Randomly crashing (SIGSEGV/SIGPIPE)

[Rebecca Palmer]

Binary package hint: firefox

I recently installed Ubuntu 6.10 in place of Debian 3.1 in the hope of ending its frequent crashes (see Debian bug http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=393650), but Firefox is still randomly closing or hanging; sometimes, but not always, it then crashes again on attempting to restart it.

I copied my bookmarks and cookies, but nothing else, from my old Firefox profile, and have no extensions beyond the default Ubuntu Firefox. JavaScript and Java are off. Connection is via college network.

The crash typically occurs once every few hours, with no pattern yet found of what triggers it, and the traces can point to a lot of different places; apart from the one included, I've had jsInterp.c (no details because it took the gdb terminal down with it before the whole trace appeared), SIGSEGV at 0xb4988dcb in ?? () from /usr/lib/firefox/components/libhtmlpars.so, SIGPIPE in a __kernel_vsyscall () <- send () <- 0xb7e62a03 in pt_Send (parameters removed by me as I was on a secure site at the time) at ptio.c:1902, and SIGSEGV at (no address) InsertPropertyTreeChild (removed) at jsscope.c:580. Because of this wide variation I do not know if this is related to the existing Firefox random crashers #24728 and #48772.

It is possible that the cause is outside Firefox (I run a commercial
on-access antivirus which doesn't officially support Ubuntu
(http://www.sophos.com/support/knowledgebase/article/4377.html) and keep getting "couldn't create backup" warnings when saving files in gedit; I am awaiting a response from them as to whether this could be the problem). Memory test (Debian memtest86+) was OK but could it be some other hardware problem?

The crash traced below happened while viewing http://www.fortunecity.com/business/director/1255/31xxxtech.html as the only tab in that window; a second window with 3 tabs also went down. Firefox then restarted normally but Bug Buddy then reported a "crash" in CD/DVD Creator (which wasn't open) as http://bugzilla.gnome.org/show_bug.cgi?id=395406

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread -1220598096 (LWP 5067)]
0xb58315c9 in imgRequestProxy::FrameChanged (this=0x8c55ce8, container=0x90e2230, newframe=0x8f3d860, dirtyRect=0xbf904ca4) at ../../../dist/include/xpcom/nsCOMPtr.h:627
627 ../../../dist/include/xpcom/nsCOMPtr.h: No such file or directory.
        in ../../../dist/include/xpcom/nsCOMPtr.h
(gdb) thread apply all bt

Thread 83 (Thread -1353098336 (LWP 5181)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deba8c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e153b3 in pt_TimedWait (cv=0x80d2984, ml=0x80d2920, timeout=<value optimized out>) at ptsynch.c:280
#3 0xb7e16241 in PR_WaitCondVar (cvar=0x80d2980, timeout=60000) at ptsynch.c:407
#4 0xb71c0567 in nsHostResolver::GetHostToLookup (this=0x80d28c0, result=0xaf595448) at nsHostResolver.cpp:556
#5 0xb71c0a04 in nsHostResolver::ThreadFunc (arg=0x80d28c0) at nsHostResolver.cpp:641
#6 0xb7e1c1ed in _pt_root (arg=0xb22a5df0) at ptthread.c:220
#7 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#8 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 82 (Thread -1296815200 (LWP 5180)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deba8c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e153b3 in pt_TimedWait (cv=0x829870c, ml=0x82986a8, timeout=<value optimized out>) at ptsynch.c:280
#3 0xb7e16241 in PR_WaitCondVar (cvar=0x8298708, timeout=60000) at ptsynch.c:407
#4 0xb718f645 in nsIOThreadPool::ThreadFunc (arg=0x8298678) at nsIOThreadPool.cpp:254
#5 0xb7e1c1ed in _pt_root (arg=0x940b328) at ptthread.c:220
#6 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#7 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 15 (Thread -1344681056 (LWP 5084)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deb816 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e162d1 in PR_WaitCondVar (cvar=0x8716038, timeout=4294967295) at ptsynch.c:405
#3 0xb7e16337 in PR_Wait (mon=0x8716078, timeout=4294967295) at ptsynch.c:584
#4 0xb7ea6971 in PL_WaitForEvent (self=0x8716008) at plevent.c:808
#5 0xb7ea8249 in nsEventQueueImpl::WaitForEvent (this=0x83ffaa0, aResult=0xafd9c440) at nsEventQueue.cpp:498
#6 0xb679cd45 in EventLoop (arg=0x0) at nsUrlClassifierDBService.cpp:1142
#7 0xb7e1c1ed in _pt_root (arg=0x8b2d200) at ptthread.c:220
#8 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#9 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 14 (Thread -1334957152 (LWP 5082)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deba8c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e153b3 in pt_TimedWait (cv=0xb2267844, ml=0xb22677e0, timeout=<value optimized out>) at ptsynch.c:280
#3 0xb7e16241 in PR_WaitCondVar (cvar=0xb2267840, timeout=250) at ptsynch.c:407
#4 0xb080b004 in nsCertVerificationThread::Run (this=0xb2267790) at nsCertVerificationThread.cpp:142
#5 0xb0809390 in nsPSMBackgroundThread::nsThreadRunner (arg=0xb2267790) at nsPSMBackgroundThread.cpp:44
#6 0xb7e1c1ed in _pt_root (arg=0xb2267880) at ptthread.c:220
#7 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#8 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 13 (Thread -1316484192 (LWP 5081)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deba8c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e153b3 in pt_TimedWait (cv=0xb2267674, ml=0xb2267610, timeout=<value optimized out>) at ptsynch.c:280
#3 0xb7e16241 in PR_WaitCondVar (cvar=0xb2267670, timeout=250) at ptsynch.c:407
#4 0xb080a287 in nsSSLThread::Run (this=0xb22675e8) at nsSSLThread.cpp:879
#5 0xb0809390 in nsPSMBackgroundThread::nsThreadRunner (arg=0xb22675e8) at nsPSMBackgroundThread.cpp:44
#6 0xb7e1c1ed in _pt_root (arg=0xb22676b0) at ptthread.c:220
#7 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#8 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 8 (Thread -1308091488 (LWP 5076)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deb816 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e162d1 in PR_WaitCondVar (cvar=0x87be650, timeout=4294967295) at ptsynch.c:405
#3 0xb2086f1d in ProcessAsyncMessages () at mozStorageAsyncIO.cpp:1525
#4 0xb2087ff9 in AsyncWriteThread::Run (this=0x87be690) at mozStorageAsyncIO.cpp:471
#5 0xb7eaa80b in nsThread::Main (arg=0x87be6a0) at nsThread.cpp:118
#6 0xb7e1c1ed in _pt_root (arg=0x87be720) at ptthread.c:220
#7 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#8 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 3 (Thread -1234052192 (LWP 5071)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb7deba8c in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7e153b3 in pt_TimedWait (cv=0x8081f44, ml=0x8081a20, timeout=<value optimized out>) at ptsynch.c:280
#3 0xb7e16241 in PR_WaitCondVar (cvar=0x8081f40, timeout=1365) at ptsynch.c:407
#4 0xb7eacb06 in TimerThread::Run (this=0x8081c20) at TimerThread.cpp:318
#5 0xb7eaa80b in nsThread::Main (arg=0x81645a8) at nsThread.cpp:118
#6 0xb7e1c1ed in _pt_root (arg=0x8164628) at ptthread.c:220
#7 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#8 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 2 (Thread -1223910496 (LWP 5070)):
#0 0xffffe410 in __kernel_vsyscall ()
#1 0xb77f5803 in poll () from /lib/tls/i686/cmov/libc.so.6
#2 0xb7e17f89 in PR_Poll (pds=0x80d1cf0, npds=1, timeout=4294967295) at ptio.c:3877
#3 0xb71b448b in nsSocketTransportService::Poll (this=0x80d1808, interval=0xb70c9428) at nsSocketTransportService2.cpp:361
#4 0xb71b4cc2 in nsSocketTransportService::Run (this=0x80d1808) at nsSocketTransportService2.cpp:577
#5 0xb7eaa80b in nsThread::Main (arg=0x80d1ff8) at nsThread.cpp:118
#6 0xb7e1c1ed in _pt_root (arg=0x80d2078) at ptthread.c:220
#7 0xb7de8504 in start_thread () from /lib/tls/i686/cmov/libpthread.so.0
#8 0xb77ff51e in clone () from /lib/tls/i686/cmov/libc.so.6

Thread 1 (Thread -1220598096 (LWP 5067)):
#0 0xb58315c9 in imgRequestProxy::FrameChanged (this=0x8c55ce8, container=0x90e2230, newframe=0x8f3d860, dirtyRect=0xbf904ca4) at ../../../dist/include/xpcom/nsCOMPtr.h:627
#1 0xb582ef17 in imgRequest::FrameChanged (this=0x8c63a38, container=0x90e2230, newframe=0x8f3d860, dirtyRect=0xbf904ca4) at imgRequest.cpp:407
#2 0xb5836728 in imgContainerGIF::Notify (this=0x90e2230, timer=0x8d1cc58) at imgContainerGIF.cpp:457
#3 0xb7eaafdd in nsTimerImpl::Fire (this=0x8d1cc58) at nsTimerImpl.cpp:397
#4 0xb7eab8a8 in handleTimerEvent (event=0xb225fea8) at nsTimerImpl.cpp:459
#5 0xb7ea6a57 in PL_HandleEvent (self=0xb225fea8) at plevent.c:688
#6 0xb7ea6d6b in PL_ProcessPendingEvents (self=0x80cc078) at plevent.c:623
#7 0xb7ea8cbe in nsEventQueueImpl::ProcessPendingEvents (this=0x80cc030) at nsEventQueue.cpp:417
#8 0xb684a5d5 in event_processor_callback (source=0x85f9580, condition=G_IO_IN, data=0x8e49d18) at nsAppShell.cpp:67
#9 0xb760ac8d in g_io_channel_unix_get_fd () from /usr/lib/libglib-2.0.so.0
#10 0xb75e1802 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#11 0xb75e47df in g_main_context_check () from /usr/lib/libglib-2.0.so.0
#12 0xb75e4b89 in g_main_loop_run () from /usr/lib/libglib-2.0.so.0
#13 0xb7bc9574 in gtk_main () from /usr/lib/libgtk-x11-2.0.so.0
#14 0xb684aa32 in nsAppShell::Run (this=0x8116788) at nsAppShell.cpp:139
#15 0xb6768ca2 in nsAppStartup::Run (this=0x813e298) at nsAppStartup.cpp:151
#16 0x0804f85c in XRE_main (argc=3, argv=0xbf9053e4, aAppData=0x805ab80) at nsAppRunner.cpp:2444
#17 0x0804ab4f in main (argc=Cannot access memory at address 0x1
) at nsBrowserApp.cpp:61
#18 0xb77478cc in __libc_start_main () from /lib/tls/i686/cmov/libc.so.6
#19 0x0804aa81 in _start ()
(gdb)

[Rebecca Palmer (rebecca-palmer)]

I now suspect this problem is not in Firefox after all, as it is now sometimes taking the whole system down rather than just Firefox (a common pattern is for X to suddenly restart for no apparent reason, get as far as logging in then restart again, and keep repeating this until X decides something is wrong and waits, at which point all I can do is restart the system; more rarely error messages appear that look like log messages but don't actually make it to the system logs, the only set I managed to cut and paste before the system went down being included below), and sometimes applications other than Firefox crash.

I suspect this is really either an anti-virus software problem (I have also reported it to Sophos) or a kernel/driver/hardware compatibility problem, or possibly a hardware fault though it doesn't seem to affect Win98; please move it to wherever you think is appropriate. I am also having video problems (https://launchpad.net/bugs/82434), though since they are new in Ubuntu while this problem was also present in Debian I suspect they are a separate problem.

Hardware as reported by Device Manager:
CS 4614/22/24 [CrystalClear SoundFusion Audio Accelerator]
Realtek RT8139 network card
Allied Data Technologies Unknown (0x0015) (I think this is my modem, which I don't use)
Intel 82371AB/EB/MB PIIX4 ISA
Intel 82371AB/EB/MB PIIX4 IDE
HDS728080PLAT20 (Primary master, partitioned approx 40GB FAT32 Win98 boot/0.5GB Swap/40GB ext3 Ubuntu boot)
ST36421A (Secondary master, approx 6GB FAT32)
OPTOWRITECD-RW CW5205 (Secondary slave)
Intel 82371AB/EB/MB PIIX4 USB
Intel 82371AB/EB/MB PIIX4 ACPI (Power management is turned off in Services but I still get an "ACPI: Unable to locate RSDP" error followed by a "GConf schema installer error, battery_low_percentage cannot be zero" on every startup)
Intel 440LX/EX - 82443LX/EX Host bridge
Intel 440LX/EX - 82443LX/EX AGP bridge
ATI 3D Rage Pro Turbo AGP 2X graphics
Mouse, floppy drive, UK keyboard
This is consistent with Win98's identification. Memory 256MB, processor 500MHz Celeron.

The following errors occurred on closing gdb after Firefox had run successfully; I then attempted to reboot the system but it went to a black screen with a text cursor, and on pressing Ctrl-Alt-Del showed "init:process rc6 (number) killed by signal 15", went to the Ubuntu screen with the progress bar stopped at 2-3 blocks from the end, then went back to the blank screen.

Program exited normally.
(gdb) quit

Message from syslogd@mert1529 at Thu Jan 11 16:52:13 2007 ...
mert1529 kernel: [32929.412974] ------------[ cut here ]------------

Message from syslogd@mert1529 at Thu Jan 11 16:52:13 2007 ...
mert1529 kernel: [32929.413008] kernel BUG at mm/rmap.c:560!

Message from syslogd@mert1529 at Thu Jan 11 16:52:13 2007 ...
mert1529 kernel: [32929.413018] invalid opcode: 0000 [#1]

Message from syslogd@mert1529 at Thu Jan 11 16:52:13 2007 ...
mert1529 kernel: [32929.413025] SMP

Message from syslogd@mert1529 at Thu Jan 11 16:52:13 2007 ...
mert1529 kernel: [32929.413210] CPU: 0

Message from syslogd@mert1529 at Thu Jan 11 16:52:13 2007 ...
mert1529 kernel: [32929.413274] EIP is at page_remove_rmap+0x27/0x40

Message from sys...

[Rebecca Palmer (rebecca-palmer)]

The attached strace is from a SIGSEGV crash while starting Firefox (before anything visible had happened). My home page is https://webmail.ox.ac.uk/; when I started Firefox again without strace it loaded successfully and displayed a "server not found" error, at which point I noticed that I had accidentally left the network unplugged.

According to Sophos, there is no indication of their anti-virus software being involved.

This strace is from kernel 2.6.17-10, but the problem is still present after the recent update to 2.6.17-11.

I have also discovered that sound isn't working, though I suspect this is a separate problem.

[John Vivirito (gnomefreak)]

Please attach the full crash report when it crashes on the kernel you are using.

[John Vivirito (gnomefreak)]

Also what version of firefox are you using. What plugins are installed? What extentions are installed? What other add-ons are installed?

[Alexander Sack (asac)]

Hi,

we need the full crash report attached as a file. Do you still have it?

[Rebecca Palmer (rebecca-palmer)]

gdb backtrace with the new kernel attached (this one was a hang, after which it would restart normally).

My Firefox is the Ubuntu 6.10 default (2.0.0.1) with all security updates; the only non-default plugin I have is RealPlayer and the problem started before I installed that.

As indicated above, I now strongly suspect that this isn't really a Firefox problem at all, since other applications also sometimes crash.

(It turned out there was nothing wrong with the sound after all-the volume just needed to be set higher than I expected to make it audible.)

[Ivan Gualandri (ivang)]

Attached you find the backtrace of System Settings application.

Ivan

[John Vivirito (gnomefreak)]

Rebecca,
Are you able to reproduce the firefox crash with lastest versions and if so can you please try running firefox -safe-mode in a terminal and see if you can reproduce it.

Ivan,
What does system-settings crash have to do with firefox or Rebecca's complaint. Please file a separate bug and attach it there, your bug is on kde-system-settings not firefox at all.

[Rebecca Palmer (rebecca-palmer)]

I last saw this recently, though I still can't reproduce it at will. It happens in both GNOME and Xfce.

Recently I also had another bug: the "You are about to view an encrypted page that contains some unencrypted information" alert sometimes appeared on https://webmail.ox.ac.uk, which hadn't triggered such a warning before and according to its technical support shouldn't.

This may be a duplicate of https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/60236; I have not tried any of their suggested workarounds.

I will try -safe-mode next time I see the bug.

[Rebecca Palmer (rebecca-palmer)]

After two crashes after which Firefox would restart normally, several more occurrences of the "You are about to view an encrypted page that contains some unencrypted information" warning, and a Nautilus crash (of which I include the backtrace as I suspect this may be a system rather than Firefox problem), the Firefox crash and won't restart reappeared today.

It still would not restart with -safe-mode; backtraces and the ends of straces with and without this are attached. As usual, reinstalling Firefox fixed the problem.

While this unusually long interval without seeing it (though with the above other problems) occurred after replacing the power supply following a fan failure, its occurrence does not appear to be related to how long the computer has been on.

[John Vivirito (gnomefreak)]

Rebecca and Ivan,
is this still an issue in newest firefox (try with new profile as well) and if it is reproducible can you please give step by step instructions on how to reproduce this bug.
Please file separate bugs for all otehr crashes but i feel this is either a profile or a gtk issue.

[Rebecca Palmer (rebecca-palmer)]

I still do not have a way to reproduce this bug; I appreciate that this can't be helpful.

My Firefox is Edgy default with all security updates. https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/60236 reports the same symptoms in Feisty and Dapper.

Since other applications are crashing as well (I originally reported the Nautilus problem as a separate bug http://bugzilla.gnome.org/show_bug.cgi?id=395406, but I suspect they may be connected) this could well be in something other than Firefox itself. I have now installed the GTK debugging symbols. (Does Firefox still use GTK when started from a non-GNOME desktop? I've also seen this in Xfce, and #60236 has been reported from KDE)

I will try a clean profile test and #60236's ssh workaround next time the bug appears.

[John Vivirito (gnomefreak)]

Im leaning towards GTK2 issue, yes firefox always uses gtk libs no matter what desktop you are in. Im gonna close this bug please file a bug on gtk with all your info in it, that should be alot more helpful to debug it. If you can run a backtrace with all debugging symbols installed you can see how to do this at: https://wiki.ubuntu.com/MozillaTeam/Bugs or at:
https://wiki.ubuntu.com/Backtrace

[Rebecca Palmer (rebecca-palmer)]

Moved to GTK2.

[John Vivirito (gnomefreak)]

Thanks but the reason i asked you to file another bug report on GTK was so the people that are getting emails for this bug will still get them and they may not want them.

[Rebecca Palmer (rebecca-palmer)]

Who? The "subscribers" panel has automatically replaced the Firefox maintainers with the GTK2 maintainers.

I don't want to start an argument, but according to https://wiki.ubuntu.com/Bugs/CommonTasks and https://wiki.ubuntu.com/MozillaTeam/Bugs/Triage it is Ubuntu policy to move rather than close bugs filed under the wrong package.

[Rebecca Palmer (rebecca-palmer)]

Firefox crashed again today (all windows disappeared, after some seconds when they weren't functioning); backtrace from apport attached. Restarting it normally worked, so I couldn't test anything else.

Something like "checking your add-ons are compatible with the new version of Firefox" had appeared on startup (presumably referring to the update to 2.0.0.5); the pages I had open when it crashed were: http://www.randi.org/education/links.html
http://ods.od.nih.gov/factsheets/calcium.asp and http://www.sciencedirect.com/science?_ob=ArticleListURL&_method=list&_ArticleListID=601294619&_sort=d&view=c&_acct=C000010360&_version=1&_urlVersion=0&_userid=126524&md5=977fb539ecd73c70756a7ceaf8448d0e (two tabs in one window)
http://www.sciencedirect.com/science?_ob=HelpURL&_file=qs_main.htm&_acct=C000010360&_version=1&_urlVersion=0&_userid=126524&md5=19dafca93101182f345c5c407c9fecd3 (in the process of loading in a new window when the crash happened)

Since this backtrace, and several others I have had, point to __kernel_vsyscall, is there anything else I need to post so you can tell which thread crashed?

[Rebecca Palmer (rebecca-palmer)]

I haven't seen the Firefox crash and won't restart since 10 Jul, or the one-off Firefox crash since 26 Jul. However due to the intermittent nature of the problem, it is impossible to be sure whether it has really gone away.

On 31 Jul, the computer was moved from a university network to an ADSL/router home network. To get it to work there IPv6 was disabled in Firefox about:config. SophosAV was also removed and replaced with ClamAV.

I have however had a gnome-panel crash today; a Bug Buddy window opened, both top and bottom bars disappeared, and a few seconds after Bug Buddy was closed the bars reappeared with the normal task buttons on the bottom but nothing on the top, but after a few more seconds another Bug Buddy window appeared and the cycle repeated. The open applications (Firefox showing uk.arxiv.org and gedit just opened showing a large tex file) appeared normal, and were successfully closed normally. The backtrace attached is from the second crash, as the first report disappeared when I right-clicked in it intending to copy and paste it. After a few cycles I rebooted the system.

[Rebecca Palmer (rebecca-palmer)]

Firefox won't restart reappeared today...

The initial problem was a hang on scrolling down to the picture in http://orbit.m6.net/Forum/default.aspx?g=posts&t=15835&p=8; I clicked the close button and selected Force Quit.

I then found I couldn't start gnome-terminal, but could start xfce4-terminal so did my testing there. A Bug Buddy window without any details also appeared during the tests, which I ignored until I had finished.

$ firefox
{nothing visible happened until I pressed Ctrl-C}
$ firefox -g
$firefox --debugger "strace -ofirefoxstrace"
{segfaulted, see next message for logs}
$ firefox -ProfileManager
{segfaulted without showing the profile selector}
$ firefox --debugger "valgrind -v --tool=memcheck --leak-check=full --num-callers=40 --log-file=/home/palmer/firefoxvalgrind.log"
/usr/bin/firefox: line 161: 4591 Killed "$@"
{several 4 byte reads/writes overlapping the end of a buffer, see attached log}
$ ssh localhost DISPLAY=:0 firefox
ssh: connect to host localhost port 22: Connection refused
{iptables -nvL (firewall stats) showed two packets out and two in per attempt, both of which were being accepted}

After these tests, I reinstalled Firefox to find this didn't fix the problem. Rebooting the system did fix it; I had to do this with Alt-SysRq-euib as Quit->Shut Down didn't work.

[Rebecca Palmer (rebecca-palmer)]

This file contains the original hang's backtrace, restart attempt backtrace and strace, and gnome-terminal backtrace.

[Sebastien Bacher (seb128)]

The backtrace has no GTK functions, that looks like a firefox bug rather, reassigning to it

[Rebecca Palmer (rebecca-palmer)]

I tried running Firefox under Valgrind when it was working, but found it to run too slowly for doing this until the bug reappeared (the only way to get a Valgrind log of the initial crash rather than a failed restart) to be practical.

The attached is a Valgrind log of opening and closing Firefox without doing anything else; I don't know if this is of any use.

Unfortunately a lot of the "errors" reported by Valgrind look like they could be false alarms from the Valgrind bugs https://bugs.launchpad.net/ubuntu/+source/valgrind/+bug/13460 and https://bugs.launchpad.net/ubuntu/+source/valgrind/+bug/39491

[Alexander Sack (asac)]

On Thu, Aug 23, 2007 at 12:42:03PM -0000, Rebecca Palmer wrote:
> I tried running Firefox under Valgrind when it was working, but found it
> to run too slowly for doing this until the bug reappeared (the only way
> to get a Valgrind log of the initial crash rather than a failed restart)
> to be practical.
>
> The attached is a Valgrind log of opening and closing Firefox without
> doing anything else; I don't know if this is of any use.
>

Please try to reproduce the crash while valgrind is running
... without the crash we cannot be sure that a clue for this bug is
actually in your log.

 - Alexander

[Rebecca Palmer (rebecca-palmer)]

I've since had the X repeated restarts at least twice (X either restarts itself or hangs so I try to restart it, then restarts over and over again, requiring an Alt-SysRq-euib system restart), at least one with Firefox not open (though Thunderbird was).

Today I started the system and immediately found that Thunderbird wouldn't start (having closed it normally last time I used it; note that I began using Thunderbird only recently, so it may have been affected earlier), then found that Firefox, gedit, gnome-terminal and nautilus wouldn't start either.

Backtraces, straces and valgrinds attached (sorry for lack of symbols, dpkg-query -W didn't find any debug packages; backtrace only for Thunderbird because its --debugger option doesn't pass arguments). All the valgrinds have the same error:

==4956== Invalid read of size 4
==4956== at 0x4148114: pthread_mutex_lock (pthread_mutex_lock.c:44)
==4956== by 0x4DD1BBD: (within /usr/lib/libORBit-2.so.0.1.0)
==4956== by 0x4DD28A8: giop_recv_buffer_get (in /usr/lib/libORBit-2.so.0.1.0)
==4956== by 0x4DD733C: ORBit_small_invoke_stub (in /usr/lib/libORBit-2.so.0.1
.0)
==4956== by 0x4DD753D: ORBit_small_invoke_stub_n (in /usr/lib/libORBit-2.so.0
.1.0)
==4956== by 0x4DE3D31: ORBit_c_stub_invoke (in /usr/lib/libORBit-2.so.0.1.0)
=

causing SIGSEGV termination, this being the first error with the address being 0xC in gedit/gnome-terminal/nautilus and the fourth with 0x346D in firefox.

[John Vivirito (gnomefreak)]

Is this still a problem in version 3.0.7 from your repos?
Changed firefox to firefox-3.0 since 2.0 has reached EOLS

[Rebecca Palmer (rebecca-palmer)]

I haven't regularly used the machine that had this bug for over a year (my current machine has never had it), so don't know if 3.0 still has it, and as I can't reproduce it there is no quick way to test for it; unless anyone else has seen it in a more recent version I suggest that it be closed.

[John Vivirito (gnomefreak)]

If you cant reproduce this than it should be closed and im doing that now

[John Vivirito (gnomefreak)]

If you see it again please use apport to report the bug.