Re: [PATCH v3] mm/page_reporting: use system_freezable_wq to fix UAF during suspend
From: Michael S. Tsirkin
Date: Fri Jul 31 2026 - 20:17:07 EST
On Fri, Jul 31, 2026 at 01:21:40PM -0700, Link Lin wrote:
> On Thu, Jul 23, 2026 at 2:15 AM Michael S. Tsirkin <mst@xxxxxxxxxx> wrote:
> >
> > I suspect rest of work items we have, e.g. update_balloon_stats_work/update_balloon_size_work
> > all have issues around freeze/restore and error handling.
> >
> > --
> > MST
> >
>
> Hi Michael,
>
> We have indeed discovered a kernel crash in update_balloon_stats_work
> after the balloon driver failed to resume and initialize in our
> production fleet.
>
> Alex Wilkins <ahwilkins@xxxxxxxxxx> diagnosed, root caused, and
> reported this vulnerability.
I wouldn't call this a "vulnerability" just a garden variety bug.
> I am currently drafting and testing a
> 2-patch series to gracefully unregister callbacks and cancel work
> items on resume failure so that the VM continues running without the
> balloon driver instead of crashing with a kernel panic.
>
> I wanted to follow up and let you know that we are actively working on
> this. I will send out the patches to the list as soon as they are
> sufficiently tested.
>
> Best,
> Link