linux-kernel - Re: [tj-wq:affinity-scopes-dbg-invalid-cpu] [workqueue] e3485721f9: WARNING:at_kernel/workqueue.c:#workqueue_online

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: <ZIdZQBJbOkuaAYvx@slm.duckdns.org>
Date:   Mon, 12 Jun 2023 07:43:28 -1000
From:   Tejun Heo <tj@...nel.org>
To:     kernel test robot <oliver.sang@...el.com>
Cc:     oe-lkp@...ts.linux.dev, lkp@...el.com, linux-kernel@...r.kernel.org
Subject: Re: [tj-wq:affinity-scopes-dbg-invalid-cpu] [workqueue] e3485721f9:
 WARNING:at_kernel/workqueue.c:#workqueue_online_cpu

On Mon, Jun 12, 2023 at 07:42:11AM -1000, Tejun Heo wrote:
> On Mon, Jun 12, 2023 at 10:04:02AM +0800, kernel test robot wrote:
> > [    3.388228][   T21] ------------[ cut here ]------------
> > [ 3.394990][ T21] WARNING: CPU: 1 PID: 21 at kernel/workqueue.c:5393 workqueue_online_cpu (kernel/workqueue.c:5393 kernel/workqueue.c:5481) 
> ...
> > [ 3.584746][ T21] ? workqueue_prepare_cpu (kernel/workqueue.c:5470) 
> > [ 3.591752][ T21] ? local_clock (kernel/sched/clock.c:270 kernel/sched/clock.c:307) 
> > [ 3.596749][ T21] cpuhp_invoke_callback (arch/x86/include/asm/jump_label.h:27 include/linux/jump_label.h:207 include/trace/events/cpuhp.h:65 kernel/cpu.c:193) 
> > [ 3.603751][ T21] ? try_to_wake_up (include/linux/sched.h:2254 kernel/sched/core.c:4312) 
> > [ 3.610751][ T21] cpuhp_thread_fun (kernel/cpu.c:815) 
> 
> This is set_cpus_allowed_ptr() failing on kworkers when workqueue tries to
> put back the per-cpu kworkers on the onlining CPU, which is rather
> unexpected. Let me include the error code in the warning message so that we
> can understand why it fails the next time.

Ah, never mind. This ->__pod_cpumaks not being initialized correctly by the
patch and is already fixed in the newer version of the code.

Thanks.

-- 
tejun