Fix load alignement when unpacking aggregates structs

Summary: Store and loads unpacked by instcombine do not always have the right alignement. This explicitely compute the alignement and set it.

Reviewers: dblaikie, majnemer, reames, hfinkel, joker.eph

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D17326

llvm-svn: 261139

2 files changed, 53 insertions, 12 deletions

diff --git a/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp b/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
index 67f7478ab23..0aa575b37ec 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
@@ -523,16 +523,17 @@ static Instruction *unpackLoadToAggregate(InstCombiner &IC, LoadInst &LI) {
   if (!T->isAggregateType())
     return nullptr;
 
+  auto Name = LI.getName();
   assert(LI.getAlignment() && "Alignment must be set at this point");
 
   if (auto *ST = dyn_cast<StructType>(T)) {
     // If the struct only have one element, we unpack.
-    unsigned Count = ST->getNumElements();
-    if (Count == 1) {
+    auto NumElements = ST->getNumElements();
+    if (NumElements == 1) {
       LoadInst *NewLoad = combineLoadToNewType(IC, LI, ST->getTypeAtIndex(0U),
                                                ".unpack");
       return IC.replaceInstUsesWith(LI, IC.Builder->CreateInsertValue(
-        UndefValue::get(T), NewLoad, 0, LI.getName()));
+        UndefValue::get(T), NewLoad, 0, Name));
     }
 
     // We don't want to break loads with padding here as we'd loose
@@ -542,23 +543,29 @@ static Instruction *unpackLoadToAggregate(InstCombiner &IC, LoadInst &LI) {
     if (SL->hasPadding())
       return nullptr;
 
-    auto Name = LI.getName();
+    auto Align = LI.getAlignment();
+    if (!Align)
+      Align = DL.getABITypeAlignment(ST);
+
     SmallString<16> LoadName = Name;
     LoadName += ".unpack";
     SmallString<16> EltName = Name;
     EltName += ".elt";
+
     auto *Addr = LI.getPointerOperand();
-    Value *V = UndefValue::get(T);
-    auto *IdxType = Type::getInt32Ty(ST->getContext());
+    auto *IdxType = Type::getInt32Ty(T->getContext());
     auto *Zero = ConstantInt::get(IdxType, 0);
-    for (unsigned i = 0; i < Count; i++) {
+
+    Value *V = UndefValue::get(T);
+    for (unsigned i = 0; i < NumElements; i++) {
       Value *Indices[2] = {
         Zero,
         ConstantInt::get(IdxType, i),
       };
-      auto *Ptr = IC.Builder->CreateInBoundsGEP(ST, Addr, makeArrayRef(Indices), EltName);
-      auto *L = IC.Builder->CreateAlignedLoad(Ptr, LI.getAlignment(),
-                                              LoadName);
+      auto *Ptr = IC.Builder->CreateInBoundsGEP(ST, Addr,
+                                            makeArrayRef(Indices), EltName);
+      auto EltAlign = MinAlign(Align, SL->getElementOffset(i));
+      auto *L = IC.Builder->CreateAlignedLoad(Ptr, EltAlign, LoadName);
       V = IC.Builder->CreateInsertValue(V, L, i);
     }
 
@@ -948,11 +955,16 @@ static bool unpackStoreToAggregate(InstCombiner &IC, StoreInst &SI) {
     if (SL->hasPadding())
       return false;
 
+    auto Align = SI.getAlignment();
+    if (!Align)
+      Align = DL.getABITypeAlignment(ST);
+
     SmallString<16> EltName = V->getName();
     EltName += ".elt";
     auto *Addr = SI.getPointerOperand();
     SmallString<16> AddrName = Addr->getName();
     AddrName += ".repack";
+
     auto *IdxType = Type::getInt32Ty(ST->getContext());
     auto *Zero = ConstantInt::get(IdxType, 0);
     for (unsigned i = 0; i < Count; i++) {
@@ -960,9 +972,11 @@ static bool unpackStoreToAggregate(InstCombiner &IC, StoreInst &SI) {
         Zero,
         ConstantInt::get(IdxType, i),
       };
-      auto *Ptr = IC.Builder->CreateInBoundsGEP(ST, Addr, makeArrayRef(Indices), AddrName);
+      auto *Ptr = IC.Builder->CreateInBoundsGEP(ST, Addr,
+                                            makeArrayRef(Indices), AddrName);
       auto *Val = IC.Builder->CreateExtractValue(V, i, EltName);
-      IC.Builder->CreateStore(Val, Ptr);
+      auto EltAlign = MinAlign(Align, SL->getElementOffset(i));
+      IC.Builder->CreateAlignedStore(Val, Ptr, EltAlign);
     }
 
     return true;
diff --git a/llvm/test/Transforms/InstCombine/unpack-fca.ll b/llvm/test/Transforms/InstCombine/unpack-fca.ll
index 435983924b7..bed3b61939d 100644
--- a/llvm/test/Transforms/InstCombine/unpack-fca.ll
+++ b/llvm/test/Transforms/InstCombine/unpack-fca.ll
@@ -151,3 +151,30 @@ define i32 @packed_alignment(%struct.S* dereferenceable(9) %s) {
   %v = extractvalue %struct.T %tv, 1
   ret i32 %v
 }
+
+%struct.U = type {i8, i8, i8, i8, i8, i8, i8, i8, i64}
+
+define void @check_alignment(%struct.U* %u, %struct.U* %v) {
+; CHECK-LABEL: check_alignment
+; CHECK: load i8, i8* {{.*}}, align 8
+; CHECK: load i8, i8* {{.*}}, align 1
+; CHECK: load i8, i8* {{.*}}, align 2
+; CHECK: load i8, i8* {{.*}}, align 1
+; CHECK: load i8, i8* {{.*}}, align 4
+; CHECK: load i8, i8* {{.*}}, align 1
+; CHECK: load i8, i8* {{.*}}, align 2
+; CHECK: load i8, i8* {{.*}}, align 1
+; CHECK: load i64, i64* {{.*}}, align 8
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 8
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 1
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 2
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 1
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 4
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 1
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 2
+; CHECK: store i8 {{.*}}, i8* {{.*}}, align 1
+; CHECK: store i64 {{.*}}, i64* {{.*}}, align 8
+  %1 = load %struct.U, %struct.U* %u
+  store %struct.U %1, %struct.U* %v
+  ret void
+}