Re: [Xen-devel] [PATCH v4 13/17] x86/VPMU: Add privileged PMU mode

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 04.02.14 at 16:53, Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx> wrote:
> On 02/04/2014 06:31 AM, Jan Beulich wrote:
>>>>> On 21.01.14 at 20:08, Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx> wrote:
>>> +            gregs = &v->arch.vpmu.xenpmu_data->pmu.r.regs;
>>> +            gregs->cs = cs;
>> And now you store a NUL selector (i.e. just the RPL bits) into the
>> output field?
>>>           }
>>> -        else if ( !is_control_domain(current->domain) &&
>>> -                 !is_idle_vcpu(current) )
>>> +        else
>>>           {
>>> -            /* PV guest */
>>> +            /* HVM guest */
>>> +            struct segment_register cs;
>>> +
>>>               gregs = guest_cpu_user_regs();
>>>               memcpy(&v->arch.vpmu.xenpmu_data->pmu.r.regs,
>>>                      gregs, sizeof(struct cpu_user_regs));
>>> +
>>> +            hvm_get_segment_register(current, x86_seg_cs, &cs);
>>> +            gregs = &v->arch.vpmu.xenpmu_data->pmu.r.regs;
>>> +            gregs->cs = cs.attr.fields.dpl;
>> And here too? If that's intended, a code comment is a must.
> 
> This is HVM-only path, PVH or PV don't go here so cs should be valid.

Isn't the reply of mine a few lines up in PV code? And why would
the selector being wrong for HVM be okay?

Jan


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