Bug 688693 - X server fatal error - "Wrong event type 0"
Summary: X server fatal error - "Wrong event type 0"
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: mesa
Version: 15
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
Assignee: Adam Jackson
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
: 682593 691135 (view as bug list)
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2011-03-17 18:13 UTC by Aaron Sowry
Modified: 2018-04-11 13:51 UTC (History)
8 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2011-06-02 02:49:12 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
Xorg.0.log.old immediately after the crash (34.32 KB, text/plain)
2011-03-18 10:07 UTC, Aaron Sowry
no flags Details
dmesg immediately after the crash (123.86 KB, text/plain)
2011-03-18 10:26 UTC, Aaron Sowry
no flags Details
/var/log/messages from today (2 crashes) (538.53 KB, text/plain)
2011-03-18 10:33 UTC, Aaron Sowry
no flags Details
dmesg (123.21 KB, text/plain)
2011-04-13 19:27 UTC, Tomas Toth
no flags Details
/var/log/messages (123.72 KB, text/plain)
2011-04-13 19:28 UTC, Tomas Toth
no flags Details
Xorg.0.log.old (31.59 KB, text/plain)
2011-04-13 19:30 UTC, Tomas Toth
no flags Details
Xorg.0.log.old with Backtrace (33.38 KB, text/plain)
2011-05-06 18:23 UTC, Tomas Toth
no flags Details
Xorg.0.log.old with Backtrace (32.91 KB, text/plain)
2011-05-09 20:19 UTC, Tomas Toth
no flags Details

Description Aaron Sowry 2011-03-17 18:13:35 UTC
Description of problem:
X server restarts itself randomly

Version-Release number of selected component (if applicable):
xorg-x11-server-Xorg-1.10.0-3.fc15.x86_64

How reproducible:
Seems to happen every couple of hours

Steps to Reproduce:
Cannot reproduce reliably yet
  
Actual results:
X server experiences fatal error and shuts down. From Xorg.0.log:

...
[ 29300.659] 
Fatal server error:
[ 29300.661] Wrong event type 0.
[ 29300.661] 
[ 29300.661] 
Please consult the Fedora Project support 
         at http://wiki.x.org
 for help. 
[ 29300.661] Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[ 29300.661] 
[ 29300.673] (II) Power Button: Close
[ 29300.673] (II) UnloadModule: "evdev"
[ 29300.673] (II) Unloading evdev
[ 29300.676] (II) Video Bus: Close
[ 29300.676] (II) UnloadModule: "evdev"
[ 29300.676] (II) Unloading evdev
[ 29300.682] (II) Sleep Button: Close
[ 29300.682] (II) UnloadModule: "evdev"
[ 29300.682] (II) Unloading evdev
[ 29300.694] (II) Integrated Camera: Close
[ 29300.694] (II) UnloadModule: "evdev"
[ 29300.694] (II) Unloading evdev
[ 29300.697] (II) AT Translated Set 2 keyboard: Close
[ 29300.697] (II) UnloadModule: "evdev"
[ 29300.697] (II) Unloading evdev
[ 29300.714] (II) UnloadModule: "synaptics"

Expected results:
X doesn't restart randomly.

Additional info:
This started happening after adding the kernel parameter "nouveau.noaccel=1" to work around bug #684608.

Comment 1 Matěj Cepl 2011-03-17 22:45:57 UTC
Thanks for the bug report.  We have reviewed the information you have provided above, and there is some additional information we require that will be helpful in our diagnosis of this issue.

Please add drm.debug=0x04 to the kernel command line, restart computer, and attach

* your X server config file (/etc/X11/xorg.conf, if available),
* X server log file (/var/log/Xorg.*.log)
* output of the dmesg command, and
* system log (/var/log/messages)

to the bug report as individual uncompressed file attachments using the bugzilla file attachment link above.

We will review this issue again once you've had a chance to attach this information.

Thanks in advance.

Comment 2 Aaron Sowry 2011-03-18 10:07:47 UTC
Created attachment 486197 [details]
Xorg.0.log.old immediately after the crash

Comment 3 Aaron Sowry 2011-03-18 10:26:18 UTC
Created attachment 486201 [details]
dmesg immediately after the crash

It looks like the dmesg buffer has been filled 10 times over by this point, so let me know if there are any earlier messages you need to see.

Comment 4 Aaron Sowry 2011-03-18 10:33:20 UTC
Created attachment 486202 [details]
/var/log/messages from today (2 crashes)

Comment 5 Aaron Sowry 2011-03-25 09:24:09 UTC
Has anyone been able to reproduce this? I'm surprised that there haven't been any me-toos yet.

Assuming it's not just me, I think this bug should be upgraded to urgent. It essentially renders F15 unusable on a decent swag of modern laptops, e.g. Latitudes, Thinkpads, EliteBooks and Macbooks.

- nouveau is unusable due to bug 684608, which may not be fixed for F15
- There are no official nVidia drivers for this chipset which work with such a recent version of X (-ignoreABI results in a segfault)
- This bug renders the fallback desktop useless (again, assuming it's not just me for some reason)

Comment 6 Ben Skeggs 2011-03-27 22:17:47 UTC
*** Bug 691135 has been marked as a duplicate of this bug. ***

Comment 7 Timon 2011-04-06 10:34:24 UTC
I catched this bug once with openchrome driver.

Comment 8 Ben Skeggs 2011-04-08 04:27:44 UTC
Yeah, I'm going to reassign this to the xserver for now.  I'm going to assume it's not a driver issue until someone tells me otherwise!

Comment 9 Tomas Toth 2011-04-13 19:27:13 UTC
Created attachment 491877 [details]
dmesg

Comment 10 Tomas Toth 2011-04-13 19:28:01 UTC
Created attachment 491878 [details]
/var/log/messages

Comment 11 Tomas Toth 2011-04-13 19:30:57 UTC
Created attachment 491879 [details]
Xorg.0.log.old

Comment 12 Tomas Toth 2011-04-13 20:08:45 UTC
Description of problem:
I hit the same or similar issue described at #comment 0, with the exception that
I do not have the "nouveau.noaccel=1" kernel parameter.


Version-Release number of selected component (if applicable):
xorg-x11-server-Xorg.x86_64    1.10.0-7.fc15


How reproducible:
Once is 10-15 minutes.


Steps to Reproduce:
No exact steps. Usually happens when browsing web with firefox. I tend to say
that it happens when I do something with the mouse, but not sure.

Created attachments, the crash is at ~725sec:
dmesg              #comment 9
/var/log/messages  #comment 10
Xorg.0.log.old     #comment 11


Additional info:
I also see the following in the /var/log/messages, do not know whether it is related:
kernel: [ 2.787518] [drm] nouveau 0000:01:00.0: 512 MiB GART (aperture)
kernel: [ 2.787532] [drm] nouveau 0000:01:00.0: PGRAPH: unsupported chipset, please report!
kernel: [ 2.788181] [drm] nouveau 0000:01:00.0: PGRAPH: unknown config: 2/0/0/0, 1
kernel: [ 2.791419] [drm] nouveau 0000:01:00.0: failed to load fuc409d

Smolt HW Profile:
http://www.smolts.org/client/show/pub_6d8c5b3a-7897-443c-bf74-3bb979abdd07

Comment 13 Adam Jackson 2011-04-18 19:04:08 UTC
"wrong event type 0" is generated by the input handling code, there's not going to be any dependence on video hardware here.

Comment 14 Tomas Toth 2011-05-03 21:37:07 UTC
During the last two evenings it crashed 18 times, during the last 2.5 hours 7 times, it is very annoying. Sometimes it crashes within 3 minutes since the previous crash.
Could you please estimate the time frame of fixing this bug? Could it get higher priority or be classified as F15Blocker, F15-accepted?

I am not pushy, just need to decide whether to revert to Fedora 14 so I can use my laptop.

Please let me know if/how I can help.

Thanks.

Comment 15 Peter Hutterer 2011-05-04 04:45:50 UTC
can you reproduce this reliably? it's a bug in the server, but I don't know what triggers it.

Comment 16 Tomas Toth 2011-05-04 20:17:16 UTC
I cannot give exact steps but it happens frequently enough to be able to generate additional reports. Let me know what you need.
I am also quite confident that it is something with the mouse / pointing device. It is HW independent, occurs with two different mouses and once it even happened while I was working with the touch-pad (no mouse attached).
I tend to say that only moving the cursor is enough, no need for clicking, I am not 100% sure.

P.S. It crashed twice while I was writing this.

Comment 17 Peter Hutterer 2011-05-06 00:51:01 UTC
There's a scratch build available at the url below. It includes an extra patch for debugging purposes which should hopefully help finding out how we get to this point. Please try it out and attach the Xorg.log (which should have a backtrace) once it crashes.

http://koji.fedoraproject.org/koji/taskinfo?taskID=3053292

Comment 18 Tomas Toth 2011-05-06 18:23:17 UTC
Created attachment 497424 [details]
Xorg.0.log.old with Backtrace

Attached please find the Xorg.0.log with Backtrace.

Comment 19 Tomas Toth 2011-05-07 18:36:41 UTC
Another two backtraces from Xorg.0.log(s):

[  4959.186] dix: invalid event type 0
[  4959.189] ff 00 00 00 06 00 00 00 
[  4959.189] 98 01 00 00 d4 ab 4b 00 
[  4959.189] 0a 00 00 00 0a 00 00 00 
[  4959.189] 00 00 00 00 de 04 00 00 
[  4959.189] 
Backtrace:
[  4959.191] 0: /usr/bin/X (xorg_backtrace+0x2f) [0x45dd7f]
[  4959.191] 1: /usr/bin/X (0x400000+0x45bf3) [0x445bf3]
[  4959.191] 2: /usr/bin/X (mieqEnqueue+0x33) [0x4a0883]
[  4959.191] 3: /usr/bin/X (xf86PostMotionEventM+0xa3) [0x47dc23]
[  4959.191] 4: /usr/bin/X (xf86PostMotionEventP+0x52) [0x47dd42]
[  4959.191] 5: /usr/lib64/xorg/modules/input/evdev_drv.so (0x7f3f4e9de000+0x498b) [0x7f3f4e9e298b]
[  4959.191] 6: /usr/bin/X (0x400000+0x6b678) [0x46b678]
[  4959.191] 7: /usr/bin/X (0x400000+0x119be0) [0x519be0]
[  4959.191] 8: /lib64/libpthread.so.0 (0x3b8d800000+0xf3c0) [0x3b8d80f3c0]
[  4959.191] 9: /lib64/libpthread.so.0 (pthread_cond_wait+0xc5) [0x3b8d80b985]
[  4959.191] 10: /usr/lib64/dri/swrast_dri.so (0x7f3f54772000+0x6bb53) [0x7f3f547ddb53]
[  4959.191] 11: /lib64/libpthread.so.0 (0x3b8d800000+0x7cd1) [0x3b8d807cd1]
[  4959.191] 12: /lib64/libc.so.6 (clone+0x6d) [0x3b8d0dfd3d]
[  4959.191] 
Fatal server error:
[  4959.191] Wrong event type 255. Aborting server

===============================================================================

[ 50769.184] dix: invalid event type 0
[ 50769.186] ff 00 00 00 06 00 00 00 
[ 50769.186] 98 01 00 00 20 ad 06 03 
[ 50769.186] 0a 00 00 00 0a 00 00 00 
[ 50769.187] 00 00 00 00 bb 05 00 00 
[ 50769.187] 
Backtrace:
[ 50769.187] 0: /usr/bin/X (xorg_backtrace+0x2f) [0x45dd7f]
[ 50769.187] 1: /usr/bin/X (0x400000+0x45bf3) [0x445bf3]
[ 50769.187] 2: /usr/bin/X (mieqEnqueue+0x33) [0x4a0883]
[ 50769.187] 3: /usr/bin/X (xf86PostMotionEventM+0xa3) [0x47dc23]
[ 50769.188] 4: /usr/bin/X (xf86PostMotionEventP+0x52) [0x47dd42]
[ 50769.188] 5: /usr/lib64/xorg/modules/input/evdev_drv.so (0x7fb69f19a000+0x498b) [0x7fb69f19e98b]
[ 50769.188] 6: /usr/bin/X (0x400000+0x6b678) [0x46b678]
[ 50769.188] 7: /usr/bin/X (0x400000+0x119be0) [0x519be0]
[ 50769.188] 8: /lib64/libpthread.so.0 (0x3b8d800000+0xf3c0) [0x3b8d80f3c0]
[ 50769.188] 9: /lib64/libpthread.so.0 (pthread_cond_wait+0xc5) [0x3b8d80b985]
[ 50769.188] 10: /usr/lib64/dri/swrast_dri.so (0x7fb6a4f2e000+0x6bb53) [0x7fb6a4f99b53]
[ 50769.188] 11: /lib64/libpthread.so.0 (0x3b8d800000+0x7cd1) [0x3b8d807cd1]
[ 50769.188] 12: /lib64/libc.so.6 (clone+0x6d) [0x3b8d0dfd3d]
[ 50769.188] 
Fatal server error:
[ 50769.188] Wrong event type 255. Aborting server

Comment 20 Tomas Toth 2011-05-07 18:44:01 UTC
Another backtrace:

[ 51729.176] dix: invalid event type 0
[ 51729.178] 00 00 00 00 00 00 00 00 
[ 51729.179] 00 00 00 00 00 00 00 00 
[ 51729.179] 00 00 00 00 00 00 00 00 
[ 51729.179] 00 00 00 00 00 00 00 00 
[ 51729.179] 
Backtrace:
[ 51729.179] 0: /usr/bin/X (xorg_backtrace+0x2f) [0x45dd7f]
[ 51729.179] 1: /usr/bin/X (0x400000+0x45bf3) [0x445bf3]
[ 51729.179] 2: /usr/bin/X (mieqProcessDeviceEvent+0x3f) [0x4a0c9f]
[ 51729.180] 3: /usr/bin/X (mieqProcessInputEvents+0xb0) [0x4a0ee0]
[ 51729.180] 4: /usr/bin/X (ProcessInputEvents+0xd) [0x46b6fd]
[ 51729.180] 5: /usr/bin/X (0x400000+0x2e99c) [0x42e99c]
[ 51729.180] 6: /usr/bin/X (0x400000+0x22e1a) [0x422e1a]
[ 51729.180] 7: /lib64/libc.so.6 (__libc_start_main+0xed) [0x3b8d02143d]
[ 51729.180] 8: /usr/bin/X (0x400000+0x23101) [0x423101]
[ 51729.180] 
Fatal server error:
[ 51729.180] Wrong event type 0. Aborting server

Comment 21 Dave Airlie 2011-05-09 10:36:17 UTC
http://kojipkgs.fedoraproject.org/scratch/airlied/task_3059966/

After some discussion today we think this might be a problem with signals and threads, there scratch build above is a new mesa build, please install and let me know if it makes any difference.

Comment 22 Tomas Toth 2011-05-09 20:19:20 UTC
Created attachment 497908 [details]
Xorg.0.log.old with Backtrace

Sorry for the bad news, it still crashes, see the attached log.

# yum list installed | egrep '^(xorg-x11-serv|mesa)'
mesa-debuginfo.x86_64              7.11-0.9.20110509.0.fc15.da  @/mesa-debuginfo-7.11-0.9.20110509.0.fc15.da.x86_64
mesa-dri-drivers.x86_64            7.11-0.9.20110509.0.fc15.da  @/mesa-dri-drivers-7.11-0.9.20110509.0.fc15.da.x86_64
mesa-dri-filesystem.x86_64         7.11-0.9.20110509.0.fc15.da  @/mesa-dri-filesystem-7.11-0.9.20110509.0.fc15.da.x86_64
mesa-dri-llvmcore.x86_64           7.11-0.9.20110509.0.fc15.da  @/mesa-dri-llvmcore-7.11-0.9.20110509.0.fc15.da.x86_64
mesa-libGL.x86_64                  7.11-0.9.20110509.0.fc15.da  @/mesa-libGL-7.11-0.9.20110509.0.fc15.da.x86_64
mesa-libGLU.x86_64                 7.11-0.9.20110509.0.fc15.da  @/mesa-libGLU-7.11-0.9.20110509.0.fc15.da.x86_64
xorg-x11-server-Xephyr.x86_64      1.10.1-14.bz688693.fc15      @/xorg-x11-server-Xephyr-1.10.1-14.bz688693.fc15.x86_64
xorg-x11-server-Xorg.x86_64        1.10.1-14.bz688693.fc15      @/xorg-x11-server-Xorg-1.10.1-14.bz688693.fc15.x86_64
xorg-x11-server-common.x86_64      1.10.1-14.bz688693.fc15      @/xorg-x11-server-common-1.10.1-14.bz688693.fc15.x86_64
xorg-x11-server-debuginfo.x86_64   1.10.1-14.bz688693.fc15      @/xorg-x11-server-debuginfo-1.10.1-14.bz688693.fc15.x86_64
xorg-x11-server-utils.x86_64       7.5-5.fc15                   @fedora

Comment 23 Tomas Toth 2011-05-09 20:21:41 UTC
And another backtrace:

[  2769.175] dix: invalid event type 0
[  2769.177] 00 00 00 00 00 00 00 00 
[  2769.177] 00 00 00 00 00 00 00 00 
[  2769.178] 0a 00 00 00 0a 00 00 00 
[  2769.178] 00 00 00 00 00 00 00 00 
[  2769.178] 
Backtrace:
[  2769.178] 0: /usr/bin/X (xorg_backtrace+0x2f) [0x45dd7f]
[  2769.178] 1: /usr/bin/X (0x400000+0x45bf3) [0x445bf3]
[  2769.178] 2: /usr/bin/X (mieqProcessDeviceEvent+0x3f) [0x4a0c9f]
[  2769.178] 3: /usr/bin/X (mieqProcessInputEvents+0xb0) [0x4a0ee0]
[  2769.178] 4: /usr/bin/X (ProcessInputEvents+0xd) [0x46b6fd]
[  2769.179] 5: /usr/bin/X (0x400000+0x2e99c) [0x42e99c]
[  2769.179] 6: /usr/bin/X (0x400000+0x22e1a) [0x422e1a]
[  2769.179] 7: /lib64/libc.so.6 (__libc_start_main+0xed) [0x3b8d02143d]
[  2769.179] 8: /usr/bin/X (0x400000+0x23101) [0x423101]
[  2769.179] 
Fatal server error:
[  2769.179] Wrong event type 0. Aborting server
[  2769.179] 
[  2769.179]

Comment 24 Peter Hutterer 2011-05-09 23:31:10 UTC
I'm wondering if there are two overlapping bugs. Look at the backtraces, one comes from the event generation code (evdev, xf86PostMotionEvent, mieqEnqueue) and one from event processing (mieqProcessInputEvents, etc.).

Have you seen both since the mesa upgrade or just the latter?

Comment 25 Dave Airlie 2011-05-11 05:12:58 UTC
I just realised I inverted the signal mask in my mesa patch, so I've done another scratch build that might have a better chance of working

http://kojipkgs.fedoraproject.org/scratch/airlied/task_3064769/

Please give this one a go.

Comment 26 Tomas Toth 2011-05-17 19:53:51 UTC
Good news. No crash since I installed the last May 11th mesa update.
I used the machine only during the last two evenings. I can report back in another couple of days.

Thanks!

Comment 27 Tomas Toth 2011-05-25 19:17:24 UTC
Still no crash. I consider the issue fixed.
Let me know, if you need anything else. 

Thanks to all!

Comment 28 Dave Airlie 2011-05-27 08:14:40 UTC
I've pushed a new mesa to updates-testing that contains the fix.

Comment 29 Tomas Toth 2011-06-01 19:01:48 UTC
The mesa from updates-testing works fine (no crashes) with xorg-x11-server from the fedora repo.

Comment 30 Peter Hutterer 2011-06-02 02:49:12 UTC
Thanks for testing. I'm closing this as CURRENTRELEASE.


http://koji.fedoraproject.org/koji/buildinfo?buildID=245099
https://admin.fedoraproject.org/updates/mesa-7.11-0.11.20110525.0.fc15

Comment 31 Peter Hutterer 2011-07-07 03:23:38 UTC
*** Bug 682593 has been marked as a duplicate of this bug. ***


Note You need to log in before you can comment on or make changes to this bug.