Re: [PATCH 2/3] thermal: sun8i: Add support for H616 THS controller

[Andre Przywara]

On Fri, 18 Aug 2023 12:54:48 +0200
Martin Botka <martin.botka@xxxxxxxxxxxxxx> wrote:

Hi Martin,

> On Fri, Aug 18 2023 at 11:29:30 AM +01:00:00, Andre Przywara 
> <andre.przywara@xxxxxxx> wrote:
> > On Fri, 18 Aug 2023 10:43:17 +0200
> > Martin Botka <martin.botka@xxxxxxxxxxxxxx> wrote:
> > 
> > Hi Martin,
> > 
> > many thanks for the time and effort for upstreaming this!
> >   
> >>  Add support for the thermal sensor found in H616 SoC
> >>  which slightly resembles the H6 thermal sensor
> >>  controller with few changes like 4 sensors.
> >> 
> >>  Signed-off-by: Martin Botka <martin.botka@xxxxxxxxxxxxxx>
> >>  ---
> >>   drivers/thermal/sun8i_thermal.c | 115 
> >> ++++++++++++++++++++++++++++++++++++++++
> >>   1 file changed, 115 insertions(+)
> >> 
> >>  diff --git a/drivers/thermal/sun8i_thermal.c 
> >> b/drivers/thermal/sun8i_thermal.c
> >>  index 195f3c5d0b38..9d73e4313ad8 100644
> >>  --- a/drivers/thermal/sun8i_thermal.c
> >>  +++ b/drivers/thermal/sun8i_thermal.c
> >>  @@ -108,6 +108,12 @@ static int sun50i_h5_calc_temp(struct 
> >> ths_device *tmdev,
> >>   		return -1590 * reg / 10 + 276000;
> >>   }
> >> 
> >>  +static int sun50i_h616_calc_temp(struct ths_device *tmdev,
> >>  +			       int id, int reg)
> >>  +{
> >>  +	return (reg + tmdev->chip->offset) * tmdev->chip->scale;  
> > 
> > So if my school maths is not letting me down, this is the same as the
> > sun8i_ths_calc_temp() function, when using:
> > scale = h616_scale * -10
> > offset = h616_offset * h616_scale
> > 
> > So we do not need this new function, when we use:
> > +	.offset = 263655,
> > +	.scale = 810,
> > below, right?  
> That looks correct. Seems like my brain has let me down once again hehe.

Well, I guess you took the code from some BSP drop, right? And "same code
written differently" is a common pattern in those vendor kernels, because
they work on this "new SoC, new Linux kernel" premise, and miss out in
factoring out the commonalities. Which on the other hand is something that
Linux maintainers tend to push for.

> > Those values are not only positive, but also seem closer to the other
> > SoC's values.
> > This of course requires some small adjustment in the calibrate 
> > function,
> > to accommodate for the changed scale, but I leave this up as an 
> > exercise
> > to the reader ;-)
> > 
> > Martin, can you confirm that this works?  
> Will do :)
> >   
> >>  +}
> >>  +
> >>   static int sun8i_ths_get_temp(struct thermal_zone_device *tz, int 
> >> *temp)
> >>   {
> >>   	struct tsensor *s = thermal_zone_device_priv(tz);
> >>  @@ -278,6 +284,64 @@ static int sun50i_h6_ths_calibrate(struct 
> >> ths_device *tmdev,
> >>   	return 0;
> >>   }
> >> 
> >>  +static int sun50i_h616_ths_calibrate(struct ths_device *tmdev,
> >>  +				   u16 *caldata, int callen)  
> > 
> > nit: alignment  
> ack
> >   
> >>  +{
> >>  +	struct device *dev = tmdev->dev;
> >>  +	int i, ft_temp;
> >>  +
> >>  +	if (!caldata[0])
> >>  +		return -EINVAL;
> >>  +
> >>  +	/*
> >>  +	 * h616 efuse THS calibration data layout:
> >>  +	 *
> >>  +	 * 0      11  16     27   32     43   48    57
> >>  +	 * +----------+-----------+-----------+-----------+
> >>  +	 * |  temp |  |sensor0|   |sensor1|   |sensor2|   |
> >>  +	 * +----------+-----------+-----------+-----------+
> >>  +	 *                      ^           ^           ^
> >>  +	 *                      |           |           |
> >>  +	 *                      |           |           sensor3[11:8]
> >>  +	 *                      |           sensor3[7:4]
> >>  +	 *                      sensor3[3:0]
> >>  +	 *
> >>  +	 * The calibration data on the H616 is the ambient temperature and
> >>  +	 * sensor values that are filled during the factory test stage.
> >>  +	 *
> >>  +	 * The unit of stored FT temperature is 0.1 degreee celusis.  
> > 
> > nit: degree Celsius  
> .... I can't even legit excuse this typo (If it even can be called 
> typo). Got it tho.
> >   
> >>  +	 */
> >>  +	ft_temp = caldata[0] & FT_TEMP_MASK;
> >>  +
> >>  +	for (i = 0; i < tmdev->chip->sensor_num; i++) {
> >>  +		int delta, cdata, offset, reg;
> >>  +
> >>  +		if (i == 3)
> >>  +			reg = (caldata[1] >> 12)
> >>  +			      | (caldata[2] >> 12 << 4)
> >>  +			      | (caldata[3] >> 12 << 8);  
> > 
> > Can you add parentheses around the (caldata[2|3] >> 12) part? Makes 
> > it a
> > bit more readable.  
> yep
> >   
> >>  +		else
> >>  +			reg = (int)caldata[i + 1] & TEMP_CALIB_MASK;
> >>  +
> >>  +		delta = (ft_temp * 100 - tmdev->chip->calc_temp(tmdev, i, reg))
> >>  +			/ tmdev->chip->scale;  
> > 
> > Looks a bit odd, can you write this as over two lines?
> > 		delta = ft_temp ...;
> > 		delta /= tmdev->chip_scale;  
> can do.
> > 
> > (And this would be the place where you adjust the calculation to use 
> > the
> > new scale value).  
> yep. Tho it is bit of a spoiler for the reader ;)
> >   
> >>  +		cdata = CALIBRATE_DEFAULT - delta;
> >>  +		if (cdata & ~TEMP_CALIB_MASK) {
> >>  +			dev_warn(dev, "sensor%d is not calibrated.\n", i);
> >>  +
> >>  +			continue;
> >>  +		}
> >>  +
> >>  +		offset = (i % 2) * 16;
> >>  +		regmap_update_bits(tmdev->regmap,
> >>  +				   SUN50I_H6_THS_TEMP_CALIB + (i / 2 * 4),
> >>  +				   0xfff << offset,  
> > 
> > That should be TEMP_CALIB_MASK << offset, compare the H6 code.  
> Agreed. Missed this one. Ty.
> >   
> >>  +				   cdata << offset);
> >>  +	}
> >>  +
> >>  +	return 0;
> >>  +}
> >>  +
> >>   static int sun8i_ths_calibrate(struct ths_device *tmdev)
> >>   {
> >>   	struct nvmem_cell *calcell;
> >>  @@ -453,6 +517,43 @@ static int sun50i_h6_thermal_init(struct 
> >> ths_device *tmdev)
> >>   	return 0;
> >>   }
> >> 
> >>  +static int sun50i_h616_thermal_init(struct ths_device *tmdev)
> >>  +{
> >>  +	int val;
> >>  +
> >>  +	/*
> >>  +	 * T_acq = 20us
> >>  +	 * clkin = 24MHz
> >>  +	 *
> >>  +	 * x = T_acq * clkin - 1
> >>  +	 *   = 479
> >>  +	 */
> >>  +	regmap_write(tmdev->regmap, SUN50I_THS_CTRL0,
> >>  +		     SUN8I_THS_CTRL0_T_ACQ0(47) | SUN8I_THS_CTRL2_T_ACQ1(479));  
> > 
> > So this whole function is the same as the H6 (diff it!), except this 
> > line.
> > Which is actually also the same, just written differently (47 == 
> > 0x2f).
> > Can you please double check this, and if you agree, remove the whole
> > function and just use the H6 version?  
> They are not the same. Yes the 47 can be replaced with the unknown 
> value from H6.
> What isnt the same is the CTRL at the end. CTRL0 in H6 and CTRL2 in 
> H616. A very
> small change :)

But the definitions of SUN8I_THS_CTRL2_T_ACQ1 and SUN50I_THS_CTRL0_T_ACQ
are the same, aren't they?
I wonder if we are missing one piece of info for the H6 here, so the
"magic" 47 isn't actually magic, but there is some formula involving
clocks, similar to the one that results in 479.
Because then this would be more similar across the different SoCs: there
are *two* timing values for each SoC, and they just happen to be the same
for the H3 (20us), but are different for the H6 and H616 (one 2us, the
other 20us). And then your H616 line would make sense for the H6 as well:
		2us				20us
	SUN8I_THS_CTRL0_T_ACQ0(47) | SUN8I_THS_CTRL2_T_ACQ1(479));

In any case I still believe the H6 line results in the very same bit
pattern than your H616 line, but I am happy to stand corrected.
In which case I would ask you to unify the code anyway, somehow, so you
better hope I am right ;-)

Cheers,
Andre

> >>  +	/* average over 4 samples */
> >>  +	regmap_write(tmdev->regmap, SUN50I_H6_THS_MFC,
> >>  +		     SUN50I_THS_FILTER_EN |
> >>  +		     SUN50I_THS_FILTER_TYPE(1));
> >>  +	/*
> >>  +	 * clkin = 24MHz
> >>  +	 * filter_samples = 4
> >>  +	 * period = 0.25s
> >>  +	 *
> >>  +	 * x = period * clkin / 4096 / filter_samples - 1
> >>  +	 *   = 365
> >>  +	 */
> >>  +	regmap_write(tmdev->regmap, SUN50I_H6_THS_PC,
> >>  +		     SUN50I_H6_THS_PC_TEMP_PERIOD(365));
> >>  +	/* enable sensor */
> >>  +	val = GENMASK(tmdev->chip->sensor_num - 1, 0);
> >>  +	regmap_write(tmdev->regmap, SUN50I_H6_THS_ENABLE, val);
> >>  +	/* thermal data interrupt enable */
> >>  +	val = GENMASK(tmdev->chip->sensor_num - 1, 0);
> >>  +	regmap_write(tmdev->regmap, SUN50I_H6_THS_DIC, val);
> >>  +
> >>  +	return 0;
> >>  +}
> >>  +
> >>   static int sun8i_ths_register(struct ths_device *tmdev)
> >>   {
> >>   	int i;
> >>  @@ -608,6 +709,19 @@ static const struct ths_thermal_chip 
> >> sun50i_h6_ths = {
> >>   	.calc_temp = sun8i_ths_calc_temp,
> >>   };
> >> 
> >>  +static const struct ths_thermal_chip sun50i_h616_ths = {
> >>  +	.sensor_num = 4,
> >>  +	.has_bus_clk_reset = true,
> >>  +	.ft_deviation = 8000,
> >>  +	.offset = -3255,
> >>  +	.scale = -81,
> >>  +	.temp_data_base = SUN50I_H6_THS_TEMP_DATA,
> >>  +	.calibrate = sun50i_h616_ths_calibrate,
> >>  +	.init = sun50i_h616_thermal_init,
> >>  +	.irq_ack = sun50i_h6_irq_ack,
> >>  +	.calc_temp = sun50i_h616_calc_temp,
> >>  +};
> >>  +
> >>   static const struct of_device_id of_ths_match[] = {
> >>   	{ .compatible = "allwinner,sun8i-a83t-ths", .data = 
> >> &sun8i_a83t_ths },
> >>   	{ .compatible = "allwinner,sun8i-h3-ths", .data = &sun8i_h3_ths },
> >>  @@ -616,6 +730,7 @@ static const struct of_device_id of_ths_match[] 
> >> = {
> >>   	{ .compatible = "allwinner,sun50i-a100-ths", .data = 
> >> &sun50i_a100_ths },
> >>   	{ .compatible = "allwinner,sun50i-h5-ths", .data = &sun50i_h5_ths 
> >> },
> >>   	{ .compatible = "allwinner,sun50i-h6-ths", .data = &sun50i_h6_ths 
> >> },
> >>  +	{ .compatible = "allwinner,sun50i-h616-ths", .data = 
> >> &sun50i_h616_ths },
> >>   	{ /* sentinel */ },
> >>   };
> >>   MODULE_DEVICE_TABLE(of, of_ths_match);
> >>   
> >   
> 
>