Messages in this thread | | | Subject | Re: [PATCH 1/2] comedi: dmm32at: Fix coding style - use BIT macro | From | Ian Abbott <> | Date | Wed, 18 Nov 2015 17:02:50 +0000 |
| |
On 18/11/15 16:45, Hartley Sweeten wrote: > On Wednesday, November 18, 2015 9:42 AM, Ian Abbott wrote: >> On 16/11/15 17:18, Ranjith Thangavel wrote: > [snip] >>> -#define DMM32AT_AI_CFG_SCINT_20US (0 << 4) >>> -#define DMM32AT_AI_CFG_SCINT_15US (1 << 4) >>> -#define DMM32AT_AI_CFG_SCINT_10US (2 << 4) >>> -#define DMM32AT_AI_CFG_SCINT_5US (3 << 4) >>> -#define DMM32AT_AI_CFG_RANGE (1 << 3) /* 0=5V 1=10V */ >>> -#define DMM32AT_AI_CFG_ADBU (1 << 2) /* 0=bipolar 1=unipolar */ >>> +#define DMM32AT_AI_CFG_SCINT_20US 0 >>> +#define DMM32AT_AI_CFG_SCINT_15US BIT(4) >>> +#define DMM32AT_AI_CFG_SCINT_10US (BIT(5) & ~BIT(4)) >> >> The `(BIT(5) & ~BIT(4))` is a bit ugly. You can just use `BIT(5)` to >> fit in with the style of your other changes. >> >> (Personally though, I don't think BIT() is appropriate for shifted, >> multi-bit values.) > > It would be more appropriate as a macro: > > #define DMM32AT_AI_CFG_SCINT(x) (((x) & 0x3) << 4) > #define DMM32AT_AI_CFG_SCINT_20US DMM32AT_AI_CFG_SCINT (0) > #define DMM32AT_AI_CFG_SCINT_15US DMM32AT_AI_CFG_SCINT (1) > #define DMM32AT_AI_CFG_SCINT_10US DMM32AT_AI_CFG_SCINT (2) > #define DMM32AT_AI_CFG_SCINT_5US DMM32AT_AI_CFG_SCINT (3)
Yes, but without the spaces in the macro calls!
It's slightly surprising no one has pushed a generic macro for this sort of thing, maybe something like:
#define MBIT(v, s) ((v) * BIT(s))
although the name sucks - maybe the problem is thinking up a decent name!
-- -=( Ian Abbott @ MEV Ltd. E-mail: <abbotti@mev.co.uk> )=- -=( Web: http://www.mev.co.uk/ )=-
| |