From owner-freebsd-current Sun Jun 23 21:49:52 2002 Delivered-To: freebsd-current@freebsd.org Received: from mail.tgd.net (mail.tgd.net [209.81.25.10]) by hub.freebsd.org (Postfix) with ESMTP id 470CA37B401 for ; Sun, 23 Jun 2002 21:49:48 -0700 (PDT) Received: by mail.tgd.net (Postfix, from userid 1001) id 3E76220F02; Sun, 23 Jun 2002 21:49:47 -0700 (PDT) Date: Sun, 23 Jun 2002 21:49:47 -0700 From: Sean Chittenden To: Shizuka Kudo Cc: freebsd-current@freebsd.org Subject: Re: buildworld failed with -O0 Message-ID: <20020623214947.J84322@ninja1.internal> References: <20020624031801.26858.qmail@web11408.mail.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <20020624031801.26858.qmail@web11408.mail.yahoo.com>; from "shizukakudo_99@yahoo.com" on Sun, Jun 23, 2002 at = 08:18:01PM X-PGP-Key: 0x1EDDFAAD X-PGP-Fingerprint: C665 A17F 9A56 286C 5CFB 1DEA 9F4F 5CEF 1EDD FAAD X-Web-Homepage: http://sean.chittenden.org/ Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG > I don't know if I'm the one that observed this. Building world > completed with default CFLAGS (-O -pipe -march=pentiumpro), but > failed with -O0 or without optimization. /usr/src cvsupped within > one day. Here's the error message (same for -O0 and without -O). > > ===> libexec/rtld-elf > cc -pipe -march=pentiumpro -Wall -DFREEBSD_ELF > -I/usr/src/libexec/rtld-elf/i386 > -I/usr/src/libexec/rtld-elf -elf -fpic -DPIC > -Wformat=2 -Wno-format-extra-args -c > /usr/src/libexec/rtld-elf/rtld.c > /usr/src/libexec/rtld-elf/rtld.c: In function > `atomic_decr_int': > /usr/src/libexec/rtld-elf/i386/rtld_machdep.h:58: > inconsistent operand constraints in an `asm' > > Is it a gcc31 problem? I ran into, what I think is the same thing as you're describing, a while ago. The short and skinny being that -O0 doesn't compile any ASM which breaks things and is why you need at least -O to compile the kernel/world. ::shrug:: I could've botched that explanation, but it took me a while to track this down and have it explained. Kind of obscure IMHO. Someone suggested testing to make sure that at least -O is specified where needed, but I'm 99% sure I haven't seen that commit. -sc -- Sean Chittenden To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message