On 14.02.2020 18:59, Andrew Cooper wrote:
> On 28/01/2020 14:10, Jan Beulich wrote:
>> On 28.01.2020 13:52, Andrew Cooper wrote:
>>> boot_cpu_has(X86_FEATURE_X2APIC) doesn't need checking to interpret
>>> APIC_BASE_EXTD.
>> Hmm, the comment you remove ...
>>
>>> --- a/xen/arch/x86/apic.c
>>> +++ b/
On 28/01/2020 14:10, Jan Beulich wrote:
> On 28.01.2020 13:52, Andrew Cooper wrote:
>> boot_cpu_has(X86_FEATURE_X2APIC) doesn't need checking to interpret
>> APIC_BASE_EXTD.
> Hmm, the comment you remove ...
>
>> --- a/xen/arch/x86/apic.c
>> +++ b/xen/arch/x86/apic.c
>> @@ -1534,18 +1534,14 @@ void
On 28.01.2020 13:52, Andrew Cooper wrote:
> boot_cpu_has(X86_FEATURE_X2APIC) doesn't need checking to interpret
> APIC_BASE_EXTD.
Hmm, the comment you remove ...
> --- a/xen/arch/x86/apic.c
> +++ b/xen/arch/x86/apic.c
> @@ -1534,18 +1534,14 @@ void __init record_boot_APIC_mode(void)
> /* Look at
On Tue, Jan 28, 2020 at 12:52:16PM +, Andrew Cooper wrote:
> boot_cpu_has(X86_FEATURE_X2APIC) doesn't need checking to interpret
> APIC_BASE_EXTD.
>
> Also take the opportunity to optimise the generated assembly by not using
> rdmsrl(). GCC isn't clever enough to spot that it can drop the shi
boot_cpu_has(X86_FEATURE_X2APIC) doesn't need checking to interpret
APIC_BASE_EXTD.
Also take the opportunity to optimise the generated assembly by not using
rdmsrl(). GCC isn't clever enough to spot that it can drop the shift and or
to put %eax in the higher half of msr_contents.
No functional