From patchwork Tue Jun 21 12:07:20 2011
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
X-Patchwork-Submitter: Ira Rosen <ira.rosen@linaro.org>
X-Patchwork-Id: 2120
Return-Path: <patch+caf_=linaro-patchwork=canonical.com@linaro.org>
X-Original-To: patchwork@peony.canonical.com
Delivered-To: patchwork@peony.canonical.com
Received: from fiordland.canonical.com (fiordland.canonical.com
 [91.189.94.145])
 by peony.canonical.com (Postfix) with ESMTP id 94E4B23E02
 for <patchwork@peony.canonical.com>;
 Tue, 21 Jun 2011 12:07:23 +0000 (UTC)
Received: from mail-vx0-f180.google.com (mail-vx0-f180.google.com
 [209.85.220.180])
 by fiordland.canonical.com (Postfix) with ESMTP id 51738A187E5
 for <linaro-patchwork@canonical.com>;
 Tue, 21 Jun 2011 12:07:23 +0000 (UTC)
Received: by vxd7 with SMTP id 7so3444989vxd.11
 for <linaro-patchwork@canonical.com>;
 Tue, 21 Jun 2011 05:07:22 -0700 (PDT)
Received: by 10.52.98.97 with SMTP id eh1mr9072435vdb.7.1308658042739;
 Tue, 21 Jun 2011 05:07:22 -0700 (PDT)
X-Forwarded-To: linaro-patchwork@canonical.com
X-Forwarded-For: patch@linaro.org linaro-patchwork@canonical.com
Delivered-To: patches@linaro.org
Received: by 10.52.183.130 with SMTP id em2cs61796vdc;
 Tue, 21 Jun 2011 05:07:22 -0700 (PDT)
Received: by 10.220.65.97 with SMTP id h33mr1757084vci.52.1308658042251;
 Tue, 21 Jun 2011 05:07:22 -0700 (PDT)
Received: from mail-pv0-f178.google.com (mail-pv0-f178.google.com
 [74.125.83.178]) by mx.google.com with ESMTPS id
 14si4026094vcy.150.2011.06.21.05.07.20
 (version=TLSv1/SSLv3 cipher=OTHER);
 Tue, 21 Jun 2011 05:07:21 -0700 (PDT)
Received-SPF: neutral (google.com: 74.125.83.178 is neither permitted nor
 denied by best guess record for domain of
 ira.rosen@linaro.org) client-ip=74.125.83.178; 
Authentication-Results: mx.google.com;
 spf=neutral (google.com: 74.125.83.178 is neither
 permitted nor denied by best guess record for domain of
 ira.rosen@linaro.org) smtp.mail=ira.rosen@linaro.org
Received: by pvg7 with SMTP id 7so2882766pvg.37
 for <patches@linaro.org>; Tue, 21 Jun 2011 05:07:20 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.142.188.14 with SMTP id l14mr981690wff.398.1308658040159;
 Tue, 21 Jun 2011 05:07:20 -0700 (PDT)
Received: by 10.143.79.13 with HTTP; Tue, 21 Jun 2011 05:07:20 -0700 (PDT)
Date: Tue, 21 Jun 2011 15:07:20 +0300
Message-ID: <BANLkTimYuAQPw+oJfUbM97QGEUuKCq-8GA@mail.gmail.com>
Subject: [patch] Fix PR tree-optimization/49478
From: Ira Rosen <ira.rosen@linaro.org>
To: gcc-patches@gcc.gnu.org
Cc: Patch Tracking <patches@linaro.org>

Hi,

This patch adds a missing support of widening multiplication by a
constant in case of DOT_PROD_EXPR: the type of the constant operand
should be set according to the other operand.

Bootstrapped and tested on powerpc64-suse-linux.
Committed.

Ira

ChangeLog:

   PR tree-optimization/49478
   * tree-vect-loop.c (vectorizable_reduction): Handle DOT_PROD_EXPR
   with constant operand.

testsuite/ChangeLog:

   PR tree-optimization/49478
   * gcc.dg/vect/pr49478.c

Index: testsuite/gcc.dg/vect/pr49478.c
===================================================================
--- testsuite/gcc.dg/vect/pr49478.c     (revision 0)
+++ testsuite/gcc.dg/vect/pr49478.c     (revision 0)
@@ -0,0 +1,24 @@
+/* { dg-do compile } */
+/* { dg-require-effective-target vect_int } */
+
+#include <stdarg.h>
+
+#define N 64
+
+unsigned char X[N] __attribute__
((__aligned__(__BIGGEST_ALIGNMENT__))) =
{0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63};
+
+unsigned int
+foo (int len) {
+  int i;
+  unsigned int result = 0;
+  unsigned short prod;
+
+  for (i=0; i<len; i++) {
+    prod = X[i] * 15;
+    result += prod;
+  }
+  return result;
+}
+
+/* { dg-final { cleanup-tree-dump "vect" } } */
+
Index: tree-vect-loop.c
===================================================================
--- tree-vect-loop.c    (revision 175252)
+++ tree-vect-loop.c    (working copy)
@@ -4591,6 +4591,25 @@ vectorizable_reduction (gimple stmt, gimple_stmt_i
       return false;
     }

+  /* In case of widenning multiplication by a constant, we update the type
+     of the constant to be the type of the other operand.  We check that the
+     constant fits the type in the pattern recognition pass.  */
+  if (code == DOT_PROD_EXPR
+      && !types_compatible_p (TREE_TYPE (ops[0]), TREE_TYPE (ops[1])))
+    {
+      if (TREE_CODE (ops[0]) == INTEGER_CST)
+        ops[0] = fold_convert (TREE_TYPE (ops[1]), ops[0]);
+      else if (TREE_CODE (ops[1]) == INTEGER_CST)
+        ops[1] = fold_convert (TREE_TYPE (ops[0]), ops[1]);
+      else
+        {
+          if (vect_print_dump_info (REPORT_DETAILS))
+            fprintf (vect_dump, "invalid types in dot-prod");
+
+          return false;
+        }
+    }
+
   if (!vec_stmt) /* transformation not required.  */