From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 6110 invoked by alias); 30 Jan 2003 17:17:49 -0000 Mailing-List: contact libc-hacker-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libc-hacker-owner@sources.redhat.com Received: (qmail 6082 invoked from network); 30 Jan 2003 17:17:48 -0000 Received: from unknown (HELO sunsite.mff.cuni.cz) (195.113.19.66) by 172.16.49.205 with SMTP; 30 Jan 2003 17:17:48 -0000 Received: (from jakub@localhost) by sunsite.mff.cuni.cz (8.11.6/8.11.6) id h0UHHWX17522; Thu, 30 Jan 2003 18:17:32 +0100 Date: Thu, 30 Jan 2003 17:17:00 -0000 From: Jakub Jelinek To: davidm@hpl.hp.com Cc: Ulrich Drepper , libc-hacker@sources.redhat.com Subject: Re: hack to get sysinfo initialized Message-ID: <20030130181732.B17910@sunsite.ms.mff.cuni.cz> Reply-To: Jakub Jelinek References: <20030130173845.W17910@sunsite.ms.mff.cuni.cz> <15929.22278.43634.52356@napali.hpl.hp.com> <3E395882.8060201@redhat.com> <15929.23049.292170.896244@napali.hpl.hp.com> <3E395BBF.8080407@redhat.com> <15929.23857.452779.385080@napali.hpl.hp.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: <15929.23857.452779.385080@napali.hpl.hp.com>; from davidm@napali.hpl.hp.com on Thu, Jan 30, 2003 at 09:13:21AM -0800 X-SW-Source: 2003-01/txt/msg00221.txt.bz2 On Thu, Jan 30, 2003 at 09:13:21AM -0800, David Mosberger wrote: > >>>>> On Thu, 30 Jan 2003 09:07:11 -0800, Ulrich Drepper said: > > Uli> LinuxThreads for x86 is meant for compatibility and adding something > Uli> like this which isn't very often just doesn't seem necessary. If you > Uli> have a new kernel you want to run a decent thread library. > > Sure, I'd love to. I thought the ia64 NPTL bits weren't quite ready yet. > But yes, it will be nice to have it. > > Uli> There is no fundamental problem but it might just open another > Uli> can of worms (i.e., it needs testing). > > I can see that, but unless I'm missing something, the sysinfo member > is always there in the thread descriptor. Leaving it uninitialized > seems more dangerous than initializing it with a sane value, no? The current state of AT_SYSINFO stuff is that NPTL libc/libpthread on i686 uses it and ld.so from both NPTL and linuxthreads builds set it up. This is so that ld.so from linuxthreads build can load both linuxthreads and NPTL libs. Jakub