Merging r195599:

------------------------------------------------------------------------
r195599 | void | 2013-11-24 21:01:21 -0800 (Sun, 24 Nov 2013) | 4 lines

Don't look past volatile loads.

A volatile load should block us from trying to coalesce stores.
PR18023
------------------------------------------------------------------------

git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_34@195600 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 20b0981..e0732fa 100644 (file)
--- a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -8703,6 +8703,11 @@ bool DAGCombiner::MergeConsecutiveStores(StoreSDNode* St) {
         Index = STn;
         break;
       } else if (LoadSDNode *Ldn = dyn_cast<LoadSDNode>(NextInChain)) {
+        if (Ldn->isVolatile()) {
+          Index = NULL;
+          break;
+        }
+
         // Save the load node for later. Continue the scan.
         AliasLoadNodes.push_back(Ldn);
         NextInChain = Ldn->getChain().getNode();

diff --git a/test/CodeGen/X86/pr18023.ll b/test/CodeGen/X86/pr18023.ll
new file mode 100644 (file)
index 0000000..1b3cf24
--- /dev/null
+++ b/test/CodeGen/X86/pr18023.ll
@@ -0,0 +1,29 @@
+; RUN: llc < %s -mtriple x86_64-apple-macosx10.9.0 | FileCheck %s
+; PR18023
+
+; CHECK: movl $1, 4(%rax)
+; CHECK: movl $1, 4(%rax)
+; CHECK: movl $1, 4(%rax)
+
+@c = common global i32 0, align 4
+@a = common global [3 x i32] zeroinitializer, align 4
+@b = common global i32 0, align 4
+@.str = private unnamed_addr constant [4 x i8] c"%d\0A\00", align 1
+
+define void @func() {
+  store i32 1, i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 1), align 4
+  store i32 0, i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 0), align 4
+  %1 = load volatile i32* @b, align 4
+  store i32 1, i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 1), align 4
+  store i32 0, i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 1), align 4
+  %2 = load volatile i32* @b, align 4
+  store i32 1, i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 1), align 4
+  store i32 0, i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 2), align 4
+  %3 = load volatile i32* @b, align 4
+  store i32 3, i32* @c, align 4
+  %4 = load i32* getelementptr inbounds ([3 x i32]* @a, i64 0, i64 1), align 4
+  %call = call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i64 0, i64 0), i32 %4)
+  ret void
+}
+
+declare i32 @printf(i8*, ...)