On Sat, 6 Sept 2025 at 02:50, John Stultz jstultz@google.com wrote:
On Fri, Sep 5, 2025 at 10:50 AM John Stultz jstultz@google.com wrote:
On Fri, Sep 5, 2025 at 6:21 AM Naresh Kamboju naresh.kamboju@linaro.org wrote:
Kernel warnings and a panic were observed on Juno-r2 while running LTP controllers (cgroup_fj_stress.sh) on the Linux next-20250904 with SCHED_PROXY_EXEC=y enabled build.
Regression Analysis:
- New regression? yes
- Reproducibility? yes
First seen on next-20250904 Bad: next-20250904 Good: next-20250822
Test regression: next-20250904 juno-r2 cgroup_fj_stress.sh kernel panic
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Thank you for the testing and the report here!
Juno-r2:
- LTP controllers
- cgroup_fj_stress.sh
Test crash: cgroup_fj_stress_net_cls_1_200_one: [ 365.917504] /usr/local/bin/kirk[402]: cgroup_fj_stress_net_cls_1_200_one: start (command: cgroup_fj_stress.sh net_cls 1 200 one) [ 374.230110] ------------[ cut here ]------------ [ 374.230132] WARNING: lib/timerqueue.c:55 at timerqueue_del+0x68/0x70, CPU#5: swapper/5/0
This looks like we are removing a timer that was already removed from the queue.
I don't see anything obvious right away in the delta that would clue me into what's going on, but I'll try to reproduce this.
So far I've not been able to reproduce this in my environment. If you are able to reproduce this easily, could you try enabling CONFIG_DEBUG_OBJECTS_TIMERS to see if it shows anything?
I have been running in CI loop for these tests to reproduce and did not find this again.
I will add this extra Kconfig CONFIG_DEBUG_OBJECTS_TIMERS and re-run to reproduce this reported regression.
- Naresh
thanks -john