From 89e7b356f270e29c2e9e18c6bbd30e5925585f06 Mon Sep 17 00:00:00 2001
From: Nadav Rotem <nrotem@apple.com>
Date: Fri, 19 Oct 2012 01:24:18 +0000
Subject: [PATCH] vectorizer: Add support for reading and writing from the same
 memory location.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@166255 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/Transforms/Vectorize/LoopVectorize.cpp | 11 +++++--
 test/Transforms/LoopVectorize/increment.ll | 35 ++++++++++++++++++++++
 2 files changed, 44 insertions(+), 2 deletions(-)
 create mode 100644 test/Transforms/LoopVectorize/increment.ll

diff --git a/lib/Transforms/Vectorize/LoopVectorize.cpp b/lib/Transforms/Vectorize/LoopVectorize.cpp
index f5c9bb31e05..5152ec11e5b 100644
--- a/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -715,6 +715,7 @@ bool LoopVectorizationLegality::canVectorizeBlock(BasicBlock &BB) {
   ValueVector Reads;
   ValueVector Writes;
 
+  SmallPtrSet<Value*, 16> AnalyzedPtrs;
   unsigned NumPhis = 0;
   for (BasicBlock::iterator it = BB.begin(), e = BB.end(); it != e; ++it) {
     Instruction *I = it;
@@ -766,7 +767,10 @@ bool LoopVectorizationLegality::canVectorizeBlock(BasicBlock &BB) {
         DEBUG(dbgs() << "LV: Found a non-simple load.\n");
         return false;
       }
-      GetUnderlyingObjects(Ld->getPointerOperand(), Reads, DL);
+
+      Value* Ptr = Ld->getPointerOperand();
+      if (AnalyzedPtrs.insert(Ptr))
+        GetUnderlyingObjects(Ptr, Reads, DL);
     }
 
     // Record store pointers. Abort on all other instructions that write to
@@ -778,7 +782,10 @@ bool LoopVectorizationLegality::canVectorizeBlock(BasicBlock &BB) {
         DEBUG(dbgs() << "LV: Found a non-simple store.\n");
         return false;
       }
-      GetUnderlyingObjects(St->getPointerOperand(), Writes, DL);
+
+      Value* Ptr = St->getPointerOperand();
+      if (AnalyzedPtrs.insert(Ptr))
+        GetUnderlyingObjects(St->getPointerOperand(), Writes, DL);
     }
 
     // We still don't handle functions.
diff --git a/test/Transforms/LoopVectorize/increment.ll b/test/Transforms/LoopVectorize/increment.ll
new file mode 100644
index 00000000000..e944a9af92d
--- /dev/null
+++ b/test/Transforms/LoopVectorize/increment.ll
@@ -0,0 +1,35 @@
+; RUN: opt < %s  -loop-vectorize -dce -instcombine -licm -S | FileCheck %s
+
+target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"
+target triple = "x86_64-apple-macosx10.8.0"
+
+@a = common global [2048 x i32] zeroinitializer, align 16
+
+; This is the loop.
+;  for (i=0; i<n; i++){
+;    a[i] += i;
+;  }
+;CHECK: @inc
+;CHECK: load <4 x i32>
+;CHECK: add <4 x i32>
+;CHECK: store <4 x i32>
+;CHECK: ret void
+define void @inc(i32 %n) nounwind uwtable noinline ssp {
+  %1 = icmp sgt i32 %n, 0
+  br i1 %1, label %.lr.ph, label %._crit_edge
+
+.lr.ph:                                           ; preds = %0, %.lr.ph
+  %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %0 ]
+  %2 = getelementptr inbounds [2048 x i32]* @a, i64 0, i64 %indvars.iv
+  %3 = load i32* %2, align 4
+  %4 = trunc i64 %indvars.iv to i32
+  %5 = add nsw i32 %3, %4
+  store i32 %5, i32* %2, align 4
+  %indvars.iv.next = add i64 %indvars.iv, 1
+  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
+  %exitcond = icmp eq i32 %lftr.wideiv, %n
+  br i1 %exitcond, label %._crit_edge, label %.lr.ph
+
+._crit_edge:                                      ; preds = %.lr.ph, %0
+  ret void
+}