[Feature] support eager mode in model runner v2 (#5210)

### What this PR does / why we need it?
#5051 only implement a basic framework for model runner v2, but there
are still some bugs for e2e functionality, this PR aim to enable basic
functionality.
model runner v2 plans:
https://github.com/vllm-project/vllm-ascend/issues/5208

- vLLM version: release/v0.13.0
- vLLM main:
ad32e3e19c
---------
Signed-off-by: Ronald1995 <ronaldautomobile@163.com>

vllm_ascend/ops/register_custom_ops.py

@@ -81,7 +81,7 @@ def _maybe_pad_and_reduce_impl(x: torch.Tensor,
     except AssertionError:
         return tensor_model_parallel_all_reduce(x)
 
-    if not forward_context.sp_enabled:
+    if not getattr(forward_context, "sp_enabled", False):
         return tensor_model_parallel_all_reduce(x)
 
     dp_metadata = forward_context.dp_metadata