Continuous incoming connections recreate the one-second sleep in run_accepted_task_runner, so completed handlers are only reaped after a one-second gap in accepts. Busy public nodes can therefore accumulate completed Tokio tasks and JoinSet entries even when their established connection count stays stable.
Keep a persistent interval outside the select loop so incoming accepts cannot reset the reaping deadline. Add a paused-clock regression test that accepts an immediately completing handler every 100 ms, checks that completed entries remain bounded, and verifies shutdown drains the set. Enable Tokio's test clock utilities only for development builds.
Co-authored-by: 225284228a-droid <239500008+225284228a-droid@users.noreply.github.com>
Fixes the DHCP allocator treating its hard-coded fallback
subnet as an interface lease when no peer IPv4 has been
observed. With DHCP enabled and no devices carrying a
virtual IPv4, the allocator now waits instead of pulling
10.126.126.0/24 onto the TUN interface.
Behavior preserved when an explicit allocator subnet is
supplied or when peers with assigned IPv4 exist.
---------
Co-authored-by: 225284228a-droid <225284228a-droid@users.noreply.github.com>
Co-authored-by: Codex <codex@users.noreply.github.com>
