On Thu, Nov 24, 2022 at 03:12:23PM +0100, Matija Glavinic Pecotic wrote: > On 24.11.2022 14:51, Andy Shevchenko wrote: > > On Thu, Nov 24, 2022 at 01:57:53PM +0100, Matija Glavinic Pecotic wrote: > >> set_rtc_noop, get_rtc_noop are used runtime, therefore need init removed. > >> Problem was observed on x86 platform where cmos rtc is unused and disabled > >> via device tree. Function was triggered from ntp: sync_hw_clock, although > >> CONFIG_RTC_SYSTOHC=n, however sync_cmos_clock doesn't honour that. > > > >> Oops: 0011 [#1] PREEMPT SMP NOPTI > >> Workqueue: events_power_efficient sync_hw_clock > >> RIP: 0010:set_rtc_noop > >> Call Trace: > >> update_persistent_clock64 > >> sync_hw_clock > > > > (see below) > > > >> process_one_work > >> worker_thread > >> ? rescuer_thread > >> kthread > >> ? kthread_create_worker_on_cpu > >> ret_from_fork > > > > This looks much better, but you can shrink even more. > > Thanks for the patience here. Would this be something shrink enough, > but still preserving all important information (such as it is oops, it > crashed in set_rtc_noop, and it is sync_hw_clock which triggered it) > Oops: 0011 [#1] PREEMPT SMP NOPTI This is questionable. That's basically what you describe as a consequences of the absence of these stubs in the commit message (i.o.w. what the current state of affairs is). > Workqueue: events_power_efficient sync_hw_clock > RIP: 0010:set_rtc_noop > Call Trace: > update_persistent_clock64 > sync_hw_clock Yes to the rest. -- With Best Regards, Andy Shevchenko
