linux-kernel - Re: [PATCH v2 3/9] mm/swap: avoid doing extra unlock error checks for direct swapin

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: <CAMgjq7AehV6tr9pKktzonHH9L_sEbB4c9V3Bi4U8wXZM9xZ4_A@mail.gmail.com>
Date: Tue, 9 Jan 2024 17:38:36 +0800
From: Kairui Song <ryncsn@...il.com>
To: "Huang, Ying" <ying.huang@...el.com>
Cc: linux-mm@...ck.org, Andrew Morton <akpm@...ux-foundation.org>, 
	Chris Li <chrisl@...nel.org>, Hugh Dickins <hughd@...gle.com>, 
	Johannes Weiner <hannes@...xchg.org>, Matthew Wilcox <willy@...radead.org>, Michal Hocko <mhocko@...e.com>, 
	Yosry Ahmed <yosryahmed@...gle.com>, David Hildenbrand <david@...hat.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 3/9] mm/swap: avoid doing extra unlock error checks for
 direct swapin

Huang, Ying <ying.huang@...el.com> 于2024年1月4日周四 16:12写道：
>
> Kairui Song <ryncsn@...il.com> writes:
>
> > From: Kairui Song <kasong@...cent.com>
> >
> > When swapping in a page, mem_cgroup_swapin_charge_folio is called for
> > new allocated folio, nothing else is referencing the folio so no need
> > to set the lock bit early. This avoided doing extra unlock checks
> > on the error path.
> >
> > Signed-off-by: Kairui Song <kasong@...cent.com>
> > ---
> >  mm/swap_state.c | 13 ++++++-------
> >  1 file changed, 6 insertions(+), 7 deletions(-)
> >
> > diff --git a/mm/swap_state.c b/mm/swap_state.c
> > index 24cb93ed5081..6130de8d5226 100644
> > --- a/mm/swap_state.c
> > +++ b/mm/swap_state.c
> > @@ -881,16 +881,15 @@ struct folio *swapin_direct(swp_entry_t entry, gfp_t gfp_mask,
> >       folio = vma_alloc_folio(GFP_HIGHUSER_MOVABLE, 0,
> >                               vma, vmf->address, false);
> >       if (folio) {
> > -             __folio_set_locked(folio);
> > -             __folio_set_swapbacked(folio);
> > -
> > -             if (mem_cgroup_swapin_charge_folio(folio,
> > -                                     vma->vm_mm, GFP_KERNEL,
> > -                                     entry)) {
> > -                     folio_unlock(folio);
> > +             if (mem_cgroup_swapin_charge_folio(folio, vma->vm_mm,
> > +                                                GFP_KERNEL, entry)) {
> >                       folio_put(folio);
> >                       return NULL;
> >               }
> > +
> > +             __folio_set_locked(folio);
> > +             __folio_set_swapbacked(folio);
> > +
> >               mem_cgroup_swapin_uncharge_swap(entry);
> >
> >               shadow = get_shadow_from_swap_cache(entry);
>
> I don't find any issue with the patch.  But another caller of
> mem_cgroup_swapin_charge_folio() in __read_swap_cache_async() setups
> newly allocated folio in the same way before the change.  Better to keep
> them same?  Because the benefit of change is small too.

OK, this is just a trivial optimization, I can drop it.