On Wednesday, July 27, 2016 1:25:33 AM CEST Rich Felker wrote:
> On Tue, Jul 26, 2016 at 02:31:29PM +0200, Arnd Bergmann wrote:
> > The CLOCKSOURCE_OF_DECLARE now takes a function that returns an 'int', but
> > a this
> > new clocksource driver has just appeared in linux-next and causes a warning
> > because
> > it has the old 'void' return value:
> >
> > In file included from ../include/linux/clocksource.h:18:0,
> > from ../include/linux/clockchips.h:13,
> > from ../drivers/clocksource/jcore-pit.c:14:
> > include/linux/of.h:1006:20: error: comparison of distinct pointer types
> > lacks a cast [-Werror]
> > .data = (fn == (fn_type)NULL) ? fn : fn }
> > ^
> >
> > This adapts the jcore_pit_init() function to the correct return type.
> >
> > Signed-off-by: Arnd Bergmann <a...@arndb.de>
> > Fixes: e0aa0655c60b ("clocksource: add J-Core timer/clocksource driver")
> > ---
> > The patch that introduces the warnign currently only exists in the linux-sh
> > git
> > tree and showed up in linux-next this week.
>
> I've removed the commit from my for-next branch since it's not
> appropriate to go upstream through my tree. Your patch should still be
> applied in the proper subsystem tree, though, I think. Daniel?
I think the way to do it now is to fold my patch into your driver and submit
it to the clocksource git on top of the API change.
Arnd
