Service port forwarding and is_connected() periodically during traffic

[edi-oai]

Problem

Continuous traffic could prevent periodic port-forwarding maintenance from running.

Conditions like:

• !vm_readable && !host_readable && !interrupt
• self.events.is_empty()

...only detect an idle poll wake, not whether the timeout interval has elapsed.

Additionally, we run is_connected() for each loop cycle, which might be too much.

Solution

Track the periodic interval with coarsetime::Instant in Poller and expose periodic_tick_due().

Port forwarding maintenance and is_connected() check now runs when the interval elapses, regardless of poll readiness events.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 61d45931bf

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Run periodic maintenance inside unbounded drain loops

When VM or host traffic arrives continuously enough that read_from_vm never observes WouldBlock, or read_from_host never observes VmnetReadNothing, these unbounded drain loops never return to this check. Since this remains the only port_forwarder.tick call in the repository, forwarding installation, updates, and lease-expiry cleanup can still be postponed indefinitely under the sustained-traffic condition this change is intended to handle; check the deadline while draining packets or otherwise bound each drain iteration.

Useful? React with 👍 / 👎.

[edi-oai]

Let's handle this separately in #199.

Again, this seems out-of-scope because this PR didn't introduce this issue in the first place.