Date:	Tue, 15 Oct 2013 14:57:43 -0700
From:	"Wendy Ng" <wendy.ng@...adcom.com>
To:	"Eduardo Valentin" <eduardo.valentin@...com>
cc:	"Rob Herring" <rob.herring@...xeda.com>,
	"Stephen Warren" <swarren@...dotorg.org>,
	devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
	linux-pm@...r.kernel.org, "Christian Daudt" <bcm@...thebug.org>,
	"Markus Mayer" <mmayer@...adcom.com>
Subject: Re: [PATCH v2 1/3] thermal: bcm281xx: Add thermal driver

On 10/13/2013 4:14 PM, Eduardo Valentin wrote:
> On 27-09-2013 18:37, Wendy Ng wrote:
>> This adds the support for reading out temperature from Broadcom bcm281xx
>> SoCs.
>>
>> Signed-off-by: Wendy Ng <wendy.ng@...adcom.com>
>> Reviewed-by: Markus Mayer <mmayer@...adcom.com>
>> Reviewed-by: Christian Daudt <csd@...adcom.com>
>> ---
>>   .../bindings/thermal/bcm-kona-thermal.txt          |   18 +++
>>   drivers/thermal/Kconfig                            |   10 ++
>>   drivers/thermal/Makefile                           |    1 +
>>   drivers/thermal/bcm_thermal.c                      |  170 ++++++++++++++++++++
>>   4 files changed, 199 insertions(+)
>>   create mode 100644 Documentation/devicetree/bindings/thermal/bcm-kona-thermal.txt
>>   create mode 100644 drivers/thermal/bcm_thermal.c
>>
>> diff --git a/Documentation/devicetree/bindings/thermal/bcm-kona-thermal.txt b/Documentation/devicetree/bindings/thermal/bcm-kona-thermal.txt
>> new file mode 100644
>> index 0000000..acca99e
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/thermal/bcm-kona-thermal.txt
>> @@ -0,0 +1,18 @@
>> +* Broadcom Kona Thermal Management Unit
>> +
>> +This version is for the BCM281xx family of SoCs.
>
> Might be worth describing the hardware a bit and also appending a link
> to its documentation for further details.
>

I will improve this part in the next version of this patch series.

>> +
>> +Required properties:
>> +- compatible : "brcm,bcm11351-thermal", "brcm,kona-thermal"
>> +- reg : Address range of the thermal register
>> +- thermal-name: this entry must be specified and it will be passed into
>> +thermal_zone_device_register().  This name will also be reported under Hwmon
>> +sysfs 'name' attribute.
>
> I guess by now you already remember my suggestion on this dt property.
> If you can drop it then we could proceed with this driver. But with it
> we are adding something specific to it that may be soon deprecated.
>

Yes, 'thermal-name' will be dropped in the next version of this patch 
series.

>> +
>> +Example:
>> +	thermal@...08000 {
>> +		compatible = "brcm,bcm11351-thermal", "brcm,kona-thermal";
>> +		reg = <0x34008000 0x0024>;
>> +		thermal-name = "bcm_kona_therm";
>> +		status = "disabled";
>> +	};
>> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
>> index dbfc390..6a5341c 100644
>> --- a/drivers/thermal/Kconfig
>> +++ b/drivers/thermal/Kconfig
>> @@ -134,6 +134,16 @@ config KIRKWOOD_THERMAL
>>   	  Support for the Kirkwood thermal sensor driver into the Linux thermal
>>   	  framework. Only kirkwood 88F6282 and 88F6283 have this sensor.
>>
>> +config BCM_THERMAL
>> +	tristate "Temperature sensor on Broadcom BCM281xx family of SoCs"
>> +	depends on ARCH_BCM_MOBILE
>
> Is the sensor IP used only on this Arch? No plans for re-usability?
>
>> +	default y
>> +	help
>> +	  If you say yes here you get support for TMU (Thermal Management
>> +	  Unit) on Broadcom BCM281xx family of SoCs. This provides thermal
>> +	  monitoring of CPU clusters, graphics, and SoC glue, but does not
>> +	  include monitoring of charger temperature.
>> +
>>   config DOVE_THERMAL
>>   	tristate "Temperature sensor on Marvell Dove SoCs"
>>   	depends on ARCH_DOVE
>> diff --git a/drivers/thermal/Makefile b/drivers/thermal/Makefile
>> index 584b363..3ea8c1c 100644
>> --- a/drivers/thermal/Makefile
>> +++ b/drivers/thermal/Makefile
>> @@ -21,6 +21,7 @@ obj-$(CONFIG_SPEAR_THERMAL)	+= spear_thermal.o
>>   obj-$(CONFIG_RCAR_THERMAL)	+= rcar_thermal.o
>>   obj-$(CONFIG_KIRKWOOD_THERMAL)  += kirkwood_thermal.o
>>   obj-y				+= samsung/
>> +obj-$(CONFIG_BCM_THERMAL)       += bcm_thermal.o
>>   obj-$(CONFIG_DOVE_THERMAL)  	+= dove_thermal.o
>>   obj-$(CONFIG_DB8500_THERMAL)	+= db8500_thermal.o
>>   obj-$(CONFIG_ARMADA_THERMAL)	+= armada_thermal.o
>> diff --git a/drivers/thermal/bcm_thermal.c b/drivers/thermal/bcm_thermal.c
>> new file mode 100644
>> index 0000000..131d3c4
>> --- /dev/null
>> +++ b/drivers/thermal/bcm_thermal.c
>> @@ -0,0 +1,170 @@
>> +/*
>> + * Copyright 2013 Broadcom Corporation.
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License, version 2,
>> + * as published by the Free Software Foundation (the "GPL").
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> + * GNU General Public License for more details.
>> + *
>> + * A copy of the GPL is available at http://www.broadcom.com/licenses/GPLv2.php,
>> + * or by writing to the Free Software Foundation, Inc.,
>> + * 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
>> + */
>> +
>> +/**
>> +*  Broadcom Thermal Management Unit - bcm_tmu
>> +*/
>> +#include <linux/module.h>
>> +#include <linux/err.h>
>> +#include <linux/kernel.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/io.h>
>> +#include <linux/thermal.h>
>> +#include <linux/of.h>
>> +
>> +/* From TMON Register Database */
>> +#define TMON_TEMP_VAL_OFFSET			0x0000001c
>> +#define TMON_TEMP_VAL_TEMP_VAL_SHIFT		0
>> +#define TMON_TEMP_VAL_TEMP_VAL_MASK		0x000003ff
>> +
>> +/* Broadcom Thermal Zone Device Structure */
>> +struct bcm_thermal_zone_priv {
>> +	char name[THERMAL_NAME_LENGTH];
>> +	void __iomem *base;
>> +};
>> +
>> +/* Temperature conversion function for TMON block */
>> +static long raw_to_mcelsius(u32 raw)
>> +{
>> +	/*
>> +	 * According to Broadcom internal Analog Module Specification
>> +	 * the formula for converting TMON block output to temperature in
>> +	 * degree Celsius is:
>> +	 *	T = 428 - (0.561 * raw)
>> +	 * Note: the valid operating range for the TMON block is -40C to 125C
>> +	 */
>
> Does 125C mean silicon is not reliable? Shall we define a critical trip
> at this level?
>

The critical trip temperature will be specified through the DTS file 
with your new thermal binding approach.

Btw, I would like to mention here about the use case of taking the 
trip-point temperature that is specified in the DTS files to configure 
hardware specific registers such that an interrupt can be fired 
according to the trip-point temperature.

Eduardo, you confirmed with me that this is a limitation with your 
thermal DT binding approach in another email thread.  Could you please 
let me know if you have a plan to address this use-case in your patch 
series:
https://lkml.org/lkml/2013/9/26/787


>> +	return 428000 - (561 * (long)raw);
>> +}
>> +
>> +/* Get temperature callback function for thermal zone */
>> +static int bcm_get_temp(struct thermal_zone_device *thermal,
>> +			unsigned long *temp)
>> +{
>> +	u32 raw;
>> +	long mcelsius;
>> +	struct bcm_thermal_zone_priv *priv = thermal->devdata;
>> +
>> +	if (!priv) {
>> +		pr_err("%s: thermal zone number %d devdata not initialized.\n",
>> +			__func__, thermal->id);
>> +		return -EINVAL;
>> +	}
>> +
>> +	raw = (readl(priv->base + TMON_TEMP_VAL_OFFSET)
>> +		& TMON_TEMP_VAL_TEMP_VAL_MASK) >> TMON_TEMP_VAL_TEMP_VAL_SHIFT;
>> +
>> +	pr_debug("%s: thermal zone number %d raw temp 0x%x\n", __func__,
>> +		thermal->id, raw);
>> +
>> +	mcelsius = raw_to_mcelsius(raw);
>> +
>> +	/*
>> +	 * Since 'mcelsius' might be negative, we need to limit it to smallest
>> +	 * unsigned value before returning it to thermal framework.
>> +	 */
>
> Do you have practical usage of negative values?

No practical usage other than taking care of the possible range that the 
thermal sensor supports.  This code is to prevent a negative value from 
reporting back to the thermal framework which is treating temperature as 
an unsigned value.

>
>> +	if (mcelsius < 0)
>> +		*temp = 0;
>> +	else
>> +		*temp = mcelsius;
>> +
>> +	pr_debug("%s: thermal zone number %d final temp %d\n", __func__,
>> +		thermal->id, (int) *temp);
>> +
>> +	return 0;
>> +}
>> +
>> +/* Operation callback functions for thermal zone */
>> +static struct thermal_zone_device_ops bcm_dev_ops = {
>> +	.get_temp = bcm_get_temp,
>> +};
>> +
>> +static const struct of_device_id bcm_tmu_match_table[] = {
>> +	{ .compatible = "brcm,kona-thermal" },
>> +	{},
>> +};
>> +MODULE_DEVICE_TABLE(of, bcm_tmu_match_table);
>> +
>> +static int bcm_tmu_probe(struct platform_device *pdev)
>> +{
>> +	struct thermal_zone_device *thermal = NULL;
>> +	struct bcm_thermal_zone_priv *priv;
>> +	struct resource *res;
>> +	const char *str;
>> +
>> +	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
>> +	if (!priv) {
>> +		dev_err(&pdev->dev, "Failed to malloc priv.\n");
>> +		return -ENOMEM;
>> +	}
>> +
>> +	/* Obtain the tmu name from device tree file */
>> +	if (of_property_read_string(pdev->dev.of_node, "thermal-name",
>> +			&str) == 0) {
>
> it is common practice to use a variable to hold the error values.
>

Thanks for pointing this out.  This line of code will be gone in the 
next version of this patch series.  We won't need to get the 
'thermal-name' from the DT binding.

>> +		strlcpy(priv->name, str, sizeof(priv->name));
>> +	} else {
>> +		dev_err(&pdev->dev, "Failed to get thermal-name from DT.\n");
>> +		return -EINVAL;
>> +	}
>> +
>> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	priv->base = devm_ioremap_resource(&pdev->dev, res);
>> +	if (IS_ERR(priv->base))
>> +		return PTR_ERR(priv->base);
>> +
>> +	thermal = thermal_zone_device_register(priv->name, 0, 0, priv,
>> +						&bcm_dev_ops, NULL, 0, 0);
>
>
> No trips, no cooling.. constraint and logic is somewhere in userland?
>

The critical trip temperature will be specified in the next version of 
this patch series.  The other passive trip points that need to bind with 
cooling devices will be added in another patch series in the near future.

>> +	if (IS_ERR(thermal)) {
>> +		dev_err(&pdev->dev,
>> +			"Failed to register Broadcom thermal zone device.\n");
>> +		return PTR_ERR(thermal);
>> +	}
>> +
>> +	platform_set_drvdata(pdev, thermal);
>> +
>> +	dev_info(&pdev->dev, "Broadcom Thermal Monitor Initialized.\n");
>
> Are you sure you want to be this wordy?

Do you see any drawback or issue with this?  I would like to understand 
your concern before changing it so I will get it right the next time. 
Thanks.

>> +
>> +	return 0;
>> +}
>> +
>> +static int bcm_tmu_remove(struct platform_device *pdev)
>> +{
>> +	struct thermal_zone_device *broadcom_thermal =
>> +		platform_get_drvdata(pdev);
>> +
>> +	thermal_zone_device_unregister(broadcom_thermal);
>> +
>> +	dev_info(&pdev->dev, "Broadcom Thermal Monitor Uninitialized.\n");
>
> ditto
>
>> +
>> +	return 0;
>> +}
>> +
>> +static struct platform_driver bcm_tmu_driver = {
>> +	.driver = {
>> +		.name = "bcm-thermal",
>> +		.owner = THIS_MODULE,
>> +		.of_match_table = bcm_tmu_match_table,
>> +	},
>> +	.probe = bcm_tmu_probe,
>> +	.remove = bcm_tmu_remove,
>> +};
>> +
>> +module_platform_driver(bcm_tmu_driver);
>> +
>> +MODULE_DESCRIPTION("Broadcom Thermal Driver");
>> +MODULE_AUTHOR("Broadcom");
>> +MODULE_LICENSE("GPL v2");
>> +MODULE_ALIAS("platform:bcm-thermal");
>>
>
>

Hello Eduardo,

Thanks for providing your feedback.  I would like to let you know that I 
have rebased my code with your latest patch series as shown in:

https://git.kernel.org/cgit/linux/kernel/git/evalenti/linux.git/log/?h=thermal_work/thermal_core/dt_parser_rfc_v4

My latest code changes are under *internal* review right now.   Once the 
review is done, then I will send it out to the LKML and you can have 
another look.  I understand your patch series is still under review, so 
I will try to keep up with it.  But once your patch series has been 
accepted, do you think you can also pick-up my patch series such that 
they get into the mainline together?  I think my patch series is another 
example to illustrate how to use your new DT binding for the thermal 
framework.

-- 
Best regards,
-Wendy

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists