[PATCH/RFC] ARM: shmobile: dts: Add common file for MSIOF1

[Simon Horman]

On Fri, Nov 21, 2014 at 01:45:32PM +0900, Magnus Damm wrote:
> Hi Simon,
> 
> On Fri, Nov 21, 2014 at 1:27 PM, Simon Horman
> <horms+renesas at verge.net.au> wrote:
> > From: Hisashi Nakamura <hisashi.nakamura.ak at renesas.com>
> >
> > The R-Car Gen 2 SoCs have MSIOF1 interfaces however they
> > are not exposed on any boards currently supported in mainline.
> 
> Ok, but that's the same as many other devices on R-Car Gen2 SoCs no?
> Basically the SoC contains a hardware block but it is not used on the
> actual board. I wonder what made MSIOF1 a target for sudden code
> consolidation?
> 
> > The purpose of r8a77xx-msiof1.dtsi is to be included
> > in a dts file if MSIOF1 is exposed by some means.
> 
> Judging by the file name this seems to be a board specific property
> made generic somehow.

Yes, it is board stuff not SoC stuff.

> 
> > Signed-off-by: Hisashi Nakamura <hisashi.nakamura.ak at renesas.com>
> > [horms: Reworked into a separate dtsi file]
> > Signed-off-by: Simon Horman <horms+renesas at verge.net.au>
> 
> [snip]
> 
> > +&pfc {
> > +       msiof1_pins: spi2 {
> > +               renesas,groups = "msiof1_clk_c", "msiof1_sync_c", "msiof1_rx_c",
> > +                                "msiof1_tx_c";
> > +               renesas,function = "msiof1";
> > +       };
> > +};
> > +
> > +&msiof1 {
> > +       pinctrl-0 = <&msiof1_pins>;
> > +       pinctrl-names = "default";
> > +
> > +       status = "okay";
> > +};
> 
> The DTS above only works on boards that are using the "C" set of pins
> for the MSIOF1. Other configurations will not work. It also assumes
> that all SoC code is using the same strings for the PFC which may or
> may not be true.
> 
> Does some common policy for the Linux kernel exist about how to break
> out board specific device support in DTS? If so we should follow it.
> 
> If not, from my conservative view point this patch looks like over
> doing code sharing with increased overhead and no apparent upside
> except more complex dependencies.

Perhaps it would be best just to drop it.

The patch that this code was based on illustrates how to enable
MSIOF1 on the koelsch. People can find that and use it if they
wish to use MSIOF1.