Revert "drop ascend scheduler" (#4580)

Reverts vllm-project/vllm-ascend#4498
- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2

vllm_ascend/worker/model_runner_v1.py

@@ -330,6 +330,10 @@ class NPUModelRunner(LoRAModelRunnerMixin):
 
         # Ascend-specific configurations
         self.ascend_config = get_ascend_config()
+        if self.ascend_config.ascend_scheduler_config.enabled:
+            self.chunked_prefill_enabled = self.scheduler_config.chunked_prefill_enabled
+        else:
+            self.chunked_prefill_enabled = True
         self.weight_prefetch_method = WeightPrefetchMethod(
             self.ascend_config.weight_prefetch_config)
         # Dump / PrecisionDebugger configuration now comes from AscendConfig
@@ -1938,6 +1942,7 @@ class NPUModelRunner(LoRAModelRunnerMixin):
 
     def _build_attn_state(self, num_reqs, num_scheduled_tokens,
                           num_valid_tokens):
+        ascend_config = get_ascend_config()
         if np.array_equal(self.seq_lens_np[:num_reqs], num_scheduled_tokens):
             attn_state = AscendAttentionState.PrefillNoCache
         # We assume it is the decode stage, where prefill occurs but only one token is not hit in cache.
@@ -1954,7 +1959,7 @@ class NPUModelRunner(LoRAModelRunnerMixin):
             else:
                 attn_state = AscendAttentionState.ChunkedPrefill
         # splitfuse
-        elif self.scheduler_config.enable_chunked_prefill:
+        elif not ascend_config.ascend_scheduler_config.enabled or self.chunked_prefill_enabled:
             attn_state = AscendAttentionState.ChunkedPrefill
         else:
             attn_state = AscendAttentionState.PrefillCacheHit