[Bug]: spec.driver.repoConfig is not mounted into openshift-driver-toolkit-ctr (DTK-mode builds can't use it)

[aqeelat]

Describe the bug

With DTK mode active, the driver build runs in openshift-driver-toolkit-ctr (via ocp_dtk_entrypoint dtk-build-driver), but the repoConfig ConfigMap is mounted only into nvidia-driver-ctr. Any dnf step in the DTK container can't see the custom repo — defeating repoConfig on platforms without RHEL entitlements (OKD/SCOS).

To Reproduce

ClusterPolicy with use_ocp_driver_toolkit: true + driver.repoConfig.configMapName. Rendered DS mounts:

nvidia-driver-ctr:            okd-gpu-kernel-packages @ /etc/yum.repos.d/okd-gpu-kernel-packages.repo
openshift-driver-toolkit-ctr: (no configMap mounts)

DTK-container build then fails (Error: Unable to find a match: kernel-6.12.0-233.el10.x86_64) since it only sees baked-in rolling CentOS Stream repos. Live-patching the same mount into the DTK container makes the identical build succeed.

Expected behavior

repoConfig (and certConfig) mounted into openshift-driver-toolkit-ctr whenever DTK mode is enabled.

Environment

• GPU Operator Version: v26.3.3 (OLM certified)
• OS: CentOS Stream CoreOS 10
• Kernel Version: 6.12.0-233.el10.x86_64
• Container Runtime: CRI-O 1.34
• Kubernetes Distro and Version: OKD 4.22.0-okd-scos.6

Debug data available on request: kubectl get pods/ds -n nvidia-gpu-operator, driver pod logs from both containers.

[rahulait]

Thanks for the report @aqeelat , we'll take a look.