[PATCH v2 0/4] thermal: mediatek: Add support for MT8365 SoC

[PATCH v2 0/4] thermal: mediatek: Add support for MT8365 SoC

[Amjad Ouled-Ameur]

This patchset adds thermal support for MT8365 SoC which contains three
thermal sensors.

Note that the V1 [0] was originally introduced by Fabien Parent [1]. He
requested that I take over this patchset.

[0]: https://lore.kernel.org/linux-arm-kernel/20220530183833.863040-1-fparent@baylibre.com/
[1]: Fabien Parent <fparent@baylibre.com>

Amjad Ouled-Ameur (1):
  thermal: mediatek: add another get_temp ops for thermal sensors

Fabien Parent (2):
  dt-bindings: thermal: mediatek: add binding documentation for MT8365
    SoC
  thermal: mediatek: add support for MT8365 SoC

Markus Schneider-Pargmann (1):
  thermal: mediatek: control buffer enablement tweaks

 .../bindings/thermal/mediatek-thermal.txt     |   1 +
 drivers/thermal/mtk_thermal.c                 | 191 +++++++++++++++---
 2 files changed, 162 insertions(+), 30 deletions(-)

-- 
2.37.1

[PATCH v2 1/4] dt-bindings: thermal: mediatek: add binding documentation for MT8365 SoC

[Amjad Ouled-Ameur]

From: Fabien Parent <fparent@baylibre.com>

Add the binding documentation for the thermal support on MT8365 SoC.

Signed-off-by: Fabien Parent <fparent@baylibre.com>
Acked-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
---
 Documentation/devicetree/bindings/thermal/mediatek-thermal.txt | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt b/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
index 5c7e7bdd029a..ba4ebffeade4 100644
--- a/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
+++ b/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
@@ -14,6 +14,7 @@ Required properties:
   - "mediatek,mt2712-thermal" : For MT2712 family of SoCs
   - "mediatek,mt7622-thermal" : For MT7622 SoC
   - "mediatek,mt8183-thermal" : For MT8183 family of SoCs
+  - "mediatek,mt8365-thermal" : For MT8365 family of SoCs
   - "mediatek,mt8516-thermal", "mediatek,mt2701-thermal : For MT8516 family of SoCs
 - reg: Address range of the thermal controller
 - interrupts: IRQ for the thermal controller
-- 
2.37.1

Re: [PATCH v2 1/4] dt-bindings: thermal: mediatek: add binding documentation for MT8365 SoC

[AngeloGioacchino Del Regno]

Il 20/07/22 20:18, Amjad Ouled-Ameur ha scritto:
> From: Fabien Parent <fparent@baylibre.com>
> 
> Add the binding documentation for the thermal support on MT8365 SoC.
> 
> Signed-off-by: Fabien Parent <fparent@baylibre.com>
> Acked-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>

Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>

Re: [PATCH v2 1/4] dt-bindings: thermal: mediatek: add binding documentation for MT8365 SoC

[Matthias Brugger]

On 20/07/2022 20:18, Amjad Ouled-Ameur wrote:
> From: Fabien Parent <fparent@baylibre.com>
> 
> Add the binding documentation for the thermal support on MT8365 SoC.
> 
> Signed-off-by: Fabien Parent <fparent@baylibre.com>
> Acked-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>

I suppose you wanted to add Signed-off-by here, at least get_maintainers.pl 
doesn't list you.

Apart from that:
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>

> ---
>   Documentation/devicetree/bindings/thermal/mediatek-thermal.txt | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt b/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
> index 5c7e7bdd029a..ba4ebffeade4 100644
> --- a/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
> +++ b/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
> @@ -14,6 +14,7 @@ Required properties:
>     - "mediatek,mt2712-thermal" : For MT2712 family of SoCs
>     - "mediatek,mt7622-thermal" : For MT7622 SoC
>     - "mediatek,mt8183-thermal" : For MT8183 family of SoCs
> +  - "mediatek,mt8365-thermal" : For MT8365 family of SoCs
>     - "mediatek,mt8516-thermal", "mediatek,mt2701-thermal : For MT8516 family of SoCs
>   - reg: Address range of the thermal controller
>   - interrupts: IRQ for the thermal controller

Re: [PATCH v2 1/4] dt-bindings: thermal: mediatek: add binding documentation for MT8365 SoC

[Amjad Ouled-Ameur]

Hi Matthias,

On 7/21/22 10:57, Matthias Brugger wrote:
>
>
> On 20/07/2022 20:18, Amjad Ouled-Ameur wrote:
>> From: Fabien Parent <fparent@baylibre.com>
>>
>> Add the binding documentation for the thermal support on MT8365 SoC.
>>
>> Signed-off-by: Fabien Parent <fparent@baylibre.com>
>> Acked-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
>
> I suppose you wanted to add Signed-off-by here, at least 
> get_maintainers.pl doesn't list you.
>
Right, will add my SoB in V3, thank you!

Regards,

Amjad

> Apart from that:
> Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
>
>> ---
>>   Documentation/devicetree/bindings/thermal/mediatek-thermal.txt | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git 
>> a/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt 
>> b/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
>> index 5c7e7bdd029a..ba4ebffeade4 100644
>> --- a/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
>> +++ b/Documentation/devicetree/bindings/thermal/mediatek-thermal.txt
>> @@ -14,6 +14,7 @@ Required properties:
>>     - "mediatek,mt2712-thermal" : For MT2712 family of SoCs
>>     - "mediatek,mt7622-thermal" : For MT7622 SoC
>>     - "mediatek,mt8183-thermal" : For MT8183 family of SoCs
>> +  - "mediatek,mt8365-thermal" : For MT8365 family of SoCs
>>     - "mediatek,mt8516-thermal", "mediatek,mt2701-thermal : For 
>> MT8516 family of SoCs
>>   - reg: Address range of the thermal controller
>>   - interrupts: IRQ for the thermal controller

[PATCH v2 2/4] thermal: mediatek: add support for MT8365 SoC

[Amjad Ouled-Ameur]

From: Fabien Parent <fparent@baylibre.com>

MT8365 is similar to the other SoCs supported by the driver. It has only
one bank and 3 actual sensors that can be multiplexed. There is another
one sensor that does not have usable data.

Signed-off-by: Fabien Parent <fparent@baylibre.com>
Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
---
 drivers/thermal/mtk_thermal.c | 68 +++++++++++++++++++++++++++++++++++
 1 file changed, 68 insertions(+)

diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
index ede94eadddda..216eb0d61c5b 100644
--- a/drivers/thermal/mtk_thermal.c
+++ b/drivers/thermal/mtk_thermal.c
@@ -31,6 +31,7 @@
 #define AUXADC_CON2_V		0x010
 #define AUXADC_DATA(channel)	(0x14 + (channel) * 4)
 
+#define APMIXED_SYS_TS_CON0	0x600
 #define APMIXED_SYS_TS_CON1	0x604
 
 /* Thermal Controller Registers */
@@ -245,6 +246,17 @@ enum mtk_thermal_version {
 /* The calibration coefficient of sensor  */
 #define MT8183_CALIBRATION	153
 
+/* MT8365 */
+#define MT8365_TEMP_AUXADC_CHANNEL 11
+#define MT8365_CALIBRATION 164
+#define MT8365_NUM_CONTROLLER 1
+#define MT8365_NUM_BANKS 1
+#define MT8365_NUM_SENSORS 3
+#define MT8365_NUM_SENSORS_PER_ZONE 3
+#define MT8365_TS1 0
+#define MT8365_TS2 1
+#define MT8365_TS3 2
+
 struct mtk_thermal;
 
 struct thermal_bank_cfg {
@@ -386,6 +398,24 @@ static const int mt7622_mux_values[MT7622_NUM_SENSORS] = { 0, };
 static const int mt7622_vts_index[MT7622_NUM_SENSORS] = { VTS1 };
 static const int mt7622_tc_offset[MT7622_NUM_CONTROLLER] = { 0x0, };
 
+/* MT8365 thermal sensor data */
+static const int mt8365_bank_data[MT8365_NUM_SENSORS] = {
+	MT8365_TS1, MT8365_TS2, MT8365_TS3
+};
+
+static const int mt8365_msr[MT8365_NUM_SENSORS_PER_ZONE] = {
+	TEMP_MSR0, TEMP_MSR1, TEMP_MSR2
+};
+
+static const int mt8365_adcpnp[MT8365_NUM_SENSORS_PER_ZONE] = {
+	TEMP_ADCPNP0, TEMP_ADCPNP1, TEMP_ADCPNP2
+};
+
+static const int mt8365_mux_values[MT8365_NUM_SENSORS] = { 0, 1, 2 };
+static const int mt8365_tc_offset[MT8365_NUM_CONTROLLER] = { 0 };
+
+static const int mt8365_vts_index[MT8365_NUM_SENSORS] = { VTS1, VTS2, VTS3 };
+
 /*
  * The MT8173 thermal controller has four banks. Each bank can read up to
  * four temperature sensors simultaneously. The MT8173 has a total of 5
@@ -460,6 +490,40 @@ static const struct mtk_thermal_data mt2701_thermal_data = {
 	.version = MTK_THERMAL_V1,
 };
 
+/*
+ * The MT8365 thermal controller has one bank, which can read up to
+ * four temperature sensors simultaneously. The MT8365 has a total of 3
+ * temperature sensors.
+ *
+ * The thermal core only gets the maximum temperature of this one bank,
+ * so the bank concept wouldn't be necessary here. However, the SVS (Smart
+ * Voltage Scaling) unit makes its decisions based on the same bank
+ * data.
+ */
+static const struct mtk_thermal_data mt8365_thermal_data = {
+	.auxadc_channel = MT8365_TEMP_AUXADC_CHANNEL,
+	.num_banks = MT8365_NUM_BANKS,
+	.num_sensors = MT8365_NUM_SENSORS,
+	.vts_index = mt8365_vts_index,
+	.cali_val = MT8365_CALIBRATION,
+	.num_controller = MT8365_NUM_CONTROLLER,
+	.controller_offset = mt8365_tc_offset,
+	.need_switch_bank = false,
+	.bank_data = {
+		{
+			.num_sensors = MT8365_NUM_SENSORS,
+			.sensors = mt8365_bank_data
+		},
+	},
+	.msr = mt8365_msr,
+	.adcpnp = mt8365_adcpnp,
+	.sensor_mux_values = mt8365_mux_values,
+	.version = MTK_THERMAL_V1,
+	.apmixed_buffer_ctl_reg = APMIXED_SYS_TS_CON0,
+	.apmixed_buffer_ctl_mask = (u32) ~GENMASK(29, 28),
+	.apmixed_buffer_ctl_set = 0,
+};
+
 /*
  * The MT2712 thermal controller has one bank, which can read up to
  * four temperature sensors simultaneously. The MT2712 has a total of 4
@@ -958,6 +1022,10 @@ static const struct of_device_id mtk_thermal_of_match[] = {
 	{
 		.compatible = "mediatek,mt8183-thermal",
 		.data = (void *)&mt8183_thermal_data,
+	},
+	{
+		.compatible = "mediatek,mt8365-thermal",
+		.data = (void *)&mt8365_thermal_data,
 	}, {
 	},
 };
-- 
2.37.1

Re: [PATCH v2 2/4] thermal: mediatek: add support for MT8365 SoC

[AngeloGioacchino Del Regno]

Il 20/07/22 20:18, Amjad Ouled-Ameur ha scritto:
> From: Fabien Parent <fparent@baylibre.com>
> 
> MT8365 is similar to the other SoCs supported by the driver. It has only
> one bank and 3 actual sensors that can be multiplexed. There is another
> one sensor that does not have usable data.
> 
> Signed-off-by: Fabien Parent <fparent@baylibre.com>
> Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> ---
>   drivers/thermal/mtk_thermal.c | 68 +++++++++++++++++++++++++++++++++++
>   1 file changed, 68 insertions(+)
> 
> diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
> index ede94eadddda..216eb0d61c5b 100644
> --- a/drivers/thermal/mtk_thermal.c
> +++ b/drivers/thermal/mtk_thermal.c
> @@ -31,6 +31,7 @@
>   #define AUXADC_CON2_V		0x010
>   #define AUXADC_DATA(channel)	(0x14 + (channel) * 4)
>   
> +#define APMIXED_SYS_TS_CON0	0x600
>   #define APMIXED_SYS_TS_CON1	0x604
>   
>   /* Thermal Controller Registers */
> @@ -245,6 +246,17 @@ enum mtk_thermal_version {
>   /* The calibration coefficient of sensor  */
>   #define MT8183_CALIBRATION	153
>   
> +/* MT8365 */
> +#define MT8365_TEMP_AUXADC_CHANNEL 11
> +#define MT8365_CALIBRATION 164
> +#define MT8365_NUM_CONTROLLER 1
> +#define MT8365_NUM_BANKS 1
> +#define MT8365_NUM_SENSORS 3
> +#define MT8365_NUM_SENSORS_PER_ZONE 3
> +#define MT8365_TS1 0
> +#define MT8365_TS2 1
> +#define MT8365_TS3 2
> +
>   struct mtk_thermal;
>   
>   struct thermal_bank_cfg {
> @@ -386,6 +398,24 @@ static const int mt7622_mux_values[MT7622_NUM_SENSORS] = { 0, };
>   static const int mt7622_vts_index[MT7622_NUM_SENSORS] = { VTS1 };
>   static const int mt7622_tc_offset[MT7622_NUM_CONTROLLER] = { 0x0, };
>   
> +/* MT8365 thermal sensor data */
> +static const int mt8365_bank_data[MT8365_NUM_SENSORS] = {
> +	MT8365_TS1, MT8365_TS2, MT8365_TS3
> +};
> +
> +static const int mt8365_msr[MT8365_NUM_SENSORS_PER_ZONE] = {
> +	TEMP_MSR0, TEMP_MSR1, TEMP_MSR2
> +};
> +
> +static const int mt8365_adcpnp[MT8365_NUM_SENSORS_PER_ZONE] = {
> +	TEMP_ADCPNP0, TEMP_ADCPNP1, TEMP_ADCPNP2
> +};
> +
> +static const int mt8365_mux_values[MT8365_NUM_SENSORS] = { 0, 1, 2 };
> +static const int mt8365_tc_offset[MT8365_NUM_CONTROLLER] = { 0 };
> +
> +static const int mt8365_vts_index[MT8365_NUM_SENSORS] = { VTS1, VTS2, VTS3 };
> +
>   /*
>    * The MT8173 thermal controller has four banks. Each bank can read up to
>    * four temperature sensors simultaneously. The MT8173 has a total of 5
> @@ -460,6 +490,40 @@ static const struct mtk_thermal_data mt2701_thermal_data = {
>   	.version = MTK_THERMAL_V1,
>   };
>   
> +/*
> + * The MT8365 thermal controller has one bank, which can read up to
> + * four temperature sensors simultaneously. The MT8365 has a total of 3
> + * temperature sensors.
> + *
> + * The thermal core only gets the maximum temperature of this one bank,
> + * so the bank concept wouldn't be necessary here. However, the SVS (Smart
> + * Voltage Scaling) unit makes its decisions based on the same bank
> + * data.
> + */
> +static const struct mtk_thermal_data mt8365_thermal_data = {
> +	.auxadc_channel = MT8365_TEMP_AUXADC_CHANNEL,
> +	.num_banks = MT8365_NUM_BANKS,
> +	.num_sensors = MT8365_NUM_SENSORS,
> +	.vts_index = mt8365_vts_index,
> +	.cali_val = MT8365_CALIBRATION,
> +	.num_controller = MT8365_NUM_CONTROLLER,
> +	.controller_offset = mt8365_tc_offset,
> +	.need_switch_bank = false,
> +	.bank_data = {
> +		{
> +			.num_sensors = MT8365_NUM_SENSORS,
> +			.sensors = mt8365_bank_data
> +		},
> +	},
> +	.msr = mt8365_msr,
> +	.adcpnp = mt8365_adcpnp,
> +	.sensor_mux_values = mt8365_mux_values,
> +	.version = MTK_THERMAL_V1,
> +	.apmixed_buffer_ctl_reg = APMIXED_SYS_TS_CON0,
> +	.apmixed_buffer_ctl_mask = (u32) ~GENMASK(29, 28),
> +	.apmixed_buffer_ctl_set = 0,

Every commit has to build on its own.

This won't compile unless moved after 'thermal: mediatek: control buffer enablement
tweaks': please move it.

Regards,
Angelo

Re: [PATCH v2 2/4] thermal: mediatek: add support for MT8365 SoC

[Krzysztof Kozlowski]

On 20/07/2022 20:18, Amjad Ouled-Ameur wrote:
> From: Fabien Parent <fparent@baylibre.com>
> 
> MT8365 is similar to the other SoCs supported by the driver. It has only
> one bank and 3 actual sensors that can be multiplexed. There is another
> one sensor that does not have usable data.
> 
> Signed-off-by: Fabien Parent <fparent@baylibre.com>
> Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>

Just like in all other cases - you miss your SoB which is necessary for
the chain.


Best regards,
Krzysztof

Re: [PATCH v2 2/4] thermal: mediatek: add support for MT8365 SoC

[Amjad Ouled-Ameur]

Hi Krzysztof,

On 7/21/22 18:37, Krzysztof Kozlowski wrote:
> On 20/07/2022 20:18, Amjad Ouled-Ameur wrote:
>> From: Fabien Parent <fparent@baylibre.com>
>>
>> MT8365 is similar to the other SoCs supported by the driver. It has only
>> one bank and 3 actual sensors that can be multiplexed. There is another
>> one sensor that does not have usable data.
>>
>> Signed-off-by: Fabien Parent <fparent@baylibre.com>
>> Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
>> Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> Just like in all other cases - you miss your SoB which is necessary for
> the chain.

Right, will add my SoB in V3, thank you!

Regards,

Amjad

>
> Best regards,
> Krzysztof

[PATCH v2 3/4] thermal: mediatek: control buffer enablement tweaks

[Amjad Ouled-Ameur]

From: Markus Schneider-Pargmann <msp@baylibre.com>

Add logic in order to be able to turn on the control buffer on MT8365.
This change now allows to have control buffer support for MTK_THERMAL_V1,
and it allows to define the register offset, and mask used to enable it.

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Signed-off-by: Fabien Parent <fparent@baylibre.com>
Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
---
 drivers/thermal/mtk_thermal.c | 23 +++++++++++++++++------
 1 file changed, 17 insertions(+), 6 deletions(-)

diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
index 216eb0d61c5b..1dc276f8c4f1 100644
--- a/drivers/thermal/mtk_thermal.c
+++ b/drivers/thermal/mtk_thermal.c
@@ -283,6 +283,9 @@ struct mtk_thermal_data {
 	bool need_switch_bank;
 	struct thermal_bank_cfg bank_data[MAX_NUM_ZONES];
 	enum mtk_thermal_version version;
+	u32 apmixed_buffer_ctl_reg;
+	u32 apmixed_buffer_ctl_mask;
+	u32 apmixed_buffer_ctl_set;
 };
 
 struct mtk_thermal {
@@ -578,6 +581,9 @@ static const struct mtk_thermal_data mt7622_thermal_data = {
 	.adcpnp = mt7622_adcpnp,
 	.sensor_mux_values = mt7622_mux_values,
 	.version = MTK_THERMAL_V2,
+	.apmixed_buffer_ctl_reg = APMIXED_SYS_TS_CON1,
+	.apmixed_buffer_ctl_mask = ~0x37,
+	.apmixed_buffer_ctl_set = 0x1,
 };
 
 /*
@@ -1031,14 +1037,18 @@ static const struct of_device_id mtk_thermal_of_match[] = {
 };
 MODULE_DEVICE_TABLE(of, mtk_thermal_of_match);
 
-static void mtk_thermal_turn_on_buffer(void __iomem *apmixed_base)
+static void mtk_thermal_turn_on_buffer(struct mtk_thermal *mt,
+				       void __iomem *apmixed_base)
 {
 	int tmp;
 
-	tmp = readl(apmixed_base + APMIXED_SYS_TS_CON1);
-	tmp &= ~(0x37);
-	tmp |= 0x1;
-	writel(tmp, apmixed_base + APMIXED_SYS_TS_CON1);
+	if (!mt->conf->apmixed_buffer_ctl_reg)
+		return;
+
+	tmp = readl(apmixed_base + mt->conf->apmixed_buffer_ctl_reg);
+	tmp &= mt->conf->apmixed_buffer_ctl_mask;
+	tmp |= mt->conf->apmixed_buffer_ctl_set;
+	writel(tmp, apmixed_base + mt->conf->apmixed_buffer_ctl_reg);
 	udelay(200);
 }
 
@@ -1138,8 +1148,9 @@ static int mtk_thermal_probe(struct platform_device *pdev)
 		goto err_disable_clk_auxadc;
 	}
 
+	mtk_thermal_turn_on_buffer(mt, apmixed_base);
+
 	if (mt->conf->version == MTK_THERMAL_V2) {
-		mtk_thermal_turn_on_buffer(apmixed_base);
 		mtk_thermal_release_periodic_ts(mt, auxadc_base);
 	}
 
-- 
2.37.1

Re: [PATCH v2 3/4] thermal: mediatek: control buffer enablement tweaks

[AngeloGioacchino Del Regno]

Il 20/07/22 20:18, Amjad Ouled-Ameur ha scritto:
> From: Markus Schneider-Pargmann <msp@baylibre.com>
> 
> Add logic in order to be able to turn on the control buffer on MT8365.
> This change now allows to have control buffer support for MTK_THERMAL_V1,
> and it allows to define the register offset, and mask used to enable it.
> 
> Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
> Signed-off-by: Fabien Parent <fparent@baylibre.com>
> Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> ---
>   drivers/thermal/mtk_thermal.c | 23 +++++++++++++++++------
>   1 file changed, 17 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
> index 216eb0d61c5b..1dc276f8c4f1 100644
> --- a/drivers/thermal/mtk_thermal.c
> +++ b/drivers/thermal/mtk_thermal.c
> @@ -283,6 +283,9 @@ struct mtk_thermal_data {
>   	bool need_switch_bank;
>   	struct thermal_bank_cfg bank_data[MAX_NUM_ZONES];
>   	enum mtk_thermal_version version;
> +	u32 apmixed_buffer_ctl_reg;
> +	u32 apmixed_buffer_ctl_mask;
> +	u32 apmixed_buffer_ctl_set;
>   };
>   
>   struct mtk_thermal {
> @@ -578,6 +581,9 @@ static const struct mtk_thermal_data mt7622_thermal_data = {
>   	.adcpnp = mt7622_adcpnp,
>   	.sensor_mux_values = mt7622_mux_values,
>   	.version = MTK_THERMAL_V2,
> +	.apmixed_buffer_ctl_reg = APMIXED_SYS_TS_CON1,
> +	.apmixed_buffer_ctl_mask = ~0x37,

This is effectively GENMASK(31, 6) | BIT(3) :-)

> +	.apmixed_buffer_ctl_set = 0x1,

...and this is BIT(0)

>   }; >
>   /*
> @@ -1031,14 +1037,18 @@ static const struct of_device_id mtk_thermal_of_match[] = {
>   };
>   MODULE_DEVICE_TABLE(of, mtk_thermal_of_match);
>   
> -static void mtk_thermal_turn_on_buffer(void __iomem *apmixed_base)
> +static void mtk_thermal_turn_on_buffer(struct mtk_thermal *mt,
> +				       void __iomem *apmixed_base)
>   {
>   	int tmp;

Since you're practically restructuring this function, please fix this wrong
variable type, as this should be a u32.

Regards,
Angelo

Re: [PATCH v2 3/4] thermal: mediatek: control buffer enablement tweaks

[Amjad Ouled-Ameur]

Hi Angelo,

On 7/21/22 10:11, AngeloGioacchino Del Regno wrote:
> Il 20/07/22 20:18, Amjad Ouled-Ameur ha scritto:
>> From: Markus Schneider-Pargmann <msp@baylibre.com>
>>
>> Add logic in order to be able to turn on the control buffer on MT8365.
>> This change now allows to have control buffer support for 
>> MTK_THERMAL_V1,
>> and it allows to define the register offset, and mask used to enable it.
>>
>> Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
>> Signed-off-by: Fabien Parent <fparent@baylibre.com>
>> Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
>> Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
>> ---
>>   drivers/thermal/mtk_thermal.c | 23 +++++++++++++++++------
>>   1 file changed, 17 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/thermal/mtk_thermal.c 
>> b/drivers/thermal/mtk_thermal.c
>> index 216eb0d61c5b..1dc276f8c4f1 100644
>> --- a/drivers/thermal/mtk_thermal.c
>> +++ b/drivers/thermal/mtk_thermal.c
>> @@ -283,6 +283,9 @@ struct mtk_thermal_data {
>>       bool need_switch_bank;
>>       struct thermal_bank_cfg bank_data[MAX_NUM_ZONES];
>>       enum mtk_thermal_version version;
>> +    u32 apmixed_buffer_ctl_reg;
>> +    u32 apmixed_buffer_ctl_mask;
>> +    u32 apmixed_buffer_ctl_set;
>>   };
>>     struct mtk_thermal {
>> @@ -578,6 +581,9 @@ static const struct mtk_thermal_data 
>> mt7622_thermal_data = {
>>       .adcpnp = mt7622_adcpnp,
>>       .sensor_mux_values = mt7622_mux_values,
>>       .version = MTK_THERMAL_V2,
>> +    .apmixed_buffer_ctl_reg = APMIXED_SYS_TS_CON1,
>> +    .apmixed_buffer_ctl_mask = ~0x37,
>
> This is effectively GENMASK(31, 6) | BIT(3) :-)
>
>> +    .apmixed_buffer_ctl_set = 0x1,
>
> ...and this is BIT(0)
>
Thank you for the catch, will update them appropriately.
>>   }; >
>>   /*
>> @@ -1031,14 +1037,18 @@ static const struct of_device_id 
>> mtk_thermal_of_match[] = {
>>   };
>>   MODULE_DEVICE_TABLE(of, mtk_thermal_of_match);
>>   -static void mtk_thermal_turn_on_buffer(void __iomem *apmixed_base)
>> +static void mtk_thermal_turn_on_buffer(struct mtk_thermal *mt,
>> +                       void __iomem *apmixed_base)
>>   {
>>       int tmp;
>
> Since you're practically restructuring this function, please fix this 
> wrong
> variable type, as this should be a u32.
>
Sure agreed.


Thank you Angelo for the review.

Regards,

Amjad

> Regards,
> Angelo

Re: [PATCH v2 3/4] thermal: mediatek: control buffer enablement tweaks

[Matthias Brugger]

On 20/07/2022 20:18, Amjad Ouled-Ameur wrote:
> From: Markus Schneider-Pargmann <msp@baylibre.com>
> 
> Add logic in order to be able to turn on the control buffer on MT8365.
> This change now allows to have control buffer support for MTK_THERMAL_V1,
> and it allows to define the register offset, and mask used to enable it.
> 
> Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
> Signed-off-by: Fabien Parent <fparent@baylibre.com>
> Reviewed-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
> Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>

Here we will need your Signed-off-by as you are sending the patch :)

Regards,
Matthias

> ---
>   drivers/thermal/mtk_thermal.c | 23 +++++++++++++++++------
>   1 file changed, 17 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
> index 216eb0d61c5b..1dc276f8c4f1 100644
> --- a/drivers/thermal/mtk_thermal.c
> +++ b/drivers/thermal/mtk_thermal.c
> @@ -283,6 +283,9 @@ struct mtk_thermal_data {
>   	bool need_switch_bank;
>   	struct thermal_bank_cfg bank_data[MAX_NUM_ZONES];
>   	enum mtk_thermal_version version;
> +	u32 apmixed_buffer_ctl_reg;
> +	u32 apmixed_buffer_ctl_mask;
> +	u32 apmixed_buffer_ctl_set;
>   };
>   
>   struct mtk_thermal {
> @@ -578,6 +581,9 @@ static const struct mtk_thermal_data mt7622_thermal_data = {
>   	.adcpnp = mt7622_adcpnp,
>   	.sensor_mux_values = mt7622_mux_values,
>   	.version = MTK_THERMAL_V2,
> +	.apmixed_buffer_ctl_reg = APMIXED_SYS_TS_CON1,
> +	.apmixed_buffer_ctl_mask = ~0x37,
> +	.apmixed_buffer_ctl_set = 0x1,
>   };
>   
>   /*
> @@ -1031,14 +1037,18 @@ static const struct of_device_id mtk_thermal_of_match[] = {
>   };
>   MODULE_DEVICE_TABLE(of, mtk_thermal_of_match);
>   
> -static void mtk_thermal_turn_on_buffer(void __iomem *apmixed_base)
> +static void mtk_thermal_turn_on_buffer(struct mtk_thermal *mt,
> +				       void __iomem *apmixed_base)
>   {
>   	int tmp;
>   
> -	tmp = readl(apmixed_base + APMIXED_SYS_TS_CON1);
> -	tmp &= ~(0x37);
> -	tmp |= 0x1;
> -	writel(tmp, apmixed_base + APMIXED_SYS_TS_CON1);
> +	if (!mt->conf->apmixed_buffer_ctl_reg)
> +		return;
> +
> +	tmp = readl(apmixed_base + mt->conf->apmixed_buffer_ctl_reg);
> +	tmp &= mt->conf->apmixed_buffer_ctl_mask;
> +	tmp |= mt->conf->apmixed_buffer_ctl_set;
> +	writel(tmp, apmixed_base + mt->conf->apmixed_buffer_ctl_reg);
>   	udelay(200);
>   }
>   
> @@ -1138,8 +1148,9 @@ static int mtk_thermal_probe(struct platform_device *pdev)
>   		goto err_disable_clk_auxadc;
>   	}
>   
> +	mtk_thermal_turn_on_buffer(mt, apmixed_base);
> +
>   	if (mt->conf->version == MTK_THERMAL_V2) {
> -		mtk_thermal_turn_on_buffer(apmixed_base);
>   		mtk_thermal_release_periodic_ts(mt, auxadc_base);
>   	}
>   

[PATCH v2 4/4] thermal: mediatek: add another get_temp ops for thermal sensors

[Amjad Ouled-Ameur]

Provide thermal zone to read thermal sensor in the SoC. We can read all the
thermal sensors value in the SoC by the node /sys/class/thermal/

In mtk_thermal_bank_temperature, return -EAGAIN instead of -EACCESS
on the first read of sensor that often are bogus values.
This can avoid following warning on boot:

  thermal thermal_zone6: failed to read out thermal zone (-13)

Signed-off-by: default avatarMichael Kao <michael.kao@mediatek.com>
Signed-off-by: default avatarHsin-Yi Wang <hsinyi@chromium.org>
Signed-off-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
Tested-by: Amjad Ouled-Ameur <aouledameur@baylibre.com>
---
 drivers/thermal/mtk_thermal.c | 100 ++++++++++++++++++++++++++--------
 1 file changed, 76 insertions(+), 24 deletions(-)

diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
index 1dc276f8c4f1..79b14ce1a08d 100644
--- a/drivers/thermal/mtk_thermal.c
+++ b/drivers/thermal/mtk_thermal.c
@@ -259,6 +259,11 @@ enum mtk_thermal_version {
 
 struct mtk_thermal;
 
+struct mtk_thermal_zone {
+	struct mtk_thermal *mt;
+	int id;
+};
+
 struct thermal_bank_cfg {
 	unsigned int num_sensors;
 	const int *sensors;
@@ -709,6 +714,32 @@ static void mtk_thermal_put_bank(struct mtk_thermal_bank *bank)
 		mutex_unlock(&mt->lock);
 }
 
+static u32 _get_sensor_temp(struct mtk_thermal *mt, int id)
+{
+	u32 raw;
+	int temp;
+
+	const struct mtk_thermal_data *conf = mt->conf;
+
+	raw = readl(mt->thermal_base + conf->msr[id]);
+
+	if (mt->conf->version == MTK_THERMAL_V1)
+		temp = raw_to_mcelsius_v1(mt, id, raw);
+	else
+		temp = raw_to_mcelsius_v2(mt, id, raw);
+
+	/*
+	 * The first read of a sensor often contains very high bogus
+	 * temperature value. Filter these out so that the system does
+	 * not immediately shut down.
+	 */
+
+	if (temp > 200000)
+		return  -EAGAIN;
+	else
+		return	temp;
+}
+
 /**
  * mtk_thermal_bank_temperature - get the temperature of a bank
  * @bank:	The bank
@@ -721,26 +752,9 @@ static int mtk_thermal_bank_temperature(struct mtk_thermal_bank *bank)
 	struct mtk_thermal *mt = bank->mt;
 	const struct mtk_thermal_data *conf = mt->conf;
 	int i, temp = INT_MIN, max = INT_MIN;
-	u32 raw;
 
 	for (i = 0; i < conf->bank_data[bank->id].num_sensors; i++) {
-		raw = readl(mt->thermal_base + conf->msr[i]);
-
-		if (mt->conf->version == MTK_THERMAL_V1) {
-			temp = raw_to_mcelsius_v1(
-				mt, conf->bank_data[bank->id].sensors[i], raw);
-		} else {
-			temp = raw_to_mcelsius_v2(
-				mt, conf->bank_data[bank->id].sensors[i], raw);
-		}
-
-		/*
-		 * The first read of a sensor often contains very high bogus
-		 * temperature value. Filter these out so that the system does
-		 * not immediately shut down.
-		 */
-		if (temp > 200000)
-			temp = 0;
+		temp = _get_sensor_temp(mt, i);
 
 		if (temp > max)
 			max = temp;
@@ -751,7 +765,8 @@ static int mtk_thermal_bank_temperature(struct mtk_thermal_bank *bank)
 
 static int mtk_read_temp(void *data, int *temperature)
 {
-	struct mtk_thermal *mt = data;
+	struct mtk_thermal_zone *tz = data;
+	struct mtk_thermal *mt = tz->mt;
 	int i;
 	int tempmax = INT_MIN;
 
@@ -770,10 +785,28 @@ static int mtk_read_temp(void *data, int *temperature)
 	return 0;
 }
 
+static int mtk_read_sensor_temp(void *data, int *temperature)
+{
+	struct mtk_thermal_zone *tz = data;
+	struct mtk_thermal *mt = tz->mt;
+	int id = tz->id - 1;
+
+	if (id < 0)
+		return  -EACCES;
+
+	*temperature = _get_sensor_temp(mt, id);
+
+	return 0;
+}
+
 static const struct thermal_zone_of_device_ops mtk_thermal_ops = {
 	.get_temp = mtk_read_temp,
 };
 
+static const struct thermal_zone_of_device_ops mtk_thermal_sensor_ops = {
+	.get_temp = mtk_read_sensor_temp,
+};
+
 static void mtk_thermal_init_bank(struct mtk_thermal *mt, int num,
 				  u32 apmixed_phys_base, u32 auxadc_phys_base,
 				  int ctrl_id)
@@ -1072,6 +1105,7 @@ static int mtk_thermal_probe(struct platform_device *pdev)
 	u64 auxadc_phys_base, apmixed_phys_base;
 	struct thermal_zone_device *tzdev;
 	void __iomem *apmixed_base, *auxadc_base;
+	struct mtk_thermal_zone *tz;
 
 	mt = devm_kzalloc(&pdev->dev, sizeof(*mt), GFP_KERNEL);
 	if (!mt)
@@ -1161,11 +1195,29 @@ static int mtk_thermal_probe(struct platform_device *pdev)
 
 	platform_set_drvdata(pdev, mt);
 
-	tzdev = devm_thermal_zone_of_sensor_register(&pdev->dev, 0, mt,
-						     &mtk_thermal_ops);
-	if (IS_ERR(tzdev)) {
-		ret = PTR_ERR(tzdev);
-		goto err_disable_clk_peri_therm;
+	for (i = 0; i < mt->conf->num_sensors + 1; i++) {
+		tz = kmalloc(sizeof(*tz), GFP_KERNEL);
+		if (!tz)
+			return -ENOMEM;
+
+		tz->mt = mt;
+		tz->id = i;
+
+		tzdev = devm_thermal_zone_of_sensor_register(&pdev->dev, i, tz, (i == 0) ?
+							     &mtk_thermal_ops :
+							     &mtk_thermal_sensor_ops);
+
+		if (IS_ERR(tzdev)) {
+			if (PTR_ERR(tzdev) == -ENODEV) {
+				dev_warn(&pdev->dev,
+					 "sensor %d not registered in thermal zone in dt\n", i);
+				continue;
+			}
+			if (PTR_ERR(tzdev) == -EACCES) {
+				ret = PTR_ERR(tzdev);
+				goto err_disable_clk_peri_therm;
+			}
+		}
 	}
 
 	ret = devm_thermal_add_hwmon_sysfs(tzdev);
-- 
2.37.1

Re: [PATCH v2 4/4] thermal: mediatek: add another get_temp ops for thermal sensors

[Christophe JAILLET]

Le 20/07/2022 à 20:18, Amjad Ouled-Ameur a écrit :
> Provide thermal zone to read thermal sensor in the SoC. We can read all the
> thermal sensors value in the SoC by the node /sys/class/thermal/
> 
> In mtk_thermal_bank_temperature, return -EAGAIN instead of -EACCESS
> on the first read of sensor that often are bogus values.
> This can avoid following warning on boot:
> 
>    thermal thermal_zone6: failed to read out thermal zone (-13)
> 
> Signed-off-by: default avatarMichael Kao <michael.kao-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
> Signed-off-by: default avatarHsin-Yi Wang <hsinyi-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
> Signed-off-by: Amjad Ouled-Ameur <aouledameur-rdvid1DuHRBWk0Htik3J/w@public.gmane.org>
> Tested-by: Amjad Ouled-Ameur <aouledameur-rdvid1DuHRBWk0Htik3J/w@public.gmane.org>
> ---
>   drivers/thermal/mtk_thermal.c | 100 ++++++++++++++++++++++++++--------
>   1 file changed, 76 insertions(+), 24 deletions(-)
> 
> diff --git a/drivers/thermal/mtk_thermal.c b/drivers/thermal/mtk_thermal.c
> index 1dc276f8c4f1..79b14ce1a08d 100644
> --- a/drivers/thermal/mtk_thermal.c
> +++ b/drivers/thermal/mtk_thermal.c
> @@ -259,6 +259,11 @@ enum mtk_thermal_version {
>   
>   struct mtk_thermal;
>   
> +struct mtk_thermal_zone {
> +	struct mtk_thermal *mt;
> +	int id;
> +};
> +
>   struct thermal_bank_cfg {
>   	unsigned int num_sensors;
>   	const int *sensors;
> @@ -709,6 +714,32 @@ static void mtk_thermal_put_bank(struct mtk_thermal_bank *bank)
>   		mutex_unlock(&mt->lock);
>   }
>   
> +static u32 _get_sensor_temp(struct mtk_thermal *mt, int id)
> +{
> +	u32 raw;
> +	int temp;
> +
> +	const struct mtk_thermal_data *conf = mt->conf;
> +
> +	raw = readl(mt->thermal_base + conf->msr[id]);
> +
> +	if (mt->conf->version == MTK_THERMAL_V1)
> +		temp = raw_to_mcelsius_v1(mt, id, raw);
> +	else
> +		temp = raw_to_mcelsius_v2(mt, id, raw);
> +
> +	/*
> +	 * The first read of a sensor often contains very high bogus
> +	 * temperature value. Filter these out so that the system does
> +	 * not immediately shut down.
> +	 */
> +
> +	if (temp > 200000)
> +		return  -EAGAIN;

This function returns a u32. Is it ok to return -EAGAIN?

There is also 2 spaces here...

> +	else
> +		return	temp;

... and a tab here.

> +}
> +
>   /**
>    * mtk_thermal_bank_temperature - get the temperature of a bank
>    * @bank:	The bank
> @@ -721,26 +752,9 @@ static int mtk_thermal_bank_temperature(struct mtk_thermal_bank *bank)
>   	struct mtk_thermal *mt = bank->mt;
>   	const struct mtk_thermal_data *conf = mt->conf;
>   	int i, temp = INT_MIN, max = INT_MIN;
> -	u32 raw;
>   
>   	for (i = 0; i < conf->bank_data[bank->id].num_sensors; i++) {
> -		raw = readl(mt->thermal_base + conf->msr[i]);
> -
> -		if (mt->conf->version == MTK_THERMAL_V1) {
> -			temp = raw_to_mcelsius_v1(
> -				mt, conf->bank_data[bank->id].sensors[i], raw);
> -		} else {
> -			temp = raw_to_mcelsius_v2(
> -				mt, conf->bank_data[bank->id].sensors[i], raw);
> -		}
> -
> -		/*
> -		 * The first read of a sensor often contains very high bogus
> -		 * temperature value. Filter these out so that the system does
> -		 * not immediately shut down.
> -		 */
> -		if (temp > 200000)
> -			temp = 0;
> +		temp = _get_sensor_temp(mt, i);

Is it ok if _get_sensor_temp() returns -EAGAIN?

>   
>   		if (temp > max)
>   			max = temp;
> @@ -751,7 +765,8 @@ static int mtk_thermal_bank_temperature(struct mtk_thermal_bank *bank)
>   
>   static int mtk_read_temp(void *data, int *temperature)
>   {
> -	struct mtk_thermal *mt = data;
> +	struct mtk_thermal_zone *tz = data;
> +	struct mtk_thermal *mt = tz->mt;
>   	int i;
>   	int tempmax = INT_MIN;
>   
> @@ -770,10 +785,28 @@ static int mtk_read_temp(void *data, int *temperature)
>   	return 0;
>   }
>   
> +static int mtk_read_sensor_temp(void *data, int *temperature)
> +{
> +	struct mtk_thermal_zone *tz = data;
> +	struct mtk_thermal *mt = tz->mt;
> +	int id = tz->id - 1;
> +
> +	if (id < 0)
> +		return  -EACCES;

2 spaces.

> +
> +	*temperature = _get_sensor_temp(mt, id);

If _get_sensor_temp() returns -EAGAIN, should this be propagated to the 
caller?

> +
> +	return 0;
> +}
> +
>   static const struct thermal_zone_of_device_ops mtk_thermal_ops = {
>   	.get_temp = mtk_read_temp,
>   };
>   
> +static const struct thermal_zone_of_device_ops mtk_thermal_sensor_ops = {
> +	.get_temp = mtk_read_sensor_temp,
> +};
> +
>   static void mtk_thermal_init_bank(struct mtk_thermal *mt, int num,
>   				  u32 apmixed_phys_base, u32 auxadc_phys_base,
>   				  int ctrl_id)
> @@ -1072,6 +1105,7 @@ static int mtk_thermal_probe(struct platform_device *pdev)
>   	u64 auxadc_phys_base, apmixed_phys_base;
>   	struct thermal_zone_device *tzdev;
>   	void __iomem *apmixed_base, *auxadc_base;
> +	struct mtk_thermal_zone *tz;
>   
>   	mt = devm_kzalloc(&pdev->dev, sizeof(*mt), GFP_KERNEL);
>   	if (!mt)
> @@ -1161,11 +1195,29 @@ static int mtk_thermal_probe(struct platform_device *pdev)
>   
>   	platform_set_drvdata(pdev, mt);
>   
> -	tzdev = devm_thermal_zone_of_sensor_register(&pdev->dev, 0, mt,
> -						     &mtk_thermal_ops);
> -	if (IS_ERR(tzdev)) {
> -		ret = PTR_ERR(tzdev);
> -		goto err_disable_clk_peri_therm;
> +	for (i = 0; i < mt->conf->num_sensors + 1; i++) {
> +		tz = kmalloc(sizeof(*tz), GFP_KERNEL);

Should this memory allocation be a devm_kmalloc(), or is this memory 
freed at some point by the framework?

(I don't know the thermal_zone API and the patch has no kfree())

CJ

> +		if (!tz)
> +			return -ENOMEM;
> +
> +		tz->mt = mt;
> +		tz->id = i;
> +
> +		tzdev = devm_thermal_zone_of_sensor_register(&pdev->dev, i, tz, (i == 0) ?
> +							     &mtk_thermal_ops :
> +							     &mtk_thermal_sensor_ops);
> +
> +		if (IS_ERR(tzdev)) {
> +			if (PTR_ERR(tzdev) == -ENODEV) {
> +				dev_warn(&pdev->dev,
> +					 "sensor %d not registered in thermal zone in dt\n", i);
> +				continue;
> +			}
> +			if (PTR_ERR(tzdev) == -EACCES) {
> +				ret = PTR_ERR(tzdev);
> +				goto err_disable_clk_peri_therm;
> +			}
> +		}
>   	}
>   
>   	ret = devm_thermal_add_hwmon_sysfs(tzdev);

Re: [PATCH v2 4/4] thermal: mediatek: add another get_temp ops for thermal sensors

[Amjad Ouled-Ameur]

Hi Christophe,

On 7/20/22 20:54, Christophe JAILLET wrote:
> Le 20/07/2022 à 20:18, Amjad Ouled-Ameur a écrit :
>> Provide thermal zone to read thermal sensor in the SoC. We can read 
>> all the
>> thermal sensors value in the SoC by the node /sys/class/thermal/
>>
>> In mtk_thermal_bank_temperature, return -EAGAIN instead of -EACCESS
>> on the first read of sensor that often are bogus values.
>> This can avoid following warning on boot:
>>
>>    thermal thermal_zone6: failed to read out thermal zone (-13)
>>
>> Signed-off-by: default avatarMichael Kao 
>> <michael.kao-NuS5LvNUpcJWk0Htik3J/w@public.gmane.org>
>> Signed-off-by: default avatarHsin-Yi Wang 
>> <hsinyi-F7+t8E8rja9g9hUCZPvPmw@public.gmane.org>
>> Signed-off-by: Amjad Ouled-Ameur 
>> <aouledameur-rdvid1DuHRBWk0Htik3J/w@public.gmane.org>
>> Tested-by: Amjad Ouled-Ameur 
>> <aouledameur-rdvid1DuHRBWk0Htik3J/w@public.gmane.org>
>> ---
>>   drivers/thermal/mtk_thermal.c | 100 ++++++++++++++++++++++++++--------
>>   1 file changed, 76 insertions(+), 24 deletions(-)
>>
>> diff --git a/drivers/thermal/mtk_thermal.c 
>> b/drivers/thermal/mtk_thermal.c
>> index 1dc276f8c4f1..79b14ce1a08d 100644
>> --- a/drivers/thermal/mtk_thermal.c
>> +++ b/drivers/thermal/mtk_thermal.c
>> @@ -259,6 +259,11 @@ enum mtk_thermal_version {
>>     struct mtk_thermal;
>>   +struct mtk_thermal_zone {
>> +    struct mtk_thermal *mt;
>> +    int id;
>> +};
>> +
>>   struct thermal_bank_cfg {
>>       unsigned int num_sensors;
>>       const int *sensors;
>> @@ -709,6 +714,32 @@ static void mtk_thermal_put_bank(struct 
>> mtk_thermal_bank *bank)
>>           mutex_unlock(&mt->lock);
>>   }
>>   +static u32 _get_sensor_temp(struct mtk_thermal *mt, int id)
>> +{
>> +    u32 raw;
>> +    int temp;
>> +
>> +    const struct mtk_thermal_data *conf = mt->conf;
>> +
>> +    raw = readl(mt->thermal_base + conf->msr[id]);
>> +
>> +    if (mt->conf->version == MTK_THERMAL_V1)
>> +        temp = raw_to_mcelsius_v1(mt, id, raw);
>> +    else
>> +        temp = raw_to_mcelsius_v2(mt, id, raw);
>> +
>> +    /*
>> +     * The first read of a sensor often contains very high bogus
>> +     * temperature value. Filter these out so that the system does
>> +     * not immediately shut down.
>> +     */
>> +
>> +    if (temp > 200000)
>> +        return  -EAGAIN;
>
> This function returns a u32. Is it ok to return -EAGAIN?
>
_get_sensor_temp() should normally return int instead u32, will fix it 
in V3.
> There is also 2 spaces here...
>
>> +    else
>> +        return    temp;
>
> ... and a tab here.
>
will fix them in V3.
>> +}
>> +
>>   /**
>>    * mtk_thermal_bank_temperature - get the temperature of a bank
>>    * @bank:    The bank
>> @@ -721,26 +752,9 @@ static int mtk_thermal_bank_temperature(struct 
>> mtk_thermal_bank *bank)
>>       struct mtk_thermal *mt = bank->mt;
>>       const struct mtk_thermal_data *conf = mt->conf;
>>       int i, temp = INT_MIN, max = INT_MIN;
>> -    u32 raw;
>>         for (i = 0; i < conf->bank_data[bank->id].num_sensors; i++) {
>> -        raw = readl(mt->thermal_base + conf->msr[i]);
>> -
>> -        if (mt->conf->version == MTK_THERMAL_V1) {
>> -            temp = raw_to_mcelsius_v1(
>> -                mt, conf->bank_data[bank->id].sensors[i], raw);
>> -        } else {
>> -            temp = raw_to_mcelsius_v2(
>> -                mt, conf->bank_data[bank->id].sensors[i], raw);
>> -        }
>> -
>> -        /*
>> -         * The first read of a sensor often contains very high bogus
>> -         * temperature value. Filter these out so that the system does
>> -         * not immediately shut down.
>> -         */
>> -        if (temp > 200000)
>> -            temp = 0;
>> +        temp = _get_sensor_temp(mt, i);
>
> Is it ok if _get_sensor_temp() returns -EAGAIN?
>
drivers/thermal/thermal_core.c:update_temperature() checks for -EAGAIN 
after

thermal_zone_get_temp() is called, thus, I think it's good to return 
-EAGAIN in

case of failure.

>>             if (temp > max)
>>               max = temp;
>> @@ -751,7 +765,8 @@ static int mtk_thermal_bank_temperature(struct 
>> mtk_thermal_bank *bank)
>>     static int mtk_read_temp(void *data, int *temperature)
>>   {
>> -    struct mtk_thermal *mt = data;
>> +    struct mtk_thermal_zone *tz = data;
>> +    struct mtk_thermal *mt = tz->mt;
>>       int i;
>>       int tempmax = INT_MIN;
>>   @@ -770,10 +785,28 @@ static int mtk_read_temp(void *data, int 
>> *temperature)
>>       return 0;
>>   }
>>   +static int mtk_read_sensor_temp(void *data, int *temperature)
>> +{
>> +    struct mtk_thermal_zone *tz = data;
>> +    struct mtk_thermal *mt = tz->mt;
>> +    int id = tz->id - 1;
>> +
>> +    if (id < 0)
>> +        return  -EACCES;
>
> 2 spaces.
>

will fix it in V3.
>> +
>> +    *temperature = _get_sensor_temp(mt, id);
>
> If _get_sensor_temp() returns -EAGAIN, should this be propagated to 
> the caller?
>
>> +
>> +    return 0;
>> +}
>> +
>>   static const struct thermal_zone_of_device_ops mtk_thermal_ops = {
>>       .get_temp = mtk_read_temp,
>>   };
>>   +static const struct thermal_zone_of_device_ops 
>> mtk_thermal_sensor_ops = {
>> +    .get_temp = mtk_read_sensor_temp,
>> +};
>> +
>>   static void mtk_thermal_init_bank(struct mtk_thermal *mt, int num,
>>                     u32 apmixed_phys_base, u32 auxadc_phys_base,
>>                     int ctrl_id)
>> @@ -1072,6 +1105,7 @@ static int mtk_thermal_probe(struct 
>> platform_device *pdev)
>>       u64 auxadc_phys_base, apmixed_phys_base;
>>       struct thermal_zone_device *tzdev;
>>       void __iomem *apmixed_base, *auxadc_base;
>> +    struct mtk_thermal_zone *tz;
>>         mt = devm_kzalloc(&pdev->dev, sizeof(*mt), GFP_KERNEL);
>>       if (!mt)
>> @@ -1161,11 +1195,29 @@ static int mtk_thermal_probe(struct 
>> platform_device *pdev)
>>         platform_set_drvdata(pdev, mt);
>>   -    tzdev = devm_thermal_zone_of_sensor_register(&pdev->dev, 0, mt,
>> -                             &mtk_thermal_ops);
>> -    if (IS_ERR(tzdev)) {
>> -        ret = PTR_ERR(tzdev);
>> -        goto err_disable_clk_peri_therm;
>> +    for (i = 0; i < mt->conf->num_sensors + 1; i++) {
>> +        tz = kmalloc(sizeof(*tz), GFP_KERNEL);
>
> Should this memory allocation be a devm_kmalloc(), or is this memory 
> freed at some point by the framework?
>
> (I don't know the thermal_zone API and the patch has no kfree())
AFAIK, thermal API does not free private data, therefore devm_kmalloc() 
should be used.
>
> CJ
>
>> +        if (!tz)
>> +            return -ENOMEM;
>> +
>> +        tz->mt = mt;
>> +        tz->id = i;
>> +
>> +        tzdev = devm_thermal_zone_of_sensor_register(&pdev->dev, i, 
>> tz, (i == 0) ?
>> +                                 &mtk_thermal_ops :
>> +                                 &mtk_thermal_sensor_ops);
>> +
>> +        if (IS_ERR(tzdev)) {
>> +            if (PTR_ERR(tzdev) == -ENODEV) {
>> +                dev_warn(&pdev->dev,
>> +                     "sensor %d not registered in thermal zone in 
>> dt\n", i);
>> +                continue;
>> +            }
>> +            if (PTR_ERR(tzdev) == -EACCES) {
>> +                ret = PTR_ERR(tzdev);
>> +                goto err_disable_clk_peri_therm;
>> +            }
>> +        }
>>       }
>>         ret = devm_thermal_add_hwmon_sysfs(tzdev);
>
Thank you Christophe for the review.

Regards,

Amjad