Subject: Re: i386 1.4Q hangs nonrandomly?
To: None <email@example.com>
From: maximum entropy <firstname.lastname@example.org>
Date: 01/25/2000 06:17:18
>Date: Mon, 24 Jan 2000 12:23:59 -0500
>From: Mason Loring Bliss <email@example.com>
>In the past couple days I've also seen some stuff like, for instance, gcc
>dying with a "signal 10." I haven't seen this very recently, though - I
>believe I'm running a new kernel since the last time I saw this.
I've had some similar mysterious program crashes.
I spent some time pondering what has changed recently on my system,
and I came up with some interesting thoughts. They may be stupid
thoughts, but I think they are interesting. Please tell me if you
think I'm totally nuts.
The recent changes to ld.elf_so enable dynamic loading of different
shared libraries based on kernel sysconf variables.
The file /usr/src/etc/etc.i386/ld.so.conf contains:
libm.so.0 machdep.fpu_present 1:libm387.so.0,libm.so.0
In other words, with these changes, if an fpu is present, libm387.so.0
is loaded, whereas before, it wasn't.
As far as I can remember, I began experiencing these weird crashes
around the same time as I installed that file into /etc. But I got
distracted by the kern_proc.c bug, and thought that was the source of
all my ills.
Is it possible that this heavier use of the math coprocessor is
tickling a latent kernel bug?
I nuked /etc/ld.so.conf on my system, and it seems healthy. But with
this sort of nondeterministic lossage it's really hard to say, as the
system could go berzerk 2.3 seconds after I send this message...
Are you running with or without this new /etc/ld.so.conf on the
machine where you're getting these errors? And does the machine have