Add Cutlass MLA attention backend (#5390)

sgl-kernel/python/sgl_kernel/attention.py

@@ -78,6 +78,7 @@ def cutlass_mla_decode(
     assert len(page_table.shape) == 2
     B_block_table, block_num = page_table.shape
     assert B_block_table == B_q
+    assert block_num > 0, f"block num must be greater than 0, got {block_num}"
     assert block_num % (128 / PAGE_SIZE) == 0
 
     # TODO(kaixih@nvidia): support fp8
@@ -109,6 +110,8 @@ def cutlass_mla_decode(
 def cutlass_mla_get_workspace_size(
     max_seq_len: int, num_batches: int, sm_count: int = 0
 ) -> int:
+    assert max_seq_len > 0, f"max_seq_len must be greater than 0, got {max_seq_len}"
+    assert num_batches > 0, f"num_batches must be greater than 0, got {num_batches}"
     return torch.ops.sgl_kernel.cutlass_mla_get_workspace_size.default(
         max_seq_len, num_batches, sm_count
     )