Re: [PATCH v3 2/2] afs: unlock the folio when vnode is marked deleted

[Jeff Layton <jlayton@kernel.org>]

On Thu, 2022-07-07 at 12:51 +0800, xiubli@redhat.com wrote:
> From: Xiubo Li <xiubli@redhat.com>
> 
> The check_write_begin() should unlock the folio if return non-zero,
> otherwise locked.
> 
> URL: https://tracker.ceph.com/issues/56423
> Signed-off-by: Xiubo Li <xiubli@redhat.com>
> ---
>  fs/afs/file.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/afs/file.c b/fs/afs/file.c
> index afacce797fb9..b23e7b5a48ad 100644
> --- a/fs/afs/file.c
> +++ b/fs/afs/file.c
> @@ -379,7 +379,13 @@ static int afs_check_write_begin(struct file *file, loff_t pos, unsigned len,
>  {
>  	struct afs_vnode *vnode = AFS_FS_I(file_inode(file));
>  
> -	return test_bit(AFS_VNODE_DELETED, &vnode->flags) ? -ESTALE : 0;
> +	if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) {
> +		folio_unlock(*folio);
> +		folio_put(*folio);

Don't you also need this?

	*folio = NULL;

> +		return -ESTALE;
> +	}
> +
> +	return 0;
>  }
>  
>  static void afs_free_request(struct netfs_io_request *rreq)

-- 
Jeff Layton <jlayton@kernel.org>