Message-ID: <88d17e4c-7dda-4e87-bb6e-58116d1b2f9f@linaro.org>
Date: Tue, 11 Feb 2025 09:27:17 +0100
From: Daniel Lezcano <daniel.lezcano@...aro.org>
To: Frank Li <Frank.li@....com>
Cc: "Rafael J. Wysocki" <rafael@...nel.org>, Zhang Rui <rui.zhang@...el.com>,
 Lukasz Luba <lukasz.luba@....com>, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Shawn Guo <shawnguo@...nel.org>,
 Sascha Hauer <s.hauer@...gutronix.de>,
 Pengutronix Kernel Team <kernel@...gutronix.de>,
 Fabio Estevam <festevam@...il.com>, Pengfei Li <pengfei.li_1@....com>,
 Marco Felsch <m.felsch@...gutronix.de>, linux-pm@...r.kernel.org,
 devicetree@...r.kernel.org, imx@...ts.linux.dev,
 linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
 Peng Fan <peng.fan@....com>
Subject: Re: [PATCH v4 2/2] thermal: imx91: Add support for i.MX91 thermal
 monitoring unit

On 31/01/2025 23:22, Frank Li wrote:
> On Fri, Jan 31, 2025 at 11:19:54AM +0100, Daniel Lezcano wrote:
>> On 16/12/2024 20:25, Frank Li wrote:
>>> From: Pengfei Li <pengfei.li_1@....com>
>>>
>>> Introduce support for the i.MX91 thermal monitoring unit, which features a
>>> single sensor for the CPU. The register layout differs from other chips,
>>> necessitating the creation of a dedicated file for this.
>>
>> Please a bit more information about the sensor (eg. resolution, I guess 1°C)
> 
> How about add
> 
> "This sensor provides a resolution of 1/64°C (6-bit fraction). For actual
> accuracy, refer to the datasheet, as it varies depending on the chip grade."

A bit of more information is welcome :) eg is it interrupt driven ?

Also Link: <link to the datasheet>

>>> Signed-off-by: Pengfei Li <pengfei.li_1@....com>
>>> Signed-off-by: Peng Fan <peng.fan@....com>
>>> Reviewed-by: Marco Felsch <m.felsch@...gutronix.de>
>>> Signed-off-by: Frank Li <Frank.Li@....com>
>>> ---
>>> Change from v3 to v4
>>> - Add Macro's review tag
>>> - Use devm_add_action()
>>> - Move pm_runtim_put before thermal_of_zone_register()
>>>
>>> change from v2 to v3
>>> - add IMX91_TMU_ prefix for register define
>>> - remove unused register define
>>> - fix missed pm_runtime_put() at error path in imx91_tmu_get_temp()
>>> - use dev variable in probe function
>>> - use pm_runtime_set_active() in probe
>>> - move START to imx91_tmu_get_temp()
>>> - use DEFINE_RUNTIME_DEV_PM_OPS()
>>> - keep set reset value because there are not sw "reset" bit in controller,
>>> uboot may change and enable tmu.
>>>
>>> change from v1 to v2
>>> - use low case for hexvalue
>>> - combine struct imx91_tmu and tmu_sensor
>>> - simplify imx91_tmu_start() and imx91_tmu_enable()
>>> - use s16 for imx91_tmu_get_temp(), which may negative value
>>> - use reverse christmas tree style
>>> - use run time pm
>>> - use oneshot to sample temp
>>> - register thermal zone after hardware init
>>> ---
>>>    drivers/thermal/Kconfig         |  10 ++
>>>    drivers/thermal/Makefile        |   1 +
>>>    drivers/thermal/imx91_thermal.c | 263 ++++++++++++++++++++++++++++++++++++++++
>>>    3 files changed, 274 insertions(+)
>>>
>>> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
>>> index d3f9686e26e71..da403ed86aeb1 100644
>>> --- a/drivers/thermal/Kconfig
>>> +++ b/drivers/thermal/Kconfig
>>> @@ -296,6 +296,16 @@ config IMX8MM_THERMAL
>>>    	  cpufreq is used as the cooling device to throttle CPUs when the passive
>>>    	  trip is crossed.
>>> +config IMX91_THERMAL
>>> +	tristate "Temperature sensor driver for NXP i.MX91 SoC"
>>> +	depends on ARCH_MXC || COMPILE_TEST
>>> +	depends on OF
>>
>> s/OF/THERMAL_OF/
>>
>>> +	help
>>> +	  Support for Temperature sensor found on NXP i.MX91 SoC.
>>> +	  It supports one critical trip point and one passive trip point. The
>>> +	  cpufreq is used as the cooling device to throttle CPUs when the passive
>>> +	  trip is crossed.
>>
>> This help message is inaccurate. It should describe the sensor not the
>> thermal configuration which is coming from the device tree for a specific
>> platform.
> 
> How about
> 
> It includes one sensor and six comparators. Each of them compares the
> temperature value (from the sensor) against the programmable threshold
> values. Support program the direction of comparison (greater than or
> less than).

Sounds good except the last sentence, I suggest " ... values. The 
direction of the comparison is configurable (greater / lesser than).

>>>    config K3_THERMAL
>>>    	tristate "Texas Instruments K3 thermal support"
>>>    	depends on ARCH_K3 || COMPILE_TEST
>>> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
>>> index 9abf43a74f2bb..08da241e6a598 100644
>>> --- a/drivers/thermal/Makefile
>>> +++ b/drivers/thermal/Makefile
>>> @@ -50,6 +50,7 @@ obj-$(CONFIG_ARMADA_THERMAL)	+= armada_thermal.o
>>>    obj-$(CONFIG_IMX_THERMAL)	+= imx_thermal.o
>>>    obj-$(CONFIG_IMX_SC_THERMAL)	+= imx_sc_thermal.o
>>>    obj-$(CONFIG_IMX8MM_THERMAL)	+= imx8mm_thermal.o
>>> +obj-$(CONFIG_IMX91_THERMAL)	+= imx91_thermal.o
>>>    obj-$(CONFIG_MAX77620_THERMAL)	+= max77620_thermal.o
>>>    obj-$(CONFIG_QORIQ_THERMAL)	+= qoriq_thermal.o
>>>    obj-$(CONFIG_DA9062_THERMAL)	+= da9062-thermal.o
>>> diff --git a/drivers/thermal/imx91_thermal.c b/drivers/thermal/imx91_thermal.c
>>> new file mode 100644
>>> index 0000000000000..ef5e8e181dd0f
>>> --- /dev/null
>>> +++ b/drivers/thermal/imx91_thermal.c
>>> @@ -0,0 +1,263 @@
>>> +// SPDX-License-Identifier: GPL-2.0
>>> +/*
>>> + * Copyright 2024 NXP.
>>> + */
>>> +
>>> +#include <linux/bitfield.h>
>>> +#include <linux/clk.h>
>>> +#include <linux/err.h>
>>> +#include <linux/iopoll.h>
>>> +#include <linux/nvmem-consumer.h>
>>> +#include <linux/module.h>
>>> +#include <linux/of.h>
>>> +#include <linux/of_device.h>
>>> +#include <linux/platform_device.h>
>>> +#include <linux/pm_runtime.h>
>>> +#include <linux/thermal.h>
>>> +
>>> +#define IMX91_TMU_STAT0				0x10
>>> +#define IMX91_TMU_STAT0_DRDY0_IF_MASK		BIT(16)
>>> +
>>> +#define IMX91_TMU_DATA0				0x20
>>> +
>>> +#define IMX91_TMU_CTRL1_SET			0x204
>>> +#define IMX91_TMU_CTRL1_CLR			0x208
>>> +#define IMX91_TMU_CTRL1_EN			BIT(31)
>>> +#define IMX91_TMU_CTRL1_START			BIT(30)
>>> +#define IMX91_TMU_CTRL1_STOP			BIT(29)
>>> +#define IMX91_TMU_CTRL1_RES_MASK		GENMASK(19, 18)
>>> +#define IMX91_TMU_CTRL1_MEAS_MODE_MASK		GENMASK(25, 24)
>>> +#define   IMX91_TMU_CTRL1_MEAS_MODE_SINGLE	0
>>> +#define   IMX91_TMU_CTRL1_MEAS_MODE_CONTINUES	1
>>> +#define   IMX91_TMU_CTRL1_MEAS_MODE_PERIODIC	2
>>> +
>>> +#define IMX91_TMU_REF_DIV			0x280
>>> +#define IMX91_TMU_DIV_EN			BIT(31)
>>> +#define IMX91_TMU_DIV_MASK			GENMASK(23, 16)
>>> +#define IMX91_TMU_DIV_MAX			255
>>> +
>>> +#define IMX91_TMU_PUD_ST_CTRL			0x2b0
>>> +#define IMX91_TMU_PUDL_MASK			GENMASK(23, 16)
>>> +
>>> +#define IMX91_TMU_TRIM1				0x2e0
>>> +#define IMX91_TMU_TRIM2				0x2f0
>>> +
>>> +#define IMX91_TMU_TEMP_LOW_LIMIT		-40000
>>> +#define IMX91_TMU_TEMP_HIGH_LIMIT		125000
>>> +
>>> +#define IMX91_TMU_DEFAULT_TRIM1_CONFIG		0xb561bc2d
>>> +#define IMX91_TMU_DEFAULT_TRIM2_CONFIG		0x65d4
>>> +
>>> +struct imx91_tmu {
>>> +	void __iomem *base;
>>> +	struct clk *clk;
>>> +	struct device *dev;
>>> +	struct thermal_zone_device *tzd;
>>
>> This field is pointless because used only in the probe function.
>>
>>> +};
>>> +
>>> +static void imx91_tmu_start(struct imx91_tmu *tmu, bool start)
>>> +{
>>> +	u32 val = start ? IMX91_TMU_CTRL1_START : IMX91_TMU_CTRL1_STOP;
>>> +
>>> +	writel_relaxed(val, tmu->base + IMX91_TMU_CTRL1_SET);
>>> +}
>>> +
>>> +static void imx91_tmu_enable(struct imx91_tmu *tmu, bool enable)
>>> +{
>>> +	u32 reg = enable ? IMX91_TMU_CTRL1_SET : IMX91_TMU_CTRL1_CLR;
>>> +
>>> +	writel_relaxed(IMX91_TMU_CTRL1_EN, tmu->base + reg);
>>> +}
>>> +
>>> +static int imx91_tmu_get_temp(struct thermal_zone_device *tz, int *temp)
>>> +{
>>> +	struct imx91_tmu *tmu = thermal_zone_device_priv(tz);
>>> +	s16 data;
>>> +	int ret;
>>> +	u32 val;
>>> +
>>> +	ret = pm_runtime_resume_and_get(tmu->dev);
>>> +	if (ret < 0)
>>> +		return ret;
>>> +
>>> +	imx91_tmu_start(tmu, true);
>>
>> Same question as [1]
>>
>> Do you really want to start and stop the sensor between two reads ?
> 
> This version use simple oneshot mode. So need call start every time.

The question remains the same. Why is one shot / single preferred than 
'continuous' ?

Stopping the measurement between two reads can have an impact on the 
accuracy and it adds an overhead when doing a poll read which is not 
desirable when reaching the thermal pressure where the mitigation kicks in.


>>> +	ret = readl_relaxed_poll_timeout(tmu->base + IMX91_TMU_STAT0, val,
>>> +					 val & IMX91_TMU_STAT0_DRDY0_IF_MASK, 1000, 40000);
>>> +	if (ret) {
>>> +		ret = -EAGAIN;
>>> +		goto out;
>>> +	}
>>> +
>>> +	/* DATA0 is 16bit signed number */
>>> +	data = readw_relaxed(tmu->base + IMX91_TMU_DATA0);
>>> +	*temp = data * 1000 / 64;
>>
>> cf units.h
>>
>> 	*temp = (data * MILLIDEGREE_PER_DEGREE) / A_LITERAL;
>>
>>> +	if (*temp < IMX91_TMU_TEMP_LOW_LIMIT || *temp > IMX91_TMU_TEMP_HIGH_LIMIT)
>>> +		ret = -EAGAIN;
>>> +
>>> +out:
>>> +	pm_runtime_put(tmu->dev);
>>> +
>>> +	return ret;
>>> +}
>>> +
>>> +static struct thermal_zone_device_ops tmu_tz_ops = {
>>> +	.get_temp = imx91_tmu_get_temp,
>>
>> Why not add the change_mode ops ?
> 
> I plan add later when add .set_trip_temp and irq support.
> 
> Frank
>>
>>> +};
>>> +
>>> +static int imx91_init_from_nvmem_cells(struct imx91_tmu *tmu)
>>> +{
>>> +	struct device *dev = tmu->dev;
>>> +	u32 trim1, trim2;
>>> +	int ret;
>>> +
>>> +	ret = nvmem_cell_read_u32(dev, "trim1", &trim1);
>>> +	if (ret)
>>> +		return ret;
>>> +
>>> +	ret = nvmem_cell_read_u32(dev, "trim2", &trim2);
>>> +	if (ret)
>>> +		return ret;
>>> +
>>> +	if (trim1 == 0 || trim2 == 0)
>>> +		return -EINVAL;
>>> +
>>> +	writel_relaxed(trim1, tmu->base + IMX91_TMU_TRIM1);
>>> +	writel_relaxed(trim2, tmu->base + IMX91_TMU_TRIM2);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static void imx91_tmu_action_remove(void *data)
>>> +{
>>> +	struct imx91_tmu *tmu = data;
>>> +
>>> +	/* disable tmu */
>>> +	imx91_tmu_enable(tmu, false);
>>> +}
>>> +
>>> +static int imx91_tmu_probe(struct platform_device *pdev)
>>> +{
>>> +	struct device *dev = &pdev->dev;
>>> +	struct imx91_tmu *tmu;
>>> +	unsigned long rate;
>>> +	u32 div;
>>> +	int ret;
>>> +
>>> +	tmu = devm_kzalloc(dev, sizeof(struct imx91_tmu), GFP_KERNEL);
>>> +	if (!tmu)
>>> +		return -ENOMEM;
>>> +
>>> +	tmu->dev = dev;
>>> +
>>> +	tmu->base = devm_platform_ioremap_resource(pdev, 0);
>>> +	if (IS_ERR(tmu->base))
>>> +		return dev_err_probe(dev, PTR_ERR(tmu->base), "failed to get io resource");
>>> +
>>> +	tmu->clk = devm_clk_get_enabled(dev, NULL);
>>> +	if (IS_ERR(tmu->clk))
>>> +		return dev_err_probe(dev, PTR_ERR(tmu->clk), "failed to get tmu clock\n");
>>> +
>>> +	platform_set_drvdata(pdev, tmu);
>>> +
>>> +	/* disable the monitor during initialization */
>>> +	imx91_tmu_enable(tmu, false);
>>> +	imx91_tmu_start(tmu, false);
>>> +
>>> +	ret = imx91_init_from_nvmem_cells(tmu);
>>> +	if (ret) {
>>> +		writel_relaxed(IMX91_TMU_DEFAULT_TRIM1_CONFIG, tmu->base + IMX91_TMU_TRIM1);
>>> +		writel_relaxed(IMX91_TMU_DEFAULT_TRIM2_CONFIG, tmu->base + IMX91_TMU_TRIM2);
>>> +	}
>>> +
>>> +	/* The typical conv clk is 4MHz, the output freq is 'rate / (div + 1)' */
>>> +	rate = clk_get_rate(tmu->clk);
>>> +	div = (rate / 4000000) - 1;
>>
>> Use literals please (eg. 4 * HZ_PER_MHZ)
>>
>>> +	if (div > IMX91_TMU_DIV_MAX)
>>> +		return dev_err_probe(dev, -EINVAL, "clock divider exceed hardware limitation");
>>> +
>>> +	/* Set divider value and enable divider */
>>> +	writel_relaxed(IMX91_TMU_DIV_EN | FIELD_PREP(IMX91_TMU_DIV_MASK, div),
>>> +		       tmu->base + IMX91_TMU_REF_DIV);
>>> +
>>> +	/* Set max power up delay: 'Tpud(ms) = 0xFF * 1000 / 4000000' */
>>> +	writel_relaxed(FIELD_PREP(IMX91_TMU_PUDL_MASK, 100U), tmu->base + IMX91_TMU_PUD_ST_CTRL);
>>> +
>>> +	/*
>>> +	 * Set resolution mode
>>> +	 * 00b - Conversion time = 0.59325 ms
>>> +	 * 01b - Conversion time = 1.10525 ms
>>> +	 * 10b - Conversion time = 2.12925 ms
>>> +	 * 11b - Conversion time = 4.17725 ms
>>> +	 */
>>> +	writel_relaxed(FIELD_PREP(IMX91_TMU_CTRL1_RES_MASK, 0x3), tmu->base + IMX91_TMU_CTRL1_CLR);
>>> +	writel_relaxed(FIELD_PREP(IMX91_TMU_CTRL1_RES_MASK, 0x1), tmu->base + IMX91_TMU_CTRL1_SET);
>>> +
>>> +	writel_relaxed(IMX91_TMU_CTRL1_MEAS_MODE_MASK, tmu->base + IMX91_TMU_CTRL1_CLR);
>>> +	writel_relaxed(FIELD_PREP(IMX91_TMU_CTRL1_MEAS_MODE_MASK, IMX91_TMU_CTRL1_MEAS_MODE_SINGLE),
>>> +		       tmu->base + IMX91_TMU_CTRL1_SET);
>>> +
>>> +	pm_runtime_set_active(dev);
>>> +	devm_pm_runtime_enable(dev);
>>> +	pm_runtime_put(dev);
>>> +
>>> +	tmu->tzd = devm_thermal_of_zone_register(dev, 0, tmu, &tmu_tz_ops);
>>> +	if (IS_ERR(tmu->tzd))
>>> +		return dev_err_probe(dev, PTR_ERR(tmu->tzd),
>>> +				     "failed to register thermal zone sensor\n");
>>> +
>>> +	ret = devm_add_action(dev, imx91_tmu_action_remove, tmu);
>>
>> Should it be moved before devm_thermal_of_zone_register(), so if the thermal
>> zone creation fails, it will stop the sensor which was previously started ?
> 
> pm_runtime_put already disable sensor before call
> devm_thermal_of_zone_register().
> 
> imx91_tmu_get_temp() will call runtime_get when devm_thermal_of_zone_register()
> call to it.
> 
> Frank
> 
>>
>>> +	if (ret)
>>> +		return dev_err_probe(dev, ret, "Failure to add action imx91_tmu_action_remove()\n");
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int imx91_tmu_runtime_suspend(struct device *dev)
>>> +{
>>> +	struct imx91_tmu *tmu = dev_get_drvdata(dev);
>>> +
>>> +	/* disable tmu */
>>> +	imx91_tmu_enable(tmu, false);
>>> +
>>> +	clk_disable_unprepare(tmu->clk);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static int imx91_tmu_runtime_resume(struct device *dev)
>>> +{
>>> +	struct imx91_tmu *tmu = dev_get_drvdata(dev);
>>> +	int ret;
>>> +
>>> +	ret = clk_prepare_enable(tmu->clk);
>>> +	if (ret)
>>> +		return ret;
>>> +
>>> +	imx91_tmu_enable(tmu, true);
>>> +
>>> +	return 0;
>>> +}
>>> +
>>> +static DEFINE_RUNTIME_DEV_PM_OPS(imx91_tmu_pm_ops, imx91_tmu_runtime_suspend,
>>> +				 imx91_tmu_runtime_resume, NULL);
>>> +
>>> +static const struct of_device_id imx91_tmu_table[] = {
>>> +	{ .compatible = "fsl,imx91-tmu", },
>>> +	{ },
>>> +};
>>> +MODULE_DEVICE_TABLE(of, imx91_tmu_table);
>>> +
>>> +static struct platform_driver imx91_tmu = {
>>> +	.driver = {
>>> +		.name	= "imx91_thermal",
>>> +		.pm	= pm_ptr(&imx91_tmu_pm_ops),
>>> +		.of_match_table = imx91_tmu_table,
>>> +	},
>>> +	.probe = imx91_tmu_probe,
>>> +};
>>> +module_platform_driver(imx91_tmu);
>>> +
>>> +MODULE_AUTHOR("Peng Fan <peng.fan@....com>");
>>> +MODULE_DESCRIPTION("i.MX91 Thermal Monitor Unit driver");
>>> +MODULE_LICENSE("GPL");
>>>
>>
>>
>> --
>> <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


-- 
<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

Powered by blists - more mailing lists