Closing X Developer Summit 2008 Notes
Phoronix: Closing X Developer Summit 2008 Notes
Aside from the short X Server 1.6 release plans and clarifying UXA+EXA, there are a few other notes to share from this afternoon's X.Org/Intel talk. - This shouldn't come as a surprise, but the release of xf86-video-intel 2.5 driver is planned for later this month. The first release candidate of this new open-source Intel driver is already available, but more test releases are expected before it's officially released...
So he's cleaning things up in X, switching it to time-based releases, and using git.
Things are about to get awesome. I'd love to see X release every 3 months... don't know whether it'd be a good idea to sync releases with The Kernel or not though..
I thought that X already had an informal time based release? Did not stop them from cutting and delaying this release.
It looks like xserver 1.6 will be an intel release...I wonder which of their customers was so anxious for the new intel features, or, maybe they're just afraid that if they wait for the next xorg release then their features will only appear in 2010 ...
For what it's worth, the proposed 1.6 plan touches on all of the features which are most important for graphics. If I were planning an "AMD/ATI release" it would be exactly the same as what Keith is proposing.
- DRI2 is critical for flicker-free 3D under Compiz (or video playback through GL)
- RandR1.3 is critical for multi-card systems (RandR1.2 can't handle >1 GPU)
I don't know about the input stuff but based on the graphics choices this seems like a good plan.
Last edited by bridgman; 09-05-2008 at 12:22 PM.
This might not work so well on multi-carded systems. It was brought up and Keith said he will not work on any multi-GPU testing at this time since Intel's is only integrated, but he did explain to others the work that would be left in making it all proper. Though he is going to start on GPU hotplugging support it sounds like for the dual GPU notebooks.
Originally Posted by bridgman
Yep, that makes sense. If nobody had time to work on RandR1.3 in the near future we were going to try to push it ahead ourselves (Matthias was already mentioned) so at the very least we should be able to help on the multi-GPU side.
There are a lot of laptops floating around with Intel integrated graphics *and* AMD/ATI discrete graphics so I think everyone would be interested.