On 8/29/25 10:20 AM, Duje Mihanović wrote: > On Friday, 29 August 2025 01:40:56 Central European Summer Time David Lechner > wrote:
... >>> + *val = raw; >>> + dev_dbg(&iio->dev, "chan: %d, raw: %d\n", chan->channel, *val); >>> + ret = IIO_VAL_INT; >>> + break; >>> + case IIO_CHAN_INFO_PROCESSED: { >> >> Unusual to have both raw and processed. What is the motivation? > > I was following what ab8500-gpadc does, no particular motivation. > Considering the above, to me it makes the most sense to limit it to > processed. > In IIO, we generally prefer do the least amount of processing on the value from the hardware, so IIO_CHAN_INFO_RAW is mostly used and processed is only used in cases where there isn't linear scaling or some other unusually thing going on. So in this driver, processed probably makes sense for the resistance channels since we have to do some calculation anyway, but for the voltage channels, raw would be preferred.