Messages in this thread | | | Date | Fri, 6 May 2022 23:49:00 +0530 | Subject | Re: [PATCH v2 2/3] x86: Remove vendor checks from prefer_mwait_c1_over_halt | From | Wyes Karny <> |
| |
On 5/6/2022 9:22 PM, Dave Hansen wrote: > On 5/6/22 02:42, Wyes Karny wrote: >> >> So, if CPUID.05H:ECX[0] = 0, MWAIT extensions are not available (not allowed) and hence >> it is safe to use MWAIT with EAX=0,ECX=0. Otherwise, we have to check CPUID.05H:EDX[bit 7:4] >> to get the number of C1 substates and this should be greater than equal to 1. > > Ahh, I misread the comment. I was confusing the CPUID leaf ECX data > with the use of ECX hints to MWAIT. > > Could you add maybe a sentence or two more in that comment to help > clarify the situation?
Sure, will do. Thanks!
>
| |