linux-kernel - Re: [PATCH -next v3] x86/dumpstack: Prevent KASAN false positive warnings in __show

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

Message-ID: <20250922122855.GBaNFBB1GeS9ao4RmU@fat_crate.local>
Date: Mon, 22 Sep 2025 14:28:55 +0200
From: Borislav Petkov <bp@...en8.de>
To: Tengda Wu <wutengda@...weicloud.com>
Cc: x86@...nel.org, jpoimboe@...nel.org,
	Andrey Ryabinin <ryabinin.a.a@...il.com>,
	Dave Hansen <dave.hansen@...ux.intel.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Alexander Potapenko <glider@...gle.com>,
	Andrey Konovalov <andreyknvl@...il.com>,
	Dmitry Vyukov <dvyukov@...gle.com>, Ingo Molnar <mingo@...hat.com>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH -next v3] x86/dumpstack: Prevent KASAN false positive
 warnings in __show_regs

On Mon, Sep 22, 2025 at 08:14:50PM +0800, Tengda Wu wrote:
> Running 'echo t > /proc/sysrq-trigger' will trigger this type of asynchronous
> stack walk, as demonstrated in the use case provided below.

So lead with that please.

> >> [332706.552324] BUG: KASAN: out-of-bounds in __show_regs+0x4b/0x340
> >> [332706.552433] Read of size 8 at addr ffff88d24999fb20 by task sysrq_t_test.sh/3977032
> >> [332706.552562]
> >> [332706.552652] CPU: 36 PID: 3977032 Comm: sysrq_t_test.sh Kdump: loaded Not tainted 6.6.0+ #20
						^^^^^^^^^^^^^^

This doesn't help - it is some random script. Trigger this with the echo 't'
... thing.

> However, the main challenge with stopping the task first is that it fundamentally
> alters the state we're trying to inspect. The primary use case for an asynchronous
> stack walk is to diagnose a task that is already misbehaving (e.g., spinning in a
> hard lockup, not responding to stops). If we need to stop a task to get its stack,
> we might not be able to stop it at all, or the act of stopping it could change the
> call stack, hiding the root cause of the issue.
> 
> This is why my implementation selectively disables KASAN precisely for the async
> walk scenario.

Add this to the commit message too. You need to explain it in such a way so
that someone can use your instructions and reproduce it herself, without
asking more questions. IOW, it needs to be perfectly clear what the issue is
and what this is fixing.

Also, you don't have to rush your next version - we have a merge window coming
up so this'll wait for the next round.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette