[PATCH] mtd: physmap_of: Add "map-indirect" DT property

[Stefan Roese]

On Monday 16 July 2012 22:41:52 David Woodhouse wrote:
> On Sat, 2012-07-07 at 09:16 +0200, Stephan Gatzka wrote:
> > this patch works very well for us (of course, because Stefan made it for
> > us). Are there any concerns against it? Probably against introducing a
> > new DT property? Of course we can make this fix much more MPC5200
> > specific but this will result in largely copying physmap_of.c and some
> > additional code for access routines and Kconfig.
> 
> I don't much like the "map-indirect" name. If it's actually unaligned
> access, or non-word-sized access, that's forbidden, then that's what the
> DT property should be.
>
> The term "map-indirect" is more a description of how the software
> currently behaves... which is exactly that DT bindings *shouldn't* be.

Okay. I've chosen "map-indirect" because it might be used by other platforms 
as well, perhaps because of different reasons (so not restricting this to the 
unaligned access problem of the MPC5200). But I have no strong feelings here, 
so I can prepare a new patch version with a different name.

How about "no-unaligned-direct-access"? Pretty long though. Any other 
suggestions here?

Thanks,
Stefan