[OpenMP] Add a test for D158802 (#70678)

In D158802 we honored user's `thread_limit` value even with the
optimization
introduced in D152014. This patch adds a simple test.

Author: shiltian

openmp/libomptarget/test/offloading/small_trip_count_thread_limit.cpp

@@ -0,0 +1,31 @@
+// clang-format off
+// RUN: %libomptarget-compilexx-generic
+// RUN: env LIBOMPTARGET_INFO=16 \
+// RUN:   %libomptarget-run-generic 2>&1 | %fcheck-generic
+
+// UNSUPPORTED: aarch64-unknown-linux-gnu
+// UNSUPPORTED: aarch64-unknown-linux-gnu-LTO
+// UNSUPPORTED: x86_64-pc-linux-gnu
+// UNSUPPORTED: x86_64-pc-linux-gnu-LTO
+
+int main(int argc, char *argv[]) {
+  constexpr const int block_size = 256;
+  constexpr const int grid_size = 4;
+  constexpr const int count = block_size * grid_size;
+
+  int *data = new int[count];
+
+#pragma omp target teams distribute parallel for thread_limit(block_size) map(from: data[0:count])
+  for (int i = 0; i < count; ++i)
+    data[i] = i;
+
+  for (int i = 0; i < count; ++i)
+    if (data[i] != i)
+      return 1;
+
+  delete[] data;
+
+  return 0;
+}
+
+// CHECK: Launching kernel {{.*}} with 4 blocks and 256 threads in SPMD mode