Re: [PATCH 08/14] mfd: remove unused aat2870 driver

On Sun, Sep 13, 2026, at 00:20, Jonathan Cameron wrote:

On Thu, 10 Sep 2026 20:14:15 +0200 "Arnd Bergmann" [off-list ref] wrote:

quoted

I tried to be careful about figuring out exactly which drivers are
unused, but I'm sure there are still a few false positive and false
negative ones in there. I've added the current list of driver files below,
let me know if you see something that shouldn't be there.

Hi Arnd

For I2C and SPI at least, there doesn't need to be explicit
device tree support as long as they have defaults when platform
data isn't there. Those two buses will happily bind based on
a dt-compatible and their i2c_device_id or spi_device_id tables
for example.

So unless they fail to probe (some might?) we don't have
a clear signal on whether they are in use via DT or not.

For vast majority of IIO drivers we don't have an upstream
DTS as there is no clear motivation for anyone to upstream
the dts for a random industrial control board or similar.

Right, that was my general rule, if I had tried to remove
all drivers that could plausibly probe with an external dtb
but have no internal users, that would have easily doubled
the 300 patches I already did.

quoted

drivers/iio/adc/ad7266.c

No to dropping this.

quoted

drivers/iio/adc/ad7791.c

No to dropping this one.

quoted

drivers/iio/adc/ad7793.c

Maybe. Analog devices ack needed.

Production part but this one indeed fails to probe.  Analog
folk, do you want to fix this one up?

quoted

drivers/iio/adc/ad7887.c

No to dropping this.

Production part - should work fine with defaults in driver.
Could like the others drop the platform data handling.

This was one patch that I wasn't sure about myself (all four
drivers together) since it looks like even when they do probe
from dtb, the feature set would be limited without a DT
binding, and they have been in the tree for a rather long time
without users.

I've dropped the patch now. I looked at dropping the platform_data
handling, but I think that only makes sense if we get a
proper DT binding first.

quoted

drivers/iio/adc/lp8788_adc.c

Probably

Sub driver of an MFD.  I'm fine with that going if we know
the mfd is no longer in use.

Ok, I posted that one as
https://lore.kernel.org/all/20260909132153.1596191-8-arnd@kernel.org/ (local)

and so far, everyone agreed on removing it.

quoted

drivers/iio/adc/lpc18xx_adc.c
drivers/iio/dac/lpc18xx_dac.c

Probably

Likewise these two.

Right, this one of course should follow the removal of
CONFIG_ARCH_LPC18XX in arch/arm/

quoted

drivers/iio/frequency/ad9523.c

Maybe. Analog Devices ack needed.

Fails to probe without platform data and marked not recomended
for new designs. If we get an Ack from Analog devices folk
I'm fine with this one going away.

This matches what I wrote in 
https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git/commit/?id=20c1515b0caec2769c0916d087326f640bfb8aab

I'll keep the patch in the series and we'll see what the
maintainers think when I post it.

Thanks a lot for taking a look!

I wonder if some of the older Analog drivers were only ever used
on Blackfin. My series still removes a few more drivers outside
of iio list that had platform_data in arch/blackfin/ until we
removed that in 2018. These all fail to probe without
platform_data, which means they also wouldn't work with the
downstream adsp-sc5xx/sc8xx port or any other upstream Arm
platform:

a4bdf847e0e1 backlight: remove unused adp8860/8870 drivers
c71829e110e9 Input: touchscreen: remove unused ad7877 driver
b4d4ca64500b Input: misc: remove unused ad714x driver
978426a3fe1a usb: remove unused sl811 driver
0860c9f297c3 mfd: remove unused adp5520 driver

       Arnd