[App/Performance] Porting masstree analytics & Major performance improvements

[crazyboycjr]

This PR mainly includes the following changes.

Applications

• Adding a masstree_analytics example. This also introduces a git submodule at src/koala_examples/masstree_analytics/
• Supporting enum and default values in mrpc-derive

Bugfix

• Fix bug of setting up concurrent connections
• Resource recycling issue (a persistent RPC server finally can shut down connections and return the resources (e.g., TCP ports, shared memory regions))
• Fix bug of latency counting in masstree analytics.

Performance improvement

• Changed the implementation of ResourceTable change to DashMap.
• Added an EngineGroup. With EngineGroup, runtime can guarantee some engines are always bundled together. This maybe renamed as EnginePackage in the future.
• Added a sequential queue, so that the communication between packaged engines barely have no cost.
• Frontend and backend query_app_addr and query_backend_addr are no-op.

Miscellaneous

• Fix the issue of bad performance in runtime when energy saving is on. Now even we don't spin the CPU, the performance shouldn't be affected.
• In the RPC API, use can not associate a token so that when the reply returns, the user can associate the reply with its corresponding request.
• Improved the concurrent hashmap benchmark to test them on multiple threads. (Leapfrog to test)
• Improved the rdma-write benchmark to support running it with multiple QPs, multiple threads, and multiple processes.
• Benchmark launcher can output to stdout and file simultaneously.
• Benchmark launcher can support launching workers as a DAG.
• Adding a feature in userland mrpc crate called timing. This optional feature is disabled by default.
• Run cargo udeps and removed some unused dependencies to speed up compilation.

Related issue number

Checks

• [ ] I've run scripts/clippy.sh to lint the changes in this PR.
• [ ] I've included any doc changes.
• Testing Strategy
  • [ ] Release tests
  • [ ] This PR is not tested :(

[crazyboycjr]

@libertyeagle This RP is ready to merge. It does have several conflicts with the upgrade branch, but they should be orthogonal. I'll start to cherry-pick features in the upgrade branch and gradually push them into main.