Re: [PATCH bisected regression] sched: rebuild sched domains at suspend/resume

[Konstantin Khlebnikov <khlebnikov@openvz.org>]

Linus Torvalds wrote:
> On Tue, Mar 6, 2012 at 5:38 PM, Peter Zijlstra<a.p.zijlstra@chello.nl>  wrote:
>> On Tue, 2012-03-06 at 16:54 -0800, Linus Torvalds wrote:
>>> I do agree that reverting is probably safer at this point, but can we
>>> get agreement on this?
>>
>> I agree with reverting, shoot it in the head :-) Do you want a git
>> thingy?
>
> Well, it's less a "git thingy" and more that there are tons of people
> involved with the original commit that haven't even piped up.
>
> Srivatsa, Ingo, Prashanth..
>
> In fact, I notice that Prashanth doesn't even seem to have been cc'd,
> even if he's the original reporter of the commit that gets reverted.
> Added (see lkml)

I forget to mention my kernel boot options. I live with them for a while, so I just forget about it.
"debug threadirqs i915.i915_enable_rc6=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1 crashkernel=128M"

So, "threadirqs" is a lost piece of the puzzle -- without it I cannot reproduce the bug.
However, I have no idea how this is connected to sched-domains. =)

(Add Thomas Gleixner to CC)