Re: [PATCH v2 1/4] xl: Add support for ignore_msrs option

[Julien Grall <julien@xen.org>]

On 22/01/2021 18:28, Boris Ostrovsky wrote:
> 
> 
> On 1/22/21 4:52 AM, Julien Grall wrote:
>> Hi Boris,
>>
>> On 20/01/2021 22:49, Boris Ostrovsky wrote:
>>> This option allows guest administrator specify what should happen when
>>> guest accesses an MSR which is not explicitly emulated by the hypervisor.
>>>
>>> Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
>>> ---
>>>    docs/man/xl.cfg.5.pod.in         | 20 +++++++++++++++++++-
>>>    tools/libs/light/libxl_types.idl |  7 +++++++
>>>    tools/xl/xl_parse.c              |  7 +++++++
>>>    3 files changed, 33 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/docs/man/xl.cfg.5.pod.in b/docs/man/xl.cfg.5.pod.in
>>> index c8e017f950de..96ce97c42cab 100644
>>> --- a/docs/man/xl.cfg.5.pod.in
>>> +++ b/docs/man/xl.cfg.5.pod.in
>>> @@ -2044,7 +2044,25 @@ Do not provide a VM generation ID.
>>>    See also "Virtual Machine Generation ID" by Microsoft:
>>>    L<https://docs.microsoft.com/en-us/windows/win32/hyperv_v2/virtual-machine-generation-identifier>
>>>    -=back
>>> +=over
>>> +
>>> +=item B<ignore_msrs="STRING">
>>> +
>>> +Determine hypervisor behavior on accesses to MSRs that are not emulated by the hypervisor.
>>
>> The description of the feature looks very x86 focus. Yet, it seems to be defined as a generic one.
>>
>> Could you clarify whether this is intended to be re-usable by other architectures?
> 
> 
> x86 only. I'll add appropriate note.

Thanks. In which case, I would suggest to move the definition of 
ignore_msrs in the idle to an x86 specific structure. Maybe "arch_x86"?

I noticed that none exists today, but we could duplicate that "arch_arm" 
one.

Wei, Ian, what do you think?

Cheers,

-- 
Julien Grall