Re: [v4 PATCH] fs/proc: task_mmu.c: don't read mapcount for migration entry

[Yang Shi]

On Thu, Apr 13, 2023 at 4:58 PM David Rientjes <rientjes@xxxxxxxxxx> wrote:
>
> On Mon, 3 Apr 2023, Yang Shi wrote:
>
> > On Mon, Apr 3, 2023 at 12:30 AM David Hildenbrand <david@xxxxxxxxxx> wrote:
> > >
> > > On 24.03.23 21:12, Yang Shi wrote:
> > > > On Fri, Mar 24, 2023 at 4:25 AM Vlastimil Babka <vbabka@xxxxxxx> wrote:
> > > >>
> > > >> On 3/23/23 21:45, Yang Shi wrote:
> > > >>> On Thu, Mar 23, 2023 at 3:11 AM Vlastimil Babka <vbabka@xxxxxxx> wrote:
> > > >>>
> > > >>> Out of curiosity, is there any public link for this CVE? Google search
> > > >>> can't find it.
> > > >>
> > > >> Only this one is live so far, AFAIK
> > > >>
> > > >> https://bugzilla.redhat.com/show_bug.cgi?id=2180936
> > > >
> > > > Thank you.
> > >
> > > There is now
> > >
> > > https://access.redhat.com/security/cve/cve-2023-1582
> >
> > Thank you.
> >
>
> Hi Yang,
>
> commit 24d7275ce2791829953ed4e72f68277ceb2571c6
> Author: Yang Shi <shy828301@xxxxxxxxx>
> Date:   Fri Feb 11 16:32:26 2022 -0800
>
>     fs/proc: task_mmu.c: don't read mapcount for migration entry
>
> is backported to 5.10 stable but not to 5.4 or earlier stable trees.  The
> commit advertises to fix a commit from 4.5.
>
> Do we need stable backports for earlier trees or are they not affected?

IIRC, it is needed. But I can't remember why it wasn't. Maybe just
because the patch failed to apply to those trees and needs to backport
the patch manually.

>
> Thanks!