Store prefetch results in LFC cache once as soon as they are received

[knizhnik]

Problem

Prefetch is performed locally, so different backers can request the same pages form PS.
Such duplicated request increase load of page server and network traffic.
Making prefetch global seems to be very difficult and undesirable, because different queries can access chunks on different speed. Storing prefetch chunks in LFC will not completely eliminate duplicates, but can minimise such requests.

The problem with storing prefetch result in LFC is that in this case page is not protected by share buffer lock.
So we will have to perform extra synchronisation at LFC side.

See:
https://neondb.slack.com/archives/C0875PUD0LC/p1736772890602029?thread_ts=1736762541.116949&cid=C0875PUD0LC

@MMeent implementation of prewarm:
See #10312

Summary of changes

Use conditional variables to sycnhronize access to LFC entry.

[github-actions]

7385 tests run: 7005 passed, 0 failed, 380 skipped (full report)

---

Flaky tests (3)

Postgres 17

• test_metrics_normal_work: release-x86-64, release-arm64

Postgres 16

• test_metrics_normal_work: release-arm64

Code coverage* (full report)

• functions: 33.6% (8445 of 25104 functions)
• lines: 49.2% (70717 of 143866 lines)

* collected from Rust tests only

---

_{The comment gets automatically updated with the latest test results
071142f at 2025-01-20T20:34:33.613Z :recycle:}

[erikgrinaker]

Thanks! I've only done a high-level review, but the approach makes sense to me. I expect the synchronization overhead would be negligible since this is on an IO path, but would be interested to see how benchmarks are affected.