X.org Lockups

ikaro · Advocate Joined: 14 Jul 2003 Posts: 2527 Location: Denmark

today i tried those new nvidia drivers..... well same thing as always.

moniek · n00b Joined: 11 Sep 2004 Posts: 71

I do not think it is a problem of nvidia.

I have got laptop with VGA compatible controller: ATI Technologies Inc Radeon R250 Lf [FireGL 9000] (rev 01)
and the same problem.

Quiet always my X loockups when I try to use tsclient to catch another screen by vnc.
It also loockups randmly.

I use udev and 2.6.11-gentoo-r4 kernel.

ikaro · Advocate Joined: 14 Jul 2003 Posts: 2527 Location: Denmark

i know the problem isnt with nvidia or ati drivers, the problem is in Xorg.
But it could be that the drivers had some fix/workaround or something else that made everything work again.
_________________
linux: #232767

ady1980 · n00b Joined: 01 Dec 2004 Posts: 19

[quote="xgrtl"]Am I the only one who had an easily reproduceable version of this bug? I usually use gnome, but my system locked up every time I tried to launch a KDE session, while trying to show the splash screen (only part of it appeared, and that was all). It ONLY happened with GLX enabled and the 1.0.7167 closed-source nvidia driver version

For more info, my system is a P4 2.0GHz, 768 MBytes RAM, Geforce 4 MX 440, gentoo with all up-to-date ~x86 packages.[/code]

Hello,
I Have setting up a new gentoo on my laptop with the 2005.0 live-cd, because the toshiba service center recovered my laptop with Win***.

I use
-xorg-x11-6.8.2-r1
-nvidia-kernel-1.0.7167-r1 and then 7174
-KDE 3.4

Ever I start kde, the splashscreen hang like yours story.

I swithced back to kde 3.3.2 but that also hangup
I switched to fluxbox and xfce4 and it run nice, I only run it at a testcase a few minutes and visited two or three sites with firefox, so I had no randomly locks.

Now, I am do a emerge --emptytree kde (kde 3.3.2), because a search hours and hours for that problem.

After my emptytree is done, I probe the older drivers.

My Laptop is a Toshiba 5200-801 with 1GB RAM an a nvidia 460 Go

I will keep you informed about my results.[code]

Meypqr · n00b Joined: 03 Apr 2005 Posts: 5

Just my experiences with this problem:
I've got two systems (1: P3 800MHz, NV 5200; 2: P4 3.2GHz, ATI Radeon 9800).

#1 is currently running kernel 2.6.5 and kde 3.3. With nv-drivers installed it locks up randomly. Removing nv-drivers "solves" the problem (no lock-ups, nothing).
#2 is running some kind of 2.4 kernel and kde 3.3 with all ati-drivers and things installed and has never had any lock-ups.

That's all.

yaneurabeya · Veteran Joined: 13 May 2004 Posts: 1754 Location: Seattle

BINGO! Those who knew me from before knew that I had issues with power management and lockups (from the other now lesser known thread), and it appears that whatever issue occurred has been solved somehow.

What I'm using:
Hardware: P4 2.4 GHz with Geforce4 Ti-4200 (Dell)
Drivers: gentoo-[dev]-sources 2.6.11-r5, nvidia 7174 driverset, xorg 6.8.2-r1
Power support compiled into kernel: ACPI (finally! they fixed the IRQ issue!), CPU fan control, the works. No APM (still might be an issue but ACPI >>> APM for my purposes).

However, considering that my system would freeze given ANY sort of power management being compiled into the kernel, I think I finally did it. Will check back in a few days after I can be positive as to whether or not its total success.

Nerdanel · Posted: Mon Apr 04, 2005 2:02 pm Post subject:

Just a while ago I had an almost repeatable lockup.

1. My random crashes had been mainly related to writing in the address bar in Konqueror. I lowered the limit of URIs to remember from 500 to 100 and saved the settings. Immediately after that, I tried to write www.povray.org (which was not in Konqueror's memory) and got only to www.povra before the machine froze and I had to reboot with magic SysRq.

2. Opened a website sure to be in recently visited (slashdot.org). It worked. Wrote www.povra-- Lockup.

3. Repeat with identical results.

4. Just one repeat before this is an official repeatable freeze. This time I'm careful to type www.povray.org very slowly with seconds between each letter. No lockup.

korngerd · Tux's lil' helper Joined: 12 Mar 2004 Posts: 104

I had my system running gentoo-sources-2.6.11-r4, since I wanted to test-drive nvidia's latest drivers (1.0.7167). At first, it seemed to be really stable. After a few days of "normal usage" though, it started to become worse and worse... First, the X.org lockup, then a complete system halt (kernel panic - not syncing: fatal exception in interrupt). After downgrading the kernel back to 2.6.10, it seems to be a bit better (still getting the X.org lockups, but no kernel panics yet). Just some advice on people trying out the 2.6.11 kernel

Gergan Penkov · Posted: Tue Apr 05, 2005 2:23 pm Post subject: lockups

I think I have some hint, but I don't have the time to test it. My xorg started to freeze when I have installed the new 2005.0 profile and gnome 2.10 and rebuilded the system. I don't now if ipv6 is the default in the USE flags for this profile, but i never used it before and now the system is recompiled with it and only firefox/mozilla-based browsers lock up the xorg server - galeon, epiphany, mozilla and firefox, some often some seldom. I use now opera and don't have freezes for 3 days. So could someone recompile the whole without ipv6 and share the experience, probably this will help ::

))

John-Boy · Posted: Tue Apr 05, 2005 6:59 pm Post subject:

Just to say that this has occurred again, after a free period, I have the latest Nvidia drivers - the latest lock-ups (two of 'em) happened whilst surfing ...

sobit · Posted: Wed Apr 06, 2005 3:22 am Post subject:

software versions:
nvidia 1.0.7174
xorg-x11-6.8.2-r1
glibc-2.3.4.20041102-r2 (nptlonly)
gcc 3.4.3.20050110-r1
gentoo-sources-2.6.11-r5 (no APM or ACPI)
mozilla-firefox-1.0.2
konqueror-3.4.0
hal, udev and holgplug

I turned off RenderAccel in xorg.conf as suggested but it still locked up. I switch to the vt I had started kde from and noticed DCOP or mcop message with my host_domain. domainname doesn't run from default runlevel in 2005.0 (at least not for me) and I haven't bothered messing with /etc/env.d/ to set it correctly. $HOME/.[d,m]cop... was looking for it and not finding it. So I edited /etc/hosts like so.

127.0.0.1 HOST localost
192.168.0.1 HOST HOST.DOMAIN

putting HOST as the first nickname. Kde3.4 has tripled in speed and I have not had any lockups in hours.

I wonder if this is a solution for other posters? I don't think this is a [d,m]cop problem but what ever is setting up tmp files.

Gergan Penkov · Posted: Thu Apr 07, 2005 6:03 am Post subject: ipv6

I don't know if someone has followed my hint to recompile the system without ipv6, but I have tried sth other. I have simply recompiled the kernel with all the tcp/ip6 enabled as modules, since than:

Nerdanel · Posted: Thu Apr 07, 2005 8:40 am Post subject:

I had noticed earlier that ipv6 had gotten added recently and removed it in case it was the problem (and because it isn't useful to me). At that time it didn't help, but just now I found out that I still had my qt compiled with ipv6 even if kde wasn't anymore. I didn't even know qt was involved with net stuff. I'm now re-emerging qt and we'll see what happens.

alaa · n00b Joined: 23 Feb 2004 Posts: 57

I went through 18 pages of this forum just to realise that this problem is still not fixed. What the hell is this problem? I don't know if this will help anyone but the majority of my crashes happen while I'm browsing the web using Firefox. But then again that's what I mostly do. My last crash happened when I clicked on a link that tried to open a popup window to display a zoomed image.

I just built gentoo-2.6.11-r5 and disabled ACPI. Still crashes. Everytime it crashes with me I can remotely ssh into the box and find that X is taking 99.9% CPU. I'm running xorg by the way. I don't have too many apps running. Just Firefox and Eterm. And sometimes I just leave my PC running and come back to find that X just decided to crash. The only code I can think of that could have run at that time is either some cron job or xscreensaver.

But it's important to note that earlier through my battles I decided to kill xscreensaver and my X still took 99.9% CPU all of a sudden.

You guys wanna hear something funny? My roommate is using the exact same hardware as mine, running the same apps and he doesn't suffer from this problem. What the hell is up with that? The only thing I can think of that is different from him and I is that we have not emerged the same packages in the same order. He has more packages installed on his gentoo than I do.

This is a nasty bug guys......good luck, and good bye (.... preparing to shoot myself ...)

ikaro · Advocate Joined: 14 Jul 2003 Posts: 2527 Location: Denmark

Hayl · Posted: Fri Apr 08, 2005 12:43 am Post subject:

i had no problems on amd64

re-emerged -e world with -fomit-frame-pointer

got random crashes of X

re-emerge -e world without -fomit-frame-pointer

not one crash since

i have the 7174 nvidia driver, and i run gnome with -qt -arts -kde in USE
_________________
"I do not fear computers. I fear lack of them." - Isaac Asimov

SnEptUne · l33t Joined: 23 Aug 2004 Posts: 656

Even after all the xorg-x11 updates and disabling power management, my system still lock up when OpenGL applications are running. Disabling RenderAccel works of course, but it makes OpenGL applications unusable (only 2 fps per seconds >_<). Additionally, after I upgraded my video card, AGPMode 4 no longer works, it freezes after time X is started. AGPMode 3, 2, or 1 have no problem though.
_________________
"There will be more joy in heaven over the tear-bathed face of a repentant sinner than over the white robes of a hundred just men." (LM, 114)

AwoL · Tux's lil' helper Joined: 11 Feb 2005 Posts: 110

I just started having this problem within the past week. For me, it usually happens when I'm typing something into a konqueror address bar, or when I'm moving files from one directory to another (using KDE, not from the command line). I recently upgraded to the xorg 6.8.2-r1. Has anyone tried reverting back to an older version? How do you do that anyway?

By the way, here are my system specs:

2004.3 profile
Xorg 6.8.2-r1
KDE 3.3.2-r2
ATE Radeon 8500
Asus A7N8X-X mobo with the nforce2 chipset

Nerdanel · Posted: Sat Apr 09, 2005 6:25 am Post subject:

I tried to revert xorg but it didn't help at all.

mhex · Posted: Sat Apr 09, 2005 11:16 am Post subject:

Here same problem. NVidia 6xxx/7xxx, KDE, Kernel 2.6.11, AMD64 3000+ 32bit Gentoo.

Starting konqueror as webbrowser freezes X. Mouse cursor works but no keyboard.
No reaction from desktop no clock update in the panel.
Login from another comp shows 99% cpu for the X process.
Killing the process with -9 brings me back to kdm.

Another application that freezes the desktop is a wine application called t@x2005.
Here it loads the splashscreen then stutters then freeze.

Running xorg with the buildin nvidia driver shows no lockups.

AwoL · Tux's lil' helper Joined: 11 Feb 2005 Posts: 110

How are you guys killing processes remotely? After I ssh into the box, I know I can use the "killall X" command, but how do you display and kill specific processes, like you can with ksysguard?

Nerdanel · Posted: Sat Apr 09, 2005 4:30 pm Post subject:

One way is to use "top" and then "k"ill the process number you want.

I'm not too much into this since I don't actually have a second computer to ssh from.

praseodymium · Tux's lil' helper Joined: 08 Dec 2004 Posts: 83

I'll add a 'me too' to this thread.

PC config: AMD 2000+/512mb RAM/nVidia GeForce 4 Ti4200 128mb AGP8x
Kernel: APM enabled. ACPI broke my soundcard because of IRQ problems.

Oh, and I have this as long as I use Gentoo which is from about December last year.

dona · n00b Joined: 09 Apr 2005 Posts: 3

I had similar problems, Xorg locked up right away, when I tried to log in. Solution for me was:
1. I disabled IPV6 in /etc/hosts and added "KDE_NO_IPV6 = 1" to /etc/env.d/99kde-env
2. I added host before localhost into /etc/hosts > 127.0.0.1 host localhost

Xorg and kde became stable and much faster.

AwoL · Tux's lil' helper Joined: 11 Feb 2005 Posts: 110

I've only had this problem for the past few days, but the odd thing is, it seems (at least since yesterday evening) to have stopped since I quit running Eterm which I had displaying tail.

Anyway, I don't really understand the IPV6 part of /etc/hosts. Mine displays: