Re: [PATCH] hwmon: amd_energy: match for supported models

[Guenter Roeck <linux@roeck-us.net>]

On 7/6/20 10:20 AM, Naveen Krishna Ch wrote:
> Hi Guenter,
> 
> On Mon, 6 Jul 2020 at 22:47, Naveen Krishna Chatradhi <nchatrad@amd.com> wrote:
>>
>> The energy counters of certain models seems to be reporting
>> inconsisten values. Hence, match for the supported models.
> Actually, the supported models could be of family 0x17 in a range
> between 0x30 ~ 0x3f. I did not find any macro or usage for a range
> of models. Could you suggest to me if i've missed an existing way to
> provide a range for models.
> 

I have no idea, sorry. Maybe match all of them, check for
actually supported models in the probe function, and return
-ENODEV if the model is not supported. Alternatively,
just list all models in the supported range.

Either case, how would you know that a future model in the presumably
supported range doesn't have a bug in the microcode that makes it
report bad data ? Assuming that this is a microcode bug, of course.

Guenter

>>
>> Signed-off-by: Naveen Krishna Chatradhi <nchatrad@amd.com>
>> ---
>>  drivers/hwmon/amd_energy.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/hwmon/amd_energy.c b/drivers/hwmon/amd_energy.c
>> index e95b7426106e..29603742c858 100644
>> --- a/drivers/hwmon/amd_energy.c
>> +++ b/drivers/hwmon/amd_energy.c
>> @@ -362,7 +362,7 @@ static struct platform_driver amd_energy_driver = {
>>  static struct platform_device *amd_energy_platdev;
>>
>>  static const struct x86_cpu_id cpu_ids[] __initconst = {
>> -       X86_MATCH_VENDOR_FAM(AMD, 0x17, NULL),
>> +       X86_MATCH_VENDOR_FAM_MODEL(AMD, 0x17, 0x31, NULL),
>>         {}
>>  };
>>  MODULE_DEVICE_TABLE(x86cpu, cpu_ids);
>> --
>> 2.17.1
>>
> 
>