Xorg crashed with SIGSEGV in CheckMotion()

[Matt Zimmerman]

This happened immediately after/while resuming from suspend to RAM. This is the first time I've seen this crash, so I think it's a regression in Jaunty.

ProblemType: Crash
Architecture: amd64
DistroRelease: Ubuntu 9.04
ExecutablePath: /usr/bin/Xorg
Package: xserver-xorg-core 2:1.5.99.3-0ubuntu3
ProcAttrCurrent: unconfined
ProcCmdline: /usr/X11R6/bin/X :0 -br -audit 0 -auth /var/lib/gdm/:0.Xauth -nolisten tcp vt9
ProcEnviron:
 PATH: custom, no user
 LANG=en_GB.UTF-8
ProcVersion: Linux version 2.6.28-3-generic (buildd@yellow) (gcc version 4.3.3 20081210 (prerelease) (Ubuntu 4.3.2-2ubuntu8) ) #4-Ubuntu SMP Fri Dec 12 22:47:31 UTC 2008

Signal: 11
SourcePackage: xorg-server
Stacktrace: #0 0x000000000045708b in CheckMotion ()
StacktraceTop: CheckMotion ()
ThreadStacktrace:

Title: Xorg crashed with SIGSEGV in CheckMotion()
Uname: Linux 2.6.28-3-generic x86_64
UserGroups:

xkbcomp:

Backtrace:
 0: /usr/X11R6/bin/X(xorg_backtrace+0x26) [0x4ef056]
 1: /usr/X11R6/bin/X(xf86SigHandler+0x41) [0x483a51]
 2: /lib/libc.so.6 [0x7fd6e659c030]
 3: /usr/X11R6/bin/X(CheckMotion+0x3b) [0x45708b]
 4: /usr/X11R6/bin/X [0x494bdc]
 5: /usr/X11R6/bin/X(NewInputDeviceRequest+0x22d) [0x494e8d]
 6: /usr/X11R6/bin/X [0x46acde]
 7: /usr/lib/libhal.so.1 [0x7fd6e778ded9]
 8: /lib/libdbus-1.so.3(dbus_connection_dispatch+0x33b) [0x7fd6e755390b]
 9: /lib/libdbus-1.so.3 [0x7fd6e7553ccb]
 10: /usr/X11R6/bin/X [0x46a07b]
 11: /usr/X11R6/bin/X(WakeupHandler+0x4b) [0x4513db]
 12: /usr/X11R6/bin/X(WaitForSomething+0x1ef) [0x4ecadf]
 13: /usr/X11R6/bin/X(Dispatch+0x80) [0x44d5a0]
 14: /usr/X11R6/bin/X(main+0x3bd) [0x4332bd]
 15: /lib/libc.so.6(__libc_start_main+0xe6) [0x7fd6e6587586]
 16: /usr/X11R6/bin/X [0x432749]

[Matt Zimmerman (mdz)]

This one is not easily reproducible; I can usually suspend and resume without problems. Hopefully the apport retracer will shed some light on it.

[Matt Zimmerman (mdz)]

It's happened again, but didn't leave a crash report behind this time. However, I retraced the original one locally to get the following:

Stacktrace:
 #0 CheckMotion (xE=0x0, pDev=0x2cd2580) at ../../dix/events.c:2554
        rootX = <value optimized out>
        rootY = <value optimized out>
        prevSpriteWin = <value optimized out>
        pSprite = (SpritePtr) 0x0

retraced .crash is attached.

[Matt Zimmerman (mdz)]

I'm attaching a syslog excerpt showing the resume process. Xorg.0.log.old (with a stack trace) is dated 10:52:55.

The only suspect thing I see in the log is:

Dec 27 10:52:54 perseus kernel: [28220.551201] input: X10 Wireless Technology Inc USB Receiver as /devices/pci0000:00/0000:00:1a.1/usb3/3-1/input/input11
Dec 27 10:52:54 perseus kernel: [28220.567136] usb 3-1: Weird data, len=1 ff 00 00 00 14 20 ...

which indicates the USB remote control went a little wonky. Note that is the same device involved in bug 311254.

[Martin Pitt (pitti)]

I get the very same crash on current jaunty i386. I can reproduce it every time.

[D (dj-lp)]

Same here. Jaunty AMD64, Intel v. 2.5.99.2 Log attached. Using 2.6.29-rc1.

[D (dj-lp)]

Oh, sorry. Posted in the wrong window.

[Martin Pitt (pitti)]

This now seems to be fixed for me on current Jaunty. I just ran the kernel suspend-resume test script successfully and also used the g-p-m applet. X came back up correctly in all cases.

Matt, does it now work for you as well again?

[Matt Zimmerman (mdz)]

On Mon, Jan 19, 2009 at 09:10:50AM -0000, Martin Pitt wrote:
> This now seems to be fixed for me on current Jaunty. I just ran the
> kernel suspend-resume test script successfully and also used the g-p-m
> applet. X came back up correctly in all cases.
>
> Matt, does it now work for you as well again?

I've never been able to reproduce it; it only happened to me one time. I
can say that the suspend/resume test is successful on the same system with
current Jaunty.

--
 - mdz

[Martin Pitt (pitti)]

Let's consider this fixed then for now.