Re: [PATCH 17/26] netfs: Fix writethrough-mode error handling

From: Jeff Layton
Date: Mon Apr 15 2024 - 08:40:40 EST


On Thu, 2024-03-28 at 16:34 +0000, David Howells wrote:
> Fix the error return in netfs_perform_write() acting in writethrough-mode
> to return any cached error in the case that netfs_end_writethrough()
> returns 0.
>
> Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
> cc: Jeff Layton <jlayton@xxxxxxxxxx>
> cc: netfs@xxxxxxxxxxxxxxx
> cc: linux-fsdevel@xxxxxxxxxxxxxxx
> ---
> fs/netfs/buffered_write.c | 10 ++++++----
> 1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/fs/netfs/buffered_write.c b/fs/netfs/buffered_write.c
> index 8e4a3fb287e3..db4ad158948b 100644
> --- a/fs/netfs/buffered_write.c
> +++ b/fs/netfs/buffered_write.c
> @@ -188,7 +188,7 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
> enum netfs_how_to_modify howto;
> enum netfs_folio_trace trace;
> unsigned int bdp_flags = (iocb->ki_flags & IOCB_SYNC) ? 0: BDP_ASYNC;
> - ssize_t written = 0, ret;
> + ssize_t written = 0, ret, ret2;
> loff_t i_size, pos = iocb->ki_pos, from, to;
> size_t max_chunk = PAGE_SIZE << MAX_PAGECACHE_ORDER;
> bool maybe_trouble = false;
> @@ -409,10 +409,12 @@ ssize_t netfs_perform_write(struct kiocb *iocb, struct iov_iter *iter,
>
> out:
> if (unlikely(wreq)) {
> - ret = netfs_end_writethrough(wreq, iocb);
> + ret2 = netfs_end_writethrough(wreq, iocb);
> wbc_detach_inode(&wbc);
> - if (ret == -EIOCBQUEUED)
> - return ret;
> + if (ret2 == -EIOCBQUEUED)
> + return ret2;
> + if (ret == 0)
> + ret = ret2;
> }
>
> iocb->ki_pos += written;
>

Should this be merged independently? It looks like a bug that's present
now.

Reviewed-by: Jeff Layton <jlayton@xxxxxxxxxx>