linux-kernel - Re: [tip:locking/futex] [futex] 56180dd20c: BUG:sleeping_function_called_from_invalid_context_at

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 for Android: free password hash cracker in your pocket

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

Message-ID: <87seikp94v.ffs@tglx>
Date: Fri, 25 Jul 2025 17:45:04 +0200
From: Thomas Gleixner <tglx@...utronix.de>
To: Oliver Sang <oliver.sang@...el.com>
Cc: Peter Zijlstra <peterz@...radead.org>, oe-lkp@...ts.linux.dev,
 lkp@...el.com, linux-kernel@...r.kernel.org, x86@...nel.org, Sebastian
 Andrzej Siewior <bigeasy@...utronix.de>, linux-mm@...ck.org,
 ltp@...ts.linux.it, oliver.sang@...el.com
Subject: Re: [tip:locking/futex] [futex]  56180dd20c:
 BUG:sleeping_function_called_from_invalid_context_at_kernel/nsproxy.c

Oliver!

On Fri, Jul 25 2025 at 17:10, Oliver Sang wrote:
> On Wed, Jul 23, 2025 at 07:22:43PM +0200, Thomas Gleixner wrote:
>> > [  286.673775][   C97] BUG: sleeping function called from invalid context at kernel/nsproxy.c:233   <---- (1)
>> > [  286.673784][   C97] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 6748, name: oom03
>> > [  286.673787][   C97] preempt_count: 7ffffffe, expected: 0
>> 
>> Ooops. That's a corrupted preempt counter, which has underflown twice.
>> 
>> Can you please enable CONFIG_DEBUG_PREEMPT, so we can see where this
>> happens?
>
> after enable CONFIG_DEBUG_PREEMPT, the config is as attached
> config-6.16.0-rc5-00002-g56180dd20c19

Thank you for trying, but I just realized that it tells us only when it
underflows, but we don't see where the actual extra decrement happens
before that.

Can you please enable CONFIG_PREEMPT_TRACER and add

'trace_event=preemptirq:preempt_disable,preemptirq:preempt_enable ftrace_dump_on_oops'

to the kernel command line. The latter will dump the recorded
preempt_enable/disable events and we can pinpoint the function which is
responsible for that.

Thanks,

        tglx