feat(gpu): add WSL CDI spec watcher and set deviceIDStrategy to index

On WSL2 hosts the NVIDIA device plugin generates CDI specs that cannot
be used directly by k3s containerd since it includes a single device
name "all" and not one based on the index or UUID of the device.

Add a background watch_cdi_specs function to cluster-entrypoint.sh that:
- detects WSL2 via /dev/dxg presence
- handles specs already present at gateway restart
- uses inotifywait to watch for new/updated specs
- transforms the spec with jq (cdiVersion=0.5.0, devices[0].name="0")

Add inotify-tools and jq to the cluster image apt-get install block to
support the watcher.

Signed-off-by: Evan Lezar <elezar@nvidia.com>

Author: elezar

deploy/docker/Dockerfile.images

@@ -244,6 +244,8 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     iptables \
     mount \
     dnsutils \
+    inotify-tools \
+    jq \
     && apt-get install -y --only-upgrade gpgv \
     && rm -rf /var/lib/apt/lists/*
 

deploy/docker/cluster-entrypoint.sh

@@ -381,6 +381,45 @@ fi
 # the k3s manifests directory so the Helm controller installs it automatically.
 # The nvidia-container-runtime binary is already on PATH (baked into the image)
 # so k3s registers the "nvidia" RuntimeClass at startup.
+CDI_SPEC_DIR="/var/run/cdi"
+CDI_WSL_INPUT="${CDI_SPEC_DIR}/k8s.device-plugin.nvidia.com-gpu.json"
+CDI_WSL_OUTPUT="${CDI_SPEC_DIR}/openshell-wsl.json"
+
+transform_wsl_cdi_spec() {
+    local tmp="${CDI_WSL_OUTPUT}.tmp.$$"
+    if jq '.cdiVersion = "0.5.0" | .devices[0].name = "0"' \
+            "$CDI_WSL_INPUT" > "$tmp" 2>/dev/null; then
+        mv "$tmp" "$CDI_WSL_OUTPUT"
+        echo "CDI: transformed WSL spec -> $CDI_WSL_OUTPUT"
+    else
+        rm -f "$tmp"
+        echo "CDI: failed to transform WSL spec (jq error)"
+    fi
+}
+
+watch_cdi_specs() {
+    if ! command -v inotifywait > /dev/null 2>&1; then
+        echo "CDI: inotifywait not found, skipping spec watcher"
+        return 1
+    fi
+
+    mkdir -p "$CDI_SPEC_DIR"
+
+    # Process spec already present at startup (e.g. gateway restart)
+    if [ -f "$CDI_WSL_INPUT" ] && grep -q '/dev/dxg' "$CDI_WSL_INPUT" 2>/dev/null; then
+        transform_wsl_cdi_spec
+    fi
+
+    # Watch for the spec to appear or be updated
+    inotifywait -m -e close_write,moved_to --format '%f' "$CDI_SPEC_DIR" 2>/dev/null \
+    | while IFS= read -r filename; do
+        if [ "$filename" = "k8s.device-plugin.nvidia.com-gpu.json" ] \
+                && grep -q '/dev/dxg' "$CDI_WSL_INPUT" 2>/dev/null; then
+            transform_wsl_cdi_spec
+        fi
+    done
+}
+
 if [ "${GPU_ENABLED:-}" = "true" ]; then
     echo "GPU support enabled — deploying NVIDIA device plugin"
 
@@ -391,6 +430,11 @@ if [ "${GPU_ENABLED:-}" = "true" ]; then
             cp "$manifest" "$K3S_MANIFESTS/"
         done
     fi
+
+    if [ -c /dev/dxg ]; then
+        echo "WSL2 GPU detected (/dev/dxg present) — starting CDI spec watcher"
+        watch_cdi_specs &
+    fi
 fi
 
 # ---------------------------------------------------------------------------

deploy/kube/gpu-manifests/nvidia-device-plugin-helmchart.yaml

@@ -16,6 +16,11 @@
 # devices are injected via CDI hooks before container start. Sandbox pods only
 # need the nvidia.com/gpu resource request — no runtimeClassName is required.
 #
+# The device plugin is set to deviceIDStrategy=index so that device names are
+# numeric indices (e.g. "0"). This simplifies the conversion of CDI specs on WSL
+# systems, where we need to rename the *.nvidia.com/gpu=all device that is
+# generated by the device plugin to *.nvidia.com/gpu=0.
+#
 # k3s auto-detects nvidia-container-runtime on PATH and registers the "nvidia"
 # RuntimeClass automatically, so no manual RuntimeClass manifest is needed.