[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMgjq7B_frEwFhUcosPL13hySk7otjThh-aOKOPALfpvUPJWLA@mail.gmail.com>
Date: Thu, 1 May 2025 02:15:06 +0800
From: Kairui Song <ryncsn@...il.com>
To: linux-mm@...ck.org
Cc: Andrew Morton <akpm@...ux-foundation.org>, Matthew Wilcox <willy@...radead.org>,
Hugh Dickins <hughd@...gle.com>, Chris Li <chrisl@...nel.org>,
David Hildenbrand <david@...hat.com>, Yosry Ahmed <yosryahmed@...gle.com>,
"Huang, Ying" <ying.huang@...ux.alibaba.com>, Nhat Pham <nphamcs@...il.com>,
Johannes Weiner <hannes@...xchg.org>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 4/6] filemap: do not use folio_contains for swap cache folios
On Thu, May 1, 2025 at 2:11 AM Kairui Song <ryncsn@...il.com> wrote:
>
> From: Kairui Song <kasong@...cent.com>
>
> Currently, none of the folio_contains callers should encounter swap
> cache folios.
>
> For fs/ callers, swap cache folios are never part of their workflow.
>
> For filemap and truncate, folio_contains is only used for sanity
> checks to verify the folio index matches the expected
> lookup / invalidation target.
>
> The swap cache does not utilize filemap or truncate helpers in ways
> that would trigger these checks, as it mostly implements its own
> cache management.
>
> Shmem won't trigger these sanity checks either unless thing went
> wrong, as it would directly trigger a BUG because swap cache index are
> unrelated and almost never matches shmem index. Shmem have to handle
> mixed values of folios, shadows, and swap entries, so it has its own
> way of handling the mapping.
>
> While some filemap helpers works for swap cache space, the swap cache
> is different from the page cache in many ways. So this particular helper
> will unlikely to work in a helpful way for swap cache folios.
>
> So make it explicit here that folio_contains should not be used for
> swap cache folios. This helps to avoid misuse, make swap cache less
> exposed and remove the folio_index usage here.
>
> Signed-off-by: Kairui Song <kasong@...cent.com>
> Acked-by: David Hildenbrand <david@...hat.com>
> ---
> include/linux/pagemap.h | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
> index af25fb640463..0c9aff5ec77f 100644
> --- a/include/linux/pagemap.h
> +++ b/include/linux/pagemap.h
> @@ -935,14 +935,14 @@ static inline struct page *folio_file_page(struct folio *folio, pgoff_t index)
> * @folio: The folio.
> * @index: The page index within the file.
> *
> - * Context: The caller should have the page locked in order to prevent
> - * (eg) shmem from moving the page between the page cache and swap cache
> - * and changing its index in the middle of the operation.
> + * Context: The caller should have the folio locked and ensure
> + * e.g., shmem did not move this folio to the swap cache.
> * Return: true or false.
> */
> static inline bool folio_contains(struct folio *folio, pgoff_t index)
> {
> - return index - folio_index(folio) < folio_nr_pages(folio);
> + VM_WARN_ON_FOLIO(folio_test_swapcache(folio), folio);
Ah, my bad, it will be better to be VM_WARN_ON_ONCE_FOLIO here as
suggested by David, other than this this series should be good.
> + return index - folio->index < folio_nr_pages(folio);
> }
>
> unsigned filemap_get_folios(struct address_space *mapping, pgoff_t *start,
> --
> 2.49.0
>
Powered by blists - more mailing lists