Re: Device tree BSP

[Grant Likely <grant.likely-s3s/WqlpOiPyB63q8FvJNQ@public.gmane.org>]

On Sun, Jul 5, 2009 at 4:10 PM, Benjamin
Herrenschmidt<benh-XVmvHMARGAS8U2dJNN8I7kB+6BGkLq7r@public.gmane.org> wrote:
> On Sun, 2009-07-05 at 15:53 -0600, Grant Likely wrote:
>> This is not a question that I've got an answer to, but
>> I've been thinking about defining a "machine" node, or something
>> similar that can be a parent of such non-addressable devices.  I'm
>> hoping Mitch, Ben, DavidG, Scott, or some other OF active folks will
>> jump in here and give their opinion before I go and define something
>> braindead.  I was thinking something like this, where the root node
>> would be the parent of this machine node:
>
> Well, device nodes can perfectly represent non addressable devices...
> One way is to have no "reg" property but that sucks as you get no unit
> address neither which can cause problems when you have multiple of them.
>
> So hosting them under some kind of node that would break that address
> translatability (have no "ranges" property) makes some amount of sense
> though I don't necessarily like the term "machine".

Neither do I, but I've got nothing better.  I'm open to suggestions.  :-)

> But if no "reg" is good for you, then stick them in the soc node with
> distinct names.

It doesn't really belong in the soc node either because it is all
stuff external to the SoC.

> Also, why "virtual," in your compatible property for the SPI node ? It's
> not really virtual...

I don't like 'virtual,' either, but its in the currently documented
binding.  (Documentation/powerpc/dts-bindings/gpio/mdio.txt).

g.

-- 
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.