From owner-freebsd-current  Wed Oct  9 14:46:11 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id OAA08689
          for current-outgoing; Wed, 9 Oct 1996 14:46:11 -0700 (PDT)
Received: from bourbon.sfc.wide.ad.jp (root@bourbon.sfc.wide.ad.jp [133.4.30.17])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id OAA08666
          for <freebsd-current@freebsd.org>; Wed, 9 Oct 1996 14:46:06 -0700 (PDT)
Received: from bourbon (max@localhost [127.0.0.1]) by bourbon.sfc.wide.ad.jp (8.7.6/3.4W409/27/96) with ESMTP id GAA04189; Thu, 10 Oct 1996 06:45:41 +0900 (JST)
Message-Id: <199610092145.GAA04189@bourbon.sfc.wide.ad.jp>
To: jdp@polstra.com
Cc: max@wide.ad.jp, freebsd-current@freebsd.org
Subject: Re: ldconfig problem 
From: Masafumi NAKANE/=?ISO-2022-JP?B?GyRCQ2Y6LDJtSjgbKEI=?= <max@wide.ad.jp>
In-Reply-To: Your message of "Wed, 09 Oct 1996 08:33:18 -0700"
References: <199610091533.IAA23308@austin.polstra.com>
X-Mailer: Mew version 1.06 on Emacs 19.28.1, Mule 2.3
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Date: Thu, 10 Oct 1996 06:45:41 +0900
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

    jdp> Hmm, these both look OK.  The shared libraries in the second
    jdp> case are mapped one page higher than in the first case, but
    jdp> that doesn't indicate a problem.  Before mapping any shared
    jdp> libraries, the dynamic linker maps the entire
    jdp> "/var/run/ld.so.hints" file into memory.  In the second case,
    jdp> that file is probably one page larger than in the first case,
    jdp> because it has the extra libraries from
    jdp> "/usr/local/lib/m3/FreeBSD2" in it.

    jdp> Does kaffe use dlopen(), or LD_PRELOAD, or any other unusual
    jdp> features of the dynamic linker?
I'm not really sure about this.  But, it looks it uses dlopen as
several files contain the string ``dlopen''.

    jdp> Can you examine the core file with a debugger and try to
    jdp> figure out what it's doing when it dies?
Guess what!  Things are getting worse.  (At least it seems to me that
way.)  I started up gdb and and did ``help set'' to find out some
stuff.  Gdb gave me the first page of the help screen as usual, but as
I hit return key to go onto the next page, it died on signal 3 with
following message after showing the second page:

    gdb internal error: Memory corruption
    Quit (core dumped)

This error is not affected by ld.so.hints change.  It happens
regardless of whether kaffe runs or not.  So, I guess now we have to
do some investigation on gdb.... ;_)

Do you think having CFLAGS=-O -pipe -m486 in /etc/make.conf coud be
causing these errors?  After I discovered gdb's error, I rebuilt
/usr/src/gnu/usr.bin/{gdb,ld} with -O and -m486 removed from
/etc/make.conf, but things don't change.  Should I do make world
without those flags and try again?

Max