Re: [PATCH v2 07/15] ARM: shmobile: Basic r8a7793 SoC support

[Simon Horman <horms@verge.net.au>]

Hi Laurent,

On Wed, May 06, 2015 at 07:20:05AM +0300, Laurent Pinchart wrote:
> Hi Ulrich,
> 
> Thank you for the patch.
> 
> On Thursday 30 April 2015 16:18:56 Ulrich Hecht wrote:
> > Minimal support without power management or SMP.
> > 
> > Signed-off-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
> > ---
> >  arch/arm/mach-shmobile/Kconfig         |  4 ++++
> >  arch/arm/mach-shmobile/Makefile        |  1 +
> >  arch/arm/mach-shmobile/setup-r8a7793.c | 33 +++++++++++++++++++++++++++++++
> >  3 files changed, 38 insertions(+)
> >  create mode 100644 arch/arm/mach-shmobile/setup-r8a7793.c
> > 
> > diff --git a/arch/arm/mach-shmobile/Kconfig b/arch/arm/mach-shmobile/Kconfig
> > index 0fb4842..140b696 100644
> > --- a/arch/arm/mach-shmobile/Kconfig
> > +++ b/arch/arm/mach-shmobile/Kconfig
> > @@ -80,6 +80,10 @@ config ARCH_R8A7791
> >  	select ARCH_RCAR_GEN2
> >  	select I2C
> > 
> > +config ARCH_R8A7793
> > +	bool "R-Car M2-N (R8A7793)"
> > +	select ARCH_RCAR_GEN2
> > +
> >  config ARCH_R8A7794
> >  	bool "R-Car E2 (R8A77940)"
> >  	select ARCH_RCAR_GEN2
> > diff --git a/arch/arm/mach-shmobile/Makefile
> > b/arch/arm/mach-shmobile/Makefile index 89e463d..e4b8fdb 100644
> > --- a/arch/arm/mach-shmobile/Makefile
> > +++ b/arch/arm/mach-shmobile/Makefile
> > @@ -13,6 +13,7 @@ obj-$(CONFIG_ARCH_R8A7778)	+= setup-r8a7778.o
> >  obj-$(CONFIG_ARCH_R8A7779)	+= setup-r8a7779.o pm-r8a7779.o
> >  obj-$(CONFIG_ARCH_R8A7790)	+= setup-r8a7790.o
> >  obj-$(CONFIG_ARCH_R8A7791)	+= setup-r8a7791.o
> > +obj-$(CONFIG_ARCH_R8A7793)	+= setup-r8a7793.o
> >  obj-$(CONFIG_ARCH_R8A7794)	+= setup-r8a7794.o
> >  obj-$(CONFIG_ARCH_EMEV2)	+= setup-emev2.o
> >  obj-$(CONFIG_ARCH_R7S72100)	+= setup-r7s72100.o
> > diff --git a/arch/arm/mach-shmobile/setup-r8a7793.c
> > b/arch/arm/mach-shmobile/setup-r8a7793.c new file mode 100644
> > index 0000000..1d2825c
> > --- /dev/null
> > +++ b/arch/arm/mach-shmobile/setup-r8a7793.c
> > @@ -0,0 +1,33 @@
> > +/*
> > + * r8a7793 processor support
> > + *
> > + * Copyright (C) 2015  Ulrich Hecht
> > + *
> > + * This program is free software; you can redistribute it and/or modify
> > + * it under the terms of the GNU General Public License as published by
> > + * the Free Software Foundation; version 2 of the License.
> > + *
> > + * This program is distributed in the hope that it will be useful,
> > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> > + * GNU General Public License for more details.
> > + */
> > +
> > +#include <linux/init.h>
> > +#include <asm/mach/arch.h>
> > +
> > +#include "common.h"
> > +#include "rcar-gen2.h"
> > +
> > +static const char *r8a7793_boards_compat_dt[] __initconst = {
> > +	"renesas,r8a7793",
> > +	NULL,
> > +};
> > +
> > +DT_MACHINE_START(R8A7793_DT, "Generic R8A7793 (Flattened Device Tree)")
> > +	.init_early	= shmobile_init_delay,
> > +	.init_time	= rcar_gen2_timer_init,
> > +	.init_late	= shmobile_init_late,
> > +	.reserve	= rcar_gen2_reserve,
> > +	.dt_compat	= r8a7793_boards_compat_dt,
> > +MACHINE_END
> 
> How about moving that to setup-rcar-gen2.c instead ? There's nothing specific 
> to r8a7793 here. We could even introduce a "renesas,rcar-gen2" DT compatible 
> entry and match on that instead of "renesas,r8a7793" (which should of course 
> be specified as well in DT). You could also add "renesas,r8a7794" to the 
> setup-rcar-gen2.c compat list and remove setup-r8a7794.c.

I believe there are two issues here:

1. Consolidating code

   From my point of view I think this is an excellent idea.
   However I would rather add this patch as-is and then
   address consolidation later.

   This avoids blocking on consolidation and I believe we have
   a pretty good record of coming back and doing consolidation
   (at least in arch/arm/mach-shmobile/) that has been suggested
   as a result of adding support for new SoCs.

2. A "renesas,rcar-gen2" DT compatible entry

   This seems to be a separate issue that depends on 1.
   I would like to treat it separately.