Re: [RFC PATCH 07/11] target/arm: Replace kvm_arm_pmu_supported by host_cpu_feature_supported

["Philippe Mathieu-Daudé" <philmd@linaro.org>]

On 12/8/25 06:49, Philippe Mathieu-Daudé wrote:
> On 12/8/25 02:48, Richard Henderson wrote:
>> On 8/12/25 03:06, Philippe Mathieu-Daudé wrote:
>>> +++ b/target/arm/kvm.c
>>> @@ -288,7 +288,7 @@ static bool 
>>> kvm_arm_get_host_cpu_features(ARMHostCPUFeatures *ahcf)
>>>                                1 << KVM_ARM_VCPU_PTRAUTH_GENERIC);
>>>       }
>>> -    if (kvm_arm_pmu_supported()) {
>>> +    if (host_cpu_feature_supported(ARM_FEATURE_PMU, false)) {
>>
>> Why is false correct here?  Alternately, in the next patch, why is it 
>> correct to pass true for the EL2 test?
> 
> I think I copied to KVM the HVF use, adapted on top of:
> https://lore.kernel.org/qemu-devel/20250808070137.48716-12- 
> mohamed@unpredictable.fr/
> 
>>
>> What is the purpose of the can_emulate parameter at all?
> 
> When using split-accel on pre-M3, we might emulate EL2:
> 
>         |   feat            |    can_emulate   |    retval
>         +   ----            +      -----       +     ----
 > M1/M2  |  ARM_FEATURE_EL2         false            false> M1/M2  |  
ARM_FEATURE_EL2         true             true
> M3/M4  |  ARM_FEATURE_EL2         any              true

For example in hvf.c:

static bool hvf_arm_get_host_cpu_features(ARMHostCPUFeatures *ahcf)
{
     ...
     if (host_cpu_feature_supported(ARM_FEATURE_EL2, true)) {
         ahcf->features |= 1ULL << ARM_FEATURE_EL2;
     }

and then only when split-accel is not enabled:

hv_return_t hvf_arch_vm_create(MachineState *ms, uint32_t pa_range)
{
     ...
     if (host_cpu_feature_supported(ARM_FEATURE_EL2, false)) {
         ret = hv_vm_config_set_el2_enabled(config, true);
         if (ret != HV_SUCCESS) {
             goto cleanup;
         }
     }