From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 31101 invoked by alias); 5 Aug 2002 22:18:37 -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 31085 invoked from network); 5 Aug 2002 22:18:37 -0000 Received: from unknown (HELO sunsite.mff.cuni.cz) (195.113.19.66) by sources.redhat.com with SMTP; 5 Aug 2002 22:18:37 -0000 Received: (from jakub@localhost) by sunsite.mff.cuni.cz (8.11.6/8.11.6) id g75MHVb22996; Tue, 6 Aug 2002 00:17:31 +0200 Date: Mon, 05 Aug 2002 15:18:00 -0000 From: Jakub Jelinek To: Roland McGrath Cc: Ulrich Drepper , Glibc hackers Subject: Re: [PATCH] Nuke another 21 .plt slots Message-ID: <20020806001731.O20867@sunsite.ms.mff.cuni.cz> Reply-To: Jakub Jelinek References: <20020805235431.N20867@sunsite.ms.mff.cuni.cz> <20020805221155.C64591BA1B@perdition.linnaean.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <20020805221155.C64591BA1B@perdition.linnaean.org>; from roland@frob.com on Mon, Aug 05, 2002 at 06:11:55PM -0400 X-SW-Source: 2002-08/txt/msg00080.txt.bz2 On Mon, Aug 05, 2002 at 06:11:55PM -0400, Roland McGrath wrote: > Thanks! I put those in. However, my PLT count does not match yours. I > get 169 in a build with TLS support, and I think that is only one more than > without TLS. > > I have been counting using this script (it gives a list of symbols; pipe it > to wc -l), which counts only PLT slots for symbols defined in libc itself, > and omits the malloc functions that are explicit exceptions. That count is > now 121. > > #!/bin/sh > > goodones='__libc_free > __libc_malloc > __libc_memalign > __libc_realloc > free > calloc > malloc > realloc' This list is too short, e.g. __fork and some others which are to be overridden by libpthread.so should be in the list too. > readelf -rW libc.so | awk '/JUMP_SLOT/ && $4 != "00000000" { print $5 }' | > fgrep -vx "$goodones" | sort I believe the difference can be that my build was with -O3, so a bunch of calls were inlined and thus not going externally. I'll try an -O2 build tomorrow. Jakub