[PATCH 04/18] ARM: at91: make ST (System Timer) soc independent

[rmallon@gmail.com (Ryan Mallon)]

On 20/02/12 14:23, Jean-Christophe PLAGNIOL-VILLARD wrote:

> On 14:16 Mon 20 Feb     , Ryan Mallon wrote:
>> On 20/02/12 14:02, Jean-Christophe PLAGNIOL-VILLARD wrote:
>>
>>> On 12:52 Mon 20 Feb     , Ryan Mallon wrote:
>>>> On 20/02/12 12:38, Jean-Christophe PLAGNIOL-VILLARD wrote:
>>>>
>>>>> On 11:22 Mon 20 Feb     , Ryan Mallon wrote:
>>>>>> On 18/02/12 04:49, Nicolas Ferre wrote:
>>>>>>
>>>>>>> From: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
>>>>>>>
>>>>>>> Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
>>>>>>> Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
>>>>>>> ---
>>>>>>>  arch/arm/mach-at91/at91rm9200.c              |    4 +-
>>>>>>>  arch/arm/mach-at91/at91rm9200_time.c         |   37 ++++++++++++++++----------
>>>>>>>  arch/arm/mach-at91/generic.h                 |    1 +
>>>>>>>  arch/arm/mach-at91/include/mach/at91_st.h    |   32 +++++++++++++++-------
>>>>>>>  arch/arm/mach-at91/include/mach/at91rm9200.h |    2 +-
>>>>>>>  drivers/watchdog/at91rm9200_wdt.c            |    8 +++---
>>>>>>>  6 files changed, 53 insertions(+), 31 deletions(-)
>>>>>>>
>>>>>>> diff --git a/arch/arm/mach-at91/at91rm9200.c b/arch/arm/mach-at91/at91rm9200.c
>>>>>>
>>>>>> Hi Jean, Nicolas,
>>>>>>
>>>>>> Patch looks mostly good, couple of points below.
>>>>>>
>>>>>> ~Ryan
>>>>>>
>>>>>> <snip>
>>>>>>
>>>>>>>  	/* Cancel any pending alarm; flush any pending IRQ */
>>>>>>> -	at91_sys_write(AT91_ST_RTAR, alm);
>>>>>>> -	(void) at91_sys_read(AT91_ST_SR);
>>>>>>> +	at91_st_write(AT91_ST_RTAR, alm);
>>>>>>> +	(void) at91_st_read(AT91_ST_SR);
>>>>>>
>>>>>>
>>>>>> Can we please remove the (void) casting of the return value when making
>>>>>> this change, especially since at91_st_read is now a macro which doesn't
>>>>>> even have a return value. Same in a few other places.
>>>>> modification done by script and it's no the scope of this patch
>>>>
>>>>
>>>> That isn't an excuse to leave incorrect code there. It is a simple fix.
>>> no (void) in c means you don't care of the return so basically it's right
>>
>>
>> Because of the way the __raw_writel is defined you are casting the
>> result of an assignment, basically you are doing this:
>>
>>    int foo, bar;
>>
>>   (void)(foo = bar);
>>
>> Which is pointless. Don't make excuses for silly, redundant code. Fix
>> it, please.
> except it's a read it has nothing to do with that


It makes no difference, it is still equally pointless. We don't cast
unused return values to void, for functions or assignments, in the kernel.

It is pointless, superfluous code, and now is a really good opportunity
to fix it. You can probably even make your script do it for you :-).

~Ryan