feat: support glm5.2 dsa sharing, cache elision and mooncake pd.

[phantomlei3]

Description

Add GLM-5.2 DSA cross-layer Top-K sharing with per-layer indexer cache elision and accurate cache estimation. Refactor Mooncake KV cache allocation and transfer to support heterogeneous layer layouts and safer RDMA registration

  - GLM-5.2 与 GLM-5.0/5.1 继续共用 glm_moe_dsa 类型；没有共享层时保持原 GLM-5 行为。
  - 检测 Prefill CP 与 DSA 跨层共享的配置冲突。
  - 按层生成 Indexer Cache 启用掩码，共享层不再分配 Indexer Cache 和 Scale。
  - 修正 KV Cache 容量估算，使其按照实际拥有 Indexer Cache 的层数计算显存。
  - Mooncake 支持每层不同数量的缓存 Buffer，并根据实际偏移生成传输 Buffer ID。
  - 补充缓存估算、混合层缓存分配和 Mooncake 变长布局测试。
  - 新增 KVCacheTensorAllocator 抽象，将缓存结构决策留在 KV Cache 工厂中，将物理内存分配策略独立出来。
  - MLU Mooncake 使用专用分配器，仅对 INDEX_SCALE 分配满足 RDMA 注册长度要求的内存。
  - Mooncake 不再自行构造 MLU KV Cache，而是通过 Worker/KV Cache 工厂统一创建。

Change Type

• Bug fix
• New feature
• Performance improvement
• Refactor
• Documentation
• Test
• Build or CI

Pull Request Checklist

Thank you for contributing to xLLM. Before requesting review, please make sure the following items are complete.

PR Title and Commit Messages

• The PR title and each commit message follow the xLLM commit format: <type>: <subject>.

Allowed types: feat, bugfix, docs, test, refactor, chore, style, revert, perf, model, build, release.
The subject should use clear English, start with a verb, include at least 4 words, and end with ..

Pre-commit Checks

• I have installed pre-commit by running pip install pre-commit or an equivalent command.
• I have installed the hooks with pre-commit install.
• I have run pre-commit run --all-files and fixed any reported issues.

If you are unsure how to set up pre-commit, see the pre-commit documentation.

Self Review

• I have self-reviewed the code according to .agents/skills/code-review/references/custom-code-style.md, especially code written or assisted by AI.
• I have rebased this PR onto the latest main branch.

Build and Test Coverage

• Tests have been added or updated as needed.
• CUDA: python setup.py build test has passed on a CUDA machine.
• NPU: python setup.py build test has passed on an NPU machine.
• MLU: python setup.py build test has passed on an MLU machine.

Reviewer Notes

[gemini-code-assist]

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.