[v1,2/2] Add PWM driver for LGM

Add PWM controller driver for Intel's Lightning Mountain(LGM) SoC.

Signed-off-by: Rahul Tanwar <rahul.tanwar@linux.intel.com>
---
 drivers/pwm/Kconfig         |   9 ++
 drivers/pwm/Makefile        |   1 +
 drivers/pwm/pwm-intel-lgm.c | 356 ++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 366 insertions(+)
 create mode 100644 drivers/pwm/pwm-intel-lgm.c

Hello,

On Fri, May 22, 2020 at 03:41:59PM +0800, Rahul Tanwar wrote:
> Add PWM controller driver for Intel's Lightning Mountain(LGM) SoC.
> 
> Signed-off-by: Rahul Tanwar <rahul.tanwar@linux.intel.com>
> ---
>  drivers/pwm/Kconfig         |   9 ++
>  drivers/pwm/Makefile        |   1 +
>  drivers/pwm/pwm-intel-lgm.c | 356 ++++++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 366 insertions(+)
>  create mode 100644 drivers/pwm/pwm-intel-lgm.c
> 
> diff --git a/drivers/pwm/Kconfig b/drivers/pwm/Kconfig
> index eebbc917ac97..a582214f50b2 100644
> --- a/drivers/pwm/Kconfig
> +++ b/drivers/pwm/Kconfig
> @@ -232,6 +232,15 @@ config PWM_IMX_TPM
>  	  To compile this driver as a module, choose M here: the module
>  	  will be called pwm-imx-tpm.
>  
> +config PWM_INTEL_LGM
> +	tristate "Intel LGM PWM support"
> +	depends on X86 || COMPILE_TEST
> +	help
> +	  Generic PWM framework driver for LGM SoC.

I wouldn't call this "framework".

> +	  To compile this driver as a module, choose M here: the module
> +	  will be called pwm-intel-lgm.
> +
>  config PWM_JZ4740
>  	tristate "Ingenic JZ47xx PWM support"
>  	depends on MACH_INGENIC
> diff --git a/drivers/pwm/Makefile b/drivers/pwm/Makefile
> index 9a475073dafc..c16a972a101d 100644
> --- a/drivers/pwm/Makefile
> +++ b/drivers/pwm/Makefile
> @@ -20,6 +20,7 @@ obj-$(CONFIG_PWM_IMG)		+= pwm-img.o
>  obj-$(CONFIG_PWM_IMX1)		+= pwm-imx1.o
>  obj-$(CONFIG_PWM_IMX27)		+= pwm-imx27.o
>  obj-$(CONFIG_PWM_IMX_TPM)	+= pwm-imx-tpm.o
> +obj-$(CONFIG_PWM_INTEL_LGM)	+= pwm-intel-lgm.o
>  obj-$(CONFIG_PWM_JZ4740)	+= pwm-jz4740.o
>  obj-$(CONFIG_PWM_LP3943)	+= pwm-lp3943.o
>  obj-$(CONFIG_PWM_LPC18XX_SCT)	+= pwm-lpc18xx-sct.o
> diff --git a/drivers/pwm/pwm-intel-lgm.c b/drivers/pwm/pwm-intel-lgm.c
> new file mode 100644
> index 000000000000..e307fd2457df
> --- /dev/null
> +++ b/drivers/pwm/pwm-intel-lgm.c
> @@ -0,0 +1,356 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (C) 2020 Intel Corporation.
> + */

Please point out limitations of your driver here similar to e.g.
drivers/pwm/pwm-sifive.c (and in the same format).

You should mention the fixed period of the hardware here (assuming I
interpreted this correctly below) and missing support for polarity.

How does the PWM behave on disable? Does the output become inactive
then? Does it complete the currently running period? What about
reconfiguration? Does changing the duty cycle complete the currently
running period?

> +#include <linux/bitfield.h>
> +#include <linux/clk.h>
> +#include <linux/module.h>
> +#include <linux/of_device.h>
> +#include <linux/pwm.h>
> +#include <linux/regmap.h>
> +#include <linux/reset.h>
> +
> +#define PWM_FAN_CON0		0x0
> +#define PWM_FAN_EN_EN		BIT(0)
> +#define PWM_FAN_EN_DIS		0x0
> +#define PWM_FAN_EN_MSK		BIT(0)
> +#define PWM_FAN_MODE_2WIRE	0x0
> +#define PWM_FAN_MODE_4WIRE	0x1
> +#define PWM_FAN_MODE_MSK	BIT(1)
> +#define PWM_FAN_PWM_DIS_DIS	0x0
> +#define PWM_FAN_PWM_DIS_MSK	BIT(2)
> +#define PWM_TACH_EN_EN		0x1
> +#define PWM_TACH_EN_MSK		BIT(4)
> +#define PWM_TACH_PLUS_2		0x0
> +#define PWM_TACH_PLUS_4		0x1
> +#define PWM_TACH_PLUS_MSK	BIT(5)
> +#define PWM_FAN_DC_MSK		GENMASK(23, 16)
> +
> +#define PWM_FAN_CON1		0x4
> +#define PWM_FAN_MAX_RPM_MSK	GENMASK(15, 0)
> +
> +#define PWM_FAN_STAT		0x10
> +#define PWM_FAN_TACH_MASK	GENMASK(15, 0)
> +
> +#define MAX_RPM			(BIT(16) - 1)
> +#define DFAULT_RPM		4000
> +#define MAX_DUTY_CYCLE		(BIT(8) - 1)
> +
> +#define FRAC_BITS		10
> +#define TWO_TENTH		204
> +
> +#define TWO_SECONDS		2000
> +#define IGNORE_FIRST_ERR	1
> +#define THIRTY_SECS_WINDOW	15
> +#define ERR_CNT_THRESHOLD	6
> +
> +struct intel_pwm_chip {
> +	struct pwm_chip		chip;
> +	struct regmap		*regmap;
> +	struct clk		*clk;
> +	struct reset_control	*rst;
> +	u32			tach_en;
> +	u32			max_rpm;
> +	u32			set_rpm;
> +	u32			set_dc;
> +	struct delayed_work	work;
> +};

I don't like aligning the member names and prefer a single space here.
(But I'm aware this is subjective and others like it the why you did
it.)

> +
> +static inline struct intel_pwm_chip *to_intel_pwm_chip(struct pwm_chip *chip)
> +{
> +	return container_of(chip, struct intel_pwm_chip, chip);
> +}
> +
> +static int pwm_update_dc(struct intel_pwm_chip *pc, u32 val)

Please use a common function prefix for all functions in your driver.

> +{
> +	return regmap_update_bits(pc->regmap, PWM_FAN_CON0, PWM_FAN_DC_MSK,
> +				  FIELD_PREP(PWM_FAN_DC_MSK, val));
> +}
> +
> +static int intel_pwm_config(struct pwm_chip *chip, struct pwm_device *pwm,
> +			    int duty_ns, int period_ns)
> +{
> +	struct intel_pwm_chip *pc = to_intel_pwm_chip(chip);
> +	u32 val;
> +
> +	val = DIV_ROUND_CLOSEST(duty_ns * MAX_DUTY_CYCLE, period_ns);
> +	val = min_t(u32, val, MAX_DUTY_CYCLE);

The period is fixed for this hardware type? If so you're supposed to
refuse requests for a period smaller than the actual period length.

Also duty_cycle is supposed to round down.

(That's something that PWM_DEBUG will tell you, too)

> +	if (pc->tach_en) {
> +		pc->set_dc = val;
> +		pc->set_rpm = val * pc->max_rpm / MAX_DUTY_CYCLE;
> +	}
> +
> +	return pwm_update_dc(pc, val);
> +}
> +
> +static int intel_pwm_enable(struct pwm_chip *chip, struct pwm_device *pwm)
> +{
> +	struct intel_pwm_chip *pc = to_intel_pwm_chip(chip);
> +	struct regmap *regmap = pc->regmap;
> +
> +	regmap_update_bits(regmap, PWM_FAN_CON0,
> +			   PWM_FAN_EN_MSK, PWM_FAN_EN_EN);
> +
> +	if (pc->tach_en)
> +		schedule_delayed_work(&pc->work, msecs_to_jiffies(10000));
> +
> +	return 0;
> +}
> +
> +static void intel_pwm_disable(struct pwm_chip *chip, struct pwm_device *pwm)
> +{
> +	struct intel_pwm_chip *pc = to_intel_pwm_chip(chip);
> +	struct regmap *regmap = pc->regmap;
> +
> +	if (pc->tach_en)
> +		cancel_delayed_work_sync(&pc->work);
> +
> +	regmap_update_bits(regmap, PWM_FAN_CON0,
> +			   PWM_FAN_EN_MSK, PWM_FAN_EN_DIS);
> +}
> +
> +static const struct pwm_ops intel_pwm_ops = {
> +	.config		= intel_pwm_config,
> +	.enable		= intel_pwm_enable,
> +	.disable	= intel_pwm_disable,
> +	.owner		= THIS_MODULE,

Please implement .apply instead of .config/.enable/.disable. Please also
enable PWM_DEBUG and fix the issues pointed out with it.

> +};
> +
> +static void tach_work(struct work_struct *work)
> +{
> +	struct intel_pwm_chip *pc = container_of(work, struct intel_pwm_chip,
> +						 work.work);
> +	struct regmap *regmap = pc->regmap;
> +	u32 fan_tach, fan_dc, val;
> +	s32 diff;
> +	static u32 fanspeed_err_cnt, time_window, delta_dc;
> +
> +	/*
> +	 * Fan speed is tracked by reading the active duty cycle of PWM output
> +	 * from the active duty cycle register. Some variance in the duty cycle
> +	 * register value is expected. So we set a time window of 30 seconds and
> +	 * if we detect inaccurate fan speed 6 times within 30 seconds then we
> +	 * mark it as fan speed problem and fix it by readjusting the duty cycle.
> +	 */

I'm a unhappy to have this in the PWM driver. The PWM driver is supposed
to be generic and I think this belongs into a dedicated driver.

> +
> +	if (fanspeed_err_cnt > IGNORE_FIRST_ERR)
> +		/*
> +		 * Ignore first time we detect inaccurate fan speed
> +		 * because it is expected during bootup.
> +		 */
> +		time_window++;
> +
> +	if (time_window == THIRTY_SECS_WINDOW) {
> +		/*
> +		 * This work is scheduled every 2 seconds i.e. each time_window
> +		 * counter step roughly mean 2 seconds. When the time window
> +		 * reaches 30 seconds, reset all the counters/logic.
> +		 */
> +		fanspeed_err_cnt = 0;
> +		delta_dc = 0;
> +		time_window = 0;
> +	}
> +
> +	regmap_read(regmap, PWM_FAN_STAT, &fan_tach);
> +	fan_tach &= PWM_FAN_TACH_MASK;
> +	if (!fan_tach)
> +		goto restart_work;
> +
> +	val = DIV_ROUND_CLOSEST(pc->set_rpm << FRAC_BITS, fan_tach);
> +	diff = val - BIT(FRAC_BITS);
> +
> +	if (abs(diff) > TWO_TENTH) {
> +		/* if duty cycle diff is more than two tenth, detect it as error */
> +		if (fanspeed_err_cnt > IGNORE_FIRST_ERR)
> +			delta_dc += val;
> +		fanspeed_err_cnt++;
> +	}
> +
> +	if (fanspeed_err_cnt == ERR_CNT_THRESHOLD) {
> +		/*
> +		 * We detected fan speed errors 6 times with 30 seconds.
> +		 * Fix the error by readjusting duty cycle and reset
> +		 * our counters/logic.
> +		 */
> +		fan_dc = pc->set_dc * delta_dc >> (FRAC_BITS + 2);
> +		fan_dc = min_t(u32, fan_dc, MAX_DUTY_CYCLE);
> +		pwm_update_dc(pc, fan_dc);
> +		fanspeed_err_cnt = 0;
> +		delta_dc = 0;
> +		time_window = 0;
> +	}
> +
> +restart_work:
> +	/*
> +	 * Fan speed doesn't need continous tracking. Schedule this work
> +	 * every two seconds so it doesn't steal too much cpu cycles.
> +	 */
> +	schedule_delayed_work(&pc->work, msecs_to_jiffies(TWO_SECONDS));
> +}
> +
> +static void pwm_init(struct intel_pwm_chip *pc)
> +{
> +	struct device *dev = pc->chip.dev;
> +	struct regmap *regmap = pc->regmap;
> +	u32 max_rpm, fan_wire, tach_plus, con0_val, con0_mask;
> +
> +	if (device_property_read_u32(dev, "intel,fan-wire", &fan_wire))
> +		fan_wire = 2; /* default is 2 wire mode */
> +
> +	con0_val = FIELD_PREP(PWM_FAN_PWM_DIS_MSK, PWM_FAN_PWM_DIS_DIS);
> +	con0_mask = PWM_FAN_PWM_DIS_MSK | PWM_FAN_MODE_MSK;
> +
> +	switch (fan_wire) {
> +	case 2 ... 3:
> +		con0_val |= FIELD_PREP(PWM_FAN_MODE_MSK, PWM_FAN_MODE_2WIRE);
> +		break;

This can be dropped as it matches the default case.

> +	case 4:
> +		con0_val |= FIELD_PREP(PWM_FAN_MODE_MSK, PWM_FAN_MODE_4WIRE) |
> +			    FIELD_PREP(PWM_TACH_EN_MSK, PWM_TACH_EN_EN);
> +		con0_mask |= PWM_TACH_EN_MSK | PWM_TACH_PLUS_MSK;
> +		pc->tach_en = 1;
> +		break;
> +	default:
> +		/* default is 2wire mode */
> +		con0_val |= FIELD_PREP(PWM_FAN_MODE_MSK, PWM_FAN_MODE_2WIRE);
> +		break;
> +	}
> +
> +	if (pc->tach_en) {
> +		if (device_property_read_u32(dev, "intel,tach-plus",
> +					     &tach_plus))
> +			tach_plus = 2;
> +
> +		switch (tach_plus) {
> +		case 2:
> +			con0_val |= FIELD_PREP(PWM_TACH_PLUS_MSK,
> +					       PWM_TACH_PLUS_2);
> +			break;
> +		case 4:
> +			con0_val |= FIELD_PREP(PWM_TACH_PLUS_MSK,
> +					       PWM_TACH_PLUS_4);
> +			break;
> +		default:
> +			con0_val |= FIELD_PREP(PWM_TACH_PLUS_MSK,
> +					       PWM_TACH_PLUS_2);
> +			break;

Similarily here, case 2: and default are identical.

> +		}
> +
> +		if (device_property_read_u32(dev, "intel,max-rpm", &max_rpm))
> +			max_rpm = DFAULT_RPM;
> +
> +		max_rpm = min_t(u32, max_rpm, MAX_RPM);
> +		if (max_rpm == 0)
> +			max_rpm = DFAULT_RPM;
> +
> +		pc->max_rpm = max_rpm;
> +		INIT_DEFERRABLE_WORK(&pc->work, tach_work);
> +		regmap_update_bits(regmap, PWM_FAN_CON1,
> +				   PWM_FAN_MAX_RPM_MSK, max_rpm);
> +	}
> +
> +	regmap_update_bits(regmap, PWM_FAN_CON0, con0_mask, con0_val);
> +}
> +
> +static const struct regmap_config pwm_regmap_config = {
> +	.reg_bits       = 32,
> +	.reg_stride     = 4,
> +	.val_bits       = 32,
> +};
> +
> +static int intel_pwm_probe(struct platform_device *pdev)
> +{
> +	struct intel_pwm_chip *pc;
> +	struct device *dev = &pdev->dev;
> +	void __iomem *io_base;
> +	int ret;
> +
> +	pc = devm_kzalloc(dev, sizeof(*pc), GFP_KERNEL);
> +	if (!pc)
> +		return -ENOMEM;
> +
> +	io_base = devm_platform_ioremap_resource(pdev, 0);
> +	if (IS_ERR(io_base))

error message here?

> +		return PTR_ERR(io_base);
> +
> +	pc->regmap = devm_regmap_init_mmio(dev, io_base, &pwm_regmap_config);
> +	if (IS_ERR(pc->regmap)) {
> +		ret = PTR_ERR(pc->regmap);
> +		dev_err(dev, "failed to init register map: %d\n", ret);

Better use %pe for the error code to get a more descriptive message.

> +		return ret;
> +	}
> +
> +	pc->clk = devm_clk_get(dev, NULL);
> +	if (IS_ERR(pc->clk)) {
> +		ret = PTR_ERR(pc->clk);
> +		dev_err(dev, "failed to get clock: %d\n", ret);
> +		return ret;
> +	}
> +
> +	pc->rst = devm_reset_control_get(dev, NULL);
> +	if (IS_ERR(pc->rst)) {
> +		ret = PTR_ERR(pc->rst);
> +		dev_err(dev, "failed to get reset control: %d\n", ret);
> +		return ret;
> +	}
> +
> +	ret = clk_prepare_enable(pc->clk);
> +	if (ret) {
> +		dev_err(dev, "failed to enable clock\n");
> +		return ret;
> +	}
> +
> +	reset_control_deassert(pc->rst);

return value check is missing.

> +	pc->chip.dev = dev;
> +	pc->chip.ops = &intel_pwm_ops;
> +	pc->chip.npwm = 1;
> +
> +	pwm_init(pc);
> +
> +	ret = pwmchip_add(&pc->chip);
> +	if (ret < 0) {
> +		dev_err(dev, "failed to add PWM chip: %d\n", ret);
> +		clk_disable_unprepare(pc->clk);
> +		return ret;
> +	}
> +
> +	platform_set_drvdata(pdev, pc);
> +	return 0;
> +}
> +
> +static int intel_pwm_remove(struct platform_device *pdev)
> +{
> +	struct intel_pwm_chip *pc = platform_get_drvdata(pdev);
> +	int ret;
> +
> +	if (pc->tach_en)
> +		cancel_delayed_work_sync(&pc->work);
> +
> +	ret = pwmchip_remove(&pc->chip);
> +	if (ret < 0)
> +		return ret;
> +
> +	reset_control_assert(pc->rst);
> +
> +	clk_disable_unprepare(pc->clk);

In .probe() you first release reset and then enable the clock. It's good
style to do it the other way round in .remove().

Best regards
Uwe

On Fri, May 22, 2020 at 10:56:13AM +0200, Uwe Kleine-König wrote:
> On Fri, May 22, 2020 at 03:41:59PM +0800, Rahul Tanwar wrote:

> > +	io_base = devm_platform_ioremap_resource(pdev, 0);
> > +	if (IS_ERR(io_base))
> 
> error message here?

platform core provides it. No need to duplicate (esp. taking into consideration
that it can issue IIRC three different error messages depending on actual error).

> 
> > +		return PTR_ERR(io_base);

Hello Andy,

On Fri, May 22, 2020 at 12:18:24PM +0300, Andy Shevchenko wrote:
> On Fri, May 22, 2020 at 10:56:13AM +0200, Uwe Kleine-König wrote:
> > On Fri, May 22, 2020 at 03:41:59PM +0800, Rahul Tanwar wrote:
> 
> > > +	io_base = devm_platform_ioremap_resource(pdev, 0);
> > > +	if (IS_ERR(io_base))
> > 
> > error message here?
> 
> platform core provides it. No need to duplicate (esp. taking into consideration
> that it can issue IIRC three different error messages depending on actual error).

Ah, missed that. Indeed that's fine as is in the patch.

Thanks
Uwe

Hi Uwe,

Thanks for review.

On 22/5/2020 4:56 pm, Uwe Kleine-König wrote:
> Hello,
>
> On Fri, May 22, 2020 at 03:41:59PM +0800, Rahul Tanwar wrote:
>> Add PWM controller driver for Intel's Lightning Mountain(LGM) SoC.
>>
>> Signed-off-by: Rahul Tanwar <rahul.tanwar@linux.intel.com>
>> ---
>>  drivers/pwm/Kconfig         |   9 ++
>>  drivers/pwm/Makefile        |   1 +
>>  drivers/pwm/pwm-intel-lgm.c | 356 ++++++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 366 insertions(+)
>>  create mode 100644 drivers/pwm/pwm-intel-lgm.c
>>
>> diff --git a/drivers/pwm/Kconfig b/drivers/pwm/Kconfig
>> index eebbc917ac97..a582214f50b2 100644
>> --- a/drivers/pwm/Kconfig
>> +++ b/drivers/pwm/Kconfig
>> @@ -232,6 +232,15 @@ config PWM_IMX_TPM
>>  	  To compile this driver as a module, choose M here: the module
>>  	  will be called pwm-imx-tpm.
>>  
>> +config PWM_INTEL_LGM
>> +	tristate "Intel LGM PWM support"
>> +	depends on X86 || COMPILE_TEST
>> +	help
>> +	  Generic PWM framework driver for LGM SoC.
[...]
>> +};
>> +
>> +static void tach_work(struct work_struct *work)
>> +{
>> +	struct intel_pwm_chip *pc = container_of(work, struct intel_pwm_chip,
>> +						 work.work);
>> +	struct regmap *regmap = pc->regmap;
>> +	u32 fan_tach, fan_dc, val;
>> +	s32 diff;
>> +	static u32 fanspeed_err_cnt, time_window, delta_dc;
>> +
>> +	/*
>> +	 * Fan speed is tracked by reading the active duty cycle of PWM output
>> +	 * from the active duty cycle register. Some variance in the duty cycle
>> +	 * register value is expected. So we set a time window of 30 seconds and
>> +	 * if we detect inaccurate fan speed 6 times within 30 seconds then we
>> +	 * mark it as fan speed problem and fix it by readjusting the duty cycle.
>> +	 */
> I'm a unhappy to have this in the PWM driver. The PWM driver is supposed
> to be generic and I think this belongs into a dedicated driver.

Well noted about all other review concerns. I will rework the driver in v2.
However, i am not very sure about the above point - of having a separate
dedicated driver for tach_work because its logic is tightly coupled with
this driver.

Regards,
Rahul

On Wed, May 27, 2020 at 02:28:53PM +0800, Tanwar, Rahul wrote:
> On 22/5/2020 4:56 pm, Uwe Kleine-König wrote:
> > On Fri, May 22, 2020 at 03:41:59PM +0800, Rahul Tanwar wrote:

...

> > I'm a unhappy to have this in the PWM driver. The PWM driver is supposed
> > to be generic and I think this belongs into a dedicated driver.
> 
> Well noted about all other review concerns. I will rework the driver in v2.
> However, i am not very sure about the above point - of having a separate
> dedicated driver for tach_work because its logic is tightly coupled with
> this driver.

Actually I agree with Uwe.
Here is layering violation, i.e. provider and consumer in the same pot. It's
not good from design perspective.

On 27/5/2020 5:15 pm, Andy Shevchenko wrote:
> On Wed, May 27, 2020 at 02:28:53PM +0800, Tanwar, Rahul wrote:
>> On 22/5/2020 4:56 pm, Uwe Kleine-König wrote:
>>> On Fri, May 22, 2020 at 03:41:59PM +0800, Rahul Tanwar wrote:
> ...
>
>>> I'm a unhappy to have this in the PWM driver. The PWM driver is supposed
>>> to be generic and I think this belongs into a dedicated driver.
>> Well noted about all other review concerns. I will rework the driver in v2.
>> However, i am not very sure about the above point - of having a separate
>> dedicated driver for tach_work because its logic is tightly coupled with
>> this driver.
> Actually I agree with Uwe.
> Here is layering violation, i.e. provider and consumer in the same pot. It's
> not good from design perspective.
>

Just to clarify, the PWM controller in our SoC serves just one purpose which
is to control the fan. Its actually named as PWM Fan Controller. There is no
other generic usage or any other consumer for this PWM driver. So separating
out this part seems redundant to me. Also, if we separate it out as a
dedicated driver, this will endup as a very small daemon which is going to be
very hard to justify while upstreaming..

Regards,
Rahul