From gilham@csl.sri.com Sat Nov 23 07:29:53 1996 Received: from csla.csl.sri.com (csla.csl.sri.com [192.12.33.2]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id HAA29608 for ; Sat, 23 Nov 1996 07:29:51 -0800 (PST) Received: from japonica.csl.sri.com (japonica.csl.sri.com [130.107.15.17]) by csla.csl.sri.com (8.7.3/8.7.3) with ESMTP id HAA14512 for ; Sat, 23 Nov 1996 07:26:53 -0800 (PST) Received: (from gilham@localhost) by japonica.csl.sri.com (8.8.2/8.7.3) id IAA00469; Sat, 23 Nov 1996 08:28:25 GMT Message-Id: <199611230828.IAA00469@japonica.csl.sri.com> Date: Sat, 23 Nov 1996 08:28:25 GMT From: Fred Gilham Reply-To: gilham@csl.sri.com To: FreeBSD-gnats-submit@freebsd.org Subject: /etc/malloc.conf and AMD X-Send-Pr-Version: 3.2 >Number: 2093 >Category: bin >Synopsis: AMD gets sig 11 when /etc/malloc.conf is linked to AJ >Confidential: no >Severity: non-critical >Priority: medium >Responsible: freebsd-bugs >State: closed >Quarter: >Keywords: >Date-Required: >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Sat Nov 23 07:30:04 PST 1996 >Closed-Date: Thu Sep 18 04:15:56 PDT 1997 >Last-Modified: Thu Sep 18 04:16:23 PDT 1997 >Originator: Fred Gilham >Release: FreeBSD 2.2-ALPHA i386 >Organization: SRI International >Environment: Pentium-Pro 200. Home directory and others NFS mounted >Description: When /etc/malloc.conf is linked to AJ (as requested for testing purposes) then AMD gets a signal 11 whenever I try to log in. Removing /etc/malloc.conf causes AMD to function normally. I was unable to track the problem down to mounting a specific remote file system. >How-To-Repeat: (See description). >Fix: (N/A) >Release-Note: >Audit-Trail: From: Poul-Henning Kamp To: gilham@csl.sri.com Cc: FreeBSD-gnats-submit@FreeBSD.ORG Subject: Re: bin/2093: /etc/malloc.conf and AMD Date: Sat, 23 Nov 1996 17:05:31 +0100 I have mailed the author of amd on this issue: To: jsp@doc.ic.ac.uk Subject: AMD malloc problem. Date: Sat, 23 Nov 1996 17:02:39 +0100 Message-ID: <12545.848764959@critter.tfs.com> From: Poul-Henning Kamp Hi! We have received the attached bug-report in the FreeBSD group. Basically what seems to happen is that if malloc(3) and friends, exhibit some nasty but legal behaviour amd falls over. What happens when you set the malloc.conf to "AJ" is that all allocations gets filled with the byte value 0xd0 before being returned from malloc() or realloc() or when returned to free() or realloc(). This generally exposes errors where the program tacitly expects the memory to be zero-filled, or if memory is used after being freed, both of which is far more common than one would expect. (amongst the previous casualities we can mention fsck, ranlib, systat &c &c) Should you wish to try to reproduce this problem, you can easily use the "phkmalloc" implementation, simply by adding the file ftp://ftp.freebsd.org/pub/FreeBSD/FreeBSD-current/\ src/lib/libc/stdlib/malloc.c to the sources of amd. The manpage is ftp://ftp.freebsd.org/pub/FreeBSD/FreeBSD-current/\ src/lib/libc/stdlib/malloc.3 I must admit that I don't know if we are up to date with amd or if a newer version exists, in which case I presume we tell the original submitter to upgrade and test it again. Thanks in advance! Poul-Henning Kamp The FreeBSD core-team. State-Changed-From-To: open->closed State-Changed-By: phk State-Changed-When: Thu Sep 18 04:15:56 PDT 1997 State-Changed-Why: Havn't heard anything on this one for a long time. >Unformatted: