Re: x86/mm: Finishing off the fix for a should_flush_tlb race

[Stephen Dolan]

On Wed, 18 Feb 2026 at 19:43, Seth Forshee <sforshee@xxxxxxxxxx> wrote:
>
> On Tue, Feb 17, 2026 at 05:21:19PM +0000, Stephen Dolan wrote:
> > On Tue, 17 Feb 2026 at 15:20, Seth Forshee <sforshee@xxxxxxxxxx> wrote:
> > > > I wanted to check on the status of the stable patches, since I see the
> > > > upstream fix went into 6.18 but there's still no fix in the 6.12 stable
> > > > tree. We've been seeing segfaults during a test case with 6.12, and
> > > > after bisecting we found that reverting both "x86/mm: Eliminate window
> > > > where TLB flushes may be inadvertently skipped" and "x86/mm/tlb: Only
> > > > trim the mm_cpumask once a second" seems to get rid of the segfaults.
> > > > I'll try to get some testing with the proposed stable patch today.
> >
> > Hmm, ok. I posted the patch to stable@xxxxxxxxxxxxxxx a while ago but
> > it looks like it never got merged. Does this patch fix the issue for
> > you?
> >
> >     https://lore.kernel.org/stable/CAHDw0oE0334gEJ=ga1PAnZ3Av8+tFkKF-MJCF7Jj3i6pBVJvFQ@xxxxxxxxxxxxxx/
>
> Without the patch we've seen a segfault within a couple of hours of
> running our test case, often within 30 minutes. We ran for 18 hours
> with the patch applied without any crashes, so it looks like it fixes
> our crash.
>
> Thanks,
> Seth

That matches our testing also. I would still like to get this patch
merged into the 6.1, 6.6 and 6.12 kernel trees to fix this tricky bug.
But I'm pretty unfamiliar with the kernel development process and not
sure how to get that done.

Stephen