From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 17587 invoked by alias); 12 Mar 2011 17:42:29 -0000 Received: (qmail 17579 invoked by uid 22791); 12 Mar 2011 17:42:28 -0000 X-SWARE-Spam-Status: No, hits=-2.2 required=5.0 tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW X-Spam-Check-By: sourceware.org Received: from mail-ww0-f43.google.com (HELO mail-ww0-f43.google.com) (74.125.82.43) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Sat, 12 Mar 2011 17:42:20 +0000 Received: by wwb17 with SMTP id 17so3252335wwb.12 for ; Sat, 12 Mar 2011 09:42:18 -0800 (PST) Received: by 10.216.142.199 with SMTP id i49mr9099365wej.33.1299951737055; Sat, 12 Mar 2011 09:42:17 -0800 (PST) Received: from [192.168.2.99] (cpc2-cmbg8-0-0-cust61.5-4.cable.virginmedia.com [82.6.108.62]) by mx.google.com with ESMTPS id f52sm2849078wes.35.2011.03.12.09.42.14 (version=SSLv3 cipher=OTHER); Sat, 12 Mar 2011 09:42:15 -0800 (PST) Message-ID: <4D7BB065.8010605@gmail.com> Date: Sat, 12 Mar 2011 17:42:00 -0000 From: Dave Korn User-Agent: Thunderbird 2.0.0.17 (Windows/20080914) MIME-Version: 1.0 To: Alen Skondro CC: binutils@sourceware.org Subject: Re: ld coredump while compiling parma polyhedra library for i686-w64-mingw32 target References: In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Mailing-List: contact binutils-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: binutils-owner@sourceware.org X-SW-Source: 2011-03/txt/msg00229.txt.bz2 On 04/03/2011 18:02, Alen Skondro wrote: > While compiling PPL v0.11.2 for i686-w64-mingw32 target (cross compiler host is > i686-pc-cygwin) ld.exe coredumps with: > > I'm using > * GNU Binutils 2.21.51.20110304 > * GCC 4.6 While I'm working on reproducing this, here's a couple of things you can try: 1) If you haven't already done so, try setting your heap_chunk_in_mb setting to 1024 or even 1536, as described here: http://cygwin.com/cygwin-ug-net/setup-maxmem.html 2) The problem I think may be related to an increase in the number of comdat sections output by 4.6 versions of gcc relative to earlier ones, perhaps in conjunction with some inefficiency in the linker's handling of them. This might be mitigated by trying a build of PPL with CFLAGS including the new "-fno-keep-inline-dllexport" flag available in 4.6, but that could also conceivably lead to an incomplete library because it's a rather blunt tool (it needs the sources to be designed to take it into account for it to be guaranteed to work completely correctly). cheers, DaveK