Re: [PATCH v2 1/9] dmaengine: st_fdma: Add STMicroelectronics FDMA DT binding documentation

[Lee Jones <lee.jones@linaro.org>]

On Sat, 12 Sep 2015, Peter Griffin wrote:

> Hi Arnd,
> 
> On Fri, 11 Sep 2015, Arnd Bergmann wrote:
> 
> > On Friday 11 September 2015 15:14:23 Peter Griffin wrote:
> > > +- st,fdma-id	: Must contain fdma controller number
> > 
> > What for?
> 
> It is used by the driver to generate a unique firmware name.
> Basically we need to know which controller instance we
> are as each controller has a different firmware which needs
> to be loaded.
> 
> Rob did say that having a index type property is undesirable
> over here, see my reply at the bottom
> http://www.spinics.net/lists/devicetree/msg92529.html.
> 
> However I can't think of any other useful properties we could add
> to derive this information.

I wouldn't use a property at all.  Why not use the compatible string?

Who chooses the naming scheme of the firmware binary?

Is there any reason they can't be:

  fdma_STiH407_audio.elf
  fdma_STiH407_app.elf
  fdma_STiH407_free_running.elf

Then you can have a different compatible for each:

  compatible = "st,stih407-fdma-mpe31-audio";
  compatible = "st,stih407-fdma-mpe31-app";
  compatible = "st,stih407-fdma-mpe31-free-running";

[...]

> > > +Example:
> > > +
> > > +	snd_uni_player2: snd-uni-player@2 {
> > > +		compatible = "st,snd_uni_player";
> > 
> > Maybe change the compatible string to "st,snd-uni-player"?
> 
> Just checking the upstream driver which got merged for v4.3
> and the compatible has changed to "st,sti-uni-player" so I
> can update the DT doc example to use this in v3.
> 
> > 
> > The string you use here would not be acceptable for a binding.
> 
> Presumably it is the use of "underscores" in the compatible which
> is not allowed? Is that correct?

Yes.

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog