Re: [RFC v2] perf: Rewrite core context handling

[Ravi Bangoria]

On 27-Jun-22 9:48 AM, Ravi Bangoria wrote:
> 
> On 13-Jun-22 8:05 PM, Peter Zijlstra wrote:
>>
>>
>> Right, so sorry for being incredibly tardy on this. Find below the
>> patch fwd ported to something recent.
>>
>> I'll reply to this with fixes and comments.
> 
> Thanks! I've resumed on this but my mind has lost all the context so
> it might take a while for me to reply to your comments. Please bear
> with me if I'm bit slow.

Sorry, it took a while for me to get started on it. Anyways, thanks for
providing fixes. I applied those and ran some tests on AMD Milan machine:

- Built in perf tests ran fine without any issues
- perf_event_tests reported one BUG_ON() and one WARN_ON(). I'll work on
  fixing those.
- Ran perf fuzzer for almost a day. It reported one softlockup but system
  recovered from it and later it reported one hardlockup but unfortunately
  my config had HARDLOCKUP_PANIC set and thus couldn't confirm whether that
  hardlockup was recoverable or not. Anyway, system was running pretty much
  fine until then.
- No lockdep warnings were observed in any of the tests.

I'll work on verifying functionality changes.

Thanks,
Ravi