On Tue, May 27, 2014 at 03:31:19AM -0700, Christoph Hellwig wrote:
> I recently ran into the issue fixed by
> "xfs: kill buffers over failed write ranges properly"
> which spams the log with lots of backtraces. Make debugging any issues
> like that easier by using WARN_ON_ONCE in the writeback code.
> Signed-off-by: Christoph Hellwig <hch@xxxxxx>
The change generally seems fine to me. IIRC, we had a point in time
where these issues were prevalent and noisy. We also had a few lingering
ones that were hard to reproduce. I do wonder whether this would make
that situation difficult to reproduce. For example, running through an
xfstests run where one test might reproduce randomly and suppress output
from a subsequent, perhaps more frequent reproducer. Am I correct to
assume that once fired, the warning wouldn't fire again before a reboot
or module reload?
Hmm, did we have asserts that covered these scenarios as well?
> fs/xfs/xfs_aops.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
> diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
> index d1b99b6..e32640e 100644
> --- a/fs/xfs/xfs_aops.c
> +++ b/fs/xfs/xfs_aops.c
> @@ -975,7 +975,7 @@ xfs_vm_writepage(
> * Given that we do not allow direct reclaim to call us, we should
> * never be called while in a filesystem transaction.
> - if (WARN_ON(current->flags & PF_FSTRANS))
> + if (WARN_ON_ONCE(current->flags & PF_FSTRANS))
> goto redirty;
> /* Is this page beyond the end of the file? */
> @@ -1225,9 +1225,9 @@ xfs_vm_releasepage(
> xfs_count_page_state(page, &delalloc, &unwritten);
> - if (WARN_ON(delalloc))
> + if (WARN_ON_ONCE(delalloc))
> return 0;
> - if (WARN_ON(unwritten))
> + if (WARN_ON_ONCE(unwritten))
> return 0;
> return try_to_free_buffers(page);
> xfs mailing list