Daniel Vetter <dan...@ffwll.ch> writes: > On Tue, Apr 02, 2013 at 06:46:32PM -0700, Eric Anholt wrote: >> This still fails, since 8192*4bpp == 32768, which is too big to use the >> blitter on. > > Meh, I should take coffee first before doing reviews ... > > Iirc on gen4+ for tiled src/dst the blt stride limit is 128k (since the > stride is supplied in dwords instead of bytes).
Good point, I could be more look just at the pitch that we would make for the temporary BO, except that our interface for calling the blitter is uint16_t bytes, so 128k would fail anyway. Something to fix later imo. > Also, with the new > snoopable bo support in the kernel v3.7 you should be able to use the > blitter also on non-llc platforms (safe for i965g/gm, the hw is busted > there), which will cut away the clflushing when accessing the cpu maps. > -Daniel Yeah, we'd need to get that hooked up correctly in libdrm so that non-llc snoopable BOs don't get returned to the BO cache.
pgpgCcw6jlqGV.pgp
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
