On 3/28/21 9:15 PM, Richard Henderson wrote:
> On 3/28/21 1:12 PM, Richard Henderson wrote:
>> On 3/26/21 1:36 PM, Claudio Fontana wrote:
>>> +++ b/target/arm/monitor.c
>>> @@ -184,9 +184,11 @@ CpuModelExpansionInfo
>>> *qmp_query_cpu_model_expansion(CpuModelExpansionType type,
>>> if (!err) {
>>> visit_check_struct(visitor, &err);
>>> }
>>> +#ifdef TARGET_AARCH64
>>> if (!err) {
>>> - arm_cpu_finalize_features(ARM_CPU(obj), &err);
>>> + aarch64_cpu_finalize_features(ARM_CPU(obj), &err);
>>> }
>>> +#endif /* TARGET_AARCH64 */
>>> visit_end_struct(visitor, NULL);
>>> visit_free(visitor);
>>> if (err) {
>>> @@ -195,7 +197,9 @@ CpuModelExpansionInfo
>>> *qmp_query_cpu_model_expansion(CpuModelExpansionType type,
>>> return NULL;
>>> }
>>> } else {
>>> - arm_cpu_finalize_features(ARM_CPU(obj), &error_abort);
>>> +#ifdef TARGET_AARCH64
>>> + aarch64_cpu_finalize_features(ARM_CPU(obj), &error_abort);
>>> +#endif /* TARGET_AARCH64 */
>>
>> These ifdefs are not an improvement.
Right, here we cannot use is_a64(), and I have not found another way other than
adding a stub for the function?
>
> And are actively wrong, since we've lost the runtime test for
> ARM_FEATURE_AARCH64.
>
> r~
>
Right will fix.
Thanks
C
