On 02/19/2016 08:09 AM, Peter Krempa wrote:
On Fri, Feb 19, 2016 at 07:19:38 -0500, John Ferlan wrote:
>
>
> On 02/17/2016 11:25 AM, Peter Krempa wrote:
>> VM startup and CPU hotplug always set the affinity regardless of cgroups
>> support. Use the same approach for the pinning API.
>> ---
>> src/qemu/qemu_driver.c | 7 +++----
>> 1 file changed, 3 insertions(+), 4 deletions(-)
>>
>> diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
>> index 24c1ca4..d1e5188 100644
>> --- a/src/qemu/qemu_driver.c
>> +++ b/src/qemu/qemu_driver.c
>> @@ -5068,12 +5068,11 @@ qemuDomainPinVcpuFlags(virDomainPtr dom,
>> goto endjob;
>> if (qemuSetupCgroupCpusetCpus(cgroup_vcpu, pcpumap) < 0)
>> goto endjob;
>> - } else {
>> - if (virProcessSetAffinity(qemuDomainGetVcpuPid(vm, vcpu),
>> - pcpumap) < 0)
>> - goto endjob;
>> }
>>
>> + if (virProcessSetAffinity(qemuDomainGetVcpuPid(vm, vcpu), pcpumap) <
0)
>> + goto endjob;
>> +
>> virBitmapFree(vcpuinfolive->cpumask);
>> vcpuinfolive->cpumask = pcpumaplive;
>> pcpumaplive = NULL;
>>
>
> Why not the same for qemuDomainPinEmulator and qemuDomainPinIOThread?
I'm too focused on the cpus in this series apparently ...
> Since qemuProcessSetEmulatorAffinity and qemuProcessSetupIOThread do the
> same as qemuProcessSetupVcpu.
Are you okay with separate patches for this?
Yes that's fine
John