[PATCH 1/5] imx8mm_evk_defconfig: Select CONFIG_IMX_TMU

[PATCH 1/5] imx8mm_evk_defconfig: Select CONFIG_IMX_TMU

[Fabio Estevam]

From: Fabio Estevam <festevam@denx.de>

Select the i.MX8MM thermal driver as it is useful for displaying
the CPU temperature and its grading:

CPU:   Commercial temperature grade (0C to 95C) at 38C

It also prevents booting when the temperature is above the alert
point.

Signed-off-by: Fabio Estevam <festevam@denx.de>
---
 configs/imx8mm_evk_defconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/configs/imx8mm_evk_defconfig b/configs/imx8mm_evk_defconfig
index ac9810fe1b..822c2fbfab 100644
--- a/configs/imx8mm_evk_defconfig
+++ b/configs/imx8mm_evk_defconfig
@@ -104,6 +104,7 @@ CONFIG_SPL_SYSRESET=y
 CONFIG_SYSRESET_PSCI=y
 CONFIG_SYSRESET_WATCHDOG=y
 CONFIG_DM_THERMAL=y
+CONFIG_IMX_TMU=y
 CONFIG_USB=y
 CONFIG_SPL_USB_HOST=y
 CONFIG_USB_EHCI_HCD=y
-- 
2.34.1

[PATCH 2/5] thermal: imx_tmu: Fix the polling default

[Fabio Estevam]

From: Fabio Estevam <festevam@denx.de>

When the 'polling-delay' property is not passed via devicetree,
pdata->polling_delay keeps at 0. This causes the imx_tmu driver to get
stuck inside the busy while() loop when the CPU temperature is above
the alert point.

Fix this problem by passing a one second polling interval, which provides
a proper delay to let the system to cool down and exit the while() loop
when the temperature is below the alert point.

Signed-off-by: Fabio Estevam <festevam@denx.de>
---
 drivers/thermal/imx_tmu.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/thermal/imx_tmu.c b/drivers/thermal/imx_tmu.c
index 97efc55044..d9a04eaf79 100644
--- a/drivers/thermal/imx_tmu.c
+++ b/drivers/thermal/imx_tmu.c
@@ -37,6 +37,7 @@ DECLARE_GLOBAL_DATA_PTR;
 #define TER_ADC_PD		0x40000000
 #define TER_ALPF		0x3
 
+#define IMX_TMU_POLLING_DELAY_MS	1000
 /*
  * i.MX TMU Registers
  */
@@ -574,6 +575,8 @@ static int imx_tmu_parse_fdt(struct udevice *dev)
 
 	dev_dbg(dev, "%s\n", __func__);
 
+	pdata->polling_delay = IMX_TMU_POLLING_DELAY_MS;
+
 	if (pdata->zone_node) {
 		pdata->regs = (union tmu_regs *)dev_read_addr_ptr(dev);
 
@@ -602,7 +605,8 @@ static int imx_tmu_parse_fdt(struct udevice *dev)
 
 	dev_dbg(dev, "args.args_count %d, id %d\n", args.args_count, pdata->id);
 
-	pdata->polling_delay = dev_read_u32_default(dev, "polling-delay", 1000);
+	pdata->polling_delay = dev_read_u32_default(dev, "polling-delay",
+						    IMX_TMU_POLLING_DELAY_MS);
 
 	trips_np = ofnode_path("/thermal-zones/cpu-thermal/trips");
 	ofnode_for_each_subnode(trips_np, trips_np) {
-- 
2.34.1

[PATCH 3/5] thermal: imx_tmu: Select LOG

[Fabio Estevam]

From: Fabio Estevam <festevam@denx.de>

Without LOG being selected, the dev_info() information
from the imx_tmu driver cannot be printed.

Select the LOG option, so that no dev_info() messages are lost.

Signed-off-by: Fabio Estevam <festevam@denx.de>
---
 drivers/thermal/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
index 681b621760..0a1d11e216 100644
--- a/drivers/thermal/Kconfig
+++ b/drivers/thermal/Kconfig
@@ -29,6 +29,7 @@ config IMX_SCU_THERMAL
 config IMX_TMU
         bool "Thermal Management Unit driver for NXP i.MX8M and iMX93"
         depends on ARCH_IMX8M || IMX93
+        select LOG
         help
           Support for Temperature sensors on NXP i.MX8M and iMX93.
           It supports one critical trip point and one passive trip point.
-- 
2.34.1

[PATCH 4/5] thermal: imx_tmu: Fix th temperature unit

[Fabio Estevam]

From: Fabio Estevam <festevam@denx.de>

The temperature unit is millidegree Celsius, so divide by 1000 to correctly
print the temperature values in Celsius.

While at it, also change a typo: "has beyond" to "is beyond".

Signed-off-by: Fabio Estevam <festevam@denx.de>
---
 drivers/thermal/imx_tmu.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/thermal/imx_tmu.c b/drivers/thermal/imx_tmu.c
index d9a04eaf79..f79b583811 100644
--- a/drivers/thermal/imx_tmu.c
+++ b/drivers/thermal/imx_tmu.c
@@ -238,8 +238,8 @@ int imx_tmu_get_temp(struct udevice *dev, int *temp)
 		return ret;
 
 	while (cpu_tmp >= pdata->alert) {
-		dev_info(dev, "CPU Temperature (%dC) has beyond alert (%dC), close to critical (%dC) waiting...\n",
-			 cpu_tmp, pdata->alert, pdata->critical);
+		dev_info(dev, "CPU Temperature (%dC) is beyond alert (%dC), close to critical (%dC) waiting...\n",
+			 cpu_tmp / 1000, pdata->alert / 1000, pdata->critical / 1000);
 		mdelay(pdata->polling_delay);
 		ret = read_temperature(dev, &cpu_tmp);
 		if (ret)
-- 
2.34.1

[PATCH 5/5] thermal: imx_tmu: Increase the polling interval

[Fabio Estevam]

From: Fabio Estevam <festevam@denx.de>

Polling every second to check whether the CPU has cooled down is
too frequent.

Allow more time for the CPU to cool down by increasing the polling
interval to 5 seconds by defaut.

This value is used in the absence of the 'polling-delay' devicetree
property.

Signed-off-by: Fabio Estevam <festevam@denx.de>
---
 drivers/thermal/imx_tmu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/thermal/imx_tmu.c b/drivers/thermal/imx_tmu.c
index f79b583811..15e01b6166 100644
--- a/drivers/thermal/imx_tmu.c
+++ b/drivers/thermal/imx_tmu.c
@@ -37,7 +37,7 @@ DECLARE_GLOBAL_DATA_PTR;
 #define TER_ADC_PD		0x40000000
 #define TER_ALPF		0x3
 
-#define IMX_TMU_POLLING_DELAY_MS	1000
+#define IMX_TMU_POLLING_DELAY_MS	5000
 /*
  * i.MX TMU Registers
  */
-- 
2.34.1

Re: [PATCH 3/5] thermal: imx_tmu: Select LOG

[Tom Rini]

[-- Attachment #1: Type: text/plain, Size: 1138 bytes --]

On Wed, Aug 23, 2023 at 02:26:35PM -0300, Fabio Estevam wrote:

> From: Fabio Estevam <festevam@denx.de>
> 
> Without LOG being selected, the dev_info() information
> from the imx_tmu driver cannot be printed.
> 
> Select the LOG option, so that no dev_info() messages are lost.
> 
> Signed-off-by: Fabio Estevam <festevam@denx.de>
> ---
>  drivers/thermal/Kconfig | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
> index 681b621760..0a1d11e216 100644
> --- a/drivers/thermal/Kconfig
> +++ b/drivers/thermal/Kconfig
> @@ -29,6 +29,7 @@ config IMX_SCU_THERMAL
>  config IMX_TMU
>          bool "Thermal Management Unit driver for NXP i.MX8M and iMX93"
>          depends on ARCH_IMX8M || IMX93
> +        select LOG
>          help
>            Support for Temperature sensors on NXP i.MX8M and iMX93.
>            It supports one critical trip point and one passive trip point.

This is now going to bring in a ton of other messages.  Perhaps the
problem is that there are critical messages in this driver which should
not be dev_info ?

-- 
Tom

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 659 bytes --]

Re: [PATCH 4/5] thermal: imx_tmu: Fix th temperature unit

[Tom Rini]

[-- Attachment #1: Type: text/plain, Size: 1421 bytes --]

On Wed, Aug 23, 2023 at 02:26:36PM -0300, Fabio Estevam wrote:
> From: Fabio Estevam <festevam@denx.de>
> 
> The temperature unit is millidegree Celsius, so divide by 1000 to correctly
> print the temperature values in Celsius.
> 
> While at it, also change a typo: "has beyond" to "is beyond".
> 
> Signed-off-by: Fabio Estevam <festevam@denx.de>
> ---
>  drivers/thermal/imx_tmu.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/thermal/imx_tmu.c b/drivers/thermal/imx_tmu.c
> index d9a04eaf79..f79b583811 100644
> --- a/drivers/thermal/imx_tmu.c
> +++ b/drivers/thermal/imx_tmu.c
> @@ -238,8 +238,8 @@ int imx_tmu_get_temp(struct udevice *dev, int *temp)
>  		return ret;
>  
>  	while (cpu_tmp >= pdata->alert) {
> -		dev_info(dev, "CPU Temperature (%dC) has beyond alert (%dC), close to critical (%dC) waiting...\n",
> -			 cpu_tmp, pdata->alert, pdata->critical);
> +		dev_info(dev, "CPU Temperature (%dC) is beyond alert (%dC), close to critical (%dC) waiting...\n",
> +			 cpu_tmp / 1000, pdata->alert / 1000, pdata->critical / 1000);
>  		mdelay(pdata->polling_delay);
>  		ret = read_temperature(dev, &cpu_tmp);
>  		if (ret)

To follow-up on the patch about select'ing LOG, this shouldn't be
dev_info, it should be higher, at least warning if not critical.  And
given the default LOGLEVEL of 4, it will be included and printed.

-- 
Tom

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 659 bytes --]