Re: [PATCH 1/2] Documentation: devicetree: m25p80: add "nor-jedec" binding

[Brian Norris <computersforpeace-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>]

On Sun, Mar 22, 2015 at 12:30:05AM +0100, Rafał Miłecki wrote:
> On 20 March 2015 at 20:12, Brian Norris <computersforpeace-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> > Yes, and that is an eventual goal I suppose, but the current list is
> > excessive and is most likely not currently relied on by any one. So I
> > don't just want to C&P the entire list into this binding immediately.
> >
> > I guess my plan looks like this:
> >
> > 1. add "nor-jedec" binding, to provide lowest common denominator binding
> > (this series)
> >
> > 2. stop adding to the m25p_ids[] table unless necessary (enabled by this
> > series)
> >
> > 3. gauge whether we can remove certain entries from m25p_ids[] (e.g., if
> > they were only used in platform_data, not DT; or if they were very
> > recently added just to synchronize with spi-nor.c)
> 
> Why we can't remove (slowly) all entries from m25p_ids that don't need
> any extra handling?
> 
> If we'll have DT with
> "spansion,m25p80", "nor-jedec"
> and then m25p80.c will handle both: "m25p80" and "nor-jedec" without
> any difference, what's the point of keeping "m25p80" entry?

ABI stability. A lot of DTBs might be using m25p80 already, and they
aren't supposed to have to update just because SW decided to drop them.
But see [1] for the official word on ABI stability. I think that leaves
room for removing most/all of these eventually.

Brian

[1] Documentation/devicetree/bindings/ABI.txt
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html