Created attachment 426614 [details] The Xorg log file Description of problem: Greetings I recently upgraded from a fully working setup in fedora 11 to fedora 13. I have a dual socket, quad core dell workstation running xeons and three nvidia graphics cards: geforce 9500GT, geforce 9400 GT and an nvidia GTX 260. The three graphics cards powered five monitors through xinerama using the nvidia driver in fedora 11 without any problems. After I installed fedora 13 and did a yum update on everything, I disabled the nouveau driver, installed the latest nvidia driver and enabled xinerama. When I move the cursor across monitors, I get the following error: [mi] EQ overflowing. The server is probably stuck in an infinite loop. The cursor starts to flicker and becomes immovable and the keyboard goes dead. My guess is at this point X is pretty much dead. I have done some research on this for a couple of days now and this error appears to be a symptom, not a bug in itself. That is to say it is a catch all error that can be triggered in a number of places. I am desperate for an answer obviously. I have been using fedora since fedora 3 and redhat 7.2 before that, any assistance you can provide is greatly appreciated. This is also the first time I file a bug report with redhat (things have been great until now) so if I miss anything please do let me know. Also, I ran into something similar when I temporarily upgraded to Fedora 12. I did not pursue the matter though and reverted back to fedora 11. If it helps, I have tried to re-apply the patch suggested here: http://www.nvnews.net/vbulletin/showthread.php?t=142656 to the VGA arbiter in the past but it did not help. I could have easily been simply doing it wrong. I do however hope that the patch made its way into the nvidia driver by now but I still trying to confirm this. Version-Release number of selected component (if applicable): * Fedora 13 * uname -a Linux defiant-sovereign 2.6.33.5-124.fc13.x86_64 #1 SMP Fri Jun 11 09:38:12 UTC 2010 x86_64 x86_64 x86_64 GNU/Linux * X -version X.Org X Server 1.8.0 Release Date: 2010-04-02 X Protocol Version 11, Revision 0 Build Operating System: x86-02 2.6.18-164.15.1.el5 Current Operating System: Linux defiant-sovereign 2.6.33.5-124.fc13.x86_64 #1 SMP Fri Jun 11 09:38:12 UTC 2010 x86_64 Kernel command line: ro root=UUID=1ef93da6-4809-4456-994b-814274216206 rd_NO_LUKS rd_NO_LVM rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us rhgb quiet rdblacklist=nouveau nomodeset intel_iommu=igfx_off Build Date: 02 May 2010 02:56:53PM Build ID: xorg-x11-server 1.8.0-12.fc13 Current version of pixman: 0.18.0 Before reporting problems, check http://bodhi.fedoraproject.org/ to make sure that you have the latest version. * Using nvidia drivers for x86_64: NVIDIA-Linux-x86_64-256.35.run from the install script not rpmfusion How reproducible: Very consistent. Whenever I move the cursor across screens it starts to flicker and then X becomes unresponsive: moving the mouse does not do anything, keyboard is dead. Steps to Reproduce: 1. Use different graphics cards, nvidia 9400 and above, connect multiple monitors 2. Install fedora 13, upgrade with yum, install nvidia drivers 3. Enable xinerama, move cursor across monitors Actual results: The cursor flickers, mouse goes dead and keyboard goes dead. Sometimes X restarts on its own sometimes not, it just hangs. Expected results: Similar behavior to Fedora 11: moving cursor across monitors does not cause X to crash Additional info:
Created attachment 426616 [details] The X configuration file The xconf file I am using
Additionally, the output of dmesg does not show anything indicative (to me at least) but, the bits that talk about NVIDIA are as follows: nvidia: module license 'NVIDIA' taints kernel. Disabling lock debugging due to kernel taint nvidia 0000:01:00.0: enabling device (0002 -> 0003) nvidia 0000:01:00.0: PCI INT A -> GSI 48 (level, low) -> IRQ 48 nvidia 0000:01:00.0: setting latency timer to 64 vgaarb: device changed decodes: PCI:0000:01:00.0,olddecodes=io+mem,decodes=none:owns=none nvidia 0000:02:00.0: enabling device (0002 -> 0003) nvidia 0000:02:00.0: PCI INT A -> GSI 52 (level, low) -> IRQ 52 nvidia 0000:02:00.0: setting latency timer to 64 vgaarb: device changed decodes: PCI:0000:02:00.0,olddecodes=io+mem,decodes=none:owns=none nvidia 0000:08:00.0: PCI INT A -> GSI 30 (level, low) -> IRQ 30 nvidia 0000:08:00.0: setting latency timer to 64 vgaarb: device changed decodes: PCI:0000:08:00.0,olddecodes=io+mem,decodes=none:owns=io+mem NVRM: loading NVIDIA UNIX x86_64 Kernel Module 256.35 Wed Jun 16 18:42:44 PDT 2010 type=1400 audit(1277364628.694:5): avc: denied { mmap_zero } for pid=827 comm="vbetool" scontext=system_u:system_r:vbetool_t:s0-s0:c0.c1023 tcontext=system_u:system_r:vbetool_t:s0-s0:c0.c1023 tclass=memprotect type=1400 audit(1277364628.694:4): avc: denied { mmap_zero } for pid=853 comm="vbetool" scontext=system_u:system_r:vbetool_t:s0-s0:c0.c1023 tcontext=system_u:system_r:vbetool_t:s0-s0:c0.c1023 tclass=memprotect type=1400 audit(1277364628.694:6): avc: denied { mmap_zero } for pid=828 comm="vbetool" scontext=system_u:system_r:vbetool_t:s0-s0:c0.c1023 tcontext=system_u:system_r:vbetool_t:s0-s0:c0.c1023 tclass=memprotect
Thanks for the report. We are sorry that we cannot help you with your problem, but we are not able to support binary-only drivers. If you would be able to reproduce this issue using only open source software, please, reopen this bug with the additional information, but in meantime I have no choice than to close this bug as CANTFIX (because we really cannot fix it). The open source 'nouveau' driver (in package xorg-x11-drv-nouveau) is the recommended alternative for users of Nvidia graphic chips. It is used by default in Fedora 11 and later if you remove any customizations that explicitly set the video driver. The older "nv" driver may be needed in some cases. It is also available in older Fedora releases. Install the packages xorg-x11-drv-nouveau or xorg-x11-drv-nv and override the X server's default choice if necessary. See https://fedoraproject.org/wiki/Features/NouveauAsDefault for more information. If you used a non-packaged version of the driver from the Nvidia website please clean your system from additional libraries and software it installed. For users who are experiencing problems installing, configuring, or using the unsupported 3rd party proprietary "nvidia" video driver, Nvidia provides indirect customer support via an online web based support forum. Nvidia monitors these web forums for commonly reported problems and passes them on to Nvidia engineers for investigation. Once they've isolated a particular problem, it is often fixed in a future video driver update. The NVNews Nvidia Linux driver forum is located at: http://www.nvnews.net/vbulletin/forumdisplay.php?s=&forumid=14 Once you have reported this issue in the Nvidia web forums, others who may have experienced the particular problem may be able to assist. If there is a real bug occuring, Nvidia will be able to determine this, and will likely resolve the issue in a future driver update for the operating system releases that they officially support. While we does not support the proprietary nvidia driver, users requiring technical support may also find the various X.Org, XFree86, and Red Hat/Fedora mailing lists helpful in finding assistance: X.Org mailing lists: http://www.freedesktop.org/XOrg/XorgMailingLists XFree86 mailing lists: http://www.xfree86.org/sos/lists.html Red Hat/Fedora mailing lists: https://listman.redhat.com/mailman/listinfo
So I found out what the problem was. It was a manifestation of this issue: http://bugs.freedesktop.org/show_bug.cgi?id=24986 updating to fedora 13 rawhide x11-server sort of fixed it. I say sort of because I am fighting more issues with instability and graphical artifacts. I honestly would use the nouveau driver if it worked properly on my setup.
(In reply to comment #4) > I honestly would use the nouveau driver if it worked properly on my setup. There is no bad will on my part for you using the nvidia blob. I can understand that noveuau might have troubles to run configurations like yours (did you try?) but the point is that we really cannot help you.
Of course there is no bad will Matej, I develop open source software too. Mine is more along the lines of HPC computing though. Anyways, yes I did try the nouveau drivers, the issue I get from them is a system freeze when the cursor crosses a screen. If you looked at the link I sent though, you will see the problem was not the nvidia driver, it was X11. The fix went in here: http://lists.freedesktop.org/archives/xorg/2010-April/050031.html So in case somebody gets this error among the dozen or or so multi-monitor users, the simplest fix is to upgrade X11 to rawhide X11. Now I have to go back and fight with gpk-application. It does not ask me about the root password anymore but tells me authentication failed.
Thanks for looking at this though. It is nice to know somebody cares :)
(In reply to comment #7) > Thanks for looking at this though. It is nice to know somebody cares :) It's my job to care ;) Anyway, have you filed a bug on that issue in comment 6?
No not yet. I am trying to get more info and found a few related bugs that might resolve it. At least I'll have something to do over lunch. Regards Yaakoub
I found the original bug report of this bug but with different symptoms so I am marking this as duplicate of. *** This bug has been marked as a duplicate of bug 567835 ***