I noticed in its DSDT that one of my tablets actually is using the GPADC
pin for temperature monitoring.
The whole axp288_adc_set_ts() function is a bit weird, in the past it was
removed because it seems to make no sense, then this was reverted because
of regressions.
So I decided to test the special GPADC pin handling on this tablet.
Conclusion: not only is axp288_adc_set_ts() necessary, we need to sleep a
bit after making the AXP288_ADC_TS_PIN_CTRL changes before sampling the
GPADC, otherwise it will often (about 80% of the time) read 0 instead of
its actual value.
It seems that there is only 1 bias current source and to be able to use it
for the GPIO0 pin in GPADC mode it must be temporarily turned off for the
TS pin, but the datasheet does not mention this.
This commit adds a sleep after disabling the TS pin bias current,
fixing the GPADC more often then not wrongly returning 0.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
static int axp288_adc_set_ts(struct regmap *regmap, unsigned int mode,
unsigned long address)
{
+ int ret;
+
/* channels other than GPADC do not need to switch TS pin */
if (address != AXP288_GP_ADC_H)
return 0;
- return regmap_write(regmap, AXP288_ADC_TS_PIN_CTRL, mode);
+ ret = regmap_write(regmap, AXP288_ADC_TS_PIN_CTRL, mode);
+ if (ret)
+ return ret;
+
+ /* When switching to the GPADC pin give things some time to settle */
+ if (mode == AXP288_ADC_TS_PIN_GPADC)
+ usleep_range(6000, 10000);
+
+ return 0;
}
static int axp288_adc_read_raw(struct iio_dev *indio_dev,