Bug#922097: inkscape stops with message complaining about an internal illegal instruction on start

[Bernhard Übelacker]

Hello Mattia Rizzolo,
I am just looking at some random crashes, and
tried to detect at which exact instruction
in libgtkmm-2.4.so.1 we are in the given backtrace.

That function contains just two instructions.
But the backtrace from the core dump tells we crashed
at ...459, which is in the middle of the nopl instruction ...

Therefore gdb may not show the line information I guess.

Maybe you can please send the output of the disassemble
and x command from gdb having loaded the core dump?

Kind regards,
Bernhard


(gdb) disassemble $pc-30,$pc+20
Dump of assembler code from 0x7ffff525b43b to 0x7ffff525b46d:
...
   0x00007ffff525b447:  nopw   0x0(%rax,%rax,1)
   0x00007ffff525b450 <Gtk::RadioButtonGroup::RadioButtonGroup()+0>:    movq   $0x0,(%rdi)
   0x00007ffff525b457 <Gtk::RadioButtonGroup::RadioButtonGroup()+7>:    retq   
   0x00007ffff525b458:  nopl   0x0(%rax,%rax,1)
   0x00007ffff525b460 <Gtk::RadioButtonGroup::RadioButtonGroup(_GSList*)+0>:    mov    %rsi,(%rdi)
...
End of assembler dump.

(gdb) x/64xb _ZN3Gtk16RadioButtonGroupC1Ev-32
0x7ffff525b430: 0x89    0xc4    0xe9    0x21    0x94    0xf4    0xff    0x49
0x7ffff525b438: 0x89    0xc4    0xe9    0x31    0x94    0xf4    0xff    0x49
0x7ffff525b440: 0x89    0xc4    0xe9    0x19    0x94    0xf4    0xff    0x66
0x7ffff525b448: 0x0f    0x1f    0x84    0x00    0x00    0x00    0x00    0x00
0x7ffff525b450: 0x48    0xc7    0x07    0x00    0x00    0x00    0x00    0xc3
0x7ffff525b458: 0x0f    0x1f    0x84    0x00    0x00    0x00    0x00    0x00
0x7ffff525b460: 0x48    0x89    0x37    0xc3    0x66    0x66    0x2e    0x0f
0x7ffff525b468: 0x1f    0x84    0x00    0x00    0x00    0x00    0x00    0x90