[DSE,MSSA] Fix crash when using tryToMergePartialOverlappingStores.

We are re-using tryToMergePartialOverlappingStores, which requires
earlier to domiante Later. In the long run,
tryToMergeParialOverlappingStores should be re-written using MemorySSA.

Fixes PR46513.

2 files changed, 51 insertions, 16 deletions

diff --git a/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp b/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
index 048b67a..261fb387 100644
--- a/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
+++ b/llvm/lib/Transforms/Scalar/DeadStoreElimination.cpp
@@ -2249,22 +2249,27 @@ bool eliminateDeadStoresMemorySSA(Function &F, AliasAnalysis &AA,
         if (EnablePartialStoreMerging && OR == OW_PartialEarlierWithFullLater) {
           auto *Earlier = dyn_cast<StoreInst>(NI);
           auto *Later = dyn_cast<StoreInst>(SI);
-          if (Constant *Merged = tryToMergePartialOverlappingStores(
-                  Earlier, Later, InstWriteOffset, DepWriteOffset, DL, &AA,
-                  &DT)) {
-
-            // Update stored value of earlier store to merged constant.
-            Earlier->setOperand(0, Merged);
-            ++NumModifiedStores;
-            MadeChange = true;
-
-            // Remove later store and remove any outstanding overlap intervals
-            // for the updated store.
-            State.deleteDeadInstruction(Later);
-            auto I = State.IOLs.find(Earlier->getParent());
-            if (I != State.IOLs.end())
-              I->second.erase(Earlier);
-            break;
+          // We are re-using tryToMergePartialOverlappingStores, which requires
+          // Earlier to domiante Later.
+          // TODO: implement tryToMergeParialOverlappingStores using MemorySSA.
+          if (Earlier && Later && DT.dominates(Earlier, Later)) {
+            if (Constant *Merged = tryToMergePartialOverlappingStores(
+                    Earlier, Later, InstWriteOffset, DepWriteOffset, DL, &AA,
+                    &DT)) {
+
+              // Update stored value of earlier store to merged constant.
+              Earlier->setOperand(0, Merged);
+              ++NumModifiedStores;
+              MadeChange = true;
+
+              // Remove later store and remove any outstanding overlap intervals
+              // for the updated store.
+              State.deleteDeadInstruction(Later);
+              auto I = State.IOLs.find(Earlier->getParent());
+              if (I != State.IOLs.end())
+                I->second.erase(Earlier);
+              break;
+            }
           }
         }
 
diff --git a/llvm/test/Transforms/DeadStoreElimination/MSSA/multiblock-overlap.ll b/llvm/test/Transforms/DeadStoreElimination/MSSA/multiblock-overlap.ll
index 5c26566..579ec8e 100644
--- a/llvm/test/Transforms/DeadStoreElimination/MSSA/multiblock-overlap.ll
+++ b/llvm/test/Transforms/DeadStoreElimination/MSSA/multiblock-overlap.ll
@@ -110,3 +110,33 @@ bb9:                                              ; preds = %bb8, %bb7
   store double 6.0, double* %tmp3, align 8
   ret void
 }
+
+; Test case from PR46513. Make sure we do not crash.
+; TODO: we should be able to shorten store i32 844283136, i32* %cast.i32 to a
+; store of i16.
+define void @overlap_no_dominance([4 x i8]* %arg, i1 %c)  {
+; CHECK-LABEL: @overlap_no_dominance(
+; CHECK-NEXT:  bb:
+; CHECK-NEXT:    br i1 [[C:%.*]], label [[BB13:%.*]], label [[BB9:%.*]]
+; CHECK:       bb9:
+; CHECK-NEXT:    [[CAST_I32:%.*]] = bitcast [4 x i8]* [[ARG:%.*]] to i32*
+; CHECK-NEXT:    store i32 844283136, i32* [[CAST_I32]], align 4
+; CHECK-NEXT:    br label [[BB13]]
+; CHECK:       bb13:
+; CHECK-NEXT:    [[CAST_I16:%.*]] = bitcast [4 x i8]* [[ARG]] to i16*
+; CHECK-NEXT:    store i16 0, i16* [[CAST_I16]], align 4
+; CHECK-NEXT:    ret void
+;
+bb:
+  br i1 %c, label %bb13, label %bb9
+
+bb9:                                              ; preds = %bb
+  %cast.i32 = bitcast [4 x i8]* %arg to i32*
+  store i32 844283136, i32* %cast.i32, align 4
+  br label %bb13
+
+bb13:                                             ; preds = %bb9, %bb
+  %cast.i16 = bitcast [4 x i8]* %arg to i16*
+  store i16 0, i16* %cast.i16, align 4
+  ret void
+}