netvsp & net_mana - instrument VTL0 & host calls at risk of hanging

[ben-zen]

There's a bunch of async operations with long or no timeout which are at risk of hanging the NetVSP code; adding spans to those actions makes them more obvious in tracing, with additional context to track down miscreant calls or make a misbehaving VM more obvious.

[Copilot]

Pull request overview

This PR improves observability around NetVSP and MANA endpoint operations that can take a long time or potentially hang, by adding tracing spans with relevant identifiers (instance/vport/VF IDs). This helps correlate stalls in async shutdown/reconfiguration paths when debugging misbehaving VMs or host/VTL0 interactions.

Changes:

• Add an info_span around NetVSP coordinator endpoint shutdown.
• Add info_span instrumentation around MANA endpoint teardown steps (RX disable, vport resource destruction, queue drain wait).
• Add info_span instrumentation around Underhill VTL0 VF offer/remove notification calls, plus an explicit log when VTL0 bus isn’t present.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
vm/devices/net/netvsp/src/lib.rs	Instruments coordinator endpoint stop().await with an info span including instance_id.
vm/devices/net/net_mana/src/lib.rs	Adds tracing::Instrument and spans around endpoint stop/cleanup operations with vport_id.
openhcl/underhill_core/src/emuplat/netvsp.rs	Adds spans around VTL0 VF removal notice and VF offer call to better pinpoint stalls and add context.

[erfrimod]

Added a comment to trace one more thing, but otherwise ready to go. :)