On Thu, Jul 05, 2012 at 11:29:10AM +0100, Chris Wilson wrote: > On Thu, 5 Jul 2012 12:17:30 +0200, Daniel Vetter <daniel.vetter at ffwll.ch> wrote: > > The tileoffset register only supports a limited offset in x/y of 4096, > > so for giant screen configuration with a shared fb we wrap around. > > > > Fix this by computing a linear offset in tiles (pages) and only use > > the tileoffset register to offset within the tile. > > > > Signed-Off-by: Daniel Vetter <daniel.vetter at ffwll.ch> > Both Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk> Thanks for the review, I've addressed your pitch concern from the other mail by merging one of Ville's patches. > We could do with validating that the CRTC is within the fb->obj, > otherwise we run foul of hanging the hardware. (Since this is a > pre-existing condition it doesn't mar these patches.) > > So the only question is whether we indicate to userspace that the kernel > is fixed? Or just kill the w/a in userspace and for reasons of sanity > strongly encourage everyone who hits this to upgrade? Since this is no > stable material, having keeping the w/a seems to make sense... As discussed I think we'll just ask ppl to upgrade their kernel and should rip out the hack in userspace ... -Daniel -- Daniel Vetter Mail: daniel at ffwll.ch Mobile: +41 (0)79 365 57 48
