On Wed, 2025-04-02 at 14:02 +0300, Dan Carpenter wrote:
> Call nfs_release_request() on this error path before returning.
>
> Fixes: c3f2235782c3 ("nfs: fold nfs_folio_find_and_lock_request into
> nfs_lock_and_join_requests")
> Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> ---
> From static analysis and untested. Pretty sure it's correct,
> though. ;)
>
> fs/nfs/write.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/fs/nfs/write.c b/fs/nfs/write.c
> index aa3d8bea3ec0..23df8b214474 100644
> --- a/fs/nfs/write.c
> +++ b/fs/nfs/write.c
> @@ -579,8 +579,10 @@ static struct nfs_page
> *nfs_lock_and_join_requests(struct folio *folio)
>
> while (!nfs_lock_request(head)) {
> ret = nfs_wait_on_request(head);
> - if (ret < 0)
> + if (ret < 0) {
> + nfs_release_request(head);
> return ERR_PTR(ret);
> + }
> }
>
> /* Ensure that nobody removed the request before we locked
> it */
Yep, that's a slam dunk. Thanks for spotting it!
--
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@xxxxxxxxxxxxxxx
