Re: [PATCH 1/2] mmc: add Device Tree properties for UHS modes

[Guennadi Liakhovetski <g.liakhovetski@gmx.de>]

On Mon, 29 Jul 2013, Pawel Moll wrote:

> On Mon, 2013-07-29 at 12:27 +0100, Guennadi Liakhovetski wrote:
> > On Mon, 29 Jul 2013, Pawel Moll wrote:
> > 
> > > On Mon, 2013-07-29 at 08:18 +0100, Guennadi Liakhovetski wrote:
> > > > A short addendum. At least with Renesas SoCs I see the situation in the 
> > > > following way: new SoC versions appear relatively frequently. 
> > > 
> > > What frequency are we talking about? Once per year? Once per month? I'm
> > > not trying to be picky, it really makes a difference...
> > 
> > Definitely not every month - not until now in the mainline at least. I 
> > currently count 9 SoCs, added since 2010, which makes about 2-3 SoCs per 
> > year.
> 
> So this is actually a slower rate that I've faced in my previous life
> working for a silicon vendor ;-) And my experience is that the IPs were
> different between the SoCs indeed but:
> 
> 1. Not all of them at the same time (so no extra compatible values for
> others).

No, that's exactly the problem. We absolutely do not want to write 
compatible="vendor,soc-v1"; in a .dts file for SoC v2 just because we 
currently think, that we don't have to distinguish between those SoC 
versions in this specific driver. I think we all know it quite well, that 
there are (practically) always differences - sometimes documented, 
sometimes undocumented. And if you later find out, that you do have to 
differentiate in the driver - it's too late. Even if we disregard the 
argument of ugliness of having to set compatibility with soc-v1, soc-v2, 
soc-v3 in different DT nodes on an SoC v4.

Thanks
Guennadi

> 2. When there was a change it required change in a driver as well (so
> adding a compatible value is not an issue).
> 3. The changes were *completely* unpredictable (so even comprehensive
> list of DT properties wouldn't help)
> 
> To summarize, count this as another vote for using compatible rather
> then "universal & future-proof" set of properties. Unless there is a
> very good rationale for it (I'm sure such cases exist)
> 
> Thanks!
> 
> Pawel
> 
> 

---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/