[panda-users] core dump at the end of replay

[Manolis Stamatogiannakis]

Hello,

I'm experimenting with the osi/osi_linux plugins.

I have added some code in the osi plugin which implements some
functionality I need. free_osiprocs() is called several times during replay
without problems. Strangely, I get a segfault when it is called for the
last time, following the end of replay (i.e. after the "Replay completed
successfully." message).

The gdb stack trace looks like this:

Core was generated by `./i386-softmmu/qemu-system-i386 -vnc :1 -panda
osi;osi_linux -replay vlassis'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x00007f6ce1d7f550 in ?? ()
(gdb) where
#0  0x00007f6ce1d7f550 in ?? ()
#1  0x00007f6ce1f87366 in free_osiprocs (ps=0x2ddd9d0) at os_intro.c:95
#2  0x00007f6ce7070f4f in __cxa_finalize (d=0x7f6ce218a540) at
cxa_finalize.c:56
#3  0x00007f6ce1f86793 in __do_global_dtors_aux ()
   from /home/mstamat/panda/qemu/i386-softmmu/panda_plugins/panda_osi.so
#4  0x00007fff109403a0 in ?? ()
#5  0x00007f6ceb72b7ba in _dl_close_worker (map=<optimized out>) at
dl-close.c:272
Backtrace stopped: frame did not save the PC

I'm not sure why this happens. My guess is that at the time of invocation
of free_osiprocs(), osi_linux has already been "unloaded", so an error
occurs.

Any hints of what actually happens and what workarounds could be used?

Thanks,
Manolis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/panda-users/attachments/20150215/3aebbe69/attachment.htm