On Fri, 4 Oct 2024 17:07:56 +0300 Antoniu Miclaus <antoniu.miclaus@xxxxxxxxxx> wrote: > Add documentation for the packet size. > > Signed-off-by: Antoniu Miclaus <antoniu.miclaus@xxxxxxxxxx> > --- > changes in v2: > - improve description for packet_format > - add kernel version > .../ABI/testing/sysfs-bus-iio-adc-ad485x | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > create mode 100644 Documentation/ABI/testing/sysfs-bus-iio-adc-ad485x > > diff --git a/Documentation/ABI/testing/sysfs-bus-iio-adc-ad485x b/Documentation/ABI/testing/sysfs-bus-iio-adc-ad485x > new file mode 100644 > index 000000000000..5d69a8d30383 > --- /dev/null > +++ b/Documentation/ABI/testing/sysfs-bus-iio-adc-ad485x > @@ -0,0 +1,16 @@ > +What: /sys/bus/iio/devices/iio:deviceX/packet_format_available > +KernelVersion: 6.13 > +Contact: linux-iio@xxxxxxxxxxxxxxx > +Description: > + Packet sizes on the CMOS or LVDS conversion data output bus. > + Reading this returns the valid values that can be written to the > + packet_format. > + > +What: /sys/bus/iio/devices/iio:deviceX/packet_format > +KernelVersion: 6.13 > +Contact: linux-iio@xxxxxxxxxxxxxxx > +Description: > + This attribute configures the frame size on conversion data > + output bus. See packet_format_available for available sizes > + based on the device used. > + Reading returns the actual size used. This needs to give some guidance to the user on 'why' they might pick a particular format. I'm also inclined to suggest that for now we pick a sensible default dependent on the other options enabled (oversampling etc) and don't expose it to the user. Eventually it looks like we may have to figure out a solution to describe metadata packed alongside the channel readings but that may take a while and I don't want to stall this driver on that discussion. Thanks, Jonathan
