Reuse shared batch read buffers

[bjmeetsfo]

Summary

• add get_shared_batch_into so retrieval callers can reuse a caller-owned result buffer
• keep get_shared_batch behavior and shared-buffer accounting intact
• switch the batch read benchmark to the reusable-buffer path and add regression coverage for capacity preservation

Validation

• cargo fmt -- --check
• cargo check --locked --no-default-features --examples -q
• cargo test --locked --no-default-features "get_shared_batch_into" -- --test-threads=1
• cargo test --locked --no-default-features -q
• python3 -m py_compile tools/validate_backend_report.py tools/compare_backend_reports.py tools/emit_backend_operator_log.py tools/gen_metrics.py tools/validate_grafana_dashboard.py
• python3 tools/validate_grafana_dashboard.py
• git diff --check
• cargo run --release --no-default-features --example batch_read_cost

Benchmark

4096 resident values of 64 bytes, memory hits only, median of 9 passes:

• batch 64: get_batch 1292.6 ns/key, shared_batch 966.8 ns/key
• batch 1024: get_batch 1564.2 ns/key, shared_batch 1489.1 ns/key