Re: [Xen-devel] [PATCH v5 1/4] x86/HVM: cancel emulation when register state got altered

[Jan Beulich <jbeulich@xxxxxxxx>]

On 03.03.2020 15:25, Durrant, Paul wrote:
>> -----Original Message-----
>> From: Jan Beulich <jbeulich@xxxxxxxx>
>> Sent: 03 March 2020 14:21
>> To: Durrant, Paul <pdurrant@xxxxxxxxxxxx>
>> Cc: xen-devel@xxxxxxxxxxxxxxxxxxxx; Andrew Cooper
>> <andrew.cooper3@xxxxxxxxxx>; Roger Pau Monné <roger.pau@xxxxxxxxxx>; Wei
>> Liu <wl@xxxxxxx>; Paul Durrant <paul@xxxxxxx>
>> Subject: RE: [EXTERNAL][Xen-devel] [PATCH v5 1/4] x86/HVM: cancel
>> emulation when register state got altered
>>
>> On 03.03.2020 14:16, Durrant, Paul wrote:
>>>> -----Original Message-----
>>>> From: Xen-devel <xen-devel-bounces@xxxxxxxxxxxxxxxxxxxx> On Behalf Of
>> Jan
>>>> Beulich
>>>> Sent: 03 March 2020 10:17
>>>> To: xen-devel@xxxxxxxxxxxxxxxxxxxx
>>>> Cc: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>; Roger Pau Monné
>>>> <roger.pau@xxxxxxxxxx>; Wei Liu <wl@xxxxxxx>; Paul Durrant
>> <paul@xxxxxxx>
>>>> Subject: [EXTERNAL][Xen-devel] [PATCH v5 1/4] x86/HVM: cancel emulation
>>>> when register state got altered
>>>>
>>>> Re-execution (after having received data from a device model) relies on
>>>> the same register state still being in place as it was when the request
>>>> was first sent to the device model. Therefore vCPU state changes
>>>> effected by remote sources need to result in no attempt of re-
>> execution.
>>>> Instead the returned data is to simply be ignored.
>>>>
>>>> Note that any such asynchronous state changes happen with the vCPU at
>>>> least paused (potentially down and/or not marked ->is_initialised), so
>>>> there's no issue with fiddling with register state behind the actively
>>>> running emulator's back. Hence the new function doesn't need to
>>>> synchronize with the core emulation logic.
>>>>
>>>> Suggested-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
>>>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
>>>
>>> Need we be concerned with any page-split I/O here? That may manifest as
>>> two separate emulations and AFAICT it would be possible for only the
>>> second part to be aborted by this change.
>>
>> I'm not sure whether e.g. INIT is recognized only on insn boundaries.
>> I.e. this may not be that different from real hardware behavior. _If_
>> we were to take care of this, how would you envision undoing the
>> first part of such an access, most notably when the access has side
>> effects?
> 
> I wasn't thinking of undoing... I was more thinking that vcpu_pause()
> ought to defer until an in-progress emulation has fully completed.

Hmm, at the first glance this looks ugly/fragile to arrange for. I'm
having a hard time thinking of a rough sketch of how such could be
made work, in particular without blocking the vcpu_pause() itself
for too long.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel