[4.0 -> main] P2P excessive logging caused by message spamming

[heifner]

Two main issues resolved by this PR in 4.0.

• "Infinite" loop of notice_message <-> handshake_message when clock skew. The clock skew caused peer to keep trying to inform other that it has the blocks already.
• Handshakes were being sent to every peer anytime any peer connection was closed. This was introduced in 4.0.1 by [4.0] Correctly handle close while syncing #1171. This PR handles this case better by resetting sync_next_expected_num on rejected block instead of spamming all connections with handshakes.

These issues were already fixed in main by work on #1072. (hence the zero diff)

Additional fixes in this 4.0 PR backported from #1072 work:

• Includes fix for using nanoseconds resolution for time_message.
• Update known head block on catch_up notice_message. The information is provided might as well use it instead of waiting for a handshake_message.

Merges release/4.0 into main including #1267.

Resolves #1240