Set --skip-module-load flag when installings .ko's with nvidia-installer

Without this flag, a subtle bug can occur if the nvidia-installer fails to unload
the NVIDIA kernel modules after the test load. The modules will remain loaded and
any custom NVIDIA module parameters configured as input to the driver container
will not be applied.

Signed-off-by: Christopher Desiniotis <[email protected]>

Author: cdesiniotis

rhel8/nvidia-driver

@@ -532,6 +532,23 @@ _install_driver() {
     if [ "${ACCEPT_LICENSE}" = "yes" ]; then
         install_args+=("--accept-license")
     fi
+
+    # Specify the --skip-module-load flag for versions of the nvidia-installer that
+    # support it. From the nvidia-installer help output:
+    #
+    #    --skip-module-load
+    #         Skip the test load of the NVIDIA kernel modules after the modules are built,
+    #         and skip loading them after installation is complete.
+    #
+    # Without this flag, a subtle bug can occur if the nvidia-installer fails to unload
+    # the NVIDIA kernel modules after the test load. The modules will remain loaded and
+    # any custom NVIDIA module parameters configured as input to the driver container
+    # will not be applied.
+    #
+    if [ "${DRIVER_BRANCH}" -ge "570" ]; then
+        install_args+=("--skip-module-load")
+    fi
+
     IGNORE_CC_MISMATCH=1 nvidia-installer --kernel-module-only --no-drm --ui=none --no-nouveau-check -m=${KERNEL_TYPE} ${install_args[@]+"${install_args[@]}"}
     # May need to add no-cc-check for Rhel, otherwise it complains about cc missing in path
     # /proc/version and lib/modules/KERNEL_VERSION/proc are different, by default installer looks at /proc/ so, added the proc-mount-point

rhel9/nvidia-driver

@@ -551,6 +551,23 @@ _install_driver() {
     if [ "${ACCEPT_LICENSE}" = "yes" ]; then
         install_args+=("--accept-license")
     fi
+
+    # Specify the --skip-module-load flag for versions of the nvidia-installer that
+    # support it. From the nvidia-installer help output:
+    #
+    #    --skip-module-load
+    #         Skip the test load of the NVIDIA kernel modules after the modules are built,
+    #         and skip loading them after installation is complete.
+    #
+    # Without this flag, a subtle bug can occur if the nvidia-installer fails to unload
+    # the NVIDIA kernel modules after the test load. The modules will remain loaded and
+    # any custom NVIDIA module parameters configured as input to the driver container
+    # will not be applied.
+    #
+    if [ "${DRIVER_BRANCH}" -ge "570" ]; then
+        install_args+=("--skip-module-load")
+    fi
+
     IGNORE_CC_MISMATCH=1 nvidia-installer --kernel-module-only --no-drm --ui=none --no-nouveau-check -m=${KERNEL_TYPE} ${install_args[@]+"${install_args[@]}"}
     # May need to add no-cc-check for Rhel, otherwise it complains about cc missing in path
     # /proc/version and lib/modules/KERNEL_VERSION/proc are different, by default installer looks at /proc/ so, added the proc-mount-point

ubuntu20.04/nvidia-driver

@@ -426,6 +426,23 @@ _install_driver() {
     if [ "${ACCEPT_LICENSE}" = "yes" ]; then
         install_args+=("--accept-license")
     fi
+
+    # Specify the --skip-module-load flag for versions of the nvidia-installer that
+    # support it. From the nvidia-installer help output:
+    #
+    #    --skip-module-load
+    #         Skip the test load of the NVIDIA kernel modules after the modules are built,
+    #         and skip loading them after installation is complete.
+    #
+    # Without this flag, a subtle bug can occur if the nvidia-installer fails to unload
+    # the NVIDIA kernel modules after the test load. The modules will remain loaded and
+    # any custom NVIDIA module parameters configured as input to the driver container
+    # will not be applied.
+    #
+    if [ "${DRIVER_BRANCH}" -ge "570" ]; then
+        install_args+=("--skip-module-load")
+    fi
+
     nvidia-installer --kernel-module-only --no-drm --ui=none --no-nouveau-check -m=${KERNEL_TYPE} ${install_args[@]+"${install_args[@]}"}
 }
 

ubuntu22.04/nvidia-driver

@@ -503,7 +503,24 @@ _install_driver() {
     if [ "${ACCEPT_LICENSE}" = "yes" ]; then
         install_args+=("--accept-license")
     fi
-    nvidia-installer --kernel-module-only --no-drm --ui=none --no-nouveau-check -m=${KERNEL_TYPE} ${install_args[@]+"${install_args[@]}"}
+
+    # Specify the --skip-module-load flag for versions of the nvidia-installer that
+    # support it. From the nvidia-installer help output:
+    #
+    #    --skip-module-load
+    #         Skip the test load of the NVIDIA kernel modules after the modules are built,
+    #         and skip loading them after installation is complete.
+    #
+    # Without this flag, a subtle bug can occur if the nvidia-installer fails to unload
+    # the NVIDIA kernel modules after the test load. The modules will remain loaded and
+    # any custom NVIDIA module parameters configured as input to the driver container
+    # will not be applied.
+    #
+    if [ "${DRIVER_BRANCH}" -ge "570" ]; then
+        install_args+=("--skip-module-load")
+    fi
+
+    nvidia-installer --kernel-module-only --no-drm --skip-module-load --ui=none --no-nouveau-check -m=${KERNEL_TYPE} ${install_args[@]+"${install_args[@]}"}
 }
 
 # Mount the driver rootfs into the run directory with the exception of sysfs.

ubuntu24.04/nvidia-driver

@@ -447,6 +447,22 @@ _install_driver() {
         install_args+=("-m=${kernel_type}")
     fi
 
+    # Specify the --skip-module-load flag for versions of the nvidia-installer that
+    # support it. From the nvidia-installer help output:
+    #
+    #    --skip-module-load
+    #         Skip the test load of the NVIDIA kernel modules after the modules are built,
+    #         and skip loading them after installation is complete.
+    #
+    # Without this flag, a subtle bug can occur if the nvidia-installer fails to unload
+    # the NVIDIA kernel modules after the test load. The modules will remain loaded and
+    # any custom NVIDIA module parameters configured as input to the driver container
+    # will not be applied.
+    #
+    if [ "${DRIVER_BRANCH}" -ge "570" ]; then
+        install_args+=("--skip-module-load")
+    fi
+
     # Install the NVIDIA driver in one step
     sh NVIDIA-Linux-$DRIVER_ARCH-$DRIVER_VERSION.run --silent \
                     --ui=none \