linux-kernel - Re: commit 3284e4adca9b causes hang on boot with CONFIG_PREEMPT

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

Message-ID: <862fa11d-0f9a-4193-91d1-ea930d2d334f@nvidia.com>
Date: Fri, 11 Jul 2025 19:06:08 -0400
From: Joel Fernandes <joelagnelf@...dia.com>
To: Bert Karwatzki <spasswolf@....de>
Cc: linux-kernel@...r.kernel.org, linux-next@...r.kernel.org,
 linux-rt-devel@...ts.linux.dev, ankur.a.arora@...cle.com,
 bobo.shaobowang@...wei.com, boqun.feng@...il.com, frederic@...nel.org,
 joel@...lfernandes.org, neeraj.upadhyay@...nel.org, paulmck@...nel.org,
 rcu@...r.kernel.org, urezki@...il.com, wangxiongfeng2@...wei.com,
 xiexiuqi@...wei.com, xiqi2@...wei.com
Subject: Re: commit 3284e4adca9b causes hang on boot with CONFIG_PREEMPT_RT=y



On 7/11/2025 7:00 PM, Bert Karwatzki wrote:
> When booting linux next-20250711 (with CONFIG_PREEMPT_RT=y) on my MSI Alpha 15 
> Laptop running debian sid amd64 the boot process hangs with the last 
> messages displayed on screen being:
> 
> fbcon: amdgpudrmfb (fb0) is primary device
> Console: switching to colour frame buffer device
> amdgpu: 0000:08:00.0: [drm]fb0: admgpudrmfb frame buffer device
> 
> after some time (about 60s) this error messages appears (hand copied
> from screen, not entirely accurate)
> 
> rcu_preempt self detected stall
> 
> with call trace
> run_irq_workd
> smpboot_thread_fn
> kthread
> ? kthreads_online_cpu
> ? kthreads_online_cpu
> ret_from_fork
> ? kthreads_online_cpu
> ret_from_fork
> 
> This only occurs when compiling with CONFIG_PREEMPT_RT=y.
> I bisected this and found the first bad commit to be
> 
> 3284e4adca9b ("rcu: Fix rcu_read_unlock() deadloop due to IRQ work")

This commit is still using old code which was fixed in the last day.

Here is the new commit:
https://web.git.kernel.org/pub/scm/linux/kernel/git/rcu/linux.git/commit/?h=next&id=2e154d164418e1eaadbf5dc58cbf19e7be8fdc67

Thanks!

 - Joel