Re: State of spidev from devicetree

[Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>]

Hi Claus,

On Tue, Oct 1, 2019 at 11:16 PM <claus.stovgaard@xxxxxxxxx> wrote:
> On tor, 2019-09-26 at 14:21 +0200, Geert Uytterhoeven wrote:
> > On Thu, Sep 26, 2019 at 12:50 PM Mark Brown <broonie@xxxxxxxxxx>
> > wrote:
> > > On Tue, Sep 24, 2019 at 02:52:59PM +0200, Claus Stovgaard wrote:
> > >
> > > > What is yours response to the idea of creating a custom-hardware
> > > > binding
> > > > for spidev, intended to be used for programmable hardware unknown
> > > > at the
> > > > devicetree time.
> > >
> > > You should use a device tree overlay to describe whatever
> > > hardware you've instantiated on your FPGA then load the overlay
> > > along with your FPGA image.
> >
> > And after that, you can bind spidev to the device in the overlay
> > using
> > sysfs, cfr. commit 5039563e7c25eccd ("spi: Add driver_override SPI
> > device
> > attribute").
>
> I know this is a bit old, but will still thank you for the replays.
>
> Regarding overlays and binding via sysfs. I understand your answer and
> proposal for solution. Though in our situation with a very basic
> embedded system it is just extra overhead and added complexity,
> compared to just having a spidev from initial devicetree. Without
> giving anything as I see it. So would still have preferred some-kind of
> custom string in the .compatible inside spidev.

You can have a custom compatible value in DT and in spidev, cfr.
drivers/spi/spidev.c:spidev_dt_ids[].
But plain "spidev" in DT is a no-go.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds