From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 16616 invoked by alias); 11 Mar 2011 04:23:42 -0000 Received: (qmail 16582 invoked by uid 22791); 11 Mar 2011 04:23:38 -0000 X-SWARE-Spam-Status: No, hits=-1.6 required=5.0 tests=AWL,BAYES_00,TW_BJ,TW_JC,T_RP_MATCHES_RCVD X-Spam-Check-By: sourceware.org Received: from mail.codesourcery.com (HELO mail.codesourcery.com) (38.113.113.100) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Fri, 11 Mar 2011 04:23:33 +0000 Received: (qmail 1544 invoked from network); 11 Mar 2011 04:23:30 -0000 Received: from unknown (HELO codesourcery.com) (froydnj@127.0.0.2) by mail.codesourcery.com with ESMTPA; 11 Mar 2011 04:23:30 -0000 From: Nathan Froyd To: gcc-patches@gcc.gnu.org Cc: Nathan Froyd , fortran@gcc.gnu.org, java-patches@gcc.gnu.org Subject: [4.7 PATCH 00/18] slim down a number of tree nodes Date: Fri, 11 Mar 2011 04:23:00 -0000 Message-Id: <1299817406-16745-1-git-send-email-froydnj@codesourcery.com> X-IsSubscribed: yes 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: 2011-03/txt/msg00549.txt.bz2 This patch series does something similar to what: http://gcc.gnu.org/ml/gcc-patches/2010-11/msg02491.html did, only it does it in a cleaner way and it addresses the problem more generally. It introduces a separate substructure (base class) for tree nodes that include TREE_TYPE so that a tree node can use TREE_TYPE without having to deal with TREE_CHAIN. It then goes about making changes where necessary to both use this new substructure and eliminate unnecessary TREE_CHAIN usage. Eliminating the block field from tree_exp would be another nice-to-have for 4.7, but that's not on my radar of things to address at the moment. (Matz, you want to do that? :) Eliminating TREE_TYPE from tree_exp as suggested on the wiki would be another interesting project, but not one I plan on tackling. The patch series touches every front-end in various places. I have CC'd the appropriate mailing lists with this introductory email, but I will only CC those mailing lists on followup patches that touch the appropriate FE. I have not rigorously measured memory savings with this patch. Based on a (very small) sample, this patch saves ~5% of tree memory according to dump_tree_statistics...though the amount of tree memory as reported by dump_tree_statistics is somewhat suspect, since it doesn't include statistics from copy_node_stat. The patch series has been bootstrapped on x86_64-unknown-linux-gnu, both in its entirety and with sub-patches along the way. Indeed, the patches were not developed in this order; the checking bits were introduced first, then any bootstrap or testsuite failures were fixed up, then the patches were committed in the proper order. Nathan Froyd (18): add typed_tree structure enforce TREE_CHAIN and TREE_TYPE accesses remove TREE_CHAIN from *_CST nodes remove TREE_CHAIN from SSA_NAME nodes remove TREE_CHAIN from CONSTRUCTOR nodes define CASE_CHAIN accessor for CASE_LABEL_EXPR generalize build_case_label to the rest of the compiler convert cp *FOR_STMTs to use private scope fields convert cp IF_STMTs to use private scope fields convert cp SWITCH_STMTs to use private scope fields mark EXPR_PACK_EXPANSION as typed only make CASE_LABEL_EXPR not abuse TREE_CHAIN move TS_EXP to be a substructure of TS_TYPED move TS_STATEMENT_LIST to be a substructure of TS_TYPED move REAL_IDENTIFIER_TYPE_VALUE to be a field of lang_identifier make TS_IDENTIFIER be a substructure of TS_BASE introduce block_chainon and use BLOCK_CHAIN more make TS_BLOCK a substructure of TS_BASE gcc/ada/gcc-interface/ada-tree.h | 2 +- gcc/ada/gcc-interface/misc.c | 16 ++++++ gcc/ada/gcc-interface/trans.c | 5 +- gcc/ada/gcc-interface/utils.c | 4 +- gcc/c-decl.c | 8 ++- gcc/c-family/c-common.c | 14 +++++- gcc/c-family/c-common.h | 20 +++++--- gcc/c-family/c-semantics.c | 28 ++++------- gcc/c-lang.c | 2 + gcc/c-parser.c | 2 +- gcc/c-typeck.c | 2 +- gcc/cp/cp-lang.c | 22 +-------- gcc/cp/cp-objcp-common.c | 74 +++++++++++++++++++++++++++ gcc/cp/cp-objcp-common.h | 1 + gcc/cp/cp-tree.def | 19 ++++--- gcc/cp/cp-tree.h | 43 ++++++++++------ gcc/cp/decl.c | 28 ++++++---- gcc/cp/decl2.c | 20 ++++--- gcc/cp/error.c | 2 +- gcc/cp/init.c | 6 +- gcc/cp/mangle.c | 28 +++++++--- gcc/cp/name-lookup.c | 10 ++-- gcc/cp/pt.c | 40 +++++++++------ gcc/cp/repo.c | 2 +- gcc/cp/rtti.c | 6 +- gcc/cp/search.c | 4 +- gcc/cp/semantics.c | 41 +++++++++------ gcc/cp/typeck.c | 4 +- gcc/except.c | 5 +- gcc/fortran/f95-lang.c | 8 ++-- gcc/fortran/trans-decl.c | 2 +- gcc/fortran/trans-io.c | 2 +- gcc/fortran/trans-stmt.c | 14 ++--- gcc/function.c | 28 ++++++++++ gcc/function.h | 1 + gcc/gimplify.c | 14 +++--- gcc/go/go-lang.c | 2 +- gcc/java/decl.c | 13 ++--- gcc/java/expr.c | 9 ++-- gcc/java/java-tree.h | 2 +- gcc/lto-streamer-in.c | 2 +- gcc/lto-streamer-out.c | 2 +- gcc/lto-streamer.c | 1 + gcc/lto/lto-tree.h | 2 +- gcc/lto/lto.c | 3 +- gcc/objc/objc-act.c | 18 +++++++ gcc/objc/objc-act.h | 1 + gcc/objc/objc-lang.c | 32 +----------- gcc/objcp/objcp-lang.c | 48 +----------------- gcc/omp-low.c | 7 +-- gcc/print-tree.c | 11 +--- gcc/tree-cfg.c | 12 ++-- gcc/tree-eh.c | 22 +++----- gcc/tree-flow.h | 2 +- gcc/tree-inline.c | 3 +- gcc/tree-iterator.c | 17 ++---- gcc/tree-ssanames.c | 28 +++------- gcc/tree.c | 96 ++++++++++++----------------------- gcc/tree.def | 2 +- gcc/tree.h | 103 +++++++++++++++++++++++++++++-------- gcc/treestruct.def | 1 + gcc/varasm.c | 1 - 62 files changed, 535 insertions(+), 432 deletions(-)