NFS: fix subtle change in COMMIT behavior

Recent work in the pgio layer made it possible for there to be more than one
request per page. This caused a subtle change in commit behavior, because
write.c:nfs_commit_unstable_pages compares the number of *pages* waiting for
writeback against the number of requests on a commit list to choose when to
send a COMMIT in a non-blocking flush.

This is probably hard to hit in normal operation - you have to be using
rsize/wsize < PAGE_SIZE, or pnfs with lots of boundaries that are not page
aligned to have a noticeable change in behavior.

Signed-off-by: Weston Andros Adamson <dros@primarydata.com>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>

1 files changed, 12 insertions, 5 deletions

diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index f83b02dc9166..d489ff3f438f 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -670,7 +670,8 @@ static void nfs_inode_add_request(struct inode *inode, struct nfs_page *req)
         nfs_lock_request(req);
 
         spin_lock(&inode->i_lock);
-        if (!nfsi->npages && NFS_PROTO(inode)->have_delegation(inode, FMODE_WRITE))
+        if (!nfsi->nrequests &&
+            NFS_PROTO(inode)->have_delegation(inode, FMODE_WRITE))
                 inode->i_version++;
         /*
          * Swap-space should not get truncated. Hence no need to plug the race
@@ -681,9 +682,11 @@ static void nfs_inode_add_request(struct inode *inode, struct nfs_page *req)
                 SetPagePrivate(req->wb_page);
                 set_page_private(req->wb_page, (unsigned long)req);
         }
-        nfsi->npages++;
+        nfsi->nrequests++;
         /* this a head request for a page group - mark it as having an
-         * extra reference so sub groups can follow suit */
+         * extra reference so sub groups can follow suit.
+         * This flag also informs pgio layer when to bump nrequests when
+         * adding subrequests. */
         WARN_ON(test_and_set_bit(PG_INODE_REF, &req->wb_flags));
         kref_get(&req->wb_kref);
         spin_unlock(&inode->i_lock);
@@ -709,7 +712,11 @@ static void nfs_inode_remove_request(struct nfs_page *req)
                         wake_up_page(head->wb_page, PG_private);
                         clear_bit(PG_MAPPED, &head->wb_flags);
                 }
-                nfsi->npages--;
+                nfsi->nrequests--;
+                spin_unlock(&inode->i_lock);
+        } else {
+                spin_lock(&inode->i_lock);
+                nfsi->nrequests--;
                 spin_unlock(&inode->i_lock);
         }
 
@@ -1735,7 +1742,7 @@ static int nfs_commit_unstable_pages(struct inode *inode, struct writeback_contr
                 /* Don't commit yet if this is a non-blocking flush and there
                  * are a lot of outstanding writes for this mapping.
                  */
-                if (nfsi->commit_info.ncommit <= (nfsi->npages >> 1))
+                if (nfsi->commit_info.ncommit <= (nfsi->nrequests >> 1))
                         goto out_mark_dirty;
 
                 /* don't wait for the COMMIT response */