Re: [PATCH 3/3] drivers: hwmon: i5k_amb: remove unnecessary #ifdef MODULE

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

On 6/5/19 4:59 PM, Enrico Weigelt, metux IT consult wrote:
> On 01.06.19 22:49, Guenter Roeck wrote:
>> On Sat, Jun 01, 2019 at 04:01:40PM +0200, Enrico Weigelt, metux IT consult wrote:
>>> The MODULE_DEVICE_TABLE() macro already checks for MODULE defined,
>>> so the extra check here is not necessary.
>>>
>>> Signed-off-by: Enrico Weigelt <info@metux.net>
>>> ---
>>>   drivers/hwmon/i5k_amb.c | 2 --
>>>   1 file changed, 2 deletions(-)
>>>
>>> diff --git a/drivers/hwmon/i5k_amb.c b/drivers/hwmon/i5k_amb.c
>>> index b09c39a..b674c2f 100644
>>> --- a/drivers/hwmon/i5k_amb.c
>>> +++ b/drivers/hwmon/i5k_amb.c
>>> @@ -482,14 +482,12 @@ static int i5k_channel_probe(u16 *amb_present, unsigned long dev_id)
>>>       { 0, 0 }
>>>   };
>>> -#ifdef MODULE
>>>   static const struct pci_device_id i5k_amb_ids[] = {
>>>       { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_5000_ERR) },
>>>       { PCI_DEVICE(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_5400_ERR) },
>>>       { 0, }
>>>   };
>>>   MODULE_DEVICE_TABLE(pci, i5k_amb_ids);
>>> -#endif
>>
>> I'd rather know what this table is used for in the first place.
> 
> Seems it's really just used for the module loader, while actual probing
> is using a different table. IMHO, the worst thing my patch could do is
> introducing a warning on unused variable (IMHO shouldn't happen when
> it's static const).
> 

You are wrong. You'll need __maybe_unused qualifiers for those variables
to avoid the warning.

Guenter

> I've just rewritten it to move everything into i5k_amb_ids ... just need
> to run build tests on it (unfortunately can't run-test, as I don't have
> that device).
> 
> 
> --mtx
>