On 28/10/2022 6:47 am, Sean Christopherson wrote:
What happens if userspace sets X86_FEATURE_PERFCTR_CORE when its not supported?
E.g. will KVM be coerced into taking a #GP on a non-existent counter?
I'm getting a bit tired of this generic issue, what does KVM need to do when the
KVM user space
sets a capability that KVM doesn't support (like cpuid). Should it change the
guest cpuid audibly
or silently ? Should it report an error when the guest uses this unsupported
capability at run time,
or should it just let the KVM report an error when user space setting the cpuid ?
For vPMU, I may prefer to report an error when setting the vpmu capability (via
cpuid or perf_cap).
Please let me know what you think and make it law as an example to others.
Thanks,
Like Xu
