t *);
 
-kmem_zone_t		*xfs_trans_zone;
+kmem_zone_t	*xfs_trans_zone;
 
 
 /*
+ * Reservation functions here avoid a huge stack in xfs_trans_init
+ * due to register overflow from temporaries in the calculations.
+ */
+
+STATIC uint
+xfs_calc_write_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_WRITE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_itruncate_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ITRUNCATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_rename_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_RENAME_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_link_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_LINK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_remove_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_REMOVE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_symlink_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_SYMLINK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_create_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_CREATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_mkdir_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_MKDIR_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_ifree_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_IFREE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_ichange_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ICHANGE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_growdata_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWDATA_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_growrtalloc_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWRTALLOC_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_growrtzero_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWRTZERO_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_growrtfree_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_GROWRTFREE_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_swrite_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_SWRITE_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_writeid_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_WRITEID_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_addafork_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ADDAFORK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_attrinval_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ATTRINVAL_LOG_RES(mp);
+}
+
+STATIC uint
+xfs_calc_attrset_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ATTRSET_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_attrrm_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_ATTRRM_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp);
+}
+
+STATIC uint
+xfs_calc_clear_agi_bucket_reservation(xfs_mount_t *mp)
+{
+	return XFS_CALC_CLEAR_AGI_BUCKET_LOG_RES(mp);
+}
+
+/*
  * Initialize the precomputed transaction reservation values
  * in the mount structure.
  */
@@ -69,39 +200,27 @@ xfs_trans_init(
 	xfs_trans_reservations_t	*resp;
 
 	resp = &(mp->m_reservations);
-	resp->tr_write =
-		(uint)(XFS_CALC_WRITE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_itruncate =
-		(uint)(XFS_CALC_ITRUNCATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_rename =
-		(uint)(XFS_CALC_RENAME_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_link = (uint)XFS_CALC_LINK_LOG_RES(mp);
-	resp->tr_remove =
-		(uint)(XFS_CALC_REMOVE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_symlink =
-		(uint)(XFS_CALC_SYMLINK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_create =
-		(uint)(XFS_CALC_CREATE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_mkdir =
-		(uint)(XFS_CALC_MKDIR_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_ifree =
-		(uint)(XFS_CALC_IFREE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_ichange =
-		(uint)(XFS_CALC_ICHANGE_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_growdata = (uint)XFS_CALC_GROWDATA_LOG_RES(mp);
-	resp->tr_swrite = (uint)XFS_CALC_SWRITE_LOG_RES(mp);
-	resp->tr_writeid = (uint)XFS_CALC_WRITEID_LOG_RES(mp);
-	resp->tr_addafork =
-		(uint)(XFS_CALC_ADDAFORK_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_attrinval = (uint)XFS_CALC_ATTRINVAL_LOG_RES(mp);
-	resp->tr_attrset =
-		(uint)(XFS_CALC_ATTRSET_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_attrrm =
-		(uint)(XFS_CALC_ATTRRM_LOG_RES(mp) + XFS_DQUOT_LOGRES(mp));
-	resp->tr_clearagi = (uint)XFS_CALC_CLEAR_AGI_BUCKET_LOG_RES(mp);
-	resp->tr_growrtalloc = (uint)XFS_CALC_GROWRTALLOC_LOG_RES(mp);
-	resp->tr_growrtzero = (uint)XFS_CALC_GROWRTZERO_LOG_RES(mp);
-	resp->tr_growrtfree = (uint)XFS_CALC_GROWRTFREE_LOG_RES(mp);
+	resp->tr_write = xfs_calc_write_reservation(mp);
+	resp->tr_itruncate = xfs_calc_itruncate_reservation(mp);
+	resp->tr_rename = xfs_calc_rename_reservation(mp);
+	resp->tr_link = xfs_calc_link_reservation(mp);
+	resp->tr_remove = xfs_calc_remove_reservation(mp);
+	resp->tr_symlink = xfs_calc_symlink_reservation(mp);
+	resp->tr_create = xfs_calc_create_reservation(mp);
+	resp->tr_mkdir = xfs_calc_mkdir_reservation(mp);
+	resp->tr_ifree = xfs_calc_ifree_reservation(mp);
+	resp->tr_ichange = xfs_calc_ichange_reservation(mp);
+	resp->tr_growdata = xfs_calc_growdata_reservation(mp);
+	resp->tr_swrite = xfs_calc_swrite_reservation(mp);
+	resp->tr_writeid = xfs_calc_writeid_reservation(mp);
+	resp->tr_addafork = xfs_calc_addafork_reservation(mp);
+	resp->tr_attrinval = xfs_calc_attrinval_reservation(mp);
+	resp->tr_attrset = xfs_calc_attrset_reservation(mp);
+	resp->tr_attrrm = xfs_calc_attrrm_reservation(mp);
+	resp->tr_clearagi = xfs_calc_clear_agi_bucket_reservation(mp);
+	resp->tr_growrtalloc = xfs_calc_growrtalloc_reservation(mp);
+	resp->tr_growrtzero = xfs_calc_growrtzero_reservation(mp);
+	resp->tr_growrtfree = xfs_calc_growrtfree_reservation(mp);
 }
 
 /*
@@ -371,7 +490,7 @@ xfs_trans_mod_sb(
 	case XFS_TRANS_SB_RES_FREXTENTS:
 		/*
 		 * The allocation has already been applied to the
-		 * in-core superblocks's counter.  This should only
+		 * in-core superblock's counter.  This should only
 		 * be applied to the on-disk superblock.
 		 */
 		ASSERT(delta < 0);
@@ -492,7 +611,7 @@ xfs_trans_apply_sb_deltas(
 
 	if (whole)
 		/*
-		 * Log the whole thing, the fields are discontiguous.
+		 * Log the whole thing, the fields are noncontiguous.
 		 */
 		xfs_trans_log_buf(tp, bp, 0, sizeof(xfs_sb_t) - 1);
 	else
@@ -550,7 +669,7 @@ xfs_trans_unreserve_and_mod_sb(
 	/*
 	 * Apply any superblock modifications to the in-core version.
 	 * The t_res_fdblocks_delta and t_res_frextents_delta fields are
-	 * explicity NOT applied to the in-core superblock.
+	 * explicitly NOT applied to the in-core superblock.
 	 * The idea is that that has already been done.
 	 */
 	if (tp->t_flags & XFS_TRANS_SB_DIRTY) {
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index d77901c07f63..100d9a4b38ee 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -354,7 +354,7 @@ typedef struct xfs_trans {
 	xfs_lsn_t		t_commit_lsn;	/* log seq num of end of
 						 * transaction. */
 	struct xfs_mount	*t_mountp;	/* ptr to fs mount struct */
-	struct xfs_dquot_acct   *t_dqinfo;	/* accting info for dquots */
+	struct xfs_dquot_acct   *t_dqinfo;	/* acctg info for dquots */
 	xfs_trans_callback_t	t_callback;	/* transaction callback */
 	void			*t_callarg;	/* callback arg */
 	unsigned int		t_flags;	/* misc flags */
@@ -380,7 +380,7 @@ typedef struct xfs_trans {
 	xfs_trans_header_t	t_header;	/* header for in-log trans */
 	unsigned int		t_busy_free;	/* busy descs free */
 	xfs_log_busy_chunk_t	t_busy;		/* busy/async free blocks */
-        xfs_pflags_t            t_pflags;       /* saved pflags state */
+	unsigned long		t_pflags;	/* saved process flags state */
 } xfs_trans_t;
 
 #endif	/* __KERNEL__ */
diff --git a/fs/xfs/xfs_trans_inode.c b/fs/xfs/xfs_trans_inode.c
index e341409172d2..7c5894d59f81 100644
--- a/fs/xfs/xfs_trans_inode.c
+++ b/fs/xfs/xfs_trans_inode.c
@@ -272,7 +272,7 @@ xfs_trans_log_inode(
 	 * This is to coordinate with the xfs_iflush() and xfs_iflush_done()
 	 * routines in the eventual clearing of the ilf_fields bits.
 	 * See the big comment in xfs_iflush() for an explanation of
-	 * this coorination mechanism.
+	 * this coordination mechanism.
 	 */
 	flags |= ip->i_itemp->ili_last_fields;
 	ip->i_itemp->ili_format.ilf_fields |= flags;
diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c
index b6ad370fab3d..504d2a80747a 100644
--- a/fs/xfs/xfs_vfsops.c
+++ b/fs/xfs/xfs_vfsops.c
@@ -55,7 +55,7 @@
 #include "xfs_clnt.h"
 #include "xfs_fsops.h"
 
-STATIC int xfs_sync(bhv_desc_t *, int, cred_t *);
+STATIC int	xfs_sync(bhv_desc_t *, int, cred_t *);
 
 int
 xfs_init(void)
@@ -77,11 +77,12 @@ xfs_init(void)
 						 "xfs_bmap_free_item");
 	xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
 					    "xfs_btree_cur");
-	xfs_inode_zone = kmem_zone_init(sizeof(xfs_inode_t), "xfs_inode");
 	xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans");
 	xfs_da_state_zone =
 		kmem_zone_init(sizeof(xfs_da_state_t), "xfs_da_state");
 	xfs_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf");
+	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
+	xfs_acl_zone_init(xfs_acl_zone, "xfs_acl");
 
 	/*
 	 * The size of the zone allocated buf log item is the maximum
@@ -93,17 +94,30 @@ xfs_init(void)
 				(((XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK) /
 				  NBWORD) * sizeof(int))),
 			       "xfs_buf_item");
-	xfs_efd_zone = kmem_zone_init((sizeof(xfs_efd_log_item_t) +
-				       ((XFS_EFD_MAX_FAST_EXTENTS - 1) * sizeof(xfs_extent_t))),
+	xfs_efd_zone =
+		kmem_zone_init((sizeof(xfs_efd_log_item_t) +
+			       ((XFS_EFD_MAX_FAST_EXTENTS - 1) *
+				 sizeof(xfs_extent_t))),
 				      "xfs_efd_item");
-	xfs_efi_zone = kmem_zone_init((sizeof(xfs_efi_log_item_t) +
-				       ((XFS_EFI_MAX_FAST_EXTENTS - 1) * sizeof(xfs_extent_t))),
+	xfs_efi_zone =
+		kmem_zone_init((sizeof(xfs_efi_log_item_t) +
+			       ((XFS_EFI_MAX_FAST_EXTENTS - 1) *
+				 sizeof(xfs_extent_t))),
 				      "xfs_efi_item");
-	xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
-	xfs_ili_zone = kmem_zone_init(sizeof(xfs_inode_log_item_t), "xfs_ili");
-	xfs_chashlist_zone = kmem_zone_init(sizeof(xfs_chashlist_t),
-					    "xfs_chashlist");
-	xfs_acl_zone_init(xfs_acl_zone, "xfs_acl");
+
+	/*
+	 * These zones warrant special memory allocator hints
+	 */
+	xfs_inode_zone =
+		kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
+					KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
+					KM_ZONE_SPREAD, NULL);
+	xfs_ili_zone =
+		kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
+					KM_ZONE_SPREAD, NULL);
+	xfs_chashlist_zone =
+		kmem_zone_init_flags(sizeof(xfs_chashlist_t), "xfs_chashlist",
+					KM_ZONE_SPREAD, NULL);
 
 	/*
 	 * Allocate global trace buffers.
@@ -176,18 +190,18 @@ xfs_cleanup(void)
 	ktrace_free(xfs_alloc_trace_buf);
 #endif
 
-	kmem_cache_destroy(xfs_bmap_free_item_zone);
-	kmem_cache_destroy(xfs_btree_cur_zone);
-	kmem_cache_destroy(xfs_inode_zone);
-	kmem_cache_destroy(xfs_trans_zone);
-	kmem_cache_destroy(xfs_da_state_zone);
-	kmem_cache_destroy(xfs_dabuf_zone);
-	kmem_cache_destroy(xfs_buf_item_zone);
-	kmem_cache_destroy(xfs_efd_zone);
-	kmem_cache_destroy(xfs_efi_zone);
-	kmem_cache_destroy(xfs_ifork_zone);
-	kmem_cache_destroy(xfs_ili_zone);
-	kmem_cache_destroy(xfs_chashlist_zone);
+	kmem_zone_destroy(xfs_bmap_free_item_zone);
+	kmem_zone_destroy(xfs_btree_cur_zone);
+	kmem_zone_destroy(xfs_inode_zone);
+	kmem_zone_destroy(xfs_trans_zone);
+	kmem_zone_destroy(xfs_da_state_zone);
+	kmem_zone_destroy(xfs_dabuf_zone);
+	kmem_zone_destroy(xfs_buf_item_zone);
+	kmem_zone_destroy(xfs_efd_zone);
+	kmem_zone_destroy(xfs_efi_zone);
+	kmem_zone_destroy(xfs_ifork_zone);
+	kmem_zone_destroy(xfs_ili_zone);
+	kmem_zone_destroy(xfs_chashlist_zone);
 }
 
 /*
@@ -258,8 +272,6 @@ xfs_start_flags(
 		mp->m_inoadd = XFS_INO64_OFFSET;
 	}
 #endif
-	if (ap->flags & XFSMNT_NOATIME)
-		mp->m_flags |= XFS_MOUNT_NOATIME;
 	if (ap->flags & XFSMNT_RETERR)
 		mp->m_flags |= XFS_MOUNT_RETERR;
 	if (ap->flags & XFSMNT_NOALIGN)
@@ -620,7 +632,7 @@ xfs_quiesce_fs(
 	xfs_mount_t		*mp)
 {
 	int			count = 0, pincount;
-		
+
 	xfs_refcache_purge_mp(mp);
 	xfs_flush_buftarg(mp->m_ddev_targp, 0);
 	xfs_finish_reclaim_all(mp, 0);
@@ -631,7 +643,7 @@ xfs_quiesce_fs(
 	 * meta data (typically directory updates).
 	 * Which then must be flushed and logged before
 	 * we can write the unmount record.
-	 */ 
+	 */
 	do {
 		xfs_syncsub(mp, SYNC_REMOUNT|SYNC_ATTR|SYNC_WAIT, 0, NULL);
 		pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1);
@@ -654,11 +666,6 @@ xfs_mntupdate(
 	xfs_mount_t	*mp = XFS_BHVTOM(bdp);
 	int		error;
 
-	if (args->flags & XFSMNT_NOATIME)
-		mp->m_flags |= XFS_MOUNT_NOATIME;
-	else
-		mp->m_flags &= ~XFS_MOUNT_NOATIME;
-
 	if (args->flags & XFSMNT_BARRIER)
 		mp->m_flags |= XFS_MOUNT_BARRIER;
 	else
@@ -814,6 +821,7 @@ xfs_statvfs(
 
 	statp->f_type = XFS_SB_MAGIC;
 
+	xfs_icsb_sync_counters_lazy(mp);
 	s = XFS_SB_LOCK(mp);
 	statp->f_bsize = sbp->sb_blocksize;
 	lsize = sbp->sb_logstart ? sbp->sb_logblocks : 0;
@@ -872,10 +880,10 @@ xfs_statvfs(
  *		       determine if they should be flushed sync, async, or
  *		       delwri.
  *      SYNC_CLOSE   - This flag is passed when the system is being
- *		       unmounted.  We should sync and invalidate everthing.
+ *		       unmounted.  We should sync and invalidate everything.
  *      SYNC_FSDATA  - This indicates that the caller would like to make
  *		       sure the superblock is safe on disk.  We can ensure
- *		       this by simply makeing sure the log gets flushed
+ *		       this by simply making sure the log gets flushed
  *		       if SYNC_BDFLUSH is set, and by actually writing it
  *		       out otherwise.
  *
@@ -900,7 +908,7 @@ xfs_sync(
  *
  * This routine supports all of the flags defined for the generic VFS_SYNC
  * interface as explained above under xfs_sync.  In the interests of not
- * changing interfaces within the 6.5 family, additional internallly-
+ * changing interfaces within the 6.5 family, additional internally-
  * required functions are specified within a separate xflags parameter,
  * only available by calling this routine.
  *
@@ -1082,7 +1090,7 @@ xfs_sync_inodes(
 		 * If this is just vfs_sync() or pflushd() calling
 		 * then we can skip inodes for which it looks like
 		 * there is nothing to do.  Since we don't have the
-		 * inode locked this is racey, but these are periodic
+		 * inode locked this is racy, but these are periodic
 		 * calls so it doesn't matter.  For the others we want
 		 * to know for sure, so we at least try to lock them.
 		 */
@@ -1221,7 +1229,7 @@ xfs_sync_inodes(
 					xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
 					error = xfs_itobp(mp, NULL, ip,
-							  &dip, &bp, 0);
+							  &dip, &bp, 0, 0);
 					if (!error) {
 						xfs_buf_relse(bp);
 					} else {
@@ -1421,7 +1429,7 @@ xfs_sync_inodes(
  *
  * This routine supports all of the flags defined for the generic VFS_SYNC
  * interface as explained above under xfs_sync.  In the interests of not
- * changing interfaces within the 6.5 family, additional internallly-
+ * changing interfaces within the 6.5 family, additional internally-
  * required functions are specified within a separate xflags parameter,
  * only available by calling this routine.
  *
@@ -1690,10 +1698,7 @@ xfs_parseargs(
 	int			iosize;
 
 	args->flags2 |= XFSMNT2_COMPAT_IOSIZE;
-
-#if 0	/* XXX: off by default, until some remaining issues ironed out */
-	args->flags |= XFSMNT_IDELETE; /* default to on */
-#endif
+	args->flags |= XFSMNT_IDELETE;
 
 	if (!options)
 		goto done;
@@ -1903,7 +1908,6 @@ xfs_showargs(
 		{ XFS_MOUNT_NOUUID,		"," MNTOPT_NOUUID },
 		{ XFS_MOUNT_NORECOVERY,		"," MNTOPT_NORECOVERY },
 		{ XFS_MOUNT_OSYNCISOSYNC,	"," MNTOPT_OSYNCISOSYNC },
-		{ XFS_MOUNT_IDELETE,		"," MNTOPT_NOIKEEP },
 		{ 0, NULL }
 	};
 	struct proc_xfs_info	*xfs_infop;
@@ -1939,6 +1943,8 @@ xfs_showargs(
 		seq_printf(m, "," MNTOPT_SWIDTH "=%d",
 				(int)XFS_FSB_TO_BB(mp, mp->m_swidth));
 
+	if (!(mp->m_flags & XFS_MOUNT_IDELETE))
+		seq_printf(m, "," MNTOPT_IKEEP);
 	if (!(mp->m_flags & XFS_MOUNT_COMPAT_IOSIZE))
 		seq_printf(m, "," MNTOPT_LARGEIO);
 	if (mp->m_flags & XFS_MOUNT_BARRIER)
diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c
index eaab355f5a89..de49601919c1 100644
--- a/fs/xfs/xfs_vnodeops.c
+++ b/fs/xfs/xfs_vnodeops.c
@@ -615,6 +615,7 @@ xfs_setattr(
 			code = xfs_igrow_start(ip, vap->va_size, credp);
 		}
 		xfs_iunlock(ip, XFS_ILOCK_EXCL);
+		vn_iowait(vp); /* wait for the completion of any pending DIOs */
 		if (!code)
 			code = xfs_itruncate_data(ip, vap->va_size);
 		if (code) {
@@ -847,7 +848,7 @@ xfs_setattr(
 	 * If this is a synchronous mount, make sure that the
 	 * transaction goes to disk before returning to the user.
 	 * This is slightly sub-optimal in that truncates require
-	 * two sync transactions instead of one for wsync filesytems.
+	 * two sync transactions instead of one for wsync filesystems.
 	 * One for the truncate and one for the timestamps since we
 	 * don't want to change the timestamps unless we're sure the
 	 * truncate worked.  Truncates are less than 1% of the laddis
@@ -1169,7 +1170,7 @@ xfs_fsync(
 
 		/*
 		 * If this inode is on the RT dev we need to flush that
-		 * cache aswell.
+		 * cache as well.
 		 */
 		if (ip->i_d.di_flags & XFS_DIFLAG_REALTIME)
 			xfs_blkdev_issue_flush(ip->i_mount->m_rtdev_targp);
@@ -1334,7 +1335,7 @@ xfs_inactive_symlink_rmt(
 	 */
 	done = 0;
 	XFS_BMAP_INIT(&free_list, &first_block);
-	nmaps = sizeof(mval) / sizeof(mval[0]);
+	nmaps = ARRAY_SIZE(mval);
 	if ((error = xfs_bmapi(tp, ip, 0, XFS_B_TO_FSB(mp, size),
 			XFS_BMAPI_METADATA, &first_block, 0, mval, &nmaps,
 			&free_list)))
@@ -1379,7 +1380,7 @@ xfs_inactive_symlink_rmt(
 	 */
 	ntp = xfs_trans_dup(tp);
 	/*
-	 * Commit the transaction containing extent freeing and EFD's.
+	 * Commit the transaction containing extent freeing and EFDs.
 	 * If we get an error on the commit here or on the reserve below,
 	 * we need to unlock the inode since the new transaction doesn't
 	 * have the inode attached.
@@ -1556,7 +1557,7 @@ xfs_release(
 			if ((error = xfs_inactive_free_eofblocks(mp, ip)))
 				return error;
 			/* Update linux inode block count after free above */
-			LINVFS_GET_IP(vp)->i_blocks = XFS_FSB_TO_BB(mp,
+			vn_to_inode(vp)->i_blocks = XFS_FSB_TO_BB(mp,
 				ip->i_d.di_nblocks + ip->i_delayed_blks);
 		}
 	}
@@ -1637,7 +1638,7 @@ xfs_inactive(
 			if ((error = xfs_inactive_free_eofblocks(mp, ip)))
 				return VN_INACTIVE_CACHE;
 			/* Update linux inode block count after free above */
-			LINVFS_GET_IP(vp)->i_blocks = XFS_FSB_TO_BB(mp,
+			vn_to_inode(vp)->i_blocks = XFS_FSB_TO_BB(mp,
 				ip->i_d.di_nblocks + ip->i_delayed_blks);
 		}
 		goto out;
@@ -2022,7 +2023,7 @@ xfs_create(
 	XFS_QM_DQRELE(mp, gdqp);
 
 	/*
-	 * Propogate the fact that the vnode changed after the
+	 * Propagate the fact that the vnode changed after the
 	 * xfs_inode locks have been released.
 	 */
 	VOP_VNODE_CHANGE(vp, VCHANGE_FLAGS_TRUNCATED, 3);
@@ -2369,7 +2370,7 @@ xfs_remove(
 	 * for a log reservation. Since we'll have to wait for the
 	 * inactive code to complete before returning from xfs_iget,
 	 * we need to make sure that we don't have log space reserved
-	 * when we call xfs_iget.  Instead we get an unlocked referece
+	 * when we call xfs_iget.  Instead we get an unlocked reference
 	 * to the inode before getting our log reservation.
 	 */
 	error = xfs_get_dir_entry(dentry, &ip);
@@ -3019,7 +3020,7 @@ xfs_rmdir(
 	 * for a log reservation.  Since we'll have to wait for the
 	 * inactive code to complete before returning from xfs_iget,
 	 * we need to make sure that we don't have log space reserved
-	 * when we call xfs_iget.  Instead we get an unlocked referece
+	 * when we call xfs_iget.  Instead we get an unlocked reference
 	 * to the inode before getting our log reservation.
 	 */
 	error = xfs_get_dir_entry(dentry, &cdp);
@@ -3186,7 +3187,7 @@ xfs_rmdir(
 
 	/* Fall through to std_return with error = 0 or the errno
 	 * from xfs_trans_commit. */
-std_return:
+ std_return:
 	if (DM_EVENT_ENABLED(dir_vp->v_vfsp, dp, DM_EVENT_POSTREMOVE)) {
 		(void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE,
 					dir_vp, DM_RIGHT_NULL,
@@ -3196,12 +3197,12 @@ std_return:
 	}
 	return error;
 
-error1:
+ error1:
 	xfs_bmap_cancel(&free_list);
 	cancel_flags |= XFS_TRANS_ABORT;
 	/* FALLTHROUGH */
 
-error_return:
+ error_return:
 	xfs_trans_cancel(tp, cancel_flags);
 	goto std_return;
 }
@@ -4310,8 +4311,10 @@ xfs_free_file_space(
 	ASSERT(attr_flags & ATTR_NOLOCK ? attr_flags & ATTR_DMI : 1);
 	if (attr_flags & ATTR_NOLOCK)
 		need_iolock = 0;
-	if (need_iolock)
+	if (need_iolock) {
 		xfs_ilock(ip, XFS_IOLOCK_EXCL);
+		vn_iowait(vp);	/* wait for the completion of any pending DIOs */
+	}
 
 	rounding = MAX((__uint8_t)(1 << mp->m_sb.sb_blocklog),
 			(__uint8_t)NBPP);