Re: [PATCH v2] iio: accel: bmc150: check reset write error

[Carlos Casadiego]

I have not verified this on BMC150 hardware.

The datasheet documents the soft reset, but I could not find a guarantee
that the I2C write is acknowledged before the reset starts.

Given that, I agree that propagating this error could cause a regression,
so I think the patch should be dropped.

Thanks for pointing this out.

En Fri, Sep 25, 2026 at 04:05:00AM +0100, Jonathan Cameron escribió:
> On Thu, 24 Sep 2026 09:48:45 -0500
> Carlos Casadiego via B4 Relay <devnull+cdcp206.gmail.com@xxxxxxxxxx> wrote:
> 
> > From: Carlos Casadiego <cdcp206@xxxxxxxxx>
> > 
> > The software reset is required to bring the device into a known
> > state before reading the chip ID.
> > 
> > Check the return value of regmap_write() and abort initialization if
> > the reset command cannot be written.
> > 
> > Fixes: 1c500840934a ("iio: accel: bmc150: reset chip at init time")
> > Acked-by: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > Signed-off-by: Carlos Casadiego <cdcp206@xxxxxxxxx>
> 
> Just to check: Was this verified against actual hardware?
> 
> Quite a few devices get going on their resets before sending
> acks on i2c which means you get an error on that particular
> write.  That is often why you don't see an error check on this
> particular operation.
> 
> 
> 
> 
> > ---
> > Changes in v2:
> > - Remove the blank line in the tag block.
> > - Add Srinivas Pandruvada's Acked-by.
> > - Link to v1: https://patch.msgid.link/20260923-iio-bmc150-check-reset-error-v1-1-b8c1ebc4c947@xxxxxxxxx
> > ---
> >  drivers/iio/accel/bmc150-accel-core.c | 8 ++++++--
> >  1 file changed, 6 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/iio/accel/bmc150-accel-core.c b/drivers/iio/accel/bmc150-accel-core.c
> > index 46bb36d9366e..43a39e1ccce5 100644
> > --- a/drivers/iio/accel/bmc150-accel-core.c
> > +++ b/drivers/iio/accel/bmc150-accel-core.c
> > @@ -1561,8 +1561,12 @@ static int bmc150_accel_chip_init(struct bmc150_accel_data *data)
> >  	 * Reset chip to get it in a known good state. A delay of 1.8ms after
> >  	 * reset is required according to the data sheets of supported chips.
> >  	 */
> > -	regmap_write(data->regmap, BMC150_ACCEL_REG_RESET,
> > -		     BMC150_ACCEL_RESET_VAL);
> > +	ret = regmap_write(data->regmap, BMC150_ACCEL_REG_RESET,
> > +			   BMC150_ACCEL_RESET_VAL);
> > +	if (ret < 0) {
> > +		dev_err(dev, "Error writing reset register\n");
> > +		return ret;
> > +	}
> >  	usleep_range(1800, 2500);
> >  
> >  	ret = regmap_read(data->regmap, BMC150_ACCEL_REG_CHIP_ID, &val);
> > 
> > ---
> > base-commit: d8c0f48f0b1583308a98401c0c7b1a65d1d43a6c
> > change-id: 20260923-iio-bmc150-check-reset-error-389b096b0ad4
> > 
> > Best regards,
> > --  
> > Carlos Casadiego <cdcp206@xxxxxxxxx>
> > 
> > 
>