feat(vm): configure gpu settings through gateway vm config

[cheese-head]

Summary

Extends the gateway VM driver config so [openshell.drivers.vm] can express the existing GPU enablement and GPU sizing knobs. The gateway now resolves those VM settings and passes explicit arguments to the openshell-driver-vm subprocess, keeping VM configuration in the gateway TOML flow.

Related Issue

N/A

Changes

• Added gpu_enabled, gpu_mem_mib, and gpu_vcpus to VmComputeConfig.
• Passed resolved GPU VM settings from the gateway to the VM driver subprocess.
• Preserved legacy OPENSHELL_VM_* environment overrides when spawning the VM driver, while preventing the child process from re-reading stale config env values.
• Added VM config and subprocess argv tests.
• Updated the gateway config reference MicroVM example.

Testing

• mise run pre-commit passes
• Unit tests added/updated
• E2E tests added/updated (if applicable)

Manual verification run:

cargo fmt -p openshell-server
cargo test -p openshell-server vm_
cargo clippy -p openshell-server --all-targets -- -D warnings
cargo test -p openshell-server
git diff --check

Checklist

• Follows Conventional Commits
• Commits are signed off (DCO)
• Architecture docs updated (if applicable)

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[elezar]

Do these need their own section?

[elezar]

Are these setting VM-specific given the guest_ prefix?

[elezar]

What does an empty image imply? Should we add a comment?

[elezar]

Out of scope for this change, but why is this not just log_level?