On Fri, Apr 3, 2009 at 11:56 AM, Kristian Høgsberg <k...@bitplanet.net> wrote: > 2009/4/2 Eric Anholt <e...@anholt.net>: >> On Thu, 2009-04-02 at 14:41 -0400, Kristian Høgsberg wrote: >>> This was only used by the i915 driver for the swapbuffer vsync tasklet. >>> That functionality is now gone and nothing uses the kernel side >>> cliprects anymore. Just stub out the ioctls, but make sure we return >>> a non-zero handle in the DRM_IOCTL_ADD_DRAW case. >> >> Have you tested old userland to make sure it doesn't freak out? Other >> than that, I love the idea. > > I did, ran gears, resized it, and it worked. glxgears is even less of > a test suite than it is a benchmark, of course, bu int this case I do > believe it tests the code paths in question. The only gotchas would > be if some userspace code relies on updating or removing an invalid > drawable handle to return -EINVAL or relies on adddraw returning > unique drawable handles. In my reading of the xf86dri, aiglx and > libgl, I've not come across any such requirement.
Do you want to pick it up then or do you want Dave to take it? thanks, Kristian ------------------------------------------------------------------------------ This SF.net email is sponsored by: High Quality Requirements in a Collaborative Environment. Download a free trial of Rational Requirements Composer Now! http://p.sf.net/sfu/www-ibm-com -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel