BrianZill

I'm not sure what is meant by this. But we should have a proper IP layer that handles datalink layer address resolution without exposing anything about it to TCP. And...

> @BrianZill if this is related to `runtime` may I transfer the issue to that repo? This issue is cross-repo. I created https://github.com/demikernel/runtime/issues/17 to cover the support needed over there...

This is good to keep around to track this issue. This should stay low-priority for now, as I think most other stacks behave similarly, barring special security settings to prevent...

See also Issue demikernel/demikernel#220.

The description for this bug seems to be talking about _sending_ RST packets, whereas my PR demikernel/inetstack#123 partially fixed the handling of _received_ RST packets. The send side of this...

PR demikernel/inetstack#135 removed all the closer closures, which at least partially helped with "Slim down TCP background workers".

Related Issue #173 extends the desired functionality further.

We also need to support splitting an existing buffer into two (preferably without copying any of the actual data).

I think we need to restructure most of this to allow things to cleanup properly. And rename TcpPeer to something that makes sense.

I removed several of the original "missing functionalities" from this list because they have their own Issue for them: TCP Timestamps is covered by Issue #182. Delayed ACKs is already...