Re: [PATCH] tty: serial: samsung_tty: remove set but not used variables

[Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>]

On 23/04/2021 12:14, Greg KH wrote:
> On Fri, Apr 23, 2021 at 05:54:16PM +0800, tiantao (H) wrote:
>>
>> 在 2021/4/23 17:47, Greg KH 写道:
>>> On Fri, Apr 23, 2021 at 05:39:00PM +0800, Tian Tao wrote:
>>>> The value of 'ret' is not used, so just delete it.

Tian Tao, please use scripts/get_maintainers.pl to get the list of
people needed for Cc.

>>>>
>>>> Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
>>>> ---
>>>>   drivers/tty/serial/samsung_tty.c | 1 -
>>>>   1 file changed, 1 deletion(-)
>>>>
>>>> diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
>>>> index d9e4b67..d269d75 100644
>>>> --- a/drivers/tty/serial/samsung_tty.c
>>>> +++ b/drivers/tty/serial/samsung_tty.c
>>>> @@ -2220,7 +2220,6 @@ static int s3c24xx_serial_probe(struct platform_device *pdev)
>>>>   			default:
>>>>   				dev_warn(&pdev->dev, "unsupported reg-io-width (%d)\n",
>>>>   						prop);
>>>> -				ret = -EINVAL;
>>> That looks odd, shouldn't you do something with this instead of ignoring
>>> it???
>>
>> How about this ？
>>
>> diff --git a/drivers/tty/serial/samsung_tty.c
>> b/drivers/tty/serial/samsung_tty.c
>> index d9e4b67..9fbc611 100644
>> --- a/drivers/tty/serial/samsung_tty.c
>> +++ b/drivers/tty/serial/samsung_tty.c
>> @@ -2220,8 +2220,7 @@ static int s3c24xx_serial_probe(struct platform_device
>> *pdev)
>>                         default:
>>                                 dev_warn(&pdev->dev, "unsupported
>> reg-io-width (%d)\n",
>>                                                 prop);
>> -                               ret = -EINVAL;
>> -                               break;
>> +                               return -EINVAL;
>>
> 
> You tell me, does the patch work for you?
> 
> Is this really a "hard error" and did you now just break devices that
> used to work properly?  Are you correctly unwinding any previously
> allocated state when you return here?
> 
> Please do some research on this, and ideally, lots of testing, before
> submitting it as a real solution.

It's a patch coming from automated tool (e.g. Coverity), so I doubt
there is any testing here. However the "return -EINVAL" looks correct here:
1. No particular unwinding is needed here,
2. It's an optional property (not used by existing DTS, only
non-upstreamed by Samsung) thus treating it as hard-error is fine.
Probably better to exit than convert it to some default value.


Best regards,
Krzysztof