refactor: address PR review comments for driver install scripts

Signed-off-by: Karthik Vetrivel <[email protected]>

Author: karthikvetrivel

rhel9/common.sh

@@ -53,14 +53,15 @@ _read_conf_file() {
 }
 
 # Build driver configuration for state comparison
+# Note: Variables are expected to be set by the sourcing script (nvidia-driver)
 _build_driver_config() {
     cat <<EOF
 DRIVER_VERSION=${DRIVER_VERSION}
-DRIVER_TYPE=${DRIVER_TYPE:-passthrough}
+DRIVER_TYPE=${DRIVER_TYPE}
 KERNEL_VERSION=$(uname -r)
-GPU_DIRECT_RDMA_ENABLED=${GPU_DIRECT_RDMA_ENABLED:-false}
-USE_HOST_MOFED=${USE_HOST_MOFED:-false}
-KERNEL_MODULE_TYPE=${KERNEL_MODULE_TYPE:-auto}
+GPU_DIRECT_RDMA_ENABLED=${GPU_DIRECT_RDMA_ENABLED}
+USE_HOST_MOFED=${USE_HOST_MOFED}
+KERNEL_MODULE_TYPE=${KERNEL_MODULE_TYPE}
 NVIDIA_MODULE_PARAMS=$(_read_conf_file /drivers/nvidia.conf)
 NVIDIA_UVM_MODULE_PARAMS=$(_read_conf_file /drivers/nvidia-uvm.conf)
 NVIDIA_MODESET_MODULE_PARAMS=$(_read_conf_file /drivers/nvidia-modeset.conf)

rhel9/nvidia-driver

@@ -8,6 +8,7 @@ PID_FILE=${RUN_DIR}/${0##*/}.pid
 DRIVER_VERSION=${DRIVER_VERSION:?"Missing DRIVER_VERSION env"}
 KERNEL_UPDATE_HOOK=/run/kernel/postinst.d/update-nvidia-driver
 NUM_VGPU_DEVICES=0
+DRIVER_TYPE="${DRIVER_TYPE:-passthrough}"
 GPU_DIRECT_RDMA_ENABLED="${GPU_DIRECT_RDMA_ENABLED:-false}"
 USE_HOST_MOFED="${USE_HOST_MOFED:-false}"
 NVIDIA_MODULE_PARAMS=()
@@ -719,7 +720,7 @@ _start_daemons() {
 }
 
 _store_driver_config() {
-    local config_file="/run/nvidia/driver-config.state"
+    local config_file="${RUN_DIR}/driver-config.state"
     echo "Storing driver configuration state..."
     _build_driver_config > "$config_file"
     echo "Driver configuration stored at $config_file"
@@ -757,7 +758,7 @@ _userspace_only_install() {
     fi
 
     _mount_rootfs
-    _start_daemons
+    _ensure_persistence
     _write_kernel_update_hook
     _store_driver_config
     echo "Userspace-only install complete"

ubuntu22.04/nvidia-driver

@@ -8,6 +8,7 @@ PID_FILE=${RUN_DIR}/${0##*/}.pid
 DRIVER_VERSION=${DRIVER_VERSION:?"Missing DRIVER_VERSION env"}
 KERNEL_UPDATE_HOOK=/run/kernel/postinst.d/update-nvidia-driver
 NUM_VGPU_DEVICES=0
+DRIVER_TYPE="${DRIVER_TYPE:-passthrough}"
 GPU_DIRECT_RDMA_ENABLED="${GPU_DIRECT_RDMA_ENABLED:-false}"
 USE_HOST_MOFED="${USE_HOST_MOFED:-false}"
 NVIDIA_MODULE_PARAMS=()
@@ -660,7 +661,7 @@ _read_conf_file() {
 _build_driver_config() {
     cat <<EOF
 DRIVER_VERSION=${DRIVER_VERSION}
-DRIVER_TYPE=${DRIVER_TYPE:-passthrough}
+DRIVER_TYPE=${DRIVER_TYPE}
 KERNEL_VERSION=$(uname -r)
 GPU_DIRECT_RDMA_ENABLED=${GPU_DIRECT_RDMA_ENABLED}
 USE_HOST_MOFED=${USE_HOST_MOFED}
@@ -673,7 +674,7 @@ EOF
 }
 
 _store_driver_config() {
-    local config_file="/run/nvidia/driver-config.state"
+    local config_file="${RUN_DIR}/driver-config.state"
     echo "Storing driver configuration state..."
     _build_driver_config > "$config_file"
     echo "Driver configuration stored at $config_file"
@@ -700,9 +701,9 @@ _install_userspace_components() {
         --x-sysconfig-path=/tmp/null
 }
 
-_copy_kernel_module_sources() {
+_move_kernel_module_sources() {
     mkdir -p /usr/src/nvidia-${DRIVER_VERSION}
-    cp -r LICENSE mkprecompiled ${KERNEL_TYPE} /usr/src/nvidia-${DRIVER_VERSION}/
+    mv LICENSE mkprecompiled ${KERNEL_TYPE} /usr/src/nvidia-${DRIVER_VERSION}/
     sed '9,${/^\(kernel\|LICENSE\)/!d}' .manifest > /usr/src/nvidia-${DRIVER_VERSION}/.manifest
 }
 
@@ -744,11 +745,9 @@ init() {
             echo "Detected matching loaded driver & config (${DRIVER_VERSION}); performing userspace-only install"
             _unmount_rootfs
             _update_package_cache
-            _resolve_kernel_version || exit 1
-            _install_prerequisites
             _install_userspace_components
             _resolve_kernel_type || exit 1
-            _copy_kernel_module_sources
+            _move_kernel_module_sources
             _mount_rootfs
             _ensure_persistenced
             _write_kernel_update_hook
@@ -764,10 +763,7 @@ init() {
     _install_userspace_components
     _resolve_kernel_type || exit 1
 
-    # Move (not copy) kernel module sources since this is the full install path
-    mkdir -p /usr/src/nvidia-${DRIVER_VERSION}
-    mv LICENSE mkprecompiled ${KERNEL_TYPE} /usr/src/nvidia-${DRIVER_VERSION}/
-    sed '9,${/^\(kernel\|LICENSE\)/!d}' .manifest > /usr/src/nvidia-${DRIVER_VERSION}/.manifest
+    _move_kernel_module_sources
 
     if _kernel_requires_package; then
         _update_ca_certificates