On Wed, Aug 2, 2017 at 2:06 PM, Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote: > On Tue, 2017-08-01 at 12:24 -0700, Dmitry Torokhov wrote: >> On Wed, Jul 26, 2017 at 01:21:37PM +0300, Andy Shevchenko wrote: >> > On Tue, 2017-07-25 at 13:58 -0700, Dmitry Torokhov wrote: >> > > When converting legacy board to use gpiod API() there migt be >> > > several >> > > lookup tables in board file, let's provide a way to register them >> > > all >> > > at >> > > once. >> > >> > Looking into the code which is using several GPIO look up tables I >> > noticed that the pattern often something like >> > >> > >> > gpiod_add_lookup_table(x_tbl); >> > platform_device_register(x_dev); >> > >> > ... >> > >> > gpiod_add_lookup_table(y_tbl); >> > platform_device_register(y_dev); >> > >> > which looks also logical. >> > >> > So, I have no strong opinion here, though it would be nice to have >> > an >> > example where it makes sense. >> >> Here is an example. It actually does not have multiple GPIO tables, >> but >> it allows to keep the structure of init sequence for the board, which >> relies on having series of arrays conditionally populated with devices >> and other data that need to be registered. > > I see your point. Okay, no objections from me. > > Please, add a stub for !GPIOLIB case. > > After addressing it, > > Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> OK I wait for the stub-inclusive patch. Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-gpio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html