On 16/05/2023 1:53 pm, Jan Beulich wrote: > On 15.05.2023 16:42, Andrew Cooper wrote: >> Extend x86_cpu_policy_fill_native() with a read of ARCH_CAPS based on the >> CPUID information just read, which removes the need handling it specially in >> calculate_raw_cpu_policy(). >> >> Extend generic_identify() to read ARCH_CAPS into x86_capability[], which is >> fed into the Host Policy. This in turn means there's no need to special case >> arch_caps in calculate_host_policy(). >> >> No practical change. >> >> Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com> > Reviewed-by: Jan Beulich <jbeul...@suse.com>
Thanks. > > I have a question though, which I think would be nice if the description > had covered: > >> --- a/xen/lib/x86/cpuid.c >> +++ b/xen/lib/x86/cpuid.c >> @@ -226,7 +226,12 @@ void x86_cpu_policy_fill_native(struct cpu_policy *p) >> p->hv_limit = 0; >> p->hv2_limit = 0; >> >> - /* TODO MSRs */ >> +#ifdef __XEN__ >> + /* TODO MSR_PLATFORM_INFO */ >> + >> + if ( p->feat.arch_caps ) >> + rdmsrl(MSR_ARCH_CAPABILITIES, p->arch_caps.raw); >> +#endif > What about non-Xen environments re-using this code? In particular the > test harnesses would be nice if they didn't run with the two fields > all blank at all times. Right now, I don't have an answer. In Linux in lockdown mode, there isn't even a way to access this info userspace, because /proc/cpu/$/msr goes away. It's only a unit test, and this doesn't break it. ~Andrew
