Re: [PATCH v2] arm64: ptrace: don't warn on missing siginfo in a group-stop
From: Will Deacon
Date: Fri Nov 13 2026 - 10:45:29 EST
On Mon, Nov 09, 2026 at 02:43:22PM +0100, Oleg Nesterov wrote:
> On 11/04, Will Deacon wrote:
> >
> > > > @@ -588,7 +584,7 @@ static void update_syscall_orig_x0_after_ptrace(struct task_struct *target)
> > > > * Don't update orig_x0 for a syscall-exit-stop, as x0 now contains the
> > > > * return value of the system call.
> > > > */
> > > > - if ((info->si_code & ~0x80) == SIGTRAP &&
> > > > + if (info && (info->si_code & ~0x80) == SIGTRAP &&
> > > > target->ptrace_message == PTRACE_EVENTMSG_SYSCALL_EXIT) {
> > > > return;
> > >
> > > LGTM, but I don't understand this arm-specific code.
> >
> > Welcome to the club :)
>
> Thank you ;)
>
> > The original rationale is described in e057b9477232 ("arm64: syscall:
> > Ensure saved x0 is kept in-sync with tracer updates"), so hopefully that
> > helps to set the scene.
>
> [...snip...]
>
> Thanks for your explanations. Yes, I tried to read that and other commits.
> The only result is that I've managed to convinced myself that more or less
> I understand why 88b839ce497c and 5a87e8c7f3702 make sense.
>
> > So that's the mess we're in and I'm all ears if there's a cleaner way to
> > sort it out without breaking anything that relies on the fragile semantics
> > we already have.
>
> No, I can't suggest anything better. I think your patches and the fix from
> Zhao are fine.
>
> ----------------------------------------------------------------------------
> Now the question...
>
> So we had CVE-2026-68171 and then it was rejected:
> https://lore.kernel.org/linux-cve-announce/2026081123-REJECTED-f1f0@gregkh/
>
> Now we have another one, CVE-2026-90347
> https://lore.kernel.org/linux-cve-announce/2026091703-CVE-2026-90347-6f2c@gregkh/
>
> Why?
>
> To me the problem looks minor (security wise). Yes, a malicious debugger can
> fool seccomp. So what? What is the real problem from the scurity perspective?
I think the security rationale probably goes back to a5cd110cb836
("arm64/ptrace: run seccomp after ptrace") so you'd need to ask Kees!
I suspect that some environments (Android?) don't want the tracer to be
able to execute arbitrary system calls in the context of the tracee.
Will