[misc][torchair] fix bugs around deepseek mtp, enable_shared_expert_dp and use_cached_kv_cache_bytes (#3074)

### What this PR does / why we need it?
This miscellaneous​ contains several small fixes:
1) fix initialization and forward bugs of DeepseekMTPLayer with
`shared_expert_dp` enabled.
2) fix a tensor shape mismatches after o_proj caused by a work-aroud
change in NPUModelRunner.
3) avoid unnecessary decline of kv_cache memory (default: 64MB) with
`use_cached_kv_cache_bytes` disabled.
4) fall back `fused_moe_state` from `MC2` to `All2All` since the padding
logic of `mc2_mask` is incompatible with input hidden_states when
`shared_expert_dp` enabled.

Once this PR is merged, users can launch disaggregated_prefill
deployments (large_ep) with `deepseek_mtp` and `shared_expert_dp` as
`v0.9.1-dev` branch. The remaining problem of kv_cache tokens decline
compared to `v0.9.1-dev` will be resolved by
https://github.com/vllm-project/vllm-ascend/pull/3073.
 
### Does this PR introduce _any_ user-facing change?

No.
### How was this patch tested?
E2E vllm serving about deepseek_mtp with torchair graph mode and
`enable_shared_expert_dp` with eager mode. Large ep deployments are also
tested with this PR.


- vLLM version: v0.10.2
- vLLM main:
5aeb925452

---------

Signed-off-by: linfeng-yuan <1102311262@qq.com>

vllm_ascend/torchair/torchair_model_runner.py

@@ -87,8 +87,8 @@ class NPUTorchairModelRunner(NPUModelRunner):
     ) -> tuple[int, Optional[torch.Tensor], bool, bool]:
         """Override from NPUModelRunner to pad num_tokens"""
         if self.enable_shared_expert_dp:
-            return super()._sync_metadata_across_dp(num_tokens, with_prefill,
-                                                    enable_dbo)
+            # Padding is not required for shared_expert_dp cases in eager mode.
+            return num_tokens, None, with_prefill, enable_dbo
         if self.dp_size == 1:
             if not with_prefill:
                 maybe_padded_num_tokens = self.select_torchair_padded_batch_size(