8.543 Crashes in CompositePicture

[FrejSoya]

Xorg.log attached, and xorg.conf attached (just switch radeon with fglrx).

Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x79) [0x80c2fe9]
1: [0xb80ca400]
2: /usr/lib/xorg/modules//glesx.so [0xb57b2936]
3: /usr/lib/xorg/modules//libxaa.so(XAAComposite+0x176) [0xb75be0c6]
4: /usr/lib/xorg/modules//libxaa.so [0xb75db4c1]
5: /usr/X11R6/bin/X [0x817718a]
6: /usr/X11R6/bin/X(CompositePicture+0x19a) [0x815fdea]
7: /usr/X11R6/bin/X [0x8165cb5]
8: /usr/X11R6/bin/X [0x81629d5]
9: /usr/X11R6/bin/X(Dispatch+0x34f) [0x808c87f]
10: /usr/X11R6/bin/X(main+0x47d) [0x8071d1d]
11: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7cbf685]
12: /usr/X11R6/bin/X [0x8071101]

[lspci]
00:00.0 Host bridge: Intel Corporation Mobile 945GM/PM/GMS, 943/940GML and 945GT Express Memory Controller Hub (rev 03)
     Subsystem: Intel Corporation Device 7270
01:00.0 VGA compatible controller: ATI Technologies Inc M56P [Radeon Mobility X1600]
     Subsystem: Apple Computer Inc. Device 0080

[FrejSoya (frej)]

Also, macbook pro (core duo, ati X1500)

[Korcsi (bukon2)]

I have same problem, Ubuntu 8.10 (upgraded from 8.04, fglrx 8.54.3)

Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x79) [0x80c2fe9]
1: [0xb7fb2400]
2: /usr/lib/xorg/modules//glesx.so [0xb56aa936]
3: /usr/lib/xorg/modules//libxaa.so(XAAComposite+0x176) [0xb74ac0c6]
4: /usr/lib/xorg/modules//libxaa.so [0xb74c94c1]
5: /usr/X11R6/bin/X [0x817718a]
6: /usr/X11R6/bin/X(CompositePicture+0x19a) [0x815fdea]
7: /usr/X11R6/bin/X [0x8165cb5]
8: /usr/X11R6/bin/X [0x81629d5]
9: /usr/X11R6/bin/X(Dispatch+0x34f) [0x808c87f]
10: /usr/X11R6/bin/X(main+0x47d) [0x8071d1d]
11: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7bb4685]
12: /usr/X11R6/bin/X [0x8071101]
Saw signal 11. Server aborting.
(II) Macintosh mouse button emulation: Close
(II) UnloadModule: "evdev"
(II) AT Translated Set 2 keyboard: Close
(II) UnloadModule: "evdev"
(II) Logitech USB-PS/2 Optical Mouse: Close
(II) UnloadModule: "evdev"
(II) AIGLX: Suspending AIGLX clients for VT switch

[Korcsi (bukon2)]

My xorg.conf

[Korcsi (bukon2)]

My xorg.conf

[Bryce Harrington (bryce)]

Hi frej,

Thank you for taking the time to report this bug and helping to make Ubuntu better. You reported this bug a while ago and there hasn't been any activity in it recently. We were wondering is this still an issue for you? Can you try with the latest development release of Ubuntu? (ISOs are available from cdimage.ubuntu.com)

If it remains an issue, could you also attach a new /var/log/Xorg.0.log?
Thanks in advance.

The output of lspci -vvnn would also be worth having.

[Bryce Harrington (bryce)]

We're closing this bug since it is has been some time with no response from the original reporter. However, if the issue still exists please feel free to reopen with the requested information. Also, if you could, please test against the latest development version of Ubuntu, since this confirms the bug is one we may be able to pass upstream for help.

[Marcus Carlson (0-launchpad-mejlamej-nu)]

I just upgraded to Intrepid and got this problem. Attached is lspci -vvnn.
Is there anything else I could try?

Part of Xorg.0.log;
Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x65) [0x480f35]
1: /lib/libc.so.6 [0x7ff4622f5060]
2: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxDisplayShadowIsShadowPixmap+0x22) [0x7ff460ee2db2]
3: /usr/lib/xorg/modules//glesx.so [0x7ff45e1374f3]
4: /usr/lib/xorg/modules//libxaa.so(XAAComposite+0x1a5) [0x7ff4600f10d5]
5: /usr/lib/xorg/modules//libxaa.so [0x7ff46010eb0d]
6: /usr/X11R6/bin/X [0x531568]
7: /usr/X11R6/bin/X [0x520a6a]
8: /usr/X11R6/bin/X(Dispatch+0x364) [0x44d6e4]
9: /usr/X11R6/bin/X(main+0x45d) [0x4336fd]
10: /lib/libc.so.6(__libc_start_main+0xe6) [0x7ff4622e0466]
11: /usr/X11R6/bin/X [0x432ad9]
Saw signal 11. Server aborting.

Some info about system; AMD Phenom(tm) 9850 Quad-Core Processor, 8 GB RAM, Ubuntu 8.10 64bit

[Marcus Carlson (0-launchpad-mejlamej-nu)]

Also tried with radeonhd driver with similar problem

From Xorg.0.log;
Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x65) [0x480f35]
1: /lib/libc.so.6 [0x7f9c57e34060]
2: /lib/libc.so.6(memcpy+0x15b) [0x7f9c57e8612b]
3: /usr/lib/xorg/modules/drivers//radeonhd_drv.so(RHDVGASave+0xff) [0x7f9c56c04b0f]
4: /usr/lib/xorg/modules/drivers//radeonhd_drv.so [0x7f9c56bf61f2]
5: /usr/lib/xorg/modules/drivers//radeonhd_drv.so [0x7f9c56bf6dd8]
6: /usr/X11R6/bin/X(AddScreen+0x1c9) [0x432d89]
7: /usr/X11R6/bin/X(InitOutput+0x241) [0x469a61]
8: /usr/X11R6/bin/X(main+0x286) [0x433526]
9: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f9c57e1f466]
10: /usr/X11R6/bin/X [0x432ad9]
Saw signal 11. Server aborting.

Could the first line indicate that this is a problem in X rather than in the driver?

But than again - ati driver is working fine (but I miss compiz and awn ;-) )

Hardy and fglrx worked.

Hope it helps...

[Brian J. Cohen (brianjcohen)]

X is crashing for me too, I have a similar backtrace. Upgraded from 8.04 to 8.10, fglrx 8.54.3, using Radeon 3650 card.

(II) Loading sub module "fglrxdrm"
(II) LoadModule: "fglrxdrm"

(II) Loading /usr/lib/xorg/modules/linux//libfglrxdrm.so
(II) Module fglrxdrm: vendor="FireGL - ATI Technologies Inc."
 compiled for 1.4.99.906, module version = 8.54.3
 ABI class: X.Org Server Extension, version 1.1
(II) fglrx(0): Using adapter: 1:0.0.

Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x79) [0x80c3009]
1: [0xb7f6b400]
2: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxPcsLoadKernelDatabase+0xbc) [0xb770b4cc]
3: /usr/lib/xorg/modules/drivers//fglrx_drv.so [0xb76e9ffa]
4: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxPreInit+0x764) [0xb76e2d44]
5: /usr/X11R6/bin/X(InitOutput+0x96f) [0x80aac9f]
6: /usr/X11R6/bin/X(main+0x279) [0x8071b19]
7: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7b6e685]
8: /usr/X11R6/bin/X [0x8071101]
Saw signal 11. Server aborting.

[ruminant1 (ruminant1)]

I got the same backtrace as Marcus on a 780G-based board (Radeon 3200HD). I tried disabling composite in xorg.conf but hit the crash anyway.

[cro (cro)]

Error also appears on AMD x86_64, 2.6.27-7-generic SMP, with Radeon 4950HD.

Going to try some source compiles today.

[snip from Xorg.0.log]
(II) Loading /usr/lib/xorg/modules//amdxmm.so
(II) Module amdxmm: vendor="X.Org Foundation"
 compiled for 1.4.99.906, module version = 1.0.0
 ABI class: X.Org Server Extension, version 1.1

Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x65) [0x480f35]
1: /lib/libc.so.6 [0x7fd29bf8c060]
2: /usr/X11R6/bin/X(xf86XvMCScreenInit+0xed) [0x49628d]
3: /usr/lib/xorg/modules//amdxmm.so(UvdFwSetup+0xa24) [0x7fd2975e4ef4]
4: /usr/lib/xorg/modules//amdxmm.so(amdxmmInit+0x130) [0x7fd2975dcb00]
5: /usr/lib/xorg/modules/drivers//fglrx_drv.so [0x7fd29ab4003d]
6: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxScreenInit+0x6b9) [0x7fd29ab386f9]
7: /usr/X11R6/bin/X(AddScreen+0x1c9) [0x432d89]
8: /usr/X11R6/bin/X(InitOutput+0x241) [0x469a61]
9: /usr/X11R6/bin/X(main+0x286) [0x433526]
10: /lib/libc.so.6(__libc_start_main+0xe6) [0x7fd29bf77466]
11: /usr/X11R6/bin/X [0x432ad9]
Saw signal 11. Server aborting.

[Edpsun (edpsun)]

I have the same issue.
I did a clean installation with 8.10. While configuring the dual head with Radeon Mobility X600, I saw the similar stack trace.
----
(II) Loading /usr/lib/xorg/modules//amdxmm.so
(II) Module amdxmm: vendor="X.Org Foundation"
 compiled for 1.4.99.906, module version = 1.0.0
 ABI class: X.Org Server Extension, version 1.1

Backtrace:
0: /usr/X11R6/bin/X(xf86SigHandler+0x79) [0x80c3009]
1: [0xb7f99400]
2: /usr/lib/xorg/modules//amdxmm.so(amdxmmInit+0x42c) [0xb63b6b2c]
3: /usr/lib/xorg/modules/drivers//fglrx_drv.so [0xb7717cac]
4: /usr/lib/xorg/modules/drivers//fglrx_drv.so(atiddxScreenInit+0x799) [0xb770e6c9]
5: /usr/X11R6/bin/X(AddScreen+0x19f) [0x807137f]
6: /usr/X11R6/bin/X(InitOutput+0x206) [0x80aa536]
7: /usr/X11R6/bin/X(main+0x279) [0x8071b19]
8: /lib/tls/i686/cmov/libc.so.6(__libc_start_main+0xe5) [0xb7b9a685]
9: /usr/X11R6/bin/X [0x8071101]
Saw signal 11. Server aborting.

[cro (cro)]

Here's an update:

I managed to resolve this issue (and will cross-post to the forums as well) by referencing this against another bug in another package/forum: http://www.virtualbox.org/ticket/2288

I found that essentially the same X error was being caused by the vboxdrv under Ubuntu 8.10/2.6.27-7 kernel, which seemed odd - since I had Virtual box installed, and dkms reported that the vboxdrv module was assigned against the earlier 2.6.24 kernel.

Disabling vboxdrv ('/etc/init.d/vboxdrv stop' followed by removing the execute flag from the file itself) caused dkms to fail to load the module at boot (which I had been seeing before), but 'dkms status' now says this:

~$ dkms status
vboxdrv, 1.6.2: added
fglrx, 8.543, 2.6.27-7-generic, x86_64: installed

and the fglrx driver works fine.

So it may be (or at least appears to be) an issue with modules loaded through dkms for older kernels not being disabled properly.

[alejandro (anicolas)]

Thanks man! :)

I confirm I had exactly the same issue and got fixed with cro workaround. fglrx driver driver is now working with Ubuntu 8.10 in dual head fashion.

fglrxinfo outputs:

display: :0.0 screen: 0
OpenGL vendor string: ATI Technologies Inc.
OpenGL renderer string: ATI Radeon HD 2400 XT
OpenGL version string: 2.1.8087 Release

The only difference after removing execution rights to vboxdrv is the output of dkms status, which for vboxdrv is "installed" instead of "added":

vboxdrv, 1.6.2, 2.6.24-19-generic, i686: installed
fglrx, 8-4, 2.6.24-19-generic, i686: built
fglrx, 8.543, 2.6.27-7-generic, i686: installed

[alejandro (anicolas)]

BTW, installing VBox kernel driver will definitely solve the issue :)

$sudo /etc/init.d/vboxdrv setup

$dmks status

vboxdrv, 1.6.2, 2.6.24-19-generic, i686: installed
vboxdrv, 2.0.4, 2.6.27-7-generic, i686: installed
fglrx, 8-4, 2.6.24-19-generic, i686: built
fglrx, 8.543, 2.6.27-7-generic, i686: installed

[Marcus Carlson (0-launchpad-mejlamej-nu)]

Maybe I'm missing something but I don't have VirtualBox installed.

dkms status gives me this
fglrx, 8.543, 2.6.24-21-generic, x86_64: installed (original_module exists)
fglrx, 8.543, 2.6.27-7-generic, x86_64: installed

which seems right - right?

[cro (cro)]

Marcus: Try removing "fglrx, 8.543, 2.6.24-21-generic, x86_64: installed (original_module exists)" if you get the 'Saw signal 11' error in your Xorg log. It appears (perhaps) to be the conflict between the module built for the old kernel that causes the module for the new kernel to fail.

Alejandro: As for reinstalling the vboxdrv driver - setup (/etc/init.d/vboxdrv setup) failed for me with a compile error that's also been seen in some versions of fglrx (although I don't have the references to hand, something about too many parameters on one file), which is what also led me to removing the modules compiled for previous kernels from dkms.

It seems, from the previous couple of entries, that it is the 'installed' status of modules for previous kernels that causes the problem. Maarcus' fails because his 2.6.24 is 'installed', but alejandro's works because it's just 'built' not 'installed'. Of course this also means that you can't revert to the old 2.6.24 kernel without backing out of these module changes, and definitely points to a conflict in the way different kernel versions of modules are handled by 2.6.27.

[Marcus Carlson (0-launchpad-mejlamej-nu)]

cro, I did try to remove the old module but with no success. Also tried to remove the current module and reinstall the fglrx-kernel-source package, but same problem.
dkms status now returns
fglrx, 8.543, 2.6.27-7-generic, x86_64: installed

Something else I could try?

[DoDoENT (dodoentertainment)]

Maybe you have this bug: https://bugs.launchpad.net/ubuntu/+source/linux-restricted-modules-2.6.24/+bug/291210

[Marcus Carlson (0-launchpad-mejlamej-nu)]

I think not, because the fglrx module seems loaded. Snip from dmesg;
dmesg | grep fglrx
[ 14.328262] fglrx: module license 'Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY' taints kernel.
[ 14.342900] [fglrx] Maximum main memory to use for locked dma buffers: 7758 MBytes.
[ 14.343291] [fglrx] vendor: 1002 device: 9501 count: 1
[ 14.343950] [fglrx] ioport: bar 4, base 0xbe00, size: 0x100
[ 14.345316] [fglrx] PAT is enabled successfully!
[ 14.358574] [fglrx] module loaded - fglrx 8.54.3 [Oct 10 2008] with 1 minors

And if the module was compiled against the wrong kernel it wouldn't get loaded...

I do however have some fglrx related files in lib - should I remove l-r-m for the 2.6.24 kernel? Even though I can't believe these would mess it up.

find /lib | grep fglrx
/lib/modules/2.6.24-21-generic/kernel/drivers/char/drm/fglrx.ko
/lib/modules/2.6.27-7-generic/updates/dkms/fglrx.ko
/lib/linux-restricted-modules/2.6.24-21-generic/fglrx
/lib/linux-restricted-modules/2.6.24-21-generic/fglrx/fglrx.mod.o
/lib/linux-restricted-modules/2.6.24-21-generic/fglrx/libfglrx_ip.a.GCC4
/lib/linux-restricted-modules/2.6.24-21-generic/fglrx/firegl_public.o

[FrejSoya (frej)]

I see my bug got hijacked

1) Bryce (or Automatic Bot disgusing as bryce), i'm concerned with the automatic aggressive pruning of bugs on launchpad. This isn't the first time i've seen it. Really somebody else had the same trace and know i need to bothered more. It's already too complicated to use apport. Using the browser is crap especially when it shows launchpad. I can only complain about launchpad since I can't fix it myself).

Please explain why a new xorg log would change anything? I havent seen any important kernel or fglrx updates, that other than pure luck whould fix the bug.

I've attached an lspci dump.

2) Everyone else but korsic posted different stack traces, it's not the same bug or doesn't appear to be. Suddenly the cause was the vbox driver which i've never used or heard of. Unless you are 100% sure open a new bug, it's easier to duplicate a bug than to split one bug into two.

Bad launchpad experience..... I guess my headline wasn't specific enough.

[FrejSoya (frej)]

Ehh, just ignore the obfuscated sentence ". Really somebody else had the same trace and know i need to bothered more". Sorry!

[cro (cro)]

My apologies - I didn't suggest that the problem was in the vboxdrv driver - I suggested that it was caused by a module compiled and installed for the previous kernel version.

It just happened that I had only one module installed for a previous kernel version (the vboxdrv module - after I removed all the old fglrx modules) and my contention was that it was this conflict between modules compiled for different kernel versions that was causing the problem. Whether it was the vboxdrv module, or some other module I don't know. What's your dkms status output? Do you have any dkms-handled driver modules compiled for an earlier version of the kernel? Could they be conflicting with the new version of X?

[Bryce Harrington (bryce)]

A script-misfire closed this bug too quickly. Reopening.

[Sven Andersen (x-launchpad-cryonet-de)]

Option "TexturedXRender" "False"

solved the problem for me on intrepid x86_64.

[supernova_hq (supernova-hq)]

For anyone who had problems running the code in alejandro's post:
The second command should be
$dkms status
not
$dmks status

[Bryce Harrington (bryce)]

> 2) Everyone else but korsic posted different stack traces, it's not the same bug or
> doesn't appear to be. Suddenly the cause was the vbox driver which i've never
> used or heard of. Unless you are 100% sure open a new bug, it's easier to duplicate
> a bug than to split one bug into two.

Looking at the stack traces I agree. The title of this bug report was so generic that a lot of people have assumed they have the same problem. I've retitled.

@Everyone other than korsic, file a new bug please. You do *not* have a duplicate of this bug. We'll only be focusing on FrejSoya's issue here.

> Bryce (or Automatic Bot disgusing as bryce), i'm concerned with the automatic aggressive pruning of bugs
> on launchpad. This isn't the first time i've seen it.

A necessary evil I'm afraid. The unfortunate fact is that with Xorg we've got around 2000 bugs open, plus around 100 new bugs per week presently. The Ubuntu-X team comprises half a dozen people - too few to stay on top of it. Mechanical pruning helps the largely volunteer team spend less time with bug triage drudgery and more on actual bug resolution.

[Bryce Harrington (bryce)]

[Reported to AMD]

[Bryce Harrington (bryce)]

AMD's response:

# More information needed

    * Steps to reproduce
    * Is issue consistently reproduced?
    * Requesting original bug-logger to indicate his concern
    * Several people with several different comments which are not necessarily relevant

Frej, are you still seeing the crash consistently with Intrepid? If so, please provide steps to reproduce it.

[jawbone (j-a-w-bone)]

Hi all,

i had the same problem on Thinkpad T60p, Centrino Duo (2x Genuine Intel(R) CPU T2600 @ 2.16GHz) and
ATI Technologies Inc M56GL [Mobility FireGL V5200] rev 0.

Renaming the glesx.so such that loading it wasn't possible anymore fixed the issue, although this is not a very
satisfactory way to do so. Strangely enough, everything seems to work fine, Video as well as GL acceleration.
Can anybody maybe shed some light on what the glesx module is good for?

As for the AMD people: I can reliably reproduce this error by giving the module its original name. The
Option "TexturedXRender" "False"
had no effect on my machine, and my xorg.conf is absolutely minimal, the only special thing is that it requests
the "glx" module.

That's what my dkms status tells me:
fglrx, 8.552, 2.6.27-9-generic, i686: installed

[Chris (a08)]

The driver uses some kind of settings database. I changed my config back and forth but it would continue crashing but used to work before...

...so I uninstalled everything fglrx related, went to the /etc folder and killed the ati folder with everything in it. After the following install of fglrx and the xorg-driver-fglrx everything startet working again.

My working xorg.conf is attached (the PC is hooked up to a beamer, hence the strange resolution) - hope this helps!

[Bryce Harrington (bryce)]

We're closing this bug since it is has been some time with no response from the original reporter. However, if the issue still exists please feel free to reopen with the requested information. Also, if you could, please test against the latest development version of Ubuntu, since this confirms the bug is one we may be able to pass upstream for help.