From owner-freebsd-hackers@FreeBSD.ORG Tue Apr 6 06:22:29 2004 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id E4C1016A4CE for ; Tue, 6 Apr 2004 06:22:29 -0700 (PDT) Received: from gwdu60.gwdg.de (gwdu60.gwdg.de [134.76.8.60]) by mx1.FreeBSD.org (Postfix) with ESMTP id 22F8743D49 for ; Tue, 6 Apr 2004 06:22:29 -0700 (PDT) (envelope-from kheuer2@gwdg.de) Received: from gwdu60.gwdg.de (localhost [127.0.0.1]) by gwdu60.gwdg.de (8.12.9p2/8.12.8) with ESMTP id i36DLCrB039184 for ; Tue, 6 Apr 2004 15:21:12 +0200 (CEST) (envelope-from kheuer2@gwdg.de) Received: from localhost (kheuer2@localhost)i36DLBNk039181 for ; Tue, 6 Apr 2004 15:21:12 +0200 (CEST) X-Authentication-Warning: gwdu60.gwdg.de: kheuer2 owned process doing -bs Date: Tue, 6 Apr 2004 15:21:11 +0200 (CEST) From: Konrad Heuer To: freebsd-hackers@freebsd.org Message-ID: <20040406152018.G30924@gwdu60.gwdg.de> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Subject: Critical mmap failure? X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 06 Apr 2004 13:22:30 -0000 I've a couple of systems running FreeBSD 4.9; they all share /usr and anything below by nfs. There are (at less) two applications that do not run or do not run correctly on one distinguished system, but correctly on each other machine. One application is mozilla; it stops shortly after invocation with the following error message: INTERNAL ERROR on Browser End: JavaPluginFactory5 init - no agent? System error?:: Cannot allocate memory Analyzing the problem with "truss" shows an error while mozilla calls "mmap": mmap(0xbfaef000,65536,0x3,0x400,-1,0x0) ERR#12 'Cannot allocate memory' mmap(0xbfade000,65536,0x3,0x400,-1,0x0) ERR#12 'Cannot allocate memory' On the other systems, there are no errors: mmap(0xbfaef000,65536,0x3,0x400,-1,0x0) = -1079054336 (0xbfaef000) mmap(0xbfade000,65536,0x3,0x400,-1,0x0) = -1079123968 (0xbfade000) * Other "mmap" calls that do not require MAP_STACK (0x400) do not lead to errors anywhere. * All kernels are compiled with -DVM_STACK since this is default on i386. * The value of the variable vm.max_proc_mmap is not higher on any of the systems which do well than on the problematic one. * The same holds for resource limits. * The machines in question are all DELL PowerEdge 2650, three with RAID controllers, one without. The main difference in hardware is that the system on wich the error occurs has 4 GB of memory, the others 2 GB. There are no significant differences in the kernel configuration files except driver entries for RAID or not. Any idea that may help is very welcome, since the other application that fails is a commercial linux binary calling "linux mmap" frequently and producing wrong data (possibly the return code of "mmap" isn't checked). This application is very important for us, and does not do anything than reading and writing data and calling "mmap". Best regards Konrad Heuer (kheuer2@gwdg.de) ____ ___ _______ GWDG / __/______ ___ / _ )/ __/ _ \ Am Fassberg / _// __/ -_) -_) _ |\ \/ // / 37077 Goettingen /_/ /_/ \__/\__/____/___/____/ Germany