fix(docs,ci): Radeon device IDs in OpenShift NFD rules + drift guard

[bhatnitish]

Summary

Deploying the GPU Operator via OLM on OpenShift with Radeon GPUs detects zero GPUs. On an 8x Radeon AI PRO R9700 (0x7551) cluster, every test failed with No nodes with AMD/GPU found in the cluster.

On OpenShift the NodeFeatureRule is created by hand from docs/installation/openshift-olm.md. Both copy-paste YAML blocks in that doc — a NodeFeatureDiscovery CR and a NodeFeatureRule CR — listed AMD Instinct device IDs only, so no node ever received feature.node.kubernetes.io/amd-gpu and nothing the operator deploys was scheduled.

Root cause

The AMD GPU PCI device-ID list is duplicated across several hand-maintained files with nothing enforcing agreement. The commit that added the Radeon AI PRO R9700 (0x7551) updated the two gpu-nfd-default-rule.yaml copies and left the documentation untouched.

Drift measured against the source of truth: 16 missing amd-gpu IDs and 2 missing amd-vgpu IDs.

Changes

File	Change
docs/installation/openshift-olm.md	+16 amd-gpu / +2 amd-vgpu device IDs, both YAML blocks
docs/troubleshooting.md	New section for a missing amd-gpu label
hack/check-nfd-device-ids.py	Drift checker (Python 3 stdlib only)
Makefile	make check-nfd-device-ids
.github/workflows/nfd-device-id-check.yml	CI, a thin caller of the make target

Source of truth is hack/k8s-patch/template-patch/gpu-nfd-default-rule.yaml. make helm-k8s runs rm -rf helm-charts-k8s and then repopulates it from template-patch, so helm-charts-k8s/templates/gpu-nfd-default-rule.yaml is a build output — drift there means the wrong file was edited, or make helm-k8s was not re-run before committing. The checker reports both directions and names the exact IDs.

The workflow carries no branch or path filter: a path-filtered workflow reports no status on unrelated PRs, which stalls merges if it is configured as a required status check. The check runs in about a second.

Validation

• Run before the doc fix, the checker failed with exactly the 16+2 IDs. It drove the doc edit, so the fix cannot be partial. It passes after.
• Both doc YAML blocks parse via pyyaml, yielding 30 amd-gpu and 11 amd-vgpu entries — the added standalone comments inside the flow sequences do not break copy-paste.
• Checker verified in both directions (missing and extra IDs), plus its parse-failure guard and its guard against a documentation code block being removed wholesale.
• make check-nfd-device-ids is green, and is unaffected by a missing docker socket.

Docs lint: passes in CI — see the Documentation / Markdown and Documentation / Spelling checks on this PR.

[spraveenio]

LGTM