Mysterious Freeze

Message

joehni · Post by **joehni** » Sat Jan 15, 2005 1:23 pm

Interesting. After logging in with ssh for my latest freeze I was able to kill X and start KDE from the remote machine again. This enabled my keyboard again ... nothing special in the logs though.

raf · Post by **raf** » Sun Jan 16, 2005 8:30 pm

Hi All,

I've been experiencing the same problem you are all discussing, however mine would NOT be as frequent as some of you describe. It would only happen about once every two-three days of heavy use and 95% of the time it would be in firefox as I'm scrolling.

Either way, I've been using my system hard for about a week (multiple instances of X, multiple instances of firefox each with 30 tabs running, kernel compiles, ssh, all at the same time) without any problems what so ever. I have recently changed two things: 1) I have really cleaned up my kernel running 2.6.10, and 2) I'm using different compiler flags.

I realize this is not a specific answer, but I could post my .config and make.conf if it would help. Hope someone can pin point this better though.

All the best,
Raf

davecs · Post by **davecs** » Sun Jan 16, 2005 11:54 pm

Clutching at straws here, but I use a 3D mouse cursor set called white_shadow. Mouse cursors? Stranger things have been known...

firephoto · Post by **firephoto** » Mon Jan 17, 2005 1:03 am

I got one of these today too when I was emerging something, came back about 30 minutes later to find the box froze.

Code: Select all

Jan 16 16:00:01 mybox /usr/sbin/cron[22836]: (root) CMD (rm -f /var/spool/cron/lastrun/cron.hourly)
Jan 16 16:21:21 mybox NVRM: Xid: 6, PE0000 0000 000faaf4 0000fdb8 e5640000 00000000
Jan 16 16:47:08 mybox syslog-ng[10227]: syslog-ng version 1.6.5 starting

Konsole was open with 2 root sessions, konqueror was running with a few tabs open, kopete, kontact, ksensors, juk (not playing anything), no auto screensaver or dpms enabled, ksim panel running, and the usuall stuff. Running the latest ~x86 nvidia, xorg, and 2.6.10-ck4 with kde-3.4.0_beta1.

Zentoo · Post by **Zentoo** » Mon Jan 17, 2005 8:36 pm

Is anyone have just try to downgrade to 6111 version of kernel-nvidia just to see if that'll crash ever. If not, so we know the nvidia driver is the cause. Isn't it ?

bunder · Post by **bunder** » Tue Jan 18, 2005 3:34 am

Tuttle wrote:Is anyone have just try to downgrade to 6111 version of kernel-nvidia just to see if that'll crash ever. If not, so we know the nvidia driver is the cause. Isn't it ?

i did, and it still does.

yaneurabeya · Post by **yaneurabeya** » Tue Jan 18, 2005 8:30 am

I've been running 6111 on both the kernel and the glx and that's not the issue for me. It still crashes regardless of whether or not I am using the proprietary nvidia drivers.

So far I've chased it down (for me) to be a kernel bug possibly and I've already submitted my bug and it's being tested by the kernel bug people, but I'm to the point that I've thrashed my fs'es enough that I don't want to try that anymore. I know this may sound silly, but has anyone thought of emulating their system via a livecd w/ those options enabled and see if that affects their setup in X? It just seems stupid to keep on trashing our files this way cause eventually we'll screw up our hard disks with this behavior =\... That's why I'm waiting until a) the bug for me is addressed in the next kernel version or b) I make a liveCD and have enough time to test out the issue.

Once again though. I think that the problems being presented in this thread may be multiple separate issues because although the initial issues sound the same, our symptoms overall tend to differ a bit from each other.

davecs · Post by **davecs** » Tue Jan 18, 2005 3:21 pm

Sunday night I recompiled my kernel to gentoo-dev-sources-2.6.10-r4. I got rid of all framebuffer stuff because I can't use it with nvidia anyway. (I also got rid of slocate from /etc/cron.daily because frankly I rarely use locate. Doubt if relevant though.)

Anyway I stayed in Gentoo last night, pushed it a couple of times, no sign of anything crashing. But then again it can be months then every 30 minutes then back to months again. Who knows?

Here's hoping ...

yaneurabeya · Post by **yaneurabeya** » Tue Jan 18, 2005 7:06 pm

I compiled r5 and am going to use the nvidia 6629 drivers. Here goes. WEE!

MickKi · Post by **MickKi** » Tue Jan 18, 2005 7:38 pm

Guys, it's not Nvidia.

I'm running an ATI 7200 AIW card on an anemic 600MHz PIII and currently using the 2.6.9-gentoo-r13 kernel. Using this kernel it just crashes every day!

After I turn the machine on, I go into screen and start emerge sync, exit screen and either work on the console, or start X and browse the internet (using Opera). Anyway, the emerge sync never finishes first time round. It will always crash. After I reboot I repeat the exercise and it won't freeze, no matter how many applications I turn on (e.g. XMMS, OOo, etc.)

Over time I have ended up blaming the different kernels. Some freeze, others don't. That said, it may be related with different versions of portage/python/gcc, etc. Change of kernel often seems to resolve it - for a while.

davecs · Post by **davecs** » Tue Jan 18, 2005 8:53 pm

MickKi, I think we all decided some time ago in this thread that GLX may be the problem. The nvidia drivers obviously enable GLX but so do other 3D drivers. Although you don't have to be doing any 3D stuff at the time.

The fault is definitely in X11 somewhere, and has been so since Xfree4.3 at least.

The main thing is: Don't reboot. Hold down ALT-SYSRQ and press "I". If that doesn't get you to a login screen, you may as well type "U", "S" and "R" to unmount, save and reboot. (Whilst holding ALT-SYSRQ.) If it does, you might find a few daemons have gone down, so reboot in the more orthodox way from a terminal.

Hopefully the xorg people are more responsive than xfree were and we may see some progress?

MickKi · Post by **MickKi** » Tue Jan 18, 2005 9:43 pm

Hi davecs,

Yes, it has plauged me on & off since the days of the xfree 4.3. However, my machine currently crashes even when I am in console (it did just earlier today) with no X server running.

If it freezes up when X is running say, during an "emerge sync", no matter what keypresses I try it won't respond:

Alt+SysRq+s
Alt+SysRq+i
Alt+SysRq+k
Alt+SysRq+b

I have to go for the plug every time.

wuzzerd · Post by **wuzzerd** » Wed Jan 19, 2005 3:24 am

Hmm this thread has been around awhile

For me it hangs with all the symptoms if I start gdm immediately after booting. The old try to close a window bug. If I grab a cuppa coffee and wait a minute or two: no problem

joehni · Post by **joehni** » Wed Jan 19, 2005 5:03 pm

beugh wrote:
Tuttle wrote:Is anyone have just try to downgrade to 6111 version of kernel-nvidia just to see if that'll crash ever. If not, so we know the nvidia driver is the cause. Isn't it ?
i did, and it still does.

Well, I did now an upgrade to 1.0.6629-r1 after it got stable, turned on preemptive in kernel (since with it the problem manifests quite steadily) and ... bang. This time I could not even log in with ssh <grrrr>

Did anyone tried different versions of xorg ?

davecs · Post by **davecs** » Wed Jan 19, 2005 9:18 pm

I think we need to split this up into a number of problems.

We have:

Lockups in X where the mouse keep moving but no keys work except SYSRQ combos. Generally happens with GLX running.

Lockups in X where the mouse freezes along with everything else

Lockups in terminals with X running

Lockups in terminals with X not running

Lockups when certain programs are run.

My problem is definitely the first in the list, and has occurred rarely but unpredictably. Though I could also freeze LBreakout2 if I didn't limit the frame rate. I think we need a few threads, but posts may have to be reallocated by the moderators...

korngerd · Post by **korngerd** » Thu Jan 20, 2005 12:36 am

davecs wrote:Lockups in X where the mouse keep moving but no keys work except SYSRQ combos. Generally happens with GLX running.

I'm definitely in this category, too. I emailed NVidia about the problem, and they responded pretty quickly. They do know that the problem exists, and suggested trying to disable AGP. I tried this out, and (although most graphical stuff slowed down) it made the system a "bit" more stable. I still get crashes, but not as often as when I had it enabled. Another strange thing is that the X server just aborts now, so X just seems to restart itself (and returns me to GDM). At least I don't have to do a cold restart (or ssh into the machine to restart X), so it might be a try. All I did in /etc/xorg.conf was:

Code: Select all

Section "Device"
        Identifier  "nVidia GeForce2"
        Driver      "nvidia"
        VendorName  "nVidia Corporation"
        BoardName   "NV15 [GeForce2 GTS/Pro]"
        BusID       "PCI:1:0:0"
        #
        # Add this NvAGP option to disable AGP completely
        #
        Option          "NvAGP" "0"
        Option          "RenderAccel" "true"
        Option          "AllowGLXWithComposite" "true"
EndSection

Since I'm not on an ATI card, I can't really suggest much to the ATI people, though.[/quote]

** Edit: Again, just for reference, I'm running on nvidia-kernel-1.0.6629-r1 and gentoo-dev-sources-2.6.10-r5. My Xorg is xorg-x11-6.8.0-r3.

korngerd · Post by **korngerd** » Thu Jan 20, 2005 4:20 am

OK, I talked too soon - it did actually do the "except mouse" freeze on me again. Maybe disabling AGP isn't actually any better...

MickKi · Post by **MickKi** » Sun Jan 23, 2005 11:04 am

davecs wrote: . . .Lockups in terminals with X not running. . .

OK, I have now rolled my first 2.6.10 series kernel (2.6.10-gentoo-r6) and tried "emerge sync". The machine did NOT freeze!

At the same time though I noticed that portage downgraded net-misc/rsync-2.6.3 to rsync-2.6.0-r3. Not sure if rync was to blame or if it is a kernel issue. As I said above it comes and goes as life and Linux progresses through different kernels and rsync versions.

I'll keep an eye on it and report back if the freezes return when I'm running X at the same time.

dibblethewrecker · Post by **dibblethewrecker** » Sun Jan 23, 2005 12:36 pm

I'm from Arch Linux but i'm just posting here as i have this problem - i am number two - total and utter lock up.

it happens to me when most xscreensavers starts - i use the preview to test for the crash. you might say it is xscreensaver but it isn't, i tried it with previous versions of xscr and gtk2, and i never had a problem before. it started happening when i moved to 2.6.10. I was using the nitro patches. so i moved back to 2.6.9 but i am still getting some crashes - it seems to be the kernel so here is run down:

2.6.10 -nitro2 crash
2.6.10 -nitro4 crash
2.6.10 -as (stock Arch linux) no crash
2.6.9 -nitro4 no crash
2.6.9 -cko3 crash

i had ZERO probs till i used the first 2.6.10 kernel - now the prob seems to have followed me back to 2.6.9 - the big bummer is i accidently deleted a load of my pre built 2.6.9 kernels when i installed 2.6.10 - so i can't test them!

i don't have nvidia but i do have an ATI card that does 3d stuff.

just my 2p

iarwain · Post by **iarwain** » Sun Jan 23, 2005 5:07 pm

Does anyone know if other UNIXs have this problem? I mean *BSD for example. If it only happens with Linux, then it must be kernel's fault.

kmare · Post by **kmare** » Mon Jan 24, 2005 1:26 pm

I'm having similar problems lately... I thought it was an nvidia driver problem... I just changed the agp to the agpgart that is in the default kernel (not the one in the binary driver). No freezes till now. We'll see... but I'm really starting to think that it is (mostly) a problem with reiserfs (or a journaling problem). I've heard from some ppl that while they download a lot of stuff (the HD usage gets high) they see system lockup, corrupted file system and such... mostly with bittorrent... I don't know what else to check. Hopefully the problem will suddenly disappear...

lookinin · Post by **lookinin** » Mon Jan 24, 2005 1:41 pm

I too am having this problem. I only see it with OpenGL apps such as xscreensaver, tuxracer, and the OpenGL spectrum analyzer in XMMS. All of those programs lock the system every time I use them and I have yet to see any log entries that might point me in a direction. I'm using xorg 6.8.1.902, gentoo-dev-sources-2.6.10-r5 (tho I've had this problem through a few kernel versions), an ATI Radeon 9000, and I'm using the xorg OpenGL implementation. This problem has driven me a little crazy over the last couple months - any help would be greatly appreciated.

Part of my kernel config:

Code: Select all

CONFIG_AGP=y
# CONFIG_AGP_ALI is not set
CONFIG_AGP_ATI=y
# CONFIG_AGP_AMD is not set
# CONFIG_AGP_AMD64 is not set
# CONFIG_AGP_INTEL is not set
# CONFIG_AGP_INTEL_MCH is not set
# CONFIG_AGP_NVIDIA is not set
# CONFIG_AGP_SIS is not set
# CONFIG_AGP_SWORKS is not set
CONFIG_AGP_VIA=y
# CONFIG_AGP_EFFICEON is not set
CONFIG_DRM=y
# CONFIG_DRM_TDFX is not set
# CONFIG_DRM_R128 is not set
CONFIG_DRM_RADEON=y
# CONFIG_DRM_MGA is not set
# CONFIG_DRM_SIS is not set
# CONFIG_MWAVE is not set
# CONFIG_RAW_DRIVER is not set
# CONFIG_HPET is not set
# CONFIG_HANGCHECK_TIMER is not set

Part of my xorg.conf:

Code: Select all

Section "Module"

    SubSection  "extmod"
      Option    "omit xfree86-dga"   # don't initialise the DGA extension
    EndSubSection

    Load        "type1"
    Load        "speedo"
    Load        "freetype"
    Load        "v4l"
    Load        "dbe"
    Load       "glx"
    Load       "dri"
    Load       "GLcore"
EndSection

...<snip>...

Section "Device"
    Identifier  "device1"
    Driver      "radeon"
    Option	"AGPMode" "4"
    Option	"EnablePageFlip" "true"
EndSection

...<snip>...

Section "DRI"
    Mode 0666
EndSection

kmare · Post by **kmare** » Mon Jan 24, 2005 9:36 pm

yours is clearly a problem with the ati opengl 3d drivers. The problem that happens to most of us is that the freeze happens randomly (even without 3d accelleration enabled...)

PS: I'm sorry but as I don't owe an ati I can't help you with your problem... I'm a happy nvidia user

lookinin · Post by **lookinin** » Mon Jan 24, 2005 9:47 pm

kmare wrote:yours is clearly a problem with the ati opengl 3d drivers. The problem that happens to most of us is that the freeze happens randomly (even without 3d accelleration enabled...)

PS: I'm sorry but as I don't owe an ati I can't help you with your problem... I'm a happy nvidia user

Thanks kmare, I'm not using the actual ATI binary drivers, but the xorg implementation instead. I'm not convinced, at least with my issue, that it's ATI specific, but more likely xorg. My symptoms are strikingly similar to some in this thread that are using NVidia on xorg as well.

Thanks for the reply

dibblethewrecker · Post by **dibblethewrecker** » Tue Jan 25, 2005 11:11 am

to update my adventures with custom kernels affecting my problem

i compiled 2.6.10-ck5 it works perfectly!

i added the fbsplash and reiser4 patches and later the clock tick speed to 200 (don't ask) and it is back to the old.

i am now recompiling again without fbsplash

Mysterious Freeze

It's not Nvidia