Add the vendor VFIO vGPU device backend

[yummybomb]

Summary

Layer 2 of the vendor VFIO vGPU stack (generalize-vgpu-device ← hypervisor-liveness ← this ← vendor-vfio-vgpu). Self-contained in lib/devices + lib/resources; nothing in the instance lifecycle calls it yet (that's the top layer).

Linux 6.8 hosts with NVIDIA R580 drop the mdev interface: vGPUs are assigned by writing a type ID to a VF's nvidia/current_vgpu_type and passed to QEMU as a plain VFIO PCI device. This adds that backend behind the framework dispatch introduced in #322:

• Discovery & placement — profile discovery from the capacity-dependent creatable_vgpu_types catalogs, least-loaded-GPU VF selection, create/verify/rollback. Profile availability counts free VFs currently advertising each type as a best-effort snapshot; creating one assignment may change sibling catalogs.
• Release guards — unlike mdev (fresh UUID per assignment), vendor VFIO reuses the same VF path across assignments, so a stale release could clear a later owner's vGPU. Release is guarded by an in-process owner map (covers the window before QEMU opens the device) and an open-VFIO-handle scan (refuses to clear a VF a running VM holds).
• Reconciliation — clears orphaned assignments on startup, skips VFs in the caller-supplied protected set, and fails closed (skips vendor VFIO entirely) when that set is unavailable, while mdev reconciliation still runs.
• Per-VF degradation — one unreadable VF is skipped with a warning instead of failing discovery or profile listing wholesale, so a flaky sysfs read cannot blank the host's advertised GPU capacity. Only when no VF is readable does discovery fail, so a wholesale outage cannot demote a vGPU host to passthrough while assignments exist. The open-handle probe stays strict on purpose: it authorizes clearing a reused VF path, so an incomplete scan fails the release rather than risk a false "not in use".
• Integration test — branched by discovered framework, extended to cover release on stop and reacquisition on start.

Testing

• go build ./..., go vet clean
• go test -race ./lib/devices/ ./lib/resources/ pass

---

Note

Medium Risk
Touches host GPU mode detection, sysfs assignment/release, and /resources admission paths; vendor VFIO release logic is security-sensitive (false “not in use” could clear another owner’s VF), though create is gated and release fails closed on incomplete VFIO scans.

Overview
Introduces dual vGPU frameworks (mdev vs vendor VFIO via current_vgpu_type / creatable_vgpu_types) with DiscoverVGPU, framework-aware profile listing, DestroyVGPU/ReconcileVGPUs, and a large Linux vendor VFIO implementation (placement, owner + open-VFIO release guards, orphan reconciliation, create rollback).

Removes DetectHostGPUMode; resources now derive vGPU vs passthrough from DiscoverVGPU (with context) and avoid mis-reporting passthrough when vGPU discovery fails transiently. Mdev discovery skips unreadable VFs instead of zeroing capacity; instance vGPU teardown passes InstanceID on assignments.

CreateVGPU still rejects vendor VFIO until durable lifecycle integration; integration test skips vendor VFIO hosts accordingly but adds framework-aware sysfs checks and stop/start release/reacquire for mdev.

Docs/OpenAPI/metrics clarify ephemeral vGPU lifecycle, QEMU-only practical vGPU support, best-effort profile availability, and vendor VFIO wedge troubleshooting.

^{Reviewed by Cursor Bugbot for commit 1f1f24f. Bugbot is set up for automated code reviews on this repo. Configure here.}

[yummybomb]

added 7f5233f: report GPUProfile.Available as the count of free VFs advertising the profile type (creatable-instance units), matching mdev's summed available_instances, the OpenAPI description, and the integration test's decrement assertion. note the gpuProfileSlots metric steps up on vendor VFIO hosts (per-GPU → per-free-VF units). go test ./lib/devices ./lib/resources green; the hardware integration test was not run locally.

[github-actions]

-->

✱ stlc build

✅ go code · compare

Your SDK build was successful.

generate ✅ → bootstrap ✅ → format ✅

116 files generated at dd54e02 (pushed)

go get github.com/kernel/hypeman-go-staging@dd54e02078ec70a34c48b149e0080ba65dde4c01

✅ python code · compare

Your SDK build was successful.

generate ✅ → bootstrap ✅ → format ✅

230 files generated at 1e0016e (pushed)

✅ typescript code · compare

Your SDK build was successful.

generate ✅ → bootstrap ✅ → format ✅

138 files generated at 5a2fd33 (pushed)

Diagnostics: ❗ 0 new / 1 total error, 💡 0 new / 5 total note

Level	Code	Message	Targets

Build metadata

Build	bd_769WR6Ls-thankful-shard
Timestamp	2026-08-18T14:46:37.661Z
stlc	8413509
Spec hash	101b9ae4ce99
Config hash	55e15f6f4434

---

This comment is auto-generated by stlc and is kept up to date as you push.
If you push new commits, re-run this workflow to update this comment.
Last updated: 2026-08-18 14:47:10 UTC

[sjmiller609]

• mdev discovery now hard-fails on one unreadable mdev_supported_types, zeroing host GPU capacity — vendor VFIO degrades per-VF. intentional asymmetry? Also that is different behavior than before the VFIO changes.
• reconcile silently skips an allocated + in-use + unprotected VF. log at warn level.