[Openmcl-devel] Segmentation fault on linuxppc

[Gary Byers]

On Tue, 10 Aug 2004, Jan Idzikowski wrote:

>
> hallo gary,
>
> i have also tested it with kernel 2.6.5 on the same hardware:
> Linux hombre 2.6.5 #1 Sun Aug 1 21:59:48 CEST 2004 ppc 604r CHRP IBM,7046-B50 GNU/Linux
>
> and openmcl is working.
> hombre ccl # ./ppccl
> Welcome to OpenMCL Version (Beta: Linux) 0.14.2-p1!
> ?
>
> i also try sbcl on my 2.6.7 linux ppc and it works.
>
> i hope you can find the differences between 2.6.5 and 2.6.7.
>
>
>

Well, I'm not sure what the difference is between your 2.6.7 and the one
that I just built, but:

[ccl-0.14/ccl] gb at z2> uname -a
Linux z2 2.6.7-gentoo-r13 #1 Tue Aug 10 17:14:03 MDT 2004 ppc 7455, altivec supported PowerBook6,3 GNU/Linux
[ccl-0.14/ccl] gb at z2> ./ppccl
Welcome to OpenMCL Version (Beta: Linux) 0.14.2-p1!
?

If you could try running this under GDB, that might tell us something.

First (to make things easier), ensure that ccl/.gdbinit is a link to
ccl/lisp-kernel/linux/.gdbinit:

shell> cd ccl
shell> ln -s lisp-kernel/linux/.gdbinit .gdbinit

Then:

shell> gdb ./ppccl
[gdb messages]

(gdb) break start_lisp
(gdb) run

If that gets to the breakpoint at 'start_lisp', that suggests that the
segfault you're getting happens after lisp kernel initialization.

If you get a segfault before reaching that breakpoint, gdb would probably
be able to say where it's happening.  Since it doesn't seem that I can
reproduce this, that'd be very helpful to know.