4 May
2009
4 May
'09
22:27
Round 2. Alex managed to find the root of the easy cursor corruption issue: the cursor mode should be the same for both CRTCs even when disabled. This is 0001. 0002 is the refactoring. It's almost trivial now that we don't need the logic to keep the cursor enabled on both CRTCs. 0003 contains the multiple-of-128px and outside-viewport code with some tweaks, and is thus different from radeon. My system that's set up with dual-screen doesn't seem to be panning capable, so the viewport part is untested; please scrutinize 0003 with extra attention. -- Yang Zhao http://yangman.ca