[RFC] Define/use pad() and pad_mask() instead of open coding it

[Matt Turner]

On Thu, Mar 11, 2010 at 8:05 PM, Peter Hutterer
<peter.hutterer at who-t.net> wrote:
> On Thu, Mar 11, 2010 at 07:33:30PM -0500, Matt Turner wrote:
>> à la the Linux Kernel's ALIGN and __ALIGN_MASK macros.
>>
>> Not even compile tested, some files might need to include "misc.h".
>>
>> CC: Peter Hutterer <peter.hutterer at who-t.net>
>> Signed-off-by: Matt Turner <mattst88 at gmail.com>
>
> ACK in principle, not tested either. :)
>
>> diff --git a/include/misc.h b/include/misc.h
>> index 62d813e..98d39bd 100644
>> --- a/include/misc.h
>> +++ b/include/misc.h
>> @@ -200,6 +200,16 @@ bytes_to_int32(const int bytes) {
>>      return (((bytes) + 3) >> 2);
>>  }
>
> + /**
> +  * please comment on what I'm supposed to be doing
> +  */

Got any suggested text?

>>
>> +static inline int
>> +pad_mask(const int bytes, const int mask) {
>> +     return (((bytes) + mask) & ~mask);
>
> you don't need the () around bytes, this isn't a macro.
> which, looking at the context goes for bytes_to_int32 as well :)
>
>> +}
>> +
>> +static inline int
>> +pad(const int bytes, const int mask) {
>
> shouldn't the second arg be called alignment or something, not mask?
> otherwise, it's a bit confusing why both pad() and pad_mask() take a mask
> argument.

Good point.

>> +     return pad_mask(bytes, mask - 1);
>> +}
>> +
>
> Cheers,
>  Peter

I've found some more too--((x + bytes - 1) / bytes) * bytes. Ugh, how nasty.

Matt