Integrate vendor VFIO vGPUs into the instance lifecycle

[yummybomb]

Summary

Top layer of the vendor VFIO vGPU stack (generalize-vgpu-device #322 ← hypervisor-liveness #363 ← vendor-vfio-backend #364 ← this). The backend itself is in #364; this layer wires it into the instance lifecycle and adds the release guards that make reusable-VF assignments safe:

• Live-instance claim guard on release — an assignment goes stale when its release succeeds but the metadata save does not (or start fails between the release and its first save). Since vendor VFIO reuses the same VF path across assignments, honoring a stale claim could clear a VF out from under the live instance that owns it now. Every release consults live instance metadata (liveness verified via Harden hypervisor process liveness checks #363's socket-ownership check) and drops stale metadata without touching the device.
• Assignment durability — assignments are tagged with the owning instance ID, persisted before booting a started instance without carrying over the previous VMM identity, and retained in metadata when rollback release fails in create/start so later release paths can still find the device.
• Fail-closed startup reconciliation — the protected set of live VFs is built from ListInstancesForReconcile, which fails on any unreadable metadata instead of silently skipping it. Recent assignments receive bounded startup protection when their PID is absent or stale; when the inventory is unavailable, vendor VFIO reconciliation is skipped entirely while mdev reconciliation still runs.
• Hypervisor scope — hypervisor selection remains a caller policy, preserving the framework layer's existing mdev behavior. The production rollout continues to place vGPU instances on QEMU.

Testing

• go build ./..., go vet ./... clean
• go test -race passes for lib/instances targeted suites, lib/devices, lib/resources, lib/builds (TestCreateInstanceWithNetwork needs image pulls + iptables and fails in this environment on the unmodified base as well)

---

Note

High Risk
Changes GPU assignment/release, startup reconciliation, and fail-closed metadata scans on paths shared by live VMs; incorrect liveness or protection logic could drop or retain the wrong VF assignment.

Overview
Wires vendor VFIO into create/start/stop/delete (replacing the previous “not integrated” gate) and treats vGPU assignments as durable host state tracked with GPUAssignedAt, instance-scoped create/destroy hooks, and metadata retention when rollback release fails.

Release safety for reused VF paths: vendor VFIO releases scan other instances’ metadata and verify hypervisor liveness before destroying a device; stale claims are dropped without touching hardware, while ambiguous or unreadable inventory fails closed so the requester keeps its assignment. Mdev releases skip that scan.

Create/start failure handling: failed rollbacks can surface VGPUCleanupPendingError to the API as vgpu_cleanup_pending (with retained vs reconcile-only guidance). Start clears stale VMM identity before assigning a new vGPU and persists assignments before boot continues.

Startup: API boot replaces bare mdev reconcile with reconcileVGPUs, building a protected VF set from ListInstancesForReconcile (strict on bad metadata), live PIDs, and a bounded grace window for assignments without a PID; it may reschedule reconcile when protection expires. Snapshot restore keeps the instance’s current vGPU fields (including GPUAssignedAt) instead of resurrecting snapshot-embedded assignments.

^{Reviewed by Cursor Bugbot for commit 6e4e796. Bugbot is set up for automated code reviews on this repo. Configure here.}

[yummybomb]

Trimmed over-engineering after an architectural review against the parent task (host kernel 6.8 / Ubuntu 24.04 support). Most of the removed weight came from earlier review-round additions rather than the original design:

[cursor]

Cursor Bugbot has reviewed your changes using high effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit f2ca548. Configure here.}