Re: [PATCH RESEND v8 1/2] sched/numa: introduce per-cgroup NUMA locality info

From: çè
Date: Sun Feb 23 2020 - 22:09:12 EST

On 2020/2/21 äå11:28, Peter Zijlstra wrote:
> On Mon, Feb 17, 2020 at 09:23:52PM +0800, çè wrote:
>> FYI, by monitoring locality, we found that the kvm vcpu thread is not
>> covered by NUMA Balancing, whatever how many maximum period passed, the
>> counters are not increasing, or very slowly, although inside guest we are
>> copying memory.
>> Later we found such task rarely exit to user space to trigger task
>> work callbacks, and NUMA Balancing scan depends on that, which help us
>> realize the importance to enable NUMA Balancing inside guest, with the
>> correct NUMA topo, a big performance risk I'll say :-P
> That's a bug in KVM, see:
> ISTR there being newer versions of that patch-set, but I can't seem to
> find them in a hurry.

Aha, that's exactly the problem we saw, will check~

Michael Wang