Re: [Qemu-devel] [PATCH 0/7 RESEND] target-i386: Simplify kvm_cpu_fill_host() and kvm_check_features_against_host()

[Eduardo Habkost <ehabkost@redhat.com>]

On Mon, Jan 20, 2014 at 02:41:07PM -0200, Eduardo Habkost wrote:
> Resend of series submitted on 24 November 2013, that didn't get any reply. Only
> change is a trivial conflict on patch 7/7.

Question: which tree is the most appropriate to get this in? qom-cpu?
kvm?


> 
> This series simplifies kvm_cpu_fill_host() and
> kvm_check_features_against_host() to simply use FeatureWord & feature_word_info
> loops to fill/check feature words.
> 
> The initial motivation for this was to avoid hacks involving the "host" CPU
> class on the forthcoming conversion of CPU models to be X86CPU subclasses.
> Instead of requiring the kvm_arch_get_supported_cpuid() results to be stored in
> the class struct for "host" (thus requiring KVM initialization hacks).
> 
> Eduardo Habkost (7):
>   target-i386: kvm_cpu_fill_host(): Kill unused code
>   target-i386: kvm_cpu_fill_host(): No need to check level
>   target-i386: kvm_cpu_fill_host(): No need to check CPU vendor
>   target-i386: kvm_cpu_fill_host(): No need to check xlevel2
>   target-i386: kvm_cpu_fill_host(): Set all feature words at end of
>     function
>   target-i386: kvm_cpu_fill_host(): Fill feature words in a loop
>   target-i386: kvm_check_features_against_host(): Kill feature word
>     array
> 
>  target-i386/cpu.c | 89 +++++++++++++------------------------------------------
>  1 file changed, 20 insertions(+), 69 deletions(-)
> 
> -- 
> 1.8.4.2
> 
> 

-- 
Eduardo