Messages in this thread | | | Date | Tue, 9 Jun 2015 09:20:09 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH] perf/x86/intel/uncore: fix IMC missing box initialization |
| |
* Vince Weaver <vincent.weaver@maine.edu> wrote:
> On Mon, 11 May 2015, Ingo Molnar wrote: > > > > > * Peter Zijlstra <peterz@infradead.org> wrote: > > > > > On Mon, Apr 27, 2015 at 03:43:32AM +0000, Liang, Kan wrote: > > > > > > > > > > > > > > > This leads me to believe that this patch: > > > > > > > > > > > > commit c05199e5a57a579fea1e8fa65e2b511ceb524ffc > > > > > > Author: Kan Liang <kan.liang@intel.com> > > > > > > Date: Tue Jan 20 04:54:25 2015 +0000 > > > > > > > > > > > > perf/x86/intel/uncore: Move uncore_box_init() out of driver > > > > > initialization > > > > > > > > > > > > If I revert it, I bet things will work again. > > > > > > > > > > Yes the initialization needs to be moved out of the IPI context. > > > > > > I'm for the clean revert I think. Crashing is bad, but hiding/delaying > > > it seems counter productive too, it'll just mean we'll only learn about > > > it later. > > > > So should I revert c05199e5a57a, with a Cc: stable? > > any progress on this issue? With 4.1-rc7 I can still quickly crash a > Haswell machine due to this bug.
This dropped through the cracks. Peter?
Thanks,
Ingo
| |