Bug 1685276 - thunderbird dumps core in versino 60.5.1-1
Summary: thunderbird dumps core in versino 60.5.1-1
Keywords:
Status: CLOSED DUPLICATE of bug 1709601
Alias: None
Product: Fedora
Classification: Fedora
Component: thunderbird
Version: 29
Hardware: i686
OS: Linux
unspecified
medium
Target Milestone: ---
Assignee: Martin Stransky
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2019-03-04 20:54 UTC by Albert Flügel
Modified: 2019-05-15 09:33 UTC (History)
11 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2019-05-15 09:33:04 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
desired debugging output with the arguments -ProfileManager -safe-mode -g (12.13 KB, text/plain)
2019-03-05 20:28 UTC, Albert Flügel
no flags Details
desired debug output with only the argument -g and the "StartupMaster" enabled (22.08 KB, text/plain)
2019-03-05 20:29 UTC, Albert Flügel
no flags Details

Description Albert Flügel 2019-03-04 20:54:33 UTC
Description of problem:
After upgrade of thunderbird from 60.4.0-1 to 60.5.1-1 it dumps core when started. Downgrade makes it work again

Version-Release number of selected component (if applicable):
60.5.1-1

How reproducible:

Steps to Reproduce:
1. Upgrade thunderbird to 60.5.1-1
2. start thunderbird
3. :-(

Actual results:
No thunderbird window shows. When starting from commandline one can see the message, that a core has been dumped. I don't find a core file, though coredumpsize is set to unlimited.

Expected results:
thunderbird starts, window becomes visible

Additional info:
the following plugins are active (don't know if that matters):
* enigmail
* StartupMaster
* Provider for Google Calendar
* Manually sort folders
* Dorando keyconfig
Version 60.4.0-1 works normally

any hint how to tighten the search for the source of the problem is appreciated.

Comment 1 Martin Stransky 2019-03-05 12:45:16 UTC
Can you please try:

1) run thunderbird in safe mode ($thunderbird -safe-mode) and disable the addons etc.
2) create a new profile and try to run TB with it ($thunderbird -ProfileManager)
3) catch the backtrace and attach it here - how it is at https://fedoraproject.org/wiki/Debugging_guidelines_for_Mozilla_products#Application_crash

Thanks!

Comment 2 Albert Flügel 2019-03-05 20:27:23 UTC
So below follows the result. Some findings in advance:
The culprit to trigger the problem seems to be the add-on "StartupMaster". When i switch it off, thunderbird starts normally. (do i still need this add-on at all ? It seems, thunderbird asks for the master vault password in the beginning anyway).
However, thunderbird also crashes when i start it this way:
thunderbird -ProfileManager -safe-mode -g -d gdb
So this way of debugging could not work, especially working with a different profile, thus i started gdb /usr/lib/thunderbird/thunderbird and then:
(gdb) run -ProfileManager -safe-mode -g
...
Using host libthread_db library "/lib/libthread_db.so.1".
warning: Error reading shared library list entry at 0xfffff800
warning: Error reading shared library list entry at 0xfffffc00
warning: Error reading shared library list entry at 0x1000
Detaching after fork from child process 11549.
warning: Error reading shared library list entry at 0x1400
warning: Error reading shared library list entry at 0x1800
warning: Error reading shared library list entry at 0x1c00
warning: Error reading shared library list entry at 0x2800
warning: Error reading shared library list entry at 0x2c00
[New Thread 0xae8bab40 (LWP 11550)]
[Thread 0xae8bab40 (LWP 11550) exited]
warning: Error reading shared library list entry at 0xffffc400
...
[Thread 0xa50fcb40 (LWP 11565) exited]
[Thread 0xa58fdb40 (LWP 11564) exited]
[Thread 0xa60feb40 (LWP 11563) exited]

Thread 1 "thunderbird" received signal SIGSEGV, Segmentation fault.
0xb01ba9a1 in gfxPlatform::UsesOffMainThreadCompositing ()
    at /usr/src/debug/thunderbird-60.5.1-1.fc28.i386/objdir/dist/include/mozilla/gfx/gfxVars.h:141
141       GFX_VARS_LIST(GFX_VAR_DECL)
Missing separate debuginfos, use: dnf debuginfo-install at-spi2-atk-2.26.2-1.fc28.i686 at-spi2-core-2.28.0-1.fc28.i686 atk-2.28.1-1.fc28.i686 bzip2-libs-1.0.6-26.fc28.i686 cairo-1.15.12-2.fc28.i686 cairo-gobject-1.15.12-2.fc28.i686 dbus-glib-0.110-2.fc28.i686 dbus-libs-1.12.10-1.fc28.i686 expat-2.2.5-3.fc28.i686 fontconfig-2.13.0-5.fc28.i686 freetype-2.8-10.fc28.i686 fribidi-1.0.2-1.fc28.i686 gdk-pixbuf2-2.36.12-1.fc28.i686 glib2-2.56.4-1.fc28.i686 graphite2-1.3.10-5.fc28.i686 gtk3-3.22.30-1.fc28.i686 gvfs-client-1.36.2-3.fc28.i686 harfbuzz-1.7.5-3.fc28.i686 libICE-1.0.9-12.fc28.i686 libSM-1.2.2-8.fc28.i686 libX11-1.6.5-7.fc28.i686 libX11-xcb-1.6.5-7.fc28.i686 libXau-1.0.8-11.fc28.i686 libXcomposite-0.4.4-12.fc28.i686 libXcursor-1.1.15-1.fc28.i686 libXdamage-1.1.4-12.fc28.i686 libXext-1.3.3-8.fc28.i686 libXfixes-5.0.3-5.fc28.i686 libXi-1.7.9-6.fc28.i686 libXinerama-1.1.3-10.fc28.i686 libXrandr-1.5.1-5.fc28.i686 libXrender-0.9.10-5.fc28.i686 libXt-1.1.5-7.fc28.i686 libblkid-2.32.1-1.fc28.i686 libcap-2.25-9.fc28.i686 libdatrie-0.2.9-7.fc28.i686 libepoxy-1.5.2-1.fc28.i686 libffi-3.1-16.fc28.i686 libgcc-8.3.1-2.fc28.i686 libgcrypt-1.8.4-1.fc28.i686 libgpg-error-1.33-1.fc28.i686 libjpeg-turbo-1.5.3-6.fc28.i686 libmount-2.32.1-1.fc28.i686 libpng-1.6.34-6.fc28.i686 libselinux-2.8-1.fc28.i686 libstdc++-8.3.1-2.fc28.i686 libthai-0.1.28-1.fc28.i686 libuuid-2.32.1-1.fc28.i686 libvpx-1.7.0-5.fc28.i686 libwayland-client-1.15.0-1.fc28.i686 libwayland-cursor-1.15.0-1.fc28.i686 libwayland-egl-1.15.0-1.fc28.i686 libxcb-1.13-1.fc28.i686 libxkbcommon-0.8.2-1.fc28.i686 lz4-libs-1.8.1.2-4.fc28.i686 nspr-4.20.0-1.fc28.i686 nss-3.42.1-1.fc28.i686 nss-util-3.42.1-1.fc28.i686 pango-1.42.4-2.fc28.i686 pcre-8.43-1.fc28.i686 pcre2-10.32-8.fc28.i686 pixman-0.34.0-8.fc28.i686 startup-notification-0.12-15.fc28.i686 systemd-libs-238-12.git07f8cd5.fc28.i686 xcb-util-0.4.0-9.fc28.i686 xz-libs-5.2.4-2.fc28.i686 zlib-1.2.11-8.fc28.i686

... entering thread apply ...

(gdb) print DumpJSStack()
there is no JSContext on the stack!
[Thread 0xa68ffb40 (LWP 12488) exited]
$1 = void

This way i obtained the file crash_bt-ProfileManager-safe-mode-g.xz , that i attach.

Then i ran thunderbird in the debugger differently:
same start of debugger, then:
(gdb) run -g
... same SEGV:
Thread 1 "thunderbird" received signal SIGSEGV, Segmentation fault.
0xb01ba9a1 in gfxPlatform::UsesOffMainThreadCompositing ()
    at /usr/src/debug/thunderbird-60.5.1-1.fc28.i386/objdir/dist/include/mozilla/gfx/gfxVars.h:141
141       GFX_VARS_LIST(GFX_VAR_DECL)
...
same stuff like above, then i do the thread apply ..., file attached
...
(gdb) print DumpJSStack()
0 openModalWindow(domWin = null, uri = "chrome://global/content/commonDialog.xul", args = [xpconnect wrapped (nsISupports, nsIWritablePropertyBag2, nsIWritablePropertyBag)]) ["jar:file:///usr/lib/thunderbird/omni.ja!/components/nsPrompter.js":353]
1 openPrompt(args = [object Object]) ["jar:file:///usr/lib/thunderbird/omni.ja!/components/nsPrompter.js":563]
    this = [object Object]
2 nsIPrompt_promptPassword(title = "Password Required", text = "Please enter your master password.", pass = [object XPCOutParam], checkLabel = null, checkValue = [object XPCOutParam]) ["jar:file:///usr/lib/thunderbird/omni.ja!/components/nsPrompter.js":781]
    this = [object Object]
3 promptPassword(null, "Please enter your master password.", [object XPCOutParam], null, [object XPCOutParam]) ["jar:file:///usr/lib/thunderbird/omni.ja!/components/nsPrompter.js":594]
    this = [object Object]
4 pk11login() ["chrome://startupmaster/content/startupmaster.jsm":42]
    this = [object Object]
5 observe(aSubject = null, aTopic = "profile-after-change", aData = "startup") ["chrome://startupmaster/content/startupmaster.jsm":85]
    this = [object Object]

[Thread 0xa01feb40 (LWP 12701) exited]
$1 = void

From this i obtained the file crash_bt-g.xz , that is also attached. This one is much longer, but this seems coincidence to me. In a previous attempt of the first way to start thunderbird i also got a log of about the same size, so it seems to me, it happens randomly and is matter of race condition, if some more threads are already started (especially java related stuff) and also part of the dump.

hope this helps.

Comment 3 Albert Flügel 2019-03-05 20:28:30 UTC
Created attachment 1541144 [details]
desired debugging output with the arguments -ProfileManager -safe-mode -g

Comment 4 Albert Flügel 2019-03-05 20:29:29 UTC
Created attachment 1541145 [details]
desired debug output with only the argument -g and the "StartupMaster" enabled

Comment 5 Albert Flügel 2019-03-05 20:32:49 UTC
The debugging guidelines need to be amended. Doing "debuginfo-install ..." does not work anymore. For a dnf maintained system one has to do: "dnf debuginfo-install ...". Without "dnf" yum is called, who complains something like:
Error: Invalid version flag: or

Comment 6 Martin Stransky 2019-03-06 09:56:00 UTC
This looks like a duplicate of Bug 1677225.

Comment 7 Martin Stransky 2019-03-06 10:05:04 UTC
So you say that the add-on "StartupMaster" do cause the issue, right?

Comment 8 Albert Flügel 2019-03-06 16:14:12 UTC
Basically yes. This add-on triggers the problem. Whether it really causes it, i would not judge. The descriptions in the referred bug imply, that there is a general issue with several instances running in parallel.

Aside of this, the SEGV also happens, when StartupMaster is disabled and thunderbird is started with -ProfileManager -safe-mode
Should i check, whether the problem with these arguments still shows up with StartupMaster completely removed ?

Comment 9 Martin Stransky 2019-03-06 16:17:21 UTC
(In reply to Albert Flügel from comment #8)
> Basically yes. This add-on triggers the problem. Whether it really causes
> it, i would not judge. The descriptions in the referred bug imply, that
> there is a general issue with several instances running in parallel.
> 
> Aside of this, the SEGV also happens, when StartupMaster is disabled and
> thunderbird is started with -ProfileManager -safe-mode
> Should i check, whether the problem with these arguments still shows up with
> StartupMaster completely removed ?

Yes, please.

Comment 10 Albert Flügel 2019-03-06 21:09:30 UTC
Also with StartupMaster removed, thunderbird segfaults with the same line indicated in gdb, when run with -ProfileManager . It does not matter, if -safe-mode is given, too, or not.
Could it be, -ProfileManager initiates an additional process like in the bug you mentioned in comment #6 ...

Comment 11 Ben Cotton 2019-05-02 19:46:09 UTC
This message is a reminder that Fedora 28 is nearing its end of life.
On 2019-May-28 Fedora will stop maintaining and issuing updates for
Fedora 28. It is Fedora's policy to close all bug reports from releases
that are no longer maintained. At that time this bug will be closed as
EOL if it remains open with a Fedora 'version' of '28'.

Package Maintainer: If you wish for this bug to remain open because you
plan to fix it in a currently maintained version, simply change the 'version' 
to a later Fedora version.

Thank you for reporting this issue and we are sorry that we were not 
able to fix it before Fedora 28 is end of life. If you would still like 
to see this bug fixed and are able to reproduce it against a later version 
of Fedora, you are encouraged  change the 'version' to a later Fedora 
version prior this bug is closed as described in the policy above.

Although we aim to fix as many bugs as possible during every release's 
lifetime, sometimes those efforts are overtaken by events. Often a 
more recent Fedora release includes newer upstream software that fixes 
bugs or makes them obsolete.

Comment 12 andreas.stoeckel 2019-05-10 21:27:38 UTC
Still a problem here after upgrading to Fedora 30. Running

thunderbird -ProfileManager -safe-mode

results in "Segmentation fault (core dumped)". "thunderbird-wayland" works fine.

Comment 13 Albert Flügel 2019-05-12 08:31:37 UTC
Confirm comment #12 and assume it's the same on 29
... and it happens as well on x86_64 and amd64

Comment 14 Martin Stransky 2019-05-15 09:31:45 UTC
New TB builds with the crashes fixed are in koji now - thunderbird-60.6.1-5. (https://koji.fedoraproject.org/koji/packageinfo?packageID=39). Thanks for your patience.

Comment 15 Martin Stransky 2019-05-15 09:33:04 UTC

*** This bug has been marked as a duplicate of bug 1709601 ***


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