Messages in this thread | | | Date | Sun, 23 Nov 2014 19:26:37 +0100 | From | Oleg Nesterov <> | Subject | Re: task_stat splat |
| |
On 11/23, Borislav Petkov wrote: > > On Sun, Nov 23, 2014 at 06:22:56PM +0100, Oleg Nesterov wrote: > > Could you identify somehow the line which matches > > task_sched_runtime+0x99 ? Probably do_task_delta_exec()... > > See http://marc.info/?l=linux-kernel&m=141675498417759&w=2 > > Chris triggers the same splat and reverting > > 6e998916dfe3 ("sched/cputime: Fix clock_nanosleep()/clock_gettime() inconsistency") > > fixes the issue for him. I'm testing that same thing reverted too, first > resume cycle looks good.
Yes, this probably confirms the theory, stop_class needs update_curr.
Oleg.
| |