Bugzilla – Bug 555619
intel [855 GM] X-Server freezes
Last modified: 2010-09-28 13:44:52 UTC
User-Agent: Mozilla/5.0 (X11; U; Linux i686; de; rv:1.9.1.4) Gecko/20091016 SUSE/3.5.4-1.1.2 Firefox/3.5.4 The display mostly freezes during the user selection or shortly after login. The mouse can still be moved. Now, my system only works with the failsafe mode. The problem is probably related to the sometimes discussed intel driver problem. Please let me know which logs etc. are necessary to solve the prob. HArdware: Laptop Fujitsu Siemens M7400 Graphics: Intel 855 GM THANKS! D. Reproducible: Always Steps to Reproduce: 1. Wait for the user selection display / shortly after 2. 3. Actual Results: Display freezes, only hardware switch stops the computer.
Please reboot once into normal mode (Xserver freeze). Then reboot again into failsafe mode, log in and attach /var/log/Xorg.0.log.old (the log file for the Xserver freeze).
Created attachment 327603 [details] as requested
Created attachment 327605 [details] as requested, vesa mode
The first posted log is with the devide mode "intel" in the xorg.conf, the second one with the vesa mode. Thanks!
Which desktop are you using? KDE4, Gnome, Xfce? Something else?
I use KDE4, but the display sometimes freezes even before I have logged in, i.e. before the desktop is chosen. If you need any log data other than posted before, just let me know! Thanks for your help, D.
I just wanted to know which desktop to install, so I can reproduce that issue. :-)
FYI: Actually, I already had to use the text mode installation since the normal mode froze on the "first configuration"-screen...
I can't reproduce that issue on my 855GM machine (HP/Compaq nx5000). Even KDE4 desktop effects work. I'm afraid such driver issues need to be addressed by Intel's Linux driver development team. Thus I would like to ask you to report this issue upstream on bugs.freedesktop.org according to http://intellinuxgraphics.org/how_to_report_bug.html and add the following Bugzilla accounts to Cc: eich@pdx.freedesktop.org kent.liu@intel.com mhopf@suse.de quanxian.wang@intel.com sndirsch@suse.de You'll need to register for an account on bugs.freedesktop.org first, in case you don't have one yet. Thanks.
Dear Stefan, I looked around in the Intel bug forum and found some bugs like https://bugs.freedesktop.org/show_bug.cgi?id=24789 which seem pretty close to my prob. Though, they dont seem to be solved yet. I will tell you when I got interesting news. Thanks! D.
Thanks a lot.
As reported in this bug report (on 11.3 M6) https://bugzilla.novell.com/show_bug.cgi?id=593463 there may be a fix for this as noted in this bug report in post#61 : https://bugs.freedesktop.org/show_bug.cgi?id=26345#c61
reopen.
Patch needs to be upstream in Linus 2.6 git tree before we consider applying it. Please let us know once this has happened. Thanks.
I don't know enough about kernel development to be able to tell when that happens (patch upstream in Linux 2.6 git tree). I note yet another bug report on the 855GM which may or may not be related, here: https://bugzilla.kernel.org/show_bug.cgi?id=15248 where the resolution is "RESOLVED PATCH_ALREADY_AVAILABLE" ... ... but whether that patch has been accepted in the the "Linux 2.6 git tree" I have no idea on who to determine this. Can you provide any hints how/where to look ?
Try to find out, in that kernel.org bug report, where the patch is that is accepted that fixes your reported issue. I don't seem to be able to find any specific patch for it, do you?
Well I hadn't known the kernel.org patch, but it seems to be this one, for FC13: https://bugzilla.kernel.org/attachment.cgi?id=25084 from: https://bugzilla.kernel.org/show_bug.cgi?id=15248#c35
This is very confusing. The Intel 855GM X-Server freeze thread is difficult to follow: (1) In the bugzilla.kernel.org thread (https://bugzilla.kernel.org/show_bug.cgi?id=15248) , as near as I can determine, user Jesse Barnes speculated was suggeted to be a LID problem with a patch noted in post#5 and post#11 which is an LID fix (Lid detection ?? ). In post #14 Daniele C. added as an attachment a modified version of Jesse Barnes' patch for latest linus tree Then it was disputed by others in that thread, noting the tread was NOT about a lid detection fix. Then in post#35 Daniele C (on 2010-02-17 10:50:20) included an attachment, which Daniele C claimed was FC13's DRM patch for i8xx devices. Post#37 possibly compares the two patches ?? That bug report 15248 has been closed with a note that "RESOLVED PATCH_ALREADY_AVAILABLE" but I do not know which patch they mean. I find this very difficult to understand. (2) Plus, to add to the confusion, the freedesktop post#61 https://bugs.freedesktop.org/show_bug.cgi?id=26345#c61 fix (Intel i845GM) puports to have a fix, albiet I do not know if that fix is specific to the i845GM or if it also applies to the i855GM. (3) then, the launchpad.net site notes (https://launchpad.net/~glasen/+archive/855gm-fix) an 855GM fix of the 855GM problem noted in the freedesktop.org (intel i855GM) problem (reported here http://bugs.freedesktop.org/show_bug.cgi?id=27187) where there is a Ubuntu community live CD with this fixed (I downloaded and burned the CD and I will give it a try soon) http://glasen-hardt.de/?p=568. (4) I think this same i855GM problem is identical to an openSUSE 11.3 bug report raised by me: https://bugzilla.novell.com/show_bug.cgi?id=593463 And if the above reads to be confusing - well it definitely is confusing to me. My apologies for the confusion, and I hope I have not added to it too much.
Reference the "Ubuntu Community" fix for what I believe to be this problem, as noted here: > (3) then, the launchpad.net site notes > (https://launchpad.net/~glasen/+archive/855gm-fix) an 855GM fix of the 855GM > problem noted in the freedesktop.org (intel i855GM) problem (reported here > http://bugs.freedesktop.org/show_bug.cgi?id=27187) where there is a Ubuntu > community live CD with this fixed (I downloaded and burned the CD and I will > give it a try soon) http://glasen-hardt.de/?p=568. I downloaded the iso file for the Ubuntu community CD, burned it, and then today booted my Fujitsu Siemens Amilo 7400M laptop with the i855GM graphics. The freeze that I experienced on openSUSE-11.2 and 11.3 M4 and M6 does NOT occur with this special Ubuntu community test version. I checked the Xorg.0.log file i that Ubuntu liveCD and the "Intel" driver (version 2.11) is being used with a Ubuntu 2.6.32-22-generic kernel (which I presume has been specially patched with the fix noted here https://launchpad.net/~glasen/+archive/855gm-fix. There are so many various fixes around, this is a bit confusing.
wrt Comment 18 above (quoted below): > Well I hadn't known the kernel.org patch, but it seems to be this one, for > FC13: https://bugzilla.kernel.org/attachment.cgi?id=25084 > from: https://bugzilla.kernel.org/show_bug.cgi?id=15248#c35 I note this thread: https://bugs.freedesktop.org/show_bug.cgi?id=25325 where the assessment is the drm-intel-big-hammer.patch from FC13 is insufficient to fix these CPU/GPU coherency issues (comment-14 in that thread) with the final comment expressing the view that the patch in https://bugs.freedesktop.org/show_bug.cgi?id=27187 thread addresses the problem better. Comment #57 on this thread: https://bugs.freedesktop.org/show_bug.cgi?id=24789 also expresses the view that the drm-intel-big-hammer.patch , while an improvement, is also seriously lacking. And comment #61 in that thread sends one to this thread https://bugs.freedesktop.org/show_bug.cgi?id=27187 again to purportedly address the problem. The https://bugs.freedesktop.org/show_bug.cgi?id=27187 fix is the same one applied to the Ubuntu community test cd that I ran yesterday for about 3 hours with no problems. Special desktop effects functioned. Xrandr functioned. again, the patch is here: https://launchpad.net/~glasen/+archive/855gm-fix But I conced this is confusing, and the above is the best I can do at the moment in trying to cut through the mass of posts on this subject.
wrt the freedesktop 27187 bug fix, post #158 http://bugs.freedesktop.org/show_bug.cgi?id=27187#c158 in turn points to this proposed bug fix: http://bugs.freedesktop.org/attachment.cgi?id=35548
> Well I hadn't known the kernel.org patch, but it seems to be this one, for > FC13: https://bugzilla.kernel.org/attachment.cgi?id=25084 > from: https://bugzilla.kernel.org/show_bug.cgi?id=15248#c35 I downloaded and booted to the FC13 liveCD, and my i855GM laptop still freezes immediately after X boot, as soon as the mouse is used. Since FC13 includes the drm-intel-big-hammer.patch (which is the one referenced) I can confirm that the drm-intel-big-hammer.patch is insufficient to prevent the freeze reported in this thread. I am still of the opinion that the closest thing we have to a fix is the freedesktop 27187 bug fix, post #158 http://bugs.freedesktop.org/show_bug.cgi?id=27187#c158 in turn points to this proposed bug fix: http://bugs.freedesktop.org/attachment.cgi?id=35548 . However I can't point to a specific Linus 2.6 git tree patch. It could be there, I just do not know how/where to look.
Is comment#165 in that thread against the Linux 2.6 git tree ? https://bugs.freedesktop.org/show_bug.cgi?id=27187#c165 or just instructions how to test (what I think is now called "fix-i855-cache-coherency-v9.patch" ) against the Linux 2.6 git tree ?
Checking in today 26-Augst on the freedesktop thread http://bugs.freedesktop.org/show_bug.cgi?id=27187 with the v9-against-latest-drm-intel-next patch and most the effort appears to be testing it against the 2.6.34 kernel and ensuring changes to the 2.6.35 kernel do not undo the benefit of the patch. I can not tell if this patch has been submitted upstream to linux git tree. As near as I can determine, this is not yet the case, but I could be wrong. If we could have someone create a "test" openSUSE-11.3 liveCD with this v9-against-latest-drm-intel-next patch in that liveCD kernel, I would test it on my Amilo 7400M, and would push on openSUSE forums to have other users with similar i855GM hardware to download and test it.
Is this still an issue with openSUSE 11.3? I guess you should additionally try with latest kernel from Kernel:HEAD (and maybe the mentioned patches on top of that 2.6.35 kernel).
As noted, in addition to Johann Paluschek who raised this bug, YES this is still an issue with openSUSE-11.3 when booting to a liveCD on my Fujitsu-Siemens Amilo 7400M laptop with 1.2GB RAM. I have a bug report on openSUSE-11.3 here: https://bugzilla.novell.com/show_bug.cgi?id=593463 where the behaviour is minimally improved (instead of an immediate crash, one gets 10 to 30 minutes of operation). Trying the Kernel:HEAD patches is a bit tricky as that laptop is used by my wife and she relies on openSUSE-11.1 which is functioning nicely. I risk her wrath if I try more than just a liveCD on that laptop. Is there a liveCD with the kernel:HEAD kernels that I could test ? I do note from discussion on openSUSE forums there are different 855GM implementations and not all 855GM users suffer from this behaviour.
Moving to 11.3.
I note the following appears to be the same bug: https://bugzilla.novell.com/show_bug.cgi?id=593463 I also note this bug is actively being investigated upstream in thread http://bugs.freedesktop.org/show_bug.cgi?id=27187 where the most promising patch has still not yet been submitted upstream to Linus git tree. The reason for not yet submitted is noted in comment 233 in that thread. However there is still effort ongoing. My apologies for not being able to try any of the kernel:HEAD kernels as the hard drive is too small on that laptop for an openSUSE-11.3 install (with a custom kernel) and the best I could do is install an openSUSE-11.1 kernel or test from a liveCD, if asked to test. I can not remove the openSUSE-11.1 from that laptop without risking the wrath of my wife, who for once in her life on that laptop running openSUSE-11.1 is happy with openSUSE on a PC. If either this bug 555619 or bug 593463 is closed in lieu of the other, I will be happy with such an approach (as I believe them to be the same bug).
Ok. Let's close it as dup of Bug #593463. *** This bug has been marked as a duplicate of bug 593463 ***