Re: [PATCH v1 0/5] KVM in-guest performance monitoring

[Jan Kiszka <jan.kiszka@siemens.com>]

On 2011-05-12 11:33, Joerg Roedel wrote:
> On Wed, May 11, 2011 at 11:55:28AM -0400, Avi Kivity wrote:
>> This not-for-merging patchset exposes an emulated version 1 architectural
>> performance monitoring unit to KVM guests.  The PMU is emulated using
>> perf_events, so the host kernel can multiplex host-wide, host-user, and the
>> guest on available resources.
>>
>> Caveats:
>> - counters that have PMI (interrupt) enabled stop counting after the
>>   interrupt is signalled.  This is because we need one-shot samples
>>   that keep counting, which perf doesn't support yet
>> - some combinations of INV and CMASK are not supported
>> - counters keep on counting in the host as well as the guest
>> - the RDPMC instruction and CR4.PCE bit are not yet emulated
>> - there is likely a bug in the implementation; running 'perf top' in
>>   a guest that spends 80% of its time in userspace shows perf itself
>>   as consuming almost all cpu
>>
>> perf maintainers: please consider the first three patches for merging (the
>> first two make sense even without the rest).  If you're familiar with the Intel
>> PMU, please review patch 5 as well - it effectively undoes all your work
>> of abstracting the PMU into perf_events by unabstracting perf_events into what
>> is hoped is a very similar PMU.
> 
> Gaah, I was just about to submit a talk about PMU virtualization for KVM
> Forum :)
> 
> Anyway, I thought about a paravirt-approach instead of implementing a
> real PMU... But there are certainly good reasons for both.

Paravirt is taking away the pressure from CPU vendors to do their virt
extensions properly - and doesn't help with unmodifiable OSes.

Jan

-- 
Siemens AG, Corporate Technology, CT T DE IT 1
Corporate Competence Center Embedded Linux