[Feature] Support to use fullgraph with eagle (#5118)

### What this PR does / why we need it?
    
We support to use full graph with eagle. 

Change list:
1. Distinguish between processing graph_params and draft_graph_params in
attention_v1.
    2. Adapt the full-graph mode in eagle_proposer, include:
        1). If use full graph, make Fullgraph Wrapper when load model.
2). Build a new meatadata, set running mode in FULL and mark attention
update in dummy_run when in Fullgraph mode.
3). Fixed and fill any attn_metadata, such as
attn_metadata.slot_mapping.
        4). Add a descriptor.
        5). Set running mode and triggered update metadata.
3. Trans is_mtp_model to is_draft_model, and add the update of
workspace.

NOTE:
When set async_scheduling=True, the draft model will enforce execution
in eager mode.

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: anon189Ty <Stari_Falcon@outlook.com>
Co-authored-by: Yizhou Liu <liu_yizhou@outlook.com>
Co-authored-by: Yizhou <136800916+yiz-liu@users.noreply.github.com>

vllm_ascend/worker/model_runner_v1.py

@@ -89,8 +89,8 @@ from vllm_ascend.attention.utils import (AscendCommonAttentionMetadata,
 # yapf conflicts with isort for this block
 # yapf: disable
 from vllm_ascend.compilation.acl_graph import (ACLGraphWrapper,
+                                               set_draft_graph_params,
                                                set_graph_params,
-                                               set_mtp_graph_params,
                                                update_attn_dcp_pcp_params,
                                                update_attn_params,
                                                update_mla_attn_dcp_pcp_params,
@@ -1104,7 +1104,8 @@ class NPUModelRunner(GPUModelRunner):
                 self.spec_decode_common_attn_metadata is None:
                 self.spec_decode_common_attn_metadata = common_attn_metadata
                 if self.speculative_config.method in ("eagle", "eagle3") and \
-                        self.vllm_config.compilation_config.cudagraph_mode.has_full_cudagraphs():
+                        (self.vllm_config.speculative_config.enforce_eager \
+                         or self.use_async_scheduling):
                     self.spec_decode_common_attn_metadata = \
                         self.spec_decode_common_attn_metadata.unpadded(
                             total_num_scheduled_tokens, base_num_reqs)
@@ -2916,7 +2917,7 @@ class NPUModelRunner(GPUModelRunner):
         # we set the graph params right before initializing the keys.
         set_graph_params(self.cudagraph_batch_sizes)
         if self.speculative_config:
-            set_mtp_graph_params(self.cudagraph_batch_sizes)
+            set_draft_graph_params(self.cudagraph_batch_sizes)
 
         self.cudagraph_dispatcher.initialize_cudagraph_keys(
             self.compilation_config.cudagraph_mode,