From owner-freebsd-current@FreeBSD.ORG Mon Jun 7 09:14:29 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 971F016A4CE for ; Mon, 7 Jun 2004 09:14:29 +0000 (GMT) Received: from itchy.rabson.org (mailgate.nlsystems.com [80.177.232.242]) by mx1.FreeBSD.org (Postfix) with ESMTP id BF4A143D58 for ; Mon, 7 Jun 2004 09:14:28 +0000 (GMT) (envelope-from dfr@nlsystems.com) Received: from ns0.nlsystems.com (ns0.nlsystems.com [80.177.232.243]) by itchy.rabson.org (8.12.11/8.12.11) with ESMTP id i579A4n8016672 for ; Mon, 7 Jun 2004 10:10:04 +0100 (BST) (envelope-from dfr@nlsystems.com) From: Doug Rabson To: freebsd-current@freebsd.org Date: Mon, 7 Jun 2004 10:14:21 +0100 User-Agent: KMail/1.6.1 References: <40C37E1C.4000402@freebsd.org> <20040606204817.GB96607@dhcp50.pn.xcllnt.net> In-Reply-To: <20040606204817.GB96607@dhcp50.pn.xcllnt.net> MIME-Version: 1.0 Content-Disposition: inline Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <200406071014.21707.dfr@nlsystems.com> X-Spam-Status: No, hits=0.0 required=5.0 tests=none autolearn=no version=2.63 X-Spam-Checker-Version: SpamAssassin 2.63 (2004-01-11) on itchy.rabson.org X-Virus-Scanned: clamd / ClamAV version 0.71, clamav-milter version 0.71 X-Virus-Status: Clean Subject: Re: HEADS UP! KSE needs more attention X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Jun 2004 09:14:29 -0000 On Sunday 06 June 2004 21:48, Marcel Moolenaar wrote: > On Sun, Jun 06, 2004 at 02:27:08PM -0600, Scott Long wrote: > > >Doug Rabson also has basic TLS support working in perforce. > > > > What platforms? My understanding was that new binutils and gcc was > > needed for sparc64 at a minimum. > > Yes. It's i386 only and not even close to being complete. In fact, > there has been discussions that the thread pointer on i386 needs to > change. Whether that's the case or not, it's likely that TLS will > complicate matters way too much to for it to ever work in 5.3. Actually its a bit better than that. It works for most use cases right now on i386 but would get confused on dlclose. I'll fix that before I move it into current. As far as the thread pointer issue goes, it turns out that our current compiler will always evaluate %gs:0 first which is fine for libpthread. I looked at gcc-3.4 sources and it can also be configured to always evaluate %gs:0 (this is the default for non-linux platforms).