[Main2Main] Upgrade vllm commit to 0120

[Meihan-chen]

What this PR does / why we need it?

1. ✅ Upgrade vllm commit to: 0115 (8471b27df97c3eb79f891802fc0e858f8f7ac6a0)
   Modify import paths due to the refactors：
   [Model Runner V2] Refactor Sampler vllm#32245
   [4/N][Attention] Move MLA common to model_executor vllm#32060
   Test result: https://github.com/vllm-project/vllm-ascend/actions/runs/21034239336/job/60490156965?pr=5913
2. ✅Upgrade vllm commit to: 0119 (9a1f16da1e423ede2c2f52a9850cbfbb39cefe96)
   Fix WorkerProc.__init__() missing 1 required positional argument: 'is_driver_worker' due to [TPU][Core] Enable Pipeline Parallelism on TPU backend vllm#28506
   https://github.com/vllm-project/vllm-ascend/actions/runs/21156263050/job/60841668755?5569
3. ✅Upgrade vllm commit to: 0120(148117ea2e689cd43df4be6892671a17cdae5833)
   1. Add skip_compiled param in set_forward_context due to [Core] Whisper support torch.compile vllm#30385
   2. Modify tests/ut/spec_decode/test_eagle_proposer.py due to feat: spec decode with draft models vllm#24322
      change self.max_num_tokens = vllm_config.scheduler_config.max_num_batched_tokens + max_batch_size
   3. Modify UT import paths due to the refactors：[4/N][Attention] Move MLA common to model_executor vllm#32060

Does this PR introduce any user-facing change?

How was this patch tested?

• vLLM version: v0.13.0
• vLLM main: vllm-project/vllm@d682094

[github-actions]

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

• A PR should do only one thing, smaller PRs enable faster reviews.
• Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
• Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.

[gemini-code-assist]

Code Review

This pull request updates the vLLM commit to 0.13.0, necessitating changes to import paths and refactoring of parallel configuration logic. The core changes involve adapting the codebase to the new vLLM library structure, particularly for attention metadata builders and sampling metadata. A significant refactoring in patch_multiproc_executor.py centralizes parallel size calculations and worker initialization, introducing version-aware compatibility. While these changes are essential for the upgrade, some aspects related to type safety, assertion preservation, and conditional logic for version compatibility require attention to ensure robustness and correctness.

[gemini-code-assist]

The use of # type: ignore can mask actual type issues that might arise from API changes in vLLM 0.13.0. It is best practice to resolve type incompatibilities explicitly or provide a more specific reason for ignoring if it's a known, harmless difference, to prevent potential runtime errors.

[gemini-code-assist]

Similar to mla_v1.py, using # type: ignore for MLACommonMetadataBuilder and AttentionCGSupport might be masking actual type issues or subtle API changes in vLLM 0.13.0. It is recommended to address the underlying type problems directly for better code maintainability and to avoid potential runtime bugs.

[gemini-code-assist]

The initialization of self.futures_queue and the call to self._post_init_executor() are moved to this position. It is critical to ensure that _post_init_executor() is called at the correct stage of the executor's initialization, especially if it has side effects or dependencies that must be met before success = True is set.

[gemini-code-assist]

The conditional logic if not vllm_version_is('0.13.0'): implies that is_driver_worker is needed for vLLM versions other than 0.13.0. If is_driver_worker is a new parameter introduced in a vLLM version after 0.13.0, this condition might be incorrect. Please clarify the exact vLLM version(s) for which is_driver_worker is required/not required to ensure proper compatibility.

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.

[github-actions]

This pull request has conflicts, please resolve those before we can evaluate the pull request.