On Mon, 2009-10-05 at 14:42 +1000, Dave Airlie wrote: > On Mon, 2009-09-28 at 10:13 +0200, Michel Dänzer wrote: > > On Mon, 2009-09-28 at 14:57 +1000, Dave Airlie wrote: > > > From: Dave Airlie <[email protected]> > > > > > > This basically stops any accel to non screen pixmap and non > > > window pixmaps when this option is enabled. It makes my > > > RN50 at least useable again under KMS. > > > > > > I'm mainly producing this to open discussions on whether > > > there is a better way. I'm defining low VRAM as < 32MB. > > > > There has to be a better way in the long run, but in the meantime this > > could be handled in the driver PixmapIsOffscreen hook? > > > > http://people.freedesktop.org/~airlied/scratch/radeon-return-FALSE-offscreen.patch > > I tried this, it fails, server crashes doing a composite operation in > pixman. > > maybe you need to explain a bit better why you think this should work, > as to me PixmapIsOffscreen seems like something you call possibly > after migrating the pixmap.
Right, I was thinking of the code in exaFinishAccess_mixed which checks IsOffscreen first, but of course that doesn't apply here. Might it be feasible to return NULL from the CreatePixmap(2) hook for non-screen pixmaps (or whichever ones you don't want in VRAM)? AFAICT the mixed pixmaps code should basically handle that, though maybe we should change exaCreateDriverPixmap_mixed to mark the pixmap as pinned to prevent repeated calls of the CreatePixmap(2) hook. -- Earthling Michel Dänzer | http://www.vmware.com Libre software enthusiast | Debian, X and DRI developer _______________________________________________ xorg-devel mailing list [email protected] http://lists.x.org/mailman/listinfo/xorg-devel
