Re: [PATCH] target/i386: Add kvm_get_one_msr helper

["Yang, Weijiang" <weijiang.yang@intel.com>]

On 1/28/2022 6:55 PM, Paolo Bonzini wrote:
> On 1/27/22 16:58, Yang Weijiang wrote:
>> @@ -135,6 +135,7 @@ static struct kvm_msr_list *kvm_feature_msrs;
>>     #define BUS_LOCK_SLICE_TIME 1000000000ULL /* ns */
>>   static RateLimit bus_lock_ratelimit_ctrl;
>> +static int kvm_get_one_msr(X86CPU *cpu, int index, uint64_t *value);
>>     int kvm_has_pit_state2(void)
>>   {
>
> ...
>
>> @@ -2734,6 +2721,25 @@ static int kvm_put_one_msr(X86CPU *cpu, int 
>> index, uint64_t value)
>>       return kvm_vcpu_ioctl(CPU(cpu), KVM_SET_MSRS, cpu->kvm_msr_buf);
>>   }
>>   +static int kvm_get_one_msr(X86CPU *cpu, int index, uint64_t *value)
>> +{
>> +    int ret;
>> +    struct {
>> +        struct kvm_msrs info;
>> +        struct kvm_msr_entry entries[1];
>> +    } msr_data = {
>> +        .info.nmsrs = 1,
>> +        .entries[0].index = index,
>> +    };
>> +
>> +    ret = kvm_vcpu_ioctl(CPU(cpu), KVM_GET_MSRS, &msr_data);
>> +    if (ret < 0) {
>> +        return ret;
>> +    }
>> +    assert(ret == 1);
>> +    *value = msr_data.entries[0].data;
>> +    return ret;
>> +}
>>   void kvm_put_apicbase(X86CPU *cpu, uint64_t value)
>>   {
>>       int ret;
>
> The patch is a good idea, but you can put the function before the 
> uses.  This way there will be no need for a forward declaration, either.

Thanks Paolo!

Is v2 version required for this?

>
> Thanks,
>
> Paolo