Messages in this thread | | | Subject | Re: [PATCH V6 2/5] perf/x86/kvm: Avoid unnecessary work in guest filtering | From | "Liang, Kan" <> | Date | Mon, 4 Feb 2019 10:39:09 -0500 |
| |
On 2/4/2019 10:17 AM, Peter Zijlstra wrote: > On Mon, Feb 04, 2019 at 04:06:23PM +0100, Peter Zijlstra wrote: >> On Mon, Jan 21, 2019 at 01:42:28PM -0800, kan.liang@linux.intel.com wrote: >> >> So what's wrong with the below? >> >> Installing a quirk for > > (typing hard..) > > but what I was going to say is that it seems overkill to sprinkle all > that stuff around. >
The microcode patch for PEBS has been applied for Icelake and future platforms. - For the platforms before Haswell, there is no microcode patch applied. The x86_pmu.pebs_no_isolation should always be 1. - Between Haswell and Kabylake, x86_pmu.pebs_no_isolation is decided by quirk table. - For Icelake and future platforms, the x86_pmu.pebs_no_isolation should always be 0.
The patch you proposed in previous Email can work well for the existing platforms. But it will be a problem when adding new support. We have to always add INTEL_CPU_DESC(INTEL_FAM6_XXX, X, 0x00000000) for each CPU ID and each stepping. That will be a disaster for maintenance.
Thanks, Kan
| |