[Openchrome-devel] 0.3.0-pre20120302 snapshot (was: Fix autoconf DRI detection (for Ubuntu Oneiric))

[James Simmons]

> > Give it a try now.
> > 
> 
> Very nice! It works!

I just created another snapshot.

xf86-video-openchrome-0.3.0-pre20120306.tbz

> It is still painfully slow though, compared to the 0.2.x version. It is
> on the limit of usable but I do have high requirements for "usable".

Currently EXA only works for hardware supported by the mainstream via
kernel drivers. With KMS mode it is also disabled. 
 
> The regression I previously reported are still there:
> http://wiki.openchrome.org/pipermail/openchrome-devel/2011-September/000578.html
> 
> I did noticed, that this happens only when I have compositing enabled
> in XFCE. If I disable compositing it works.

Yes I also tracked it down over the weekend. In the latest snapshot I 
disabled compositing. Most likely I will not fix it before the release.

> I also wonder if it would be an idea to get the kernel part into
> staging area in kernel. Then it might be possible to push it out for
> testing earlier. If not, is it easy to backport it to some of the
> current kernels?

I hate to push it just yet. I really like to have xorg driver out there
supporting KMS first. I have to say my current testing shows trunk works
well with drm-openchrome. To push the new kernel code I like to see LVDS
support and EXA being able to work correctly with the new GEM/TTM apis.