lkml.org 
[lkml]   [2012]   [Mar]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] xen/acpi-processor: C and P-state driver that uploads said data to hypervisor.
>>> On 12.03.12 at 16:07, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> wrote:
> On Mon, Mar 12, 2012 at 11:01:49AM +0000, Jan Beulich wrote:
>> >>> On 10.03.12 at 17:05, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> wrote:
>> Rather than arbitrarily limiting this, you could call into Xen at startup
>> and scan over all CPUs' ACPI IDs to find the maximum. I think that
>> would even cover statically declared hotplug ones, but adding some
>> slack may still be necessary to cover dynamic hotplug ones (implying
>> that pCPU hotplug patches will make it in at some point).
>
> This would be the XENPF_get_cpuinfo call right? I do plan on looking at the

Yes.

> pCPU hotplug, but I think I need some fancy hardware to test it correctly?

No, you don't really - for testing these code paths you can soft offline
and then online back CPUs in Xen.

>> > + register_hotcpu_notifier(&xen_cpu_notifier);
>>
>> This is pointless - you'd get notified of vCPU-s arrival/departure only.
>> Without pCPU hotplug code in place, there's just nothing you can (and
>> need to) do here.
>
> I hadn't actually looked in details on the pCPU hotplug code to see how it
> works.
>
> I presume I need special hardware for this to work as well as the ACPI
> is involved in triggering the hotplug CPU up calls?

For exercising the ACPI code paths - yes, you'll need suitable
hardware. But for just exercising the notifier logic, as said above,
you don't.

Jan



\
 
 \ /
  Last update: 2012-03-12 16:29    [from the cache]
©2003-2011 Jasper Spaans