From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 10436 invoked by alias); 20 Mar 2005 15:39:27 -0000 Mailing-List: contact binutils-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: binutils-owner@sources.redhat.com Received: (qmail 10369 invoked from network); 20 Mar 2005 15:39:24 -0000 Received: from unknown (HELO nevyn.them.org) (66.93.172.17) by sourceware.org with SMTP; 20 Mar 2005 15:39:24 -0000 Received: from drow by nevyn.them.org with local (Exim 4.44 #1 (Debian)) id 1DD2WK-0006sW-Ux; Sun, 20 Mar 2005 10:39:17 -0500 Date: Sun, 20 Mar 2005 19:45:00 -0000 From: Daniel Jacobowitz To: "Peter S. Mazinger" Cc: binutils@sourceware.org Subject: Re: elf32-arm.c corrections Message-ID: <20050320153916.GA26320@nevyn.them.org> Mail-Followup-To: "Peter S. Mazinger" , binutils@sourceware.org References: <20050320023224.GA12444@nevyn.them.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.6+20040907i X-SW-Source: 2005-03/txt/msg00596.txt.bz2 On Sun, Mar 20, 2005 at 10:26:12AM +0100, Peter S. Mazinger wrote: > On Sat, 19 Mar 2005, Daniel Jacobowitz wrote: > > > On Sun, Mar 20, 2005 at 02:15:24AM +0100, Peter S. Mazinger wrote: > > > Hello! > > > > > > add_dynamic_entry: changes !info->shared to info->executable (PIE) > > > corrects typo, and syncs up with other archs (some others could do the > > > same). For !relocs the hole part would be omitted, probably some speed gain. > > > > Um, why are disabling the setting of DT_TEXTREL for shared libraries? > > The 1 line removal is because none of the archs has that. > The i386 implementation has everything within if (relocs), the other > archs have it outside, but DT_TEXTREL is only valid for if (relocs), so we > would omit that part gaining some speed Ah, not enough context in the diff. Makes sense. > > > Should the other patch (*3) for allocate_dynrelocs be applied (as done for > > > ppc32)? > > > > Probably. > > I am asking it, because no other arch has that, only ppc32, so there must > be some other solution to that as well. > > > > > > Why is ELIMINATE_COPY_RELOCS not used for arm? > > > > Because no one implemented it. > > Would the implementation make the binaries smaller? Not appreciably. -- Daniel Jacobowitz CodeSourcery, LLC