Re: [Xen-devel] Is: axe read_tscp pvops call. Was: Re: [RFC] ACPI S3 and Xen (suprisingly small\!).

[David Vrabel <david.vrabel@citrix.com>]

On 17/10/12 17:54, Konrad Rzeszutek Wilk wrote:
> On Wed, Oct 17, 2012 at 09:50:11AM -0700, H. Peter Anvin wrote:
>> On 10/17/2012 09:10 AM, Konrad Rzeszutek Wilk wrote:
>>> On Wed, Oct 17, 2012 at 09:03:12AM -0700, H. Peter Anvin wrote:
>>>> On 10/17/2012 06:49 AM, Konrad Rzeszutek Wilk wrote:
>>>>>
>>>>> Note: These are the other patches that went in 3.7-rc1:
>>>>> xen/bootup: allow {read|write}_cr8 pvops call [https://lkml.org/lkml/2012/10/10/339]
>>>>> xen/bootup: allow read_tscp call for Xen PV guests. [https://lkml.org/lkml/2012/10/10/340]
>>>>>
>>>>
>>>> So WTF do we have a read_tscp PV call?  Again, if there isn't a user
>>>> we should just axe it...
>>>
>>> Let me spin off a patch to see if that can be done.
>>>
>>
>> Could you do an audit for other pvops calls that have no users?  If
>> the *only* user is lguest, we should talk about it, too...
> 
> I can do that - but I don't want to be hasty here. There is a bit of
> danger here - for example the read_pmc (or read_tsc) is not in use right
> now. But it might be when one starts looking at making perf be able to
> analyze the hypervisor (hand-waving the implementation details). So while
> removing read_pmc now sounds good, it might be needed in the future.

I don't see any reason why would ever need a PV-specific implementation
of either read_pmc or read_tsc.  And I certainly agree with hpa that
leaving them around 'just in case' isn't useful.

As for 'perf', since Xen already provides a virtual PMU for HVM guests
It's not clear why we would spend the effort to implement another
mechanism for PV guests (instead of using the virtual PMU from a PVH guest).

David