[PATCH for 3.0] AT91: Change nand buswidth logic to match hardware default configuration

[Arnd Bergmann]

On Monday 04 July 2011, Nicolas Ferre wrote:
> > In this case, studying the patch more closely shows that it's
> > very harmless, but I'd rather not have to look that closely.
> 
> Well, in fact it is a fix against what was introduced in a 3.0 patch
> which I found to be wrong.
> The reason because I do not want to be in next kernel is the fact that
> it can puzzle the user (people that want to use kernel without changing
> the system_rev between 2.6.39 -> 3.0 and again revert their changes for
> 3.0 -> 3.1).
> 
> > Am I correct that the bug is a regression against 2.6.39?
> 
> No, in fact it was introduced during 3.0 early -rc.

That's what I mean by 'regression against 2.6.39': it was working in 2.6.39,
but later kernels are broken without this fix.

	Arnd