Date: Mon, 02 Oct 2006 09:17:22 -0400
From: Steve Dickson <SteveD@redhat.com>
Subject: [RHEL5/FC6] [PATCH] fsstress on nfs trigger xmon on rhel5 beta1

The attached upstream patch stop a known regression
in the direct I/O path in NFS. I have verified and
IBM has also verified that this patch does indeed
fix the oops..

See bz:  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=207040

steved.

If the RPC call tanked, we should not be checking the return value
of data->res.verf->committed, since it is unlikely to even be
initialised.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
---
 fs/nfs/direct.c |   12 +++++++-----
 1 files changed, 7 insertions(+), 5 deletions(-)

--- linux-2.6.18.i686/fs/nfs/direct.c.dio	2006-09-28 14:08:30.816760000 -0400
+++ linux-2.6.18.i686/fs/nfs/direct.c	2006-10-02 08:53:06.772120000 -0400
@@ -534,10 +534,12 @@
 
 	spin_lock(&dreq->lock);
 
-	if (likely(status >= 0))
-		dreq->count += data->res.count;
-	else
-		dreq->error = task->tk_status;
+	if (unlikely(status < 0)) {
+		dreq->error = status;
+		goto out_unlock;
+	}
+
+	dreq->count += data->res.count;
 
 	if (data->res.verf->committed != NFS_FILE_SYNC) {
 		switch (dreq->flags) {
@@ -552,7 +554,7 @@
 				}
 		}
 	}
-
+out_unlock:
 	spin_unlock(&dreq->lock);
 }