linux-kernel - Re: [GIT PULL] runtime verification: Updates for 6.17

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAFLxGvyBcaf76F=O2w7q7pJGEGLtS4GwgOjfN=vLV3=80CfWnw@mail.gmail.com>
Date: Tue, 2 Sep 2025 22:51:02 +0200
From: Richard Weinberger <richard.weinberger@...il.com>
To: "Russell King (Oracle)" <linux@...linux.org.uk>
Cc: Steven Rostedt <rostedt@...dmis.org>, Linus Torvalds <torvalds@...ux-foundation.org>, 
	LKML <linux-kernel@...r.kernel.org>, Gabriele Monaco <gmonaco@...hat.com>, 
	Nam Cao <namcao@...utronix.de>, John Kacur <jkacur@...hat.com>, 
	Tomas Glozar <tglozar@...hat.com>, linux-arm-kernel@...ts.infradead.org
Subject: Re: [GIT PULL] runtime verification: Updates for 6.17

On Mon, Aug 18, 2025 at 10:41 AM Richard Weinberger
<richard.weinberger@...il.com> wrote:
>
> On Sat, Aug 16, 2025 at 9:26 PM Richard Weinberger
> <richard.weinberger@...il.com> wrote:
> > > Also please check whether you're using frame pointers or the unwinder
> > > (CONFIG_UNWINDER_FRAME_POINTER or CONFIG_ARM_UNWIND).
> >
> > With CONFIG_UNWINDER_FRAME_POINTER the stack trace is sane,
> > so only CONFIG_ARM_UNWIND is broken.
> > I kind of expected it the other way around...
>
> I found time to gather more details, maybe it rings a bell...
>
> The problem is related to va_list. As soon as va_list is passed as
> function argument the
> unwinder is no longer able to unwind correctly.
> Unwinding vpanic() does not lead to a correct frame anymore.
> Maybe because of va_list the function arguments are passed via stack?

Time to post an update, there is a problem in the ARM unwinder.
When the last instruction of a function is BL, LR points outside of
the function,
this confuses the unwinder.

This branch contains two WIP patches from Russel and myself to address
the issue:
https://git.kernel.org/pub/scm/linux/kernel/git/rw/misc.git/log/?h=arm_unwind_fix

-- 
Thanks,
//richard