Re: bug report: signedness issue in ds3232_update_alarm()

[Lan Chunhe-B25806 <B25806@freescale.com>]

On Thu, 28 Oct 2010 16:55:22 +0800, Dan Carpenter <error27@gmail.com>  
wrote:

> Jack's email <jack.lan@freescale.com> is dead.  B25806, could you take
> a look?

    Sorry, you can use b25806@freescale.com OR chunhe.lan@gmail.com.

> regards,
> dan carpenter
>
> On Thu, Oct 28, 2010 at 10:35:03AM +0200, Dan Carpenter wrote:
>> Hi Jack,
>>
>> This code was added in f46418c5cadf "drivers/rtc/rtc-ds3232.c: add alarm
>> function".  Smatch complains because bcd2bin() returns an unsigned so
>> it's never less than zero and the whole block is a no-op.  I'm not sure
>> what was intended here, could you take a look?

       Of course, bcd2bin()/Time is never less than zero. But it is fault  
tolerant.
       When time has exceptional data, it set time to the default value.

>> drivers/rtc/rtc-ds3232.c +292 ds3232_update_alarm(13)
>> 	warn: unsigned 'bcd2bin(buf[0])' is never less than zero.
>>    291
>>    292          buf[0] = bcd2bin(buf[0]) < 0 || (ds3232->rtc->irq_data  
>> & RTC_UF) ?
>>    293									0x80 : buf[0];
>>    294          buf[1] = bcd2bin(buf[1]) < 0 || (ds3232->rtc->irq_data  
>> & RTC_UF) ?
>>    295									0x80 : buf[1];
>>    296          buf[2] = bcd2bin(buf[2]) < 0 || (ds3232->rtc->irq_data  
>> & RTC_UF) ?
>>    297									0x80 : buf[2];
>>    298          buf[3] = bcd2bin(buf[3]) < 0 || (ds3232->rtc->irq_data  
>> & RTC_UF) ?
>>    299									0x80 : buf[3];
>>    300
>>
>> regards,
>> dan carpenter

      Thanks.

      Jack Lan