Felipe Reyes wrote: > On Thu, 2008-02-07 at 19:13 +0100, Xavier Bachelot wrote: >> Andrey Melentyev wrote: >>>> Okay, in SVN AGP DMA is disabled by default now on the K8M800. >>>> Let's wait for Xavier to include this in a 0.2.902 release. >>> Its not only K8M800 but VN800 too. I thought I did mention it >>> somewhere in ticket comments. >>> >> I'm not particularly happy with the AGPDMA change for K8M800, it used to >> work and I believe the pb is not in openchrome code. Andrey's follow-up >> is the perfect example of what will happen next if we follow this path, >> disable AGPDMA for VM800. And all chipsets if we continue on this road... >> I suspect the problem could be in the kernel (either AGP or DRM), but as >> I'm not affected by the pb, I can't try to narrow it down. As it seems >> pretty easy to trigger the bad behaviour for those affected, it would be >> really interesting to have someone try with various kernel release, say >> from 2.6.18 up to 2.6.23 and see if/when the pb disappear. >> > I own a K8M800 and when i have AGPDMA enabled the unichrome_dri.so > produce the hang up of my machine (in debian sid provided by > libgl1-mesa-dri). This behaviour i experimented with 2.6.20, 2.6.21, > 2.6.22 and 2.6.23 (kernels compiled by myself and also the debian > packaged kernels) > The hangs with 3D on K8M800 are well known and probably unrelated to AGPDMA being enabled or not. Can you reproduce the hang with AGPDMA enabled and unichrome_dri.so desactivated (rename it or whatever) ?
To follow up on the VM800 problem, it seems it is a screen repaint problem, not a hang problem. Andrey, can you please confirm what affected you, hard lockup or screen repaint ? Regards, Xavier _______________________________________________ openchrome-users mailing list [email protected] http://wiki.openchrome.org/mailman/listinfo/openchrome-users Main page: http://www.openchrome.org Wiki: http://wiki.openchrome.org User Forum: http://wiki.openchrome.org/tikiwiki/tiki-view_forum.php?forumId=1
