X.org Lockups

Message

frilled · Post by **frilled** » Wed Jan 19, 2005 5:21 pm

Just to get us into a new direction: Does anybody get those X freezes with mouse moving without any action on keyboard or mouse? I recall it always happened whan I was doing something, most of the time switching virtual desktops.

I *might* be a problem with event processing on a low level.

I recently dug out a BIOS update for the mainboard of my "crashbox" (it's like three or four years old, so I was very surprised to find it). Flashed it about two weeks ago, re-enabled all the stuff that I took out (like APIC and DRI) and have so far been stable. This does not mean the box *is* stable (I was let down by my expectations more than once), but two weeks is the best I got until now.

So maybe there is some weird ACPI or device management function not working correctly in combination with some boards, I have no idea. Just wanted to add in a new hint.

frilled · Post by **frilled** » Wed Jan 19, 2005 5:24 pm

FreeFly42 wrote:So it makes me wonder if it's actually an NPTL problem?

I was thinking about that, too. But I am very confident that at the time the crashes started I was still running XFree and glibc without NPTL. I later rebuilt the box with NPTL enabled from bootstrap, using xorg with no noticable differences,

praseodymium · Post by **praseodymium** » Wed Jan 19, 2005 5:28 pm

My time is always set to the future when Xorg freezes...

And I also had the problem before the I enabled NPTL/linux-kernel headers.

drowbot · Post by **drowbot** » Wed Jan 19, 2005 5:31 pm

My Xorg freezes with the mouse moving and no action from the keyboard or mouse. I can start compiling something or play an MP3 or video and walk away and it is guarenteed to freeze.

cayenne · Post by **cayenne** » Wed Jan 19, 2005 10:37 pm

drowbot wrote:My Xorg freezes with the mouse moving and no action from the keyboard or mouse. I can start compiling something or play an MP3 or video and walk away and it is guarenteed to freeze.

I've seen this too. Actually, I think last time mine did this (iBook), I had Kmail open...walked out of the room to do something for a few minutes.....came back and it was locked up. Mouse would move...but, nothing responded.

drowbot · Post by **drowbot** » Wed Jan 19, 2005 10:43 pm

Funny thing is, I think that the system is still going, just X is hosed. Because, like I said, I can be compiling something when this happens, but my HDD light is still flickering on and off, like the compiler is still cranking away. And again, the mouse still moves, but I can't click on anything and the kbd won't respond.

Eskarel · Post by **Eskarel** » Thu Jan 20, 2005 2:09 am

Just thought I'd say that I fixed mine, it was a kernel panic caused by bad wireless NIC drivers. The drivers are now better and I xorg/my pc no longer freezes.

Ricky · Post by **Ricky** » Thu Jan 20, 2005 4:40 am

I get a complete lock-up when I try and start xorg with both monitors (two radeon cards: AGP All-in-Wonder (7000), PCI Radeon 7200) with DRI loaded. I can get both monitors w DRI commented out and DRI on one monitor. It completely locks up my system and I have to do a hard reboot.

Any thoughts?

My xorg.conf:
http://s88382932.onlinehome.us/files/xorg.conf
and as far as Xorg.0.log gets before total meltdown:
http://s88382932.onlinehome.us/files/Xorg.0.log

I'm currently on xorg-x11-6.8.1.902, but this has been a problem since 6.8 came out.

Thanks!

frilled · Post by **frilled** » Thu Jan 20, 2005 2:06 pm

Ricky wrote:Any thoughts?

My xorg.conf:
http://s88382932.onlinehome.us/files/xorg.conf

Hm, does not look too bad I think. There are some caveats in your config, though.

As far as I know, GLcore is deprecated, "dri" and "glx" should suffice.

Try disabling AGPFastWrite on a general basis. There is no penalty in terms of speed or whatever. Benefits are that your machine might become more stable since FastWrites are known to cause problems for quite some hardware combinations.

Disable RenderAccel for a start and see if it gets better.

Do not load RENDER and Composite Extensions and see if it gets beter.

If it does, start re-enabling one by one and see when it starts crashing.

LinuxRocks · Post by **LinuxRocks** » Thu Jan 20, 2005 2:59 pm

In reply to the time change...

I don't have lockups too often (But when I do its at a most critical time (Like updateing a server or something) but when it has happend, my time will change to Moscow time (That is what it says on the K menu anyway).

Anyone running SuperKaramba? I have NEVER had it lock up in Fluxbox, every time it has locked up was in KDE.

Joe

drowbot · Post by **drowbot** » Thu Jan 20, 2005 4:20 pm

I've branch this off into another thread, hopefully to narrow the problem down: http://forums.gentoo.org/viewtopic.php? ... 41#2000141

Ricky · Post by **Ricky** » Fri Jan 21, 2005 8:44 am

wgi wrote:
Ricky wrote:Any thoughts?

My xorg.conf:
http://s88382932.onlinehome.us/files/xorg.conf
Hm, does not look too bad I think. There are some caveats in your config, though.

As far as I know, GLcore is deprecated, "dri" and "glx" should suffice.

Try disabling AGPFastWrite on a general basis. There is no penalty in terms of speed or whatever. Benefits are that your machine might become more stable since FastWrites are known to cause problems for quite some hardware combinations.

Disable RenderAccel for a start and see if it gets better.

Do not load RENDER and Composite Extensions and see if it gets beter.

If it does, start re-enabling one by one and see when it starts crashing.

Thanks for the suggestions! Unfortunately they didn't seem to do the trick...

I commented out glcore and composite and pretty much all of the driver options, and still got the same hard-lock.

When it locks up, it seems to have initialized the first screen, but it locks before I get a cursor or the second screen comes to life.

Any more ideas?

frilled · Post by **frilled** » Fri Jan 21, 2005 3:34 pm

Ricky wrote:Any more ideas?

It's getting thinner now...

dmesg | grep -i mtrr -- If you get errors, it is probably because of vesafb. I got around it by speficying vtotal=<VideoMemInMBs> as an option (was called vram= until recently) of the video:vesafb parameter. There are forum threads about manipulating the MTRR entries, too.

Look for strange messages and errors in /var/log/messages, especially containing the keyword PCI.

Try with NOAPIC kernel option. Try with PCI=noacpi.

Phew. No more really bright ideas.

Try without DualHead first, maybe.

drowbot · Post by **drowbot** » Fri Jan 21, 2005 4:23 pm

Well, mine is finally fixed. I took AGPGART out of the kernel, recompiled, and set NvAGP to 2 in my xorg.conf. Now my system is 100% stable. This may not work for everybody, since not everybody has an NVIDIA card. But there should be similar options for other cards, I would think.

Jimmy Jazz · Post by **Jimmy Jazz** » Fri Jan 21, 2005 8:04 pm

Ricky,

same problem here with a geforce Ti4200, as lon as i try to configure xorg in dualhead or twinview mode.

The problem seems to affect both nvida and ati cards

Jj

Ricky · Post by **Ricky** » Sat Jan 22, 2005 12:33 am

Well, unfortunately none of that seems tomake any difference.

I'm currently running w/o the framebuffer. It does work with just a single display, and either display works individually, I just can't get both to load up w/ DRI. I also tried it as root, and it's the same thing.

vola · Post by **vola** » Sun Jan 23, 2005 11:20 am

When i'm using ndiswrapper it happens for me too.
Without ndiswrapper all works fine.

Regards

TaTaE · Post by **TaTaE** » Mon Jan 24, 2005 8:17 am

Hi everyone. I'm glad to tell you that i had no desktop freeze or crash for a week now. I have kernel2.6.10-r4/xorg681/kde332/nvidia(tm) module and i got rid of freezes after i did two things:
1. I disabled nvclock.

2. I got rid of plastik style, which seemed to have a stability problem with dragging scrollbars. I used other several styles which had now problem at all by now. Now I have enabled alloy.
I was experiencing crashes several times a day and it seems they're gone now.
My advice: anyone not using (I am) pure konsole, disable gpm or make
a softlevel for konsole. Another thing is that vesafb is not really necessary when you only need kde+nvidia combination, so disable your bootsplash/gensplash (saving it in grub for your m$-fan friends

).

Cheers.

Ricky · Post by **Ricky** » Wed Feb 09, 2005 6:35 pm

Well, for the record, xorg-x11-6.8.1.904 doesn't seem to help any.

In case 6.8.2 doesn't clear this up, are there any cvs ebuilds for xorg floating around? I'm also kind of interested in playing around w/ the gatos drivers that have been merged into cvs.

Ricky · Post by **Ricky** » Fri Feb 11, 2005 9:59 pm

Well, looks like 6.8.2 isn't doing any better...

cvs ebuild anyone?

Hieronymus · Post by **Hieronymus** » Sat Feb 12, 2005 4:04 pm

Ricky wrote:Well, looks like 6.8.2 isn't doing any better...

cvs ebuild anyone?

Try this page:
http://xorg.freedesktop.org/wiki/X11R68 ... 15a9f4c696

grz Hieronymus

MrMoustache · Post by **MrMoustache** » Sun Feb 13, 2005 2:13 am

Total noob here, be gentle. I've got lockups too.

System is an EPoX 8KHA+ (VIA KT266A chipset), Athlon XP 1900+, 512M, GeForce 4 MX440, SB Live!, 3COM 905B ethernet, Hauppauge PVr250 card. Running kernel 2.6.10-r6, KDE 3.3.

This is box I built to run MythTV. A couple months ago it was running SuSE 9.0 but with a Hauppauge WinTV card and a different NIC. When I updated SuSE from the web (lots of new packages, new KDE), I started getting random lockups. I thought it was the NIC because I saw "No TX..." messages on some reboots, or the boot would hang getting a DHCP address. The 3COM NIC worked better, but the SuSE update broke a lot of other things (for example, getting errors running apps in KDE).

So I installed the PVR250, wiped it and began loading Gentoo. Today I've been in KDE, reading docs in FireFox, playing Mahjongg

and emerging packages in a Konsole window. I've been doing this for a couple hours now when I got my first lockup.

I rebooted, restarted Firefox, restarted the emerge. A couple minutes later I get the second lockup. So I reboot again, open only a single Konsole window and start the emerge. Within about 20 seconds I get a lockup. Another reboot, and this time I'm not starting X, and just do the emerge from the terminal window. Good ol' 80x25. No problems emerging the myth package I was on. Sounds like a possible heat problem? Well, as I said, I was in KDE for a couple hours before the lockups started, surfing in FireFix, playing a few games.

After the emerge was done, I started X and walked away with no apps open. A couple minutes later I come back and it's locked up again.

The last things I emerged were corefonts, mythtv (mySQL took a long time to compile) then it bombed and wanted me to re-emerge Qt which I did (smoking gun?). Then I emerged mythvideo, followed by mythdvd... which is where the lockups started.

I hope that helps.

juniper · Post by **juniper** » Fri Feb 18, 2005 6:17 pm

Hello all,

I have an ati radeon 9200se and i am experiencing some xorg lockups (hard ones, sometimes you can log in via ssh and reboot and sometimes not). when i can log in X is taking up 99% of the cpu. anyways, what happens is that X runs for about a day, and it crashes (with the xscreensaver on) overnight.

i have kernel-2.6.9, but these lockups also occurred with a 2.6.10 kernel. i am using the native xorg radeon drivers (for the 9200 card it gives 3d). further i am using xorg-6.7.0-r2.

the reason why i am using this version of X is that i have an ati 9000 in another desktop and it is rock solid stable with that X and its native drivers.

i don't know what else to say. i checked the various logs and they were clean (/var/log/messages and /var/log/Xorg.0.log). the only log that says anything of interest is the users .xsession-errors log and it says

*** attempt to put segment in horiz list twice

here is my xorg.conf file

Section "ServerLayout"
Identifier "X.org Configured"
Screen 0 "Screen0" 0 0
InputDevice "Mouse0" "CorePointer"
InputDevice "Keyboard0" "CoreKeyboard"
EndSection

Section "Files"
RgbPath "/usr/X11R6/lib/X11/rgb"
ModulePath "/usr/X11R6/lib/modules"
FontPath "/usr/share/fonts/misc/"
FontPath "/usr/share/fonts/TTF/"
FontPath "/usr/share/fonts/Speedo/"
FontPath "/usr/share/fonts/Type1/"
#FontPath "/usr/share/fonts/CID/"
FontPath "/usr/share/fonts/75dpi/"
FontPath "/usr/share/fonts/100dpi/"
EndSection

Section "Module"
Load "extmod"
Load "dri"
Load "dbe"
Load "record"
#Load "xtrap"
Load "glx"
Load "speedo"
Load "type1"
Load "freetype"
EndSection

Section "InputDevice"
Identifier "Keyboard0"
Driver "keyboard"
EndSection

Section "InputDevice"
Identifier "Mouse0"
Driver "mouse"
#Option "Protocol" "auto"
Option "Protocol" "IMPS/2"
Option "ZAxisMapping" "4 5"
Option "Device" "/dev/input/mice"
EndSection

Section "Monitor"
#DisplaySize 330 240 # mm
Identifier "Monitor0"
VendorName "DEL"
ModelName "DELL P780"
HorizSync 30.0 - 85.0
VertRefresh 48.0 - 120.0
Option "DPMS"
EndSection

Section "Device"
### Available Driver options are:-
### Values: : integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz"
### [arg]: arg optional
#Option "NoAccel" # [<bool>]
#Option "SWcursor" # [<bool>]
#Option "Dac6Bit" # [<bool>]
#Option "Dac8Bit" # [<bool>]
#Option "ForcePCIMode" # [<bool>]
#Option "BusType" # [<str>]
#Option "CPPIOMode" # [<bool>]
#Option "CPusecTimeout" # 
#Option "AGPMode" # 
#Option "AGPFastWrite" # [<bool>]
#Option "AGPSize" # 
#Option "GARTSize" # 
#Option "RingSize" # 
#Option "BufferSize" # 
#Option "EnableDepthMoves" # [<bool>]
#Option "EnablePageFlip" # [<bool>]
#Option "NoBackBuffer" # [<bool>]
#Option "DRIReinit" # [<bool>]
#Option "PanelOff" # [<bool>]
#Option "DDCMode" # [<bool>]
#Option "MonitorLayout" # [<str>]
#Option "IgnoreEDID" # [<bool>]
#Option "OverlayOnCRTC2" # [<bool>]
#Option "CloneMode" # [<str>]
#Option "CloneHSync" # [<str>]
#Option "CloneVRefresh" # [<str>]
#Option "UseFBDev" # [<bool>]
#Option "VideoKey" # 
#Option "DisplayPriority" # [<str>]
#Option "PanelSize" # [<str>]
#Option "ForceMinDotClock" # <freq>
Identifier "Card0"
Driver "radeon"
VendorName "ATI Technologies Inc"
BoardName "Unknown Board"
BusID "PCI:1:0:0"
EndSection

Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
DefaultColorDepth 24
SubSection "Display"
Viewport 0 0
Depth 8
Modes "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
SubSection "Display"
Viewport 0 0
Depth 16
Modes "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
SubSection "Display"
Viewport 0 0
Depth 24
Modes "1280x1024" "1152x864" "1024x768" "800x600"
EndSubSection
EndSection

Section "dri"
Mode 0666
EndSection

i recently changed the driver to 'radeon' from 'ati' (i don't think that made a difference, i think the same modules are loaded, in particular the radeon module was loaded in both cases. oddly, when it was supposed to call ati the radeon module was loaded, there was no ati module loaded as indicated by lsmod) and commented out the "Load "xtrap"" under modules to load. i also unmerge xscreensaver. it hasn't crashed yet, but it has only been running for a couple of hours. I guess i will know over night...

does anyone have any obvious suggestions??? something crazy in my xorg file???

thanks.

j

juniper · Post by **juniper** » Sun Feb 20, 2005 4:30 pm

well, the changes i made did something, there are no more crashes.

i suspect that the fact that i unmerged xscreensaver is the one that did it, which leads me to suspect that opengl apps are crashing the system. that sucks.

has anyone tried the a newer xorg with the new ati proprietary drivers and the radeon 9200se card? maybe that will help.

j

meowsqueak · Post by **meowsqueak** » Mon Feb 21, 2005 12:13 am

I get this problem (started last week probably after one of my regular world updates) several times a day now. It would appear that X jumps up to ~99% CPU usage and I am unable to move my mouse cursor across to my second monitor (dual head). Clicking on things does nothing, nor does typing anything.

A few observations:

- It seems to happen at random intervals - once it happened halfway thru X starting up, leaving everything partially drawn and no wallpaper.

- It seems unrelated to me actually doing anything. I haven't noticed myself dragging/clicking/typing anything when it happens. I just come to move the mouse and it's gone to lunch.

- I can ssh into the machine - it's fine, just X running at ~99% CPU. If I kill X, everything returns to normal except the system has no control over the video any more and I may as well reboot or live in ssh-land permanently

- CTRL+ALT+BKSPC doesn't kill X, however the magic sysreq combo (S, U, B) works.

- I have tried Linux 2.6.8.1 and 2.6.10 - no difference.

- I have just re-emerged Xorg - we'll see if that helps...

- Perhaps related, perhaps not, but when I install the SwitchProxy Firefox plugin it causes Firefox to completely lock up. However X is still fine and I can open another window and kill Firefox. The only reason I wonder if this is related is that it started happening around the same time as the lockups began.

I'm using Nvidia with dual-head (seperate screens, not xinerama) with fluxbox ~x86 and xorg 6.8.2.

Could this be a font issue?

X.org Lockups

Two more cents

I've got lockups too.