From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 20929 invoked by alias); 30 Sep 2009 17:20:29 -0000 Received: (qmail 20917 invoked by uid 22791); 30 Sep 2009 17:20:27 -0000 X-SWARE-Spam-Status: No, hits=-1.6 required=5.0 tests=AWL,BAYES_00,SARE_MSGID_LONG40,SPF_HELO_PASS,SPF_PASS X-Spam-Check-By: sourceware.org Received: from smtp-out.google.com (HELO smtp-out.google.com) (216.239.33.17) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 30 Sep 2009 17:20:23 +0000 Received: from zps19.corp.google.com (zps19.corp.google.com [172.25.146.19]) by smtp-out.google.com with ESMTP id n8UHKK0e017617 for ; Wed, 30 Sep 2009 18:20:21 +0100 Received: from gxk24 (gxk24.prod.google.com [10.202.11.24]) by zps19.corp.google.com with ESMTP id n8UHJRlt009665 for ; Wed, 30 Sep 2009 10:20:18 -0700 Received: by gxk24 with SMTP id 24so3947851gxk.5 for ; Wed, 30 Sep 2009 10:20:17 -0700 (PDT) MIME-Version: 1.0 Received: by 10.91.183.4 with SMTP id k4mr91543agp.41.1254331217780; Wed, 30 Sep 2009 10:20:17 -0700 (PDT) Date: Wed, 30 Sep 2009 18:53:00 -0000 Message-ID: <522e93240909301020q5b772da9xb529cb4749c6d67a@mail.gmail.com> Subject: [lipo]: merge from trunk From: Xinliang David Li To: GCC Patches Content-Type: text/plain; charset=ISO-8859-1 X-System-Of-Record: true 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: 2009-09/txt/msg02250.txt.bz2 This is a merge from trunk 151395:151806 Along with the merge, several fixes to resolve the conflicts: 1) Disallow auxiliary functions in ipa early sra. The aux functions may get deleted. 2) Change the assembler name for sra transformed function to avoid name conflict in assembly file -- the conflict may come from a transformed function and another function in aux module which is emitted (e.g, in comdat) 3) Due to the phase ordering change of thunk emission, thunks for external auxilary functions are conditionally emitted -- leading to inconsistent comdat group (with the defining primary module) -- this can result in unsats. 4) during cfg fixup for sra transformation, the eh landing pad can be missing. This is due to the wrong order of eh edge manipulation. Testing: SPEC2k int with LIPO -- all clean SPEC06 (C/C++) with LIPO -- all clean except deallI with an outstanding problem: missing ipa summary for a new generated function. David