On Wed, 23 Mar 2011 10:23:46 -0700, Eric Anholt <e...@anholt.net> wrote: > On Tue, 22 Mar 2011 13:51:58 +0000, Chris Wilson <ch...@chris-wilson.co.uk> > wrote: > > Signed-off-by: Chris Wilson <ch...@chris-wilson.co.uk> > > Again, needs more justification. For example, I would expect us not to > bother with the GTT on SNB pread/pwrite, and just read the backing pages > (they're in our cache, no need to clflush or anything) like we already > have code to do. And one note below.
Right, so snooped pages goes through the shmem paths, nothing changes. Not everything is as nice as SNB and keeping the pages in the GTT is a benefit for everything else. [snip] > So, previously if I pread a tiled object, I got the tiled data. Now > you're giving me maybe the tiled data, maybe the data untiled by reading > through a fence. No, it clears the fence before reading to maintain the existing API that isn't that useful in current practice. -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
