Re: [Qemu-devel] [PATCH 00/11] cpu model bug fixes and definition corrections (v2)

[Eduardo Habkost]

(CCing Marcelo, Avi, and kvm mailing list, so they can help answering
the uq/master patch flow question)

On Fri, Jun 03, 2011 at 12:51:42AM +0200, Jan Kiszka wrote:
> On 2011-06-02 21:34, Eduardo Habkost wrote:
> > Ouch, the subject prefix is completely wrong because of broken
> > git-send-email config on my side, sorry.
> > 
> > Please ignore the 'RHEL6 qemu-kvm' prefix, it is actually supposed to go
> > to the main Qemu tree.
> 
> Some of my review comments on John's original version still apply. Same
> for the advice on the patch flow (uq/master for kvm stuff).

Just to make sure I didn't miss anything:

1) uq/master flow: considering that most of the series is not
  KVM-specific but depends on patch 02/11 (Allow an optional
  qemu_early_init_vcpu()) what is the best approach? Should the whole
  series go through uq/master, or just patch 02/11? In the case of the
  latter, shall the rest of the series wait for the patch to be merged
  upstream, or should patch 02/11 go to both branches at the same time?

2) Reviewing cpu_x86_cpuid() cpuid hacking code & dropping
  -enable-nesting: should it hold the series, or may it be addressed
  after this series enter the tree?

3) Other recommendations for the qemu_early_init_vcpu() code
   (checkpatch.sh, return code evaluation, KVMState vs. VCPU): I will
   address those issues and send a new version.

Something else I may have missed?

-- 
Eduardo