Re: [PATCH 9/9] Allow setting the SVME bit v5

[Anthony Liguori <anthony@codemonkey.ws>]

Alexander Graf wrote:
>
> On 30.10.2008, at 20:16, Anthony Liguori wrote:
>>
>> Instead of doing this as a module parameter, we could either avoid 
>> advertising SVM support in cpuid from userspace or we could disable 
>> SVM in the BIOS.
>>
>> I really like the later approach because it gives a better error 
>> message within the guest.  We could also keep the module parameter 
>> but it would be nice to support disabling SVM in userspace too.
>
> That's a really good idea. We'd only have to check the MSR value on 
> EFER set, to know if setting SVME is allowed. That would allow for a 
> really easy control over the SVM feature on a per-guest basis and from 
> userspace.

And with Gleb's BIOS configuration changes that are now upstream, it's 
really easy to communicate to the BIOS whether SVM should be enabled.

Regards,

Anthony LIguori

> Cool,
>
> Alex
>
>> Regards,
>>
>> Anthony Liguori
>>
>>> static void kvm_reput_irq(struct vcpu_svm *svm);
>>> static void svm_flush_tlb(struct kvm_vcpu *vcpu);
>>> @@ -457,6 +460,9 @@ static __init int svm_hardware_setup(void)
>>>     if (boot_cpu_has(X86_FEATURE_NX))
>>>         kvm_enable_efer_bits(EFER_NX);
>>> +    if (nested)
>>> +        kvm_enable_efer_bits(MSR_EFER_SVME_MASK);
>>> +
>>>     for_each_online_cpu(cpu) {
>>>         r = svm_cpu_init(cpu);
>>>         if (r)
>>>
>>
>