Re: [PATCH 2/3] thermal: Add Mediatek thermal controller support

From: Eduardo Valentin
Date: Tue Nov 10 2015 - 13:26:39 EST


On Tue, Nov 10, 2015 at 12:05:54PM +0000, Javi Merino wrote:
> On Mon, Nov 09, 2015 at 11:13:32AM +0100, Sascha Hauer wrote:

<cut>

> > +
> > +/*
> > + * The MT8173 thermal controller has four banks. Each bank can read up to
> > + * four temperature sensors simultaneously. The MT8173 has a total of 5
> > + * temperature sensors. We use each bank to measure a certain area of the
> > + * SoC. Since TS2 is located centrally in the SoC it is influenced by multiple
> > + * areas, hence is used in different banks.
> > + */
> > +static const struct mtk_thermal_bank_cfg bank_data[] = {
> > + {
> > + .num_sensors = 2,
> > + .sensors = { MT8173_TS2, MT8173_TS3 },
> > + }, {
> > + .num_sensors = 2,
> > + .sensors = { MT8173_TS2, MT8173_TS4 },
> > + }, {
> > + .num_sensors = 3,
> > + .sensors = { MT8173_TS1, MT8173_TS2, MT8173_TSABB },
> > + }, {
> > + .num_sensors = 1,
> > + .sensors = { MT8173_TS2 },
> > + },
> > +};

Would it make sense to simply expose all sensors and let the
configuration of their aggregation be done by DT?

There is already ongoing effort to get aggregation functions
generalized.

> > +
> > +struct mtk_thermal_sense_point {
> > + int msr;
> > + int adcpnp;
> > +};
> > +
> > +static const struct mtk_thermal_sense_point
> > + sensing_points[MT8173_NUM_SENSORS_PER_ZONE] = {
> > + {
> > + .msr = TEMP_MSR0,
> > + .adcpnp = TEMP_ADCPNP0,
> > + }, {
> > + .msr = TEMP_MSR1,
> > + .adcpnp = TEMP_ADCPNP1,
> > + }, {
> > + .msr = TEMP_MSR2,
> > + .adcpnp = TEMP_ADCPNP2,
> > + }, {
> > + .msr = TEMP_MSR3,
> > + .adcpnp = TEMP_ADCPNP3,
> > + },
> > +};
> > +
> > +/**
> > + * raw_to_mcelsius - convert a raw ADC value to mcelsius
> > + * @mt: The thermal controller
> > + * @raw: raw ADC value
> > + *
> > + * This converts the raw ADC value to mcelsius using the SoC specific
> > + * calibration constants
> > + */
> > +static int raw_to_mcelsius(struct mtk_thermal *mt, int sensno, s32 raw)
> > +{
> > + s32 tmp;
> > +
> > + raw &= 0xfff;
> > +
> > + tmp = 203450520 << 3;
> > + tmp /= 165 + mt->o_slope;
> > + tmp /= 10000 + mt->adc_ge;
> > + tmp *= raw - mt->vts[sensno] - 3350;
> > + tmp >>= 3;
> > +
> > + return mt->degc_cali * 500 - tmp;
> > +}
> > +
> > +/**
> > + * mtk_thermal_get_bank - get bank
> > + * @bank: The bank
> > + *
> > + * The bank registers are banked, we have to select a bank in the
> > + * PTPCORESEL register to access it.
> > + */
> > +static void mtk_thermal_get_bank(struct mtk_thermal_bank *bank)
> > +{
> > + struct mtk_thermal *mt = bank->mt;
> > + u32 val;
> > +
> > + mutex_lock(&mt->lock);
> > +
> > + val = readl(mt->thermal_base + PTPCORESEL);
> > + val &= ~0xf;
> > + val |= bank->id;
> > + writel(val, mt->thermal_base + PTPCORESEL);
> > +}
> > +
> > +/**
> > + * mtk_thermal_put_bank - release bank
> > + * @bank: The bank
> > + *
> > + * release a bank previously taken with mtk_thermal_get_bank,
> > + */
> > +static void mtk_thermal_put_bank(struct mtk_thermal_bank *bank)
> > +{
> > + struct mtk_thermal *mt = bank->mt;
> > +
> > + mutex_unlock(&mt->lock);
> > +}
> > +
> > +/**
> > + * mtk_thermal_bank_temperature - get the temperature of a bank
> > + * @bank: The bank
> > + *
> > + * The temperature of a bank is considered the maximum temperature of
> > + * the sensors associated to the bank.
> > + */
> > +static int mtk_thermal_bank_temperature(struct mtk_thermal_bank *bank)
> > +{
> > + struct mtk_thermal *mt = bank->mt;
> > + int temp, i, max;
> > + u32 raw;
> > +
> > + temp = max = INT_MIN;
> > +
> > + for (i = 0; i < bank_data[bank->id].num_sensors; i++) {
> > + raw = readl(mt->thermal_base + sensing_points[i].msr);
> > +
> > + temp = raw_to_mcelsius(mt, 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;
> > +
> > + if (temp > max)
> > + max = temp;
> > + }
> > +
> > + return max;
> > +}
> > +
> > +static int mtk_read_temp(void *data, int *temperature)
> > +{
> > + struct mtk_thermal *mt = data;
> > + int i;
> > + int tempmax = INT_MIN;
> > +
> > + for (i = 0; i < MT8173_NUM_ZONES; i++) {
> > + struct mtk_thermal_bank *bank = &mt->banks[i];
> > + int t;
> > +
> > + mtk_thermal_get_bank(bank);
> > +
> > + t = mtk_thermal_bank_temperature(bank);
>
> IIUIC, when you had multiple thermal zones
> mtk_thermal_bank_temperature() made sense, but now it looks like
> you're just doing the maximum of all sensors. Why bother with the
> banks any more? Aren't you just calculating the maximum of all
> sensors? As TS2 is present in all banks, there's no point in reading
> it four times just to get the maximum of all sensors.
>

Yeah, agreed here. If that is your intention, maybe read each sensor one
time, then compute the max of each subset from memory instead.

> > + mtk_thermal_put_bank(bank);
> > +
> > + if (t > tempmax)
> > + tempmax = t;
> > + }
> > +
> > + *temperature = tempmax;
> > +
> > + return 0;
> > +}
> > +
> > +static const struct thermal_zone_of_device_ops mtk_thermal_ops = {
> > + .get_temp = mtk_read_temp,
> > +};
>
> Cheers,
> Javi
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/