[JDEV] AIX 4.3.3 compilation of 1.4.1 server.

[David Clissold]

Yes, I've seen the same error, initially with jabber 1.4.0 and pth 1.3.7.
(Check the mail archives for mid-March; I posted info on what I was
seeing at the time, but there were no responses.)
I've tried with 1.4.1 as well, and again with the new version 1.4.0
of pth instead of the 1.3.7 that is provided along with jabber.
The dbx traceback is a little different with the new 1.4.0 libpth, but
failure occurs in the same place:

Floating point exception in swapcontext at 0xd0251348
0xd0251348 (swapcontext+0x2c) d9c90070       stfd   fr14,0x70(r9)
(dbx) t
swapcontext() at 0xd0251348
pth_yield() at 0xd0ffb330
pth_wait() at 0xd0ffd8a4
mtq_main(0x20104e98) at 0x10038d14
pth_spawn_trampoline() at 0xd0ffab94
_uc_trampoline() at 0xd0251208
(dbx)

And when I look at the registers, I see $fr14: 0x0000000000000000
though I don't know why that is a problem, unless 0x70(r9) is
a bad memory location.  ($r9:0x2010589c).

I am at something of a loss here.
The last log entry I get is different than yours, however:
mio.c:507 calling the connect handler for mio object 200F11C8
is *always* the last jabber log entry for me.  I've never
seen it after the "MIO is starting up" message like your entry.

Someone else had posted notes about the server dying after a SIGUSR2;
I did try their workarounds even though I never saw a SIGUSR2 -- so
not surprisingly, it didn't help (different problem, I'm sure).  I would
love to get past this error, and am willing to experiment with it if you
or anyone have any ideas.

(I was able to debug it a little easier via dbx by switching from gcc
to the VAC c compiler, though I had to do a little source patching.  The
error happens with either compiler.  But it goes through the same pth code
many times before it actually happens -- it isn't consistent).
----
David Clissold
cliss at austin.ibm.com


>When I try to run the 1.4.1 jabber server on AIX 4.3.3, I get the
>following message:
>
>$ ./jabberd/jabberd -D
>Wed May 23 10:09:14 2001  mio.c:568 MIO is starting up
>Signal 65535
>$
>
>I looked at the core with dbx and get a whole page. I think the line that
>sticks out the most is this:
>
>Floating point exception in swapcontext at 0xd0262918
>0xd0262918 (swapcontext+0x2c) d9c90070       stfd   fr14,0x70(r9)
>
>Has anyone seen this before and came up with a solution? 
>
>Thanks!
>
>Jonathan Siegle                 Center for Academic Computing
>jsiegle at psu.edu                 Penn State University
>814-865-5840                    University Park, Pa 16802
>