[libvirt] [PATCH 0/3] fix some issue around host-passthough cpu model

[Luyao Huang]

On 05/05/2015 08:30 PM, John Ferlan wrote:
>
> On 03/30/2015 04:49 AM, Luyao Huang wrote:
>> We support add feature in host-passthough cpu model, but there are
>> some place need fix.
>>
>> Luyao Huang (3):
>>    cpu:x86: fix cannot keep cpu feature after migrate/restore
>>    qemu: add a check for non-migratable cpu flags
>>    docs: fix a small xml error in docs
>>
>>   docs/formatdomain.html.in | 2 +-
>>   src/cpu/cpu_x86.c         | 9 ++++-----
>>   src/qemu/qemu_migration.c | 2 +-
>>   3 files changed, 6 insertions(+), 7 deletions(-)
>>
>
> Looks like these were orphaned ;-)...

Thanks a lot for 'saving' my patches again and again ;-)

>
> Looking through history of the changes, it seems you're essentially
> undoing what Jan changed for
> https://bugzilla.redhat.com/show_bug.cgi?id=1147584 in commit id
> 'f53bb1a'. I'm assuming these were missed on the initial pass, but
> bumping them should grab his attention.

Yes, as your said i try to undoing the Jan's changes, but we support add 
features line with host-passthough cpu model, some code for 
host-passthrough model work not well host-passthough with features, and 
there is a new problem around the  vm save/restore with host-passthough 
with features (you can see this bug 
https://bugzilla.redhat.com/show_bug.cgi?id=1207095).

Also as host-passthough can have the features now, I think we shouldn't 
skip the cpu features check for host-passthough model.

But seems my patches will cause some regression issue, i will recheck 
and retest them.

> I think patch 3/3 is ACK-able and I'll push that separately since it's
> unrelated.

Yes, it is a small issue.

> Your patch 4/3 undoes commit id 'ec5f817f'.

Thanks for pointing out that, i will recheck these patches and write a 
new version these days.


Thanks a lot for your review.

>
>
> John

Luyao