touchscreen: sun4i-ts: A10 (sun4i) has a different temperature curve

Testing has revealed that the temperature in the rtp controller of the A10 (sun4i) SoC has a different curve then on the A13 (sun5i) and later models.

Add a new sun5i-a13-ts compatible to differentiate the newer models and set the curve based on the compatible string.

The new curve is still not ideal on all A10-s, that seems to have to do with there being a large spread between different A10-s out there, the new curve us based on callibration results on 4 completely different models:
raw min raw max temp min temp max stepsize offset

Tong Zhang's hackberry    2402    2680    45.0     80.0    0.125   -255.3
Hansg's Cubieboard        2207    2300    36.0     45.0    0.096   -175.8
Olliver's lime 1 (*):     2258    2537    48.3     87.1    0.139   -265.7
Olliver's lime 2 (*):     2222    2486    46.7     91.7    0.170   -331.0

*) from: http://linux-sunxi.org/Temperature_Calibration

Average all 4:                                             0.133   -257.0
Average without outliers (middle 2):                       0.132   -261.0

Since it is better to slightly overreport the temperature this patch uses the average of all 4 as curve.

This fixes the temperature reported on the A10 being much higher then expected.

Cc Tong Zhang <[email protected]>

Cc Olliver Schinagl <[email protected]>

Reported-by Tong Zhang <[email protected]>

Signed-off-by Hans de Goede <[email protected]>

---
 .../devicetree/bindings/input/touchscreen/sun4i.txt      |  3 ++-
 drivers/input/touchscreen/sun4i-ts.c                     | 16 +++++++++++++---
 2 files changed, 15 insertions(+), 4 deletions(-)

Documentation/devicetree/bindings/input/touchscreen/sun4i.txt +2 -1

index 433332d..d59d252 100644
@@ -2,7 +2,8 @@ sun4i resistive touchscreen controller
 --------------------------------------
 
 Required properties:
- - compatible: "allwinner,sun4i-a10-ts" or "allwinner,sun6i-a31-ts"
+ - compatible: "allwinner,sun4i-a10-ts", "allwinner,sun5i-a13-ts" or
+   "allwinner,sun6i-a31-ts"
  - reg: mmio address range of the chip
  - interrupts: interrupt to which the chip is connected
  - #thermal-sensor-cells: shall be 0

drivers/input/touchscreen/sun4i-ts.c +13 -3

index b93a28b..66ccd5a 100644
@@ -258,6 +258,15 @@ static int sun4i_ts_probe(struct platform_device *pdev)
 		/* Allwinner SDK has temperature = -271 + (value / 6) (C) */
 		ts->temp_offset = 1626;
 		ts->temp_step = 167;
+	} else if (of_device_is_compatible(np, "allwinner,sun4i-a10-ts")) {
+		/*
+		 * The A10 temperature sensor has quite a wide spread, these
+		 * parameters are based on the averaging of the calibration
+		 * results of 4 completely different boards, with a spread of
+		 * temp_step from 96 - 170 and temp_offset from 1758 - 3310.
+		 */
+		ts->temp_offset = 2570;
+		ts->temp_step = 133;
 	} else {
 		/*
 		 * The user manuals do not contain the formula for calculating

@@ -330,10 +339,10 @@ static int sun4i_ts_probe(struct platform_device *pdev)
 	 * finally enable tp mode.
 	 */
 	reg = STYLUS_UP_DEBOUN(5) | STYLUS_UP_DEBOUN_EN(1);
-	if (of_device_is_compatible(np, "allwinner,sun4i-a10-ts"))
-		reg |= TP_MODE_EN(1);
-	else
+	if (of_device_is_compatible(np, "allwinner,sun6i-a31-ts"))
 		reg |= SUN6I_TP_MODE_EN(1);
+	else
+		reg |= TP_MODE_EN(1);
 	writel(reg, ts->base + TP_CTRL1);
 
 	/*

@@ -383,6 +392,7 @@ static int sun4i_ts_remove(struct platform_device *pdev)
 
 static const struct of_device_id sun4i_ts_of_match[] = {
 	{ .compatible = "allwinner,sun4i-a10-ts", },
+	{ .compatible = "allwinner,sun5i-a13-ts", },
 	{ .compatible = "allwinner,sun6i-a31-ts", },
 	{ /* sentinel */ }
 };

Signature · 1 line

2.3.1

Quoted 26 lines · Hans de Goede

Testing has revealed that the temperature in the rtp controller of the A10 (sun4i) SoC has a different curve then on the A13 (sun5i) and later models.

Add a new sun5i-a13-ts compatible to differentiate the newer models and set the curve based on the compatible string.

The new curve is still not ideal on all A10-s, that seems to have to do with there being a large spread between different A10-s out there, the new curve us based on callibration results on 4 completely different models:
raw min raw max temp min temp max stepsize offset

Tong Zhang's hackberry    2402    2680    45.0     80.0    0.125   -255.3
Hansg's Cubieboard        2207    2300    36.0     45.0    0.096   -175.8
Olliver's lime 1 (*):     2258    2537    48.3     87.1    0.139   -265.7
Olliver's lime 2 (*):     2222    2486    46.7     91.7    0.170   -331.0

*) from: http://linux-sunxi.org/Temperature_Calibration

Average all 4:                                             0.133   -257.0
Average without outliers (middle 2):                       0.132   -261.0

Since it is better to slightly overreport the temperature this patch uses the average of all 4 as curve.

This fixes the temperature reported on the A10 being much higher then expected.

Quoted 4 lines

Cc Tong Zhang <[email protected]>

Cc Olliver Schinagl <[email protected]>

Reported-by Tong Zhang <[email protected]>

Signed-off-by Hans de Goede <[email protected]>

Applied, thank you.

Quoted 65 lines

---
  .../devicetree/bindings/input/touchscreen/sun4i.txt      |  3 ++-
  drivers/input/touchscreen/sun4i-ts.c                     | 16 +++++++++++++---
  2 files changed, 15 insertions(+), 4 deletions(-)

Documentation/devicetree/bindings/input/touchscreen/sun4i.txt +2 -1

index 433332d..d59d252 100644
@@ -2,7 +2,8 @@ sun4i resistive touchscreen controller
  --------------------------------------
 
  Required properties:
- - compatible: "allwinner,sun4i-a10-ts" or "allwinner,sun6i-a31-ts"
+ - compatible: "allwinner,sun4i-a10-ts", "allwinner,sun5i-a13-ts" or
+   "allwinner,sun6i-a31-ts"
   - reg: mmio address range of the chip
   - interrupts: interrupt to which the chip is connected
   - #thermal-sensor-cells: shall be 0

drivers/input/touchscreen/sun4i-ts.c +13 -4

index b93a28b..66ccd5a 100644
@@ -258,6 +258,15 @@ static int sun4i_ts_probe(struct platform_device *pdev)
  		/* Allwinner SDK has temperature = -271 + (value / 6) (C) */
  		ts->temp_offset = 1626;
  		ts->temp_step = 167;
+	} else if (of_device_is_compatible(np, "allwinner,sun4i-a10-ts")) {
+		/*
+		 * The A10 temperature sensor has quite a wide spread, these
+		 * parameters are based on the averaging of the calibration
+		 * results of 4 completely different boards, with a spread of
+		 * temp_step from 96 - 170 and temp_offset from 1758 - 3310.
+		 */
+		ts->temp_offset = 2570;
+		ts->temp_step = 133;
  	} else {
  		/*
  		 * The user manuals do not contain the formula for calculating

@@ -330,10 +339,10 @@ static int sun4i_ts_probe(struct platform_device *pdev)
  	 * finally enable tp mode.
  	 */
  	reg = STYLUS_UP_DEBOUN(5) | STYLUS_UP_DEBOUN_EN(1);
-	if (of_device_is_compatible(np, "allwinner,sun4i-a10-ts"))
-		reg |= TP_MODE_EN(1);
-	else
+	if (of_device_is_compatible(np, "allwinner,sun6i-a31-ts"))
  		reg |= SUN6I_TP_MODE_EN(1);
+	else
+		reg |= TP_MODE_EN(1);
  	writel(reg, ts->base + TP_CTRL1);
 
  	/*

@@ -383,6 +392,7 @@ static int sun4i_ts_remove(struct platform_device *pdev)
 
  static const struct of_device_id sun4i_ts_of_match[] = {
  	{ .compatible = "allwinner,sun4i-a10-ts", },
+	{ .compatible = "allwinner,sun5i-a13-ts", },
  	{ .compatible = "allwinner,sun6i-a31-ts", },
  	{ /* sentinel */ }
  };
--

2.3.1

The touchscreen controller in the A13 and later has a different temperature curve than the one in the original A10, change the compatible for the A13 and later so that the kernel will use the correct curve.

Reported-by Tong Zhang <[email protected]>

Signed-off-by Hans de Goede <[email protected]>

Hi Dmitry, Maxime, et al.

Here is a patch-set for the sun4i-ts driver temperature sensor support, it consists of 2 fixes:

1) Use a better temperature curve for A10 SoCs to avoid reporting a way to high temperature.

2) Start with reporting an "estimated" temperature of 40 degrees rather then returning -EAGAIN until we get the first temperature ready interrupt to avoid the kernel tempzone core logging errors because of our -EAGAIN return.

Dmitry can you please queue up the 2 sun4i-ts patches for 4.1, and Maxime, can you please queue up the matching dts change ?

Thanks & Regards,

Hans

Hi Hans,

Quoted 8 lines · Hans de Goede

Hi Dmitry, Maxime, et al.

Here is a patch-set for the sun4i-ts driver temperature sensor support, it consists of 2 fixes:

1) Use a better temperature curve for A10 SoCs to avoid reporting a way to high temperature.

Quoted 3 lines

2) Start with reporting an "estimated" temperature of 40 degrees rather then returning -EAGAIN until we get the first temperature ready interrupt to avoid the kernel tempzone core logging errors because of our -EAGAIN return.

No, I think we better teach thermal core not to complain about -EAGAINs.