Re: [RFC] new timeofday core subsystem (v.A1)

[Martin Waitz]

hoi :)

On Tue, Dec 07, 2004 at 05:56:38PM -0800, john stultz wrote:
> +struct timesource_t {

usually only typedefs end with _t

> +	char* name;
> +	int priority;
> +	enum {
> +		TIMESOURCE_FUNCTION,
> +		TIMESOURCE_MMIO_32,
> +		TIMESOURCE_MMIO_64
> +	} type;
> +	cycle_t (*read_fnct)(void);
> +	void* ptr;

This could be made __iomem if it is intended to point to an IO memory region.
Hmm, but then there is no ioread64, so I guess I'm wrong.

> +	cycle_t mask;
> +	u32 mult;
> +	u32 shift;
> +};

> +static inline nsec_t cyc2ns(struct timesource_t* ts, cycle_t cycles, cycle_t* rem)
> +{
> +	u64 ret;
> +	ret = (u64)cycles;
> +	ret *= ts->mult;
> +	ret >>= ts->shift;
> +	if (rem) /* XXX we still need to do remainder math */
> +		*rem = (cycle_t)0;
> +	return (nsec_t)ret;
> +}

well, the math is simple:
	if (rem) *rem = ret & (1 << ts->shift -1);
	ret >>= ts->shift;


-- 
Martin Waitz

Attachment: signature.asc
Description: Digital signature