Re: [PATCH v2 08/13] x86/PMU: Interface for setting PMU mode and flags

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

On 09/25/2013 10:11 AM, Jan Beulich wrote:
>>>> On 20.09.13 at 11:42, Boris Ostrovsky <boris.ostrovsky@oracle.com> wrote:
>> --- a/xen/include/public/xen.h
>> +++ b/xen/include/public/xen.h
>> @@ -101,6 +101,7 @@ DEFINE_XEN_GUEST_HANDLE(xen_ulong_t);
>>   #define __HYPERVISOR_kexec_op             37
>>   #define __HYPERVISOR_tmem_op              38
>>   #define __HYPERVISOR_xc_reserved_op       39 /* reserved for XenClient */
>> +#define __HYPERVISOR_xenpmu_op            40
>>   
>>   /* Architecture-specific hypercall definitions. */
>>   #define __HYPERVISOR_arch_0               48
> I wonder whether you wouldn't better use an arch-specific hypercall
> number here - there's really very little that's generic in what I have
> seen so far.

Some of xenpmu commands (querying PMU mode/flags, PMU 
initialization/de-initialization)
seem generic to me. They may call arch-specific routines.

>> +/* Parameters structure for HYPERVISOR_xenpmu_op call */
>> +struct xenpmu_params {
>> +    /* IN/OUT parameters */
>> +    union {
>> +        struct version {
>> +            uint8_t maj;
>> +            uint8_t min;
>> +        } version;
>> +        uint64_t pad;
>> +    };
>> +    union {
>> +        uint64_t val;
>> +        void *valp;
> Without there also being a handle here I can't see how you could
> make use of the pointer.

Tha't really a placeholder, valp is not used anywhere. I can replace it
with XEN_GUEST_HANDLE.

-boris