LoopVersioning: Respect convergent

This changes the standalone pass only. Arguably the utility class
itself should assert there are no convergent calls. However, a target
pass with additional context may still be able to version a loop if
all of the dynamic conditions are sufficiently uniform.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363165 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Transforms/Utils/LoopVersioning.cpp b/lib/Transforms/Utils/LoopVersioning.cpp
index d59076c2a9f2ae04a4ea13f09f3f41c87dd90a4c..a9a480a4b7f9c552a71bf0422de8436bc1637548 100644 (file)
--- a/lib/Transforms/Utils/LoopVersioning.cpp
+++ b/lib/Transforms/Utils/LoopVersioning.cpp
@@ -280,8 +280,9 @@ public:
     bool Changed = false;
     for (Loop *L : Worklist) {
       const LoopAccessInfo &LAI = LAA->getInfo(L);
-      if (L->isLoopSimplifyForm() && (LAI.getNumRuntimePointerChecks() ||
-          !LAI.getPSE().getUnionPredicate().isAlwaysTrue())) {
+      if (L->isLoopSimplifyForm() && !LAI.hasConvergentOp() &&
+          (LAI.getNumRuntimePointerChecks() ||
+           !LAI.getPSE().getUnionPredicate().isAlwaysTrue())) {
         LoopVersioning LVer(LAI, L, LI, DT, SE);
         LVer.versionLoop();
         LVer.annotateLoopWithNoAlias();

diff --git a/test/Transforms/LoopVersioning/convergent.ll b/test/Transforms/LoopVersioning/convergent.ll
new file mode 100644 (file)
index 0000000..48c2725
--- /dev/null
+++ b/test/Transforms/LoopVersioning/convergent.ll
@@ -0,0 +1,40 @@
+; RUN: opt -basicaa -loop-versioning -S < %s | FileCheck %s
+
+target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
+
+; Do not version this loop because of a convergent operation
+
+; CHECK-LABEL: @f(
+; CHECK: call i32 @llvm.convergent(
+; CHECK-NOT: call i32 @llvm.convergent(
+define void @f(i32* %a, i32* %b, i32* %c) #0 {
+entry:
+  br label %for.body
+
+for.body:                                         ; preds = %for.body, %entry
+  %ind = phi i64 [ 0, %entry ], [ %add, %for.body ]
+
+  %arrayidxA = getelementptr inbounds i32, i32* %a, i64 %ind
+  %loadA = load i32, i32* %arrayidxA, align 4
+
+  %arrayidxB = getelementptr inbounds i32, i32* %b, i64 %ind
+  %loadB = load i32, i32* %arrayidxB, align 4
+  %convergentB = call i32 @llvm.convergent(i32 %loadB)
+
+  %mulC = mul i32 %loadA, %convergentB
+
+  %arrayidxC = getelementptr inbounds i32, i32* %c, i64 %ind
+  store i32 %mulC, i32* %arrayidxC, align 4
+
+  %add = add nuw nsw i64 %ind, 1
+  %exitcond = icmp eq i64 %add, 20
+  br i1 %exitcond, label %for.end, label %for.body
+
+for.end:                                          ; preds = %for.body
+  ret void
+}
+
+declare i32 @llvm.convergent(i32) #1
+
+attributes #0 = { nounwind convergent }
+attributes #1 = { nounwind readnone convergent }