Re: Random system freezes with Linux 6.12 + Xenomai 3.3.1 on Intel Ultra 5 125H

On 31.07.26 14:16, 刘杨 wrote:
> Hi
>      I used the following changes
> 
> sudo sysctl -w kernel.panic_on_rcu_stall=1
> sudo sysctl -w kernel.ftrace_dump_on_oops=1
> sudo sysctl -w kernel.panic=0
> sudo trace-cmd start \
>   -e 'cobalt*' \
>   -e 'sched:*' \
>   -e 'irq:*' \
>   -e 'irq_vectors:*' \
>   -e 'signal:*' \
>   -e 'rcu:rcu_stall_warning'
> echo 'snapshot' | sudo tee /sys/kernel/debug/tracing/events/rcu/rcu_stall_warning/trigger
> echo 'traceoff' | sudo tee /sys/kernel/debug/tracing/events/rcu/rcu_stall_warning/trigger

Just to be clear: This approach if for the kernel/setup where you still
have ssh access afterwards. You need to be able to extract the frozen
trace after the stall was detected.

> 
> echo 10 > /sys/module/rcupdate/parameters/rcu_cpu_stall_timeout
> 
> Kernel config add 
> CONFIG_DEBUG_KERNEL=y 
> CONFIG_PROVE_LOCKING=y 
> CONFIG_PROVE_RAW_LOCK_NESTING=y 
> CONFIG_DEBUG_PREEMPT=y
> 
> The log obtained is as follows:
> 
> fiveages@W2-0000:~$ [18611.389806] rcu: INFO: rcu_preempt detected stalls on CPUs/tasks:
> [18611.395936] rcu:     6-...0: (2 ticks this GP) idle=9d34/0/0x3 softirq=7383494/7383494 fqs=1876
> [18611.404400] rcu:     11-...0: (1 ticks this GP) idle=3a5c/1/0x4000000000000002 softirq=7395092/7395092 fqs=1876
> [18611.414250] rcu:     (detected by 1, t=15004 jiffies, g=14237693, q=73 ncpus=12)
> [18611.423416] rcu: rcu_preempt kthread starved for 7500 jiffies! g14237693 f0x0 RCU_GP_DOING_FQS(6) ->state=0x0 ->cpu=0
> [18611.434054] rcu:     Unless rcu_preempt kthread gets sufficient CPU time, OOM is now expected behavior.
> [18611.443210] rcu: RCU grace-period kthread stack dump:
> [18611.448322] rcu: Stack dump where RCU GP kthread last ran:
> [18611.454839] Kernel panic - not syncing: RCU Stall

So, lockdep is not reporting anything before this?

Jan

-- 
Siemens AG, Foundational Technologies
Linux Expert Center