[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v4 00/15] PVH VCPU hotplug support



>>> On 29.11.16 at 17:40, <boris.ostrovsky@xxxxxxxxxx> wrote:
> On 11/29/2016 11:11 AM, Jan Beulich wrote:
>>>>> On 29.11.16 at 16:33, <boris.ostrovsky@xxxxxxxxxx> wrote:
>>> This series adds support for ACPI-based VCPU hotplug for unprivileged
>>> PVH guests.
>>>
>>> Changes in v4:
>>> * Replaced XEN_DOMCTL_set_avail_vcpus with XEN_DOMCTL_acpi_access,
>>>   toolstack is expected to issue two ACPI "writes" to trigger the hotplug
>>> * Moved pm1a registers from PMTState to hvm_domain. gpe0 registers are
>>>   added there.
>>> * Both pm1a and gpe0 registers are now represented as uint16_t for status
>>>   and enable (and as result no new public macros for those registers' length
>>>   are needed). Ths is partly to avoid changes in HVM code (i.e. pmtimer)
>>>   and partly to simplify code. I don't expect larger registers will be
>>>   needed any time soon.
>>> * ACPI handling now lives in xen/arch/x86/hvm/acpi.c
>>> * Moved definition of XEN_ACPI_CPU_MAP/XEN_GPE0_CPUHP_BIT from ioreq.h
>>>   to public/arch-x86/xen.h. This is the best I could come up with but
>>>   ioreq is the wrong place, now that everything happens in acpi.c
>> Without having looked at the patch(es) - why not e.g.
>> public/arch-x86/hvm/acpi.h?
> 
> It's just 3 defines so I am not sure it's worth a whole new include file.

Are you reasonably convinced that nothing else will want to be put
there soon? There were quite a few other ACPI additions in the
previous version of your series iirc, some or all of which may also
better fit here ...

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.