Bug 15650 - Hard freeze when starting X with GeForce 6800 GS (NV43)
Summary: Hard freeze when starting X with GeForce 6800 GS (NV43)
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-04-22 07:12 UTC by Andrew McNabb
Modified: 2011-02-15 02:59 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
xorg.conf (619 bytes, text/plain)
2008-04-22 07:12 UTC, Andrew McNabb
no flags Details
Xorg.0.log (101.24 KB, text/plain)
2008-04-22 07:13 UTC, Andrew McNabb
no flags Details
/var/log/messages (40.81 KB, text/plain)
2008-04-22 11:06 UTC, Andrew McNabb
no flags Details
Xorg.0.log for most recent source. (102.71 KB, text/x-log)
2008-04-22 11:17 UTC, Andrew McNabb
no flags Details
kernel log with better log level (42.77 KB, text/plain)
2008-04-22 12:11 UTC, Andrew McNabb
no flags Details
Xorg.0.log after AGP init change (110.48 KB, text/x-log)
2008-04-22 13:34 UTC, Andrew McNabb
no flags Details
Xorg.0.log; this time AGP init thing was recompiled properly (110.39 KB, text/x-log)
2008-04-22 13:48 UTC, Andrew McNabb
no flags Details
Kernel log from X startup to X crash, which was triggered by urxvt (136.59 KB, application/octet-stream)
2008-04-23 07:20 UTC, Andrew McNabb
no flags Details

Description Andrew McNabb 2008-04-22 07:12:13 UTC
Created attachment 16099 [details]
xorg.conf

I installed Fedora Rawhide, which included the package xorg-x11-drv-nouveau-0.0.10-2.20080408git0991281.fc9.x86_64.  When starting X, the mouse cursor appears in the middle of the screen, and then the machine freezes.  It's not just X that freezes; the computer is no longer reachable over the network.

This is my first time trying Nouveau, so I'm not sure which information is most helpful, but I'll attach Xorg.0.log and xorg.conf.  Here is the information about the card from `lspci -v`:

05:00.0 VGA compatible controller: nVidia Corporation NV43 [GeForce 6800 GS] (rev a2) (prog-if 00 [VGA controller])
	Subsystem: XFX Pine Group Inc. XFX GeForce 6800 XTreme 256MB DDR3 AGP
	Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 11
	Memory at fd000000 (32-bit, non-prefetchable) [size=16M]
	Memory at d0000000 (32-bit, prefetchable) [size=256M]
	Memory at fc000000 (32-bit, non-prefetchable) [size=16M]
	Expansion ROM at feae0000 [disabled] [size=128K]
	Capabilities: [60] Power Management version 2
	Capabilities: [44] AGP version 3.0
	Kernel modules: nvidiafb
Comment 1 Andrew McNabb 2008-04-22 07:13:01 UTC
Created attachment 16100 [details]
Xorg.0.log
Comment 2 Andrew McNabb 2008-04-22 08:39:51 UTC
I just tried installing from source to see if it made any difference.  I checked out the latest sources from git (following the InstallNouveau instructions on the wiki).  Unfortunately, the same problem occurred.

I also tried doing `startx` instead of using gdm.  It turns out that I can move the mouse until the first X client connects to the server, at which point the machine locks up.  Even a simple X client like `xterm` is enough to freeze the computer.
Comment 3 Andrew McNabb 2008-04-22 11:06:46 UTC
Created attachment 16102 [details]
/var/log/messages

I just ran it again.  This time I did `insmod drm.ko debug=1`.  I'm attaching /var/log/messages (from boot to crash).  It doesn't look very informative, but it might be helpful.
Comment 4 Andrew McNabb 2008-04-22 11:17:13 UTC
Created attachment 16103 [details]
Xorg.0.log for most recent source.
Comment 5 Andrew McNabb 2008-04-22 12:11:49 UTC
Created attachment 16107 [details]
kernel log with better log level

I was able to get a kernel log with more messages.  This new kernel log is with / mounted with sync, and with the kernel booted with nosmp.
Comment 6 Stephane Marchesin 2008-04-22 13:21:12 UTC
This is a bridged card (AGP). This might or might not matter.
Comment 7 Andrew McNabb 2008-04-22 13:34:13 UTC
Created attachment 16109 [details]
Xorg.0.log after AGP init change

Replaced #if !defined(__powerpc__) && !defined(__ia64__)
with #if 0
Comment 8 Andrew McNabb 2008-04-22 13:48:01 UTC
Created attachment 16110 [details]
Xorg.0.log; this time AGP init thing was recompiled properly
Comment 9 Andrew McNabb 2008-04-23 07:19:56 UTC
When we replaced:

#if !defined(__powerpc__) && !defined(__ia64__)

with 

#if 0

on line 498 of nouveau_mem.c, clients were able to connect without X crashing.  I was able to start gnome-terminal, xterm, Firefox, etc.

However, when I started urxvt (a simple terminal with unicode support), X crashed and I got the following errors in .xsession-errors:

gnome-session: Fatal IO error 11 (Resource temporarily unavailable) on X server : \0.0.
gnome-terminal: Fatal IO error 11 (Resource temporarily unavailable) on X server :\0.0.
gnome-panel: Fatal IO error 11 (Resource temporarily unavailable) on X server :0. \0.
bluetooth-applet: Fatal IO error 11 (Resource temporarily unavailable) on X       \server :0.0.
nm-applet: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.0.
pam-panel-icon: Fatal IO error 11 (Resource temporarily unavailable) on X server :\0.0.
firefox: Fatal IO error 104 (Connection reset by peer) on X server :0.0.
Window manager warning: Fatal IO error 11 (Resource temporarily unavailable) on   \display ':0.0'.
kerneloops-applet: Fatal IO error 11 (Resource temporarily unavailable) on X      \server :0.0.
applet.py: Fatal IO error 11 (Resource temporarily unavailable) on X server :0.0.
gpk-update-icon: Fatal IO error 11 (Resource temporarily unavailable) on X        \server :0.0.


I checked Xorg.0.log, and it didn't give any additional output.  I also tried with Composite both enabled and disabled, and both showed the same problem.  I will attach the kernel log (with drm output) from the time X started to the time it crashed.

This seems related to the earlier crashes, but it might not be.  You would probably know better than I do.

Thanks for your help.
Comment 10 Andrew McNabb 2008-04-23 07:20:55 UTC
Created attachment 16129 [details]
Kernel log from X startup to X crash, which was triggered by urxvt
Comment 11 Andrew McNabb 2008-04-28 12:22:18 UTC
It turns out that the urxvt crash happens with the nv driver as well, so I think it's a general xserver problem.  It's certainly not related to this bug, so I filed a separate bug report.

The workaround for this bug (of disabling the AGP initialization) seems to work great.

Thanks.
Comment 12 Daniel Qarras 2008-08-19 10:08:38 UTC
FWIW, I've also seen freezes with nouveau / nv drivers on x86-64 with NV43 (but nvidia blob works all ok). Full details are at:

https://bugzilla.redhat.com/show_bug.cgi?id=459405
Comment 13 Andrew McNabb 2008-09-11 09:41:22 UTC
Daniel, I think yours is actually a separate bug.  This one pertains specifically to bridged AGP cards.

By the way, this has been open a long time.  I was just wondering whether anyone has had a chance to look at it?
Comment 14 Lucas Stach 2011-02-15 02:59:29 UTC
Closing this as there were no status updates for about two years. If this is still a problem with recent nouveau, please reopen with new logs.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.