Re: Coding style details (checkpatch)

[khalasa@piap.pl (Krzysztof Hałasa)]

Joe Perches <joe@perches.com> writes:

> It might be better to use some base + index macro
> as it could be smaller object code.
>
> Something like:
>
> #define REG_NO(base, multiplier, index)	(base + (multiplier * index))
>
> 	reg_write(vc->dev, REG_NO(0x10, 1, vc->ch), dma_cfg);
> or
>
> #define VDMA_CHANNEL_CONFIG	0x10
>
> 	reg_write(vc->dev, REG_NO(VDMA_CHANNEL_CONFIG, 1, vc->ch), dma_cfg);

Wouldn't work, the register map is a bit messy.
E.g.

#define DMA_PAGE_TABLE0_ADDR    ((const u16[8]){0x08, 0xD0, 0xD2, 0xD4, 0xD6, 0xD8, 0xDA, 0xDC})
#define DMA_PAGE_TABLE1_ADDR    ((const u16[8]){0x09, 0xD1, 0xD3, 0xD5, 0xD7, 0xD9, 0xDB, 0xDD})

also

#define VDREG8(a0) ((const u16[8]){                     \
        a0 + 0x000, a0 + 0x010, a0 +0x020, a0 + 0x030,  \
        a0 + 0x100, a0 + 0x110, a0 +0x120, a0 + 0x130})
#define VIDSTAT                 VDREG8(0x100)
#define BRIGHT                  VDREG8(0x101)
#define CONTRAST                VDREG8(0x102)

etc.

One would have to remember (writing .c code) which scheme applies to
which access. The tables, while probably less than optimal WRT CPU
cycles used, are consistent, and the addressing details are grouped in
one place - the *regs.h file.
-- 
Krzysztof Halasa

Industrial Research Institute for Automation and Measurements PIAP
Al. Jerozolimskie 202, 02-486 Warsaw, Poland
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
Please read the FAQ at  http://www.tux.org/lkml/