Fix Linear3D Indexing if Iteration Space Rank < 3

The Linear3D launch config setup did not pad its work_items lambda with ones if the iteration space rank was less than three. This MR fixes and tests this behavior.

tests/kernelcreation/test_gpu.py

@@ -99,17 +99,17 @@ def test_indexing_options_2d(
     ast = create_kernel(asm, cfg)
     kernel = ast.compile()
 
-    src_arr = cp.ones((18, 34))
+    src_arr = cp.ones((18, 42))
     dst_arr = cp.zeros_like(src_arr)
 
     if manual_grid:
         match indexing_scheme:
             case "linear3d":
                 kernel.launch_config.block_size = (10, 8, 1)
-                kernel.launch_config.grid_size = (4, 4, 1)
+                kernel.launch_config.grid_size = (4, 2, 1)
             case "blockwise4d":
                 kernel.launch_config.block_size = (40, 1, 1)
-                kernel.launch_config.grid_size = (32, 1, 1)
+                kernel.launch_config.grid_size = (16, 1, 1)
 
     elif indexing_scheme == "linear3d":
         kernel.launch_config.block_size = (10, 8, 1)