Re: [Qemu-devel] [PATCH v3] vl.c: Support multiple CPU ranges on -numa option

[Bandan Das]

Wanlong Gao <address@hidden> writes:

> On 06/21/2013 12:02 AM, Bandan Das wrote:
>> Paolo Bonzini <address@hidden> writes:
>> 
>>> Il 20/06/2013 15:26, Eduardo Habkost ha scritto:
>>>> On Thu, Jun 20, 2013 at 11:52:42AM +0200, Paolo Bonzini wrote:
>>>>> Il 20/06/2013 11:30, Igor Mammedov ha scritto:
>>>>>>>>>>>> So, basically the format seemed easier to work with if we are 
>>>>>>>>>>>> thinking 
>>>>>>>>>>>> of using QemuOpts for -numa. Using -cpu rather than cpus probably
>>>>>>>>>>>> makes it less ambiguous as well IMO. However, it's probably not a 
>>>>>>>>>>>> good idea
>>>>>>>>>>>> if the current syntax is well established ?
>>>>>>>>
>>>>>>>> libvirt uses the "cpus" option already, so we have to keep it working.
>>>>>> Sure, we can leave it as it's now for some time while a new interface is
>>>>>> introduced/adopted. And than later deprecate "cpus".
>>>>>
>>>>> So, you used a new name because the new behavior of "-numa
>>>>> node,cpus=1-2,cpus=3-4" would be incompatible with the old.
>>>>
>>>> I don't think anybody uses "cpus=1-2,cpus=3-4" today, so I believe we
>>>> can change its behavior. The problem was to get agreement on the syntax
>>>> to represent multiple CPU ranges.
>>>
>>> Ok.  I think almost everyone agreed on "cpus=1-2,cpus=3-4", which is
>>> basically what Bandan's patch does minus s/cpu/cpus/.  It matches what
>>> already happens with other options (SLIRP), so it's hardly surprising.
>> 
>> Good, so should I spin a new version with "cpus" ?
>
> I already merged your patch to my patch set "Add support for binding guest 
> numa nodes to host numa nodes"
> since I should base on that.
>
> Thanks,
> Wanlong Gao

Oh, great! Thank you for taking care of the "cpus" change.


>> 
>> Also note that this patch actually doesn't add any extra code to support 
>> multiple cpus arguments. It all happens automatically as part of conversion 
>> to
>> QemuOpts. So, if we need to revisit the syntax later, we can always do that.
>> 
>> Bandan
>>> Let's go on with that.
>>>
>>> Paolo
>>>
>>>>> Personally I don't think that's a problem, but I remember a long
>>>>> discussion in the past.  Igor/Eduardo, do you remember the conclusions?
>>>>
>>>> I don't remember seeing the discussion reach any conclusion,
>>>> unfortunately.
>>>>
>> 
>>