Re: New crashes walking proc with Saturday's git

[Chris Mason <clm@fb.com>]

On Sun, Nov 23, 2014 at 11:32 AM, Borislav Petkov <bp@alien8.de> wrote:
> On Sun, Nov 23, 2014 at 11:16:51AM -0500, Chris Mason wrote:
>>  It must be:
>> 
>>  commit 6e998916dfe327e785e7c2447959b2c1a3ea4930
>>  Author: Stanislaw Gruszka <sgruszka@redhat.com>
>>  Date:   Wed Nov 12 16:58:44 2014 +0100
>> 
>>     sched/cputime: Fix clock_nanosleep()/clock_gettime() 
>> inconsistency
>> 
>>  I'll do two runs to confirm, but it's the only related patch 
>> between rc5 and
>>  now.
> 
> Cool, let me know and I'll revert here too to verify. It might take a
> while to confirm as it happens only during resume but I could try the
> fork bomb if I must.
> 
> From looking at it I'm not getting any smarter. We iterate over the
> threads under a rcu_read_lock() so we should be protected. Unless the
> sched changes have somehow effect on RCU grace periods but WTH do I
> know... This is crazy code.

+               p->sched_class->update_curr(rq);

Total guess while I'm waiting for the test, but somehow this is NULL.

-chris