From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 16120 invoked by alias); 18 Jun 2008 04:17:43 -0000 Received: (qmail 16111 invoked by uid 22791); 18 Jun 2008 04:17:42 -0000 X-Spam-Check-By: sourceware.org Received: from yw-out-1718.google.com (HELO yw-out-1718.google.com) (74.125.46.156) by sourceware.org (qpsmtpd/0.31) with ESMTP; Wed, 18 Jun 2008 04:17:23 +0000 Received: by yw-out-1718.google.com with SMTP id 5so40116ywm.26 for ; Tue, 17 Jun 2008 21:17:21 -0700 (PDT) Received: by 10.150.134.21 with SMTP id h21mr265082ybd.156.1213762641099; Tue, 17 Jun 2008 21:17:21 -0700 (PDT) Received: from ?10.255.254.191? ( [66.135.114.72]) by mx.google.com with ESMTPS id p30sm5678849qbp.18.2008.06.17.21.17.19 (version=TLSv1/SSLv3 cipher=RC4-MD5); Tue, 17 Jun 2008 21:17:19 -0700 (PDT) Message-ID: <48588C3A.5020703@naturalbridge.com> Date: Wed, 18 Jun 2008 05:34:00 -0000 From: Kenneth Zadeck User-Agent: Thunderbird 2.0.0.14 (X11/20080421) MIME-Version: 1.0 To: Steven Bosscher CC: gcc-patches Subject: Re: [patch] Make df_ref point to a df_insn_info instead of an insn rtx References: <571f6b510806171620i519d4501r67aa365de6c0ceec@mail.gmail.com> In-Reply-To: <571f6b510806171620i519d4501r67aa365de6c0ceec@mail.gmail.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2008-06/txt/msg01131.txt.bz2 Steven Bosscher wrote: > Hi, > > This is the patch to change how df_refs point to their insns. Without > this patch, a df_ref points directly to the insn's rtx. And from the > insn, we get the df_insn_info via DF_INSN_GET, which points to the > insn def and use refs. > > But if we are going to have fake insns that have defs and uses, but no > real insns (e.g. PHI nodes), we need a different way to find the > sibling defs/uses for a given reference. The easiest way is to just > point directly at the df_insn_info for the insn that contains the > reference. This way, we can then still point at a df_insn_info for > fake insns. > > Attached patch was bootstrapped&tested on ia64-unknown-linux-gnu (and > an earlier equivalent version on powerpc). Kenny, what do you think > about this? OK for trunk? > > Gr. > Steven > one comment: ChangeLog: s/df_code.c/df_core.c/ fine for commit to trunk. i bootstrapped this on ia-64 just to make sure that the ia-64 only passes were properly changed. kenny