From patchwork Tue Nov  9 15:18:52 2010
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Patchwork-Submitter: Paul Koning <Paul_Koning@Dell.com>
X-Patchwork-Id: 70551
Return-Path: 
 <gcc-patches-return-277821-incoming=patchwork.ozlabs.org@gcc.gnu.org>
X-Original-To: incoming@patchwork.ozlabs.org
Delivered-To: patchwork-incoming@bilbo.ozlabs.org
Received: from sourceware.org (server1.sourceware.org [209.132.180.131])
	by ozlabs.org (Postfix) with SMTP id 27079B6EF0
	for <incoming@patchwork.ozlabs.org>;
	Wed, 10 Nov 2010 02:19:10 +1100 (EST)
Received: (qmail 17131 invoked by alias); 9 Nov 2010 15:19:04 -0000
Received: (qmail 17120 invoked by uid 22791); 9 Nov 2010 15:19:02 -0000
X-SWARE-Spam-Status: No, hits=-1.7 required=5.0	tests=AWL, BAYES_00, TW_SV,
	T_RP_MATCHES_RCVD
X-Spam-Check-By: sourceware.org
Received: from ausc60pc101.us.dell.com (HELO ausc60pc101.us.dell.com)
	(143.166.85.206) by sourceware.org (qpsmtpd/0.43rc1) with
	ESMTP; Tue, 09 Nov 2010 15:18:56 +0000
X-Loopcount0: from 10.152.240.141
Mime-Version: 1.0 (Apple Message framework v1081)
Subject: Re: [PATCH] pdp11: fix wrong code
From: Paul Koning <paul_koning@dell.com>
In-Reply-To: <086E06BA-0C4E-433C-A172-C9AA17BA466D@dell.com>
Date: Tue, 9 Nov 2010 10:18:52 -0500
Message-Id: <CC5C764B-EFD8-4CBF-BB2E-B92B1CACD15A@dell.com>
References: <086E06BA-0C4E-433C-A172-C9AA17BA466D@dell.com>
To: gcc-patches <gcc-patches@gcc.gnu.org>
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Unsubscribe: 
 <mailto:gcc-patches-unsubscribe-incoming=patchwork.ozlabs.org@gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
Delivered-To: mailing list gcc-patches@gcc.gnu.org

On Nov 8, 2010, at 8:15 PM, Paul Koning wrote:

> This is another wrong-code fix, in logical right shift.  The original code tried to optimize the case for shift by 1, but ended up unconditionally shifting by 1 even if the request was for a different shift.
> 
> Tested by build and inspection of the output.  Committed.
> 
> 	paul

Here is a correction to that patch, needed because subtract constant should be encoded as an add.

Tested by build, make check.  Committed.

	paul

ChangeLog:

2010-11-09  Paul Koning  <ni1d@arrl.net>

	* config/pdp11/pdp11.md (lshrsi3, lshrhi3): Use add for
	decrement. 

aaa-5v57ooxfyrm:gcc pkoning$ svn update ChangeLog

Index: config/pdp11/pdp11.md
===================================================================
--- config/pdp11/pdp11.md	(revision 166480)
+++ config/pdp11/pdp11.md	(working copy)
@@ -897,7 +897,7 @@
   if (GET_CODE (operands[2]) != CONST_INT)
     {
       r = gen_reg_rtx (HImode);
-      emit_insn (gen_subhi3 (r, operands [2], GEN_INT (1)));
+      emit_insn (gen_addhi3 (r, operands [2], GEN_INT (-1)));
       emit_insn (gen_ashrsi3 (operands[0], operands[0], r));
     }
   else if ((unsigned) INTVAL (operands[2]) != 1)
@@ -1043,7 +1043,7 @@
   if (GET_CODE (operands[2]) != CONST_INT)
     {
       r = gen_reg_rtx (HImode);
-      emit_insn (gen_subhi3 (r, operands [2], GEN_INT (1)));
+      emit_insn (gen_addhi3 (r, operands [2], GEN_INT (-1)));
       emit_insn (gen_ashrhi3 (operands[0], operands[0], r));
     }
   else if ((unsigned) INTVAL (operands[2]) != 1)