[PATCH v6 06/43] arm64: RME: Check for RME support at KVM init

[Gavin Shan]

On 12/19/24 3:44 PM, Aneesh Kumar K.V wrote:
> Steven Price <steven.price at arm.com> writes:
> 
>> +static int rmi_check_version(void)
>> +{
>> +	struct arm_smccc_res res;
>> +	int version_major, version_minor;
>> +	unsigned long host_version = RMI_ABI_VERSION(RMI_ABI_MAJOR_VERSION,
>> +						     RMI_ABI_MINOR_VERSION);
>> +
>> +	arm_smccc_1_1_invoke(SMC_RMI_VERSION, host_version, &res);
>> +
>> +	if (res.a0 == SMCCC_RET_NOT_SUPPORTED)
>> +		return -ENXIO;
>> +
>> +	version_major = RMI_ABI_VERSION_GET_MAJOR(res.a1);
>> +	version_minor = RMI_ABI_VERSION_GET_MINOR(res.a1);
>> +
>> +	if (res.a0 != RMI_SUCCESS) {
>> +		kvm_err("Unsupported RMI ABI (v%d.%d) we want v%d.%d\n",
>> +			version_major, version_minor,
>> +			RMI_ABI_MAJOR_VERSION,
>> +			RMI_ABI_MINOR_VERSION);
>> +		return -ENXIO;
>> +	}
>> +
>> +	kvm_info("RMI ABI version %d.%d\n", version_major, version_minor);
>> +
>> +	return 0;
>> +}
>> +
> 
> Should we include both high and low version numbers in the kvm_err
> message on error? ie,
> 
> 	high_version_major = RMI_ABI_VERSION_GET_MAJOR(res.a2);
> 	high_version_minor = RMI_ABI_VERSION_GET_MINOR(res.a2);
> 

I think so since a range of supported versions are returned in the failing case.

Besides, 'unsigned short' is more suitable for the local variable version_{major, minor}
since both are 16-bits in width. 'unsigned short' explicitly indicates their
width.

Thanks,
Gavin