On 03/07/2016 07:20 AM, Paolo Bonzini wrote:
On 06/03/2016 08:15, Xiao Guangrong wrote:
+ /* cpuid 7.0.ecx*/
+ const u32 kvm_supported_word11_x86_features = F(PKU) | 0 /*OSPKE*/;
+
cpuid_mask(&entry->ebx, 9);
// TSC_ADJUST is emulated
entry->ebx |= F(TSC_ADJUST);
- } else
+ entry->ecx &= kvm_supported_word11_x86_features;
+ cpuid_mask(&entry->ecx, 13);
Can we use a meaningful name defined by cpuid_leafs instead of the raw
number?
This is not a cpuid leaf number, it's the Linux feature word number. In
other words, the "11" in kvm_supported_word11_x86_features should have
matched the "13" in cpuid_mask(&entry->ecx, 13).
But i noticed kernel uses the more meaningful name to update
boot_cpu_data.x86_capability[]:
$ git grep x86_capability arch/x86/
......
arch/x86/kernel/cpu/centaur.c: c->x86_capability[CPUID_C000_0001_EDX] = cpuid_edx(0xC0000001);
arch/x86/kernel/cpu/common.c: c->x86_capability[CPUID_1_ECX] = ecx;
arch/x86/kernel/cpu/common.c: c->x86_capability[CPUID_1_EDX] = edx;
arch/x86/kernel/cpu/common.c: c->x86_capability[CPUID_7_0_EBX] = ebx;
arch/x86/kernel/cpu/common.c: c->x86_capability[CPUID_6_EAX] = cpuid_eax(0x00000006);
arch/x86/kernel/cpu/common.c: c->x86_capability[CPUID_D_1_EAX] = eax;
arch/x86/kernel/cpu/common.c: c->x86_capability[CPUID_F_0_EDX] = edx;
.....
Why not follow it in KVM?
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
