Re: kvm guest: hrtimer: interrupt too slow

[Michael Tokarev <mjt@tls.msk.ru>]

Marcelo Tosatti wrote:
[]
> You should see min_delta_ns increase to a much smaller value, hopefully in
> the 2000-10000 range.

That's what I see:

Oct  4 16:13:09 paltus kernel: [boot]
Oct  5 12:35:51 paltus kernel: hrtimer: interrupt too slow, forcing clock min delta to 1500 ns
Oct  5 12:47:29 paltus kernel: hrtimer: interrupt too slow, forcing clock min delta to 2250 ns

I tried to stress-load the machine yesterday (Oct-4), including massive
network usage (I think it's the network that triggers the issue most
reliable), but wasn't able to trigger it.  Today, at Monday, at normal
system load, it triggered "by its own", increasing min_delta to 50%
as in the patch.  The system runs smoothly so far, but it's to be
expected (till it increases min delta to some 10-digit number :).

But I'm unsure still if that's the right approach...  The root
cause seems to be the problem, and this patch just tries to
work around it in a better-than-before way.  But it might be
just me... ;)

I'll continue watching it, since the machine(s) in question are
important for us.

Thanks!

/mjt