DONOTMERGE Force NVIDIA driver to use the same gcc version used to compile the kernel

Signed-off-by: Giacomo Marciani <[email protected]>

Author: gmarciani

cookbooks/aws-parallelcluster-platform/resources/nvidia_driver/nvidia_driver_ubuntu20+.rb renamed to cookbooks/aws-parallelcluster-platform/resources/nvidia_driver/nvidia_driver_ubuntu20.rb

@@ -13,7 +13,7 @@
 # See the License for the specific language governing permissions and limitations under the License.
 
 provides :nvidia_driver, platform: 'ubuntu' do |node|
-  node['platform_version'].to_i >= 20
+  node['platform_version'].to_i == 20
 end
 
 use 'partial/_nvidia_driver_common.rb'

cookbooks/aws-parallelcluster-platform/resources/nvidia_driver/nvidia_driver_ubuntu22.rb

@@ -0,0 +1,45 @@
+# frozen_string_literal: true
+
+# Copyright:: 2024 Amazon.com, Inc. or its affiliates. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License").
+# You may not use this file except in compliance with the License.
+# A copy of the License is located at
+#
+# http://aws.amazon.com/apache2.0/
+#
+# or in the "LICENSE.txt" file accompanying this file.
+# This file is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, express or implied.
+# See the License for the specific language governing permissions and limitations under the License.
+
+provides :nvidia_driver, platform: 'ubuntu' do |node|
+  node['platform_version'].to_i == 22
+end
+
+use 'partial/_nvidia_driver_common.rb'
+
+# def set_compiler?
+#   # Ubuntu22.04 with Kernel 6.8.x needs to set CC to /usr/bin/gcc12 using dkms override
+#   node['kernel']['release'].split('.')[0].to_i == 6
+#   node['kernel']['release'].split('.')[1].to_i == 8
+# end
+
+def rebuild_initramfs?
+  true
+end
+
+# def compiler_version
+#   'gcc'
+# end
+
+def compiler_path
+  gcc_version = get_gcc_version_used_by_kernel
+
+  if gcc_version.nil?
+    return ""
+  end
+
+  gcc_major_version = gcc_version.split('.')[0].to_i
+
+  "CC=/usr/bin/gcc-#{gcc_major_version}"
+end

cookbooks/aws-parallelcluster-platform/resources/nvidia_driver/partial/_nvidia_driver_common.rb

@@ -73,14 +73,13 @@
   end
 
   # Install driver
-  # TODO remove --no-cc-version-check when we can update ubuntu 22 images
   bash 'nvidia.run advanced' do
     user 'root'
     group 'root'
     cwd '/tmp'
     code <<-NVIDIA
       set -e
-      #{compiler_path} ./nvidia.run --silent --dkms --disable-nouveau --no-cc-version-check -m=#{nvidia_kernel_module}
+      #{compiler_path} ./nvidia.run --silent --dkms --disable-nouveau -m=#{nvidia_kernel_module}
       rm -f /tmp/nvidia.run
     NVIDIA
     creates '/usr/bin/nvidia-smi'
@@ -120,6 +119,10 @@ def compiler_path
   ""
 end
 
+def extra_packages
+  []
+end
+
 def nvidia_kernel_module
   if ['false', 'no', false].include?(node['cluster']['nvidia']['kernel_open'])
     "kernel"

cookbooks/aws-parallelcluster-shared/libraries/helpers.rb

@@ -106,3 +106,14 @@ def wait_sync_file(path)
     timeout 5
   end
 end
+
+def get_gcc_version_used_by_kernel
+  begin
+    gcc_version = shell_out!("awk '{print $8}' /proc/version | tr -d ',' | cut -d '.' -f 1").stdout.strip
+  rescue => error
+    Chef::Log.error("Cannot detect gcc version used to compile the kernel: #{error}")
+    return nil
+  end
+  Chef::Log.info("Detected version of gcc used to compile the kernel is: #{gcc_version}")
+  gcc_version
+end