Re: [Qemu-devel] [PATCH 7/7] Implement cpu hot-add using device_add monitor command

[Andreas Färber]

Am 16.02.2012 10:33, schrieb Igor Mammedov:
> On 02/16/2012 12:35 AM, Anthony Liguori wrote:
>> On 02/15/2012 05:16 PM, Igor Mammedov wrote:
>>> Adds ability to hot-add cpus if guest was started with options
>>> -smp X,maxcpus=Y
>>> where X< Y.
>>> For simplicity sake this implementation doesn't allow to add specific
>>> cpu but rather it adds the next not yet plugged cpu.
>>>
>>> For adding cpu just execute following command in monitor:
>>> device_add cpu-pc
>>> If guest was started with a specific cpu model, then add option model
>>> to cmonitor command. for example:
>>> device_add cpu-pc,model="host"
>>>
>>> Signed-off-by: Igor Mammedov<address@hidden>
>>
>> I don't see how the code matches this description, but unless you've
>> got qdev-ification patches in the earlier ones (haven't made the list
>> yet),
>> device_add is not the right place for this.
> Indeed, I've attempted qdev-ify cpu first (patch 2/7)
> 
>>
>> There's no such thing as "cpu-pc" either. You should start with a
>> cpu_add command.
> "cpu-pc" is coming from 2/7 where cpu was qdev-ifed and I needed to call
> it somehow.
> Looking at other attempts to qdev-ify cpus i.e. "cpu-ppc", "cpu-mips",
> I've just
> followed trend. Is there any suggestions on how it should be called?

There's a CPU QOM'ification series of mine on the list with which your
series conflicts conceptually. Unfortunately due to name conflicts I can
no longer incrementally post a v4 so I am converting all targets now one
after another, x86 will by the current name scheme get an X86CPU
(TYPE_X86_CPU).

Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg