Messages in this thread | | | Date | Wed, 6 Jun 2012 07:35:36 -0700 | From | Andi Kleen <> | Subject | Re: [PATCH 1/5] perf, x86: Don't assume the alternative cycles encoding is architectural |
| |
On Wed, Jun 06, 2012 at 04:28:30PM +0200, Peter Zijlstra wrote: > On Wed, 2012-06-06 at 07:23 -0700, Andi Kleen wrote: > > > > But it's very clear this cannot be done without a model check. > > I don't understand how you can even argue against that. > > Who said I fixed it without a model check?
Do you mean the explicit PEBS disable? That's only for one model, but it did not address the underlying problem that Linux did non architecturally guaranteed things just based on ArchPerfmon.
The PEBS disable still is needed even with my patch of course.
> > But simply disabling it for a model isn't how you do things.
Do you want it enabled per model? I can turn the flag around. Anything else?
-Andi
-- ak@linux.intel.com -- Speaking for myself only
| |