Messages in this thread | | | Date | Thu, 09 Dec 2010 12:07:32 -0500 | From | Rik van Riel <> | Subject | Re: [RFC PATCH 3/3] kvm: use yield_to instead of sleep in kvm_vcpu_on_spin |
| |
On 12/09/2010 05:28 AM, Avi Kivity wrote: > On 12/09/2010 12:38 AM, Rik van Riel wrote: >> >>>> - /* Sleep for 100 us, and hope lock-holder got scheduled */ >>>> - expires = ktime_add_ns(ktime_get(), 100000UL); >>>> - schedule_hrtimeout(&expires, HRTIMER_MODE_ABS); >>>> + if (first_round&& last_boosted_vcpu == kvm->last_boosted_vcpu) { >>>> + /* We have not found anyone yet. */ >>>> + first_round = 0; >>>> + goto again; >>> >>> Need to guarantee termination. >> >> We do that by setting first_round to 0 :) >> >> We at most walk N+1 VCPUs in a VM with N VCPUs, with >> this patch. >> > > Right. May be clearer by using a for () loop instead of the goto.
And open coding kvm_for_each_vcpu ?
Somehow I suspect that won't add to clarity...
-- All rights reversed
| |