Re: [PATCH v6 13/19] x86/VPMU: Handle PMU interrupts for PV guests

["Jan Beulich" <JBeulich@suse.com>]

>>> On 22.05.14 at 19:25, <boris.ostrovsky@oracle.com> wrote:
> On 05/22/2014 11:30 AM, Jan Beulich wrote:
>>
>>> @@ -99,14 +120,87 @@ int vpmu_do_rdmsr(unsigned int msr, uint64_t 
> *msr_content)
>>>           return 0;
>>>   
>>>       if ( vpmu->arch_vpmu_ops && vpmu->arch_vpmu_ops->do_rdmsr )
>>> -        return vpmu->arch_vpmu_ops->do_rdmsr(msr, msr_content);
>>> +    {
>>> +        int ret = vpmu->arch_vpmu_ops->do_rdmsr(msr, msr_content);
>>> +
>>> +        if ( !is_hvm_domain(current->domain) &&
>>> +            (current->arch.vpmu.xenpmu_data->pmu_flags & PMU_CACHED) )
>> Wouldn't the same comment as in WRMSR handling apply here too?
>> If so, either replicate it or add a brief comment referring to the
>> other one.
> 
> 
> Yes, the comment is applicable here too and the next patch merges the 
> two routines.

When I saw that the next patch merges the two, I realized that my
request here was sort of pointless. Question in such cases of course
is whether the merging wouldn't then better be done up front.

Jan