Re: [PATCH v8 2/2] pwm: sifive: Add a driver for SiFive SoC PWM

[Uwe Kleine-König]

Hello,

On Fri, Mar 01, 2019 at 04:23:19PM +0530, Yash Shah wrote:
> Adds a PWM driver for PWM chip present in SiFive's HiFive Unleashed SoC.
> 
> Signed-off-by: Wesley W. Terpstra <wesley@xxxxxxxxxx>
> [Atish: Various fixes and code cleanup]
> Signed-off-by: Atish Patra <atish.patra@xxxxxxx>
> Signed-off-by: Yash Shah <yash.shah@xxxxxxxxxx>
> ---
>  drivers/pwm/Kconfig      |  11 ++
>  drivers/pwm/Makefile     |   1 +
>  drivers/pwm/pwm-sifive.c | 345 +++++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 357 insertions(+)
>  create mode 100644 drivers/pwm/pwm-sifive.c
> 
> diff --git a/drivers/pwm/Kconfig b/drivers/pwm/Kconfig
> index a8f47df..4a61d1a 100644
> --- a/drivers/pwm/Kconfig
> +++ b/drivers/pwm/Kconfig
> @@ -380,6 +380,17 @@ config PWM_SAMSUNG
>  	  To compile this driver as a module, choose M here: the module
>  	  will be called pwm-samsung.
>  
> +config PWM_SIFIVE
> +	tristate "SiFive PWM support"
> +	depends on OF
> +	depends on COMMON_CLK
> +	depends on RISCV || COMPILE_TEST
> +	help
> +	  Generic PWM framework driver for SiFive SoCs.
> +
> +	  To compile this driver as a module, choose M here: the module
> +	  will be called pwm-sifive.
> +
>  config PWM_SPEAR
>  	tristate "STMicroelectronics SPEAr PWM support"
>  	depends on PLAT_SPEAR
> diff --git a/drivers/pwm/Makefile b/drivers/pwm/Makefile
> index 9c676a0..30089ca 100644
> --- a/drivers/pwm/Makefile
> +++ b/drivers/pwm/Makefile
> @@ -37,6 +37,7 @@ obj-$(CONFIG_PWM_RCAR)		+= pwm-rcar.o
>  obj-$(CONFIG_PWM_RENESAS_TPU)	+= pwm-renesas-tpu.o
>  obj-$(CONFIG_PWM_ROCKCHIP)	+= pwm-rockchip.o
>  obj-$(CONFIG_PWM_SAMSUNG)	+= pwm-samsung.o
> +obj-$(CONFIG_PWM_SIFIVE)	+= pwm-sifive.o
>  obj-$(CONFIG_PWM_SPEAR)		+= pwm-spear.o
>  obj-$(CONFIG_PWM_STI)		+= pwm-sti.o
>  obj-$(CONFIG_PWM_STM32)		+= pwm-stm32.o
> diff --git a/drivers/pwm/pwm-sifive.c b/drivers/pwm/pwm-sifive.c
> new file mode 100644
> index 0000000..6679ec7
> --- /dev/null
> +++ b/drivers/pwm/pwm-sifive.c
> @@ -0,0 +1,345 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (C) 2017-2018 SiFive
> + * For SiFive's PWM IP block documentation please refer Chapter 14 of
> + * Reference Manual : https://static.dev.sifive.com/FU540-C000-v1.0.pdf
> + *
> + * Limitations:
> + * - When changing both duty cycle and period, we cannot prevent in
> + *   software that the output might produce a period with mixed
> + *   settings (new period length and old duty cycle).
> + * - The hardware cannot generate a 100% duty cycle.
> + * - The hardware generates only inverted output.
> + */
> +#include <linux/clk.h>
> +#include <linux/io.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/pwm.h>
> +#include <linux/slab.h>
> +
> +/* Register offsets */
> +#define PWM_SIFIVE_PWMCFG		0x0
> +#define PWM_SIFIVE_PWMCOUNT		0x8
> +#define PWM_SIFIVE_PWMS			0x10
> +#define PWM_SIFIVE_PWMCMP0		0x20
> +
> +/* PWMCFG fields */
> +#define PWM_SIFIVE_PWMCFG_SCALE		0
> +#define PWM_SIFIVE_PWMCFG_STICKY	8
> +#define PWM_SIFIVE_PWMCFG_ZERO_CMP	9
> +#define PWM_SIFIVE_PWMCFG_DEGLITCH	10
> +#define PWM_SIFIVE_PWMCFG_EN_ALWAYS	BIT(12)
> +#define PWM_SIFIVE_PWMCFG_EN_ONCE	13
> +#define PWM_SIFIVE_PWMCFG_CENTER	16
> +#define PWM_SIFIVE_PWMCFG_GANG		24
> +#define PWM_SIFIVE_PWMCFG_IP		28

It's a bit inconsistent to have one of them use BIT and the others not.
For consistency please use BIT for all defines. (They are unused
anyhow.) For PWM_SIFIVE_PWMCFG_SCALE use:

	#define PWM_SIFIVE_PWMCFG_SCALE GENMASK(3, 0)

and then FIELD_GET and FIELD_PREP to access the values.

> +/* PWM_SIFIVE_SIZE_PWMCMP is used to calculate offset for pwmcmpX registers */
> +#define PWM_SIFIVE_SIZE_PWMCMP		4
> +#define PWM_SIFIVE_CMPWIDTH		16
> +#define PWM_SIFIVE_DEFAULT_PERIOD	10000000
> +
> +struct pwm_sifive_ddata {
> +	struct pwm_chip	chip;
> +	struct mutex lock; /* lock to protect user_count */
> +	struct notifier_block notifier;
> +	struct clk *clk;
> +	void __iomem *regs;
> +	unsigned int real_period;
> +	int user_count;
> +};
> +
> +static inline
> +struct pwm_sifive_ddata *pwm_sifive_chip_to_ddata(struct pwm_chip *c)
> +{
> +	return container_of(c, struct pwm_sifive_ddata, chip);
> +}
> +
> +static int pwm_sifive_request(struct pwm_chip *chip, struct pwm_device *dev)
> +{
> +	struct pwm_sifive_ddata *pwm = pwm_sifive_chip_to_ddata(chip);
> +
> +	mutex_lock(&pwm->lock);
> +	pwm->user_count++;
> +	mutex_unlock(&pwm->lock);
> +
> +	return 0;
> +}
> +
> +static void pwm_sifive_free(struct pwm_chip *chip, struct pwm_device *dev)
> +{
> +	struct pwm_sifive_ddata *pwm = pwm_sifive_chip_to_ddata(chip);
> +
> +	mutex_lock(&pwm->lock);
> +	pwm->user_count--;
> +	mutex_unlock(&pwm->lock);
> +}
> +
> +static void pwm_sifive_update_clock(struct pwm_sifive_ddata *pwm,
> +				    unsigned long rate)
> +{
> +	u32 val;
> +	unsigned long long num;
> +	/* (1 << (PWM_SIFIVE_CMPWIDTH+scale)) * 10^9/rate = real_period */
> +	unsigned long scale_pow =
> +			div64_ul(pwm->real_period * (u64)rate, NSEC_PER_SEC);
> +	int scale = clamp(ilog2(scale_pow) - PWM_SIFIVE_CMPWIDTH, 0, 0xf);

I tried for some time to verify the code here and this would have been
easier with a more verbose comment. Something like:

/*
 * The PWM unit is used with pwmzerocmp=0, so the only way to modify the
 * period length is using pwmscale which provides the number of bits the
 * counter is shifted before being feed to the comparators. A period
 * lasts (1 << (PWM_SIFIVE_CMPWIDTH + pwmscale)) clock ticks.
 */

There is a bad rounding effect here. I don't know the machine's details,
and so will consider a parent clock running at 250 MHz. So one clock tick
is 4 ns long and the smallest period length is 4 ns << 16 == 262144 ns.
Consider further an initial target period of 10000000 ns (which is
PWM_SIFIVE_DEFAULT_PERIOD).

The calculation here results in scale_pow = 2500000 and so scale = 5.

> +	val = PWM_SIFIVE_PWMCFG_EN_ALWAYS | (scale << PWM_SIFIVE_PWMCFG_SCALE);
> +	writel(val, pwm->regs + PWM_SIFIVE_PWMCFG);
> +
> +	/* As scale <= 15 the shift operation cannot overflow. */
> +	num = 1000000000ULL << (PWM_SIFIVE_CMPWIDTH + scale);
> +	pwm->real_period = div64_ul(num, rate);
> +	dev_dbg(pwm->chip.dev, "New real_period = %u ns\n", pwm->real_period);
> +}

Then real_period ends up being 8388608 ns. If now the input clk increases
to 750 MHz it is 8388608 ns which is being used as target period length
and the calculation results in:

	scale_pow = 6291456
	scale = 6
	real_period = 5592405

I'd claim it would be better to use scale = 7 here which results in
11184810 which is nearer to the initially targeted 10000000 ns. (But we
cannot be sure as there is no rounding guide for the PWM framework.)

But worse than that is that if the input clock goes back to 250 MHz we
start with real_period = 5592405 and we get

	scale_pow = 1398101
	scale = 4
	real_period = 4194304

so we're not going back to the state we had when the clk was initially
running at 250 MHz.

To get the result independent of the prior configuration you better use
the real targeted period length as input instead of the last configured
approximation.

> +static void pwm_sifive_get_state(struct pwm_chip *chip, struct pwm_device *dev,
> +				 struct pwm_state *state)
> +{
> +	struct pwm_sifive_ddata *pwm = pwm_sifive_chip_to_ddata(chip);
> +	u32 duty, val;
> +	unsigned long long num;
> +	int ret;
> +
> +	ret = clk_enable(pwm->clk);
> +	if (ret)
> +		return;

Ideally we'd report state->enabled = 0 if the clk was off. I don't know
how this could be done reliably though.

> +	duty = readl(pwm->regs + PWM_SIFIVE_PWMCMP0 +
> +		     dev->hwpwm * PWM_SIFIVE_SIZE_PWMCMP);
> +
> +	state->enabled = duty > 0;

If duty is bigger than 0 but PWM_SIFIVE_PWMCFG_EN_ALWAYS isn't set you
should report enabled = false, too.

> +	val = readl(pwm->regs + PWM_SIFIVE_PWMCFG);
> +	val &= 0x0F;

Maybe name that "scale" instead of "val"?

> +	num = 1000000000ULL << (PWM_SIFIVE_CMPWIDTH + val);

Is this (unsigned long long)NSEC_PER_SEC? (Ditto above in
pwm_sifive_update_clock().)

> +	pwm->real_period = div64_ul(num, clk_get_rate(pwm->clk));
> +
> +	state->period = pwm->real_period;
> +	state->duty_cycle =
> +		(u64)duty * pwm->real_period >> PWM_SIFIVE_CMPWIDTH;
> +	state->polarity = PWM_POLARITY_INVERSED;
> +
> +	clk_disable(pwm->clk);
> +}
> +
> +static int pwm_sifive_enable(struct pwm_chip *chip, bool enable)
> +{
> +	struct pwm_sifive_ddata *pwm = pwm_sifive_chip_to_ddata(chip);
> +	int ret;
> +
> +	if (enable) {
> +		ret = clk_enable(pwm->clk);
> +		if (ret) {
> +			dev_err(pwm->chip.dev, "Enable clk failed:%d\n", ret);
> +			return ret;
> +		}
> +	}
> +
> +	if (!enable)
> +		clk_disable(pwm->clk);
> +
> +	return 0;
> +}
> +
> +static int pwm_sifive_apply(struct pwm_chip *chip, struct pwm_device *dev,
> +			    struct pwm_state *state)
> +{
> +	struct pwm_sifive_ddata *pwm = pwm_sifive_chip_to_ddata(chip);
> +	unsigned int duty_cycle, x;
> +	u32 frac;
> +	struct pwm_state cur_state;
> +	bool enabled;
> +	int ret = 0;
> +	unsigned long num;
> +
> +	if (state->polarity != PWM_POLARITY_INVERSED)
> +		return -EINVAL;
> +
> +	mutex_lock(&pwm->lock);
> +	pwm_get_state(dev, &cur_state);
> +	enabled = cur_state.enabled;
> +
> +	if (state->period != cur_state.period) {

Did you test this with more than one consumer? For sure the following
should work:

	pwm1 = pwm_get(.. the first ..);
	pwm_apply_state(pwm1, { .enabled = true, .period = 10000000, .... });

	pwm2 = pwm_get(.. the second ..);
	pwm_apply_state(pwm2, { .enabled = true, .period = 10000000, .... });

but for the second pwm_apply_state() run state->period is likely not
exactly 10000000.

> +		if (pwm->user_count != 1) {
> +			ret = -EINVAL;

EBUSY?

> +			goto exit;
> +		}
> +		pwm->real_period = state->period;
> +		pwm_sifive_update_clock(pwm, clk_get_rate(pwm->clk));

It's not ensured that pwm->clk is enabled here which is a pre-condition
to be allowed to call clk_get_rate().

> +	}
> +
> +	duty_cycle = state->duty_cycle;
> +	if (!state->enabled)
> +		duty_cycle = 0;
> +
> +	x = 1U << PWM_SIFIVE_CMPWIDTH;

"x" is a bad name.

> +	num = (u64)duty_cycle * x + x / 2;
> +	frac = div_u64(num, state->period);

I don't understand the "+ x / 2" part. Should this better be
"+ state->period / 2"? Something like

#define div_u64_round(a, b) ({typeof(b) __b = b; div_u64(a + __b / 2, __b)})

would make this less error prone.

> +	/* The hardware cannot generate a 100% duty cycle */
> +	frac = min(frac, x - 1);
> +
> +	writel(frac, pwm->regs + PWM_SIFIVE_PWMCMP0 +
> +	       dev->hwpwm * PWM_SIFIVE_SIZE_PWMCMP);
> +
> +	if (!state->enabled && enabled) {
> +		ret = pwm_sifive_enable(chip, false);
> +		if (ret)
> +			goto exit;
> +		enabled = false;
> +	}
> +
> +	if (state->enabled && !enabled) {
> +		ret = pwm_sifive_enable(chip, state->enabled);
> +		if (ret)
> +			goto exit;
> +	}

These two ifs can be combined to:

	if (state->enabled != enabled)
		ret = pwm_sifive_enable(chip, state->enabled);

> +
> +exit:
> +	mutex_unlock(&pwm->lock);
> +	return ret;
> +}
> +
> +static const struct pwm_ops pwm_sifive_ops = {
> +	.request = pwm_sifive_request,
> +	.free = pwm_sifive_free,
> +	.get_state = pwm_sifive_get_state,
> +	.apply = pwm_sifive_apply,
> +	.owner = THIS_MODULE,
> +};
> +
> +static int pwm_sifive_clock_notifier(struct notifier_block *nb,
> +				     unsigned long event, void *data)
> +{
> +	struct clk_notifier_data *ndata = data;
> +	struct pwm_sifive_ddata *pwm =
> +		container_of(nb, struct pwm_sifive_ddata, notifier);
> +
> +	if (event == POST_RATE_CHANGE)
> +		pwm_sifive_update_clock(pwm, ndata->new_rate);
> +
> +	return NOTIFY_OK;
> +}
> +
> +static int pwm_sifive_probe(struct platform_device *pdev)
> +{
> +	struct device *dev = &pdev->dev;
> +	struct pwm_sifive_ddata *pwm;
> +	struct pwm_chip *chip;
> +	struct resource *res;
> +	int ret, ch;
> +	bool is_enabled = false;
> +
> +	pwm = devm_kzalloc(dev, sizeof(*pwm), GFP_KERNEL);
> +	if (!pwm)
> +		return -ENOMEM;
> +
> +	mutex_init(&pwm->lock);
> +	chip = &pwm->chip;
> +	chip->dev = dev;
> +	chip->ops = &pwm_sifive_ops;
> +	chip->of_pwm_n_cells = 3;
> +	chip->base = -1;
> +	chip->npwm = 4;
> +
> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	pwm->regs = devm_ioremap_resource(dev, res);
> +	if (IS_ERR(pwm->regs)) {
> +		dev_err(dev, "Unable to map IO resources\n");
> +		return PTR_ERR(pwm->regs);
> +	}
> +
> +	pwm->clk = devm_clk_get(dev, NULL);
> +	if (IS_ERR(pwm->clk)) {
> +		if (PTR_ERR(pwm->clk) != -EPROBE_DEFER)
> +			dev_err(dev, "Unable to find controller clock\n");
> +		return PTR_ERR(pwm->clk);
> +	}
> +
> +	ret = clk_prepare_enable(pwm->clk);
> +	if (ret) {
> +		dev_err(dev, "failed to enable clock for pwm: %d\n", ret);
> +		return ret;
> +	}
> +
> +	/* Watch for changes to underlying clock frequency */
> +	pwm->notifier.notifier_call = pwm_sifive_clock_notifier;
> +	ret = clk_notifier_register(pwm->clk, &pwm->notifier);
> +	if (ret) {
> +		dev_err(dev, "failed to register clock notifier: %d\n", ret);
> +		goto disable_clk;
> +	}
> +
> +	ret = pwmchip_add(chip);
> +	if (ret < 0) {
> +		dev_err(dev, "cannot register PWM: %d\n", ret);
> +		goto unregister_clk;
> +	}

After pwmchip_add is called the first consumer might appear...

> +	/* Initialize PWM */
> +	pwm->real_period = PWM_SIFIVE_DEFAULT_PERIOD;
> +	pwm_sifive_update_clock(pwm, clk_get_rate(pwm->clk));
> +
> +	for (ch = 0; ch < pwm->chip.npwm; ch++) {
> +		if (pwm_is_enabled(&pwm->chip.pwms[ch])) {
> +			is_enabled = true;
> +			break;
> +		}
> +	}
> +	if (!is_enabled)
> +		clk_disable(pwm->clk);

... so this should better be called after these initialisations.
(This also means you must not use pwm_is_enabled(), but I'd consider that
an upside, because this function is for PWM consumers, not
implementors.)

> +	platform_set_drvdata(pdev, pwm);
> +	dev_dbg(dev, "SiFive PWM chip registered %d PWMs\n", chip->npwm);
> +
> +	return 0;
> +
> +unregister_clk:
> +	clk_notifier_unregister(pwm->clk, &pwm->notifier);
> +disable_clk:
> +	clk_disable_unprepare(pwm->clk);
> +
> +	return ret;
> +}

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |