2011/2/26 Sangbeom Kim <sbkim73@xxxxxxxxxxx>:
> (...)
> +static void s5p_clockevent_init(void)
> +{
> + unsigned long pclk;
> + unsigned long clock_rate;
> + unsigned int irq_number;
> + struct clk *tscaler;
> +
> + pclk = clk_get_rate(timerclk);
> +
> + tscaler = clk_get_parent(tdiv_event);
> +
> + clk_set_rate(tscaler, pclk / 2);
> + clk_set_rate(tdiv_event, pclk / 2);
> + clk_set_parent(tin_event, tdiv_event);
> +
> + clock_rate = clk_get_rate(tin_event);
> + clock_count_per_tick = clock_rate / HZ;
> +
> + time_event_device.mult =
> + div_sc(clock_rate, NSEC_PER_SEC, time_event_device.shift);
> + time_event_device.max_delta_ns =
> + clockevent_delta2ns(-1, &time_event_device);
> + time_event_device.min_delta_ns =
> + clockevent_delta2ns(1, &time_event_device);
This is a very complicated and inprecise way of doing this nowadays.
Skip hardcoding the shift value and calculating mult like that and use
/* Be able to sleep for atleast 4 seconds (usually more) */
#define EVT_MIN_RANGE 4
clockevents_calc_mult_shift(&time_event_device,
clock_rate, EVT_MIN_RANGE);
> (...)
> +static void s5p_clocksource_init(void)
> +{
> + unsigned long pclk;
> + unsigned long clock_rate;
> +
> + pclk = clk_get_rate(timerclk);
> +
> + clk_set_rate(tdiv_source, pclk / 2);
> + clk_set_parent(tin_source, tdiv_source);
> +
> + clock_rate = clk_get_rate(tin_source);
> +
> + s5p_time_setup(timer_source.source_id, TCNT_MAX);
> + s5p_time_start(timer_source.source_id, PERIODIC);
> +
> + if (clocksource_register_hz(&time_clocksource, clock_rate))
> + panic("%s: can't register clocksource\n", time_clocksource.name);
> +}
This is more like it :-)
But you probably also want to add a sched_clock hook for this
platform too, so you get some nice scheduling resolution.
I suggest you look at the simple straight-forward driver
for U300 in arch/arm/mach-u300/timer.c for inspiration.
It's using the same timer that is used for clocksource for
sched_clock().
Yours,
Linus Walleij
--
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
