Add tests for the NVPTXLowerAggrCopies pass.

Note: not testing memmove lowering for now, as it's broken
[see https://llvm.org/bugs/show_bug.cgi?id=24056]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@241736 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/test/CodeGen/NVPTX/lower-aggr-copies.ll b/test/CodeGen/NVPTX/lower-aggr-copies.ll
new file mode 100644 (file)
index 0000000..ff706b5
--- /dev/null
+++ b/test/CodeGen/NVPTX/lower-aggr-copies.ll
@@ -0,0 +1,34 @@
+; RUN: llc < %s -march=nvptx -mcpu=sm_35 | FileCheck %s
+
+; Verify that the NVPTXLowerAggrCopies pass works as expected - calls to
+; llvm.mem* intrinsics get lowered to loops.
+
+declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture readonly, i64, i32, i1) #1
+declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) #1
+
+define i8* @memcpy_caller(i8* %dst, i8* %src, i64 %n) #0 {
+entry:
+  tail call void @llvm.memcpy.p0i8.p0i8.i64(i8* %dst, i8* %src, i64 %n, i32 1, i1 false)
+  ret i8* %dst
+; CHECK-LABEL: .visible .func (.param .b32 func_retval0) memcpy_caller
+; CHECK: LBB[[LABEL:[_0-9]+]]:
+; CHECK:      ld.u8 %rs[[REG:[0-9]+]]
+; CHECK:      st.u8 [%r{{[0-9]+}}], %rs[[REG]]
+; CHECK:      add.s64 %rd[[COUNTER:[0-9]+]], %rd[[COUNTER]], 1
+; CHECK-NEXT: setp.lt.u64 %p[[PRED:[0-9]+]], %rd[[COUNTER]], %rd
+; CHECK-NEXT: @%p[[PRED]] bra LBB[[LABEL]]
+}
+
+define i8* @memset_caller(i8* %dst, i32 %c, i64 %n) #0 {
+entry:
+  %0 = trunc i32 %c to i8
+  tail call void @llvm.memset.p0i8.i64(i8* %dst, i8 %0, i64 %n, i32 1, i1 false)
+  ret i8* %dst
+; CHECK-LABEL: .visible .func (.param .b32 func_retval0) memset_caller(
+; CHECK:      ld.param.u8 %rs[[REG:[0-9]+]]
+; CHECK:      LBB[[LABEL:[_0-9]+]]:
+; CHECK:      st.u8 [%r{{[0-9]+}}], %rs[[REG]]
+; CHECK:      add.s64 %rd[[COUNTER:[0-9]+]], %rd[[COUNTER]], 1
+; CHECK-NEXT: setp.lt.u64 %p[[PRED:[0-9]+]], %rd[[COUNTER]], %rd
+; CHECK-NEXT: @%p[[PRED]] bra LBB[[LABEL]]
+}