[PATCH v2 00/18] OMAP4: PM data big spring cleanup and fixes

[Tony Lindgren]

* Rajendra Nayak <rnayak at ti.com> [110708 10:17]:
> On 7/8/2011 12:11 AM, Paul Walmsley wrote:
> >On Thu, 7 Jul 2011, Martin Fouts wrote:
> >
> >>From: Tony Lindgren [tony at atomide.com]
> >>
> >>>The second problem we have here is "why does adding 4460 support depend
> >>>on a cosmetic clean-up patch". That dependency should not exist at all
> >>>as it seems the 4460 patches should work even without this patch.
> >>
> >>I agree. Had the original submitter had the foresight to realize that
> >>the code should work for all 44xx family processors, we would have no
> >>issue at all.
> >
> >Uhhh...
> >
> >The original 4430 data, which is mostly what we're talking about here, was
> >added in 2009.  Maybe a few people inside TI knew what was going to change
> >and what was going to be the same for future OMAP4 parts.  But even if
> >someone did know, the decision of what to call a chip often isn't up to
> >engineers, it's up to marketing, which picks whatever name they like.
> >You know, like Linux 2.6.40^H^H^H^H^H^H3.0.
> >
> >So back in 2009, the submitter and maintainers were faced with a choice:
> >
> >Option 1. Submit patches with facts.  "OMAP4430".  Don't speculate what
> >future, as-yet-nonexistent products will be numbered, and what their
> >feature set will be.  Plan to generalize later once it is known exactly
> >what needs to be generalized.
> >
> >Option 2. Try to predict what marketing will call the next chip, and what
> >features will still be present, then put that into the codebase.
> >"OMAP44XX".  Hope you guess right so you don't have to change them all if
> >marketing or engineering comes up with something different.
> >
> >So, what's the right answer?
> >
> >I probably can't tell you that, but I can tell you that in 2009, option 1
> >seemed more technically conservative.  So that's what we did.  Maybe that
> >isn't the right answer, though.
> 
> I completely agree with Paul. What if the 4460 of today was called
> 4640 for some reason. (Well we do have a 3630, don't we)
> Would the OMAP44XX work then, no. We would need a OMAP4XXX instead.

The right way to fix issues like this is just to mark things omap2,
omap3, omap4 and so on. And to initialize the device list based on the
SoC detection. There really should not be any need to repeat the CHIP_IS
flag in all devices.

Regards,

Tony