diff --git a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
--- a/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
+++ b/llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp
@@ -1291,8 +1291,9 @@
       O << ".param .b" << size << " _";
     } else if (isa<PointerType>(retTy)) {
       O << ".param .b" << PtrVT.getSizeInBits() << " _";
-    } else if (retTy->isAggregateType() || retTy->isVectorTy() || retTy->isIntegerTy(128)) {
-      auto &DL = CS.getCalledFunction()->getParent()->getDataLayout();
+    } else if (retTy->isAggregateType() || retTy->isVectorTy() ||
+               retTy->isIntegerTy(128)) {
+      auto &DL = CS.getParent()->getModule()->getDataLayout();
       O << ".param .align " << retAlignment << " .b8 _["
         << DL.getTypeAllocSize(retTy) << "]";
     } else {
diff --git a/llvm/test/CodeGen/NVPTX/bug41651.ll b/llvm/test/CodeGen/NVPTX/bug41651.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/NVPTX/bug41651.ll
@@ -0,0 +1,13 @@
+; RUN: llc -filetype=asm -o - %s | FileCheck %s
+target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
+target triple = "nvptx64-nvidia-cuda"
+
+%func = type { i32 (i32, i32)** }
+
+; CHECK: foo
+; CHECK: call
+; CHECK: ret
+define void @foo() {
+  %call = call %func undef(i32 0, i32 1)
+  ret void
+}