[libvirt] [PATCHv2 0/5] Basic Enablement of s390

[Viktor Mihajlovski]

On 06/25/2012 04:43 PM, Michal Privoznik wrote:
> On 25.06.2012 15:54, Daniel Veillard wrote:
>> On Mon, Jun 25, 2012 at 09:43:41PM +0800, Daniel Veillard wrote:
>>> On Mon, Jun 25, 2012 at 03:08:09PM +0200, Viktor Mihajlovski wrote:
>>>> On 06/25/2012 10:41 AM, Daniel Veillard wrote:
>>>>> On Fri, Jun 22, 2012 at 01:50:09PM +0200, Viktor Mihajlovski wrote:
>>>>>> Resending as a thread, same content. Please ignore previous submission.
>>>>>>
>>>>>> The first series of patches aimed to enable libvirt to manage
>>>>>> qemu/kvm domains hosted on the System z platform.
>>>>>> The patches in this series address s390 specifics of node and hypervisor.
>>>>>>
>>>>>> Patches 2,3 and 5 are written by Thang Pham<thang.pham at us.ibm.com>
>>>>>> the others are my doing. Since they all should go together, I am
>>>>>> sending in Thang's as well, hoping to facilitate the review a bit
>>>>>> this way.
>>>>>>
>>>>>> Thang Pham (3):
>>>>>>    S390: CPU support for s390(x)
>>>>>>    S390: Fixed Parser for /proc/cpuinfo needs to be adapted for your
>>>>>>      architecture
>>>>>>    S390: Added sysinfo for host on s390x.
>>>>>>
>>>>>> Viktor Mihajlovski (2):
>>>>>>    S390: Override QEMU_CAPS_NO_ACPI for s390x
>>>>>>    S390: Fixed core identification for s390
>>>>>>
>>>>>>   src/Makefile.am              |    1 +
>>>>>>   src/cpu/cpu.c                |    2 +
>>>>>>   src/cpu/cpu_s390x.c          |   80 ++++++++++++++++++++++
>>>>>>   src/cpu/cpu_s390x.h          |   31 +++++++++
>>>>>>   src/nodeinfo.c               |   11 +++
>>>>>>   src/qemu/qemu_capabilities.c |    5 ++
>>>>>>   src/util/sysinfo.c           |  154 ++++++++++++++++++++++++++++++++++++++++++
>>>>>>   7 files changed, 284 insertions(+), 0 deletions(-)
>>>>>>   create mode 100644 src/cpu/cpu_s390x.c
>>>>>>   create mode 100644 src/cpu/cpu_s390x.h
>>>>>
>>>>>    In general I like seeing additional platform support. If done well
>>>>> this is an opportunity to improve the code and bust out bug not raised
>>>>> on the main architecture.
>>>>>    I was tempted to get such patch in before the freeze for 0.9.13, but
>>>>> it's a bit too late. So let's try to clean an improve the patch set(s) and
>>>>> push starting next week after the freeze.
>>>>>    Ideally we should make sure we do thing uniformly for s390 s390x ppc64
>>>>> and try to be ready if/when arm64 shows up !
>>>>>
>>>>>    thanks !
>>>>>
>>>>> Daniel
>>>>>
>>>>
>>>> thanks for reviewing the patches, we will send out a cleaned up
>>>> version shortly.
>>>
>>>    Well I reviewed only 2 of the set, lemme try to get a bit further :-)
>>
>>    Actually now that I went though all 5 and since Dan Berrange reviewed
>> them too, it seems to me this code should not generate portability
>> troubles for other platforms (which was my main concern at this point).
>> So I would agree with trying to push that patch set in rc2 if you fix
>> the few issues which were raised, and send the new set within a couple
>> of days.
>>
>>    Thanks !
>>
>> Daniel
>>
>
> Well, I went ahead and cleaned up some nits you guys pointed our or I've
> spotted myself. Viktor, can you please give it a try and send any follow
> up patch if you find anything not working?
>
> Michal
>
Hi Michal,

thanks for fixing the issues :-). I rebased my local branch to upstream 
and everything is working as expected.
Please note that my other patch set (virtio support for s390) is needed 
to start s390 guests with I/O devices. As this touches common areas in 
the qemu driver, I'll hold off until 0.9.13 is out of the door.

-- 

Mit freundlichen Grüßen/Kind Regards
    Viktor Mihajlovski

IBM Deutschland Research & Development GmbH
Vorsitzender des Aufsichtsrats: Martin Jetter
Geschäftsführung: Dirk Wittkopp
Sitz der Gesellschaft: Böblingen
Registergericht: Amtsgericht Stuttgart, HRB 243294