GUI failure (SHM?)

[marmarek]

Reported by joanna on 9 May 2010 09:53 UTC
Got a flash crash in AppVM (dmesg output):

npviewer.bin[segfault at 8 ip 00007f7259147a9b sp 00007fff1793b6f8 error 4 in libflashplayer.so[7f7258b15000+8d2000](1073]:)

And then the GUI for this particular AppVM has frozen -- i.e. the !AppViewer windows are still present, but are not operational, not only cannot operate their contents, but also cannot close them.

Both guid in Dom0, and gui_agent in the AppVM are still running:

• GUI daemon:

(gdb) bt
#0  0x00007fce51c2c3c3 in __select_nocancel () from /lib64/libc.so.6
#1  0x0000000000405fb3 in wait_for_vchan_or_argfd_once ()
#2  0x000000000040616c in wait_for_vchan_or_argfd ()
#3  0x00000000004055d8 in main ()

• GUI agent:

(gdb) bt
#0  0x00007f07fafab570 in __read_nocancel () from /lib64/libc.so.6
#1  0x00000000004022d6 in send_pixmap_mfns ()
#2  0x00000000004026e6 in process_xevent_configure ()
#3  0x0000000000402bbf in process_xevent ()
#4  0x00000000004039b4 in main ()

The GUI daemon keeps outputing the following into the logs:

ErrorHandler: BadValue (integer parameter out of range for operation)
                 Major opcode: 139 (MIT-SHM)
                 Minor opcode: 3 (X_ShmPutImage)
                 Value:        0x4fa
                 Failed serial number:  103729
                 Current serial number: 103730

even after the freeze, i.e. even now when the !AppViewer windows are not operational, still when I try to e.g. resize them, or just focus on them, I get the above messages in the Dom0 log.

No idea how to reproduce it: got it for the first time after several weeks of flawless operation.

Migrated-From: https://wiki.qubes-os.org/ticket/29

[marmarek]

Comment by joanna on 30 May 2010 13:51 UTC
Cannot reproduce it reliably, moving to Beta 1 milestone.

[marmarek]

Comment by joanna on 30 May 2010 13:51 UTC
Cannot reproduce it reliably, moving to Beta 1 milestone.

[marmarek]

Comment by joanna on 16 Jun 2010 13:08 UTC
Perhaps this was just some internal VM/firefox problem? Closing this ticket for now. If it still occurs with the new F13-based templates, we will reopen it then.

[marmarek]

Comment by joanna on 16 Jun 2010 13:08 UTC
Perhaps this was just some internal VM/firefox problem? Closing this ticket for now. If it still occurs with the new F13-based templates, we will reopen it then.

[marmarek]

Modified by joanna on 16 Sep 2010 14:35 UTC

[marmarek]

Modified by joanna on 16 Sep 2010 14:35 UTC