Re: [PATCH] iio: temperature: mlx90614: Fix dual channel probe condition check

[Jonathan Cameron <jic23@kernel.org>]

On Thu, 17 Sep 2026 14:16:15 +0100
Salah Triki <salah.triki@gmail.com> wrote:

> Hi Crt,
> 
> Thanks for your feedback.
> 
> Actually, the patch primarily fixes the behavior for the MLX90614.
> 
> In the current codebase, MLX90614 has `chip_info->dual_channel = true`.
> Because `mlx90614_probe_num_ir_sensors()` checks
> `if (chip_info->dual_channel)`, it returns 0 immediately for MLX90614
> without reading the EEPROM CONFIG1 register. As a result, dual-channel
> MLX90614 sensors are incorrectly registered as single-channel devices.
> 
> For MLX90615 (`dual_channel = false`), the condition was bypassed, causing
> an unnecessary SMBus read to EEPROM CONFIG1, even though MLX90615 only
> supports a single object channel.
> 
> By changing the check to `if (!chip_info->dual_channel)`, we fix both
> issues:
> 1. MLX90614 correctly reads EEPROM CONFIG1 to detect whether it operates
> in single or dual channel mode.
> 2. MLX90615 skips the EEPROM read entirely.
> 
> I don't have physical hardware to test this on; this bug was spotted via
> code inspection.

Explanation makes sense to me and as Crt hasn't come back with
a reply, I'll assuming Crt has no objections (shout if not!)

Applied and marked stable

Thanks,

Jonathan

> 
> Best regards,
> Salah
> 
> On Thu, Sep 17, 2026 at 01:27:48PM +0200, Crt Mori wrote:
> > So this fix is for the MLX90615, which is registered as single channel
> > operation? Did you test on 90614 as well?
> > 
> > Best regards,
> > Crt
> > 
> > Crt Mori
> > 
> > Melexis Technologies NV
> > Transportstraat 1
> > 3980 Tessenderlo
> > 
> > Mobile: +32 492 46 22 15
> > E-mail: cmo@melexis.com
> > Website: www.melexis.com
> > 
> > ----------------------------------------------------------
> > The contents of this e-mail are CONFIDENTIAL AND PROPRIETARY. Please
> > read our disclaimer at http://www.melexis.com/mailpolicy
> > 
> > 
> > On Thu, 17 Sept 2026 at 13:01, Salah Triki <salah.triki@gmail.com> wrote:  
> > >
> > > In mlx90614_probe_num_ir_sensors(), the check on chip_info->dual_channel
> > > is inverted.
> > >
> > > Currently, if dual_channel is true (MLX90614), the function returns 0
> > > immediately without checking op_eeprom_config1. This forces all MLX90614
> > > devices to be registered as single sensor (num_channels = 2), making the
> > > second object temperature channel inaccessible.
> > >
> > > Conversely, if dual_channel is false (MLX90615), the function skips the
> > > early return and performs an unnecessary SMBus read on EEPROM CONFIG1 even
> > > though MLX90615 does not support dual channel operation.
> > >
> > > Fix this by negating the condition so that single-channel chips return
> > > early with 0, while dual-channel capable chips (MLX90614) proceed to read
> > > the EEPROM configuration register to detect whether 1 or 2 object channels
> > > are present.
> > >
> > > Fixes: 3d5ead238bc8 ("iio: mlx90614: Factor our register IO and constants into model specific descriptor")
> > > Assisted-by: LLM
> > > Signed-off-by: Salah Triki <salah.triki@gmail.com>
> > > ---
> > >  drivers/iio/temperature/mlx90614.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/iio/temperature/mlx90614.c b/drivers/iio/temperature/mlx90614.c
> > > index 27d6ab5f5d7a..4f79da1f905e 100644
> > > --- a/drivers/iio/temperature/mlx90614.c
> > > +++ b/drivers/iio/temperature/mlx90614.c
> > > @@ -565,7 +565,7 @@ static int mlx90614_probe_num_ir_sensors(struct i2c_client *client)
> > >         const struct mlx_chip_info *chip_info = data->chip_info;
> > >         s32 ret;
> > >
> > > -       if (chip_info->dual_channel)
> > > +       if (!chip_info->dual_channel)
> > >                 return 0;
> > >
> > >         ret = i2c_smbus_read_word_data(client, chip_info->op_eeprom_config1);
> > > --
> > > 2.43.0
> > >