linux-kernel - Re: korg-clang-19-lkftconfig-hardening: TI x15 board - PC is at edma

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: <CA+G9fYuf6A+3gjNfifnd5cRL5oOU6HA2j-KXaaP90mskNVrV9Q@mail.gmail.com>
Date: Thu, 28 Nov 2024 15:30:56 +0530
From: Naresh Kamboju <naresh.kamboju@...aro.org>
To: Nathan Chancellor <nathan@...nel.org>
Cc: open list <linux-kernel@...r.kernel.org>, clang-built-linux <llvm@...ts.linux.dev>, 
	dmaengine@...r.kernel.org, lkft-triage@...ts.linaro.org, 
	peter.ujfalusi@...il.com, Vinod Koul <vkoul@...nel.org>, 
	Dan Carpenter <dan.carpenter@...aro.org>, Anders Roxell <anders.roxell@...aro.org>, 
	Arnd Bergmann <arnd@...db.de>, kees@...nel.org, linux-hardening@...r.kernel.org
Subject: Re: korg-clang-19-lkftconfig-hardening: TI x15 board - PC is at
 edma_probe (drivers/dma/ti/edma.c

On Mon, 25 Nov 2024 at 21:53, Nathan Chancellor <nathan@...nel.org> wrote:
>
> Hi Naresh,
>
> + Kees and linux-hardening, since this is a hardening configuration.
>
> On Mon, Nov 25, 2024 at 07:34:22PM +0530, Naresh Kamboju wrote:
> > The arm TI x15 board boot has failed with the Linux next, mainline
> > and the Linux stable. Please find boot log and build links.
> >
> > The boot failed with clang tool chain and PASS with gcc-13.
> >
> > Device: TI x15 device
> > Boot pass: gcc-13
>
> Are the UBSAN options getting enabled with GCC as well? I am somewhat
> surprised that they are not agreeing here, unless this is a __counted_by
> related issue? Does this not happen with GCC 14 as well? It would be
> nice if we could get a copy of GCC 15 with __counted_by into TuxMake for
> validation of __counted_by between the two compilers easily. I did not
> see any obvious instances of __counted_by in edma_probe() but
> admittedly, I did not look too hard.
>
> > Boot failed: clang-19
> > Configs: korg-clang-19-lkftconfig-hardening
> > Boot pass: qemu-armv7 (Additional info)
> >
> > This is always reproducible.
> >
> > x15 beagleboard:
> >   boot:
> >     * clang-nightly-lkftconfig-hardening
> >     * korg-clang-19-lkftconfig-hardening
> >
> > Reported-by: Linux Kernel Functional Testing <lkft@...aro.org>
> >
> > Log details:
> > ------------
> > [    0.000000] Booting Linux on physical CPU 0x0
> > [    0.000000] Linux version 6.12.0 (tuxmake@...make) (ClangBuiltLinux
> > clang version 19.1.4 (https://github.com/llvm/llvm-project.git
> > aadaa00de76ed0c4987b97450dd638f63a385bed), ClangBuiltLinux LLD 19.1.4
> > (https://github.com/llvm/llvm-project.git
> > aadaa00de76ed0c4987b97450dd638f63a385bed)) #1 SMP @1732428891
> > [    0.000000] CPU: ARMv7 Processor [412fc0f2] revision 2 (ARMv7), cr=30c5387d
> > <>
> > [    3.543395] pcieport 0000:00:00.0: PME: Signaling with IRQ 136
> > [    3.556976] Internal error: Oops - undefined instruction: 0 [#1] SMP ARM
>
> Can you turn off UBSAN_TRAP and see if that gives us any indication as
> to where exactly UBSAN_BOUNDS is triggering here? It is not entirely
> clear because we do not seem to have a line number in edma.c from
> decoding the stacktrace.

Anders made a new build with UBSAN_TRAP turn off and the boot pass
 CONFIG_UBSAN_TRAP=n
on arm TI x15 device.

Link Log:
--------
 - https://lkft.validation.linaro.org/scheduler/job/8008096#L1846
 - https://storage.tuxsuite.com/public/linaro/anders/builds/2pR6MSNK0MqztHMe45eAqWYlIaf/

- Naresh