Re: [PATCH] 8250: uniphier: allow modular build with 8250 console

[Arnd Bergmann]

On Thursday 21 January 2016 20:39:03 Masahiro Yamada wrote:
> 2016-01-14 7:33 GMT+09:00 Arnd Bergmann <arnd@xxxxxxxx>:
> > The recently added uniphier 8250 port driver supports early console
> > probing, and it supports being built as a module, but the combination
> > of the two fails to link:
> >
> > ERROR: "early_serial8250_setup" [drivers/tty/serial/8250/8250_uniphier.ko] undefined!
> >
> > Given that earlycon support in a loadable module makes no sense,
> > making that code conditional on 'MODULE' is a correct solution.
> >
> > Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> > Fixes: b8d20e06eaad ("serial: 8250_uniphier: add earlycon support")
> >
> > diff --git a/drivers/tty/serial/8250/8250_uniphier.c b/drivers/tty/serial/8250/8250_uniphier.c
> > index bab6b3ae2540..1b7bd26555b7 100644
> > --- a/drivers/tty/serial/8250/8250_uniphier.c
> > +++ b/drivers/tty/serial/8250/8250_uniphier.c
> > @@ -35,7 +35,7 @@ struct uniphier8250_priv {
> >         spinlock_t atomic_write_lock;
> >  };
> >
> > -#ifdef CONFIG_SERIAL_8250_CONSOLE
> > +#if defined(CONFIG_SERIAL_8250_CONSOLE) && !defined(MODULE)
> >  static int __init uniphier_early_console_setup(struct earlycon_device *device,
> >                                                const char *options)
> >  {
> >
> >
> 
> 
> If my patch (https://patchwork.kernel.org/patch/8029331/) is accepted,
> this build error will be fixed as well.  But I doubt it is accepted soon...
> 
> So, I am OK with this workaround to fix "make allmodconfig".
> 
> Acked-by: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
> 

Good point. Adding the DT maintainers to Cc. Guys, any chance of you
picking up Masahiro's patch for 4.5?

It can't hurt to have both patches merged, so it would still be good
to have mine in the serial driver fixes, but we can live with either
of the two to fix the build.

	Arnd