NFS: Ensure we immediately start writeback on rescheduled writes
If the writes are being rescheduled due to a pNFS error, then we really
want to immediately start a new flush. The O_DIRECT code already does
this, so we only need to worry about buffered writes.
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index a057b4f..586726a 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1406,6 +1406,8 @@ static void nfs_async_write_error(struct list_head *head)
static void nfs_async_write_reschedule_io(struct nfs_pgio_header *hdr)
{
nfs_async_write_error(&hdr->pages);
+ filemap_fdatawrite_range(hdr->inode->i_mapping, hdr->args.offset,
+ hdr->args.offset + hdr->args.count - 1);
}
static const struct nfs_pgio_completion_ops nfs_async_write_completion_ops = {