Re: [syzbot] WARNING: locking bug in __perf_event_task_sched_in (2)

[Jakub Kicinski]

On Tue, 10 Jan 2023 12:50:48 -0800 syzbot wrote:
>  <TASK>
>  lock_acquire kernel/locking/lockdep.c:5668 [inline]
>  lock_acquire+0x1e3/0x630 kernel/locking/lockdep.c:5633
>  rcu_lock_acquire include/linux/rcupdate.h:325 [inline]
>  rcu_read_lock include/linux/rcupdate.h:764 [inline]
>  perf_event_context_sched_in kernel/events/core.c:3913 [inline]
>  __perf_event_task_sched_in+0xe2/0x6c0 kernel/events/core.c:3980
>  perf_event_task_sched_in include/linux/perf_event.h:1328 [inline]
>  finish_task_switch.isra.0+0x5e5/0xc80 kernel/sched/core.c:5118
>  context_switch kernel/sched/core.c:5247 [inline]
>  __schedule+0xb92/0x5450 kernel/sched/core.c:6555
>  schedule+0xde/0x1b0 kernel/sched/core.c:6631
>  schedule_preempt_disabled+0x13/0x20 kernel/sched/core.c:6690
>  __mutex_lock_common kernel/locking/mutex.c:679 [inline]
>  __mutex_lock+0xa48/0x1360 kernel/locking/mutex.c:747
>  devl_lock net/devlink/core.c:54 [inline]
>  devlink_pernet_pre_exit+0x10a/0x220 net/devlink/core.c:301
>  ops_pre_exit_list net/core/net_namespace.c:159 [inline]
>  cleanup_net+0x455/0xb10 net/core/net_namespace.c:594
>  process_one_work+0x9bf/0x1710 kernel/workqueue.c:2289
>  worker_thread+0x669/0x1090 kernel/workqueue.c:2436
>  kthread+0x2e8/0x3a0 kernel/kthread.c:376
>  ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:308
>  </TASK>

Yes, I pooped it. We need to keep the mutex around as well as 
the devlink instance memory, otherwise locked screams.
Fix building..