[InstCombine] Transform to undef incorrect atomic unordered mem intrinsics

[skatkov]

Summary

According to LangRef:
If len is not a positive integer multiple of element_size, then the behaviour of the intrinsic is undefined.

Add InstCombine rule to transform intrinsic to undef operation.

This is a follow-up for D76116

[reames]

LGTM

llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp

Instruction *InstCombiner::visitCallInst(CallInst &CI) {
   if (CI.getFunction()->doesNotThrow() && !CI.doesNotThrow()) {
     CI.setDoesNotThrow();
     return &CI;
   }
 
   IntrinsicInst *II = dyn_cast<IntrinsicInst>(&CI);
   if (!II) return visitCallBase(CI);
 
+  // For atomic unordered mem intrinsics if len is not a positive or
+  // not a multiple of element size then behavior is undefined.
+  if (auto *AMI = dyn_cast<AtomicMemIntrinsic>(II))
+    if (ConstantInt *NumBytes = dyn_cast<ConstantInt>(AMI->getLength()))
+      if (NumBytes->getSExtValue() < 0 ||
+          (NumBytes->getZExtValue() % AMI->getElementSizeInBytes() != 0)) {
+        CreateNonTerminatorUnreachable(AMI);
+        assert(AMI->getType()->isVoidTy() &&
+               "non void atomic unordered mem intrinsic");
+        return eraseInstFromFunction(*AMI);
+      }
+
   // Intrinsics cannot occur in an invoke or a callbr, so handle them here
   // instead of in visitCallBase.
   if (auto *MI = dyn_cast<AnyMemIntrinsic>(II)) {
     bool Changed = false;
 
     // memmove/cpy/set of zero bytes is a noop.
     if (Constant *NumBytes = dyn_cast<Constant>(MI->getLength())) {
       if (NumBytes->isNullValue())

llvm/test/Transforms/InstCombine/element-atomic-memintrins.ll

;
   call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 1, i32 1)
   call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 2, i32 2)
   call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 4, i32 4)
   call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 8, i32 8)
   call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 16, i32 16)
   ret void
 }
 
+define void @test_undefined(i8* %dest, i8* %src) {
+; CHECK-LABEL: @test_undefined(
+entry:
+  br i1 undef, label %ok, label %undefined
+undefined:
+; CHECK: undefined:
+; CHECK-NEXT:    store i1 true, i1* undef
+; CHECK-NEXT:    br label %ok
+  call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 7, i32 4)
+  call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 -8, i32 4)
+  call void @llvm.memmove.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 7, i32 4)
+  call void @llvm.memmove.element.unordered.atomic.p0i8.p0i8.i32(i8* align 16 %dest, i8* align 16 %src, i32 -8, i32 4)
+  call void @llvm.memset.element.unordered.atomic.p0i8.i32(i8* align 16 %dest, i8 1, i32 7, i32 4)
+  call void @llvm.memset.element.unordered.atomic.p0i8.i32(i8* align 16 %dest, i8 1, i32 -8, i32 4)
+  br label %ok
+ok:
+  ret void
+}
+
 declare void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i32(i8* nocapture writeonly, i8* nocapture readonly, i32, i32) nounwind argmemonly