Skip to contentSkip to conversation

thermal: bcm2835: add thermal driver for bcm2835 soc

PATCH Superseded kernel at martin.sperl.org 8 participants 23 messages Jan 7, 2017 – Feb 9, 2017 4 messages not in the archive

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

---
 drivers/thermal/Kconfig           |   8 +
 drivers/thermal/Makefile          |   1 +
 drivers/thermal/bcm2835_thermal.c | 354 ++++++++++++++++++++++++++++++++++++++
 3 files changed, 363 insertions(+)
 create mode 100644 drivers/thermal/bcm2835_thermal.c

ChangeLog:
V1 -> V2: added specific settings depending on compatiblity
added trip point based on register
setting up ctrl-register if HW is not enabled by firmware
as per recommendation of Eric (untested)
check that clock frequency is in range
(1.9 - 5MHz - as per comment in clk-bcm2835.c)
V2 -> V4: moved back to thermal (not using bcm sub-directory)
set polling interval to 1second (was 0ms, so interrupt driven)
V5 -> V6: added correct depends in KConfig
removed defined default for RESET_DELAY
removed obvious comments
clarify HW setup comments if not set up by FW already move clk_prepare_enable to an earlier stage and add error handling clarify warning when TS-clock runs out of recommended range clk_disable_unprepare added in bcm2835_thermal_remove added comment on recommended temperature ranges for SOC
V6 -> V7: removed depends on ARCH_BCM2836 || ARCH_BCM2837 in Kconfig V7 -> V8: rebased
V8 -> V9: moved to use the thermal framework offset and slope in
thermal_zone_parameters as per request

A few additional notes on this latest patch:
* all the other portions of the patch (dt, bindings, defconf) have already been merged into 4.10.0-rc2 - this only leaves the driver itself which is now just a single patch to get applied.
* the driver has been moved to use thermal_zone_parameters->offset and slope making use of the thermal_zone_get_offset and thermal_zone_get_slope methods where relevant
Note that we cannot use those during initial setup of the HW - this code section is typically not used anyway, as the FW sets up the device already. This so modified patch actually increases code size by 14 lines:

    drivers/thermal/bcm2835_thermal.c | 42 ++++++++++++++++++++++++++-------------
    1 file changed, 28 insertions(+), 14 deletions(-)

* /sys/class/thermal/*/ contains now slope and offset files
and those expose the correct values
* for some reason on module load the kernel emits now the message:
(NULL device *): hwmon_device_register() is deprecated. Please convert the driver to use hwmon_device_register_with_info().
But this seems framework specific and needs to get addressed there.

drivers/thermal/Kconfig +8 -0
index c2c056c..18f2de6 100644
@@ -436,4 +436,12 @@ depends on (ARCH_QCOM && OF) || COMPILE_TEST
 source "drivers/thermal/qcom/Kconfig"
 endmenu
 
+config BCM2835_THERMAL
+	tristate "Thermal sensors on bcm2835 SoC"
+	depends on ARCH_BCM2835 || COMPILE_TEST
+	depends on HAS_IOMEM
+	depends on OF
+	help
+	  Support for thermal sensors on Broadcom bcm2835 SoCs.
+
 endif
drivers/thermal/Makefile +1 -0
index 6a3d7b5..677c6d9 100644
@@ -56,3 +56,4 @@ obj-$(CONFIG_TEGRA_SOCTHERM)	+= tegra/
 obj-$(CONFIG_HISI_THERMAL)     += hisi_thermal.o
 obj-$(CONFIG_MTK_THERMAL)	+= mtk_thermal.o
 obj-$(CONFIG_GENERIC_ADC_THERMAL)	+= thermal-generic-adc.o
+obj-$(CONFIG_BCM2835_THERMAL)	+= bcm2835_thermal.o
drivers/thermal/bcm2835_thermal.c new file +354 -0

354 lines not on this page · raw message

--
2.1.4

#1 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Hello Martin,

Quoted 11 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

<cut>

Quoted 18 lines
drivers/thermal/bcm2835_thermal.c new file +18 -0
+
+static const struct of_device_id bcm2835_thermal_of_match_table[] = {
+	{
+		.compatible = "brcm,bcm2835-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2836-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},

drivers/thermal/bcm2835_thermal.c line 338

Quoted 10 lines
drivers/thermal/bcm2835_thermal.c new file +10 -0
+	{
+		.compatible = "brcm,bcm2837-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			/* the bcm2837 needs adjustment of +5C */
+			.offset = 407000 + 5000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{},

Just for the same of clarification, is there anything preventing this driver of using of-thermal API? the above data (slope, offset, and trip_temps) would be in DT the place where they are supposed to be, instead of code.

Quoted 18 lines
drivers/thermal/bcm2835_thermal.c new file +16 -1
+};
+MODULE_DEVICE_TABLE(of, bcm2835_thermal_of_match_table);
+
+static struct platform_driver bcm2835_thermal_driver = {
+	.probe = bcm2835_thermal_probe,
+	.remove = bcm2835_thermal_remove,
+	.driver = {
+		.name = "bcm2835_thermal",
+		.of_match_table = bcm2835_thermal_of_match_table,
+	},
+};
+module_platform_driver(bcm2835_thermal_driver);
+
+MODULE_AUTHOR("Martin Sperl");
+MODULE_DESCRIPTION("Thermal driver for bcm2835 chip");
+MODULE_LICENSE("GPL");
--

2.1.4

#2 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Quoted 10 lines · Eduardo Valentin

Hello Martin,

Quoted 7 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

Quoted 6 lines
Quoted 4 lines

<cut>

Also, I am getting this warn from sparse:
drivers/thermal/bcm2835_thermal.c:110:16: warning: cast truncates bits from constant value (3ffffffffff00 becomes ffffff00) drivers/thermal/bcm2835_thermal.c:134:16: warning: cast truncates bits from constant value (3ffffffffff becomes ffffffff)

Have you seen this?

#3 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Quoted 22 lines · Eduardo Valentin
Quoted 10 lines · Eduardo Valentin

Hello Martin,

Quoted 7 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

Quoted 6 lines
Quoted 4 lines

<cut>

Quoted 7 lines

Also, I am getting this warn from sparse:
drivers/thermal/bcm2835_thermal.c:110:16: warning: cast truncates bits from constant value (3ffffffffff00 becomes ffffff00) drivers/thermal/bcm2835_thermal.c:134:16: warning: cast truncates bits from constant value (3ffffffffff becomes ffffffff)

Have you seen this?

No, I have not checked sparse.

These values are defined via GENMASK on line 47 and 57 respectively and should actually compute to the following values:
for line 110 (line 47 has the define):

    GENMASK(BCM2835_TS_TSENSCTL_THOLD_BITS +     \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT - 1, \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT)
    = GENMASK(10 + 8 - 1, 8) 
    = GENMASK(17, 8)
    = (((~0UL) << (8)) & (~0UL >> (32 - 1 - (10 + 8 - 1))))
    = 0x3ff00
  for line 134 (line 57 has the define):
    GENMASK(BCM2835_TS_TSENSCTL_THOLD_BITS +     \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT - 1, \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT)
    = GENMASK(10 + 0 - 1, 0)
    = GENMASK(9, 0)
    = (((~0UL) << (0)) & (~0UL >> (32 - 1 - (10 + 0 - 1))))
    = 0x003ff

Note that the preprocessor expansions have been verified by looking at the preprocessed driver source (drivers/thermal/bcm2835_thermal.i)

I wonder why sparse is computing these GENMASK values as:
0x3ffffffffff00 and 0x3ffffffffff

Martin

#4 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Hello Martin,

Quoted 51 lines · kernel at martin.sperl.org
Quoted 22 lines · Eduardo Valentin
Quoted 10 lines · Eduardo Valentin

Hello Martin,

Quoted 7 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

Quoted 6 lines
Quoted 4 lines

<cut>

Quoted 7 lines

Also, I am getting this warn from sparse:
drivers/thermal/bcm2835_thermal.c:110:16: warning: cast truncates bits from constant value (3ffffffffff00 becomes ffffff00) drivers/thermal/bcm2835_thermal.c:134:16: warning: cast truncates bits from constant value (3ffffffffff becomes ffffffff)

Have you seen this?

No, I have not checked sparse.

These values are defined via GENMASK on line 47 and 57 respectively and should actually compute to the following values:
for line 110 (line 47 has the define):

    GENMASK(BCM2835_TS_TSENSCTL_THOLD_BITS +     \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT - 1, \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT)
    = GENMASK(10 + 8 - 1, 8) 
    = GENMASK(17, 8)
    = (((~0UL) << (8)) & (~0UL >> (32 - 1 - (10 + 8 - 1))))
    = 0x3ff00
  for line 134 (line 57 has the define):
    GENMASK(BCM2835_TS_TSENSCTL_THOLD_BITS +     \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT - 1, \
            BCM2835_TS_TSENSCTL_THOLD_SHIFT)
    = GENMASK(10 + 0 - 1, 0)
    = GENMASK(9, 0)
    = (((~0UL) << (0)) & (~0UL >> (32 - 1 - (10 + 0 - 1))))
    = 0x003ff
Quoted 3 lines

Note that the preprocessor expansions have been verified by looking at the preprocessed driver source (drivers/thermal/bcm2835_thermal.i)

OK then.

Quoted 3 lines

I wonder why sparse is computing these GENMASK values as:
0x3ffffffffff00 and 0x3ffffffffff

In the case you can confirm that the values are correct, I believe this could be a false positive report on sparse, in this case.

Quoted 2 lines

Martin

#5 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Quoted 61 lines · Eduardo Valentin

Hello Martin,

Quoted 51 lines · On Fri, Jan 20, 2017 at 09:43:02AM +0100, [email protected] wrote:
Quoted 22 lines · Eduardo Valentin
Quoted 10 lines · Eduardo Valentin

Hello Martin,

Quoted 7 lines · On Sat, Jan 07, 2017 at 04:55:45PM +0000, [email protected] wrote:

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

Quoted 6 lines
Quoted 4 lines

<cut>

Quoted 7 lines

Also, I am getting this warn from sparse:
drivers/thermal/bcm2835_thermal.c:110:16: warning: cast truncates bits from constant value (3ffffffffff00 becomes ffffff00) drivers/thermal/bcm2835_thermal.c:134:16: warning: cast truncates bits from constant value (3ffffffffff becomes ffffffff)

Have you seen this?

No, I have not checked sparse.

These values are defined via GENMASK on line 47 and 57 respectively and should actually compute to the following values:
for line 110 (line 47 has the define):

   GENMASK(BCM2835_TS_TSENSCTL_THOLD_BITS +     \
           BCM2835_TS_TSENSCTL_THOLD_SHIFT - 1, \
           BCM2835_TS_TSENSCTL_THOLD_SHIFT)
   = GENMASK(10 + 8 - 1, 8) 
   = GENMASK(17, 8)
   = (((~0UL) << (8)) & (~0UL >> (32 - 1 - (10 + 8 - 1))))
   = 0x3ff00
 for line 134 (line 57 has the define):
   GENMASK(BCM2835_TS_TSENSCTL_THOLD_BITS +     \
           BCM2835_TS_TSENSCTL_THOLD_SHIFT - 1, \
           BCM2835_TS_TSENSCTL_THOLD_SHIFT)
   = GENMASK(10 + 0 - 1, 0)
   = GENMASK(9, 0)
   = (((~0UL) << (0)) & (~0UL >> (32 - 1 - (10 + 0 - 1))))
   = 0x003ff
Quoted 3 lines

Note that the preprocessor expansions have been verified by looking at the preprocessed driver source (drivers/thermal/bcm2835_thermal.i)

Quoted 8 lines

OK then.

Quoted 3 lines

I wonder why sparse is computing these GENMASK values as:
0x3ffffffffff00 and 0x3ffffffffff

In the case you can confirm that the values are correct, I believe this could be a false positive report on sparse, in this case.

The correct values are the ones in my email: 0x3ff00 and 0x003ff

The ones reported by sparse (0x3ffffffffff00 and 0x3ffffffffff) are not calculated correctly - to me it looks as if it is possibly some sort of 64 bit issue of sparse in conjunction with the generic macro GENMASK.

#6 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

In reply to Eduardo Valentin
Quoted 45 lines · Eduardo Valentin

Hello Martin,

Quoted 11 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

<cut>

Quoted 25 lines
drivers/thermal/bcm2835_thermal.c new file +25 -0
+
+static const struct of_device_id bcm2835_thermal_of_match_table[] = {
+	{
+		.compatible = "brcm,bcm2835-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2836-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2837-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			/* the bcm2837 needs adjustment of +5C */
+			.offset = 407000 + 5000,
+			.slope = -538,
+			.trip_temp = 80000
Quoted 9 lines

drivers/thermal/bcm2835_thermal.c line 338

Quoted 3 lines
drivers/thermal/bcm2835_thermal.c new file +3 -0
+		}
+	},
+	{},

Just for the same of clarification, is there anything preventing this driver of using of-thermal API? the above data (slope, offset, and trip_temps) would be in DT the place where they are supposed to be, instead of code.

As the DT changes, that only define compatible strings, have already gone in without any such properties set, we need to define defaults for the slope/offset and trip_temp values.

I guess (for newer SOC) you still can use the values in the DT, as (I guess) these are parsed and set in thermal_zone_device_register after the defaults are set in thermal_zone_params.

Martin

#7 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Hello Martin,

Quoted 55 lines · kernel at martin.sperl.org
Quoted 45 lines · Eduardo Valentin

Hello Martin,

Quoted 11 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

<cut>

Quoted 25 lines
drivers/thermal/bcm2835_thermal.c new file +25 -0
+
+static const struct of_device_id bcm2835_thermal_of_match_table[] = {
+	{
+		.compatible = "brcm,bcm2835-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2836-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2837-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			/* the bcm2837 needs adjustment of +5C */
+			.offset = 407000 + 5000,
+			.slope = -538,
+			.trip_temp = 80000
Quoted 9 lines

drivers/thermal/bcm2835_thermal.c line 338

Quoted 3 lines
drivers/thermal/bcm2835_thermal.c new file +3 -0
+		}
+	},
+	{},

Just for the same of clarification, is there anything preventing this driver of using of-thermal API? the above data (slope, offset, and trip_temps) would be in DT the place where they are supposed to be, instead of code.

Quoted 4 lines

As the DT changes, that only define compatible strings, have already gone in without any such properties set, we need to define defaults for the slope/offset and trip_temp values.

These properties won't go into the same node you are referring to. They go into the thermal-zone node you would create, which would then refer to the node you referred (already merged). Therefore, I do not see anything blocking a proper of-thermal usage to cover for the above data.

Quoted 3 lines

I guess (for newer SOC) you still can use the values in the DT, as (I guess) these are parsed and set in thermal_zone_device_register after the defaults are set in thermal_zone_params.

Not sure what you meant here, but these values, when correctly used in DT, they would come as part of the thermal_zone_params and in the thermal trips of the thermal zones, as the of-thermal code with already deal with those for you.

Please have a look at:
a. Documentation/devicetree/bindings/thermal/thermal.txt b. drivers/thermal/of-thermal.c

And let me know if you see anything that would prevent this driver of using the correct API to describe hardware data with DT.

BR,

Quoted 2 lines

Martin

#8 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Quoted 80 lines · Eduardo Valentin

Hello Martin,

Quoted 55 lines · kernel at martin.sperl.org
Quoted 45 lines · Eduardo Valentin

Hello Martin,

Quoted 11 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

<cut>

Quoted 25 lines
drivers/thermal/bcm2835_thermal.c new file +25 -0
+
+static const struct of_device_id bcm2835_thermal_of_match_table[] = {
+	{
+		.compatible = "brcm,bcm2835-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2836-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2837-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			/* the bcm2837 needs adjustment of +5C */
+			.offset = 407000 + 5000,
+			.slope = -538,
+			.trip_temp = 80000
Quoted 9 lines

drivers/thermal/bcm2835_thermal.c line 338

Quoted 3 lines
drivers/thermal/bcm2835_thermal.c new file +3 -0
+		}
+	},
+	{},

Just for the same of clarification, is there anything preventing this driver of using of-thermal API? the above data (slope, offset, and trip_temps) would be in DT the place where they are supposed to be, instead of code.

Quoted 4 lines

As the DT changes, that only define compatible strings, have already gone in without any such properties set, we need to define defaults for the slope/offset and trip_temp values.

These properties won't go into the same node you are referring to. They go into the thermal-zone node you would create, which would then refer to the node you referred (already merged). Therefore, I do not see anything blocking a proper of-thermal usage to cover for the above data.

Quoted 3 lines

I guess (for newer SOC) you still can use the values in the DT, as (I guess) these are parsed and set in thermal_zone_device_register after the defaults are set in thermal_zone_params.

Not sure what you meant here, but these values, when correctly used in DT, they would come as part of the thermal_zone_params and in the thermal trips of the thermal zones, as the of-thermal code with already deal with those for you.

Quoted 6 lines

Please have a look at:
a. Documentation/devicetree/bindings/thermal/thermal.txt b. drivers/thermal/of-thermal.c

And let me know if you see anything that would prevent this driver of using the correct API to describe hardware data with DT.

I guess you miss my point:
The argument is that we have DT in the 4.10.0-rc2 kernel right now that look like this:

                thermal at 7e212000 {
                        compatible = "brcm,bcm2835-thermal";
                        clocks = <0x6 0x1b>;
                        status = "okay";
                        reg = <0x7e212000 0x8>;
                }

so we still need to be compatible with those without any extra defines.

Hence we need to define those slopes and offsets in the driver itself to stay compatible with those older device-trees.

As for if it works with the framework - I have to admit I do not have the slightest clue - it looks way to complicated for the soc right now.

As a note: afaiu the trip_temp register is the temperature at which the soc will reboot on its own - similar to a watchdog, but for temperatures. (main reason for the assumption is because there is no interrupt that can get assigned a handler to catch this situation).

Martin

#9 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Hello Martin,

Quoted 84 lines · kernel at martin.sperl.org
Quoted 84 lines · Eduardo Valentin

Hello Martin,

Quoted 55 lines · kernel at martin.sperl.org
Quoted 45 lines · Eduardo Valentin

Hello Martin,

Quoted 11 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

<cut>

Quoted 25 lines
drivers/thermal/bcm2835_thermal.c new file +25 -0
+
+static const struct of_device_id bcm2835_thermal_of_match_table[] = {
+	{
+		.compatible = "brcm,bcm2835-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2836-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2837-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			/* the bcm2837 needs adjustment of +5C */
+			.offset = 407000 + 5000,
+			.slope = -538,
+			.trip_temp = 80000
Quoted 9 lines

drivers/thermal/bcm2835_thermal.c line 338

Quoted 3 lines
drivers/thermal/bcm2835_thermal.c new file +3 -0
+		}
+	},
+	{},

Just for the same of clarification, is there anything preventing this driver of using of-thermal API? the above data (slope, offset, and trip_temps) would be in DT the place where they are supposed to be, instead of code.

Quoted 4 lines

As the DT changes, that only define compatible strings, have already gone in without any such properties set, we need to define defaults for the slope/offset and trip_temp values.

These properties won't go into the same node you are referring to. They go into the thermal-zone node you would create, which would then refer to the node you referred (already merged). Therefore, I do not see anything blocking a proper of-thermal usage to cover for the above data.

Quoted 3 lines

I guess (for newer SOC) you still can use the values in the DT, as (I guess) these are parsed and set in thermal_zone_device_register after the defaults are set in thermal_zone_params.

Not sure what you meant here, but these values, when correctly used in DT, they would come as part of the thermal_zone_params and in the thermal trips of the thermal zones, as the of-thermal code with already deal with those for you.

Please have a look at:
a. Documentation/devicetree/bindings/thermal/thermal.txt b. drivers/thermal/of-thermal.c

Quoted 5 lines
Quoted 2 lines

And let me know if you see anything that would prevent this driver of using the correct API to describe hardware data with DT.

I guess you miss my point:

Maybe you missed, or did not read the doc I pointed you...

Quoted 9 lines

The argument is that we have DT in the 4.10.0-rc2 kernel right now that look like this:

                thermal at 7e212000 {
                        compatible = "brcm,bcm2835-thermal";
                        clocks = <0x6 0x1b>;
                        status = "okay";
                        reg = <0x7e212000 0x8>;
                }

so we still need to be compatible with those without any extra defines.

Yes, but the above DT entry will still be valid if you add the correct of-thermal support. In fact, you would add in your DTS a thermal-zones node, and in one of the defined zone, you would then reference the node you already got into mainline. Below is a copy of the doc I mentioned before:

#include <dt-bindings/thermal/thermal.h>

ocp {
	...
	/*
	 * A simple IC with several bandgap temperature sensors.
	 */
	bandgap0: bandgap at 0x0000ED00 {
		...
		#thermal-sensor-cells = <1>;
	};
};
thermal-zones {
	cpu_thermal: cpu-thermal {
		polling-delay-passive = <250>; /* milliseconds */
		polling-delay = <1000>; /* milliseconds */

				/* sensor       ID */
		thermal-sensors = <&bandgap0     0>;

		trips {
			/* each zone within the SoC may have its own trips */
			cpu_alert: cpu-alert {
				temperature = <100000>; /* millicelsius */
				hysteresis = <2000>; /* millicelsius */
				type = "passive";
			};
			cpu_crit: cpu-crit {
				temperature = <125000>; /* millicelsius */
				hysteresis = <2000>; /* millicelsius */
				type = "critical";
			};
		};

		cooling-maps {
			/* each zone within the SoC may have its own cooling */
			...
		};
	};
Quoted 3 lines

Hence we need to define those slopes and offsets in the driver itself to stay compatible with those older device-trees.

not really, as long as we do not merge the driver with the missing of-thermal support, I see no need to have both supports in your driver, i.e., if we start correct for the beggining there is no need to have offsets and slopes data in your driver code.

Quoted 4 lines

As for if it works with the framework - I have to admit I do not have the slightest clue - it looks way to complicated for the soc right now.

Well, there is the documentation I mentioned, which several other drivers used as base for their support. You can also look at other thermal zones already defined in the existing DTS(I)'s.

Quoted 6 lines

As a note: afaiu the trip_temp register is the temperature at which the soc will reboot on its own - similar to a watchdog, but for temperatures. (main reason for the assumption is because there is no interrupt that can get assigned a handler to catch this situation).

OK. But that does not prevent you to have other trips so your running system can act before the shutdown trip is crossed.

Quoted 3 lines

Martin

BR,

Eduardo Valentin

#10 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Quoted 175 lines · Eduardo Valentin

Hello Martin,

Quoted 84 lines · kernel at martin.sperl.org
Quoted 84 lines · Eduardo Valentin

Hello Martin,

Quoted 55 lines · kernel at martin.sperl.org
Quoted 45 lines · Eduardo Valentin

Hello Martin,

Quoted 11 lines · kernel at martin.sperl.org

From: Martin Sperl <[email protected]>

Add basic thermal driver for bcm2835 SOC.

This driver currently relies on the firmware setting up the tsense HW block and does not set it up itself.

<cut>

Quoted 25 lines
drivers/thermal/bcm2835_thermal.c new file +25 -0
+
+static const struct of_device_id bcm2835_thermal_of_match_table[] = {
+	{
+		.compatible = "brcm,bcm2835-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2836-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			.offset = 407000,
+			.slope = -538,
+			.trip_temp = 80000
+		}
+	},
+	{
+		.compatible = "brcm,bcm2837-thermal",
+		.data = &(struct bcm2835_thermal_info) {
+			/* the bcm2837 needs adjustment of +5C */
+			.offset = 407000 + 5000,
+			.slope = -538,
+			.trip_temp = 80000
Quoted 9 lines

drivers/thermal/bcm2835_thermal.c line 338

Quoted 3 lines
drivers/thermal/bcm2835_thermal.c new file +3 -0
+		}
+	},
+	{},

Just for the same of clarification, is there anything preventing this driver of using of-thermal API? the above data (slope, offset, and trip_temps) would be in DT the place where they are supposed to be, instead of code.

Quoted 4 lines

As the DT changes, that only define compatible strings, have already gone in without any such properties set, we need to define defaults for the slope/offset and trip_temp values.

These properties won't go into the same node you are referring to. They go into the thermal-zone node you would create, which would then refer to the node you referred (already merged). Therefore, I do not see anything blocking a proper of-thermal usage to cover for the above data.

Quoted 3 lines

I guess (for newer SOC) you still can use the values in the DT, as (I guess) these are parsed and set in thermal_zone_device_register after the defaults are set in thermal_zone_params.

Not sure what you meant here, but these values, when correctly used in DT, they would come as part of the thermal_zone_params and in the thermal trips of the thermal zones, as the of-thermal code with already deal with those for you.

Please have a look at:
a. Documentation/devicetree/bindings/thermal/thermal.txt b. drivers/thermal/of-thermal.c

Quoted 5 lines
Quoted 2 lines

And let me know if you see anything that would prevent this driver of using the correct API to describe hardware data with DT.

I guess you miss my point:

Maybe you missed, or did not read the doc I pointed you...

Quoted 9 lines

The argument is that we have DT in the 4.10.0-rc2 kernel right now that look like this:

               thermal at 7e212000 {
                       compatible = "brcm,bcm2835-thermal";
                       clocks = <0x6 0x1b>;
                       status = "okay";
                       reg = <0x7e212000 0x8>;
               }

so we still need to be compatible with those without any extra defines.

Yes, but the above DT entry will still be valid if you add the correct of-thermal support. In fact, you would add in your DTS a thermal-zones node, and in one of the defined zone, you would then reference the node you already got into mainline. Below is a copy of the doc I mentioned before:

#include <dt-bindings/thermal/thermal.h>

ocp {
	...
	/*
	 * A simple IC with several bandgap temperature sensors.
	 */
	bandgap0: bandgap at 0x0000ED00 {
		...
		#thermal-sensor-cells = <1>;
	};
};
thermal-zones {
	cpu_thermal: cpu-thermal {
		polling-delay-passive = <250>; /* milliseconds */
		polling-delay = <1000>; /* milliseconds */

				/* sensor       ID */
		thermal-sensors = <&bandgap0     0>;

		trips {
			/* each zone within the SoC may have its own trips */
			cpu_alert: cpu-alert {
				temperature = <100000>; /* millicelsius */
				hysteresis = <2000>; /* millicelsius */
				type = "passive";
			};
			cpu_crit: cpu-crit {
				temperature = <125000>; /* millicelsius */
				hysteresis = <2000>; /* millicelsius */
				type = "critical";
			};
		};

		cooling-maps {
			/* each zone within the SoC may have its own cooling */
			...
		};
	};
Quoted 3 lines

Hence we need to define those slopes and offsets in the driver itself to stay compatible with those older device-trees.

not really, as long as we do not merge the driver with the missing of-thermal support, I see no need to have both supports in your driver, i.e., if we start correct for the beggining there is no need to have offsets and slopes data in your driver code.

Quoted 4 lines

As for if it works with the framework - I have to admit I do not have the slightest clue - it looks way to complicated for the soc right now.

Well, there is the documentation I mentioned, which several other drivers used as base for their support. You can also look at other thermal zones already defined in the existing DTS(I)'s.

Quoted 1 line
Quoted 12 lines
Quoted 5 lines

As a note: afaiu the trip_temp register is the temperature at which the soc will reboot on its own - similar to a watchdog, but for temperatures. (main reason for the assumption is because there is no interrupt that can get assigned a handler to catch this situation).

OK. But that does not prevent you to have other trips so your running system can act before the shutdown trip is crossed.

Quoted 3 lines

Martin

So how does this change/impact the driver code itself?

Defining a thermal zone in the dts is just an additional feature that only impacts the device tree.
The DT as is works fine and at least allows to read the current temperature.

So can we just merge the driver now and look into the DT separately?

Martin

#11 lore raw

PATCH 1/2 dt-bindings: Add thermal zone to bcm2835-thermal example

Add a thermal zone in order to make the example complete.

---
 .../bindings/thermal/brcm,bcm2835-thermal.txt      |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
Documentation/devicetree/bindings/thermal/brcm,bcm2835-thermal.txt +19 -0
index 474531d..680aca9 100644
@@ -10,6 +10,25 @@ clocks: 	Phandle of the clock used by the thermal sensor.
 
 Example:
 
+thermal-zones {
+	gpu_thermal: gpu-thermal {
+		polling-delay-passive = <0>;
+		polling-delay = <1000>;
+
+		thermal-sensors = <&thermal>;
+
+		trips {
+			cpu-crit {
+				temperature	= <80000>;
+				hysteresis	= <0>;
+				type		= "critical";
+			};
+		};
+		cooling-maps {
+		};
+	};
+};
+
 thermal: thermal@7e212000 {
 	compatible = "brcm,bcm2835-thermal";
 	reg = <0x7e212000 0x8>;
Signature · 1 line
1.7.9.5
#12 lore raw

PATCH 2/2 ARM: dts: bcm283x: Add critical thermal zone for GPU

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +19 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;
+
+			trips {
+				cpu-crit {
+					temperature	= <80000>;
+					hysteresis	= <0>;
+					type		= "critical";
+				};
+			};
+			cooling-maps {
+			};
+		};
+	};
+
 	soc {
 		compatible = "simple-bus";
 		#address-cells = <1>;
Signature · 1 line
1.7.9.5

A message that is not in the archive (sent off-list or not yet received): 1486217787-15703-2-git-send-email-stefan.wahren-eS4NqCHxEME@public.gmane.org

#14 lore raw

Re: [PATCH 2/2] ARM: dts: bcm283x: Add critical thermal zone for GPU

Quoted 26 lines · Stefan Wahren

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +9 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;
+
+			trips {
+				cpu-crit {

arch/arm/boot/dts/bcm283x.dtsi line 40

Quoted 10 lines
arch/arm/boot/dts/bcm283x.dtsi +10 -0
+					temperature	= <80000>;
+					hysteresis	= <0>;
+					type		= "critical";
+				};
+			};
+			cooling-maps {
+			};
+		};
+	};
+

This is fine with me. Checking the other patches in the series.

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +0 -0
 	soc {
 		compatible = "simple-bus";
 		#address-cells = <1>;
Signature · 1 line
1.7.9.5

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

A message that is not in the archive (sent off-list or not yet received): 20170208041929.GA6809-bi+AKbBUZKY6gyzm1THtWbp2dZbC/[email protected]

#16 lore raw

Re: [PATCH 2/2] ARM: dts: bcm283x: Add critical thermal zone for GPU

Quoted 18 lines · Eduardo Valentin
Quoted 17 lines · Stefan Wahren

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +0 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
Quoted 6 lines

arch/arm/boot/dts/bcm283x.dtsi line 27

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +6 -0
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;

Just came to my mind, dont you need to have an id to specify with sensor points to gpu?

Quoted 7 lines
Quoted 7 lines
arch/arm/boot/dts/bcm283x.dtsi +7 -0
+
+			trips {
+				cpu-crit {
+					temperature	= <80000>;
+					hysteresis	= <0>;
+					type		= "critical";
+				};
Quoted 15 lines

arch/arm/boot/dts/bcm283x.dtsi line 40

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +6 -0
+			};
+			cooling-maps {
+			};
+		};
+	};
+

This is fine with me. Checking the other patches in the series.

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +0 -0
 	soc {
 		compatible = "simple-bus";
 		#address-cells = <1>;
Signature · 1 line
1.7.9.5

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

A message that is not in the archive (sent off-list or not yet received): 20170208042351.GB6809-bi+AKbBUZKY6gyzm1THtWbp2dZbC/[email protected]

#18 lore raw

Re: [PATCH 2/2] ARM: dts: bcm283x: Add critical thermal zone for GPU

Am 08.02.2017 um 05:23 schrieb Eduardo Valentin:

Quoted 19 lines · Eduardo Valentin
Quoted 18 lines · Stefan Wahren
Quoted 17 lines

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +0 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
Quoted 8 lines
Quoted 6 lines

arch/arm/boot/dts/bcm283x.dtsi line 27

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +6 -0
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;

Just came to my mind, dont you need to have an id to specify with sensor points to gpu?

Sorry, i don't know the exact setup of the single thermal sensor on the SoC (datasheet doesn't provide any helpful information). I adapted the Renesas R-Car thermal binding because i think it would be the best match.

@Eric: What's your opinion?

Quoted 16 lines
Quoted 16 lines
Quoted 4 lines
arch/arm/boot/dts/bcm283x.dtsi +3 -0
+
+			trips {
+				cpu-crit {

arch/arm/boot/dts/bcm283x.dtsi line 40

Quoted 10 lines
arch/arm/boot/dts/bcm283x.dtsi +10 -0
+					temperature	= <80000>;
+					hysteresis	= <0>;
+					type		= "critical";
+				};
+			};
+			cooling-maps {
+			};
+		};
+	};
+

This is fine with me. Checking the other patches in the series.

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +0 -0
 	soc {
 		compatible = "simple-bus";
 		#address-cells = <1>;
Signature · 1 line
1.7.9.5

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#19 lore raw

Re: [PATCH 2/2] ARM: dts: bcm283x: Add critical thermal zone for GPU

Quoted 29 lines · Stefan Wahren

Am 08.02.2017 um 05:23 schrieb Eduardo Valentin:

Quoted 19 lines · Eduardo Valentin
Quoted 18 lines · Stefan Wahren
Quoted 17 lines

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +0 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
Quoted 8 lines
Quoted 6 lines

arch/arm/boot/dts/bcm283x.dtsi line 27

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +6 -0
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;

Just came to my mind, dont you need to have an id to specify with sensor points to gpu?

Quoted 5 lines

Sorry, i don't know the exact setup of the single thermal sensor on the SoC (datasheet doesn't provide any helpful information). I adapted the Renesas R-Car thermal binding because i think it would be the best match.

@Eric: What's your opinion?

I don't understand the question.

#20 lore raw

Re: [PATCH 2/2] ARM: dts: bcm283x: Add critical thermal zone for GPU

Hi Eric,

Quoted 34 lines

Eric Anholt <[email protected]> hat am 8. Februar 2017 um 20:50 geschrieben:

Quoted 29 lines · Stefan Wahren

Am 08.02.2017 um 05:23 schrieb Eduardo Valentin:

Quoted 27 lines · Eduardo Valentin
Quoted 18 lines · Stefan Wahren
Quoted 17 lines

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +0 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
Quoted 6 lines

arch/arm/boot/dts/bcm283x.dtsi line 27

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +6 -0
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;

Just came to my mind, dont you need to have an id to specify with sensor points to gpu?

Quoted 7 lines
Quoted 5 lines

Sorry, i don't know the exact setup of the single thermal sensor on the SoC (datasheet doesn't provide any helpful information). I adapted the Renesas R-Car thermal binding because i think it would be the best match.

@Eric: What's your opinion?

I don't understand the question.

i hope to get it right. We are talking about defining the thermal zone [1].

There are 4 examples:

(a) - CPU thermal zone with one internal sensor
(b) - IC with several internal sensors
(c) - Several sensors within one single thermal zone
(d) - Board thermal

I decided to choose (a) for the patch and Eduardo tends to (b).

Here are the questions:

Where is the thermal sensor TSENS located (ARM core or VideoCore 4)?

Do we expect several internal sensors?

[1] - http://lxr.free-electrons.com/source/Documentation/devicetree/bindings/thermal/thermal.txt

A message that is not in the archive (sent off-list or not yet received): 124007607.358509.1486662532562-7tX72C7vayboQLBSYMtkGA@public.gmane.org

#22 lore raw

Re: [PATCH 2/2] ARM: dts: bcm283x: Add critical thermal zone for GPU

Quoted 53 lines · Stefan Wahren

Hi Eric,

Quoted 34 lines

Eric Anholt <[email protected]> hat am 8. Februar 2017 um 20:50 geschrieben:

Quoted 29 lines · Stefan Wahren

Am 08.02.2017 um 05:23 schrieb Eduardo Valentin:

Quoted 27 lines · Eduardo Valentin
Quoted 18 lines · Stefan Wahren
Quoted 17 lines

As suggested by Eduardo Valentin this adds the thermal zone for the bcm2835 SoC. Since we currently don't have any cooling devices leave this section empty.

---
 arch/arm/boot/dts/bcm283x.dtsi |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
arch/arm/boot/dts/bcm283x.dtsi +0 -0
index 9c7ec7e..359da9d 100644
@@ -19,6 +19,25 @@
 		bootargs = "earlyprintk console=ttyAMA0";
 	};
 
Quoted 6 lines

arch/arm/boot/dts/bcm283x.dtsi line 27

Quoted 6 lines
arch/arm/boot/dts/bcm283x.dtsi +6 -0
+	thermal-zones {
+		gpu_thermal: gpu-thermal {
+			polling-delay-passive = <0>;
+			polling-delay = <1000>;
+
+			thermal-sensors = <&thermal>;

Just came to my mind, dont you need to have an id to specify with sensor points to gpu?

Quoted 7 lines
Quoted 5 lines

Sorry, i don't know the exact setup of the single thermal sensor on the SoC (datasheet doesn't provide any helpful information). I adapted the Renesas R-Car thermal binding because i think it would be the best match.

@Eric: What's your opinion?

I don't understand the question.

i hope to get it right. We are talking about defining the thermal zone [1].

There are 4 examples:

(a) - CPU thermal zone with one internal sensor
(b) - IC with several internal sensors
(c) - Several sensors within one single thermal zone
(d) - Board thermal

Quoted 7 lines

I decided to choose (a) for the patch and Eduardo tends to (b).

Here are the questions:

Where is the thermal sensor TSENS located (ARM core or VideoCore 4)?

Do we expect several internal sensors?

There is only one temperature sensor on the chip, and no others on the board that I know of. The docs for tsens don't say if the actual sensor is closer to the center of the ARM or some VC4 component within the chip.

#24 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

In reply to kernel at martin.sperl.org
Quoted 7 lines · kernel at martin.sperl.org

So how does this change/impact the driver code itself?

Defining a thermal zone in the dts is just an additional feature that only impacts the device tree.
The DT as is works fine and at least allows to read the current temperature.

Well, your driver is still half finished. It is a DT only driver, which does not follow the DT spec on thermal. The impact on your code is that it wont need to carry your hardware data in the source code.

Also, having the data described in DT allows you porting your driver without patching it, in case you need, or any other system engineer, to have different thermal data, trips values, number or trips, offset and slope per sensor, on different boards, or even on different chip version.

Not to say that having the data described in DT also allows system engineers to deploy different thermal zones, based on your driver/sensor, to extrapolate different hotspots.

Quoted 2 lines

Martin

BR,

Eduardo Valentin

#25 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

Quoted 22 lines · Eduardo Valentin
Quoted 7 lines · kernel at martin.sperl.org

So how does this change/impact the driver code itself?

Defining a thermal zone in the dts is just an additional feature that only impacts the device tree.
The DT as is works fine and at least allows to read the current temperature.

Well, your driver is still half finished. It is a DT only driver, which does not follow the DT spec on thermal. The impact on your code is that it wont need to carry your hardware data in the source code.

Also, having the data described in DT allows you porting your driver without patching it, in case you need, or any other system engineer, to have different thermal data, trips values, number or trips, offset and slope per sensor, on different boards, or even on different chip version.

Quoted 4 lines

Not to say that having the data described in DT also allows system engineers to deploy different thermal zones, based on your driver/sensor, to extrapolate different hotspots.

That is all true and as far as I understand you can do all of that using the current driver - see the patch-sets by Stefan Wahren that incorporates this into the dts.

My point is still: there are dtb?s that come with 4.10-rcX that do not have all of those defined

But as the mantra for DT is: "it is an API?, we have to be backwards compatible from the driver perspective.
So we need those values in the driver to ensure the older version of dtb?s are working correctly. That is why it is still included.

Also the settings of the trip are for the HW trip point for the case that a future version of the firmware does not set up the thermal HW block.

Do you really want to break this Mantra of backwards compatibility by having those compatiblity ?defines? stripped out with all the possible negative side-effects if things are not set up correctly?

Thanks,
Martin

P.s: I find it strange that V3 of the driver was sent out in May 2016 (way before the DTS changes got merged) and you only started to give feedback in November only in V8 (that primarily incorporated minor changes and a rebase to 4.9) by which time the changes to the dts have already been merged.

#26 lore raw

PATCH v9 1/1 thermal: bcm2835: add thermal driver for bcm2835 soc

In reply to Eduardo Valentin

Hi Eduardo,

Quoted 40 lines

Eduardo Valentin <[email protected]> hat am 2. Februar 2017 um 05:29 geschrieben:

Hello Martin,

...

#include <dt-bindings/thermal/thermal.h>

ocp {
	...
	/*
	 * A simple IC with several bandgap temperature sensors.
	 */
	bandgap0: bandgap at 0x0000ED00 {
		...
		#thermal-sensor-cells = <1>;
	};
};
thermal-zones {
	cpu_thermal: cpu-thermal {
		polling-delay-passive = <250>; /* milliseconds */
		polling-delay = <1000>; /* milliseconds */

				/* sensor       ID */
		thermal-sensors = <&bandgap0     0>;

		trips {
			/* each zone within the SoC may have its own trips */
			cpu_alert: cpu-alert {
				temperature = <100000>; /* millicelsius */
				hysteresis = <2000>; /* millicelsius */
				type = "passive";
			};
			cpu_crit: cpu-crit {
				temperature = <125000>; /* millicelsius */
				hysteresis = <2000>; /* millicelsius */
				type = "critical";
Quoted 9 lines
			};
		};

		cooling-maps {
			/* each zone within the SoC may have its own cooling */
			...
		};
	};

if i get it right the device tree binding requires also a cooling map. But how should we model this without a fan or a DVFS driver? Is there something like a placeholder? Do you have an example?

Regards
Stefan