2017-12-21 15:25 GMT+01:00 Andy Shevchenko <andy.shevche...@gmail.com>: > On Thu, Dec 21, 2017 at 4:18 PM, Bartosz Golaszewski <b...@bgdev.pl> wrote: >> 2017-12-21 15:08 GMT+01:00 Andy Shevchenko <andy.shevche...@gmail.com>: >>> On Thu, Dec 21, 2017 at 3:48 PM, Bartosz Golaszewski <b...@bgdev.pl> wrote: >>>> "atmel,spd" is reported by checkpatch as undocumented in the device >>>> tree bindings. Add it to the list of supported compatible strings. >>> >>>> "atmel,24c00", >>>> "atmel,24c01", >>>> "atmel,24c02", >>>> + "atmel,spd", >>>> "atmel,24c04", >>>> "atmel,24c08", >>>> "atmel,24c16", >>> >>> Sounds alogical to me to make a split by this new record. >>> Can you find better line to inject? > >> They are actually ordered by memory size. I want to keep it like this >> in the driver and I prefer that the DT reflect it. > > So, I just disagree on the above. Rationale I described at one of the comment. > > At the end it's your call, but from my p.o.v. it makes life harder to > read and catch the chips which are (un)supported. > > -- > With Best Regards, > Andy Shevchenko
Yes, I prefer this version. Another reason to keep it like this is to group chip variants together. E.g: atmel,24cs02, atmel,spd, atmel,24mac402, atmel,24mac602 are all variants of atmel,24c02. Thanks, Bartosz
