tentatively merged into cifs-2.6.git for-next pending testing
On Mon, Jul 18, 2022 at 2:07 PM Matthew Wilcox (Oracle)
<willy@xxxxxxxxxxxxx> wrote:
>
> If we hit the 'index == next_cached' case, we leak a refcount on the
> struct page. Fix this by using readahead_folio() which takes care of
> the refcount for you.
>
> Fixes: 0174ee9947bd ("cifs: Implement cache I/O by accessing the cache directly")
> Cc: David Howells <dhowells@xxxxxxxxxx>
> Cc: Jeff Layton <jlayton@xxxxxxxxxx>
> Signed-off-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx>
> ---
> fs/cifs/file.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/fs/cifs/file.c b/fs/cifs/file.c
> index e64cda7a7610..6985710e14c2 100644
> --- a/fs/cifs/file.c
> +++ b/fs/cifs/file.c
> @@ -4459,10 +4459,10 @@ static void cifs_readahead(struct readahead_control *ractl)
> * TODO: Send a whole batch of pages to be read
> * by the cache.
> */
> - page = readahead_page(ractl);
> - last_batch_size = 1 << thp_order(page);
> + struct folio *folio = readahead_folio(ractl);
> + last_batch_size = folio_nr_pages(folio);
> if (cifs_readpage_from_fscache(ractl->mapping->host,
> - page) < 0) {
> + &folio->page) < 0) {
> /*
> * TODO: Deal with cache read failure
> * here, but for the moment, delegate
> @@ -4470,7 +4470,7 @@ static void cifs_readahead(struct readahead_control *ractl)
> */
> caching = false;
> }
> - unlock_page(page);
> + folio_unlock(folio);
> next_cached++;
> cache_nr_pages--;
> if (cache_nr_pages == 0)
> --
> 2.35.1
>
--
Thanks,
Steve
