Readd a default lowering rule for cumsum et al.

A previous change removed the only non-constrained lowering rule, breaking lowering for platforms without explicit lowering rules

PiperOrigin-RevId: 633297839

jax/_src/lax/control_flow/loops.py

@@ -2449,14 +2449,14 @@ def register_lowering(fn, platform=None):
 
   if xla_extension_version >= 263:
     # In XLA, there's a rewriter for an O(N^2) reduce-window implementation.
+    # TODO(https://github.com/llvm/llvm-project/issues/91883): enable rewrite
+    # for CPU once the vectorizer crash is fixed..
     for platform in ['cuda', 'rocm', 'tpu']:
       register_lowering(
           partial(cumred_reduce_window_impl, reduce_window_fn), platform
       )
 
-    # TODO(https://github.com/llvm/llvm-project/issues/91883) Re-enable rewrite
-    # for CPU once the vectorizer crash is fixed..
-    register_lowering(partial(associative_scan, reduce_fn), 'cpu')
+    register_lowering(partial(associative_scan, reduce_fn))
   else:
     # Older XLA versions only have this rewrite for TPU.
     register_lowering(