Hi Jacopo, On Tue, Jun 5, 2018 at 10:12 AM, jacopo mondi <jacopo@xxxxxxxxxx> wrote: > On Tue, Jun 05, 2018 at 09:49:38AM +0200, Simon Horman wrote: >> On Mon, Jun 04, 2018 at 02:23:25PM +0200, Niklas Söderlund wrote: >> > On 2018-05-29 17:05:59 +0200, Jacopo Mondi wrote: >> > > The 'bus-width' and 'pclk-sample' properties are not parsed by the VIN >> > > driver and only confuse users. Remove them in all Gen2 SoC that use >> > > them. >> > > >> > > Signed-off-by: Jacopo Mondi <jacopo+renesas@xxxxxxxxxx> >> > >> > The more I think about this the more I lean towards that this patch >> > should be dropped. The properties accurately describes the hardware and >> > I think there is value in that. That the driver currently don't parse or >> > make use of them don't in my view reduce there value. Maybe you should >> > break out this patch to a separate series? >> >> I also think there is value in describing the hardware not the state of the >> driver at this time. Is there any missmatch between these properties and >> the bindings? > > Niklas and I discussed a bit offline on this yesterday. My main > concern, and sorry for being pedant on this, is that changing those > properties value does not change the interface behaviour, and this > could cause troubles when integrating image sensor not known to be > working on the VIN interface. > > This said, the documentation of those (and all other) properties is in the > generic "video-interfaces.txt" file and it is my understanding, but I think > Laurent and Rob agree on this as well from their replies to my previous series, > that each driver should list which properties it actually supports, as s/driver/device-specific binding/ > some aspects are very implementation specific, like default values and > what happens if the property is not specified [1]. Nonetheless, all In se defaults are not (Linux) implementation-specific, but fixed in the DT bindings. > properties describing hardware features and documented in the generic > file should be accepted in DTS, as those aims to be OS-independent and > even independent from the single driver implementation. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds