On Thu, 2018-06-28 at 11:05 +0200, Lorenzo Bianconi wrote:
> > AV_CONF register (RH & TEMP. oversampling ratio's) and CTRL1 register
> > (ODR & BDU settings) values are lost after suspend.
> > 
> > While the change in AV_CONF updates the sensor resolution modes
> > (overriding the user configuration before the device went to suspend);
> > loss of the contents of the CTRL1 register leads to failure in reading
> > sensor output.
> > 
> > This patch restores the AV_CONF & CTRL1 registers after resume.
> > 
> 
> Hi Shrirang,
> 
> I still suspect we have some hw issue (IoT Gateway) here (cc ST folks
> to get more info) and anyway I guess the proposed patch will fix the issue
> completely since OD and HL configuration is not restored properly.
> Could you please double check HL/OD regs are properly preserved after resume
> phase?
Yes, certainly. Will include the i2c register dumps before and after the resume to support
the patch.

Regards,
Shrirang
> 
> Regards,
> Lorenzo
> 
> > Changes from v1:
> > - Don't drop enabling the sensor during resume
> > - Restore AV_CONF register along with CTRL1
> >   (As demonstrated with i2c register dumps of hts221 captured on Dell
> >   IoT Gateways 300x before suspend & after resume [1])
> > 
> > Fixes: ffebe74b7c95 (iio: humidity: hts221: avoid useless ODR reconfiguration)
> > Signed-off-by: Shrirang Bagul <shrirang.bagul@canonical.com>
> > 
> > [v1] https://marc.info/?l=linux-iio&m=152506543000442&w=2
> > [1] https://marc.info/?l=linux-iio&m=152534455701742&w=2
> > 
> > This patch is based on:
> > git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git -b testing
> > ---
> >  drivers/iio/humidity/hts221_core.c | 31 +++++++++++++++++++++++++++++-
> >  1 file changed, 30 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/iio/humidity/hts221_core.c b/drivers/iio/humidity/hts221_core.c
> > index 166946d4978d..fc5599497f55 100644
> > --- a/drivers/iio/humidity/hts221_core.c
> > +++ b/drivers/iio/humidity/hts221_core.c
> > @@ -656,14 +656,43 @@ static int __maybe_unused hts221_resume(struct device *dev)
> >  {
> >  	struct iio_dev *iio_dev = dev_get_drvdata(dev);
> >  	struct hts221_hw *hw = iio_priv(iio_dev);
> > +	const struct hts221_avg *avg;
> > +	u8 data, idx;
> >  	int err = 0;
> >  
> > +	/* Restore contents of AV_CONF (RH & TEMP. oversampling ratio's) */
> > +	avg = &hts221_avg_list[HTS221_SENSOR_H];
> > +	idx = hw->sensors[HTS221_SENSOR_H].cur_avg_idx;
> > +	data = avg->avg_avl[idx];
> > +	err = hts221_update_avg(hw, HTS221_SENSOR_H, data);
> > +	if (err < 0)
> > +		goto fail_err;
> > +
> > +	avg = &hts221_avg_list[HTS221_SENSOR_T];
> > +	idx = hw->sensors[HTS221_SENSOR_T].cur_avg_idx;
> > +	data = avg->avg_avl[idx];
> > +	err = hts221_update_avg(hw, HTS221_SENSOR_T, data);
> > +	if (err < 0)
> > +		goto fail_err;
> 
> just return err, no need for goto statement
Okay.
> 
> > +
> > +	/* Restore contents of CTRL1 (BDU & ODR) */
> > +	err = regmap_update_bits(hw->regmap, HTS221_REG_CNTRL1_ADDR,
> > +				 HTS221_BDU_MASK,
> > +				 FIELD_PREP(HTS221_BDU_MASK, 1));
> > +	if (err < 0)
> > +		goto fail_err;
> > +
> > +	err = hts221_update_odr(hw, hw->odr);
> > +	if (err < 0)
> > +		goto fail_err;
> > +
> 
> just return err, no need for goto statement
Okay.
> 
> >  	if (hw->enabled)
> >  		err = regmap_update_bits(hw->regmap, HTS221_REG_CNTRL1_ADDR,
> >  					 HTS221_ENABLE_MASK,
> >  					 FIELD_PREP(HTS221_ENABLE_MASK,
> >  						    true));
> > -	return err;
> > +fail_err:
> > +	return err < 0 ? err : 0;
> >  }
> >  
> >  const struct dev_pm_ops hts221_pm_ops = {
> > -- 
> > 2.17.1
> >