[X1650 Pro] Xorg lockup when using xserver-xorg-video-ati 1:6.9.0+git20080826.a3cc1d7a-2ubuntu3

[Tim Besard]

Binary package hint: xserver-xorg

After working a while, the Xserver sometimes hangs with following errors spammed in the Xorg.log:
[mi] EQ overflowing. The server is probably stuck in an infinite loop.
[mi] mieqEnequeue: out-of-order valuator event; dropping.

When such an error occurs, my music just keeps on playing, the mouse moves too, but all other activity is prohibited. I need to disconnect my PS2 keyboard (due to unresponsive), connect it to the USB port, and execute the magic SEIUB keystrokes to safely reboot.

The error has been reported by other people at the ubuntu fora: http://ubuntuforums.org/showthread.php?p=5875714

System details:
Ubuntu Intrepid, fully updated
ii xserver-xorg 1:7.4~2ubuntu4 the X.Org X server
ii xserver-xorg-video-ati 1:6.9.0+git20080826.a3cc1d7a-2ubuntu3 X.Org X server -- ATI display driver wrapper

No .crash file is generated, as the xserver doesn't crashes, but only loops at a certain level.
[lspci]
00:00.0 Host bridge [0600]: Intel Corporation 82865G/PE/P DRAM Controller/Host-Hub Interface [8086:2570] (rev 02)
     Subsystem: Intel Corporation 82865G/PE/P DRAM Controller/Host-Hub Interface [8086:2570]
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon X1650 Pro [1002:71c1] (rev 9e)
     Subsystem: PC Partner Limited Device [174b:0850]

[Wouter Stomp (wouterstomp-deactivatedaccount)]

Do you experience the same problem when running hardy or only on intrepid? Also could you attach your /etc/X11/xorg.conf file and the output of lspci -vvnn? Thanks

[Tim Besard (maleadt)]

Only on intreprid, and only during the last week. Can't pinpoint an exact update in it though.

I am using, forgot to mention it, an ATI X1650 AGP. Other user with this problem is using an ATI X1600, so it could be an xserver-xorg-video-radeon R500 bug.

[Tim Besard (maleadt)]

I indeed have EXA enabled, this to support video playback when having compiz enabled.

[Bryce Harrington (bryce)]

Thanks. (I declined the Intrepid task since we do not ship EXA turned on by default, due to concerns about bugs like this one. But we hope to switch it on maybe for Jaunty, so this bug will be worth tracking.)

[Martin Nowack (martin-nowack)]

Sorry to say but it also happens without EXA enabled, see attached log.

[dkasak (dan-entropy)]

I'm the user who originally reported the bug in the forums. I was going to log a bug myself, but someone beat me to it ... unfortunately getting some important details wrong.

This bug most certainly also affects XAA ( which I did clearly note in the forums ). Could someone PLEASE at least rename this bug and remove the 'with EXA enabled' bit. Alternatively I can enter a separate bug for the default ( XAA ) case.

> I declined the Intrepid task since we do not ship EXA turned on by default, due to concerns about bugs like this one.

Frankly I'm a bit perplexed at this comment. Consensus with the radeon developers is that EXA is more stable than XAA at this point, fixing some lockups when starting some OpenGL apps, amongst other things. I have used EXA on my Radeon for at least a year. Anyway ... I'm not too concerned about Ubuntu not 'supporting' EXA at this point ... but I am concerned that I can't start X without DRI disabled.

So I've done some 'git bisect' work with xf86-video-ati. I went back a couple of months. Note that I've been running 'git current' of xserver, mesa, and friends on Gentoo for the past couple of years ( yeah, cvs current before the git migration ). Anyway, I *can't* find a particular point in the xf86-video-ati history that broke things. ie I don't think the issue is in xf86-video-ati. I have also git-bisected libdrm and the radeon module, back as far as I can actually build things. No luck here either. Now I'm building a 2.6.26 kernel. If this doesn't work out, then I'm betting the problem is in xserver. I will of course report any successes I have.

But in the meantime, someone please rename this bug.

[dkasak (dan-entropy)]

Lockup also occurs with a vanilla 2.6.26 kernel ( with included DRI modules ). At this point, I give up. As noted earlier, I'm starting to think this might be in xserver. I talked to some devs in the #radeon channel at freenode, and they said that it's definitely *NOT* in mesa, as mesa isn't involved in X startup.

I think for now I'll switch back to Gentoo where I can easily block the upgrade to xserver-1.5, and maybe do some more git-bisect work on xserver when I get some time ( note: I don't get much spare time ).

[ChrisTomalty (christomalty-deactivatedaccount)]

I'm running a default Intrepid release downloaded 13.10.08. This occurs in a fresh install. I don't know what it means, but it does. Video Card: ATI Mobility FireGL V3100 w/ 64MB VRAM. Music continues to play and mouse is moveable. CTRL-ALT-BKSPC doesn't work.

[dkasak (dan-entropy)]

LegoAddict: I have some bad news for you. My errors turned out to be because of catastrophic hardware failure. Rebooting into Windows confirmed this - my Windows desktop became corrupted almost instantly, and it would flash on & off, turning down the graphics acceleration until it became stable ( I wish X & drivers had done this ). In the end, Windows refused to activate the 3D accelerator at all, and disabled almost all 2D acceleration.

As for exactly what caused this, I'm not sure. Perhaps my hardware overheated ( possible ). Perhaps it was the result of the very strange display corruption I always got when switching from X <==> VT or back ( lockup, with dark colours that fade to light colours over time ). It always looked kinda scary. I'm a little suspicious it was bad for the hardware.

Anyway, I would check that you can still activate ( and use ) the 3D accelerator in Windows before continuing. As people in the #DRI channel told me, [mi] EQ overflowing & [mi] mieqEnequeue: out-of-order valuator event; dropping is a symptom of the hardware locking up. It can be caused by multiple things, and unfortunately hardware failure is one of them.

[Bryce Harrington (bryce)]

Thanks for the analysis dkasak. maleadt I'm going to assume your issue is also hardware related like dkasak's; feel free to reopen if you can pinpoint it to software (if you can revert to an older or newer version and make the issue go away, that would be a sufficient indication that it's software).

[Tim Besard (maleadt)]

dkasak: please don't blaim me for writing an incorrect bug report, as in my situation the bug only triggered with EXA enabled.

Bryce Harrington: I will continue to analyse every X crash which occurs, but it has been a while I saw the EQ overflow in my logs. Maybe an update corrected this issue? Because I doubt this is a case of hardware failure, as the ATI card I'm using is brand new (replacement model). I will report back in if I get another case of this bug.

[Laurent Bigonville (bigon)]

I'm hit by this bug too with a mobility 9700

[Laurent Bigonville (bigon)]

I've rebooted on a feisty livecd and no problem with the X server.
If I disable the hw acceleration with the "NoAccel" flag no problem too

[loopx (loop-x)]

HI,

I'm running Intrepid Beta and I get same lockup. I'm using "radeon" driver on a radeon mobility 9700 128Mb.

I'm using KDE4.1.2 with 3D desktop on (kde, not compiz).

I don't think my problem is a hardware failure; this problem is certainly X related.

[Brian Fallik (bfallik)]

I also recently encountered this bug after an Intrepid update. Using Intrepid over the last few weeks ago did not exhibit this at all. After an update yesterday (of around 250 packages), it's happened twice in ~2 hours.

I have no reason to believe it is a hardware failure. I'm on a Thinkpad T41 with ATI M22 [Mobility Radeon X300] standard install (no fglrx). Anything else I can provide? This may be a showstopper for me using Intrepid, so I'd like to help any way I can.

[Daniel Holm (danielholm)]

My computer freezes every other minute. And then is runs as usual for a while and then freezes again. But only the screen. Music still plays and I can still write.

I also got it from a big upgrade the other day.

[Brian Fallik (bfallik)]

What was the reason for marking this "Declined for Intrepid"? Is there any way to get this accepted (and fixed) for Intrepid?

[dkasak (dan-entropy)]

It was declined because the subject mentions EXA ( see my post above on this ). I don't have permission to change the subject line. Up until the point when I discovered my hardware was fried, I was going to report another bug and make sure I didn't mention EXA. If others are having the same problem as me, then I'd urge _extreme_ caution ( especially for Mobility users ), as I'm more than a little suspicious that it was software that fried my hardware. All my problems started immediately after an X & xf86-video-ati upgrade.

[Tim Wright (timw)]

Same issue on a vanilla Intrepid upgrade and machine-generated xorg.conf (ati driver) on a Radeon 9800Pro.
As such, this needs to be fixed for Intrepid. I'm not running some experimental/non-standard configuration, I am running the default configuration, and a card that was previously stable in 8.04 now hangs in Intrepid Beta with the "EQ overflowing" errors in Xorg.0.log.

[Brian Fallik (bfallik)]

My issue has gone away as of another large update on 10/23-ish. I haven't experienced the freeze after many hours of testing spread across several days. Very strange.

[Laurent Bigonville (bigon)]

Seem fixed here too since last Xorg upgrade

[Tim Wright (timw)]

It hasn't happened on my system for a few days either. It's always difficult to prove a negative, i.e. how long is needed to feel confident that the problem has gone away, but it's looking good so far.

[Martin Nowack (martin-nowack)]

I was hit two times today.
And I've got an up to date system.

[Tim Wright (timw)]

Yep, still broken here,
came back to login this morning and the screen was non-responsive. ssh in and in /var/log/Xorg.0.log, I see
[mi] EQ overflowing. The server is probably stuck in an infinite loop.
[mi] mieqEnequeue: out-of-order valuator event; dropping.
repeated multiple times.

This is xserver-xorg-video-radeon 1:6.9.0+git20081003.f9826a56-0ubuntu2 which I believe is the latest.

[Laurent Bigonville (bigon)]

I was hit again by this issue too

I discover that the X server lockup only when my external usb mouse is disconnected

USB mouse connected -> no lockup
USB mouse disconnected -> lockup

[Daniel Holm (danielholm)]

I also still got this problem. But I do not get the EQ error in the Xorg0.log, actually. But the system is as unstable as when I got them. I can write for a while without the text appearing on the screen. And I can scroll without anything happends for a few seconds, and thing like that. It completely freezes for a while.

And also I got a nVIDIA card, not a ATI(AMD).

[Lorenco Trichardt (trichalo)]

I also have this problem ..... (But I also get hard lockups ?)

(II) intel(0): Modeline "1400x1050"x0.0 108.00 1400 1448 1560 1688 1050 1051 1054 1066 -hsync -vsync (64.0 kHz) (II) intel(0): Modeline "1400x1050"x0.0 89.97 1400 1448 1560 1688 1050 1051 1054 1066 -hsync -vsync (53.3 kHz) (II) intel(0): EDID vendor "LEN", prod id 16418
[mi] EQ overflowing. The server is probably stuck in an infinite loop.
[mi] mieqEnequeue: out-of-order valuator event; dropping.
[mi] EQ overflowing. The server is probably stuck in an infinite loop.
[mi] mieqEnequeue: out-of-order valuator event; dropping.

[Tim Wright (timw)]

Thank Lorenco.
Looking at the output Lorenco sent, he appears to have an Intel video chipset. This is clearly not an ati server problem, but a core Xserver bug. It needs to be reclassified.

[Tim Wright (timw)]

I hope this is the correct package for the core X Server. It's a bit hard to ttell since the name is arbitrarily different to the '.deb' package that gets installed (xserver-xorg), and the exploded contents in the dropdown are truncated.

[TheBus (tw2006)]

i have the same issue with openchrome driver (VIA KN8000) on a benq laptop. If i use vesa on xorg.conf it work but the resolution is very bad. With 8.04 it work perfect

[Ellsworth Wente (ell-wente)]

I have this same problem since upgrading to intrepid. Running two Nvidia 8400GS's with three heads and the nvidia x86_64-177.80 drivers... The system has run flawlessly for 6 months on heron.

[dkasak (dan-entropy)]

Oh gee golly gosh I wish people would PAY ATTENTION to what they're doing. This has nothing to do with nvidia, intel, via. It has nothing to do with camels, horses or pigs. It's about an ATI lockup. Got a different bug? It needs to be reported separately.

[Ellsworth Wente (ell-wente)]

perhaps there's an ATI issue that triggers this same hang... my reason for posting my experience with nvidia was merely to point out that this same mieqEnqueue overflow error seems to be common to multiple graphics cards, and may actually be xserver related. I see a duplicate bug along these lines in #259808, so apparently it's being seen in many different environments. gnome, kde on nvidia, ati, the common theme (to me) is, this is showing up with the current xorg release that shipped with intrepid... perhaps we need yet another bug report? my apologies if it was inappropriate to post in this thread, just trying to help...

[Tim Wright (timw)]

On what basis do you make the assertion that this is an ATI bug Dan? The fact that it *affects* the ATI driver does not mean that it is a bug in the ATI driver. It could very easily be a core Xserver bug unless you have evidence to the contrary. The fact that multiple different GPU drivers are exhibiting the same issue on Intrepid on hardware that worked flawlessly on Hardy would tend to point towards a common issue, no?

[dkasak (dan-entropy)]

'EQ overflowing' is a symptom of the graphics card locking up, which is almost always caused by the drivers. This advise came from a number of devs in the #radeon IRC channel. As for why I believe it's an ATI bug, this previous comment, together with the fact that the bug first appeared when I updated xf86-video-ati & mesa, makes me pretty certain it's an ATI bug. Also note that after a couple of such video card lockups my hardware is now FRIED ( I can only start X with DRI disabled, and Windows locks hard when booting - I have now bought a new laptop and use my old one as a server ). I find it hard to believe that it's an xserver issue, as the X server doesn't talk directly to the video card - this is what the drivers do.

[Laurent Bigonville (bigon)]

And what about the fact that the lockup happens only if I remove my usb mouse? I can reproduce that 100% of the time

[dkasak (dan-entropy)]

Who knows? As I mentioned before, 'EQ overflowing' is a symptom of a hardware lockup. But speaking generally, if you've got a *different* set of symptoms, then this indicates a *different* bug. Anyway, I'm no longer a Ubuntu user; in fact not even using xf86-video-ati any more ( got me a Radeon 3100 or something, need fglrx ). So by all means, pile on reports for unrelated bugs, and see how quickly things progress from here - this will become a bit-bucket that no-one can even hope to solve with so many different hardware combos, triggers and symptoms.

[Tim Wright (timw)]

Well, mine is definitely an ATI card card (Radeon 9800Pro). The lockup happens with both XAA and EXA (I tried turning on EXA to see if that made a difference), Interestingly, I didn't get the "EQ overflowing" messages when I had EXA enabled. My lockups have, so far, always happened when the machine had gone into screensaver mode and subsequently DPMS powersave mode. Moving the mouse brings the display out of powersave, and the "X" cursor is visible and movable, but the keyboard is non-responsive, and it's necessary to ssh in and kill the server to get things back. Not sure how feasible it is to try to get a backrev version (a la current Hardy) of the ati driver onto my system to check. I would imagine I'd have to try to find the relevant source and build it?

[dkasak (dan-entropy)]

> Interestingly, I didn't get the "EQ overflowing"
> messages when I had EXA enabled.

At first I noticed the same kind of thing. If I kept trying ( which I did ), I eventually got the 'EQ overflowing' with EXA enabled - maybe 30% or the time or something like that. I assumed things went bad a little bit earlier with EXA enabled, mainly because with XAA I'd get GDM outputting *something* ( http://entropy.homelinux.org/gdm_xaa.jpg ), but with EXA I'd just get a back screen with a cursor.

[Lorenco Trichardt (trichalo)]

I have a complete broken session now after the kernel upgrades this morning.
The whole screen goes corrupt.... here I am sitting in a "failsafe terminal" starting firefox.

I can see garbage forming where the boxes should be. It starts normal and then as soon as compiz kicks in it goes south....
Help!

[dkasak (dan-entropy)]

Lorenco: Sure. I hear what you're saying. The thing is ... my bike is only getting 5.5 litres / 100km these days, and this all started happening after Windows Vista was released. Do you think this has anything to do with this bug?

[Bryce Harrington (bryce)]

The "EQ overflowing" error message is a generic error that occurs after the Xserver has hung for any old reason. So it is most likely that each reporter here is seeing an xserver hang for a *different* reason, but since this esoteric error message is printed in each case, it is confusing and hard to tell the lockups apart. (See bug #259808)

Since it sounds like maleadt's original issue has gone away, I think that may mean the original bug is gone. So I'd like to ask each person still seeing lockups to enter a NEW bug on your issue, and I'll close this one as fixed.

@maleadt, if you start seeing this issue again, please feel free to reopen. Everyone else, file a NEW bug, don't reopen this one.

[Lorenco Trichardt (trichalo)]

I had a chuckle about the Vista comment. Fair comment, btw, I fixed it by disabling compiz in safe xterm mode ;)

Question: I have seen a lockup on intel driver and there is a current bug open for that. Is that related to this? (Hard lockup) Or do you want me to open a new bug for my overflow error ?

[Bryce Harrington (bryce)]

On Fri, Nov 07, 2008 at 04:05:30AM -0000, Lorenco Trichardt wrote:
> I had a chuckle about the Vista comment. Fair comment, btw, I fixed it
> by disabling compiz in safe xterm mode ;)
>
> Question: I have seen a lockup on intel driver and there is a current
> bug open for that. Is that related to this? (Hard lockup) Or do you
> want me to open a new bug for my overflow error ?

Nope, file a new bug. :-)

If you suspect it might be a dupe of another bug, mention that bug ID,
and someone else can review and make the decision about duping it.
Duping bugs is cheap and easy. Untangling two unrelated bugs in a bug
report is hard.

In general, when we troubleshoot bugs we usually only consider the
original reporter. If you "me-too" your report onto someone else's,
your issue probably won't be investigated.

Also please note you're not having an "overflow error". That error
message is just a side effect. You are having an "X lockup", and that
is the issue that has to be troubleshot.

So yeah, file a new bug. One bug report per issue, one bug report per
person. That's easiest for us. :-)

[Jared Evans (jarednevans)]

I am seeing the same errors on Intrepid with standard updating (not bleeding-edge) and a Nvidia video card. Running Nvidia drivers and compiz. I'm seeing hard lock ups on X. Have three so far today and never saw this happening before today.

Xorg.0.log reports these errors before the hard lockup requiring a hard reboot:

[mi] EQ overflowing. The server is probably stuck in an infinite loop.
[mi] mieqEnequeue: out-of-order valuator event; dropping.

[Tim Besard (maleadt)]

Op vrijdag 07-11-2008 om 00:45 uur [tijdzone +0000], schreef Bryce
Harrington:
> Since it sounds like maleadt's original issue has gone away, I think
> that may mean the original bug is gone. So I'd like to ask each person
> still seeing lockups to enter a NEW bug on your issue, and I'll close
> this one as fixed.
>
> @maleadt, if you start seeing this issue again, please feel free to
> reopen. Everyone else, file a NEW bug, don't reopen this one.

Fair enough, as it has been a while since I came across the EQ
overflowing issue and I haven't been able to find what causes it.

maleadt

[Jared Evans (jarednevans)]

After some investigation and testing... it looks like it's related to the X input devices update I did last night. USB mouse may be causing it. I am testing this right now by leaving my USB mouse disconnected and see if it freezes up again.

[Tormod Volden (tormodvolden)]

To the me-too'ers: I uploaded a test build of xorg-server 1.5.3 for Intrepid to the ~xorg-edgers PPA <https://edge.launchpad.net/~xorg-edgers/+archive>. This version will drop a backtrace in Xorg.0.log when it detects an infinite loop, and this might help to find the real issue behind it. You can include this information in your own new bug reports. Please also install the xserver-xorg-core-dbg package.

Please disable the PPA in your Software Sources after upgrading to this version, since new, more experimental versions might be uploaded later, and it's not sure you want them.

[Xavier Gnata (xavier-gnata-gmail)]

Here is the backtrace:

Backtrace:
0: /usr/bin/X(xorg_backtrace+0x26) [0x4ee236]
1: /usr/bin/X(mieqEnqueue+0x23f) [0x4cebbf]
2: /usr/bin/X(xf86PostMotionEventP+0xc4) [0x478754]
3: /usr/bin/X(xf86PostMotionEvent+0xb1) [0x478931]
4: /usr/lib/xorg/modules/input//evdev_drv.so [0x7f96d50ba83f]
5: /usr/bin/X [0x497c45]
6: /usr/bin/X [0x46f047]
7: /lib/libpthread.so.0 [0x7f96dac610f0]
8: /usr/lib/xorg/modules/drivers//nvidia_drv.so [0x7f96d6a786b7]
9: /usr/lib/libpixman-1.so.0 [0x7f96da3c8c6e]
10: /usr/lib/libpixman-1.so.0 [0x7f96da3dcbdb]
11: /usr/lib/libpixman-1.so.0 [0x7f96da3db96a]
12: /usr/lib/libpixman-1.so.0 [0x7f96da3e0984]
13: /usr/lib/libpixman-1.so.0(pixman_image_composite+0x79f) [0x7f96da3e05ef]
14: /usr/lib/xorg/modules//libwfb.so(wfbComposite+0x1ba) [0x7f96d633bc5a]
15: /usr/lib/xorg/modules/drivers//nvidia_drv.so [0x7f96d6a89556]
16: /usr/bin/X [0x531e98]
17: /usr/bin/X [0x5210da]
18: /usr/bin/X(Dispatch+0x364) [0x44d754]
19: /usr/bin/X(main+0x45d) [0x43376d]
20: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f96d9203466]
21: /usr/bin/X [0x432b49]
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
and so on and so on (more than 100lines "ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.")

or please also see https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/294941 (Xpost is a bad idea but here I thing it could help).

Xavier

[Tim Besard (maleadt)]

Lockups have returned in here, I'm installing the Xserver with backtrace mod now.

I'm still wondering though why the kernel doesn't accept the magic REISUB keystroke anymore. My keyboard is an USB one, connected to PS2 through a convertor.

[Antonio Costantino (anto-costantino)]

I originally commented on the duplicate Bug #259808. I have to say I experienced hard locks, forcing me to reboot from the power button because alt-sysreq-REISUB doesn't work on my laptop. On a fully updated kubuntu 8.10, open source ati driver for my X700, and EXA mode, I wasn't able to reproduce the lock-ups. It just happened randomly. Then I tried to enable backports, but no success: still lockups. Finally, I switched desktop effects off and no more lock-ups since then. Is anyone experiencing the same? Are desktop effects triggering Xorg to lock-up in some way?

[Gareth Bult (gareth-encryptec)]

Mmm,

The main fix for me was to turn Xinerama off, but I did experience some other weird stuff so effects are now off too .. I can confirm that with Xinerama + Effect I was locking on average every 10 mins .. now, not for a couple of weeks.

hth

----- Original Message -----
From: "Antonio C. Costantino" <email address hidden>
To: <email address hidden>
Sent: Sunday, 16 November, 2008 2:10:44 PM GMT +00:00 GMT Britain, Ireland, Portugal
Subject: [Bug 276518] Re: Xorg hangs due to EQ overflowing

I originally commented on the duplicate Bug #259808. I have to say I
experienced hard locks, forcing me to reboot from the power button
because alt-sysreq-REISUB doesn't work on my laptop. On a fully updated
kubuntu 8.10, open source ati driver for my X700, and EXA mode, I wasn't
able to reproduce the lock-ups. It just happened randomly. Then I tried
to enable backports, but no success: still lockups. Finally, I switched
desktop effects off and no more lock-ups since then. Is anyone
experiencing the same? Are desktop effects triggering Xorg to lock-up in
some way?

--
Xorg hangs due to EQ overflowing
https://bugs.launchpad.net/bugs/276518
You received this bug notification because you are a direct subscriber
of a duplicate bug.

--
Managing Director, Encryptec Limited
Tel: [phone], Mob: [phone]
Email: <email address hidden>
Statements made are at all times subject to Encryptec's Terms and Conditions of Business, which are available upon request.

[Tormod Volden (tormodvolden)]

Antonio, I filed bug #295904 for hangs on X700, also happens more often with compiz, likely because compiz uses 3D.

[Simon Skrede (simon-skrede)]

I experienced this problem consistently up until yesterday when I installed the new beta 180.06 nvidia driver. It is now completely gone in my environment (Dell laptop with Quadro FX 1600M).

[Chow Loong Jin (hyperair)]

I just got hit by it on Intel GMA965. This is the first time. Some extra things worth considering:
1. I'm using a multi-head setup (1280x800+0+224 + 1280x1024+1280+0) with Compiz running.
2. I'm using a patched mesa, to break the 2048x2048 false 'hardware' limit placed on Intel GMA965. Patch found here: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/146298/comments/16

[kreggz (jasonkregting)]

I upgraded to Intrepid on my Acer when it first came out, unfortunately I had to roll back to Hardy because of this bug. I have now reinstalled Intrepid and the bug is back.

When it locks up I can move my mouse and music plays. Unfortunately the keyboard and mouse clicks are unresponsive, my cacti logging continues so its not a full crash.

(II) RADEON(0): EDID vendor "ACR", prod id 44370
(II) RADEON(0): I2C device "DVI-0:ddc2" registered at address 0xA0.
(II) RADEON(0): I2C device "DVI-0:ddc2" removed.
(II) RADEON(0): Output: DVI-0, Detected Monitor Type: 0
(II) RADEON(0): Output: S-video, Detected Monitor Type: 0
[mi] EQ overflowing. The server is probably stuck in an infinite loop.
[mi] mieqEnequeue: out-of-order valuator event; dropping.

I have an ATI card but this only occurs in Intrepid. Going back to Hardy is the only way to fix his at the moment. I have disabled Compiz for now to see if it stops.

[Tormod Volden (tormodvolden)]

maleadt, have you got any backtrace?

All you others, please read comment 47 and comment 49.

[Tim Besard (maleadt)]

Sadly / Luckily, no. Having installed the Xorg with backtrace
modification, I'm carefully monitoring my system for possibly related
events.

Note though I have also upgraded xf86-video-ati to a newer version,
which I found on the following PPA:
http://ppa.launchpad.net/tormodvolden/ubuntu/ . This might not have been
the smartest thing to obtain a backtrace, as I haven't encountered any
EQ overflow ever since (and thus it <might> have fixed the original
issue).

Op dinsdag 25-11-2008 om 20:30 uur [tijdzone +0000], schreef Tormod
Volden:
> maleadt, have you got any backtrace?
>
> All you others, please read comment 47 and comment 49.
>
> ** Changed in: xorg (Ubuntu)
> Assignee: (unassigned) => Tormod Volden (tormodvolden)
> Status: New => Incomplete
>

[Tormod Volden (tormodvolden)]

If you can confirm that a newer -ati driver helps against hangs, that would also be very useful information. Can you please downgrade to the official version and see if you get hangs again? Then try the new version in intrepid-proposed.

[Antonio Costantino (anto-costantino)]

Hi Tormod, I've seen bug #295904. I think you may be on the right track, as since I disabled desktop effect in KWin I haven't been hit yet. Unfortunately I can't help more, as I don't use other 3d applications so much (but 3d is enabled). As regards the backtrace, I'm in the middle of a new job and I can't mess up with my system at the moment, so I can be only a spectator so far. If I got hit using a 3d app I will post it here and in #295904. I appreciate your effort to sort this issue out.

[Tim Besard (maleadt)]

I have tested the newer -ati driver plus newer X server (from your ppa)
for a full week, and did not experience a single crash. Right now I
cannot revert to the old driver, as I have for the moment switched to
Fedora 10. However, on Fedora I haven't experienced such a crash either,
using Xserver 1.5.3 and xf86-video-ati 6.9.0-54 (which seems to be the
git master from 16/10 with a whole bunch of KMS patches).

I'd want to reinstall Ubuntu for a while on a spare partition, but the
problem is I only triggered the bug after working on it for some days.

Op woensdag 26-11-2008 om 18:12 uur [tijdzone +0000], schreef Tormod
Volden:
> If you can confirm that a newer -ati driver helps against hangs, that
> would also be very useful information. Can you please downgrade to the
> official version and see if you get hangs again? Then try the new
> version in intrepid-proposed.
>

[Szilárd Páll (sin-pecado)]

I have recently reported a similar problem here:
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/275809

Although my system has intel video card the symptoms are quite similar. Is there any possible solution/fix (temporary or definite) for the bug-flavor reported here? Do the recently release Intrepid updates for xorg and xorg-server provide a fix? When will be available an update for Hardy?

I'd be glad to hear some good news because this problem is really frustrating!

[Bryce Harrington (bryce)]

Moving to -ati awaiting backtrace.

@maleadt, let us know if you are able to test the old version of the driver on ubuntu again.

@everyone else - again, the overflow error is a _generic_ error message and doesn't mean you're seeing the same bug as maleadt. Please don't comment on maleadt's bug report since it just distracts from his issue, and your issue won't get reviewed. Report a NEW bug, and reference this bug 276518 if you like, but most of the time these bugs are due to unrelated causes.

[Michael (michaeljt)]

I have seen this lockup twice now on a relatively vanilla intrepid install (but with the kernel 2.6.27.10 from proposed). Nvidia, not ATI:

(--) PCI:*(0@1:0:0) nVidia Corporation Quadro FX 570M rev 161, Mem @ 0xd6000000/16777216, 0xe0000000/268435456, 0xd4000000/33554432, I/O @ 0x00002000/128
...
(II) NVIDIA GLX Module 177.80 Wed Oct 1 15:09:29 PDT 2008

I'm afraid that I didn't get a backtrace, but I have installed the dbgsyms for the X server and will get so if I see it again. I don't want to open a new bug report yet, as the lockup lines in Xorg.0.log could really mean almost anything, and I don't really have enough information to warrant a new report. Other than a backtrace, is there any other useful information I could provide? I have a bit of experience debugging the X server if that helps.

[Michael (michaeljt)]

Forgot to say, I had an external USB mouse plugged in at the time.

[emms (emms007)]

I am having the same issue.

I'm using packages from Tormod's PPA
- Intel driver (v2.5.99.1)
- wserver-xorg-core 1.5.3+git20081216

My system runs Kwin with desktop effects enabled. I have no external mouse nor usb keyboard, it's all built in the laptop.
I happens randomly, sometimes while I'm using the computer or some other times while the computer is idle.
The system does not completely hang when it happens, I can remote connect by SSH, kill Xorg and it's back to normal (until the next lockup)

Here is the backtrace :

Backtrace:
0: /usr/bin/X(xorg_backtrace+0x3b) [0x813164b]
1: /usr/bin/X(mieqEnqueue+0x237) [0x8110ba7]
2: /usr/bin/X(xf86PostKeyboardEvent+0xa7) [0x80d18d7]
3: /usr/lib/xorg/modules/input//evdev_drv.so [0xa74d00e9]
4: /usr/bin/X [0x80cce47]
5: /usr/bin/X [0x80b45dc]
6: [0xb7fb7400]
7: /usr/local/lib/libdrm.so.2(drmWaitVBlank+0x28) [0xb7a73928]
8: /usr/lib/dri/i965_dri.so [0xa74fe05d]
9: /usr/lib/dri/i965_dri.so(driWaitForVBlank+0xfb) [0xa74fe2eb]
10: /usr/lib/dri/i965_dri.so(intelSwapBuffers+0xc7) [0xa7508647]
11: /usr/lib/dri/i965_dri.so [0xa74ff607]
12: /usr/lib/xorg/modules/extensions//libglx.so [0xb7ad8b74]
13: /usr/lib/xorg/modules/extensions//libglx.so [0xb7acb2ce]
14: /usr/lib/xorg/modules/extensions//libglx.so [0xb7acec0a]
15: /usr/bin/X(Dispatch+0x34f) [0x808c8ef]
16: /usr/bin/X(main+0x47d) [0x8071d6d]
17: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7bb0685]
18: /usr/bin/X [0x8071151]
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
ERROR: Server Lockup! Stuck in an infinite loop. See backtrace above.
[...]
(EE) ACPI Virtual Keyboard Device: Read error: No such device

[Gareth Bult (gareth-encryptec)]

Hi, just in case anyone thought this has been remotely nailed down / fixed;

I've been running find for nearly two months after reporting this problem.
My documented fix was to disable xinerama.

I thought I'd try today to see if it was any better, so I ticked xinerama in my nvidia config, logged out then logged in again ... lasted less than an hour before the mouse locked up.

Keyboard was still working, but the mouse would do nothing.
Keyboar driven logout cured it.

Couldn't see any traces of anything in the log files except a few;
"Xorg.0.log.old:(II) USB Keyboard + 2P Hub: Device reopened after 10 attempts."

It's almost as if Xinerama is fouling the mouse in some way ...
*nothing* else changed, so it does seem to hinge on xinerama .. (!)
(I've had and still have Compiz turned off)

[Tormod Volden (tormodvolden)]

Guys, this is maleadt's bug report. Don't post here. Please read comment #68 again. And comment #47 and comment #49.

[Bryce Harrington (bryce)]

[Still awaiting response from original reporter]

[Tim Besard (maleadt)]

Op donderdag 08-01-2009 om 10:33 uur [tijdzone +0000], schreef Bryce
Harrington:
> [Still awaiting response from original reporter]

I am using Ubuntu again for two weeks now, and have only once triggered
the bug, which however did not lock up my machine but kept it running
nicely.

I have however changed my hardware layout a bit (PS2 mouse to USB
mouse), which I'll try to replace for a week in order to have exactly
the same situation as I had two months ago.

[Tim Besard (maleadt)]

Got it twice today, using the -ati driver from Tormod's PPA. No
backtrace mentioned though. Log included.

[Tormod Volden (tormodvolden)]

maleadt, it seems that you are using the xserver-xorg-core 1.5.2, without the backtrace on detected loop.

[Bryce Harrington (bryce)]

maleadt,
Would you mind testing to see if this bug is still an issue in Jaunty? CD ISOs are available from http://cdimages.ubuntu.com/releases/jaunty/ . If it is still an issue, then perhaps we can forward it upstream.

[Tim Besard (maleadt)]

It took me some weeks, but I have now triggered this lockup with bleeding-edge xorg (xorg-edgers's ppa) and xserver-xorg-video-ati (tormod volden's ppa). Also discard the comment from 2009-01-02, as that partial lockup (xorg complaining about stuck server, but still remaining responsive) does not seem to be related. The lockup I am talking about right now, is the same as described in the bug report and most of my comments; a lockup in which case only the mouse moves, but the xserver doesn't respond on any action at all.

Regarding the various comments about mouse/keybord possibly causing the problem when being connected through usb: this does not seem to be related to this lockup, as both my mouse and keyboard now were connected over PS/2 (through an USB->PS/2 connector, but this shouldn't matter).

Apart from confirming this lockup with updated drivers, I can also confirm it with a fully updated Intrepid Ibex. I triggered the bug fairly quickly after having Ubuntu reinstalled and updated.

I don't have the time to test Jaunty right now, but I might test it later on.

Xorg log included, backtrace included with xorg & xserver-xorg-video-ati from ppa's.

[Martin Nowack (martin-nowack)]

I've got the same stack trace.

@maleadt, can you try "System Settings"->"Desktop"-> and disable! "Enable desktop effects" and then logout and work again.
Since switching the effects off, my system never locked again. Otherwise it would lock at least one time a day.

Therefore, for me it seams DRM related.
Hope that helps you too.

[Antonio Costantino (anto-costantino)]

I don't have the time neither the ability to produce an helpful backtrace, however I can confirm that disabling the "Desktop Effects (compositing right?)" solves the issue for me. I was working without it for quite a lot now, but after kde 4.2 went to the backports I wished to try it with the effects... Result: (almost) immediate lock. I disabled it again and everything works good again. I certainly can't speak technically, but it has something to do with effects for sure.

This bug, at least for ati users with open source driver, wipes out most of the benefits of the new kde... Really hope is solved for Jaunty... Don't give up ;-)

[Bryce Harrington (bryce)]

[Still need this tested against jaunty before we can proceed]

[Kai Krakow (hurikhan77)]

I just posted on bgo, maybe related:
http://bugs.gentoo.org/show_bug.cgi?id=263057

To shorten this: Although I'm on gentoo, the bug is similar/the same. And I discovered that the probability of this bug may be related to how much work is put on the gpu while using the input device. So it still happens with desktop effects disabled but much more seldom. I have two identical systems (software versions, not hardware - hardware is different, see link). One shows these problems, the other not (both nvidia).

[lhotari (lartsa)]

This helped for me: http://www.nvnews.net/vbulletin/showthread.php?t=118088

my xorg.conf options (not sure if these should be here):
 Option "PixmapCacheSize" "1000000"
 Option "AllowSHMPixmaps" "0"
 Option "PixmapCache" "1"
 Option "PixmapCacheRoundSizeKB" "1024"
 Option "InitialPixmapPlacement" "2"
 Option "GlyphCache" "1"

from command line:
nvidia-settings -a InitialPixmapPlacement=2 -a GlyphCache=1

Also fixed the x-server restart problem (https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-177/+bug/258357).

[xenophed (xenophed)]

This is related to the opengl screensavers so I would think that it is a problem to gl mode operations

[xenophed (xenophed)]

it is not dependent on graphics card but by using a non gl screensaver it stops
system config core dual intel 3.7ghz 750mb 333 ram 120gb hd 945 graphics chipset

[Bryce Harrington (bryce)]

Lockups are always GPU specific things, but the symptoms of a freeze are almost identical so it's hard to tell them apart. The error message is also a generic one, it just means that something noticed the GPU had locked up. So just because you got that message doesn't mean it's the same bug.

I've written a troubleshooting guide for these kinds of problems, here: https://wiki.ubuntu.com/X/Troubleshooting/Freeze

Since it appears a number of people have 'me-too-ed' onto this bug report that don't have the same hardware and thus almost certainly don't have the same actual bug, I'm going to close this bug, and ask that each person report a NEW bug with their problem, against their video driver. See the above link for additional troubleshooting tips. The most important thing is to provide detailed steps to reproduce your freeze.

In Karmic we hope to be providing some tools that will make it easier to report and troubleshoot lockups. Meanwhile, if you include the word "freeze" in the title of your bug report, we'll be able to find and follow up on it with you when the tools become available.

[Bryce Harrington (bryce)]

[Closing as per my last comment]

[Cinquero (cinquero)]

Quite similar bug here:

[mi] EQ overflowing. The server is probably stuck in an infinite loop.

Backtrace:
0: /usr/bin/X (xorg_backtrace+0x28) [0x4a3238]
1: /usr/bin/X (mieqEnqueue+0x1f4) [0x4a2ab4]
2: /usr/bin/X (xf86PostMotionEventP+0xc4) [0x47ce94]
3: /usr/lib/xorg/modules/input/evdev_drv.so (0x7fe6668d1000+0x53cf) [0x7fe6668d63cf]
4: /usr/bin/X (0x400000+0x6fca7) [0x46fca7]
5: /usr/bin/X (0x400000+0x11d1b3) [0x51d1b3]
6: /lib/libpthread.so.0 (0x7fe66adf1000+0xf8f0) [0x7fe66ae008f0]
7: /usr/bin/X (0x400000+0x65700) [0x465700]
8: /lib/libpthread.so.0 (0x7fe66adf1000+0xf8f0) [0x7fe66ae008f0]
9: /usr/lib/libpixman-1.so.0 (0x7fe66a756000+0x42f90) [0x7fe66a798f90]
10: /usr/lib/libpixman-1.so.0 (0x7fe66a756000+0x431c8) [0x7fe66a7991c8]
11: /usr/lib/libpixman-1.so.0 (pixman_blt+0x78) [0x7fe66a785818]
12: /usr/lib/xorg/modules/libfb.so (fbCopyNtoN+0x26a) [0x7fe667280b6a]
13: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7fe667cac000+0x61168) [0x7fe667d0d168]
14: /usr/lib/xorg/modules/drivers/intel_drv.so (0x7fe667cac000+0x6539d) [0x7fe667d1139d]
15: /usr/bin/X (0x400000+0xd8110) [0x4d8110]
16: /usr/bin/X (0x400000+0xd182e) [0x4d182e]
17: /usr/bin/X (0x400000+0x30c3c) [0x430c3c]
18: /usr/bin/X (0x400000+0x261aa) [0x4261aa]
19: /lib/libc.so.6 (__libc_start_main+0xfd) [0x7fe669ae9c4d]
20: /usr/bin/X (0x400000+0x25d59) [0x425d59]

Steps to reproduce:

Latest (with all the latest updates as of right now) Ubuntu Lucid (Stable, 10.04).

1. start Firefox 3.6.3
2. open a second window (not tab!)
3. go to http://www.ardmediathek.de/ard/servlet/
4. Xorg hangs

in most cases, the X screen is then locked with the backtrace shown above.

Seems to be a Flash plugin concurrency issue in Firefox.

    npwrapper.libflashplayer.so
    Version: Shockwave Flash 10.1 r53