Re: [PATCH] xen/x86: Adjust stack pointer in xen_sysexit

[Joao Martins <joao.m.martins@oracle.com>]

On 11/16/2015 09:04 PM, Andy Lutomirski wrote:
> On Mon, Nov 16, 2015 at 1:03 PM, Konrad Rzeszutek Wilk
> <konrad.wilk@oracle.com> wrote:
>> On Mon, Nov 16, 2015 at 12:50:19PM -0800, Andy Lutomirski wrote:
>>> On Mon, Nov 16, 2015 at 12:48 PM, Boris Ostrovsky
>>> <boris.ostrovsky@oracle.com> wrote:
>>>> On 11/16/2015 03:22 PM, Borislav Petkov wrote:
>>>>>
>>>>> On Mon, Nov 16, 2015 at 12:11:11PM -0800, Andy Lutomirski wrote:
>>>>>
>>>>>> Are there really multiple feature bits for this stuff?  I'd like to
>>>>>> imagine that the entry code is all either Xen PV or native/PVH/PVHVM
>>>>>> -- i.e. I assumed that PVH works like native for all entries.
>>>>
>>>>
>>>>
>>>> Almost. For PVH we will have a small stub to set up bootparams and such but
>>>> then we jump to startup_{32|64} code.
>>>>
>>>>
>>>>> I just reacted to Boris' statement:
>>>>>
>>>>> "We don't currently have a Xen-specific CPU feature. We could, in
>>>>> principle, add it but we can't replace all of current paravirt patching
>>>>> with a single feature since PVH guests use a subset of existing pv ops
>>>>> (and in the future it may become even more fine-grained)."
>>>>
>>>>
>>>> Actually, nevermind this --- I was thinking about APIC ops and they are not
>>>> pv ops.
>>>>
>>>> Note though that there are other users of pv ops --- lguest and looks like
>>>> KVM (for one op) use them too.
>>>>
>>>
>>> Honestly, I think we should just delete lguest some time soon.  And
>>> KVM uses this stuff so lightly that we shouldn't need all of the pvop
>>> stuff.  (In fact, I'm slowly working on removing KVM_GUEST's
>>> dependency on PARAVIRT.)
>>
>> Even for the pvclock?
>>
>> (sorry for stealing this thread on this subject).
> 
> I don't think that pvclock uses any of the paravirt infrastructure.
> It's just another clock source AFAIK.
> 
Yeah, but pv_time_ops is still used on both Xen and KVM. Even though it looks
that on KVM some of it's used only when pvclock isn't marked as stable (i.e. no
PVCLOCK_TSC_STABLE_BIT).

Joao

> --Andy
>