Re: [PATCH] KVM: halt-polling: poll if emulated lapic timer will fire soon

From: Christian Borntraeger
Date: Thu May 19 2016 - 07:56:16 EST


On 05/19/2016 01:48 PM, Wanpeng Li wrote:
> 2016-05-19 19:42 GMT+08:00 Christian Borntraeger <borntraeger@xxxxxxxxxx>:
>> On 05/19/2016 01:35 PM, Wanpeng Li wrote:
>>> 2016-05-19 19:23 GMT+08:00 Christian Borntraeger <borntraeger@xxxxxxxxxx>:
>>>> On 05/19/2016 11:26 AM, Wanpeng Li wrote:
>>>>
>>>> I think in general a good idea to poll if a timer will expire soon.
>>>>
>>>> Some patch comments:
>>>>
>>>> Same for all non-x86 archs:
>>>>> +static inline unsigned int kvm_arch_timer_remaining(struct kvm_vcpu *vcpu) {}
>>>>
>>>> A function returning int, without a return statement?
>>>> That gives at least a compiler warning.
>>>
>>> How about return 0 for all non-x86 archs?
>>
>> We will provide an s390 implementation soon, but until then a proper
>> default would be good.
>>
>> [....]
>>>>> + if (vcpu->halt_poll_ns || (remaining < halt_poll_ns_base)) {
>>
>> but then remaining is 0 and the 2nd condition will always be true, no?
>
> Nice catch!
>
> How about something like below:
>
> + if (vcpu->halt_poll_ns ||
> + (remaining != 0 && remaining < halt_poll_ns_base)) {

Maybe just use -1UL to have a "will never expire" and change the return value into u64
while changing that.