[libvirt] [PATCH v2 1/3] Add per-guest S3/S4 state configuration

Martin Kletzander mkletzan at redhat.com
Fri Aug 3 07:45:23 UTC 2012


On 08/03/2012 07:16 AM, Doug Goldstein wrote:
> On Thu, Aug 2, 2012 at 3:36 PM, Eric Blake <eblake at redhat.com> wrote:
>> On 08/02/2012 06:05 AM, Martin Kletzander wrote:
>>> There is a new <pm/> element implemented that can control what ACPI
>>> sleeping states will be advertised by BIOS and allowed to be switched
>>> to by libvirt. The default keeps defaults on hypervisor, otherwise
>>> forces chosen setting.
>>
>> You are proposing /domain/pm; but we also have /domain/os/bios, would
>> this be better as a subelement /domain/os/bios/pm, since it is related
>> to bios options?
> 
> I would say that /domain/os/bios/pm isn't the correct place because
> /domain/os relates only to OS booting while Power Management is
> outside of the scope of booting. In fact /domain/os/bios simply
> controls whether the BIOS output will be over VGA or over serial.
> Features specific to the system are typically kept as top level items,
> which matches /domain/pm. I would however say that using the terms
> ACPI, s3 and s4 isn't good design for the XML. You can leave that to
> documentation but /domain/pm/suspend-to-mem and
> /domain/pm/suspend-to-disk are better so that the same XML structure
> can be used for PPC. They have concepts similar to s3 and s4 (I only
> briefly Googled this) but they're named completely differently. The
> only under place I could see this would be /domain/features/acpi. If
> its there then s3 and s4 potentially make sense, but then we're left
> without a place for PPC to tie in.
> 

Thanks both of you for the tips. I also see I haven't added this to the
documentation, which I have to fix in next version of the patch. I have
no problem adapting the naming in XML to whatever will suit you. I was
thinking to put this into /domain/sysinfo, but it looked like it doesn't
fit there and I couldn't make it tristate in /domain/features, so that's
why I've chosen this.




More information about the libvir-list mailing list