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.
• 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 sysfs assignment and release with VFIO in-use checks, but vendor VFIO instance creation remains disabled until a follow-up integration layer lands.

Overview
Adds NVIDIA vendor VFIO vGPU support for newer hosts (assign via current_vgpu_type / creatable_vgpu_types) alongside the existing mdev path, unified behind DiscoverVGPU and framework-aware profile listing, create/destroy, and reconciliation.

vendor_vfio_linux.go implements sysfs discovery, profile availability from free VFs, least-loaded-GPU placement, create/verify/rollback, and teardown that refuses to clear a VF when open VFIO handles exist or a stale instance does not own the assignment. ReconcileVGPUs clears orphaned vendor assignments while honoring a protected device-path set.

CreateVGPU still errors on vendor VFIO until instance lifecycle integration lands; destroy/reconcile and resources reporting already route through the new stack. GetGPUStatus now takes context, probes vGPU before falling back to passthrough, and documents profile availability as a best-effort VF snapshot.

The vGPU integration test is framework-aware (sysfs checks, stop/release, start/reacquire) but skips vendor VFIO hosts until create is wired. Docs/OpenAPI reflect mdev vs vendor VFIO and ephemeral assignment lifecycle.

^{Reviewed by Cursor Bugbot for commit 9769a33. 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 ✅

112 files generated at 6927b75 (pushed)

go get github.com/kernel/hypeman-go-staging@6927b75ff19ed53011cc2047f9d61c23cc1cc694

✅ typescript code · compare

Your SDK build was successful.

generate ✅ → bootstrap ✅ → format ✅

134 files generated at 6a2532f (pushed)

Diagnostics: 💡 0 new / 5 total note

Level	Code	Message	Targets

Build metadata

Build	bd_766sHtSJ-forked-cradle
Timestamp	2026-08-10T20:45:58.247Z
stlc	8413509
Spec hash	c36f98252380
Config hash	9125eb22cb4c

---

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: 2026-08-10 20:46:15 UTC