NFS: Fix two bugs in the O_DIRECT write code

From: Trond Myklebust
Date: Sat Apr 14 2007 - 19:13:39 EST


Do not flag an error if the COMMIT call fails and we decide to resend the
writes. Let the resend flag the error if it fails.

If a write has failed, then nfs_direct_write_result should not attempt to
send a commit. It should just exit asap and return the error to the user.

Signed-off-by: Trond Myklebust <Trond.Myklebust@xxxxxxxxxx>
Cc: Chuck Lever <chuck.lever@xxxxxxxxxx>
---

fs/nfs/direct.c | 11 +++++++----
1 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index b1c98ea..2877744 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -432,10 +432,10 @@ static void nfs_direct_commit_result(struct rpc_task *task, void *calldata)
if (NFS_PROTO(data->inode)->commit_done(task, data) != 0)
return;
if (unlikely(task->tk_status < 0)) {
- dreq->error = task->tk_status;
+ dprintk("NFS: %5u commit failed with error %d.\n",
+ task->tk_pid, task->tk_status);
dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
- }
- if (memcmp(&dreq->verf, &data->verf, sizeof(data->verf))) {
+ } else if (memcmp(&dreq->verf, &data->verf, sizeof(data->verf))) {
dprintk("NFS: %5u commit verify failed\n", task->tk_pid);
dreq->flags = NFS_ODIRECT_RESCHED_WRITES;
}
@@ -531,9 +531,12 @@ static void nfs_direct_write_result(struct rpc_task *task, void *calldata)

spin_lock(&dreq->lock);

+ if (unlikely(dreq->error != 0))
+ goto out_unlock;
if (unlikely(status < 0)) {
+ /* An error has occured, so we should not commit */
+ dreq->flags = 0;
dreq->error = status;
- goto out_unlock;
}

dreq->count += data->res.count;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/