Request For Testing - Intel Graphics Drivers in -current
SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
FWIW, it seems like KMS is the culprit regardless of versions of intel-driver and kernel.
I have tried all intel-drivers from your site, together with kernel 2.6.29.{5,6}, 2.6.30.1 and 2.6.31-rc2.
As long as I have latest mesa it all works great without KMS.
However, with KMS enabled, the intel driver causes a crash in X on KDE 4.2.4 logout, forcing me to switch to console (which is very fast though :-) ) , do a telinit 3;sleep 5;telinit 4;logout before I can login again.
I see that xorg-server 1.6.2 has a few memory corruption fixes, but haven't dared testing it in fear of contaminating the test-data...
Has anyone else got KMS to work properly with the GM45 chipset on laptop?
Reason why I'm super-keen on KMS in 2.6.31 when it's released is the DisplayPort support so I can finally ditch the pre-historic RGB-cable to the brand new monitor :-)
Thanks again for your great work here - while it's not yet a 100% for all new features, it's come a long way!!
OK - some more news - now I got it to be relatively stable on my end.
Current setup is that I have KMS enabled (needed for my DisplayPort monitor), kernel 2.6.31-rc2 and xf86-video-intel-2.8rc1.
Other configurations that work include kernels 2.6.29.6 and 2.6.30.1, all with KMS and xf86-video-intel-2.8-beta from Robby's test-area.
add TerminateServer=true
to /etc/kde/kdm/kdmrc under [X-*-Core]
DISCLAIMER: It's been working nice for two days only, and with the latest intel-RC-driver, only for a few hours. This was the last issue however resolved by the above fix, so for me I cross my fingers the X11-woes are now over for this time
-y1 (now happily running my first-ever RC-kernel in "production-mode" with kernel modesetting and working X :-) )
Intel Releases xf86-video-intel 2.8 RC Driver
Posted by Michael Larabel on July 13, 2009
In preparation for Intel's quarterly DDX driver update, Carl Worth has announced the first release candidate of the xf86-video-intel 2.8 driver. This driver is significant in that it completely abolishes DRI1 and EXA support in favor of only supporting DRI2 and UXA, respectively.
The UMA Acceleration Architecture is derived from EXA and was supposed to be merged back into EXA, but that didn't end up happening. UXA basically takes the EXA API but internally now uses the Graphics Execution Manager for managing the memory. As our UXA benchmarks have shown, UXA is faster than EXA on the Intel Linux graphics stack, but it isn't yet completely stabilized and is rather common for people to run into screen artifacts, stability issues, and other problems. The Intel driver is currently the only one implementing UXA for its means of 2D acceleration.
In the xf86-video-intel 2.8 release candidate are various fixes (in particular, some for X-Video, UXA, kernel mode-setting, and various other areas throughout) along with support for their unreleased, next-generation chipset. Support for this unreleased Intel IGP is already in the Linux kernel and the DDX driver has picked up a new shader compiler and other features for this chipset simply known as "IGDNG" right now, or Intel Graphics Device Next Generation.
The complete list of changes for the first release candidate in the Intel X.Org 2.8 driver series along with a source download link is available from the intel-gfx mailing list. The final release of the xf86-video-intel 2.8.0 driver will likely come this month or next.
I got KMS working with the Intel-2.7.99-901 package and a 2.6.30 kernel. Had to take out the vga line in my lilo, though. Not sure if there was another method or not to getting it started for the terminals.
Hardware: Asus EeePC 901 with Intel Corporation Mobile 945GME
Software: uptodate Slackware-current
Kernel 2.6.30.1-smp without KMS and xf86-video-intel-2.7.1
glxgears works.
no problems with truncated window-decorations, buttons etc.
switching to virtual terminals works.
Kernel 2.6.30.1-smp without KMS and xf86-video-intel-2.7.9.901 and 902
glxgears does not work: truncated output, but frame counter works.
no problems with truncated window-decorations, buttons etc.
switching to virtual terminals works.
Kernel 2.6.30.2-smp with KMS and xf86-video-intel-2.7.1
glxgears works.
no problems with truncated window-decorations, buttons etc.
switching to virtual terminals does not refresh the display content, so nothing can be done until switching back to x.
Kernel 2.6.30.2-smp with KMS and xf86-video-intel-2.7.9.901 and 902
glxgears does not work: truncated output, but frame counter works.
no problems with truncated window-decorations, buttons etc.
switching to virtual terminals works.
With each combination i need to manually call init 4 to use X. If i boot into init 4 directly i do not get any display output and only the power button shutdowns (is this word correct?) the system. This problem did not appear with X.org 1.4.2 and appears since X.org 1.6.1.
Xorg.log says exactly nothing and it looks like the start of X runs perfectly. Maybe a timing problem?
Hate to nag, just putting it out there that Intel released the 2.8 drivers. Is it possible for a new package to be put out there in /extra?
For those who can't wait for an official Intel driver version 2.8.0, I put up my highly unofficial version on http://hiawata.com/linux/
It's built with Pat's build script, just replacing the old source file with the new one.
I've been using it myself since it came out, together with kernel 2.6.31rc4 and apart from compositing still not working (see previous posts in this thread) the stability is actually great and also works with DisplayPort :-)
Based on what I've read from the xorg mail list and various bug reports and feedback from intel guys, I think the 2.8.0 driver pretty much requires mesa-7.5 and an xorg-server-1.6.2 compiled against that mesa release. At this stage of the game, I suspect Pat is going to be *VERY* resistant to doing that. I personally am starting to think that we should have stayed with xorg-server-1.6.1...
I probably need to go back and read xorg-devel again, because from my understanding the 2.8 driver requires new xserver and friends, but as far as I can see not mesa 7.5.
I've been running it on vanilla -current since it came out, and it does appear more stable than the previous release candidates. It only works with UXA though, and is significantly more stable on kernel 2.6.31rc than 2.6.29.
Having said that, on vanilla -current, I still don't get any compositing or nice effects since the last X-related update in current. Downgrading only pixman didn't help
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.
