Re: [PATCH 1/5] Revert "x86/fpu: Refine and simplify the magic number check during signal return"

[Andrei Vagin]

On Wed, May 27, 2026 at 12:49 PM Sasha Levin <sashal@xxxxxxxxxx> wrote:
>
> > This reverts commit dc8aa31a7ac2 ("x86/fpu: Refine and simplify the
> > magic number check during signal return").
> >
> > The reverted commit broke applications that construct signal frames in
> > userspace (such as CRIU and gVisor) if the frame's xstate size is
> > smaller than the kernel's fpstate->user_size.
>
> Holding this off on the stable side until the revert (and the rest of
> the series) lands in mainline. Once it's upstream, please ping with the
> mainline SHAs and the list of trees you want it on, and I'll queue it.

Will do. Thanks.

Thomas and Ingo, this revert is critical for CRIU. We've received
reports of silent memory corruption caused by the original change. Could
we please get this merged with high priority? This patch has been
pending on the mailing list for a month now; it is identical to the one
sent here:
https://lore.kernel.org/all/20260429000623.3356606-1-avagin@xxxxxxxxxx/

Thanks,
Andrei