From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 22825 invoked by alias); 3 Dec 2008 20:03:10 -0000 Received: (qmail 22814 invoked by uid 22791); 3 Dec 2008 20:03:10 -0000 X-Spam-Level: * X-Spam-Status: No, hits=-2.5 required=5.0 tests=AWL,BAYES_00 X-Spam-Check-By: sourceware.org Received: from NaN.false.org (HELO nan.false.org) (208.75.86.248) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 03 Dec 2008 20:02:30 +0000 Received: from nan.false.org (localhost [127.0.0.1]) by nan.false.org (Postfix) with ESMTP id 94A3210EB6; Wed, 3 Dec 2008 20:02:27 +0000 (GMT) Received: from caradoc.them.org (209.195.188.212.nauticom.net [209.195.188.212]) by nan.false.org (Postfix) with ESMTP id 7DE2B10EB1; Wed, 3 Dec 2008 20:02:27 +0000 (GMT) Received: from drow by caradoc.them.org with local (Exim 4.69) (envelope-from ) id 1L7xvW-0007Wm-Hs; Wed, 03 Dec 2008 15:02:26 -0500 Date: Wed, 03 Dec 2008 20:03:00 -0000 From: Daniel Jacobowitz To: Catherine Moore , prelink@sourceware.org, rdsandiford@googlemail.com Subject: Re: [patch] Add support for new mips relocs Message-ID: <20081203200226.GA28907@caradoc.them.org> References: <4936D3D3.6030306@codesourcery.com> <87r64pcapp.fsf@firetop.home> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <87r64pcapp.fsf@firetop.home> User-Agent: Mutt/1.5.17 (2008-05-11) X-IsSubscribed: yes Mailing-List: contact prelink-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: prelink-owner@sourceware.org X-SW-Source: 2008-q4/txt/msg00003.txt.bz2 Catherine, please preserve attribution when you can; I wrote all this. Thanks. On Wed, Dec 03, 2008 at 07:36:34PM +0000, Richard Sandiford wrote: > I'm probably missing something, sorry, but how does the patch handle > the property mentioned in the original comment? Relocations must never > resolve to another object's (traditional) lazy-binding stub. This does happen, but the code that does it is far away: in this case, in glibc. The dynamic loader will output which object it bound the relocation to, and it will be one that contains a definition rather than a lazy binding stub. As you may remember, we also have cross prelink support in our local tree (I'm willing to take a stab at submitting this if Jakub is interested). That has basically a copy of the glibc resolver in it. And that's where the STO_MIPS_PLT check lives for cross operation. -- Daniel Jacobowitz CodeSourcery