[PATCH] new radeon memory map fixes

[Benjamin Herrenschmidt]

> There should be no need to check for info->cursor_offset == 0 in the
> cursor functions. Longer term, I think we should just reserve a static
> FB region for the cursor upfront instead of going through all these
> hoops with EXA.

I had some dodgy stuff happening at things like
shutdown/entervt/leavevt, so I preferred being extra-safe there, though
they might indeed not be necessary.

> Also, unless I'm missing something, you're removing the code that forces
> the display priority to high for Radeon 7200.

Oops ? Where did I miss that ? (which bit of code specifically ? If it's
the hack that was in SetFBLocation, it's now in the bandwidth calc)
> 
> > http://gate.crashing.org/~benh/radeon-memmap-drm-3.diff
> 
> The way you handle backwards compatibility here is brilliant, thanks.
> The only minor issue I see is that the setparam ioctl can be called by
> unprivileged clients, but that applies to the existing colour tiling
> part as well, and it may not be a problem thanks to the offset fixups.

I'm not 100% sure yet of whta the clients may or may not do here, I'd be
very happy if you could double check that part :)

Ben.