On 01/28/2011 07:33 AM, Paolo Bonzini wrote:
On 01/28/2011 11:00 AM, Michael Tokarev wrote:
28.01.2011 11:06, Yoshiaki Tamura wrote:
[]
I would like Kemari to be included for 0.14. Thanks to many
valuable comments from various reviewers, it got better than
ever, I believe. For those who may object, it wouldn't affect
any functions including live migration unless one turns it on.
Besides, it is a good application of live migration. By having
merged, I would be able to focus on increasing functionality and
optimization.
It's somewhat unexpected to see merge request for any new
functionality into a "stable" branch. Having nothing bad
against Kemari, I still think we shouldn't include any new
functionality into 0.14, which were planned to be released
before new year... ;)
I mean, if it were not merged so far, for reasons that can
be discussed separately etc, regardless of its good shape
and so on... let's don't do any last-minute merges.
To be fair to Yoshiaki, the first 17 patches in his 19-patch series
have no impact on the operation of QEMU, and even the last two are
Kemari-only. I know almost nothing about Kemari so I cannot comment
on the technical side of those patches, only that it w^Hshouldn't be
destabilizing. I made some comments on a couple of patches, so I'm
inclined to say no as well.
No, I'd prefer not to take Kemari before the 0.14 freeze. It gives us a
full release cycle for testing plus gives a nice headline feature for 0.15.
http://permalink.gmane.org/gmane.comp.emulators.qemu/91096 should be
applied in any case, as it is a regression from 0.12.
Thanks.
Regards,
Anthony Liguori
Paolo
