From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 18 Jul 2002 21:41:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 18 Jul 2002 21:40:43 -0400 Received: from samba.sourceforge.net ([198.186.203.85]:9890 "HELO lists.samba.org") by vger.kernel.org with SMTP id ; Thu, 18 Jul 2002 21:40:41 -0400 From: Rusty Russell To: William Lee Irwin III Cc: linux-kernel@vger.kernel.org, Kai Germaschewski , sfr@canb.auug.org.au Subject: Re: [PATCH] Initcall depends automagic In-reply-to: Your message of "Thu, 18 Jul 2002 00:44:21 MST." <20020718074421.GN1096@holomorphy.com> Date: Fri, 19 Jul 2002 11:29:13 +1000 Message-Id: <20020719014425.7F9F5415E@lists.samba.org> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org In message <20020718074421.GN1096@holomorphy.com> you write: > On Thu, Jul 18, 2002 at 05:36:09PM +1000, Rusty Russell wrote: > > This patch generates initcall ordering based on the theory that if A.o > > references a symbol in B.o, then B.o's initcall must preceed A.o's. > > Any chance you can post the strongly connected components of the > dependency graph? Hmm, I don't have that at the moment, but apply the patch and do a build, and you'll get a ".defs" file for each .o file, eg kernel/futex.o.defs which contains U and T lines (as per nm output) showing what symbols it defines and requires. Then you can do all the analysis you need 8) Good luck! Rusty. -- Anyone who quotes me in their sig is an idiot. -- Rusty Russell.