[PATCH 1/2] clocksource: stm32: rework driver to use only one timer
Benjamin Gaignard
benjamin.gaignard at linaro.org
Tue Sep 19 00:59:03 PDT 2017
2017-09-18 23:30 GMT+02:00 Daniel Lezcano <daniel.lezcano at linaro.org>:
> On 14/09/2017 09:56, Benjamin Gaignard wrote:
>> Rework driver code to use only one timer for both clocksource
>> and clockevent.
>> This patch also forbids to use 16 bits timers because they are
>> not enough accurate.
>> Do some clean up in structures and functions names too.
>>
>> Signed-off-by: Ludovic Barre <ludovic.barre at st.com>
>> Signed-off-by: Benjamin Gaignard <benjamin.gaignard at linaro.org>
>
> Hi Benjamin,
>
> I have a few comments below. Can you when reposting split this patch
> into smaller changes ?
Not so easy because I change the way of how the hardware is used to be able
to provide clocksource and clockevent on the same hardware block.
>
> Also, can you consider using the timer-of API ?
Is it just about using TIMER_OF_DECLARE ? or do you have something
else in mind ?
>
> Thanks.
>
> -- Daniel
>
>> ---
>> drivers/clocksource/timer-stm32.c | 259 +++++++++++++++++++++++---------------
>> 1 file changed, 155 insertions(+), 104 deletions(-)
>>
>> diff --git a/drivers/clocksource/timer-stm32.c b/drivers/clocksource/timer-stm32.c
>> index 8f24237..648c10a 100644
>> --- a/drivers/clocksource/timer-stm32.c
>> +++ b/drivers/clocksource/timer-stm32.c
>> @@ -16,175 +16,226 @@
>> #include <linux/of_irq.h>
>> #include <linux/clk.h>
>> #include <linux/reset.h>
>> +#include <linux/sched_clock.h>
>> +#include <linux/slab.h>
>>
>> #define TIM_CR1 0x00
>> #define TIM_DIER 0x0c
>> #define TIM_SR 0x10
>> #define TIM_EGR 0x14
>> +#define TIM_CNT 0x24
>> #define TIM_PSC 0x28
>> #define TIM_ARR 0x2c
>> +#define TIM_CCR1 0x34
>>
>> #define TIM_CR1_CEN BIT(0)
>> -#define TIM_CR1_OPM BIT(3)
>> +#define TIM_CR1_UDIS BIT(1)
>> #define TIM_CR1_ARPE BIT(7)
>>
>> -#define TIM_DIER_UIE BIT(0)
>> -
>> -#define TIM_SR_UIF BIT(0)
>> +#define TIM_DIER_CC1IE BIT(1)
>>
>> #define TIM_EGR_UG BIT(0)
>>
>> -struct stm32_clock_event_ddata {
>> +struct stm32_clock_event {
>> struct clock_event_device evtdev;
>> unsigned periodic_top;
>> - void __iomem *base;
>> + void __iomem *regs;
>> };
>>
>> static int stm32_clock_event_shutdown(struct clock_event_device *evtdev)
>> {
>> - struct stm32_clock_event_ddata *data =
>> - container_of(evtdev, struct stm32_clock_event_ddata, evtdev);
>> - void *base = data->base;
>> + struct stm32_clock_event *ce =
>> + container_of(evtdev, struct stm32_clock_event, evtdev);
>> +
>> + writel_relaxed(0, ce->regs + TIM_DIER);
>>
>> - writel_relaxed(0, base + TIM_CR1);
>
> Why this change? TIM_CR1 -> TIM_DIER? A 16b to 32b change?
No it is because I use the interrupt from the comparator instead of the counter.
With this patch clocksource will use the 32 bits counter of the hardware block
to provide the clock and comparator is generate interrupt for the event.
That change quite a lot the code, sorry.
>
>> return 0;
>> }
>>
>> -static int stm32_clock_event_set_periodic(struct clock_event_device *evtdev)
>> +static int stm32_clock_event_set_next_event(unsigned long evt,
>> + struct clock_event_device *evtdev)
>> {
>> - struct stm32_clock_event_ddata *data =
>> - container_of(evtdev, struct stm32_clock_event_ddata, evtdev);
>> - void *base = data->base;
>> + struct stm32_clock_event *ce =
>> + container_of(evtdev, struct stm32_clock_event, evtdev);
>> + unsigned long cnt;
>> +
>> + cnt = readl_relaxed(ce->regs + TIM_CNT);
>> + writel_relaxed(cnt + evt, ce->regs + TIM_CCR1);
>> + writel_relaxed(TIM_DIER_CC1IE, ce->regs + TIM_DIER);
>>
>> - writel_relaxed(data->periodic_top, base + TIM_ARR);
>> - writel_relaxed(TIM_CR1_ARPE | TIM_CR1_CEN, base + TIM_CR1);
>> return 0;
>> }
>>
>> -static int stm32_clock_event_set_next_event(unsigned long evt,
>> - struct clock_event_device *evtdev)
>> +static int stm32_clock_event_set_periodic(struct clock_event_device *evtdev)
>> {
>> - struct stm32_clock_event_ddata *data =
>> - container_of(evtdev, struct stm32_clock_event_ddata, evtdev);
>> + struct stm32_clock_event *ce =
>> + container_of(evtdev, struct stm32_clock_event, evtdev);
>>
>> - writel_relaxed(evt, data->base + TIM_ARR);
>> - writel_relaxed(TIM_CR1_ARPE | TIM_CR1_OPM | TIM_CR1_CEN,
>> - data->base + TIM_CR1);
>> + return stm32_clock_event_set_next_event(ce->periodic_top, evtdev);
>> +}
>>
>> - return 0;
>> +static int stm32_clock_event_set_oneshot(struct clock_event_device *evtdev)
>> +{
>> + return stm32_clock_event_set_next_event(0, evtdev);
>> }
>>
>> static irqreturn_t stm32_clock_event_handler(int irq, void *dev_id)
>> {
>> - struct stm32_clock_event_ddata *data = dev_id;
>> + struct stm32_clock_event *ce = dev_id;
>> +
>> + writel_relaxed(0, ce->regs + TIM_SR);
>>
>> - writel_relaxed(0, data->base + TIM_SR);
>> + if (clockevent_state_periodic(&ce->evtdev))
>> + stm32_clock_event_set_periodic(&ce->evtdev);
>
> nit: else condition to prevent an extra check
OK
>
>> - data->evtdev.event_handler(&data->evtdev);
>> + if (clockevent_state_oneshot(&ce->evtdev))
>> + stm32_clock_event_shutdown(&ce->evtdev);
>> +
>> + ce->evtdev.event_handler(&ce->evtdev);
>>
>> return IRQ_HANDLED;
>> }
>>
>> -static struct stm32_clock_event_ddata clock_event_ddata = {
>> - .evtdev = {
>> - .name = "stm32 clockevent",
>> - .features = CLOCK_EVT_FEAT_ONESHOT | CLOCK_EVT_FEAT_PERIODIC,
>> - .set_state_shutdown = stm32_clock_event_shutdown,
>> - .set_state_periodic = stm32_clock_event_set_periodic,
>> - .set_state_oneshot = stm32_clock_event_shutdown,
>> - .tick_resume = stm32_clock_event_shutdown,
>> - .set_next_event = stm32_clock_event_set_next_event,
>> - .rating = 200,
>> - },
>> -};
>> +static int __init stm32_clockevent_init(struct device_node *np,
>> + void __iomem *base,
>> + struct clk *clk, int irq)
>> +{
>> + struct stm32_clock_event *ce;
>> + unsigned long rate;
>> + int err;
>> +
>> + ce = kzalloc(sizeof(*ce), GFP_KERNEL);
>> + if (!ce)
>> + return -ENOMEM;
>> +
>> + ce->regs = base;
>> + ce->evtdev.name = "stm32_clockevent";
>> + ce->evtdev.features = CLOCK_EVT_FEAT_ONESHOT | CLOCK_EVT_FEAT_PERIODIC;
>> + ce->evtdev.set_state_shutdown = stm32_clock_event_shutdown;
>> + ce->evtdev.set_state_periodic = stm32_clock_event_set_periodic;
>> + ce->evtdev.set_state_oneshot = stm32_clock_event_set_oneshot;
>> + ce->evtdev.tick_resume = stm32_clock_event_shutdown;
>> + ce->evtdev.set_next_event = stm32_clock_event_set_next_event;
>> + ce->evtdev.rating = 200;
>>
>> -static int __init stm32_clockevent_init(struct device_node *np)
>> + rate = clk_get_rate(clk);
>> + ce->periodic_top = DIV_ROUND_CLOSEST(rate, HZ);
>> +
>> + writel_relaxed(0, ce->regs + TIM_DIER);
>> + writel_relaxed(0, ce->regs + TIM_SR);
>> +
>> + err = request_irq(irq, stm32_clock_event_handler, IRQF_TIMER,
>> + "stm32 clockevent", ce);
>> + if (err) {
>> + kfree(ce);
>> + return err;
>> + }
>> +
>> + clockevents_config_and_register(&ce->evtdev, rate, 0x60, ~0U);
>> +
>> + return 0;
>> +}
>> +
>> +static void __iomem *stm32_timer_cnt __read_mostly;
>> +static u64 notrace stm32_read_sched_clock(void)
>> +{
>> + return readl_relaxed(stm32_timer_cnt);
>> +}
>> +
>> +static int __init stm32_clocksource_init(struct device_node *node,
>> + void __iomem *regs,
>> + struct clk *clk)
>> +{
>> + unsigned long rate;
>> +
>> + rate = clk_get_rate(clk);
>> +
>> + writel_relaxed(~0U, regs + TIM_ARR);
>> + writel_relaxed(0, regs + TIM_PSC);
>> + writel_relaxed(0, regs + TIM_SR);
>> + writel_relaxed(0, regs + TIM_DIER);
>> + writel_relaxed(0, regs + TIM_SR);
>> + writel_relaxed(TIM_CR1_ARPE | TIM_CR1_UDIS, regs + TIM_CR1);
>> +
>> + /* Make sure that registers are updated */
>> + writel_relaxed(TIM_EGR_UG, regs + TIM_EGR);
>> +
>> + /* Enable controller */
>> + writel_relaxed(TIM_CR1_ARPE | TIM_CR1_UDIS | TIM_CR1_CEN,
>> + regs + TIM_CR1);
>> +
>> + stm32_timer_cnt = regs + TIM_CNT;
>> + sched_clock_register(stm32_read_sched_clock, 32, rate);
>> +
>> + return clocksource_mmio_init(stm32_timer_cnt, "stm32_timer",
>> + rate, 250, 32, clocksource_mmio_readl_up);
>> +}
>> +
>> +static int __init stm32_timer_init(struct device_node *node)
>> {
>> - struct stm32_clock_event_ddata *data = &clock_event_ddata;
>> - struct clk *clk;
>> struct reset_control *rstc;
>> - unsigned long rate, max_delta;
>> - int irq, ret, bits, prescaler = 1;
>> + void __iomem *timer_base;
>> + unsigned long max_arr;
>> + struct clk *clk;
>> + int irq, err;
>>
>> - clk = of_clk_get(np, 0);
>> - if (IS_ERR(clk)) {
>> - ret = PTR_ERR(clk);
>> - pr_err("failed to get clock for clockevent (%d)\n", ret);
>> - goto err_clk_get;
>> + timer_base = of_io_request_and_map(node, 0, of_node_full_name(node));
>> + if (IS_ERR(timer_base)) {
>> + pr_err("Can't map registers\n");
>> + goto out;
>> }
>>
>> - ret = clk_prepare_enable(clk);
>> - if (ret) {
>> - pr_err("failed to enable timer clock for clockevent (%d)\n",
>> - ret);
>> - goto err_clk_enable;
>> + irq = irq_of_parse_and_map(node, 0);
>> + if (irq <= 0) {
>> + pr_err("Can't parse IRQ\n");
>> + goto out_unmap;
>> }
>>
>> - rate = clk_get_rate(clk);
>
> Why not pass the rate to clkevt_init and clksrc_init instead of clk? So
> clk_get_rate() is not called twice.
I will change that in v3
>
>> + clk = of_clk_get(node, 0);
>> + if (IS_ERR(clk)) {
>> + pr_err("Can't get timer clock\n");
>> + goto out_unmap;
>> + }
>>
>> - rstc = of_reset_control_get(np, NULL);
>> + rstc = of_reset_control_get(node, NULL);
>> if (!IS_ERR(rstc)) {
>> reset_control_assert(rstc);
>> reset_control_deassert(rstc);
>> }
>>
>> - data->base = of_iomap(np, 0);
>> - if (!data->base) {
>> - ret = -ENXIO;
>> - pr_err("failed to map registers for clockevent\n");
>> - goto err_iomap;
>> - }
>> -
>> - irq = irq_of_parse_and_map(np, 0);
>> - if (!irq) {
>> - ret = -EINVAL;
>> - pr_err("%pOF: failed to get irq.\n", np);
>> - goto err_get_irq;
>> + err = clk_prepare_enable(clk);
>> + if (err) {
>> + pr_err("Couldn't enable parent clock\n");
>> + goto out_clk;
>> }
>>
>> /* Detect whether the timer is 16 or 32 bits */
>> - writel_relaxed(~0U, data->base + TIM_ARR);
>> - max_delta = readl_relaxed(data->base + TIM_ARR);
>> - if (max_delta == ~0U) {
>> - prescaler = 1;
>> - bits = 32;
>> - } else {
>> - prescaler = 1024;
>> - bits = 16;
>> + writel_relaxed(~0U, timer_base + TIM_ARR);
>> + max_arr = readl_relaxed(timer_base + TIM_ARR);
>> + if (max_arr != ~0U) {
>> + err = -EINVAL;
>> + pr_err("32 bits timer is needed\n");
>> + goto out_unprepare;
>> }
>> - writel_relaxed(0, data->base + TIM_ARR);
>> -
>> - writel_relaxed(prescaler - 1, data->base + TIM_PSC);
>> - writel_relaxed(TIM_EGR_UG, data->base + TIM_EGR);
>> - writel_relaxed(TIM_DIER_UIE, data->base + TIM_DIER);
>> - writel_relaxed(0, data->base + TIM_SR);
>> -
>> - data->periodic_top = DIV_ROUND_CLOSEST(rate, prescaler * HZ);
>>
>> - clockevents_config_and_register(&data->evtdev,
>> - DIV_ROUND_CLOSEST(rate, prescaler),
>> - 0x1, max_delta);
>> + err = stm32_clocksource_init(node, timer_base, clk);
>> + if (err)
>> + goto out_unprepare;
>>
>> - ret = request_irq(irq, stm32_clock_event_handler, IRQF_TIMER,
>> - "stm32 clockevent", data);
>> - if (ret) {
>> - pr_err("%pOF: failed to request irq.\n", np);
>> - goto err_get_irq;
>> - }
>> -
>> - pr_info("%pOF: STM32 clockevent driver initialized (%d bits)\n",
>> - np, bits);
>> + err = stm32_clockevent_init(node, timer_base, clk, irq);
>> + if (err)
>> + goto out_unprepare;
>>
>> - return ret;
>> + return 0;
>>
>> -err_get_irq:
>> - iounmap(data->base);
>> -err_iomap:
>> +out_unprepare:
>> clk_disable_unprepare(clk);
>> -err_clk_enable:
>> +out_clk:
>> clk_put(clk);
>> -err_clk_get:
>> - return ret;
>> +out_unmap:
>> + iounmap(timer_base);
>> +out:
>> + return err;
>> }
>>
>> -TIMER_OF_DECLARE(stm32, "st,stm32-timer", stm32_clockevent_init);
>> +CLOCKSOURCE_OF_DECLARE(stm32, "st,stm32-timer", stm32_timer_init);
>
> CLOCKSOURCE_OF_DECLARE is deprecated, keep using TIMER_OF_DECLARE.
OK
>
>
> --
> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>
> Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog
>
More information about the linux-arm-kernel
mailing list