On 09/02/2015 22:00, Christian Borntraeger wrote:
> I can confirm that this also helps uperf with a 1/1 byte round trip work load
> between guests on s390. And I can confirm the higher CPU load. This is
> normally
> a no-go for the typical s390 users, which utilize their systems as much as
> po
Am 06.02.2015 um 13:48 schrieb Paolo Bonzini:
> This patch introduces a new module parameter for the KVM module; when it
> is present, KVM attempts a bit of polling on every HLT before scheduling
> itself out via kvm_vcpu_block.
>
> This parameter helps a lot for latency-bound workloads---in parti
On Fri, Feb 6, 2015 at 4:48 AM, Paolo Bonzini wrote:
> This patch introduces a new module parameter for the KVM module; when it
> is present, KVM attempts a bit of polling on every HLT before scheduling
> itself out via kvm_vcpu_block.
>
[...]
>
> Signed-off-by: Paolo Bonzini
> ---
Looks good, t
2015-02-09 17:10+0100, Paolo Bonzini:
> On 09/02/2015 16:21, Radim Krčmář wrote:
> > 2015-02-06 13:48+0100, Paolo Bonzini:
> >> +TRACE_EVENT(kvm_vcpu_wakeup,
> >> + TP_PROTO(__u64 ns, bool waited),
> >
> > (__u64 is preferred here?)
>
> Preferred to what?
To 'u64'. (The header file shouldn
On 09/02/2015 16:21, Radim Krčmář wrote:
> 2015-02-06 13:48+0100, Paolo Bonzini:
> [...]
>> Signed-off-by: Paolo Bonzini
>> ---
>
> Reviewed-by: Radim Krčmář
>
> Noticed changes since RFC:
> - polling is used in more situations
> - new tracepoint
> - module parameter in nanoseconds
> - pr
2015-02-06 13:48+0100, Paolo Bonzini:
[...]
> Signed-off-by: Paolo Bonzini
> ---
Reviewed-by: Radim Krčmář
Noticed changes since RFC:
- polling is used in more situations
- new tracepoint
- module parameter in nanoseconds
- properly handled time
- no polling with overcommit
> diff --git a
On 02/09/2015 05:06 PM, Paolo Bonzini wrote:
On 09/02/2015 09:22, Xiao Guangrong wrote:
On 02/06/2015 08:48 PM, Paolo Bonzini wrote:
+unsigned int halt_poll_ns = 0;
+module_param(halt_poll_ns, uint, S_IRUGO | S_IWUSR);
+
Can we make this parameter be changeable? So that we can tune it
On 09/02/2015 09:22, Xiao Guangrong wrote:
>
>
> On 02/06/2015 08:48 PM, Paolo Bonzini wrote:
>
>>
>> +unsigned int halt_poll_ns = 0;
>> +module_param(halt_poll_ns, uint, S_IRUGO | S_IWUSR);
>> +
>
> Can we make this parameter be changeable? So that we can tune it
> on the fly.
It is changea
On 02/06/2015 08:48 PM, Paolo Bonzini wrote:
+unsigned int halt_poll_ns = 0;
+module_param(halt_poll_ns, uint, S_IRUGO | S_IWUSR);
+
Can we make this parameter be changeable? So that we can tune it
on the fly.
finish_wait(&vcpu->wq, &wait);
+ cur = ktime_get();
We can move
This patch introduces a new module parameter for the KVM module; when it
is present, KVM attempts a bit of polling on every HLT before scheduling
itself out via kvm_vcpu_block.
This parameter helps a lot for latency-bound workloads---in particular
I tested it with O_DSYNC writes with a battery-bac
10 matches
Mail list logo
