feat(p2p-07): add read acceleration baseline (#671)

[jlon]

Summary

• Add the first feature-flagged P2P read path for remote block reads.
• Preserve worker fallback and local chunk caching so disabled mode and miss handling stay correct.

Design

• Key read reuse by file_id + version_epoch + block_id + off to avoid stale cross-version hits.
• Probe local cache first, then P2P, then worker reads, with single-flight protection around shared chunk fetches.

Implementation

• Thread file identity and mtime into BlockReader and FsReaderBase.
• Publish worker-fetched chunks into P2pService, populate local read cache, and add registry/cache integration tests.

Verification

• cargo test -p curvine-client worker_chunk_read_populates_local_cache_and_p2p_registry
• cargo test -p curvine-client running_services_share_published_chunks_and_cleanup_on_stop

[jlon]

Part 7/17 of the split P2P series that supersedes #671.

This PR targets , so the visible diff is cumulative. Review the slice described in this PR body instead of treating the whole diff as new.

Intended review delta is p2p-07 on top of #757.

[jlon]

Review note: this is part 7/17 of the split P2P series that supersedes #671.

This PR targets main, so the visible diff is cumulative. Please review the slice described in this PR body.

Review after #757.