Re: [PATCH v2 1/3] x86/vpmu: Add get/put_vpmu() and VPMU_AVAILABLE

[Boris Ostrovsky <boris.ostrovsky@oracle.com>]

On 02/21/2017 06:00 AM, Andrew Cooper wrote:
> On 17/02/17 17:40, Boris Ostrovsky wrote:
>> @@ -509,15 +498,63 @@ void vpmu_initialise(struct vcpu *v)
>>      if ( ret )
>>          printk(XENLOG_G_WARNING "VPMU: Initialization failed for %pv\n", v);
>>  
>> -    /* Intel needs to initialize VPMU ops even if VPMU is not in use */
>> -    if ( !is_priv_vpmu &&
>> -         (ret || (vpmu_mode == XENPMU_MODE_OFF) ||
>> -          (vpmu_mode == XENPMU_MODE_ALL)) )
>> +    return ret;
>> +}
>> +
>> +static void get_vpmu(struct vcpu *v)
>> +{
>> +    spin_lock(&vpmu_lock);
>> +
>> +    /*
>> +     * Count active VPMUs so that we won't try to change vpmu_mode while
>> +     * they are in use.
>> +     * vpmu_mode can be safely updated while dom0's VPMUs are active and
>> +     * so we don't need to include it in the count.
>> +     */
>> +    if ( !is_hardware_domain(v->domain) &&
>> +        (vpmu_mode & (XENPMU_MODE_SELF | XENPMU_MODE_HV)) )
>> +    {
>> +        vpmu_count++;
>> +        vpmu_set(vcpu_vpmu(v), VPMU_AVAILABLE);
>> +    }
>> +    else if ( is_hardware_domain(v->domain) &&
>> +              (vpmu_mode != XENPMU_MODE_OFF) )
>> +        vpmu_set(vcpu_vpmu(v), VPMU_AVAILABLE);
> Why do we recalculate AVAILABLE on every get()?
>
> (In the absence of toolstack control from the CPUID side), surely it
> should be set (or not) exactly once during domain creation, and be
> unchanged thereafter?

Yes, that's exactly what's happening -- get_vpmu() is only called from
vcpu_initialise() -> vpmu_initialise() and so we set VPMU_AVAILABLE only
once for a VCPU.

I am not sure what you are asking.

-boris


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel