On Thu, May 13, 2010 at 11:02 AM, Ben Dooks <ben-linux@xxxxxxxxx> wrote:
> On Thu, May 13, 2010 at 10:49:29AM +0900, Jassi Brar wrote:
>> On Thu, May 13, 2010 at 9:28 AM, Kukjin Kim <kgene.kim@xxxxxxxxxxx> wrote:
>> > From: Thomas Abraham <thomas.ab@xxxxxxxxxxx>
>> > --- a/arch/arm/mach-s5p6440/clock.c
>> > +++ b/arch/arm/mach-s5p6440/clock.c
>> > @@ -357,121 +357,121 @@ static struct clk init_clocks_disable[] = {
>> > .id = -1,
>> > .parent = &clk_hclk.clk,
>> > .enable = s5p6440_mem_ctrl,
>> > - .ctrlbit = S5P_CLKCON_MEM0_HCLK_NFCON,
>> > + .ctrlbit = (1 << 2),
>> Peculiar. I have never seen defines dropped in favor of magic numbers.
>
> I know it seems a little odd at first, but people seem to be clinging
> on to writing it down in a header file and then using it once as some
> form of high law of programming. Whilst talking with Kukjin and others
> last year and looking at this, we came to the following conclusions
> about single-use defines:
>
> 1) It takes two lines of code, where one is sufficient.
>
> 2) You only have to look in the relevant .c file to find out the
> value instead of tracking down a header. This makes it easier
> to verify the value against the manual and easier to compare
> against simialr code.
Then define it at c code and use the macro. I also don't like the
hard-coded values.
Thank you,
Kyungmin Park
--
To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
