Re: [PATCH 1/3] thermal: bcm281xx: Add thermal driver

From: Wendy Ng
Date: Mon Oct 07 2013 - 19:11:36 EST



On 10/7/2013 3:32 PM, Eduardo Valentin wrote:
On 07-10-2013 13:52, Wendy Ng wrote:
On 10/7/2013 9:12 AM, Eduardo Valentin wrote:
On 07-10-2013 10:52, Eduardo Valentin wrote:
On 04-10-2013 18:17, Wendy Ng wrote:
Hello Eduardo,

Hello Wendy,

I have rebased my work on the top of yours and tried out the new way of
registering the thermal zone. The new method is certainly making
things easier for me to create a new thermal zone device driver. But I
have some questions and I hope to hear back from you before I upload my
new patch set that is based on your changes.
Great that you gave it a shot. Thanks for that.

1. I found that the hwmon sysfs interface is no longer created when I
call thermal_zone_of_sensor_register() to register my thermal
sensor. I
believe it is because you purposely hard-coded 'no_hwmon' to true.
Would it be possible to let the users to control this option as part of
the sensor register function?
Hmmm.. I also thought of allowing the thermal to hwmon interface
configurable. But the interface gets created in different path from the
sensor registration, which makes things a bit harder. The zones are
probed while parsing the DT file. While the sensors get linked,
thermal_zone_of_sensor_register(), typically when sensors are probed.

Do you have a real use case for using the thermal to hwmon interface?
Durgadoss (original author) and Rui were considering deprecating and
finally removing it. If yes, it is time to rise your voice here.

2. The of_thermal_get_temp() output parameter 'temp' is an unsigned
value in order to conforms with the thermal framework 'get_temp'
function prototype. But the new sensor interface 'get_temp' callback
function (defined in __thermal_zone struct) expects 'temp' to be
signed. So technically, a negative value can be returned from the
underlying sensor function. Should some checks be added to
of_thermal_get_temp() to ensure that a negative value (i.e. a very
large
unsigned value) won't be passed back to the thermal framework? Or
should the 'get_temp' function prototype in __thermal_zone struct be
changed with 'unsigned' temperature value and it will be the
responsibility of the underlying function to do the check?

I could for sure, add a warning inside of_thermal_get_temp(). But in
general, I believe the thermal framework has to cover for signed
temperature values, even if typical usage is on positive temperature
range. Mainly for correct temperature representation.


3. Your documentation (thermal.txt) seems to suggest that the
'cooling-attachments' are one of the required properties for
'thermal-zones' node in the device tree binding files. However, I
don't
Hmm... That was not the intention.

have any cooling device ready in my current patch series yet and
therefore I tried to leave out the 'cooling-attachments' node from the
'thermal-zones' node. It seems to be working just find and I can
register the sensor and read out the temperature once the target
platform has been booted up. Is this your expectation that people can
leave the 'cooling-attachments' node out? I want to make sure I can
still submit a new patch set for this series with just the
functionality
of reading out the temperature from the thermal sensor.
So, let me try to clarify a couple of things. First thing is that the
cooling-attachments has been renamed to cooling-maps, just a new naming
convention. Also, there are thermal zones that do not have
cooling-maps, they just describe thermal shutdown by software, thermal
zones with only CRITICAL trip points, for instance. For this reason,
thermal zones must be probed properly without cooling-maps. I need to
double check the documentation in this case, I believe.
Just to clarify a bit, this does not mean we should allow having thermal
zones only for monitoring. If you are generating a thermal zone, it
means you want to cover thermal limits. The example I gave was a thermal
zone that does not have cooling-maps, but it does have a thermal
constraint, a trip point which is CRITICAL, meaning, the silicon state
is not reliable anymore.
Hi Eduardo,

Thanks for your clarification. Yes, there are thermal limits that I
need to specify through the trip points. The issue I am facing is that
I don't have the cooling device that needs to be associated with the
trip point in this patch series yet. I was hoping to get the thermal
sensor driver code into the mainline first. The addition for the
cooling device and the "cooling-maps" node will be added in another
patch series.
Wendy,

Let me try to understand it properly. What cooling device is missing? Is
it only the kernel code or the its DT bindings too? Does it prevent you
to define the bindings? I believe it should be fine to have the bindings
without the kernel code (as long as the bindings are correct).

I have also tried this patch series on top of a systems with missing
cpufreq support. That means, I had all the thermal bindings, but the
cooling device was not loaded in the system. Unfortunately, the zones
would only have the critical trip points as runtime effect. But nothing
really broke dramatically.

Hi Eduardo,

Sorry, I probably have confused you. You are correct -- things seem to work just fine for me when I have all the thermal binding specified in the device tree file (*.dts/dtsi) despite the missing cpufreq support in my kernel code.

I even tried to remove the 'cooling-attachments' (a.k.a cooling-maps) from the device tree files and the thermal sensor can still be registered to the thermal zone. So my original question about the documentation is just to get a clarification on what is the expected behavior when I left out what is specified as the *required* properties in the thermal-zones node in the device tree file.


Since this patch series is now gated by your work, I am wondering if you
are targeting to get your patch series
(http://lkml.org/lkml/2013/9/15/122) into 3.13? I would like to get an
idea of the timeline such that I can plan out my work as well.
Regarding timing, I cannot promise you that my patch series is making
3.13, because I cannot proceed / queue this work without a proper review
and acks from device tree maintainers. I would really appreciate if we
could make it for 3.13, but that is not only on my hands. As Mark R.
said, it is a bit complex binding and it is better to have more than one
device tree maintainer review/ack.

Regarding your driver, my major concern on it is just the fact that you
are using one driver specific device tree property, which will get soon
deprecated. That is why I have kindly requested you to wait on my work,
this way we avoid compatibility issues between DT and kernel version.
Thanks!
Thanks,
-Wendy

On 9/23/2013 4:46 PM, Eduardo Valentin wrote:
On 23-09-2013 18:43, Wendy Ng wrote:
On 9/23/2013 11:19 AM, Eduardo Valentin wrote:
On 23-09-2013 13:51, Wendy Ng wrote:
This adds the support for reading out temperature from Broadcom
bcm281xx
SoCs.

Signed-off-by: Wendy Ng <wendy.ng@xxxxxxxxxxxx>
Reviewed-by: Markus Mayer <mmayer@xxxxxxxxxxxx>
Reviewed-by: Christian Daudt <csd@xxxxxxxxxxxx>
---
.../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.
+
+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.
+
+Example:
+ thermal@34008000 {
+ 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..7f823f0 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
+ 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
+ */
+ 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.
+ */
+ 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",
Hello Wendy,

I would prefer we wait until the work for thermal data [1] gets
accepted
before accepting this driver, specially because you are adding
device
specific DT entry to help in your registration with the thermal
framework. With the mentioned work you wont need it at all.

All best,

[1] - http://lkml.org/lkml/2013/9/15/122
Hello Eduardo,

Since your changes are quite extensive, would you recommend me to get
your 16 patches into my local workspace and try to integrate the code
from my patches on the top of your patches at this time? Or should I
wait until your patches get accepted.
Hello again Wendy,

The only part I am concerned is the device tree entry you are creating
for your device/driver. That could be obsolete in near future with the
work I am doing currently.

The work is, as you mentioned, quite extensive, but should make things
easier for device driver writer. If you want, of course, it would be
great if you could rebase your work on top of mine and try it out on
your side and let me know the outcome.

Comments on the proposed binding and documentation is also welcome.
Let
me know if there is something that it is not clear.

Thanks,
-Wendy
+ &str) == 0) {
+ 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);
+ 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");
+
+ 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");
+
+ 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");



--
Best regards,
-Wendy


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