On Wednesday 24 November 2010, 15:02:37 Marcus Meissner wrote:

On Wed, Nov 24, 2010 at 12:24:12PM +0100, Hans-Peter Jansen wrote:

...

Hi,

I have to handle a (now outdated) diskless openSUSE 11.1 desktop installation of about 30 systems in a company network. One of the customizations is Firefox 3.6.12 from the mozilla repo, consisting of MozillaFirefox-3.6.12-1.2, mozilla-xulrunner192-1.9.2.12-1.2, ... Kernel is a 2.6.31.13 also based on the openSUSE trees on gitorious.

Unfortunately, I'm getting complains of Firefox crashing often during the day without any user interaction, just displaying occasionly updated sites (without flash content), and I do suffer from these crashes, too.

Previous versions from that repo doesn't show this behavior. In that installation, only Firefox, flash and acrobat reader were updated lately, everything else was being kept as is for a few month running (rock solid!), therefor I can say, that this problem is triggered by changes in firefox related packages.

Last night, I've kept Firefox running under gdb control with just the build system monitor (where I was logged in) and a few of my project pages. The only "active" content was the monitor page occasionly updating. Today, I found:

Program received signal SIGPIPE, Broken pipe. [Switching to Thread 0xb43ffb90 (LWP 8683)] 0xffffe424 in __kernel_vsyscall () (gdb) bt

This is just a read thing, it waits for sending SSL traffic to somewhere on close and the remote socket might got closed too early.

...

Does this rings a bell for somebody? Any advice, how to circumvent this problem?

Perhaps other threads of firefox need to be inspected, check: info threads thread 1 bt thread 2 bt thread 3

to get all backtraces. (Not sure if "bt all" should work, it did not work for me).

Ahh, I see. Since it will crash again as safe as the Bank of England¹, I will follow up with a more detailed crash report. Thanks for the hints, Marcus. Pete ¹) I really ask myself: Is such an expression politically correct in these times? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org

Am 09.12.2010 11:17, schrieb Hans-Peter Jansen:

Thread 1 (Thread 0xb7c236d0 (LWP 11475)): #0 0xb7880a0c in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #1 0xb7881e37 in js_LookupProperty () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #2 0xb7881ea6 in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #3 0xb7881c9d in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #4 0xb785f384 in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #5 0xb7870f4b in js_Invoke () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #6 0xb66295d3 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #7 0xb6624f83 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so ---Type <return> to continue, or q <return> to quit--- #8 0xb6f4e198 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #9 0xb66edc65 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #10 0xb6f40772 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #11 0xb6f0f00c in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #12 0xb6e840f6 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #13 0xb6edc99c in MessageLoop::RunInternal() () from /usr/lib/xulrunner-1.9.2.12/libxul.so #14 0xb6edc9c0 in MessageLoop::RunHandler() () from /usr/lib/xulrunner-1.9.2.12/libxul.so #15 0xb6edca37 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #16 0xb6dfc3bc in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #17 0xb6cc0d30 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #18 0xb65f2e73 in XRE_main () from /usr/lib/xulrunner-1.9.2.12/libxul.so #19 0x08049f6b in ?? () #20 0xb7c45705 in __libc_start_main (main=0x80499a4 , argc=1, ubp_av=0xbfffebd4, init=0x8058400, fini=0x80583f0, rtld_fini=0xb7fef210 <_dl_fini>, stack_end=0xbfffebcc) at libc-start.c:220 #21 0x080498e1 in ?? ()

Here's, what I've installed right now: mozilla-xulrunner192-1.9.2.12-1.2 mozilla-xulrunner192-buildsymbols-1.9.2.12-1.2 mozilla-xulrunner192-debuginfo-1.9.2.12-1.2 mozilla-xulrunner192-debugsource-1.9.2.12-1.2 mozilla-xulrunner192-gnome-1.9.2.12-1.2 mozilla-xulrunner192-translations-common-1.9.2.12-1.2

You need mozilla-js192 and its debuginfo as well.

BTW, what's up with the buildsymbols packages? A new debug symbols variant?

Yes, but it's not used for local gdb debugging but only for automatic retracing of minidumps sent to crash-reports.mozilla.com

Why am I not surprised, that it crashed somewhere in js (if my analysis is sound), probably because it wrongly does not ignore the SIGPIPE. The question is, what can trigger SIGPIPEs in the JS interpreter?

Would be really interesting to see more symbols here. Wolfgang -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org

Am 09.12.2010 20:50, schrieb Hans-Peter Jansen:

Dear Wolfgang,

thanks for your answer and sorry for the delay.

On Thursday 09 December 2010, 11:41:46 Wolfgang Rosenauer wrote:

...

Am 09.12.2010 11:17, schrieb Hans-Peter Jansen:

...

Thread 1 (Thread 0xb7c236d0 (LWP 11475)): #0 0xb7880a0c in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #1 0xb7881e37 in js_LookupProperty () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #2 0xb7881ea6 in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #3 0xb7881c9d in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #4 0xb785f384 in ?? () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #5 0xb7870f4b in js_Invoke () from /usr/lib/xulrunner-1.9.2.12/libmozjs.so #6 0xb66295d3 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #7 0xb6624f83 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so ---Type <return> to continue, or q <return> to quit--- #8 0xb6f4e198 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #9 0xb66edc65 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #10 0xb6f40772 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #11 0xb6f0f00c in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #12 0xb6e840f6 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #13 0xb6edc99c in MessageLoop::RunInternal() () from /usr/lib/xulrunner-1.9.2.12/libxul.so #14 0xb6edc9c0 in MessageLoop::RunHandler() () from /usr/lib/xulrunner-1.9.2.12/libxul.so #15 0xb6edca37 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #16 0xb6dfc3bc in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #17 0xb6cc0d30 in ?? () from /usr/lib/xulrunner-1.9.2.12/libxul.so #18 0xb65f2e73 in XRE_main () from /usr/lib/xulrunner-1.9.2.12/libxul.so #19 0x08049f6b in ?? () #20 0xb7c45705 in __libc_start_main (main=0x80499a4 , argc=1, ubp_av=0xbfffebd4, init=0x8058400, fini=0x80583f0, rtld_fini=0xb7fef210 <_dl_fini>, stack_end=0xbfffebcc) at libc-start.c:220 #21 0x080498e1 in ?? ()

Here's, what I've installed right now: mozilla-xulrunner192-1.9.2.12-1.2 mozilla-xulrunner192-buildsymbols-1.9.2.12-1.2 mozilla-xulrunner192-debuginfo-1.9.2.12-1.2 mozilla-xulrunner192-debugsource-1.9.2.12-1.2 mozilla-xulrunner192-gnome-1.9.2.12-1.2 mozilla-xulrunner192-translations-common-1.9.2.12-1.2

You need mozilla-js192 and its debuginfo as well.

# LANG=C zypper se -s mozilla-js192 Loading repository data... Reading installed packages...

S | Name | Type | Version | Arch | Repository --+---------------+---------+----------------+------+--------------------------------------- v | mozilla-js192 | package | 1.9.2.13-3.1 | i586 | Mozilla based projects (openSUSE_11.1) v | mozilla-js192 | package | 1.9.2.12-0.1.1 | i586 | openSUSE-11.1-Updates v | mozilla-js192 | package | 1.9.2.10-0.4.1 | i586 | openSUSE-11.1-Updates v | mozilla-js192 | package | 1.9.2.8-1.1.1 | i586 | openSUSE-11.1-Updates i | mozilla-js192 | package | 1.9.2.12-1.2 | i586 | (System Packages)

There's no such debug package, at least not for 11.1:

http://download.opensuse.org/repositories/mozilla/openSUSE_11.1/i586/

Build issue?

Hmm, no. I can find | mozilla-js192-debuginfo | package | 1.9.2.13-3.1 | i586 | mozilla (note the new version which was released today) Wolfgang -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org

Am 10.12.2010 00:21, schrieb Hans-Peter Jansen:

...

...

There's no such debug package, at least not for 11.1:

http://download.opensuse.org/repositories/mozilla/openSUSE_11.1/i586/

Build issue?

Hmm, no. I can find | mozilla-js192-debuginfo | package | 1.9.2.13-3.1 | i586 | mozilla

With all due respect, yes. Directly from:

Hmm, I haven't expected a difference between 11.3 and 11.1 here. But now that you point it out I _think_ that debuginfo packages for subpackages were not supported in 11.1 and the contents do live in the one debuginfo available. So the needed data should be on your system already. Sorry for the confusion.

https://build.opensuse.org/package/rawlog?arch=i586&package=mozilla-xulrunner192&project=mozilla&repository=openSUSE_11.1

The spec doesn't show any salience in this respect either.

Because everything is done automatically and therefore differently automatically between 11.3 and 11.1. Wolfgang -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org

On Friday 10 December 2010, 08:32:36 Hans-Peter Jansen wrote:

On Friday 10 December 2010, 00:27:18 Wolfgang Rosenauer wrote:

...

Am 10.12.2010 00:21, schrieb Hans-Peter Jansen:

...

...

...

There's no such debug package, at least not for 11.1:

http://download.opensuse.org/repositories/mozilla/openSUSE_11.1 /i 586/

Build issue?

Hmm, no. I can find

| mozilla-js192-debuginfo | package | 1.9.2.13-3.1 | | i586 |

mozilla

With all due respect, yes. Directly from:

Hmm, I haven't expected a difference between 11.3 and 11.1 here. But now that you point it out I _think_ that debuginfo packages for subpackages were not supported in 11.1 and the contents do live in the one debuginfo available. So the needed data should be on your system already. Sorry for the confusion.

...

https://build.opensuse.org/package/rawlog?arch=i586&package=mozil la -xulrunner192&project=mozilla&repository=openSUSE_11.1

The spec doesn't show any salience in this respect either.

Because everything is done automatically and therefore differently automatically between 11.3 and 11.1.

:) I see, updating to current and rearming gdb now.

Here we go: Program received signal SIGPIPE, Broken pipe. Thread 1 (Thread 0xb7c236d0 (LWP 18172)): #0 0xb6f37062 in nsCycleCollector::Suspect2 (this=0xb7a10000, n=0xa33eadf0) at /usr/src/debug/mozilla/xpcom/base/nsCycleCollector.cpp:2298 #1 0xb6f37117 in NS_CycleCollectorSuspect2_P (n=0xa33eadf0) at /usr/src/debug/mozilla/xpcom/base/nsCycleCollector.cpp:3115 #2 0xb67f745a in nsCycleCollectingAutoRefCnt::decr (this=0xa33eae28, owner=0xa33eadf0) at ../../dist/include/nsISupportsImpl.h:198 #3 0xb6a3edbd in nsGlobalWindow::Release (this=0xa33eadc0) at /usr/src/debug/mozilla/dom/base/nsGlobalWindow.cpp:1073 #4 0xb6ef5c39 in ~nsCOMPtr_base (this=<value optimized out>) at nsCOMPtr.cpp:81 #5 0xb6947f4c in nsEventDispatcher::Dispatch (aTarget=0xad42a200, aPresContext=0xa3ff0800, aEvent=0xbfff8738, aDOMEvent=0x0, aEventStatus=0xbfff8b3c, aCallback=0xbfff8650, aTargets=0x0) at ../../../dist/include/nsCOMPtr.h:865 #6 0xb67a0878 in PresShell::HandleEventInternal (this=0xa310f300, aEvent=0xbfff8738, aView=0x0, aStatus=0xbfff8b3c) at /usr/src/debug/mozilla/layout/base/nsPresShell.cpp:6525 #7 0xb67a0d90 in PresShell::HandleEventWithTarget (this=0xa310f300, aEvent=0xbfff8738, aFrame=0xa5fd4160, aContent=0xad42a200, aStatus=0xbfff8b3c) at /usr/src/debug/mozilla/layout/base/nsPresShell.cpp:6381 #8 0xb6933839 in nsEventStateManager::CheckForAndDispatchClick (this=0xb174f230, aPresContext=0xa3ff0800, aEvent=0xbfff8c80, aStatus=0xbfff8b3c) at /usr/src/debug/mozilla/content/events/src/nsEventStateManager.cpp:3994 #9 0xb6939963 in nsEventStateManager::PostHandleEvent (this=0xb174f230, aPresContext=0xa3ff0800, aEvent=0xbfff8c80, aTargetFrame=0xa5fd4160, aStatus=0xbfff8b3c, aView=0xa7ef1460) at /usr/src/debug/mozilla/content/events/src/nsEventStateManager.cpp:2947 #10 0xb67a08d7 in PresShell::HandleEventInternal (this=0xa310f300, aEvent=0xbfff8c80, aView=0xa7ef1460, aStatus=0xbfff8b3c) at /usr/src/debug/mozilla/layout/base/nsPresShell.cpp:6548 #11 0xb67a0e70 in PresShell::HandlePositionedEvent (this=0xa310f300, aView=0xa7ef1460, aTargetFrame=0xa5fd4160, aEvent=0xbfff8c80, aEventStatus=0xbfff8b3c) at /usr/src/debug/mozilla/layout/base/nsPresShell.cpp:6364 #12 0xb67a199f in PresShell::HandleEvent (this=0xa310f300, aView=0xa7ef1460, aEvent=0xbfff8c80, aEventStatus=0xbfff8b3c) at /usr/src/debug/mozilla/layout/base/nsPresShell.cpp:6228 #13 0xb6a1a8ef in nsViewManager::HandleEvent (this=0xa74ad460, aView=0xa7ef1460, aPoint= {x = -1073771552, y = -1073771392}, aEvent=0xbfff8c80, aCaptured=0) at /usr/src/debug/mozilla/view/src/nsViewManager.cpp:1224 #14 0xb6a1c955 in nsViewManager::DispatchEvent (this=0xa74ad460, aEvent=0xbfff8c80, aView=0xa7ef1460, aStatus=0xbfff8c2c) at /usr/src/debug/mozilla/view/src/nsViewManager.cpp:1203 #15 0xb6a17445 in HandleEvent (aEvent=0xbfff8c80) at /usr/src/debug/mozilla/view/src/nsView.cpp:167 #16 0xb6dc4e68 in nsWindow::DispatchEvent (this=0xa3f30580, aEvent=0xbfff8c80, aStatus=@0xbfff8ce4) at /usr/src/debug/mozilla/widget/src/gtk2/nsWindow.cpp:626 #17 0xb6dccf4a in nsWindow::OnButtonReleaseEvent (this=0xa3f30580, aWidget=0xa1ed3060, aEvent=0xa2cc9f60) at /usr/src/debug/mozilla/widget/src/gtk2/nsWindow.cpp:3227 #18 0xb6dccfcf in button_release_event_cb (widget=0xa1ed3060, event=0xa2cc9f60) at /usr/src/debug/mozilla/widget/src/gtk2/nsWindow.cpp:5981 #19 0xb5b494b6 in ?? () from /usr/lib/libgtk-x11-2.0.so.0 #20 0xb5e1afbb in g_closure_invoke () from /usr/lib/libgobject-2.0.so.0 #21 0xb5e312bf in signal_emit_unlocked_R () from /usr/lib/libgobject-2.0.so.0 #22 0xb5e327af in g_signal_emit_valist () from /usr/lib/libgobject-2.0.so.0 #23 0xb5e32d96 in g_signal_emit () from /usr/lib/libgobject-2.0.so.0 #24 0xb5c5e9fe in ?? () from /usr/lib/libgtk-x11-2.0.so.0 ---Type <return> to continue, or q <return> to quit--- #25 0xb5b41fcc in gtk_propagate_event () from /usr/lib/libgtk-x11-2.0.so.0 #26 0xb5b43377 in gtk_main_do_event () from /usr/lib/libgtk-x11-2.0.so.0 #27 0xb59036ca in ?? () from /usr/lib/libgdk-x11-2.0.so.0 #28 0xb6046998 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0 #29 0xb604a203 in g_main_context_iterate () from /usr/lib/libglib-2.0.so.0 #30 0xb604a388 in g_main_context_iteration () from /usr/lib/libglib-2.0.so.0 #31 0xb6dd1858 in nsAppShell::ProcessNextNativeEvent (this=0xb444c1a0, mayWait=1) at /usr/src/debug/mozilla/widget/src/gtk2/nsAppShell.cpp:147 #32 0xb6de77d0 in nsBaseAppShell::DoProcessNextNativeEvent (this=0xb444c1a0, mayWait=1) at /usr/src/debug/mozilla/widget/src/xpwidgets/nsBaseAppShell.cpp:155 #33 0xb6de7959 in nsBaseAppShell::OnProcessNextEvent (this=0xb444c1a0, thr=0xb7ab7150, mayWait=1, recursionDepth=0) at /usr/src/debug/mozilla/widget/src/xpwidgets/nsBaseAppShell.cpp:311 #34 0xb6f2be5e in nsThread::ProcessNextEvent (this=0xb7ab7150, mayWait=1, result=0xbfff92c4) at /usr/src/debug/mozilla/xpcom/threads/nsThread.cpp:508 #35 0xb6efa744 in NS_ProcessNextEvent_P (thread=0xb7c23b48, mayWait=1) at nsThreadUtils.cpp:250 #36 0xb6e6f82e in mozilla::ipc::MessagePump::Run (this=0xb7aaa6a0, aDelegate=0xb7a6d7c0) at /usr/src/debug/mozilla/ipc/glue/MessagePump.cpp:134 #37 0xb6ec80d4 in MessageLoop::RunInternal (this=0xb7a6d7c0) at /usr/src/debug/mozilla/ipc/chromium/src/base/message_loop.cc:216 #38 0xb6ec80f8 in MessageLoop::RunHandler (this=0xb7a6d7c0) at /usr/src/debug/mozilla/ipc/chromium/src/base/message_loop.cc:199 #39 0xb6ec816f in MessageLoop::Run (this=0xb7a6d7c0) at /usr/src/debug/mozilla/ipc/chromium/src/base/message_loop.cc:173 #40 0xb6de7ad8 in nsBaseAppShell::Run (this=0xb444c1a0) at /usr/src/debug/mozilla/widget/src/xpwidgets/nsBaseAppShell.cpp:174 #41 0xb6cac3d8 in nsAppStartup::Run (this=0xb2a542e0) at /usr/src/debug/mozilla/toolkit/components/startup/src/nsAppStartup.cpp:183 #42 0xb65ddca3 in XRE_main (argc=1, argv=0xbfffebd4, aAppData=0xb7a22980) at /usr/src/debug/mozilla/toolkit/xre/nsAppRunner.cpp:3483 #43 0x08049f6b in main (argc=1, argv=0xbfffebd4) at /usr/src/debug/mozilla/xulrunner/stub/nsXULStub.cpp:583 This was the only thread _not_ hanging in __kernel_vsyscall. This time, I'm not 100% confident, that it was triggered by the Build Service, since I've had another session open (bots, locally installed: http://bots.sourceforge.net/), that is running on top of a a django instance, and also using JS everywhere... I've used it, before the SIGPIPE arrived. Anyway, this shouldn't crash, and I've no idea, how nsCycleCollector::Suspect2 get up to trigger a SIGPIPE. Wolfgang, do you? TIA, Pete -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org