Merge branch 'for-2.6.37' of git://linux-nfs.org/~bfields/linux

* 'for-2.6.37' of git://linux-nfs.org/~bfields/linux: (99 commits)
  svcrpc: svc_tcp_sendto XPT_DEAD check is redundant
  svcrpc: no need for XPT_DEAD check in svc_xprt_enqueue
  svcrpc: assume svc_delete_xprt() called only once
  svcrpc: never clear XPT_BUSY on dead xprt
  nfsd4: fix connection allocation in sequence()
  nfsd4: only require krb5 principal for NFSv4.0 callbacks
  nfsd4: move minorversion to client
  nfsd4: delay session removal till free_client
  nfsd4: separate callback change and callback probe
  nfsd4: callback program number is per-session
  nfsd4: track backchannel connections
  nfsd4: confirm only on succesful create_session
  nfsd4: make backchannel sequence number per-session
  nfsd4: use client pointer to backchannel session
  nfsd4: move callback setup into session init code
  nfsd4: don't cache seq_misordered replies
  SUNRPC: Properly initialize sock_xprt.srcaddr in all cases
  SUNRPC: Use conventional switch statement when reclassifying sockets
  sunrpc/xprtrdma: clean up workqueue usage
  sunrpc: Turn list_for_each-s into the ..._entry-s
  ...

Fix up trivial conflicts (two different deprecation notices added in
separate branches) in Documentation/feature-removal-schedule.txt

23 files changed, 635 insertions, 461 deletions

diff --git a/fs/Makefile b/fs/Makefile
index e6ec1d309b1d..26956fcec917 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -29,10 +29,7 @@ obj-$(CONFIG_EVENTFD)		+= eventfd.o
 obj-$(CONFIG_AIO)               += aio.o
 obj-$(CONFIG_FILE_LOCKING)      += locks.o
 obj-$(CONFIG_COMPAT)		+= compat.o compat_ioctl.o
-
-nfsd-$(CONFIG_NFSD)		:= nfsctl.o
-obj-y				+= $(nfsd-y) $(nfsd-m)
-
+obj-$(CONFIG_NFSD_DEPRECATED)	+= nfsctl.o
 obj-$(CONFIG_BINFMT_AOUT)	+= binfmt_aout.o
 obj-$(CONFIG_BINFMT_EM86)	+= binfmt_em86.o
 obj-$(CONFIG_BINFMT_MISC)	+= binfmt_misc.o
diff --git a/fs/compat.c b/fs/compat.c
index 0644a154672b..f03abdadc401 100644
--- a/fs/compat.c
+++ b/fs/compat.c
@@ -1963,7 +1963,7 @@ asmlinkage long compat_sys_ppoll(struct pollfd __user *ufds,
 }
 #endif /* HAVE_SET_RESTORE_SIGMASK */
 
-#if defined(CONFIG_NFSD) || defined(CONFIG_NFSD_MODULE)
+#if (defined(CONFIG_NFSD) || defined(CONFIG_NFSD_MODULE)) && !defined(CONFIG_NFSD_DEPRECATED)
 /* Stuff for NFS server syscalls... */
 struct compat_nfsctl_svc {
 	u16			svc32_port;
diff --git a/fs/lockd/host.c b/fs/lockd/host.c
index bb464d12104c..25e21e4023b2 100644
--- a/fs/lockd/host.c
+++ b/fs/lockd/host.c
@@ -353,6 +353,7 @@ nlm_bind_host(struct nlm_host *host)
 			.to_retries	= 5U,
 		};
 		struct rpc_create_args args = {
+			.net		= &init_net,
 			.protocol	= host->h_proto,
 			.address	= nlm_addr(host),
 			.addrsize	= host->h_addrlen,
diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c
index e3015464fbab..e0c918949644 100644
--- a/fs/lockd/mon.c
+++ b/fs/lockd/mon.c
@@ -69,6 +69,7 @@ static struct rpc_clnt *nsm_create(void)
 		.sin_addr.s_addr	= htonl(INADDR_LOOPBACK),
 	};
 	struct rpc_create_args args = {
+		.net			= &init_net,
 		.protocol		= XPRT_TRANSPORT_UDP,
 		.address		= (struct sockaddr *)&sin,
 		.addrsize		= sizeof(sin),
diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
index f1bacf1a0391..b13aabc12298 100644
--- a/fs/lockd/svc.c
+++ b/fs/lockd/svc.c
@@ -206,7 +206,7 @@ static int create_lockd_listener(struct svc_serv *serv, const char *name,
 
 	xprt = svc_find_xprt(serv, name, family, 0);
 	if (xprt == NULL)
-		return svc_create_xprt(serv, name, family, port,
+		return svc_create_xprt(serv, name, &init_net, family, port,
 						SVC_SOCK_DEFAULTS);
 	svc_xprt_put(xprt);
 	return 0;
diff --git a/fs/lockd/svc4proc.c b/fs/lockd/svc4proc.c
index 031c6569a134..a336e832475d 100644
--- a/fs/lockd/svc4proc.c
+++ b/fs/lockd/svc4proc.c
@@ -230,9 +230,7 @@ static void nlm4svc_callback_exit(struct rpc_task *task, void *data)
 
 static void nlm4svc_callback_release(void *data)
 {
-	lock_kernel();
 	nlm_release_call(data);
-	unlock_kernel();
 }
 
 static const struct rpc_call_ops nlm4svc_callback_ops = {
diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c
index 84055d31bfc5..6f1ef000975a 100644
--- a/fs/lockd/svclock.c
+++ b/fs/lockd/svclock.c
@@ -52,12 +52,13 @@ static const struct rpc_call_ops nlmsvc_grant_ops;
  * The list of blocked locks to retry
  */
 static LIST_HEAD(nlm_blocked);
+static DEFINE_SPINLOCK(nlm_blocked_lock);
 
 /*
  * Insert a blocked lock into the global list
  */
 static void
-nlmsvc_insert_block(struct nlm_block *block, unsigned long when)
+nlmsvc_insert_block_locked(struct nlm_block *block, unsigned long when)
 {
 	struct nlm_block *b;
 	struct list_head *pos;
@@ -87,6 +88,13 @@ nlmsvc_insert_block(struct nlm_block *block, unsigned long when)
 	block->b_when = when;
 }
 
+static void nlmsvc_insert_block(struct nlm_block *block, unsigned long when)
+{
+	spin_lock(&nlm_blocked_lock);
+	nlmsvc_insert_block_locked(block, when);
+	spin_unlock(&nlm_blocked_lock);
+}
+
 /*
  * Remove a block from the global list
  */
@@ -94,7 +102,9 @@ static inline void
 nlmsvc_remove_block(struct nlm_block *block)
 {
 	if (!list_empty(&block->b_list)) {
+		spin_lock(&nlm_blocked_lock);
 		list_del_init(&block->b_list);
+		spin_unlock(&nlm_blocked_lock);
 		nlmsvc_release_block(block);
 	}
 }
@@ -651,7 +661,7 @@ static int nlmsvc_grant_deferred(struct file_lock *fl, struct file_lock *conf,
 	struct nlm_block *block;
 	int rc = -ENOENT;
 
-	lock_kernel();
+	spin_lock(&nlm_blocked_lock);
 	list_for_each_entry(block, &nlm_blocked, b_list) {
 		if (nlm_compare_locks(&block->b_call->a_args.lock.fl, fl)) {
 			dprintk("lockd: nlmsvc_notify_blocked block %p flags %d\n",
@@ -665,13 +675,13 @@ static int nlmsvc_grant_deferred(struct file_lock *fl, struct file_lock *conf,
 			} else if (result == 0)
 				block->b_granted = 1;
 
-			nlmsvc_insert_block(block, 0);
+			nlmsvc_insert_block_locked(block, 0);
 			svc_wake_up(block->b_daemon);
 			rc = 0;
 			break;
 		}
 	}
-	unlock_kernel();
+	spin_unlock(&nlm_blocked_lock);
 	if (rc == -ENOENT)
 		printk(KERN_WARNING "lockd: grant for unknown block\n");
 	return rc;
@@ -803,7 +813,7 @@ static void nlmsvc_grant_callback(struct rpc_task *task, void *data)
 
 	dprintk("lockd: GRANT_MSG RPC callback\n");
 
-	lock_kernel();
+	spin_lock(&nlm_blocked_lock);
 	/* if the block is not on a list at this point then it has
 	 * been invalidated. Don't try to requeue it.
 	 *
@@ -825,19 +835,20 @@ static void nlmsvc_grant_callback(struct rpc_task *task, void *data)
 		/* Call was successful, now wait for client callback */
 		timeout = 60 * HZ;
 	}
-	nlmsvc_insert_block(block, timeout);
+	nlmsvc_insert_block_locked(block, timeout);
 	svc_wake_up(block->b_daemon);
 out:
-	unlock_kernel();
+	spin_unlock(&nlm_blocked_lock);
 }
 
+/*
+ * FIXME: nlmsvc_release_block() grabs a mutex.  This is not allowed for an
+ * .rpc_release rpc_call_op
+ */
 static void nlmsvc_grant_release(void *data)
 {
 	struct nlm_rqst		*call = data;
-
-	lock_kernel();
 	nlmsvc_release_block(call->a_block);
-	unlock_kernel();
 }
 
 static const struct rpc_call_ops nlmsvc_grant_ops = {
diff --git a/fs/lockd/svcproc.c b/fs/lockd/svcproc.c
index 0f2ab741ae7c..c3069f38d602 100644
--- a/fs/lockd/svcproc.c
+++ b/fs/lockd/svcproc.c
@@ -260,9 +260,7 @@ static void nlmsvc_callback_exit(struct rpc_task *task, void *data)
 
 static void nlmsvc_callback_release(void *data)
 {
-	lock_kernel();
 	nlm_release_call(data);
-	unlock_kernel();
 }
 
 static const struct rpc_call_ops nlmsvc_callback_ops = {
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index e17b49e2eabd..aeec017fe814 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -109,7 +109,7 @@ nfs4_callback_up(struct svc_serv *serv)
 {
 	int ret;
 
-	ret = svc_create_xprt(serv, "tcp", PF_INET,
+	ret = svc_create_xprt(serv, "tcp", &init_net, PF_INET,
 				nfs_callback_set_tcpport, SVC_SOCK_ANONYMOUS);
 	if (ret <= 0)
 		goto out_err;
@@ -117,7 +117,7 @@ nfs4_callback_up(struct svc_serv *serv)
 	dprintk("NFS: Callback listener port = %u (af %u)\n",
 			nfs_callback_tcpport, PF_INET);
 
-	ret = svc_create_xprt(serv, "tcp", PF_INET6,
+	ret = svc_create_xprt(serv, "tcp", &init_net, PF_INET6,
 				nfs_callback_set_tcpport, SVC_SOCK_ANONYMOUS);
 	if (ret > 0) {
 		nfs_callback_tcpport6 = ret;
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index fd6f0a70021b..0870d0d4efc0 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -605,6 +605,7 @@ static int nfs_create_rpc_client(struct nfs_client *clp,
 {
 	struct rpc_clnt		*clnt = NULL;
 	struct rpc_create_args args = {
+		.net		= &init_net,
 		.protocol	= clp->cl_proto,
 		.address	= (struct sockaddr *)&clp->cl_addr,
 		.addrsize	= clp->cl_addrlen,
diff --git a/fs/nfs/dns_resolve.c b/fs/nfs/dns_resolve.c
index dba50a5625db..a6e711ad130f 100644
--- a/fs/nfs/dns_resolve.c
+++ b/fs/nfs/dns_resolve.c
@@ -167,7 +167,7 @@ static int nfs_dns_show(struct seq_file *m, struct cache_detail *cd,
 		return 0;
 	}
 	item = container_of(h, struct nfs_dns_ent, h);
-	ttl = (long)item->h.expiry_time - (long)get_seconds();
+	ttl = item->h.expiry_time - seconds_since_boot();
 	if (ttl < 0)
 		ttl = 0;
 
@@ -239,7 +239,7 @@ static int nfs_dns_parse(struct cache_detail *cd, char *buf, int buflen)
 	ttl = get_expiry(&buf);
 	if (ttl == 0)
 		goto out;
-	key.h.expiry_time = ttl + get_seconds();
+	key.h.expiry_time = ttl + seconds_since_boot();
 
 	ret = -ENOMEM;
 	item = nfs_dns_lookup(cd, &key);
@@ -301,7 +301,7 @@ static int do_cache_lookup_nowait(struct cache_detail *cd,
 		goto out_err;
 	ret = -ETIMEDOUT;
 	if (!test_bit(CACHE_VALID, &(*item)->h.flags)
-			|| (*item)->h.expiry_time < get_seconds()
+			|| (*item)->h.expiry_time < seconds_since_boot()
 			|| cd->flush_time > (*item)->h.last_refresh)
 		goto out_put;
 	ret = -ENOENT;
diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c
index d610203d95c6..eceafe74f473 100644
--- a/fs/nfs/mount_clnt.c
+++ b/fs/nfs/mount_clnt.c
@@ -153,6 +153,7 @@ int nfs_mount(struct nfs_mount_request *info)
 		.rpc_resp	= &result,
 	};
 	struct rpc_create_args args = {
+		.net		= &init_net,
 		.protocol	= info->protocol,
 		.address	= info->sap,
 		.addrsize	= info->salen,
@@ -224,6 +225,7 @@ void nfs_umount(const struct nfs_mount_request *info)
 		.to_retries = 2,
 	};
 	struct rpc_create_args args = {
+		.net		= &init_net,
 		.protocol	= IPPROTO_UDP,
 		.address	= info->sap,
 		.addrsize	= info->salen,
diff --git a/fs/nfsd/Kconfig b/fs/nfsd/Kconfig
index 7cf4ddafb4ab..31a78fce4732 100644
--- a/fs/nfsd/Kconfig
+++ b/fs/nfsd/Kconfig
@@ -29,6 +29,18 @@ config NFSD
 
 	  If unsure, say N.
 
+config NFSD_DEPRECATED
+	bool "Include support for deprecated syscall interface to NFSD"
+	depends on NFSD
+	default y
+	help
+	  The syscall interface to nfsd was obsoleted in 2.6.0 by a new
+	  filesystem based interface.  The old interface is due for removal
+	  in 2.6.40.  If you wish to remove the interface before then
+	  say N.
+
+	  In unsure, say Y.
+
 config NFSD_V2_ACL
 	bool
 	depends on NFSD
diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c
index c2a4f71d87dd..c0fcb7ab7f6d 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -28,9 +28,6 @@
 typedef struct auth_domain	svc_client;
 typedef struct svc_export	svc_export;
 
-static void		exp_do_unexport(svc_export *unexp);
-static int		exp_verify_string(char *cp, int max);
-
 /*
  * We have two caches.
  * One maps client+vfsmnt+dentry to export options - the export map
@@ -802,6 +799,7 @@ exp_find_key(svc_client *clp, int fsid_type, u32 *fsidv, struct cache_req *reqp)
 	return ek;
 }
 
+#ifdef CONFIG_NFSD_DEPRECATED
 static int exp_set_key(svc_client *clp, int fsid_type, u32 *fsidv,
 		       struct svc_export *exp)
 {
@@ -852,6 +850,7 @@ exp_get_fsid_key(svc_client *clp, int fsid)
 
 	return exp_find_key(clp, FSID_NUM, fsidv, NULL);
 }
+#endif
 
 static svc_export *exp_get_by_name(svc_client *clp, const struct path *path,
 				     struct cache_req *reqp)
@@ -893,6 +892,7 @@ static struct svc_export *exp_parent(svc_client *clp, struct path *path)
 	return exp;
 }
 
+#ifdef CONFIG_NFSD_DEPRECATED
 /*
  * Hashtable locking. Write locks are placed only by user processes
  * wanting to modify export information.
@@ -925,6 +925,19 @@ exp_writeunlock(void)
 {
 	up_write(&hash_sem);
 }
+#else
+
+/* hash_sem not needed once deprecated interface is removed */
+void exp_readlock(void) {}
+static inline void exp_writelock(void){}
+void exp_readunlock(void) {}
+static inline void exp_writeunlock(void){}
+
+#endif
+
+#ifdef CONFIG_NFSD_DEPRECATED
+static void		exp_do_unexport(svc_export *unexp);
+static int		exp_verify_string(char *cp, int max);
 
 static void exp_fsid_unhash(struct svc_export *exp)
 {
@@ -935,10 +948,9 @@ static void exp_fsid_unhash(struct svc_export *exp)
 
 	ek = exp_get_fsid_key(exp->ex_client, exp->ex_fsid);
 	if (!IS_ERR(ek)) {
-		ek->h.expiry_time = get_seconds()-1;
+		sunrpc_invalidate(&ek->h, &svc_expkey_cache);
 		cache_put(&ek->h, &svc_expkey_cache);
 	}
-	svc_expkey_cache.nextcheck = get_seconds();
 }
 
 static int exp_fsid_hash(svc_client *clp, struct svc_export *exp)
@@ -973,10 +985,9 @@ static void exp_unhash(struct svc_export *exp)
 
 	ek = exp_get_key(exp->ex_client, inode->i_sb->s_dev, inode->i_ino);
 	if (!IS_ERR(ek)) {
-		ek->h.expiry_time = get_seconds()-1;
+		sunrpc_invalidate(&ek->h, &svc_expkey_cache);
 		cache_put(&ek->h, &svc_expkey_cache);
 	}
-	svc_expkey_cache.nextcheck = get_seconds();
 }
 	
 /*
@@ -1097,8 +1108,7 @@ out:
 static void
 exp_do_unexport(svc_export *unexp)
 {
-	unexp->h.expiry_time = get_seconds()-1;
-	svc_export_cache.nextcheck = get_seconds();
+	sunrpc_invalidate(&unexp->h, &svc_export_cache);
 	exp_unhash(unexp);
 	exp_fsid_unhash(unexp);
 }
@@ -1150,6 +1160,7 @@ out_unlock:
 	exp_writeunlock();
 	return err;
 }
+#endif /* CONFIG_NFSD_DEPRECATED */
 
 /*
  * Obtain the root fh on behalf of a client.
@@ -1459,25 +1470,43 @@ static void show_secinfo_flags(struct seq_file *m, int flags)
 	show_expflags(m, flags, NFSEXP_SECINFO_FLAGS);
 }
 
+static bool secinfo_flags_equal(int f, int g)
+{
+	f &= NFSEXP_SECINFO_FLAGS;
+	g &= NFSEXP_SECINFO_FLAGS;
+	return f == g;
+}
+
+static int show_secinfo_run(struct seq_file *m, struct exp_flavor_info **fp, struct exp_flavor_info *end)
+{
+	int flags;
+
+	flags = (*fp)->flags;
+	seq_printf(m, ",sec=%d", (*fp)->pseudoflavor);
+	(*fp)++;
+	while (*fp != end && secinfo_flags_equal(flags, (*fp)->flags)) {
+		seq_printf(m, ":%d", (*fp)->pseudoflavor);
+		(*fp)++;
+	}
+	return flags;
+}
+
 static void show_secinfo(struct seq_file *m, struct svc_export *exp)
 {
 	struct exp_flavor_info *f;
 	struct exp_flavor_info *end = exp->ex_flavors + exp->ex_nflavors;
-	int lastflags = 0, first = 0;
+	int flags;
 
 	if (exp->ex_nflavors == 0)
 		return;
-	for (f = exp->ex_flavors; f < end; f++) {
-		if (first || f->flags != lastflags) {
-			if (!first)
-				show_secinfo_flags(m, lastflags);
-			seq_printf(m, ",sec=%d", f->pseudoflavor);
-			lastflags = f->flags;
-		} else {
-			seq_printf(m, ":%d", f->pseudoflavor);
-		}
+	f = exp->ex_flavors;
+	flags = show_secinfo_run(m, &f, end);
+	if (!secinfo_flags_equal(flags, exp->ex_flags))
+		show_secinfo_flags(m, flags);
+	while (f != end) {
+		flags = show_secinfo_run(m, &f, end);
+		show_secinfo_flags(m, flags);
 	}
-	show_secinfo_flags(m, lastflags);
 }
 
 static void exp_flags(struct seq_file *m, int flag, int fsid,
@@ -1532,6 +1561,7 @@ const struct seq_operations nfs_exports_op = {
 	.show	= e_show,
 };
 
+#ifdef CONFIG_NFSD_DEPRECATED
 /*
  * Add or modify a client.
  * Change requests may involve the list of host addresses. The list of
@@ -1563,7 +1593,7 @@ exp_addclient(struct nfsctl_client *ncp)
 	/* Insert client into hashtable. */
 	for (i = 0; i < ncp->cl_naddr; i++) {
 		ipv6_addr_set_v4mapped(ncp->cl_addrlist[i].s_addr, &addr6);
-		auth_unix_add_addr(&addr6, dom);
+		auth_unix_add_addr(&init_net, &addr6, dom);
 	}
 	auth_unix_forget_old(dom);
 	auth_domain_put(dom);
@@ -1621,6 +1651,7 @@ exp_verify_string(char *cp, int max)
 	printk(KERN_NOTICE "nfsd: couldn't validate string %s\n", cp);
 	return 0;
 }
+#endif /* CONFIG_NFSD_DEPRECATED */
 
 /*
  * Initialize the exports module.
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 014482c4e57d..143da2eecd7b 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -247,10 +247,11 @@ encode_cb_recall(struct xdr_stream *xdr, struct nfs4_delegation *dp,
 }
 
 static void
-encode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_cb_sequence *args,
+encode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_callback *cb,
 		   struct nfs4_cb_compound_hdr *hdr)
 {
 	__be32 *p;
+	struct nfsd4_session *ses = cb->cb_clp->cl_cb_session;
 
 	if (hdr->minorversion == 0)
 		return;
@@ -258,8 +259,8 @@ encode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_cb_sequence *args,
 	RESERVE_SPACE(1 + NFS4_MAX_SESSIONID_LEN + 20);
 
 	WRITE32(OP_CB_SEQUENCE);
-	WRITEMEM(args->cbs_clp->cl_sessionid.data, NFS4_MAX_SESSIONID_LEN);
-	WRITE32(args->cbs_clp->cl_cb_seq_nr);
+	WRITEMEM(ses->se_sessionid.data, NFS4_MAX_SESSIONID_LEN);
+	WRITE32(ses->se_cb_seq_nr);
 	WRITE32(0);		/* slotid, always 0 */
 	WRITE32(0);		/* highest slotid always 0 */
 	WRITE32(0);		/* cachethis always 0 */
@@ -279,18 +280,18 @@ nfs4_xdr_enc_cb_null(struct rpc_rqst *req, __be32 *p)
 
 static int
 nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, __be32 *p,
-		struct nfs4_rpc_args *rpc_args)
+		struct nfsd4_callback *cb)
 {
 	struct xdr_stream xdr;
-	struct nfs4_delegation *args = rpc_args->args_op;
+	struct nfs4_delegation *args = cb->cb_op;
 	struct nfs4_cb_compound_hdr hdr = {
-		.ident = args->dl_ident,
-		.minorversion = rpc_args->args_seq.cbs_minorversion,
+		.ident = cb->cb_clp->cl_cb_ident,
+		.minorversion = cb->cb_minorversion,
 	};
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_cb_compound_hdr(&xdr, &hdr);
-	encode_cb_sequence(&xdr, &rpc_args->args_seq, &hdr);
+	encode_cb_sequence(&xdr, cb, &hdr);
 	encode_cb_recall(&xdr, args, &hdr);
 	encode_cb_nops(&hdr);
 	return 0;
@@ -338,15 +339,16 @@ decode_cb_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected)
  * with a single slot.
  */
 static int
-decode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_cb_sequence *res,
+decode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_callback *cb,
 		   struct rpc_rqst *rqstp)
 {
+	struct nfsd4_session *ses = cb->cb_clp->cl_cb_session;
 	struct nfs4_sessionid id;
 	int status;
 	u32 dummy;
 	__be32 *p;
 
-	if (res->cbs_minorversion == 0)
+	if (cb->cb_minorversion == 0)
 		return 0;
 
 	status = decode_cb_op_hdr(xdr, OP_CB_SEQUENCE);
@@ -362,13 +364,12 @@ decode_cb_sequence(struct xdr_stream *xdr, struct nfsd4_cb_sequence *res,
 	READ_BUF(NFS4_MAX_SESSIONID_LEN + 16);
 	memcpy(id.data, p, NFS4_MAX_SESSIONID_LEN);
 	p += XDR_QUADLEN(NFS4_MAX_SESSIONID_LEN);
-	if (memcmp(id.data, res->cbs_clp->cl_sessionid.data,
-		   NFS4_MAX_SESSIONID_LEN)) {
+	if (memcmp(id.data, ses->se_sessionid.data, NFS4_MAX_SESSIONID_LEN)) {
 		dprintk("%s Invalid session id\n", __func__);
 		goto out;
 	}
 	READ32(dummy);
-	if (dummy != res->cbs_clp->cl_cb_seq_nr) {
+	if (dummy != ses->se_cb_seq_nr) {
 		dprintk("%s Invalid sequence number\n", __func__);
 		goto out;
 	}
@@ -392,7 +393,7 @@ nfs4_xdr_dec_cb_null(struct rpc_rqst *req, __be32 *p)
 
 static int
 nfs4_xdr_dec_cb_recall(struct rpc_rqst *rqstp, __be32 *p,
-		struct nfsd4_cb_sequence *seq)
+		struct nfsd4_callback *cb)
 {
 	struct xdr_stream xdr;
 	struct nfs4_cb_compound_hdr hdr;
@@ -402,8 +403,8 @@ nfs4_xdr_dec_cb_recall(struct rpc_rqst *rqstp, __be32 *p,
 	status = decode_cb_compound_hdr(&xdr, &hdr);
 	if (status)
 		goto out;
-	if (seq) {
-		status = decode_cb_sequence(&xdr, seq, rqstp);
+	if (cb) {
+		status = decode_cb_sequence(&xdr, cb, rqstp);
 		if (status)
 			goto out;
 	}
@@ -472,30 +473,34 @@ static int max_cb_time(void)
 /* Reference counting, callback cleanup, etc., all look racy as heck.
  * And why is cl_cb_set an atomic? */
 
-int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *cb)
+int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *conn)
 {
 	struct rpc_timeout	timeparms = {
 		.to_initval	= max_cb_time(),
 		.to_retries	= 0,
 	};
 	struct rpc_create_args args = {
-		.protocol	= XPRT_TRANSPORT_TCP,
-		.address	= (struct sockaddr *) &cb->cb_addr,
-		.addrsize	= cb->cb_addrlen,
+		.net		= &init_net,
+		.address	= (struct sockaddr *) &conn->cb_addr,
+		.addrsize	= conn->cb_addrlen,
 		.timeout	= &timeparms,
 		.program	= &cb_program,
-		.prognumber	= cb->cb_prog,
 		.version	= 0,
 		.authflavor	= clp->cl_flavor,
 		.flags		= (RPC_CLNT_CREATE_NOPING | RPC_CLNT_CREATE_QUIET),
-		.client_name    = clp->cl_principal,
 	};
 	struct rpc_clnt *client;
 
-	if (!clp->cl_principal && (clp->cl_flavor >= RPC_AUTH_GSS_KRB5))
-		return -EINVAL;
-	if (cb->cb_minorversion) {
-		args.bc_xprt = cb->cb_xprt;
+	if (clp->cl_minorversion == 0) {
+		if (!clp->cl_principal && (clp->cl_flavor >= RPC_AUTH_GSS_KRB5))
+			return -EINVAL;
+		args.client_name = clp->cl_principal;
+		args.prognumber	= conn->cb_prog,
+		args.protocol = XPRT_TRANSPORT_TCP;
+		clp->cl_cb_ident = conn->cb_ident;
+	} else {
+		args.bc_xprt = conn->cb_xprt;
+		args.prognumber = clp->cl_cb_session->se_cb_prog;
 		args.protocol = XPRT_TRANSPORT_BC_TCP;
 	}
 	/* Create RPC client */
@@ -505,7 +510,7 @@ int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *cb)
 			PTR_ERR(client));
 		return PTR_ERR(client);
 	}
-	nfsd4_set_callback_client(clp, client);
+	clp->cl_cb_client = client;
 	return 0;
 
 }
@@ -518,7 +523,7 @@ static void warn_no_callback_path(struct nfs4_client *clp, int reason)
 
 static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs4_client *clp = calldata;
+	struct nfs4_client *clp = container_of(calldata, struct nfs4_client, cl_cb_null);
 
 	if (task->tk_status)
 		warn_no_callback_path(clp, task->tk_status);
@@ -527,6 +532,8 @@ static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata)
 }
 
 static const struct rpc_call_ops nfsd4_cb_probe_ops = {
+	/* XXX: release method to ensure we set the cb channel down if
+	 * necessary on early failure? */
 	.rpc_call_done = nfsd4_cb_probe_done,
 };
 
@@ -542,38 +549,42 @@ int set_callback_cred(void)
 	return 0;
 }
 
+static struct workqueue_struct *callback_wq;
 
-void do_probe_callback(struct nfs4_client *clp)
+static void do_probe_callback(struct nfs4_client *clp)
 {
-	struct rpc_message msg = {
-		.rpc_proc       = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL],
-		.rpc_argp       = clp,
-		.rpc_cred	= callback_cred
-	};
-	int status;
+	struct nfsd4_callback *cb = &clp->cl_cb_null;
 
-	status = rpc_call_async(clp->cl_cb_client, &msg,
-				RPC_TASK_SOFT | RPC_TASK_SOFTCONN,
-				&nfsd4_cb_probe_ops, (void *)clp);
-	if (status)
-		warn_no_callback_path(clp, status);
+	cb->cb_op = NULL;
+	cb->cb_clp = clp;
+
+	cb->cb_msg.rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_NULL];
+	cb->cb_msg.rpc_argp = NULL;
+	cb->cb_msg.rpc_resp = NULL;
+	cb->cb_msg.rpc_cred = callback_cred;
+
+	cb->cb_ops = &nfsd4_cb_probe_ops;
+
+	queue_work(callback_wq, &cb->cb_work);
 }
 
 /*
- * Set up the callback client and put a NFSPROC4_CB_NULL on the wire...
+ * Poke the callback thread to process any updates to the callback
+ * parameters, and send a null probe.
  */
-void nfsd4_probe_callback(struct nfs4_client *clp, struct nfs4_cb_conn *cb)
+void nfsd4_probe_callback(struct nfs4_client *clp)
 {
-	int status;
+	set_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_cb_flags);
+	do_probe_callback(clp);
+}
 
+void nfsd4_change_callback(struct nfs4_client *clp, struct nfs4_cb_conn *conn)
+{
 	BUG_ON(atomic_read(&clp->cl_cb_set));
 
-	status = setup_callback_client(clp, cb);
-	if (status) {
-		warn_no_callback_path(clp, status);
-		return;
-	}
-	do_probe_callback(clp);
+	spin_lock(&clp->cl_lock);
+	memcpy(&clp->cl_cb_conn, conn, sizeof(struct nfs4_cb_conn));
+	spin_unlock(&clp->cl_lock);
 }
 
 /*
@@ -584,8 +595,7 @@ void nfsd4_probe_callback(struct nfs4_client *clp, struct nfs4_cb_conn *cb)
 static int nfsd41_cb_setup_sequence(struct nfs4_client *clp,
 		struct rpc_task *task)
 {
-	struct nfs4_rpc_args *args = task->tk_msg.rpc_argp;
-	u32 *ptr = (u32 *)clp->cl_sessionid.data;
+	u32 *ptr = (u32 *)clp->cl_cb_session->se_sessionid.data;
 	int status = 0;
 
 	dprintk("%s: %u:%u:%u:%u\n", __func__,
@@ -597,14 +607,6 @@ static int nfsd41_cb_setup_sequence(struct nfs4_client *clp,
 		status = -EAGAIN;
 		goto out;
 	}
-
-	/*
-	 * We'll need the clp during XDR encoding and decoding,
-	 * and the sequence during decoding to verify the reply
-	 */
-	args->args_seq.cbs_clp = clp;
-	task->tk_msg.rpc_resp = &args->args_seq;
-
 out:
 	dprintk("%s status=%d\n", __func__, status);
 	return status;
@@ -616,13 +618,13 @@ out:
  */
 static void nfsd4_cb_prepare(struct rpc_task *task, void *calldata)
 {
-	struct nfs4_delegation *dp = calldata;
+	struct nfsd4_callback *cb = calldata;
+	struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall);
 	struct nfs4_client *clp = dp->dl_client;
-	struct nfs4_rpc_args *args = task->tk_msg.rpc_argp;
-	u32 minorversion = clp->cl_cb_conn.cb_minorversion;
+	u32 minorversion = clp->cl_minorversion;
 	int status = 0;
 
-	args->args_seq.cbs_minorversion = minorversion;
+	cb->cb_minorversion = minorversion;
 	if (minorversion) {
 		status = nfsd41_cb_setup_sequence(clp, task);
 		if (status) {
@@ -639,19 +641,20 @@ static void nfsd4_cb_prepare(struct rpc_task *task, void *calldata)
 
 static void nfsd4_cb_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs4_delegation *dp = calldata;
+	struct nfsd4_callback *cb = calldata;
+	struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall);
 	struct nfs4_client *clp = dp->dl_client;
 
 	dprintk("%s: minorversion=%d\n", __func__,
-		clp->cl_cb_conn.cb_minorversion);
+		clp->cl_minorversion);
 
-	if (clp->cl_cb_conn.cb_minorversion) {
+	if (clp->cl_minorversion) {
 		/* No need for lock, access serialized in nfsd4_cb_prepare */
-		++clp->cl_cb_seq_nr;
+		++clp->cl_cb_session->se_cb_seq_nr;
 		clear_bit(0, &clp->cl_cb_slot_busy);
 		rpc_wake_up_next(&clp->cl_cb_waitq);
 		dprintk("%s: freed slot, new seqid=%d\n", __func__,
-			clp->cl_cb_seq_nr);
+			clp->cl_cb_session->se_cb_seq_nr);
 
 		/* We're done looking into the sequence information */
 		task->tk_msg.rpc_resp = NULL;
@@ -661,7 +664,8 @@ static void nfsd4_cb_done(struct rpc_task *task, void *calldata)
 
 static void nfsd4_cb_recall_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs4_delegation *dp = calldata;
+	struct nfsd4_callback *cb = calldata;
+	struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall);
 	struct nfs4_client *clp = dp->dl_client;
 	struct rpc_clnt *current_rpc_client = clp->cl_cb_client;
 
@@ -706,7 +710,8 @@ static void nfsd4_cb_recall_done(struct rpc_task *task, void *calldata)
 
 static void nfsd4_cb_recall_release(void *calldata)
 {
-	struct nfs4_delegation *dp = calldata;
+	struct nfsd4_callback *cb = calldata;
+	struct nfs4_delegation *dp = container_of(cb, struct nfs4_delegation, dl_recall);
 
 	nfs4_put_delegation(dp);
 }
@@ -717,8 +722,6 @@ static const struct rpc_call_ops nfsd4_cb_recall_ops = {
 	.rpc_release = nfsd4_cb_recall_release,
 };
 
-static struct workqueue_struct *callback_wq;
-
 int nfsd4_create_callback_queue(void)
 {
 	callback_wq = create_singlethread_workqueue("nfsd4_callbacks");
@@ -733,57 +736,88 @@ void nfsd4_destroy_callback_queue(void)
 }
 
 /* must be called under the state lock */
-void nfsd4_set_callback_client(struct nfs4_client *clp, struct rpc_clnt *new)
+void nfsd4_shutdown_callback(struct nfs4_client *clp)
 {
-	struct rpc_clnt *old = clp->cl_cb_client;
-
-	clp->cl_cb_client = new;
+	set_bit(NFSD4_CLIENT_KILL, &clp->cl_cb_flags);
 	/*
-	 * After this, any work that saw the old value of cl_cb_client will
-	 * be gone:
+	 * Note this won't actually result in a null callback;
+	 * instead, nfsd4_do_callback_rpc() will detect the killed
+	 * client, destroy the rpc client, and stop:
 	 */
+	do_probe_callback(clp);
 	flush_workqueue(callback_wq);
-	/* So we can safely shut it down: */
-	if (old)
-		rpc_shutdown_client(old);
 }
 
-/*
- * called with dp->dl_count inc'ed.
- */
-static void _nfsd4_cb_recall(struct nfs4_delegation *dp)
+void nfsd4_release_cb(struct nfsd4_callback *cb)
 {
-	struct nfs4_client *clp = dp->dl_client;
-	struct rpc_clnt *clnt = clp->cl_cb_client;
-	struct nfs4_rpc_args *args = &dp->dl_recall.cb_args;
-	struct rpc_message msg = {
-		.rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_RECALL],
-		.rpc_cred = callback_cred
-	};
+	if (cb->cb_ops->rpc_release)
+		cb->cb_ops->rpc_release(cb);
+}
 
-	if (clnt == NULL) {
-		nfs4_put_delegation(dp);
-		return; /* Client is shutting down; give up. */
+void nfsd4_process_cb_update(struct nfsd4_callback *cb)
+{
+	struct nfs4_cb_conn conn;
+	struct nfs4_client *clp = cb->cb_clp;
+	int err;
+
+	/*
+	 * This is either an update, or the client dying; in either case,
+	 * kill the old client:
+	 */
+	if (clp->cl_cb_client) {
+		rpc_shutdown_client(clp->cl_cb_client);
+		clp->cl_cb_client = NULL;
 	}
+	if (test_bit(NFSD4_CLIENT_KILL, &clp->cl_cb_flags))
+		return;
+	spin_lock(&clp->cl_lock);
+	/*
+	 * Only serialized callback code is allowed to clear these
+	 * flags; main nfsd code can only set them:
+	 */
+	BUG_ON(!clp->cl_cb_flags);
+	clear_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_cb_flags);
+	memcpy(&conn, &cb->cb_clp->cl_cb_conn, sizeof(struct nfs4_cb_conn));
+	spin_unlock(&clp->cl_lock);
 
-	args->args_op = dp;
-	msg.rpc_argp = args;
-	dp->dl_retries = 1;
-	rpc_call_async(clnt, &msg, RPC_TASK_SOFT, &nfsd4_cb_recall_ops, dp);
+	err = setup_callback_client(clp, &conn);
+	if (err)
+		warn_no_callback_path(clp, err);
 }
 
 void nfsd4_do_callback_rpc(struct work_struct *w)
 {
-	/* XXX: for now, just send off delegation recall. */
-	/* In future, generalize to handle any sort of callback. */
-	struct nfsd4_callback *c = container_of(w, struct nfsd4_callback, cb_work);
-	struct nfs4_delegation *dp = container_of(c, struct nfs4_delegation, dl_recall);
+	struct nfsd4_callback *cb = container_of(w, struct nfsd4_callback, cb_work);
+	struct nfs4_client *clp = cb->cb_clp;
+	struct rpc_clnt *clnt;
 
-	_nfsd4_cb_recall(dp);
-}
+	if (clp->cl_cb_flags)
+		nfsd4_process_cb_update(cb);
 
+	clnt = clp->cl_cb_client;
+	if (!clnt) {
+		/* Callback channel broken, or client killed; give up: */
+		nfsd4_release_cb(cb);
+		return;
+	}
+	rpc_call_async(clnt, &cb->cb_msg, RPC_TASK_SOFT | RPC_TASK_SOFTCONN,
+			cb->cb_ops, cb);
+}
 
 void nfsd4_cb_recall(struct nfs4_delegation *dp)
 {
+	struct nfsd4_callback *cb = &dp->dl_recall;
+
+	dp->dl_retries = 1;
+	cb->cb_op = dp;
+	cb->cb_clp = dp->dl_client;
+	cb->cb_msg.rpc_proc = &nfs4_cb_procedures[NFSPROC4_CLNT_CB_RECALL];
+	cb->cb_msg.rpc_argp = cb;
+	cb->cb_msg.rpc_resp = cb;
+	cb->cb_msg.rpc_cred = callback_cred;
+
+	cb->cb_ops = &nfsd4_cb_recall_ops;
+	dp->dl_retries = 1;
+
 	queue_work(callback_wq, &dp->dl_recall.cb_work);
 }
diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c
index c78dbf493424..f0695e815f0e 100644
--- a/fs/nfsd/nfs4idmap.c
+++ b/fs/nfsd/nfs4idmap.c
@@ -482,109 +482,26 @@ nfsd_idmap_shutdown(void)
 	cache_unregister(&nametoid_cache);
 }
 
-/*
- * Deferred request handling
- */
-
-struct idmap_defer_req {
-       struct cache_req		req;
-       struct cache_deferred_req deferred_req;
-       wait_queue_head_t	waitq;
-       atomic_t			count;
-};
-
-static inline void
-put_mdr(struct idmap_defer_req *mdr)
-{
-	if (atomic_dec_and_test(&mdr->count))
-		kfree(mdr);
-}
-
-static inline void
-get_mdr(struct idmap_defer_req *mdr)
-{
-	atomic_inc(&mdr->count);
-}
-
-static void
-idmap_revisit(struct cache_deferred_req *dreq, int toomany)
-{
-	struct idmap_defer_req *mdr =
-		container_of(dreq, struct idmap_defer_req, deferred_req);
-
-	wake_up(&mdr->waitq);
-	put_mdr(mdr);
-}
-
-static struct cache_deferred_req *
-idmap_defer(struct cache_req *req)
-{
-	struct idmap_defer_req *mdr =
-		container_of(req, struct idmap_defer_req, req);
-
-	mdr->deferred_req.revisit = idmap_revisit;
-	get_mdr(mdr);
-	return (&mdr->deferred_req);
-}
-
-static inline int
-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);
-	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 *),
-			struct ent *key, struct cache_detail *detail,
-			struct ent **item)
-{
-	int ret = -ENOMEM;
-
-	*item = lookup_fn(key);
-	if (!*item)
-		goto out_err;
-	ret = -ETIMEDOUT;
-	if (!test_bit(CACHE_VALID, &(*item)->h.flags)
-			|| (*item)->h.expiry_time < get_seconds()
-			|| detail->flush_time > (*item)->h.last_refresh)
-		goto out_put;
-	ret = -ENOENT;
-	if (test_bit(CACHE_NEGATIVE, &(*item)->h.flags))
-		goto out_put;
-	return 0;
-out_put:
-	cache_put(&(*item)->h, detail);
-out_err:
-	*item = NULL;
-	return ret;
-}
-
 static int
 idmap_lookup(struct svc_rqst *rqstp,
 		struct ent *(*lookup_fn)(struct ent *), struct ent *key,
 		struct cache_detail *detail, struct ent **item)
 {
-	struct idmap_defer_req *mdr;
 	int ret;
 
-	mdr = kzalloc(sizeof(*mdr), GFP_KERNEL);
-	if (!mdr)
+	*item = lookup_fn(key);
+	if (!*item)
 		return -ENOMEM;
-	atomic_set(&mdr->count, 1);
-	init_waitqueue_head(&mdr->waitq);
-	mdr->req.defer = idmap_defer;
-	ret = do_idmap_lookup(lookup_fn, key, detail, item, mdr);
-	if (ret == -EAGAIN) {
-		wait_event_interruptible_timeout(mdr->waitq,
-			test_bit(CACHE_VALID, &(*item)->h.flags), 1 * HZ);
-		ret = do_idmap_lookup_nowait(lookup_fn, key, detail, item);
+ retry:
+	ret = cache_check(detail, &(*item)->h, &rqstp->rq_chandle);
+
+	if (ret == -ETIMEDOUT) {
+		struct ent *prev_item = *item;
+		*item = lookup_fn(key);
+		if (*item != prev_item)
+			goto retry;
+		cache_put(&(*item)->h, detail);
 	}
-	put_mdr(mdr);
 	return ret;
 }
 
diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c
index 59ec449b0c7f..0cdfd022bb7b 100644
--- a/fs/nfsd/nfs4proc.c
+++ b/fs/nfsd/nfs4proc.c
@@ -1031,8 +1031,11 @@ nfsd4_proc_compound(struct svc_rqst *rqstp,
 	resp->cstate.session = NULL;
 	fh_init(&resp->cstate.current_fh, NFS4_FHSIZE);
 	fh_init(&resp->cstate.save_fh, NFS4_FHSIZE);
-	/* Use the deferral mechanism only for NFSv4.0 compounds */
-	rqstp->rq_usedeferral = (args->minorversion == 0);
+	/*
+	 * Don't use the deferral mechanism for NFSv4; compounds make it
+	 * too hard to avoid non-idempotency problems.
+	 */
+	rqstp->rq_usedeferral = 0;
 
 	/*
 	 * According to RFC3010, this takes precedence over all other errors.
diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index a7292fcf7718..9019e8ec9dc8 100644
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -207,7 +207,6 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f
 {
 	struct nfs4_delegation *dp;
 	struct nfs4_file *fp = stp->st_file;
-	struct nfs4_cb_conn *cb = &stp->st_stateowner->so_client->cl_cb_conn;
 
 	dprintk("NFSD alloc_init_deleg\n");
 	/*
@@ -234,7 +233,6 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_stateid *stp, struct svc_f
 	nfs4_file_get_access(fp, O_RDONLY);
 	dp->dl_flock = NULL;
 	dp->dl_type = type;
-	dp->dl_ident = cb->cb_ident;
 	dp->dl_stateid.si_boot = boot_time;
 	dp->dl_stateid.si_stateownerid = current_delegid++;
 	dp->dl_stateid.si_fileid = 0;
@@ -535,171 +533,258 @@ gen_sessionid(struct nfsd4_session *ses)
  */
 #define NFSD_MIN_HDR_SEQ_SZ  (24 + 12 + 44)
 
+static void
+free_session_slots(struct nfsd4_session *ses)
+{
+	int i;
+
+	for (i = 0; i < ses->se_fchannel.maxreqs; i++)
+		kfree(ses->se_slots[i]);
+}
+
 /*
- * Give the client the number of ca_maxresponsesize_cached slots it
- * requests, of size bounded by NFSD_SLOT_CACHE_SIZE,
- * NFSD_MAX_MEM_PER_SESSION, and nfsd_drc_max_mem. Do not allow more
- * than NFSD_MAX_SLOTS_PER_SESSION.
- *
- * If we run out of reserved DRC memory we should (up to a point)
+ * We don't actually need to cache the rpc and session headers, so we
+ * can allocate a little less for each slot:
+ */
+static inline int slot_bytes(struct nfsd4_channel_attrs *ca)
+{
+	return ca->maxresp_cached - NFSD_MIN_HDR_SEQ_SZ;
+}
+
+static int nfsd4_sanitize_slot_size(u32 size)
+{
+	size -= NFSD_MIN_HDR_SEQ_SZ; /* We don't cache the rpc header */
+	size = min_t(u32, size, NFSD_SLOT_CACHE_SIZE);
+
+	return size;
+}
+
+/*
+ * XXX: If we run out of reserved DRC memory we could (up to a point)
  * re-negotiate active sessions and reduce their slot usage to make
  * rooom for new connections. For now we just fail the create session.
  */
-static int set_forechannel_drc_size(struct nfsd4_channel_attrs *fchan)
+static int nfsd4_get_drc_mem(int slotsize, u32 num)
 {
-	int mem, size = fchan->maxresp_cached;
+	int avail;
 
-	if (fchan->maxreqs < 1)
-		return nfserr_inval;
+	num = min_t(u32, num, NFSD_MAX_SLOTS_PER_SESSION);
 
-	if (size < NFSD_MIN_HDR_SEQ_SZ)
-		size = NFSD_MIN_HDR_SEQ_SZ;
-	size -= NFSD_MIN_HDR_SEQ_SZ;
-	if (size > NFSD_SLOT_CACHE_SIZE)
-		size = NFSD_SLOT_CACHE_SIZE;
-
-	/* bound the maxreqs by NFSD_MAX_MEM_PER_SESSION */
-	mem = fchan->maxreqs * size;
-	if (mem > NFSD_MAX_MEM_PER_SESSION) {
-		fchan->maxreqs = NFSD_MAX_MEM_PER_SESSION / size;
-		if (fchan->maxreqs > NFSD_MAX_SLOTS_PER_SESSION)
-			fchan->maxreqs = NFSD_MAX_SLOTS_PER_SESSION;
-		mem = fchan->maxreqs * size;
-	}
+	spin_lock(&nfsd_drc_lock);
+	avail = min_t(int, NFSD_MAX_MEM_PER_SESSION,
+			nfsd_drc_max_mem - nfsd_drc_mem_used);
+	num = min_t(int, num, avail / slotsize);
+	nfsd_drc_mem_used += num * slotsize;
+	spin_unlock(&nfsd_drc_lock);
 
+	return num;
+}
+
+static void nfsd4_put_drc_mem(int slotsize, int num)
+{
 	spin_lock(&nfsd_drc_lock);
-	/* bound the total session drc memory ussage */
-	if (mem + nfsd_drc_mem_used > nfsd_drc_max_mem) {
-		fchan->maxreqs = (nfsd_drc_max_mem - nfsd_drc_mem_used) / size;
-		mem = fchan->maxreqs * size;
-	}
-	nfsd_drc_mem_used += mem;
+	nfsd_drc_mem_used -= slotsize * num;
 	spin_unlock(&nfsd_drc_lock);
+}
 
-	if (fchan->maxreqs == 0)
-		return nfserr_jukebox;
+static struct nfsd4_session *alloc_session(int slotsize, int numslots)
+{
+	struct nfsd4_session *new;
+	int mem, i;
 
-	fchan->maxresp_cached = size + NFSD_MIN_HDR_SEQ_SZ;
-	return 0;
+	BUILD_BUG_ON(NFSD_MAX_SLOTS_PER_SESSION * sizeof(struct nfsd4_slot *)
+			+ sizeof(struct nfsd4_session) > PAGE_SIZE);
+	mem = numslots * sizeof(struct nfsd4_slot *);
+
+	new = kzalloc(sizeof(*new) + mem, GFP_KERNEL);
+	if (!new)
+		return NULL;
+	/* allocate each struct nfsd4_slot and data cache in one piece */
+	for (i = 0; i < numslots; i++) {
+		mem = sizeof(struct nfsd4_slot) + slotsize;
+		new->se_slots[i] = kzalloc(mem, GFP_KERNEL);
+		if (!new->se_slots[i])
+			goto out_free;
+	}
+	return new;
+out_free:
+	while (i--)
+		kfree(new->se_slots[i]);
+	kfree(new);
+	return NULL;
 }
 
-/*
- * fchan holds the client values on input, and the server values on output
- * sv_max_mesg is the maximum payload plus one page for overhead.
- */
-static int init_forechannel_attrs(struct svc_rqst *rqstp,
-				  struct nfsd4_channel_attrs *session_fchan,
-				  struct nfsd4_channel_attrs *fchan)
+static void init_forechannel_attrs(struct nfsd4_channel_attrs *new, struct nfsd4_channel_attrs *req, int numslots, int slotsize)
 {
-	int status = 0;
-	__u32   maxcount = nfsd_serv->sv_max_mesg;
+	u32 maxrpc = nfsd_serv->sv_max_mesg;
 
-	/* headerpadsz set to zero in encode routine */
+	new->maxreqs = numslots;
+	new->maxresp_cached = slotsize + NFSD_MIN_HDR_SEQ_SZ;
+	new->maxreq_sz = min_t(u32, req->maxreq_sz, maxrpc);
+	new->maxresp_sz = min_t(u32, req->maxresp_sz, maxrpc);
+	new->maxops = min_t(u32, req->maxops, NFSD_MAX_OPS_PER_COMPOUND);
+}
 
-	/* Use the client's max request and max response size if possible */
-	if (fchan->maxreq_sz > maxcount)
-		fchan->maxreq_sz = maxcount;
-	session_fchan->maxreq_sz = fchan->maxreq_sz;
+static void free_conn(struct nfsd4_conn *c)
+{
+	svc_xprt_put(c->cn_xprt);
+	kfree(c);
+}
 
-	if (fchan->maxresp_sz > maxcount)
-		fchan->maxresp_sz = maxcount;
-	session_fchan->maxresp_sz = fchan->maxresp_sz;
+static void nfsd4_conn_lost(struct svc_xpt_user *u)
+{
+	struct nfsd4_conn *c = container_of(u, struct nfsd4_conn, cn_xpt_user);
+	struct nfs4_client *clp = c->cn_session->se_client;
 
-	/* Use the client's maxops if possible */
-	if (fchan->maxops > NFSD_MAX_OPS_PER_COMPOUND)
-		fchan->maxops = NFSD_MAX_OPS_PER_COMPOUND;
-	session_fchan->maxops = fchan->maxops;
+	spin_lock(&clp->cl_lock);
+	if (!list_empty(&c->cn_persession)) {
+		list_del(&c->cn_persession);
+		free_conn(c);
+	}
+	spin_unlock(&clp->cl_lock);
+}
 
-	/* FIXME: Error means no more DRC pages so the server should
-	 * recover pages from existing sessions. For now fail session
-	 * creation.
-	 */
-	status = set_forechannel_drc_size(fchan);
+static struct nfsd4_conn *alloc_conn(struct svc_rqst *rqstp, u32 flags)
+{
+	struct nfsd4_conn *conn;
 
-	session_fchan->maxresp_cached = fchan->maxresp_cached;
-	session_fchan->maxreqs = fchan->maxreqs;
+	conn = kmalloc(sizeof(struct nfsd4_conn), GFP_KERNEL);
+	if (!conn)
+		return NULL;
+	svc_xprt_get(rqstp->rq_xprt);
+	conn->cn_xprt = rqstp->rq_xprt;
+	conn->cn_flags = flags;
+	INIT_LIST_HEAD(&conn->cn_xpt_user.list);
+	return conn;
+}
 
-	dprintk("%s status %d\n", __func__, status);
-	return status;
+static void __nfsd4_hash_conn(struct nfsd4_conn *conn, struct nfsd4_session *ses)
+{
+	conn->cn_session = ses;
+	list_add(&conn->cn_persession, &ses->se_conns);
 }
 
-static void
-free_session_slots(struct nfsd4_session *ses)
+static void nfsd4_hash_conn(struct nfsd4_conn *conn, struct nfsd4_session *ses)
 {
-	int i;
+	struct nfs4_client *clp = ses->se_client;
 
-	for (i = 0; i < ses->se_fchannel.maxreqs; i++)
-		kfree(ses->se_slots[i]);
+	spin_lock(&clp->cl_lock);
+	__nfsd4_hash_conn(conn, ses);
+	spin_unlock(&clp->cl_lock);
 }
 
-/*
- * We don't actually need to cache the rpc and session headers, so we
- * can allocate a little less for each slot:
- */
-static inline int slot_bytes(struct nfsd4_channel_attrs *ca)
+static void nfsd4_register_conn(struct nfsd4_conn *conn)
 {
-	return ca->maxresp_cached - NFSD_MIN_HDR_SEQ_SZ;
+	conn->cn_xpt_user.callback = nfsd4_conn_lost;
+	register_xpt_user(conn->cn_xprt, &conn->cn_xpt_user);
 }
 
-static int
-alloc_init_session(struct svc_rqst *rqstp, struct nfs4_client *clp,
-		   struct nfsd4_create_session *cses)
+static __be32 nfsd4_new_conn(struct svc_rqst *rqstp, struct nfsd4_session *ses)
 {
-	struct nfsd4_session *new, tmp;
-	struct nfsd4_slot *sp;
-	int idx, slotsize, cachesize, i;
-	int status;
+	struct nfsd4_conn *conn;
+	u32 flags = NFS4_CDFC4_FORE;
 
-	memset(&tmp, 0, sizeof(tmp));
+	if (ses->se_flags & SESSION4_BACK_CHAN)
+		flags |= NFS4_CDFC4_BACK;
+	conn = alloc_conn(rqstp, flags);
+	if (!conn)
+		return nfserr_jukebox;
+	nfsd4_hash_conn(conn, ses);
+	nfsd4_register_conn(conn);
+	return nfs_ok;
+}
 
-	/* FIXME: For now, we just accept the client back channel attributes. */
-	tmp.se_bchannel = cses->back_channel;
-	status = init_forechannel_attrs(rqstp, &tmp.se_fchannel,
-					&cses->fore_channel);
-	if (status)
-		goto out;
+static void nfsd4_del_conns(struct nfsd4_session *s)
+{
+	struct nfs4_client *clp = s->se_client;
+	struct nfsd4_conn *c;
 
-	BUILD_BUG_ON(NFSD_MAX_SLOTS_PER_SESSION * sizeof(struct nfsd4_slot)
-		     + sizeof(struct nfsd4_session) > PAGE_SIZE);
+	spin_lock(&clp->cl_lock);
+	while (!list_empty(&s->se_conns)) {
+		c = list_first_entry(&s->se_conns, struct nfsd4_conn, cn_persession);
+		list_del_init(&c->cn_persession);
+		spin_unlock(&clp->cl_lock);
 
-	status = nfserr_jukebox;
-	/* allocate struct nfsd4_session and slot table pointers in one piece */
-	slotsize = tmp.se_fchannel.maxreqs * sizeof(struct nfsd4_slot *);
-	new = kzalloc(sizeof(*new) + slotsize, GFP_KERNEL);
-	if (!new)
-		goto out;
+		unregister_xpt_user(c->cn_xprt, &c->cn_xpt_user);
+		free_conn(c);
 
-	memcpy(new, &tmp, sizeof(*new));
+		spin_lock(&clp->cl_lock);
+	}
+	spin_unlock(&clp->cl_lock);
+}
 
-	/* allocate each struct nfsd4_slot and data cache in one piece */
-	cachesize = slot_bytes(&new->se_fchannel);
-	for (i = 0; i < new->se_fchannel.maxreqs; i++) {
-		sp = kzalloc(sizeof(*sp) + cachesize, GFP_KERNEL);
-		if (!sp)
-			goto out_free;
-		new->se_slots[i] = sp;
+void free_session(struct kref *kref)
+{
+	struct nfsd4_session *ses;
+	int mem;
+
+	ses = container_of(kref, struct nfsd4_session, se_ref);
+	nfsd4_del_conns(ses);
+	spin_lock(&nfsd_drc_lock);
+	mem = ses->se_fchannel.maxreqs * slot_bytes(&ses->se_fchannel);
+	nfsd_drc_mem_used -= mem;
+	spin_unlock(&nfsd_drc_lock);
+	free_session_slots(ses);
+	kfree(ses);
+}
+
+static struct nfsd4_session *alloc_init_session(struct svc_rqst *rqstp, struct nfs4_client *clp, struct nfsd4_create_session *cses)
+{
+	struct nfsd4_session *new;
+	struct nfsd4_channel_attrs *fchan = &cses->fore_channel;
+	int numslots, slotsize;
+	int status;
+	int idx;
+
+	/*
+	 * Note decreasing slot size below client's request may
+	 * make it difficult for client to function correctly, whereas
+	 * decreasing the number of slots will (just?) affect
+	 * performance.  When short on memory we therefore prefer to
+	 * decrease number of slots instead of their size.
+	 */
+	slotsize = nfsd4_sanitize_slot_size(fchan->maxresp_cached);
+	numslots = nfsd4_get_drc_mem(slotsize, fchan->maxreqs);
+
+	new = alloc_session(slotsize, numslots);
+	if (!new) {
+		nfsd4_put_drc_mem(slotsize, fchan->maxreqs);
+		return NULL;
 	}
+	init_forechannel_attrs(&new->se_fchannel, fchan, numslots, slotsize);
 
 	new->se_client = clp;
 	gen_sessionid(new);
-	idx = hash_sessionid(&new->se_sessionid);
-	memcpy(clp->cl_sessionid.data, new->se_sessionid.data,
-	       NFS4_MAX_SESSIONID_LEN);
 
+	INIT_LIST_HEAD(&new->se_conns);
+
+	new->se_cb_seq_nr = 1;
 	new->se_flags = cses->flags;
+	new->se_cb_prog = cses->callback_prog;
 	kref_init(&new->se_ref);
+	idx = hash_sessionid(&new->se_sessionid);
 	spin_lock(&client_lock);
 	list_add(&new->se_hash, &sessionid_hashtbl[idx]);
 	list_add(&new->se_perclnt, &clp->cl_sessions);
 	spin_unlock(&client_lock);
 
-	status = nfs_ok;
-out:
-	return status;
-out_free:
-	free_session_slots(new);
-	kfree(new);
-	goto out;
+	status = nfsd4_new_conn(rqstp, new);
+	/* whoops: benny points out, status is ignored! (err, or bogus) */
+	if (status) {
+		free_session(&new->se_ref);
+		return NULL;
+	}
+	if (!clp->cl_cb_session && (cses->flags & SESSION4_BACK_CHAN)) {
+		struct sockaddr *sa = svc_addr(rqstp);
+
+		clp->cl_cb_session = new;
+		clp->cl_cb_conn.cb_xprt = rqstp->rq_xprt;
+		svc_xprt_get(rqstp->rq_xprt);
+		rpc_copy_addr((struct sockaddr *)&clp->cl_cb_conn.cb_addr, sa);
+		clp->cl_cb_conn.cb_addrlen = svc_addr_len(sa);
+		nfsd4_probe_callback(clp);
+	}
+	return new;
 }
 
 /* caller must hold client_lock */
@@ -731,21 +816,6 @@ unhash_session(struct nfsd4_session *ses)
 	list_del(&ses->se_perclnt);
 }
 
-void
-free_session(struct kref *kref)
-{
-	struct nfsd4_session *ses;
-	int mem;
-
-	ses = container_of(kref, struct nfsd4_session, se_ref);
-	spin_lock(&nfsd_drc_lock);
-	mem = ses->se_fchannel.maxreqs * slot_bytes(&ses->se_fchannel);
-	nfsd_drc_mem_used -= mem;
-	spin_unlock(&nfsd_drc_lock);
-	free_session_slots(ses);
-	kfree(ses);
-}
-
 /* must be called under the client_lock */
 static inline void
 renew_client_locked(struct nfs4_client *clp)
@@ -812,6 +882,13 @@ static struct nfs4_client *alloc_client(struct xdr_netobj name)
 static inline void
 free_client(struct nfs4_client *clp)
 {
+	while (!list_empty(&clp->cl_sessions)) {
+		struct nfsd4_session *ses;
+		ses = list_entry(clp->cl_sessions.next, struct nfsd4_session,
+				se_perclnt);
+		list_del(&ses->se_perclnt);
+		nfsd4_put_session(ses);
+	}
 	if (clp->cl_cred.cr_group_info)
 		put_group_info(clp->cl_cred.cr_group_info);
 	kfree(clp->cl_principal);
@@ -838,15 +915,12 @@ release_session_client(struct nfsd4_session *session)
 static inline void
 unhash_client_locked(struct nfs4_client *clp)
 {
+	struct nfsd4_session *ses;
+
 	mark_client_expired(clp);
 	list_del(&clp->cl_lru);
-	while (!list_empty(&clp->cl_sessions)) {
-		struct nfsd4_session  *ses;
-		ses = list_entry(clp->cl_sessions.next, struct nfsd4_session,
-				 se_perclnt);
-		unhash_session(ses);
-		nfsd4_put_session(ses);
-	}
+	list_for_each_entry(ses, &clp->cl_sessions, se_perclnt)
+		list_del_init(&ses->se_hash);
 }
 
 static void
@@ -875,7 +949,7 @@ expire_client(struct nfs4_client *clp)
 		sop = list_entry(clp->cl_openowners.next, struct nfs4_stateowner, so_perclient);
 		release_openowner(sop);
 	}
-	nfsd4_set_callback_client(clp, NULL);
+	nfsd4_shutdown_callback(clp);
 	if (clp->cl_cb_conn.cb_xprt)
 		svc_xprt_put(clp->cl_cb_conn.cb_xprt);
 	list_del(&clp->cl_idhash);
@@ -960,6 +1034,8 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir,
 	if (clp == NULL)
 		return NULL;
 
+	INIT_LIST_HEAD(&clp->cl_sessions);
+
 	princ = svc_gss_principal(rqstp);
 	if (princ) {
 		clp->cl_principal = kstrdup(princ, GFP_KERNEL);
@@ -976,8 +1052,9 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir,
 	INIT_LIST_HEAD(&clp->cl_strhash);
 	INIT_LIST_HEAD(&clp->cl_openowners);
 	INIT_LIST_HEAD(&clp->cl_delegations);
-	INIT_LIST_HEAD(&clp->cl_sessions);
 	INIT_LIST_HEAD(&clp->cl_lru);
+	spin_lock_init(&clp->cl_lock);
+	INIT_WORK(&clp->cl_cb_null.cb_work, nfsd4_do_callback_rpc);
 	clp->cl_time = get_seconds();
 	clear_bit(0, &clp->cl_cb_slot_busy);
 	rpc_init_wait_queue(&clp->cl_cb_waitq, "Backchannel slot table");
@@ -986,7 +1063,7 @@ static struct nfs4_client *create_client(struct xdr_netobj name, char *recdir,
 	clp->cl_flavor = rqstp->rq_flavor;
 	copy_cred(&clp->cl_cred, &rqstp->rq_cred);
 	gen_confirm(clp);
-
+	clp->cl_cb_session = NULL;
 	return clp;
 }
 
@@ -1098,7 +1175,7 @@ find_unconfirmed_client_by_str(const char *dname, unsigned int hashval,
 static void
 gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se, u32 scopeid)
 {
-	struct nfs4_cb_conn *cb = &clp->cl_cb_conn;
+	struct nfs4_cb_conn *conn = &clp->cl_cb_conn;
 	unsigned short expected_family;
 
 	/* Currently, we only support tcp and tcp6 for the callback channel */
@@ -1111,24 +1188,23 @@ gen_callback(struct nfs4_client *clp, struct nfsd4_setclientid *se, u32 scopeid)
 	else
 		goto out_err;
 
-	cb->cb_addrlen = rpc_uaddr2sockaddr(se->se_callback_addr_val,
+	conn->cb_addrlen = rpc_uaddr2sockaddr(se->se_callback_addr_val,
 					    se->se_callback_addr_len,
-					    (struct sockaddr *) &cb->cb_addr,
-					    sizeof(cb->cb_addr));
+					    (struct sockaddr *)&conn->cb_addr,
+					    sizeof(conn->cb_addr));
 
-	if (!cb->cb_addrlen || cb->cb_addr.ss_family != expected_family)
+	if (!conn->cb_addrlen || conn->cb_addr.ss_family != expected_family)
 		goto out_err;
 
-	if (cb->cb_addr.ss_family == AF_INET6)
-		((struct sockaddr_in6 *) &cb->cb_addr)->sin6_scope_id = scopeid;
+	if (conn->cb_addr.ss_family == AF_INET6)
+		((struct sockaddr_in6 *)&conn->cb_addr)->sin6_scope_id = scopeid;
 
-	cb->cb_minorversion = 0;
-	cb->cb_prog = se->se_callback_prog;
-	cb->cb_ident = se->se_callback_ident;
+	conn->cb_prog = se->se_callback_prog;
+	conn->cb_ident = se->se_callback_ident;
 	return;
 out_err:
-	cb->cb_addr.ss_family = AF_UNSPEC;
-	cb->cb_addrlen = 0;
+	conn->cb_addr.ss_family = AF_UNSPEC;
+	conn->cb_addrlen = 0;
 	dprintk(KERN_INFO "NFSD: this client (clientid %08x/%08x) "
 		"will not receive delegations\n",
 		clp->cl_clientid.cl_boot, clp->cl_clientid.cl_id);
@@ -1415,7 +1491,9 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 {
 	struct sockaddr *sa = svc_addr(rqstp);
 	struct nfs4_client *conf, *unconf;
+	struct nfsd4_session *new;
 	struct nfsd4_clid_slot *cs_slot = NULL;
+	bool confirm_me = false;
 	int status = 0;
 
 	nfs4_lock_state();
@@ -1438,7 +1516,6 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 				cs_slot->sl_seqid, cr_ses->seqid);
 			goto out;
 		}
-		cs_slot->sl_seqid++;
 	} else if (unconf) {
 		if (!same_creds(&unconf->cl_cred, &rqstp->rq_cred) ||
 		    !rpc_cmp_addr(sa, (struct sockaddr *) &unconf->cl_addr)) {
@@ -1451,25 +1528,10 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 		if (status) {
 			/* an unconfirmed replay returns misordered */
 			status = nfserr_seq_misordered;
-			goto out_cache;
+			goto out;
 		}
 
-		cs_slot->sl_seqid++; /* from 0 to 1 */
-		move_to_confirmed(unconf);
-
-		if (cr_ses->flags & SESSION4_BACK_CHAN) {
-			unconf->cl_cb_conn.cb_xprt = rqstp->rq_xprt;
-			svc_xprt_get(rqstp->rq_xprt);
-			rpc_copy_addr(
-				(struct sockaddr *)&unconf->cl_cb_conn.cb_addr,
-				sa);
-			unconf->cl_cb_conn.cb_addrlen = svc_addr_len(sa);
-			unconf->cl_cb_conn.cb_minorversion =
-				cstate->minorversion;
-			unconf->cl_cb_conn.cb_prog = cr_ses->callback_prog;
-			unconf->cl_cb_seq_nr = 1;
-			nfsd4_probe_callback(unconf, &unconf->cl_cb_conn);
-		}
+		confirm_me = true;
 		conf = unconf;
 	} else {
 		status = nfserr_stale_clientid;
@@ -1477,22 +1539,30 @@ nfsd4_create_session(struct svc_rqst *rqstp,
 	}
 
 	/*
+	 * XXX: we should probably set this at creation time, and check
+	 * for consistent minorversion use throughout:
+	 */
+	conf->cl_minorversion = 1;
+	/*
 	 * We do not support RDMA or persistent sessions
 	 */
 	cr_ses->flags &= ~SESSION4_PERSIST;
 	cr_ses->flags &= ~SESSION4_RDMA;
 
-	status = alloc_init_session(rqstp, conf, cr_ses);
-	if (status)
+	status = nfserr_jukebox;
+	new = alloc_init_session(rqstp, conf, cr_ses);
+	if (!new)
 		goto out;
-
-	memcpy(cr_ses->sessionid.data, conf->cl_sessionid.data,
+	status = nfs_ok;
+	memcpy(cr_ses->sessionid.data, new->se_sessionid.data,
 	       NFS4_MAX_SESSIONID_LEN);
+	cs_slot->sl_seqid++;
 	cr_ses->seqid = cs_slot->sl_seqid;
 
-out_cache:
 	/* cache solo and embedded create sessions under the state lock */
 	nfsd4_cache_create_session(cr_ses, cs_slot, status);
+	if (confirm_me)
+		move_to_confirmed(conf);
 out:
 	nfs4_unlock_state();
 	dprintk("%s returns %d\n", __func__, ntohl(status));
@@ -1546,8 +1616,11 @@ nfsd4_destroy_session(struct svc_rqst *r,
 
 	nfs4_lock_state();
 	/* wait for callbacks */
-	nfsd4_set_callback_client(ses->se_client, NULL);
+	nfsd4_shutdown_callback(ses->se_client);
 	nfs4_unlock_state();
+
+	nfsd4_del_conns(ses);
+
 	nfsd4_put_session(ses);
 	status = nfs_ok;
 out:
@@ -1555,6 +1628,36 @@ out:
 	return status;
 }
 
+static struct nfsd4_conn *__nfsd4_find_conn(struct svc_xprt *xpt, struct nfsd4_session *s)
+{
+	struct nfsd4_conn *c;
+
+	list_for_each_entry(c, &s->se_conns, cn_persession) {
+		if (c->cn_xprt == xpt) {
+			return c;
+		}
+	}
+	return NULL;
+}
+
+static void nfsd4_sequence_check_conn(struct nfsd4_conn *new, struct nfsd4_session *ses)
+{
+	struct nfs4_client *clp = ses->se_client;
+	struct nfsd4_conn *c;
+
+	spin_lock(&clp->cl_lock);
+	c = __nfsd4_find_conn(new->cn_xprt, ses);
+	if (c) {
+		spin_unlock(&clp->cl_lock);
+		free_conn(new);
+		return;
+	}
+	__nfsd4_hash_conn(new, ses);
+	spin_unlock(&clp->cl_lock);
+	nfsd4_register_conn(new);
+	return;
+}
+
 __be32
 nfsd4_sequence(struct svc_rqst *rqstp,
 	       struct nfsd4_compound_state *cstate,
@@ -1563,11 +1666,20 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 	struct nfsd4_compoundres *resp = rqstp->rq_resp;
 	struct nfsd4_session *session;
 	struct nfsd4_slot *slot;
+	struct nfsd4_conn *conn;
 	int status;
 
 	if (resp->opcnt != 1)
 		return nfserr_sequence_pos;
 
+	/*
+	 * Will be either used or freed by nfsd4_sequence_check_conn
+	 * below.
+	 */
+	conn = alloc_conn(rqstp, NFS4_CDFC4_FORE);
+	if (!conn)
+		return nfserr_jukebox;
+
 	spin_lock(&client_lock);
 	status = nfserr_badsession;
 	session = find_in_sessionid_hashtbl(&seq->sessionid);
@@ -1599,6 +1711,9 @@ nfsd4_sequence(struct svc_rqst *rqstp,
 	if (status)
 		goto out;
 
+	nfsd4_sequence_check_conn(conn, session);
+	conn = NULL;
+
 	/* Success! bump slot seqid */
 	slot->sl_inuse = true;
 	slot->sl_seqid = seq->seqid;
@@ -1613,6 +1728,7 @@ out:
 		nfsd4_get_session(cstate->session);
 		atomic_inc(&session->se_client->cl_refcount);
 	}
+	kfree(conn);
 	spin_unlock(&client_lock);
 	dprintk("%s: return %d\n", __func__, ntohl(status));
 	return status;
@@ -1747,6 +1863,11 @@ nfsd4_setclientid(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 			goto out;
 		gen_clid(new);
 	}
+	/*
+	 * XXX: we should probably set this at creation time, and check
+	 * for consistent minorversion use throughout:
+	 */
+	new->cl_minorversion = 0;
 	gen_callback(new, setclid, rpc_get_scope_id(sa));
 	add_to_unconfirmed(new, strhashval);
 	setclid->se_clientid.cl_boot = new->cl_clientid.cl_boot;
@@ -1807,7 +1928,8 @@ nfsd4_setclientid_confirm(struct svc_rqst *rqstp,
 			status = nfserr_clid_inuse;
 		else {
 			atomic_set(&conf->cl_cb_set, 0);
-			nfsd4_probe_callback(conf, &unconf->cl_cb_conn);
+			nfsd4_change_callback(conf, &unconf->cl_cb_conn);
+			nfsd4_probe_callback(conf);
 			expire_client(unconf);
 			status = nfs_ok;
 
@@ -1841,7 +1963,7 @@ nfsd4_setclientid_confirm(struct svc_rqst *rqstp,
 			}
 			move_to_confirmed(unconf);
 			conf = unconf;
-			nfsd4_probe_callback(conf, &conf->cl_cb_conn);
+			nfsd4_probe_callback(conf);
 			status = nfs_ok;
 		}
 	} else if ((!conf || (conf && !same_verf(&conf->cl_confirm, &confirm)))
@@ -2944,7 +3066,11 @@ nfs4_preprocess_stateid_op(struct nfsd4_compound_state *cstate,
 	if (STALE_STATEID(stateid)) 
 		goto out;
 
-	status = nfserr_bad_stateid;
+	/*
+	 * We assume that any stateid that has the current boot time,
+	 * but that we can't find, is expired:
+	 */
+	status = nfserr_expired;
 	if (is_delegation_stateid(stateid)) {
 		dp = find_delegation_stateid(ino, stateid);
 		if (!dp)
@@ -2964,6 +3090,7 @@ nfs4_preprocess_stateid_op(struct nfsd4_compound_state *cstate,
 		stp = find_stateid(stateid, flags);
 		if (!stp)
 			goto out;
+		status = nfserr_bad_stateid;
 		if (nfs4_check_fh(current_fh, stp))
 			goto out;
 		if (!stp->st_stateowner->so_confirmed)
@@ -3038,8 +3165,9 @@ nfs4_preprocess_seqid_op(struct nfsd4_compound_state *cstate, u32 seqid,
 		 * a replayed close:
 		 */
 		sop = search_close_lru(stateid->si_stateownerid, flags);
+		/* It's not stale; let's assume it's expired: */
 		if (sop == NULL)
-			return nfserr_bad_stateid;
+			return nfserr_expired;
 		*sopp = sop;
 		goto check_replay;
 	}
@@ -3304,6 +3432,7 @@ nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	status = nfserr_bad_stateid;
 	if (!is_delegation_stateid(stateid))
 		goto out;
+	status = nfserr_expired;
 	dp = find_delegation_stateid(inode, stateid);
 	if (!dp)
 		goto out;
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index 1a468bbd330f..f35a94a04026 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -1805,19 +1805,23 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp,
 				goto out_nfserr;
 		}
 	}
-	if ((buflen -= 16) < 0)
-		goto out_resource;
 
-	if (unlikely(bmval2)) {
+	if (bmval2) {
+		if ((buflen -= 16) < 0)
+			goto out_resource;
 		WRITE32(3);
 		WRITE32(bmval0);
 		WRITE32(bmval1);
 		WRITE32(bmval2);
-	} else if (likely(bmval1)) {
+	} else if (bmval1) {
+		if ((buflen -= 12) < 0)
+			goto out_resource;
 		WRITE32(2);
 		WRITE32(bmval0);
 		WRITE32(bmval1);
 	} else {
+		if ((buflen -= 8) < 0)
+			goto out_resource;
 		WRITE32(1);
 		WRITE32(bmval0);
 	}
@@ -1828,15 +1832,17 @@ nfsd4_encode_fattr(struct svc_fh *fhp, struct svc_export *exp,
 		u32 word1 = nfsd_suppattrs1(minorversion);
 		u32 word2 = nfsd_suppattrs2(minorversion);
 
-		if ((buflen -= 12) < 0)
-			goto out_resource;
 		if (!aclsupport)
 			word0 &= ~FATTR4_WORD0_ACL;
 		if (!word2) {
+			if ((buflen -= 12) < 0)
+				goto out_resource;
 			WRITE32(2);
 			WRITE32(word0);
 			WRITE32(word1);
 		} else {
+			if ((buflen -= 16) < 0)
+				goto out_resource;
 			WRITE32(3);
 			WRITE32(word0);
 			WRITE32(word1);
diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c
index 06fa87e52e82..d6dc3f61f8ba 100644
--- a/fs/nfsd/nfsctl.c
+++ b/fs/nfsd/nfsctl.c
@@ -22,6 +22,7 @@
  */
 enum {
 	NFSD_Root = 1,
+#ifdef CONFIG_NFSD_DEPRECATED
 	NFSD_Svc,
 	NFSD_Add,
 	NFSD_Del,
@@ -29,6 +30,7 @@ enum {
 	NFSD_Unexport,
 	NFSD_Getfd,
 	NFSD_Getfs,
+#endif
 	NFSD_List,
 	NFSD_Export_features,
 	NFSD_Fh,
@@ -54,6 +56,7 @@ enum {
 /*
  * write() for these nodes.
  */
+#ifdef CONFIG_NFSD_DEPRECATED
 static ssize_t write_svc(struct file *file, char *buf, size_t size);
 static ssize_t write_add(struct file *file, char *buf, size_t size);
 static ssize_t write_del(struct file *file, char *buf, size_t size);
@@ -61,6 +64,7 @@ static ssize_t write_export(struct file *file, char *buf, size_t size);
 static ssize_t write_unexport(struct file *file, char *buf, size_t size);
 static ssize_t write_getfd(struct file *file, char *buf, size_t size);
 static ssize_t write_getfs(struct file *file, char *buf, size_t size);
+#endif
 static ssize_t write_filehandle(struct file *file, char *buf, size_t size);
 static ssize_t write_unlock_ip(struct file *file, char *buf, size_t size);
 static ssize_t write_unlock_fs(struct file *file, char *buf, size_t size);
@@ -76,6 +80,7 @@ static ssize_t write_recoverydir(struct file *file, char *buf, size_t size);
 #endif
 
 static ssize_t (*write_op[])(struct file *, char *, size_t) = {
+#ifdef CONFIG_NFSD_DEPRECATED
 	[NFSD_Svc] = write_svc,
 	[NFSD_Add] = write_add,
 	[NFSD_Del] = write_del,
@@ -83,6 +88,7 @@ static ssize_t (*write_op[])(struct file *, char *, size_t) = {
 	[NFSD_Unexport] = write_unexport,
 	[NFSD_Getfd] = write_getfd,
 	[NFSD_Getfs] = write_getfs,
+#endif
 	[NFSD_Fh] = write_filehandle,
 	[NFSD_FO_UnlockIP] = write_unlock_ip,
 	[NFSD_FO_UnlockFS] = write_unlock_fs,
@@ -121,6 +127,14 @@ static ssize_t nfsctl_transaction_write(struct file *file, const char __user *bu
 
 static ssize_t nfsctl_transaction_read(struct file *file, char __user *buf, size_t size, loff_t *pos)
 {
+	static int warned;
+	if (file->f_dentry->d_name.name[0] == '.' && !warned) {
+		printk(KERN_INFO
+		       "Warning: \"%s\" uses deprecated NFSD interface: %s."
+		       "  This will be removed in 2.6.40\n",
+		       current->comm, file->f_dentry->d_name.name);
+		warned = 1;
+	}
 	if (! file->private_data) {
 		/* An attempt to read a transaction file without writing
 		 * causes a 0-byte write so that the file can return
@@ -187,6 +201,7 @@ static const struct file_operations pool_stats_operations = {
  * payload - write methods
  */
 
+#ifdef CONFIG_NFSD_DEPRECATED
 /**
  * write_svc - Start kernel's NFSD server
  *
@@ -402,7 +417,7 @@ static ssize_t write_getfs(struct file *file, char *buf, size_t size)
 
 	ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, &in6);
 
-	clp = auth_unix_lookup(&in6);
+	clp = auth_unix_lookup(&init_net, &in6);
 	if (!clp)
 		err = -EPERM;
 	else {
@@ -465,7 +480,7 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size)
 
 	ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, &in6);
 
-	clp = auth_unix_lookup(&in6);
+	clp = auth_unix_lookup(&init_net, &in6);
 	if (!clp)
 		err = -EPERM;
 	else {
@@ -482,6 +497,7 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size)
  out:
 	return err;
 }
+#endif /* CONFIG_NFSD_DEPRECATED */
 
 /**
  * write_unlock_ip - Release all locks used by a client
@@ -1000,12 +1016,12 @@ static ssize_t __write_ports_addxprt(char *buf)
 	if (err != 0)
 		return err;
 
-	err = svc_create_xprt(nfsd_serv, transport,
+	err = svc_create_xprt(nfsd_serv, transport, &init_net,
 				PF_INET, port, SVC_SOCK_ANONYMOUS);
 	if (err < 0)
 		goto out_err;
 
-	err = svc_create_xprt(nfsd_serv, transport,
+	err = svc_create_xprt(nfsd_serv, transport, &init_net,
 				PF_INET6, port, SVC_SOCK_ANONYMOUS);
 	if (err < 0 && err != -EAFNOSUPPORT)
 		goto out_close;
@@ -1356,6 +1372,7 @@ static ssize_t write_recoverydir(struct file *file, char *buf, size_t size)
 static int nfsd_fill_super(struct super_block * sb, void * data, int silent)
 {
 	static struct tree_descr nfsd_files[] = {
+#ifdef CONFIG_NFSD_DEPRECATED
 		[NFSD_Svc] = {".svc", &transaction_ops, S_IWUSR},
 		[NFSD_Add] = {".add", &transaction_ops, S_IWUSR},
 		[NFSD_Del] = {".del", &transaction_ops, S_IWUSR},
@@ -1363,6 +1380,7 @@ static int nfsd_fill_super(struct super_block * sb, void * data, int silent)
 		[NFSD_Unexport] = {".unexport", &transaction_ops, S_IWUSR},
 		[NFSD_Getfd] = {".getfd", &transaction_ops, S_IWUSR|S_IRUSR},
 		[NFSD_Getfs] = {".getfs", &transaction_ops, S_IWUSR|S_IRUSR},
+#endif
 		[NFSD_List] = {"exports", &exports_operations, S_IRUGO},
 		[NFSD_Export_features] = {"export_features",
 					&export_features_operations, S_IRUGO},
diff --git a/fs/nfsd/nfsd.h b/fs/nfsd/nfsd.h
index b76ac3a82e39..6b641cf2c19a 100644
--- a/fs/nfsd/nfsd.h
+++ b/fs/nfsd/nfsd.h
@@ -249,7 +249,7 @@ extern time_t nfsd4_grace;
 #define	COMPOUND_SLACK_SPACE		140    /* OP_GETFH */
 #define COMPOUND_ERR_SLACK_SPACE	12     /* OP_SETATTR */
 
-#define NFSD_LAUNDROMAT_MINTIMEOUT      10   /* seconds */
+#define NFSD_LAUNDROMAT_MINTIMEOUT      1   /* seconds */
 
 /*
  * The following attributes are currently not supported by the NFSv4 server:
diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c
index e2c43464f237..2bae1d86f5f2 100644
--- a/fs/nfsd/nfssvc.c
+++ b/fs/nfsd/nfssvc.c
@@ -16,6 +16,7 @@
 #include <linux/lockd/bind.h>
 #include <linux/nfsacl.h>
 #include <linux/seq_file.h>
+#include <net/net_namespace.h>
 #include "nfsd.h"
 #include "cache.h"
 #include "vfs.h"
@@ -186,12 +187,12 @@ static int nfsd_init_socks(int port)
 	if (!list_empty(&nfsd_serv->sv_permsocks))
 		return 0;
 
-	error = svc_create_xprt(nfsd_serv, "udp", PF_INET, port,
+	error = svc_create_xprt(nfsd_serv, "udp", &init_net, PF_INET, port,
 					SVC_SOCK_DEFAULTS);
 	if (error < 0)
 		return error;
 
-	error = svc_create_xprt(nfsd_serv, "tcp", PF_INET, port,
+	error = svc_create_xprt(nfsd_serv, "tcp", &init_net, PF_INET, port,
 					SVC_SOCK_DEFAULTS);
 	if (error < 0)
 		return error;
diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
index 322518c88e4b..39adc27b0685 100644
--- a/fs/nfsd/state.h
+++ b/fs/nfsd/state.h
@@ -35,6 +35,7 @@
 #ifndef _NFSD4_STATE_H
 #define _NFSD4_STATE_H
 
+#include <linux/sunrpc/svc_xprt.h>
 #include <linux/nfsd/nfsfh.h>
 #include "nfsfh.h"
 
@@ -64,19 +65,12 @@ typedef struct {
 	(s)->si_fileid, \
 	(s)->si_generation
 
-struct nfsd4_cb_sequence {
-	/* args/res */
-	u32			cbs_minorversion;
-	struct nfs4_client	*cbs_clp;
-};
-
-struct nfs4_rpc_args {
-	void				*args_op;
-	struct nfsd4_cb_sequence	args_seq;
-};
-
 struct nfsd4_callback {
-	struct nfs4_rpc_args cb_args;
+	void *cb_op;
+	struct nfs4_client *cb_clp;
+	u32 cb_minorversion;
+	struct rpc_message cb_msg;
+	const struct rpc_call_ops *cb_ops;
 	struct work_struct cb_work;
 };
 
@@ -91,7 +85,6 @@ struct nfs4_delegation {
 	u32			dl_type;
 	time_t			dl_time;
 /* For recall: */
-	u32			dl_ident;
 	stateid_t		dl_stateid;
 	struct knfsd_fh		dl_fh;
 	int			dl_retries;
@@ -103,8 +96,8 @@ struct nfs4_cb_conn {
 	/* SETCLIENTID info */
 	struct sockaddr_storage	cb_addr;
 	size_t			cb_addrlen;
-	u32                     cb_prog;
-	u32			cb_minorversion;
+	u32                     cb_prog; /* used only in 4.0 case;
+					    per-session otherwise */
 	u32                     cb_ident;	/* minorversion 0 only */
 	struct svc_xprt		*cb_xprt;	/* minorversion 1 only */
 };
@@ -160,6 +153,15 @@ struct nfsd4_clid_slot {
 	struct nfsd4_create_session	sl_cr_ses;
 };
 
+struct nfsd4_conn {
+	struct list_head cn_persession;
+	struct svc_xprt *cn_xprt;
+	struct svc_xpt_user cn_xpt_user;
+	struct nfsd4_session *cn_session;
+/* CDFC4_FORE, CDFC4_BACK: */
+	unsigned char cn_flags;
+};
+
 struct nfsd4_session {
 	struct kref		se_ref;
 	struct list_head	se_hash;	/* hash by sessionid */
@@ -169,6 +171,9 @@ struct nfsd4_session {
 	struct nfs4_sessionid	se_sessionid;
 	struct nfsd4_channel_attrs se_fchannel;
 	struct nfsd4_channel_attrs se_bchannel;
+	struct list_head	se_conns;
+	u32			se_cb_prog;
+	u32			se_cb_seq_nr;
 	struct nfsd4_slot	*se_slots[];	/* forward channel slots */
 };
 
@@ -221,24 +226,32 @@ struct nfs4_client {
 	clientid_t		cl_clientid;	/* generated by server */
 	nfs4_verifier		cl_confirm;	/* generated by server */
 	u32			cl_firststate;	/* recovery dir creation */
+	u32			cl_minorversion;
 
 	/* for v4.0 and v4.1 callbacks: */
 	struct nfs4_cb_conn	cl_cb_conn;
+#define NFSD4_CLIENT_CB_UPDATE	1
+#define NFSD4_CLIENT_KILL	2
+	unsigned long		cl_cb_flags;
 	struct rpc_clnt		*cl_cb_client;
+	u32			cl_cb_ident;
 	atomic_t		cl_cb_set;
+	struct nfsd4_callback	cl_cb_null;
+	struct nfsd4_session	*cl_cb_session;
+
+	/* for all client information that callback code might need: */
+	spinlock_t		cl_lock;
 
 	/* for nfs41 */
 	struct list_head	cl_sessions;
 	struct nfsd4_clid_slot	cl_cs_slot;	/* create_session slot */
 	u32			cl_exchange_flags;
-	struct nfs4_sessionid	cl_sessionid;
 	/* number of rpc's in progress over an associated session: */
 	atomic_t		cl_refcount;
 
 	/* for nfs41 callbacks */
 	/* We currently support a single back channel with a single slot */
 	unsigned long		cl_cb_slot_busy;
-	u32			cl_cb_seq_nr;
 	struct rpc_wait_queue	cl_cb_waitq;	/* backchannel callers may */
 						/* wait here for slots */
 };
@@ -440,12 +453,13 @@ extern int nfs4_in_grace(void);
 extern __be32 nfs4_check_open_reclaim(clientid_t *clid);
 extern void nfs4_free_stateowner(struct kref *kref);
 extern int set_callback_cred(void);
-extern void nfsd4_probe_callback(struct nfs4_client *clp, struct nfs4_cb_conn *);
+extern void nfsd4_probe_callback(struct nfs4_client *clp);
+extern void nfsd4_change_callback(struct nfs4_client *clp, struct nfs4_cb_conn *);
 extern void nfsd4_do_callback_rpc(struct work_struct *);
 extern void nfsd4_cb_recall(struct nfs4_delegation *dp);
 extern int nfsd4_create_callback_queue(void);
 extern void nfsd4_destroy_callback_queue(void);
-extern void nfsd4_set_callback_client(struct nfs4_client *, struct rpc_clnt *);
+extern void nfsd4_shutdown_callback(struct nfs4_client *);
 extern void nfs4_put_delegation(struct nfs4_delegation *dp);
 extern __be32 nfs4_make_rec_clidname(char *clidname, struct xdr_netobj *clname);
 extern void nfsd4_init_recdir(char *recdir_name);