[PATCH 06/12] drm: fix __alpha__ PCI lookup

[Daniel Vetter]

On Wed, Jul 23, 2014 at 09:29:48PM +0200, Daniel Vetter wrote:
> On Wed, Jul 23, 2014 at 05:26:41PM +0200, David Herrmann wrote:
> > Testing the return value of list_entry() for NULL is a no-op (as it is
> > just a fancy container_of() / offsetof()). Drop the superfluous if-clause
> > and instead verify the actual root-node is available. This is probably
> > what it was meant to test for from the beginning, anyway.
> > 
> > Signed-off-by: David Herrmann <dh.herrmann at gmail.com>
> 
> Oh drm handling of alpha. I've tried to dig out how this all works
> together and it's dangerous. drm totally ignored this thing called dma api
> so needed to handle cpu phys address vs. bus address offsets itself. If
> you munge around also in the agp code (which is really only used by drm)
> then you'll notice that it's handled completely differently on alpha vs.
> ppc.
> 
> Tbh I'd wait until this is all officially dead and then rip it all out.

I dare you for a patch to rip out alpha drm handling completely though ;-)
ppc probably needs to stay a bit longer, there's still lots of apple
hardware with radeons around.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch