BTW trying with https://www.google.com/finance?q=nyse%3Arht, there is definitely some segfault in plugin-container:

Mar 31 07:07:30 unused-4-232.brq.redhat.com kernel: plugin-containe[22980]: segfault at 0 ip 00007f0b14ae3c33 sp 00007ffdd13c06a8 error 6 in libmozgtk.so[7f0b14ae2000+3000]

Can you please test X session?

I have tested on X session with SELinux set to permissive. Fedora 24.

Mar 31 00:36:02 rawhide kernel: plugin-containe[9684]: segfault at 0 ip 00007f1b840b2c33 sp 00007ffecd35a0a8 error 6 in libmozgtk.so[7f1b840b1000+3000]
Mar 31 00:36:02 rawhide abrt-hook-ccpp[9687]: Process 9684 (plugin-container) of user 1000 killed by SIGSEGV - ignoring (repeated crash)
Mar 31 00:36:02 rawhide firefox.desktop[9420]: [9420] WARNING: pipe error (85): Connection reset by peer: file /builddir/build/BUILD/firefox-45.0.1/firefox-45.0.1/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 459
Mar 31 00:36:02 rawhide audit[9688]: ANOM_ABEND auid=1000 uid=1000 gid=1000 ses=4 subj=unconfined_u:unconfined_r:mozilla_plugin_t:s0-s0:c0.c1023 pid=9688 comm="plugin-containe" exe="/usr/lib64/firefox/plugin-container" sig=11
Mar 31 00:36:02 rawhide kernel: plugin-containe[9688]: segfault at 0 ip 00007f75ea409c33 sp 00007fff2719ea98 error 6 in libmozgtk.so[7f75ea408000+3000]
Mar 31 00:36:02 rawhide abrt-hook-ccpp[9691]: Process 9688 (plugin-container) of user 1000 killed by SIGSEGV - ignoring (repeated crash)

Hm, I don't see that, flash works fine for me, SELinux enabled.

I got the same problem(?) on X. When running firefox from terminal there are a lot of messages like that:

[23060] WARNING: pipe error (84): Connection reset by peer: file /builddir/build/BUILD/firefox-45.0.1/firefox-45.0.1/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 459
[23060] WARNING: waitpid failed pid:23211 errno:10: file /builddir/build/BUILD/firefox-45.0.1/firefox-45.0.1/ipc/chromium/src/base/process_util_posix.cc, line 267

(In reply to Heiko Adams from comment #5)
> I got the same problem(?) on X. When running firefox from terminal there are
> a lot of messages like that:
> 
> [23060] WARNING: pipe error (84): Connection reset by peer: file
> /builddir/build/BUILD/firefox-45.0.1/firefox-45.0.1/ipc/chromium/src/chrome/
> common/ipc_channel_posix.cc, line 459
> [23060] WARNING: waitpid failed pid:23211 errno:10: file
> /builddir/build/BUILD/firefox-45.0.1/firefox-45.0.1/ipc/chromium/src/base/
> process_util_posix.cc, line 267

That just means that the plugin crashed inside plugin-container and firefox can't connect to it.

I'm also running into this same segfault on Fedora 24/firefox-45.0.1-4.fc24. In my situation it is with the Citrix ICA plugin instead of Flash:

apr 04 10:50:43 foo.bar audit[3180]: ANOM_ABEND auid=1000 uid=1000 gid=1000 ses=2 subj=unconfined_u:unconfined_r:mozilla_plugin_t:s0-s0:c0.c1023 pid=3180 comm="plugin-containe" exe="/usr/lib64/firefox/plugin-container" sig=11
apr 04 10:50:43 foo.bar kernel: plugin-containe[3180]: segfault at 0 ip 00007fc2cdc88c33 sp 00007ffcb26a0d48 error 6 in libmozgtk.so[7fc2cdc87000+3000]
apr 04 10:50:43 foo.bar abrt-hook-ccpp[3183]: Process 3180 (plugin-container) of user 1000 killed by SIGSEGV - dumping core
apr 04 10:50:44 foo.bar firefox.desktop[3089]: [3089] WARNING: pipe error (55): Verbinding is weggevallen: file /builddir/build/BUILD/firefox-45.0.1/firefox-45.0.1/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 459

Setting SELinux to permissive does not resolve the issue. It also doesn't matter if I'm running from a Wayland or a X11 session

I managed to workaround it by downgrading to firefox-44.0.2-3.fc24

Can you please test upstream Firefox from mozilla.com? Thanks.

With upstream Firefox 45.0.1 the Citrix ICA plugin works fine without segfaults

Please note that I am experiencing it only with firefox-45.0.1-4.fc24.x86_64.rpm. Currently I am using firefox-45.0-4.fc24.x86_64 with which flash works fine.

I've also tried upstream 45.0.1, found no problem with flash and no error messages at terminal regarding plugin-container.

Hi,

I'm seeing this too, like the other I too am getting segfaults in the plugin-container, dmesg says the segfaults are in libmozgtk.so, which does suggest an issue with firefox.

Parallel installing libicu and libvpx from F23 (using --replacefiles to work around conflicts under /usr/share/doc) and then downgrading to the F23 firefox build fixes this. So chances are this is an issue caused by building firefox with gcc6 for F24.

Note: all requested info has been provided in earlier comments -> clearing needinfo.

Regards,

Hans

This is backtrace I was able to get using GDB:


(gdb) set follow-fork-mode child
(gdb) c
Continuing.
[Thread 0x7fb86e5ef700 (LWP 5071) exited]
[New Thread 0x7fb86e5ef700 (LWP 5113)]
[New process 5114]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
process 5114 is executing new program: /usr/lib64/firefox/plugin-container
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
[New Thread 0x7f0d287ff700 (LWP 5116)]
[Thread 0x7f0d287ff700 (LWP 5116) exited]
[New Thread 0x7f0d287ff700 (LWP 5117)]

Thread 2.1 "plugin-containe" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7f0d3cae2ac0 (LWP 5114)]
0x00007f0d36d4ec33 in gtk_button_get_type () from /usr/lib64/firefox/gtk2/libmozgtk.so
Missing separate debuginfos, use: dnf debuginfo-install firefox-45.0.1-4.fc25.x86_64
(gdb) bt
#0  0x00007f0d36d4ec33 in gtk_button_get_type () at /usr/lib64/firefox/gtk2/libmozgtk.so
#1  0x00007f0d25eebfbb in gtk_module_init () at /usr/lib64/gtk-2.0/modules/libcanberra-gtk-module.so
#2  0x00007f0d2f415650 in default_display_notify_cb () at /lib64/libgtk-x11-2.0.so.0
#3  0x00007f0d319803e8 in g_closure_invoke () at /lib64/libgobject-2.0.so.0
#4  0x00007f0d31992442 in signal_emit_unlocked_R () at /lib64/libgobject-2.0.so.0
#5  0x00007f0d3199b08c in g_signal_emit_valist () at /lib64/libgobject-2.0.so.0
#6  0x00007f0d3199b45f in g_signal_emit () at /lib64/libgobject-2.0.so.0
#7  0x00007f0d319847b4 in g_object_dispatch_properties_changed () at /lib64/libgobject-2.0.so.0
#8  0x00007f0d31986cc4 in g_object_notify () at /lib64/libgobject-2.0.so.0
#9  0x00007f0d2f013b0a in gdk_display_open_default_libgtk_only () at /lib64/libgdk-x11-2.0.so.0
#10 0x00007f0d2f3f8a65 in gtk_init () at /lib64/libgtk-x11-2.0.so.0
#11 0x00007f0d3a27a61e in mozilla::plugins::PluginModuleChild::InitGraphics() () at /usr/lib64/firefox/libxul.so
#12 0x00007f0d3a27add5 in mozilla::plugins::PluginModuleChild::InitForChrome(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, int, MessageLoop*, IPC::Channel*) ()
    at /usr/lib64/firefox/libxul.so
#13 0x00007f0d3a281ff1 in mozilla::plugins::PluginProcessChild::Init() () at /usr/lib64/firefox/libxul.so
#14 0x00007f0d3aa43a9b in XRE_InitChildProcess () at /usr/lib64/firefox/libxul.so
#15 0x00005601f46c0795 in content_process_main(int, char**) ()
#16 0x00007f0d371fe0c1 in __libc_start_main () at /lib64/libc.so.6
#17 0x00005601f46c0379 in _start ()
(gdb)

I am glad that I am not the one who is facing this issue. I am still not able to play flash widgets. I am using firefox-45.0.1-4.fc24 build. Previously I thought its flash issue but after few times removal/reinstall still firefox can't show flash widgets.

Thanks, the backtrace is very useful. Can you please set "GTK_DEBUG=all", run it in gdb and attach console output? I expect some gtk warnings. 

You can also run FF with "--g-fatal-warnings" param which stops gdb when some gtk critical event happens. 

Thanks.

btw. the /usr/lib64/gtk-2.0/modules/libcanberra-gtk-module.so seems to be a part of MATE DE. Do you run MATE or some different DE? Like gnome-shell, Cinamon, KDE?

(In reply to Martin Stransky from comment #15)
> btw. the /usr/lib64/gtk-2.0/modules/libcanberra-gtk-module.so seems to be a
> part of MATE DE. Do you run MATE or some different DE? Like gnome-shell,
> Cinamon, KDE?

Negative, neither I have ever tried them. I am using just plain Gnome 3.20.

(In reply to Vít Ondruch from comment #16)
> (In reply to Martin Stransky from comment #15)
> > btw. the /usr/lib64/gtk-2.0/modules/libcanberra-gtk-module.so seems to be a
> > part of MATE DE. Do you run MATE or some different DE? Like gnome-shell,
> > Cinamon, KDE?
> 
> Negative, neither I have ever tried them. I am using just plain Gnome 3.20.

This seems to be required by xchat-gnome I am using:



$ sudo dnf remove `rpm -qf /usr/lib64/gtk-2.0/modules/libcanberra-gtk-module.so`
Dependencies resolved.
=====================================================================================
 Package              Arch       Version                          Repository    Size
=====================================================================================
Removing:
 libcanberra-gtk2     x86_64     0.30-11.fc24                     @rawhide      46 k
 xchat-gnome          x86_64     0.26.2-22.git40c5bf988.fc25      @rawhide     6.2 M

Transaction Summary
=====================================================================================
Remove  2 Packages

Installed size: 6.3 M
Is this ok [y/N]:

Well, yes, it's really a problem with the mate component.

Added to firefox-45.0.1-6.fc24, please retest when the builds are done in koji.

Wow it was simple fix for me on Fedora 24 Workstation. I just removed
$ rpm -qf /usr/lib64/gtk-2.0/modules/libcanberra-gtk-module.so
libcanberra-gtk2-0.30-11.fc24.x86_64
$ sudo dnf remove libcanberra-gtk2

And flash widgets started working fine. Thanks Martin and Vit for your inputs. 

But I don't understand howcome firefox flash widgets are affected by this libcanberra-gtk-module.so ?

I can also confirm that removing libcanberra-gtk2-0.30-11.fc24.x86_64 package makes firefox-45.0.1-4.fc24.x86_64.rpm work. It is strange however because it is not required by any other package and it seems it was installed with the base desktop system.

By looking at history, it was installed for me with "Fedora Workstation" group this was the command line for dnf.

Command Line   : --disablerepo=* --enablerepo=rawhide --nogpgcheck --installroot=/mnt groupinstall Fedora Workstation
--disablerepo=* --enablerepo=rawhide --nogpgcheck --installroot=/mnt groupinstall Fedora Workstation

(In reply to Martin Stransky from comment #19)
> Added to firefox-45.0.1-6.fc24, please retest when the builds are done in
> koji.

This fixes the issue for me, although I have two cursors shown above the Google Finance graph. Will open separate issue for that. Thx for the fix.

firefox-45.0.1-6.fc24 has been submitted as an update to Fedora 24. https://bodhi.fedoraproject.org/updates/FEDORA-2016-4c2341fb85

firefox-45.0.1-6.fc24 has been pushed to the Fedora 24 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2016-4c2341fb85

(In reply to Parag Nemade from comment #20)
> And flash widgets started working fine. Thanks Martin and Vit for your
> inputs. 
> 
> But I don't understand howcome firefox flash widgets are affected by this
> libcanberra-gtk-module.so ?

It's because flash search missing gtk2 symbols in libcanberra-gtk-module.so which is obviously wrong. That was caused by wrong symbols in the mozillagtk2 module.

firefox-45.0.1-6.fc24 has been pushed to the Fedora 24 stable repository. If problems still persist, please make note of it in this bug report.