From a0b0a6e39bd1bb4a0922086feee73627cbd53ba4 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Mon, 17 Sep 2012 17:12:15 -0400
Subject: NFS: Clean up the pNFS layoutget interface

Ensure that we do return errors from nfs4_proc_layoutget() and that we
don't mark the layout as having failed if the error was due to a
signal or resource problem on the client side.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/pnfs.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 745aa1b39e7c..d51ef888e71b 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -172,7 +172,7 @@ extern int nfs4_proc_getdevicelist(struct nfs_server *server,
 				   struct pnfs_devicelist *devlist);
 extern int nfs4_proc_getdeviceinfo(struct nfs_server *server,
 				   struct pnfs_device *dev);
-extern void nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags);
+extern struct pnfs_layout_segment* nfs4_proc_layoutget(struct nfs4_layoutget *lgp, gfp_t gfp_flags);
 extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp);
 
 /* pnfs.c */
@@ -192,7 +192,7 @@ void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *, struct nfs_page
 int pnfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc);
 bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req);
 void pnfs_set_lo_fail(struct pnfs_layout_segment *lseg);
-int pnfs_layout_process(struct nfs4_layoutget *lgp);
+struct pnfs_layout_segment *pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
 void pnfs_destroy_layout(struct nfs_inode *);
 void pnfs_destroy_all_layouts(struct nfs_client *);
-- 
cgit v1.2.3


From 49a85061b0bc9cb26361096482c81172c666c937 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Tue, 18 Sep 2012 20:43:31 -0400
Subject: NFSv4.1: Cleanup add a "pnfs_" prefix to mark_matching_lsegs_invalid

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/callback_proc.c | 4 ++--
 fs/nfs/pnfs.c          | 6 +++---
 fs/nfs/pnfs.h          | 2 +-
 3 files changed, 6 insertions(+), 6 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c
index 1b5d809a105e..57b8bda0f4e9 100644
--- a/fs/nfs/callback_proc.c
+++ b/fs/nfs/callback_proc.c
@@ -158,7 +158,7 @@ static u32 initiate_file_draining(struct nfs_client *clp,
 	ino = lo->plh_inode;
 	spin_lock(&ino->i_lock);
 	if (test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags) ||
-	    mark_matching_lsegs_invalid(lo, &free_me_list,
+	    pnfs_mark_matching_lsegs_invalid(lo, &free_me_list,
 					&args->cbl_range))
 		rv = NFS4ERR_DELAY;
 	else
@@ -211,7 +211,7 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 		ino = lo->plh_inode;
 		spin_lock(&ino->i_lock);
 		set_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags);
-		if (mark_matching_lsegs_invalid(lo, &free_me_list, &range))
+		if (pnfs_mark_matching_lsegs_invalid(lo, &free_me_list, &range))
 			rv = NFS4ERR_DELAY;
 		list_del_init(&lo->plh_bulk_recall);
 		spin_unlock(&ino->i_lock);
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 3a7ac97020df..aea2e5256fe4 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -390,7 +390,7 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg,
  * after call.
  */
 int
-mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
+pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 			    struct list_head *tmp_list,
 			    struct pnfs_layout_range *recall_range)
 {
@@ -458,7 +458,7 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
 	lo = nfsi->layout;
 	if (lo) {
 		lo->plh_block_lgets++; /* permanently block new LAYOUTGETs */
-		mark_matching_lsegs_invalid(lo, &tmp_list, NULL);
+		pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL);
 	}
 	spin_unlock(&nfsi->vfs_inode.i_lock);
 	pnfs_free_lseg_list(&tmp_list);
@@ -651,7 +651,7 @@ _pnfs_return_layout(struct inode *ino)
 	/* Reference matched in nfs4_layoutreturn_release */
 	get_layout_hdr(lo);
 	empty = list_empty(&lo->plh_segs);
-	mark_matching_lsegs_invalid(lo, &tmp_list, NULL);
+	pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL);
 	/* Don't send a LAYOUTRETURN if list was initially empty */
 	if (empty) {
 		spin_unlock(&ino->i_lock);
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index d51ef888e71b..6af518934e4c 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -203,7 +203,7 @@ void pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo,
 int pnfs_choose_layoutget_stateid(nfs4_stateid *dst,
 				  struct pnfs_layout_hdr *lo,
 				  struct nfs4_state *open_state);
-int mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
+int pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 				struct list_head *tmp_list,
 				struct pnfs_layout_range *recall_range);
 bool pnfs_roc(struct inode *ino);
-- 
cgit v1.2.3


From 70c3bd2bdf9a3c7c9282c362a4ec9ec88c713e13 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Tue, 18 Sep 2012 20:51:13 -0400
Subject: NFSv4.1: Cleanup; add "pnfs_" prefix to get_layout_hdr() and
 put_layout_hdr()

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/callback_proc.c |  8 ++++----
 fs/nfs/nfs4proc.c      |  2 +-
 fs/nfs/pnfs.c          | 30 +++++++++++++++---------------
 fs/nfs/pnfs.h          |  4 ++--
 4 files changed, 22 insertions(+), 22 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c
index 57b8bda0f4e9..24252fea2c9c 100644
--- a/fs/nfs/callback_proc.c
+++ b/fs/nfs/callback_proc.c
@@ -122,7 +122,7 @@ static struct pnfs_layout_hdr * get_layout_by_fh_locked(struct nfs_client *clp,
 			ino = igrab(lo->plh_inode);
 			if (!ino)
 				continue;
-			get_layout_hdr(lo);
+			pnfs_get_layout_hdr(lo);
 			return lo;
 		}
 	}
@@ -166,7 +166,7 @@ static u32 initiate_file_draining(struct nfs_client *clp,
 	pnfs_set_layout_stateid(lo, &args->cbl_stateid, true);
 	spin_unlock(&ino->i_lock);
 	pnfs_free_lseg_list(&free_me_list);
-	put_layout_hdr(lo);
+	pnfs_put_layout_hdr(lo);
 	iput(ino);
 	return rv;
 }
@@ -198,7 +198,7 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 		list_for_each_entry(lo, &server->layouts, plh_layouts) {
 			if (!igrab(lo->plh_inode))
 				continue;
-			get_layout_hdr(lo);
+			pnfs_get_layout_hdr(lo);
 			BUG_ON(!list_empty(&lo->plh_bulk_recall));
 			list_add(&lo->plh_bulk_recall, &recall_list);
 		}
@@ -216,7 +216,7 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 		list_del_init(&lo->plh_bulk_recall);
 		spin_unlock(&ino->i_lock);
 		pnfs_free_lseg_list(&free_me_list);
-		put_layout_hdr(lo);
+		pnfs_put_layout_hdr(lo);
 		iput(ino);
 	}
 	return rv;
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 1c8656f8745c..bdacb8c21a33 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -6375,7 +6375,7 @@ static void nfs4_layoutreturn_release(void *calldata)
 	struct nfs4_layoutreturn *lrp = calldata;
 
 	dprintk("--> %s\n", __func__);
-	put_layout_hdr(lrp->args.layout);
+	pnfs_put_layout_hdr(lrp->args.layout);
 	kfree(calldata);
 	dprintk("<-- %s\n", __func__);
 }
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index aea2e5256fe4..512c8632bf36 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -190,7 +190,7 @@ EXPORT_SYMBOL_GPL(pnfs_unregister_layoutdriver);
 
 /* Need to hold i_lock if caller does not already hold reference */
 void
-get_layout_hdr(struct pnfs_layout_hdr *lo)
+pnfs_get_layout_hdr(struct pnfs_layout_hdr *lo)
 {
 	atomic_inc(&lo->plh_refcount);
 }
@@ -221,14 +221,14 @@ destroy_layout_hdr(struct pnfs_layout_hdr *lo)
 }
 
 static void
-put_layout_hdr_locked(struct pnfs_layout_hdr *lo)
+pnfs_put_layout_hdr_locked(struct pnfs_layout_hdr *lo)
 {
 	if (atomic_dec_and_test(&lo->plh_refcount))
 		destroy_layout_hdr(lo);
 }
 
 void
-put_layout_hdr(struct pnfs_layout_hdr *lo)
+pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
 {
 	struct inode *inode = lo->plh_inode;
 
@@ -254,8 +254,8 @@ static void free_lseg(struct pnfs_layout_segment *lseg)
 	struct inode *ino = lseg->pls_layout->plh_inode;
 
 	NFS_SERVER(ino)->pnfs_curr_ld->free_lseg(lseg);
-	/* Matched by get_layout_hdr in pnfs_insert_layout */
-	put_layout_hdr(NFS_I(ino)->layout);
+	/* Matched by pnfs_get_layout_hdr in pnfs_insert_layout */
+	pnfs_put_layout_hdr(NFS_I(ino)->layout);
 }
 
 static void
@@ -268,7 +268,7 @@ put_lseg_common(struct pnfs_layout_segment *lseg)
 	if (list_empty(&lseg->pls_layout->plh_segs)) {
 		set_bit(NFS_LAYOUT_DESTROYED, &lseg->pls_layout->plh_flags);
 		/* Matched by initial refcount set in alloc_init_layout_hdr */
-		put_layout_hdr_locked(lseg->pls_layout);
+		pnfs_put_layout_hdr_locked(lseg->pls_layout);
 	}
 	rpc_wake_up(&NFS_SERVER(inode)->roc_rpcwaitq);
 }
@@ -404,7 +404,7 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 		NFS_I(lo->plh_inode)->write_io = 0;
 		NFS_I(lo->plh_inode)->read_io = 0;
 		if (!test_and_set_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags))
-			put_layout_hdr_locked(lo);
+			pnfs_put_layout_hdr_locked(lo);
 		return 0;
 	}
 	list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
@@ -649,13 +649,13 @@ _pnfs_return_layout(struct inode *ino)
 	}
 	stateid = nfsi->layout->plh_stateid;
 	/* Reference matched in nfs4_layoutreturn_release */
-	get_layout_hdr(lo);
+	pnfs_get_layout_hdr(lo);
 	empty = list_empty(&lo->plh_segs);
 	pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL);
 	/* Don't send a LAYOUTRETURN if list was initially empty */
 	if (empty) {
 		spin_unlock(&ino->i_lock);
-		put_layout_hdr(lo);
+		pnfs_put_layout_hdr(lo);
 		dprintk("NFS: %s no layout segments to return\n", __func__);
 		goto out;
 	}
@@ -672,7 +672,7 @@ _pnfs_return_layout(struct inode *ino)
 		set_bit(NFS_LAYOUT_RW_FAILED, &lo->plh_flags);
 		set_bit(NFS_LAYOUT_RO_FAILED, &lo->plh_flags);
 		pnfs_clear_layout_returned(lo);
-		put_layout_hdr(lo);
+		pnfs_put_layout_hdr(lo);
 		goto out;
 	}
 
@@ -709,7 +709,7 @@ bool pnfs_roc(struct inode *ino)
 	if (!found)
 		goto out_nolayout;
 	lo->plh_block_lgets++;
-	get_layout_hdr(lo); /* matched in pnfs_roc_release */
+	pnfs_get_layout_hdr(lo); /* matched in pnfs_roc_release */
 	spin_unlock(&ino->i_lock);
 	pnfs_free_lseg_list(&tmp_list);
 	return true;
@@ -726,7 +726,7 @@ void pnfs_roc_release(struct inode *ino)
 	spin_lock(&ino->i_lock);
 	lo = NFS_I(ino)->layout;
 	lo->plh_block_lgets--;
-	put_layout_hdr_locked(lo);
+	pnfs_put_layout_hdr_locked(lo);
 	spin_unlock(&ino->i_lock);
 }
 
@@ -819,7 +819,7 @@ pnfs_insert_layout(struct pnfs_layout_hdr *lo,
 		__func__, lseg, lseg->pls_range.iomode,
 		lseg->pls_range.offset, lseg->pls_range.length);
 out:
-	get_layout_hdr(lo);
+	pnfs_get_layout_hdr(lo);
 
 	dprintk("%s:Return\n", __func__);
 }
@@ -1058,7 +1058,7 @@ pnfs_update_layout(struct inode *ino,
 		goto out_unlock;
 	atomic_inc(&lo->plh_outstanding);
 
-	get_layout_hdr(lo);
+	pnfs_get_layout_hdr(lo);
 	if (list_empty(&lo->plh_segs))
 		first = true;
 
@@ -1091,7 +1091,7 @@ pnfs_update_layout(struct inode *ino,
 		spin_unlock(&clp->cl_lock);
 	}
 	atomic_dec(&lo->plh_outstanding);
-	put_layout_hdr(lo);
+	pnfs_put_layout_hdr(lo);
 out:
 	dprintk("%s end, state 0x%lx lseg %p\n", __func__,
 		nfsi->layout ? nfsi->layout->plh_flags : -1, lseg);
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 6af518934e4c..2af681f0a491 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -176,7 +176,7 @@ extern struct pnfs_layout_segment* nfs4_proc_layoutget(struct nfs4_layoutget *lg
 extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp);
 
 /* pnfs.c */
-void get_layout_hdr(struct pnfs_layout_hdr *lo);
+void pnfs_get_layout_hdr(struct pnfs_layout_hdr *lo);
 void put_lseg(struct pnfs_layout_segment *lseg);
 
 void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *,
@@ -196,7 +196,7 @@ struct pnfs_layout_segment *pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
 void pnfs_destroy_layout(struct nfs_inode *);
 void pnfs_destroy_all_layouts(struct nfs_client *);
-void put_layout_hdr(struct pnfs_layout_hdr *lo);
+void pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo);
 void pnfs_set_layout_stateid(struct pnfs_layout_hdr *lo,
 			     const nfs4_stateid *new,
 			     bool update_barrier);
-- 
cgit v1.2.3


From 9369a431bce1e985597eda32992960c969b27c5b Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Tue, 18 Sep 2012 20:57:08 -0400
Subject: NFSv4.1: Cleanup; add "pnfs_" prefix to put_lseg() and get_lseg()

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c | 16 ++++++++--------
 fs/nfs/nfs4proc.c       |  2 +-
 fs/nfs/pnfs.c           | 36 ++++++++++++++++++------------------
 fs/nfs/pnfs.h           |  8 ++++----
 4 files changed, 31 insertions(+), 31 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index 53f94d915bd1..77cd1151ef0d 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -453,7 +453,7 @@ static void filelayout_commit_release(void *calldata)
 	struct nfs_commit_data *data = calldata;
 
 	data->completion_ops->completion(data);
-	put_lseg(data->lseg);
+	pnfs_put_lseg(data->lseg);
 	nfs_put_client(data->ds_clp);
 	nfs_commitdata_release(data);
 }
@@ -931,7 +931,7 @@ filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 	nfs_init_cinfo(&cinfo, pgio->pg_inode, pgio->pg_dreq);
 	status = filelayout_alloc_commit_info(pgio->pg_lseg, &cinfo, GFP_NOFS);
 	if (status < 0) {
-		put_lseg(pgio->pg_lseg);
+		pnfs_put_lseg(pgio->pg_lseg);
 		pgio->pg_lseg = NULL;
 		goto out_mds;
 	}
@@ -985,7 +985,7 @@ filelayout_clear_request_commit(struct nfs_page *req,
 out:
 	nfs_request_remove_commit_list(req, cinfo);
 	spin_unlock(cinfo->lock);
-	put_lseg(freeme);
+	pnfs_put_lseg(freeme);
 }
 
 static struct list_head *
@@ -1018,7 +1018,7 @@ filelayout_choose_commit_list(struct nfs_page *req,
 		 * off due to a rewrite, in which case it will be done in
 		 * filelayout_clear_request_commit
 		 */
-		buckets[i].wlseg = get_lseg(lseg);
+		buckets[i].wlseg = pnfs_get_lseg(lseg);
 	}
 	set_bit(PG_COMMIT_TO_DS, &req->wb_flags);
 	cinfo->ds->nwritten++;
@@ -1128,7 +1128,7 @@ filelayout_scan_ds_commit_list(struct pnfs_commit_bucket *bucket,
 		if (list_empty(src))
 			bucket->wlseg = NULL;
 		else
-			get_lseg(bucket->clseg);
+			pnfs_get_lseg(bucket->clseg);
 	}
 	return ret;
 }
@@ -1159,12 +1159,12 @@ static void filelayout_recover_commit_reqs(struct list_head *dst,
 
 	/* NOTE cinfo->lock is NOT held, relying on fact that this is
 	 * only called on single thread per dreq.
-	 * Can't take the lock because need to do put_lseg
+	 * Can't take the lock because need to do pnfs_put_lseg
 	 */
 	for (i = 0, b = cinfo->ds->buckets; i < cinfo->ds->nbuckets; i++, b++) {
 		if (transfer_commit_list(&b->written, dst, cinfo, 0)) {
 			BUG_ON(!list_empty(&b->written));
-			put_lseg(b->wlseg);
+			pnfs_put_lseg(b->wlseg);
 			b->wlseg = NULL;
 		}
 	}
@@ -1200,7 +1200,7 @@ alloc_ds_commits(struct nfs_commit_info *cinfo, struct list_head *list)
 		if (list_empty(&bucket->committing))
 			continue;
 		nfs_retry_commit(&bucket->committing, bucket->clseg, cinfo);
-		put_lseg(bucket->clseg);
+		pnfs_put_lseg(bucket->clseg);
 		bucket->clseg = NULL;
 	}
 	/* Caller will clean up entries put on list */
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index bdacb8c21a33..e605d417a006 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -6549,7 +6549,7 @@ static void nfs4_layoutcommit_release(void *calldata)
 		list_del_init(&lseg->pls_lc_list);
 		if (test_and_clear_bit(NFS_LSEG_LAYOUTCOMMIT,
 				       &lseg->pls_flags))
-			put_lseg(lseg);
+			pnfs_put_lseg(lseg);
 	}
 
 	clear_bit_unlock(NFS_INO_LAYOUTCOMMITTING, bitlock);
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 512c8632bf36..498af8779959 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -259,7 +259,7 @@ static void free_lseg(struct pnfs_layout_segment *lseg)
 }
 
 static void
-put_lseg_common(struct pnfs_layout_segment *lseg)
+pnfs_put_lseg_common(struct pnfs_layout_segment *lseg)
 {
 	struct inode *inode = lseg->pls_layout->plh_inode;
 
@@ -274,7 +274,7 @@ put_lseg_common(struct pnfs_layout_segment *lseg)
 }
 
 void
-put_lseg(struct pnfs_layout_segment *lseg)
+pnfs_put_lseg(struct pnfs_layout_segment *lseg)
 {
 	struct inode *inode;
 
@@ -288,13 +288,13 @@ put_lseg(struct pnfs_layout_segment *lseg)
 	if (atomic_dec_and_lock(&lseg->pls_refcount, &inode->i_lock)) {
 		LIST_HEAD(free_me);
 
-		put_lseg_common(lseg);
+		pnfs_put_lseg_common(lseg);
 		list_add(&lseg->pls_list, &free_me);
 		spin_unlock(&inode->i_lock);
 		pnfs_free_lseg_list(&free_me);
 	}
 }
-EXPORT_SYMBOL_GPL(put_lseg);
+EXPORT_SYMBOL_GPL(pnfs_put_lseg);
 
 static inline u64
 end_offset(u64 start, u64 len)
@@ -378,7 +378,7 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg,
 		dprintk("%s: lseg %p ref %d\n", __func__, lseg,
 			atomic_read(&lseg->pls_refcount));
 		if (atomic_dec_and_test(&lseg->pls_refcount)) {
-			put_lseg_common(lseg);
+			pnfs_put_lseg_common(lseg);
 			list_add(&lseg->pls_list, tmp_list);
 			rv = 1;
 		}
@@ -914,7 +914,7 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo,
 	list_for_each_entry(lseg, &lo->plh_segs, pls_list) {
 		if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags) &&
 		    is_matching_lseg(&lseg->pls_range, range)) {
-			ret = get_lseg(lseg);
+			ret = pnfs_get_lseg(lseg);
 			break;
 		}
 		if (lseg->pls_range.offset > range->offset)
@@ -1135,7 +1135,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp)
 	}
 	init_lseg(lo, lseg);
 	lseg->pls_range = res->range;
-	get_lseg(lseg);
+	pnfs_get_lseg(lseg);
 	pnfs_insert_layout(lo, lseg);
 
 	if (res->return_on_close) {
@@ -1369,12 +1369,12 @@ pnfs_do_multiple_writes(struct nfs_pageio_descriptor *desc, struct list_head *he
 		if (trypnfs == PNFS_NOT_ATTEMPTED)
 			pnfs_write_through_mds(desc, data);
 	}
-	put_lseg(lseg);
+	pnfs_put_lseg(lseg);
 }
 
 static void pnfs_writehdr_free(struct nfs_pgio_header *hdr)
 {
-	put_lseg(hdr->lseg);
+	pnfs_put_lseg(hdr->lseg);
 	nfs_writehdr_free(hdr);
 }
 EXPORT_SYMBOL_GPL(pnfs_writehdr_free);
@@ -1389,17 +1389,17 @@ pnfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc)
 	whdr = nfs_writehdr_alloc();
 	if (!whdr) {
 		desc->pg_completion_ops->error_cleanup(&desc->pg_list);
-		put_lseg(desc->pg_lseg);
+		pnfs_put_lseg(desc->pg_lseg);
 		desc->pg_lseg = NULL;
 		return -ENOMEM;
 	}
 	hdr = &whdr->header;
 	nfs_pgheader_init(desc, hdr, pnfs_writehdr_free);
-	hdr->lseg = get_lseg(desc->pg_lseg);
+	hdr->lseg = pnfs_get_lseg(desc->pg_lseg);
 	atomic_inc(&hdr->refcnt);
 	ret = nfs_generic_flush(desc, hdr);
 	if (ret != 0) {
-		put_lseg(desc->pg_lseg);
+		pnfs_put_lseg(desc->pg_lseg);
 		desc->pg_lseg = NULL;
 	} else
 		pnfs_do_multiple_writes(desc, &hdr->rpc_list, desc->pg_ioflags);
@@ -1524,12 +1524,12 @@ pnfs_do_multiple_reads(struct nfs_pageio_descriptor *desc, struct list_head *hea
 		if (trypnfs == PNFS_NOT_ATTEMPTED)
 			pnfs_read_through_mds(desc, data);
 	}
-	put_lseg(lseg);
+	pnfs_put_lseg(lseg);
 }
 
 static void pnfs_readhdr_free(struct nfs_pgio_header *hdr)
 {
-	put_lseg(hdr->lseg);
+	pnfs_put_lseg(hdr->lseg);
 	nfs_readhdr_free(hdr);
 }
 EXPORT_SYMBOL_GPL(pnfs_readhdr_free);
@@ -1545,17 +1545,17 @@ pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc)
 	if (!rhdr) {
 		desc->pg_completion_ops->error_cleanup(&desc->pg_list);
 		ret = -ENOMEM;
-		put_lseg(desc->pg_lseg);
+		pnfs_put_lseg(desc->pg_lseg);
 		desc->pg_lseg = NULL;
 		return ret;
 	}
 	hdr = &rhdr->header;
 	nfs_pgheader_init(desc, hdr, pnfs_readhdr_free);
-	hdr->lseg = get_lseg(desc->pg_lseg);
+	hdr->lseg = pnfs_get_lseg(desc->pg_lseg);
 	atomic_inc(&hdr->refcnt);
 	ret = nfs_generic_pagein(desc, hdr);
 	if (ret != 0) {
-		put_lseg(desc->pg_lseg);
+		pnfs_put_lseg(desc->pg_lseg);
 		desc->pg_lseg = NULL;
 	} else
 		pnfs_do_multiple_reads(desc, &hdr->rpc_list);
@@ -1608,7 +1608,7 @@ pnfs_set_layoutcommit(struct nfs_write_data *wdata)
 	}
 	if (!test_and_set_bit(NFS_LSEG_LAYOUTCOMMIT, &hdr->lseg->pls_flags)) {
 		/* references matched in nfs4_layoutcommit_release */
-		get_lseg(hdr->lseg);
+		pnfs_get_lseg(hdr->lseg);
 	}
 	if (end_pos > nfsi->layout->plh_lwb)
 		nfsi->layout->plh_lwb = end_pos;
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 2af681f0a491..04958797fad0 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -177,7 +177,7 @@ extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp);
 
 /* pnfs.c */
 void pnfs_get_layout_hdr(struct pnfs_layout_hdr *lo);
-void put_lseg(struct pnfs_layout_segment *lseg);
+void pnfs_put_lseg(struct pnfs_layout_segment *lseg);
 
 void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *,
 			   const struct nfs_pgio_completion_ops *);
@@ -281,7 +281,7 @@ static inline int lo_fail_bit(u32 iomode)
 }
 
 static inline struct pnfs_layout_segment *
-get_lseg(struct pnfs_layout_segment *lseg)
+pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
 	if (lseg) {
 		atomic_inc(&lseg->pls_refcount);
@@ -406,12 +406,12 @@ static inline void pnfs_destroy_layout(struct nfs_inode *nfsi)
 }
 
 static inline struct pnfs_layout_segment *
-get_lseg(struct pnfs_layout_segment *lseg)
+pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
 	return NULL;
 }
 
-static inline void put_lseg(struct pnfs_layout_segment *lseg)
+static inline void pnfs_put_lseg(struct pnfs_layout_segment *lseg)
 {
 }
 
-- 
cgit v1.2.3


From b9e028fd89d6834558aa2a5bb30e5cff5c6c1059 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Tue, 18 Sep 2012 16:41:18 -0400
Subject: NFSv4.1: Add helpers for setting/reading the I/O fail bit

...and make them local to the pnfs.c file.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/pnfs.c | 38 ++++++++++++++++++++++++++------------
 fs/nfs/pnfs.h |  6 ------
 2 files changed, 26 insertions(+), 18 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index df45acaf91f7..f46f9bc4f767 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -238,6 +238,27 @@ pnfs_put_layout_hdr(struct pnfs_layout_hdr *lo)
 	}
 }
 
+static int
+pnfs_iomode_to_fail_bit(u32 iomode)
+{
+	return iomode == IOMODE_RW ?
+		NFS_LAYOUT_RW_FAILED : NFS_LAYOUT_RO_FAILED;
+}
+
+static void
+pnfs_layout_io_set_failed(struct pnfs_layout_hdr *lo, u32 iomode)
+{
+	set_bit(pnfs_iomode_to_fail_bit(iomode), &lo->plh_flags);
+	dprintk("%s Setting layout IOMODE_%s fail bit\n", __func__,
+			iomode == IOMODE_RW ?  "RW" : "READ");
+}
+
+static bool
+pnfs_layout_io_test_failed(struct pnfs_layout_hdr *lo, u32 iomode)
+{
+	return test_bit(pnfs_iomode_to_fail_bit(iomode), &lo->plh_flags) != 0;
+}
+
 static void
 init_lseg(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg)
 {
@@ -612,7 +633,7 @@ send_layoutget(struct pnfs_layout_hdr *lo,
 			break;
 		default:
 			/* remember that LAYOUTGET failed and suspend trying */
-			set_bit(lo_fail_bit(range->iomode), &lo->plh_flags);
+			pnfs_layout_io_set_failed(lo, range->iomode);
 		}
 		return NULL;
 	}
@@ -669,8 +690,8 @@ _pnfs_return_layout(struct inode *ino)
 	lrp = kzalloc(sizeof(*lrp), GFP_KERNEL);
 	if (unlikely(lrp == NULL)) {
 		status = -ENOMEM;
-		set_bit(NFS_LAYOUT_RW_FAILED, &lo->plh_flags);
-		set_bit(NFS_LAYOUT_RO_FAILED, &lo->plh_flags);
+		pnfs_layout_io_set_failed(lo, IOMODE_RW);
+		pnfs_layout_io_set_failed(lo, IOMODE_READ);
 		pnfs_clear_layout_returned(lo);
 		pnfs_put_layout_hdr(lo);
 		goto out;
@@ -1019,7 +1040,6 @@ pnfs_update_layout(struct inode *ino,
 		.length = count,
 	};
 	unsigned pg_offset;
-	struct nfs_inode *nfsi = NFS_I(ino);
 	struct nfs_server *server = NFS_SERVER(ino);
 	struct nfs_client *clp = server->nfs_client;
 	struct pnfs_layout_hdr *lo;
@@ -1044,7 +1064,7 @@ pnfs_update_layout(struct inode *ino,
 	}
 
 	/* if LAYOUTGET already failed once we don't try again */
-	if (test_bit(lo_fail_bit(iomode), &nfsi->layout->plh_flags))
+	if (pnfs_layout_io_test_failed(lo, iomode))
 		goto out_unlock;
 
 	/* Check to see if the layout for the given range already exists */
@@ -1585,13 +1605,7 @@ static void pnfs_list_write_lseg(struct inode *inode, struct list_head *listp)
 
 void pnfs_set_lo_fail(struct pnfs_layout_segment *lseg)
 {
-	if (lseg->pls_range.iomode == IOMODE_RW) {
-		dprintk("%s Setting layout IOMODE_RW fail bit\n", __func__);
-		set_bit(lo_fail_bit(IOMODE_RW), &lseg->pls_layout->plh_flags);
-	} else {
-		dprintk("%s Setting layout IOMODE_READ fail bit\n", __func__);
-		set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
-	}
+	pnfs_layout_io_set_failed(lseg->pls_layout, lseg->pls_range.iomode);
 }
 EXPORT_SYMBOL_GPL(pnfs_set_lo_fail);
 
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 04958797fad0..e3eb7d1b17a8 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -274,12 +274,6 @@ pnfs_test_layout_returned(struct pnfs_layout_hdr *lo)
 	return test_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags);
 }
 
-static inline int lo_fail_bit(u32 iomode)
-{
-	return iomode == IOMODE_RW ?
-			 NFS_LAYOUT_RW_FAILED : NFS_LAYOUT_RO_FAILED;
-}
-
 static inline struct pnfs_layout_segment *
 pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
-- 
cgit v1.2.3


From 25c7533357a4c4a9311d40cc92e9648c8a7e763e Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Tue, 18 Sep 2012 17:01:12 -0400
Subject: NFSv4.1: Retry pNFS after a 2 minute timeout

If we had to fall back to read/write through MDS, then assume that we should
retry pNFS after a suitable timeout period.
The following patch sets a timeout of 2 minutes.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/pnfs.c | 15 ++++++++++++++-
 fs/nfs/pnfs.h |  1 +
 2 files changed, 15 insertions(+), 1 deletion(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index f46f9bc4f767..2c59da5511db 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -35,6 +35,7 @@
 #include "iostat.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PNFS
+#define PNFS_LAYOUTGET_RETRY_TIMEOUT (120*HZ)
 
 /* Locking:
  *
@@ -248,6 +249,7 @@ pnfs_iomode_to_fail_bit(u32 iomode)
 static void
 pnfs_layout_io_set_failed(struct pnfs_layout_hdr *lo, u32 iomode)
 {
+	lo->plh_retry_timestamp = jiffies;
 	set_bit(pnfs_iomode_to_fail_bit(iomode), &lo->plh_flags);
 	dprintk("%s Setting layout IOMODE_%s fail bit\n", __func__,
 			iomode == IOMODE_RW ?  "RW" : "READ");
@@ -256,7 +258,18 @@ pnfs_layout_io_set_failed(struct pnfs_layout_hdr *lo, u32 iomode)
 static bool
 pnfs_layout_io_test_failed(struct pnfs_layout_hdr *lo, u32 iomode)
 {
-	return test_bit(pnfs_iomode_to_fail_bit(iomode), &lo->plh_flags) != 0;
+	unsigned long start, end;
+	if (test_bit(pnfs_iomode_to_fail_bit(iomode), &lo->plh_flags) == 0)
+		return false;
+	end = jiffies;
+	start = end - PNFS_LAYOUTGET_RETRY_TIMEOUT;
+	if (!time_in_range(lo->plh_retry_timestamp, start, end)) {
+		/* It is time to retry the failed layoutgets */
+		clear_bit(NFS_LAYOUT_RW_FAILED, &lo->plh_flags);
+		clear_bit(NFS_LAYOUT_RO_FAILED, &lo->plh_flags);
+		return false;
+	}
+	return true;
 }
 
 static void
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index e3eb7d1b17a8..bc8e5001203d 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -140,6 +140,7 @@ struct pnfs_layout_hdr {
 	atomic_t		plh_outstanding; /* number of RPCs out */
 	unsigned long		plh_block_lgets; /* block LAYOUTGET if >0 */
 	u32			plh_barrier; /* ignore lower seqids */
+	unsigned long		plh_retry_timestamp;
 	unsigned long		plh_flags;
 	loff_t			plh_lwb; /* last write byte for layoutcommit */
 	struct rpc_cred		*plh_lc_cred; /* layoutcommit cred */
-- 
cgit v1.2.3


From 1dfed2737d8cfe2f2378fddfb3bed126ff5474e7 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Tue, 18 Sep 2012 19:51:12 -0400
Subject: NFSv4.1: pNFS data servers may be temporarily offline

In cases where the pNFS data server is just temporarily out of service,
we want to mark it as such, and then try again later. Typically that will
be in cases of network connection errors etc.
This patch allows us to mark the devices as being "unavailable" for such
transient errors, and will make them available for retries after a
2 minute timeout period.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c    | 22 +++++++++++++++++++---
 fs/nfs/nfs4filelayout.h    |  8 ++------
 fs/nfs/nfs4filelayoutdev.c | 15 +++++++--------
 fs/nfs/pnfs.h              |  4 ++++
 fs/nfs/pnfs_dev.c          | 27 +++++++++++++++++++++++++++
 5 files changed, 59 insertions(+), 17 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index af6ee4ad3f15..dac2162c3ac4 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -205,7 +205,7 @@ static int filelayout_async_handle_error(struct rpc_task *task,
 	case -EPIPE:
 		dprintk("%s DS connection error %d\n", __func__,
 			task->tk_status);
-		filelayout_mark_devid_invalid(devid);
+		nfs4_mark_deviceid_unavailable(devid);
 		clear_bit(NFS_INO_LAYOUTCOMMIT, &NFS_I(inode)->flags);
 		_pnfs_return_layout(inode);
 		rpc_wake_up(&tbl->slot_tbl_waitq);
@@ -269,6 +269,22 @@ filelayout_set_layoutcommit(struct nfs_write_data *wdata)
 		(unsigned long) NFS_I(hdr->inode)->layout->plh_lwb);
 }
 
+bool
+filelayout_test_devid_unavailable(struct nfs4_deviceid_node *node)
+{
+	return filelayout_test_devid_invalid(node) ||
+		nfs4_test_deviceid_unavailable(node);
+}
+
+static bool
+filelayout_reset_to_mds(struct pnfs_layout_segment *lseg)
+{
+	struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg);
+
+	return filelayout_test_layout_invalid(lseg->pls_layout) ||
+		filelayout_test_devid_unavailable(node);
+}
+
 /*
  * Call ops for the async read/write cases
  * In the case of dense layouts, the offset needs to be reset to its
@@ -613,8 +629,8 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 			goto out;
 	} else
 		dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node);
-	/* Found deviceid is being reaped */
-	if (test_bit(NFS_DEVICEID_INVALID, &dsaddr->id_node.flags))
+	/* Found deviceid is unavailable */
+	if (filelayout_test_devid_unavailable(&dsaddr->id_node))
 			goto out_put;
 
 	fl->dsaddr = dsaddr;
diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h
index 11053c425a61..10b0f134400b 100644
--- a/fs/nfs/nfs4filelayout.h
+++ b/fs/nfs/nfs4filelayout.h
@@ -140,12 +140,8 @@ filelayout_test_devid_invalid(struct nfs4_deviceid_node *node)
 	return test_bit(NFS_DEVICEID_INVALID, &node->flags);
 }
 
-static inline bool
-filelayout_reset_to_mds(struct pnfs_layout_segment *lseg)
-{
-	return filelayout_test_devid_invalid(FILELAYOUT_DEVID_NODE(lseg)) ||
-		filelayout_test_layout_invalid(lseg->pls_layout);
-}
+extern bool
+filelayout_test_devid_unavailable(struct nfs4_deviceid_node *node);
 
 extern struct nfs_fh *
 nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j);
diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c
index b85a29df20ae..3336d5eaf879 100644
--- a/fs/nfs/nfs4filelayoutdev.c
+++ b/fs/nfs/nfs4filelayoutdev.c
@@ -804,13 +804,14 @@ nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx)
 	struct nfs4_pnfs_ds *ds = dsaddr->ds_list[ds_idx];
 	struct nfs4_deviceid_node *devid = FILELAYOUT_DEVID_NODE(lseg);
 
-	if (filelayout_test_devid_invalid(devid))
+	if (filelayout_test_devid_unavailable(devid))
 		return NULL;
 
 	if (ds == NULL) {
 		printk(KERN_ERR "NFS: %s: No data server for offset index %d\n",
 			__func__, ds_idx);
-		goto mark_dev_invalid;
+		filelayout_mark_devid_invalid(devid);
+		return NULL;
 	}
 
 	if (!ds->ds_clp) {
@@ -818,14 +819,12 @@ nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx)
 		int err;
 
 		err = nfs4_ds_connect(s, ds);
-		if (err)
-			goto mark_dev_invalid;
+		if (err) {
+			nfs4_mark_deviceid_unavailable(devid);
+			return NULL;
+		}
 	}
 	return ds;
-
-mark_dev_invalid:
-	filelayout_mark_devid_invalid(devid);
-	return NULL;
 }
 
 module_param(dataserver_retrans, uint, 0644);
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index bc8e5001203d..9735031e1e1a 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -234,6 +234,7 @@ struct nfs4_threshold *pnfs_mdsthreshold_alloc(void);
 /* nfs4_deviceid_flags */
 enum {
 	NFS_DEVICEID_INVALID = 0,       /* set when MDS clientid recalled */
+	NFS_DEVICEID_UNAVAILABLE,	/* device temporarily unavailable */
 };
 
 /* pnfs_dev.c */
@@ -243,6 +244,7 @@ struct nfs4_deviceid_node {
 	const struct pnfs_layoutdriver_type *ld;
 	const struct nfs_client		*nfs_client;
 	unsigned long 			flags;
+	unsigned long			timestamp_unavailable;
 	struct nfs4_deviceid		deviceid;
 	atomic_t			ref;
 };
@@ -255,6 +257,8 @@ void nfs4_init_deviceid_node(struct nfs4_deviceid_node *,
 			     const struct nfs4_deviceid *);
 struct nfs4_deviceid_node *nfs4_insert_deviceid_node(struct nfs4_deviceid_node *);
 bool nfs4_put_deviceid_node(struct nfs4_deviceid_node *);
+void nfs4_mark_deviceid_unavailable(struct nfs4_deviceid_node *node);
+bool nfs4_test_deviceid_unavailable(struct nfs4_deviceid_node *node);
 void nfs4_deviceid_purge_client(const struct nfs_client *);
 
 static inline void
diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c
index 73f701f1f4d3..d35b62e83ea6 100644
--- a/fs/nfs/pnfs_dev.c
+++ b/fs/nfs/pnfs_dev.c
@@ -40,6 +40,8 @@
 #define NFS4_DEVICE_ID_HASH_SIZE	(1 << NFS4_DEVICE_ID_HASH_BITS)
 #define NFS4_DEVICE_ID_HASH_MASK	(NFS4_DEVICE_ID_HASH_SIZE - 1)
 
+#define PNFS_DEVICE_RETRY_TIMEOUT (120*HZ)
+
 static struct hlist_head nfs4_deviceid_cache[NFS4_DEVICE_ID_HASH_SIZE];
 static DEFINE_SPINLOCK(nfs4_deviceid_lock);
 
@@ -218,6 +220,30 @@ nfs4_put_deviceid_node(struct nfs4_deviceid_node *d)
 }
 EXPORT_SYMBOL_GPL(nfs4_put_deviceid_node);
 
+void
+nfs4_mark_deviceid_unavailable(struct nfs4_deviceid_node *node)
+{
+	node->timestamp_unavailable = jiffies;
+	set_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags);
+}
+EXPORT_SYMBOL_GPL(nfs4_mark_deviceid_unavailable);
+
+bool
+nfs4_test_deviceid_unavailable(struct nfs4_deviceid_node *node)
+{
+	if (test_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags)) {
+		unsigned long start, end;
+
+		end = jiffies;
+		start = end - PNFS_DEVICE_RETRY_TIMEOUT;
+		if (time_in_range(node->timestamp_unavailable, start, end))
+			return true;
+		clear_bit(NFS_DEVICEID_UNAVAILABLE, &node->flags);
+	}
+	return false;
+}
+EXPORT_SYMBOL_GPL(nfs4_test_deviceid_unavailable);
+
 static void
 _deviceid_purge_client(const struct nfs_client *clp, long hash)
 {
@@ -276,3 +302,4 @@ nfs4_deviceid_mark_client_invalid(struct nfs_client *clp)
 	}
 	rcu_read_unlock();
 }
+
-- 
cgit v1.2.3


From 7fdab069b7172f2348cf3d87e19c6c24340292bf Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Thu, 20 Sep 2012 20:15:57 -0400
Subject: NFSv4.1: Fix a race in the pNFS return-on-close code

If we sleep after dropping the inode->i_lock, then we are no longer
atomic with respect to the rpc_wake_up() call in pnfs_layout_remove_lseg().

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4proc.c |  8 +++-----
 fs/nfs/pnfs.c     | 22 ++++++++++++----------
 fs/nfs/pnfs.h     |  4 ++--
 3 files changed, 17 insertions(+), 17 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index e605d417a006..6d5750cabd8d 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -2137,6 +2137,7 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data)
 {
 	struct nfs4_closedata *calldata = data;
 	struct nfs4_state *state = calldata->state;
+	struct inode *inode = calldata->inode;
 	int call_close = 0;
 
 	dprintk("%s: begin!\n", __func__);
@@ -2170,16 +2171,13 @@ static void nfs4_close_prepare(struct rpc_task *task, void *data)
 	if (calldata->arg.fmode == 0) {
 		task->tk_msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE];
 		if (calldata->roc &&
-		    pnfs_roc_drain(calldata->inode, &calldata->roc_barrier)) {
-			rpc_sleep_on(&NFS_SERVER(calldata->inode)->roc_rpcwaitq,
-				     task, NULL);
+		    pnfs_roc_drain(inode, &calldata->roc_barrier, task))
 			goto out;
-		}
 	}
 
 	nfs_fattr_init(calldata->res.fattr);
 	calldata->timestamp = jiffies;
-	if (nfs4_setup_sequence(NFS_SERVER(calldata->inode),
+	if (nfs4_setup_sequence(NFS_SERVER(inode),
 				&calldata->arg.seq_args,
 				&calldata->res.seq_res,
 				task))
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 9ee3bd705b94..8b32f8745337 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -807,27 +807,29 @@ void pnfs_roc_set_barrier(struct inode *ino, u32 barrier)
 	spin_unlock(&ino->i_lock);
 }
 
-bool pnfs_roc_drain(struct inode *ino, u32 *barrier)
+bool pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task)
 {
 	struct nfs_inode *nfsi = NFS_I(ino);
+	struct pnfs_layout_hdr *lo;
 	struct pnfs_layout_segment *lseg;
+	u32 current_seqid;
 	bool found = false;
 
 	spin_lock(&ino->i_lock);
 	list_for_each_entry(lseg, &nfsi->layout->plh_segs, pls_list)
 		if (test_bit(NFS_LSEG_ROC, &lseg->pls_flags)) {
+			rpc_sleep_on(&NFS_SERVER(ino)->roc_rpcwaitq, task, NULL);
 			found = true;
-			break;
+			goto out;
 		}
-	if (!found) {
-		struct pnfs_layout_hdr *lo = nfsi->layout;
-		u32 current_seqid = be32_to_cpu(lo->plh_stateid.seqid);
+	lo = nfsi->layout;
+	current_seqid = be32_to_cpu(lo->plh_stateid.seqid);
 
-		/* Since close does not return a layout stateid for use as
-		 * a barrier, we choose the worst-case barrier.
-		 */
-		*barrier = current_seqid + atomic_read(&lo->plh_outstanding);
-	}
+	/* Since close does not return a layout stateid for use as
+	 * a barrier, we choose the worst-case barrier.
+	 */
+	*barrier = current_seqid + atomic_read(&lo->plh_outstanding);
+out:
 	spin_unlock(&ino->i_lock);
 	return found;
 }
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 9735031e1e1a..aa9fa1b1ff4a 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -210,7 +210,7 @@ int pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 bool pnfs_roc(struct inode *ino);
 void pnfs_roc_release(struct inode *ino);
 void pnfs_roc_set_barrier(struct inode *ino, u32 barrier);
-bool pnfs_roc_drain(struct inode *ino, u32 *barrier);
+bool pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task);
 void pnfs_set_layoutcommit(struct nfs_write_data *wdata);
 void pnfs_cleanup_layoutcommit(struct nfs4_layoutcommit_data *data);
 int pnfs_layoutcommit_inode(struct inode *inode, bool sync);
@@ -442,7 +442,7 @@ pnfs_roc_set_barrier(struct inode *ino, u32 barrier)
 }
 
 static inline bool
-pnfs_roc_drain(struct inode *ino, u32 *barrier)
+pnfs_roc_drain(struct inode *ino, u32 *barrier, struct rpc_task *task)
 {
 	return false;
 }
-- 
cgit v1.2.3


From 965938b83b19aeffdc1d16ce9947c8c127e8f59b Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Thu, 20 Sep 2012 15:07:45 -0400
Subject: NFSv4.1: Get rid of pNFS layout state "NFS_LAYOUT_INVALID"

In all cases where we set NFS_LAYOUT_INVALID, we also set NFS_LAYOUT_DESTROYED.
Furthermore, in all cases where we test for NFS_LAYOUT_INVALID, we should
also be testing for NFS_LAYOUT_DESTROYED, since the latter means that
we hold no valid layout segments.
Ergo the two are redundant.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c | 4 +---
 fs/nfs/nfs4filelayout.h | 6 ------
 fs/nfs/pnfs.h           | 7 ++++++-
 3 files changed, 7 insertions(+), 10 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index dac2162c3ac4..6cce57e7fe55 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -190,8 +190,6 @@ static int filelayout_async_handle_error(struct rpc_task *task,
 		 * i/o and all i/o waiting on the slot table to the MDS until
 		 * layout is destroyed and a new valid layout is obtained.
 		 */
-		set_bit(NFS_LAYOUT_INVALID,
-				&NFS_I(inode)->layout->plh_flags);
 		pnfs_destroy_layout(NFS_I(inode));
 		rpc_wake_up(&tbl->slot_tbl_waitq);
 		goto reset;
@@ -281,7 +279,7 @@ filelayout_reset_to_mds(struct pnfs_layout_segment *lseg)
 {
 	struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg);
 
-	return filelayout_test_layout_invalid(lseg->pls_layout) ||
+	return pnfs_test_layout_destroyed(lseg->pls_layout) ||
 		filelayout_test_devid_unavailable(node);
 }
 
diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h
index 10b0f134400b..dca47d786710 100644
--- a/fs/nfs/nfs4filelayout.h
+++ b/fs/nfs/nfs4filelayout.h
@@ -128,12 +128,6 @@ filelayout_mark_devid_invalid(struct nfs4_deviceid_node *node)
 	set_bit(NFS_DEVICEID_INVALID, &node->flags);
 }
 
-static inline bool
-filelayout_test_layout_invalid(struct pnfs_layout_hdr *lo)
-{
-	return test_bit(NFS_LAYOUT_INVALID, &lo->plh_flags);
-}
-
 static inline bool
 filelayout_test_devid_invalid(struct nfs4_deviceid_node *node)
 {
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index aa9fa1b1ff4a..aacda7fbb536 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -63,7 +63,6 @@ enum {
 	NFS_LAYOUT_BULK_RECALL,		/* bulk recall affecting layout */
 	NFS_LAYOUT_ROC,			/* some lseg had roc bit set */
 	NFS_LAYOUT_DESTROYED,		/* no new use of layout allowed */
-	NFS_LAYOUT_INVALID,		/* layout is being destroyed */
 	NFS_LAYOUT_RETURNED,		/* layout has already been returned */
 };
 
@@ -279,6 +278,12 @@ pnfs_test_layout_returned(struct pnfs_layout_hdr *lo)
 	return test_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags);
 }
 
+static inline bool
+pnfs_test_layout_destroyed(struct pnfs_layout_hdr *lo)
+{
+	return test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags);
+}
+
 static inline struct pnfs_layout_segment *
 pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
-- 
cgit v1.2.3


From 8006bfba36d42b6976ed92979f51e5f9bef2625c Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Fri, 21 Sep 2012 14:48:04 -0400
Subject: NFSv4.1: Get rid of the NFS_LAYOUT_DESTROYED state

We already have a mechanism for blocking LAYOUTGET by means of the
plh_block_lgets counter. The only "service" that NFS_LAYOUT_DESTROYED
provides at this point is to block layoutget once the layout segment
list is empty, which basically means that you have to wait until
the pnfs_layout_hdr is destroyed before you can do pNFS on that file
again.

This patch enables the reuse of the pnfs_layout_hdr if the layout
segment list is empty.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/nfs4filelayout.c | 3 +--
 fs/nfs/pnfs.c           | 9 +--------
 fs/nfs/pnfs.h           | 7 -------
 3 files changed, 2 insertions(+), 17 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c
index 6cce57e7fe55..52d847212066 100644
--- a/fs/nfs/nfs4filelayout.c
+++ b/fs/nfs/nfs4filelayout.c
@@ -279,8 +279,7 @@ filelayout_reset_to_mds(struct pnfs_layout_segment *lseg)
 {
 	struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg);
 
-	return pnfs_test_layout_destroyed(lseg->pls_layout) ||
-		filelayout_test_devid_unavailable(node);
+	return filelayout_test_devid_unavailable(node);
 }
 
 /*
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index fcc72ecf2fd7..bda88a275071 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -331,8 +331,6 @@ pnfs_layout_remove_lseg(struct pnfs_layout_hdr *lo,
 	list_del_init(&lseg->pls_list);
 	/* Matched by pnfs_get_layout_hdr in pnfs_layout_insert_lseg */
 	atomic_dec(&lo->plh_refcount);
-	if (list_empty(&lo->plh_segs))
-		set_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags);
 	rpc_wake_up(&NFS_SERVER(inode)->roc_rpcwaitq);
 }
 
@@ -463,10 +461,8 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 
 	dprintk("%s:Begin lo %p\n", __func__, lo);
 
-	if (list_empty(&lo->plh_segs)) {
-		set_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags);
+	if (list_empty(&lo->plh_segs))
 		return 0;
-	}
 	list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
 		if (!recall_range ||
 		    should_free_lseg(&lseg->pls_range, recall_range)) {
@@ -590,7 +586,6 @@ pnfs_layoutgets_blocked(struct pnfs_layout_hdr *lo, nfs4_stateid *stateid,
 	    (int)(lo->plh_barrier - be32_to_cpu(stateid->seqid)) >= 0)
 		return true;
 	return lo->plh_block_lgets ||
-		test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags) ||
 		test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags) ||
 		(list_empty(&lo->plh_segs) &&
 		 (atomic_read(&lo->plh_outstanding) > lget));
@@ -912,8 +907,6 @@ pnfs_find_alloc_layout(struct inode *ino,
 	dprintk("%s Begin ino=%p layout=%p\n", __func__, ino, nfsi->layout);
 
 	if (nfsi->layout) {
-		if (test_bit(NFS_LAYOUT_DESTROYED, &nfsi->layout->plh_flags))
-			return NULL;
 		pnfs_get_layout_hdr(nfsi->layout);
 		return nfsi->layout;
 	}
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index aacda7fbb536..92f6ce6532ba 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -62,7 +62,6 @@ enum {
 	NFS_LAYOUT_RW_FAILED,		/* get rw layout failed stop trying */
 	NFS_LAYOUT_BULK_RECALL,		/* bulk recall affecting layout */
 	NFS_LAYOUT_ROC,			/* some lseg had roc bit set */
-	NFS_LAYOUT_DESTROYED,		/* no new use of layout allowed */
 	NFS_LAYOUT_RETURNED,		/* layout has already been returned */
 };
 
@@ -278,12 +277,6 @@ pnfs_test_layout_returned(struct pnfs_layout_hdr *lo)
 	return test_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags);
 }
 
-static inline bool
-pnfs_test_layout_destroyed(struct pnfs_layout_hdr *lo)
-{
-	return test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags);
-}
-
 static inline struct pnfs_layout_segment *
 pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
-- 
cgit v1.2.3


From e5929f3cff05e84f20c68df235f4768920e2e89e Mon Sep 17 00:00:00 2001
From: Trond Myklebust <Trond.Myklebust@netapp.com>
Date: Fri, 21 Sep 2012 16:37:02 -0400
Subject: NFSv4.1: Remove the NFS_LAYOUT_RETURNED state

It serves no purpose that the test for whether or not we have valid
layout segments doesn't already serve.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/pnfs.c |  7 +------
 fs/nfs/pnfs.h | 19 -------------------
 2 files changed, 1 insertion(+), 25 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 174c51a5001c..20a1b6222ff6 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -692,7 +692,7 @@ _pnfs_return_layout(struct inode *ino)
 
 	spin_lock(&ino->i_lock);
 	lo = nfsi->layout;
-	if (!lo || pnfs_test_layout_returned(lo)) {
+	if (!lo) {
 		spin_unlock(&ino->i_lock);
 		dprintk("NFS: %s no layout to return\n", __func__);
 		goto out;
@@ -710,7 +710,6 @@ _pnfs_return_layout(struct inode *ino)
 		goto out;
 	}
 	lo->plh_block_lgets++;
-	pnfs_mark_layout_returned(lo);
 	spin_unlock(&ino->i_lock);
 	pnfs_free_lseg_list(&tmp_list);
 
@@ -721,7 +720,6 @@ _pnfs_return_layout(struct inode *ino)
 		status = -ENOMEM;
 		pnfs_layout_io_set_failed(lo, IOMODE_RW);
 		pnfs_layout_io_set_failed(lo, IOMODE_READ);
-		pnfs_clear_layout_returned(lo);
 		pnfs_put_layout_hdr(lo);
 		goto out;
 	}
@@ -1111,9 +1109,6 @@ pnfs_update_layout(struct inode *ino,
 	if (list_empty(&lo->plh_segs))
 		first = true;
 
-	/* Enable LAYOUTRETURNs */
-	pnfs_clear_layout_returned(lo);
-
 	spin_unlock(&ino->i_lock);
 	if (first) {
 		/* The lo must be on the clp list if there is any
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 92f6ce6532ba..6cede2c6c961 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -62,7 +62,6 @@ enum {
 	NFS_LAYOUT_RW_FAILED,		/* get rw layout failed stop trying */
 	NFS_LAYOUT_BULK_RECALL,		/* bulk recall affecting layout */
 	NFS_LAYOUT_ROC,			/* some lseg had roc bit set */
-	NFS_LAYOUT_RETURNED,		/* layout has already been returned */
 };
 
 enum layoutdriver_policy_flags {
@@ -259,24 +258,6 @@ void nfs4_mark_deviceid_unavailable(struct nfs4_deviceid_node *node);
 bool nfs4_test_deviceid_unavailable(struct nfs4_deviceid_node *node);
 void nfs4_deviceid_purge_client(const struct nfs_client *);
 
-static inline void
-pnfs_mark_layout_returned(struct pnfs_layout_hdr *lo)
-{
-	set_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags);
-}
-
-static inline void
-pnfs_clear_layout_returned(struct pnfs_layout_hdr *lo)
-{
-	clear_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags);
-}
-
-static inline bool
-pnfs_test_layout_returned(struct pnfs_layout_hdr *lo)
-{
-	return test_bit(NFS_LAYOUT_RETURNED, &lo->plh_flags);
-}
-
 static inline struct pnfs_layout_segment *
 pnfs_get_lseg(struct pnfs_layout_segment *lseg)
 {
-- 
cgit v1.2.3


From 6296556f0b31eaff29f2a3aee2c17b7eae895b98 Mon Sep 17 00:00:00 2001
From: Peng Tao <bergwolf@gmail.com>
Date: Tue, 25 Sep 2012 14:55:57 +0800
Subject: NFS41: send real write size in layoutget

For buffer write, block layout client scan inode mapping to find
next hole and use offset-to-hole as layoutget length. Object
layout client uses offset-to-isize as layoutget length.

For direct write, both block layout and object layout use dreq->bytes_left.

Signed-off-by: Peng Tao <tao.peng@emc.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/blocklayout/blocklayout.c | 38 +++++++++++++++++++++++++++++++++++---
 fs/nfs/direct.c                  |  7 +++++++
 fs/nfs/internal.h                |  1 +
 fs/nfs/objlayout/objio_osd.c     |  9 ++++++++-
 fs/nfs/pnfs.c                    |  6 ++++--
 fs/nfs/pnfs.h                    |  3 ++-
 6 files changed, 57 insertions(+), 7 deletions(-)

(limited to 'fs/nfs/pnfs.h')

diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
index a34014a7f9a5..f1027b06a1a9 100644
--- a/fs/nfs/blocklayout/blocklayout.c
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -37,6 +37,7 @@
 #include <linux/bio.h>		/* struct bio */
 #include <linux/buffer_head.h>	/* various write calls */
 #include <linux/prefetch.h>
+#include <linux/pagevec.h>
 
 #include "../pnfs.h"
 #include "../internal.h"
@@ -1204,14 +1205,45 @@ bl_pg_test_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 	return pnfs_generic_pg_test(pgio, prev, req);
 }
 
+/*
+ * Return the number of contiguous bytes for a given inode
+ * starting at page frame idx.
+ */
+static u64 pnfs_num_cont_bytes(struct inode *inode, pgoff_t idx)
+{
+	struct address_space *mapping = inode->i_mapping;
+	pgoff_t end;
+
+	/* Optimize common case that writes from 0 to end of file */
+	end = DIV_ROUND_UP(i_size_read(inode), PAGE_CACHE_SIZE);
+	if (end != NFS_I(inode)->npages) {
+		rcu_read_lock();
+		end = radix_tree_next_hole(&mapping->page_tree, idx + 1, ULONG_MAX);
+		rcu_read_unlock();
+	}
+
+	if (!end)
+		return i_size_read(inode) - (idx << PAGE_CACHE_SHIFT);
+	else
+		return (end - idx) << PAGE_CACHE_SHIFT;
+}
+
 static void
 bl_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
 	if (pgio->pg_dreq != NULL &&
-	    !is_aligned_req(req, PAGE_CACHE_SIZE))
+	    !is_aligned_req(req, PAGE_CACHE_SIZE)) {
 		nfs_pageio_reset_write_mds(pgio);
-	else
-		pnfs_generic_pg_init_write(pgio, req);
+	} else {
+		u64 wb_size;
+		if (pgio->pg_dreq == NULL)
+			wb_size = pnfs_num_cont_bytes(pgio->pg_inode,
+						      req->wb_index);
+		else
+			wb_size = nfs_dreq_bytes_left(pgio->pg_dreq);
+
+		pnfs_generic_pg_init_write(pgio, req, wb_size);
+	}
 }
 
 static bool
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index 4be8673ee18d..cae26cbd59ee 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -46,6 +46,7 @@
 #include <linux/kref.h>
 #include <linux/slab.h>
 #include <linux/task_io_accounting_ops.h>
+#include <linux/module.h>
 
 #include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
@@ -191,6 +192,12 @@ static void nfs_direct_req_release(struct nfs_direct_req *dreq)
 	kref_put(&dreq->kref, nfs_direct_req_free);
 }
 
+ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq)
+{
+	return dreq->bytes_left;
+}
+EXPORT_SYMBOL_GPL(nfs_dreq_bytes_left);
+
 /*
  * Collects and returns the final error value/byte-count.
  */
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 89a795dc3027..59b133c5d652 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -464,6 +464,7 @@ static inline void nfs_inode_dio_wait(struct inode *inode)
 {
 	inode_dio_wait(inode);
 }
+extern ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq);
 
 /* nfs4proc.c */
 extern void __nfs4_read_done_cb(struct nfs_read_data *);
diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c
index ea6d111b03e9..be731e6b7b9c 100644
--- a/fs/nfs/objlayout/objio_osd.c
+++ b/fs/nfs/objlayout/objio_osd.c
@@ -41,6 +41,7 @@
 #include <scsi/osd_ore.h>
 
 #include "objlayout.h"
+#include "../internal.h"
 
 #define NFSDBG_FACILITY         NFSDBG_PNFS_LD
 
@@ -606,8 +607,14 @@ static bool aligned_on_raid_stripe(u64 offset, struct ore_layout *layout,
 void objio_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
 	unsigned long stripe_end = 0;
+	u64 wb_size;
 
-	pnfs_generic_pg_init_write(pgio, req);
+	if (pgio->pg_dreq == NULL)
+		wb_size = i_size_read(pgio->pg_inode) - req_offset(req);
+	else
+		wb_size = nfs_dreq_bytes_left(pgio->pg_dreq);
+
+	pnfs_generic_pg_init_write(pgio, req, wb_size);
 	if (unlikely(pgio->pg_lseg == NULL))
 		return; /* Not pNFS */
 
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index bd9769296e4e..9a2bcce45282 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -1240,7 +1240,8 @@ pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *r
 EXPORT_SYMBOL_GPL(pnfs_generic_pg_init_read);
 
 void
-pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
+pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
+			   struct nfs_page *req, u64 wb_size)
 {
 	BUG_ON(pgio->pg_lseg != NULL);
 
@@ -1248,10 +1249,11 @@ pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *
 		nfs_pageio_reset_write_mds(pgio);
 		return;
 	}
+
 	pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
 					   req->wb_context,
 					   req_offset(req),
-					   req->wb_bytes,
+					   wb_size,
 					   IOMODE_RW,
 					   GFP_NOFS);
 	/* If no lseg, fall back to write through mds */
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index 6cede2c6c961..2d722dba1111 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -186,7 +186,8 @@ void set_pnfs_layoutdriver(struct nfs_server *, const struct nfs_fh *, u32);
 void unset_pnfs_layoutdriver(struct nfs_server *);
 void pnfs_generic_pg_init_read(struct nfs_pageio_descriptor *, struct nfs_page *);
 int pnfs_generic_pg_readpages(struct nfs_pageio_descriptor *desc);
-void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *, struct nfs_page *);
+void pnfs_generic_pg_init_write(struct nfs_pageio_descriptor *pgio,
+			        struct nfs_page *req, u64 wb_size);
 int pnfs_generic_pg_writepages(struct nfs_pageio_descriptor *desc);
 bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req);
 void pnfs_set_lo_fail(struct pnfs_layout_segment *lseg);
-- 
cgit v1.2.3