2010/5/27 Michel D?nzer <michel at daenzer.net>: > On Die, 2010-05-25 at 19:09 -0400, Alex Deucher wrote: >> >> Also, on r6xx/r7xx, we don't enable tiling of the front buffer as >> that would require the use of wfb or tiled to untiled blits for CPU >> access. Tiled to untiled blits works (I've tested it), but shows an >> approximately 40% performance hit for software fallbacks (tested with >> non-aa text in x11perf). ?I'm not sure what how that compared to sw >> detiling with wfb since the algo is pretty complex. ?In practice, I >> didn't notice any difference in performance on the desktop. > > How exactly did you implement the (un)tiling blits? I'd suggest doing it > similarly to the way we're handling byte swapping with KMS on big endian > machines: return FALSE from the PrepareAccess hook and handle the > (un)tiling in the DownloadFrom/UploadToScreen hooks. >
I used that exact method. > In my experience, wfb is excruciatingly inefficient (one callback per > byte in the worst case). It's horrible. There's also no way to handle multiple cards sanely as there's nothing to attach a driver priv to. Alex > > > -- > Earthling Michel D?nzer ? ? ? ? ? | ? ? ? ? ? ? ? ?http://www.vmware.com > Libre software enthusiast ? ? ? ? | ? ? ? ? ?Debian, X and DRI developer >
