Re: [PATCH] spi: tegra114: add spi driver

[Stephen Warren <swarren@wwwdotorg.org>]

On 02/20/2013 10:31 AM, Mark Brown wrote:
> On Wed, Feb 20, 2013 at 10:25:13AM -0700, Stephen Warren wrote:
> 
>> But, please do think this approach through fully. The DT binding
>> needs to define which clock-names the driver requires to be
>> present, and any optional clock names. DT bindings are supposed
>> to be immutable, or perhaps extendible in a completely
>> backwards-compatible fashion. This implies that you need to have
>> thought through the entire list of clocks that the driver might
>> want in the DT clock-names property when you first write the DT
>> binding documentation...
> 
> Since we can extend the list of clocks it doesn't seem like there's
> much issue here, especially if some of them are optional?

Yes, there's certainly a way to extend the binding in a
backwards-compatible way.

However, I have seen in Rob and/or Grant push back on not fully
defining bindings in the past - i.e. actively planning to initially
create a minimal binding and extend it in the future, rather than
completely defining it up-front.

I don't know how strong of a rule they intend that to be though. If we
get to the point of moving the DT bindings out of the kernel, it'd be
good to get a concrete definition of what can and can't be changed in
bindings.

> Though in general it seems like this sort of mux really should be
> in the clock stuff anyway.

How do you see that working: something automatic inside clk_set_rate()
seeing that some other parent could provide the rate, so the clock
could be reparented, or ...?