On 15/01/17 17:06, Marek Vasut wrote: > On 01/15/2017 05:28 PM, Jonathan Cameron wrote: >> >> >> On 15 January 2017 15:47:31 GMT+00:00, Marek Vasut <[email protected]> >> wrote: >>> On 01/15/2017 03:05 PM, Jonathan Cameron wrote: >>>> On 15/01/17 00:14, Marek Vasut wrote: >>>>> Add IIO driver for the Renesas RCar GyroADC block. This block is a >>>>> simple 4/8-channel ADC which samples 12/15/24 bits of data every >>>>> cycle from all channels. >>>>> >>>>> Signed-off-by: Marek Vasut <[email protected]> >>>>> Cc: Geert Uytterhoeven <[email protected]> >>>>> Cc: Simon Horman <[email protected]> >>>>> Cc: Jonathan Cameron <[email protected]> >>>>> Cc: [email protected] >>>> One question about the name attribute and what it gives you. Looks >>> like another >>>> case of what Lars has been pointing out this morning. >>>> >>>> We had a load of these sneak in. Changing them would break userspace >>> code >>>> so we can't fix them, but that attribute ought to give you something >>>> related to the part number. >>>> >>>> Here it might even be cute to give an indication of what is connected >>>> to it as part of the name. >>> >>> You can extract that information from the of_node, so I don't think >>> that >>> is a good idea, it'd only bring in duplication: >>> >>> ~ # cat /sys/bus/iio/devices/iio\:device0/of_node/adc@*/compatible >>> maxim,max11100 >>> >> Fair enough >>>> J >>> >>> [...] >>> >>>>> + ret = rcar_gyroadc_parse_subdevs(indio_dev); >>>>> + if (ret) >>>>> + return ret; >>>>> + >>>>> + ret = rcar_gyroadc_init_supplies(indio_dev); >>>>> + if (ret) >>>>> + return ret; >>>>> + >>>>> + priv->model = (enum rcar_gyroadc_model)of_id->data; >>>>> + >>>>> + platform_set_drvdata(pdev, indio_dev); >>>>> + >>>>> + indio_dev->name = dev_name(dev); >>>> What do you then get when you read /sys/bus/iio/iio\:deviceX/name? >>> >>> e6e54000.adc >>> >>>> Should be something like the part number, rather than the id of the >>>> parent device (I'm not 100% sure what that even is in this case!) >>> >>> You mean like "rcar-gyroadc" ? That works for me, sure. >> Yes, that will do nicely. > > > Done and V6 is out. > Cool. I'm going to let this sit for a few days at least. It's unusual enough that I want to give it a bit more time on the list.
Jonathan
