[PATCH 0/13] Hypervisor-mode KVM on POWER7

[Alexander Graf]

On 21.05.2011, at 19:00, Alexander Graf wrote:

> 
> On 21.05.2011, at 18:41, Alexander Graf wrote:
> 
>> 
>> On 19.05.2011, at 07:22, Paul Mackerras wrote:
>> 
>>> On Tue, May 17, 2011 at 02:42:08PM +0300, Avi Kivity wrote:
>>>> On 05/17/2011 02:38 PM, Alexander Graf wrote:
>>>>>> 
>>>>>> What would be the path for these patches to get upstream?  Would this
>>>>>> stuff normally go through Avi's tree?  There is a bit of a
>>>>>> complication in that they are based on Ben's next branch.  Would Avi
>>>>>> pull Ben's next branch, or would they go in via Ben's tree?
>>>>> 
>>>>> Usually the ppc tree gets merged into Avi's tree and goes on from
>>>>> there. When we have interdependencies, we can certainly do it
>>>>> differently though. We can also shove them through Ben's tree this
>>>>> time around, as there are more dependencies on ppc code than KVM
>>>>> code.
>>>>> 
>>>> 
>>>> Yes, both options are fine.  If it goes through kvm.git I can merge
>>>> Ben's tree (provided it is append-only) and apply the kvm-ppc
>>>> patches on top.
>>> 
>>> OK, the easiest thing is for them to go via Ben's tree, I think, since
>>> they depend so much on other stuff in Ben's tree.
>>> 
>>> Alex, could you give Ben an acked-by for patches 1-8 of the series?
>>> There haven't been any changes requested for them.
>> 
>> With ben's tree merged into avi's tree and your patches applied on top, MOL breaks. I'll have to track down why exactly though.
> 
> In fact, qemu-system-ppc64 -enable-kvm also breaks.

Only ben's tree merged into avi's tree works fine. So it's definitely one of your patches from 1-8 (except for 2,3 which are already in Ben's tree).


Alex