linux-kernel - Re: [PATCH v2 03/15] mm, swap: fix swap cahe index error when retrying reclaim

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAMgjq7CyXEy3+98JdATscgPu9b9oLbE6CM7-2pFpiQP2QrsrpQ@mail.gmail.com>
Date: Sat, 6 Sep 2025 14:28:17 +0800
From: Kairui Song <ryncsn@...il.com>
To: Chris Li <chrisl@...nel.org>
Cc: linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>, 
	Matthew Wilcox <willy@...radead.org>, Hugh Dickins <hughd@...gle.com>, Barry Song <baohua@...nel.org>, 
	Baoquan He <bhe@...hat.com>, Nhat Pham <nphamcs@...il.com>, 
	Kemeng Shi <shikemeng@...weicloud.com>, Baolin Wang <baolin.wang@...ux.alibaba.com>, 
	Ying Huang <ying.huang@...ux.alibaba.com>, Johannes Weiner <hannes@...xchg.org>, 
	David Hildenbrand <david@...hat.com>, Yosry Ahmed <yosryahmed@...gle.com>, 
	Lorenzo Stoakes <lorenzo.stoakes@...cle.com>, Zi Yan <ziy@...dia.com>, 
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 03/15] mm, swap: fix swap cahe index error when
 retrying reclaim

On Sat, Sep 6, 2025 at 11:19 AM Chris Li <chrisl@...nel.org> wrote:
>
> Hi Kairui,
>
> The patch looks obviously correct to me with some very minor nitpicks following.
>
> Acked-by: Chris Li <chrisl@...nel.org>
>
> On Fri, Sep 5, 2025 at 12:14 PM Kairui Song <ryncsn@...il.com> wrote:
> >
> > From: Kairui Song <kasong@...cent.com>
> >
> > The allocator will reclaim cached slots while scanning. Currently, it
> > will try again if the reclaim found a folio that is already removed from
> > the swap cache due to a race. But the following lookup will be using the
> > wrong index. It won't cause any OOB issue since the swap cache index is
> > truncated upon lookup, but it may lead to reclaiming of an irrelevant
> > folio.
> >
> > This should not cause a measurable issue, but we should fix it.
> >
> > Fixes: fae8595505313 ("mm, swap: avoid reclaiming irrelevant swap cache")
> > Signed-off-by: Kairui Song <kasong@...cent.com>
> > ---
> >  mm/swapfile.c | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/mm/swapfile.c b/mm/swapfile.c
> > index 4b8ab2cb49ca..4c63fc62f4cb 100644
> > --- a/mm/swapfile.c
> > +++ b/mm/swapfile.c
> > @@ -240,13 +240,13 @@ static int __try_to_reclaim_swap(struct swap_info_struct *si,
> >          * Offset could point to the middle of a large folio, or folio
> >          * may no longer point to the expected offset before it's locked.
> >          */
> > -       entry = folio->swap;
> Nitpick:
> This and the following reuse the folio->swap dereference and
> swp_offset() many times.
> You can use some local variables to cache the value into a register
> and less function calls. I haven't looked into if the compiler will do
> the same expression elimination on this, a good compiler should. The
> following looks less busy and doesn't need the compiler to optimize it
> for you.
>
>            fe = folio->swap;
>            eoffset = swp_offset(fe);
>            if (offset < eoffset ) || offset >= eoffset + nr_pages) {
> ...
>            }
>            offset = eoffset;
>
> This might generate better code due to less function code. If the
> compiler does the perfect jobs the original code can generate the same
> optimized code as well.

Right, this part of the code will be gone soon so I think maybe better
to keep the change minimal, and it's not a hot path.

>
> > -       if (offset < swp_offset(entry) || offset >= swp_offset(entry) + nr_pages) {
> > +       if (offset < swp_offset(folio->swap) ||
> > +           offset >= swp_offset(folio->swap) + nr_pages) {
> >                 folio_unlock(folio);
> >                 folio_put(folio);
> >                 goto again;
> >         }
> > -       offset = swp_offset(entry);
> > +       offset = swp_offset(folio->swap);
>
> So the first entry is only assigned once in the function and never changed?
>
> You can use const to declare it.

That's a very good point, thanks!

>
> Chris
>
> >
> >         need_reclaim = ((flags & TTRS_ANYWAY) ||
> >                         ((flags & TTRS_UNMAPPED) && !folio_mapped(folio)) ||
> > --
> > 2.51.0
> >
> >
>