Merge ~bladernr/plainbox-provider-certification-server:1879965-move-GPGPU-to-own-provider into plainbox-provider-certification-server:master

[Jeff Lane ]

LP: 1879965 - Moved GPGPU test files into their own provider.

LP: 1879965 - Moved GPGPU test files into their own provider.

[Sylvain Pineau (sylvain-pineau)]

+1

review: Approve

diff --git a/launcher/test-gpgpu b/launcher/test-gpgpu
 deleted file mode 100755
index 4669bf0..0000000
--- a/launcher/test-gpgpu
+++ /dev/null
@@ -1,19 +0,0 @@
-#!/usr/bin/env checkbox-cli
-
-[launcher]
-app_id = com.canonical.certification:certification-server
-launcher_version = 1
-stock_reports = submission_files, certification
-
-[test plan]
-unit = com.canonical.certification::gpgpu-only
-forced = yes
-
-[test selection]
-forced = yes
-
-[ui]
-output = hide-resource-and-attachment
-
-[config]
-config_filename = canonical-certification.conf
diff --git a/tools/canonical-certification-precheck b/tools/canonical-certification-precheck
index 210a5b3..c8ba2c4 100755
--- a/tools/canonical-certification-precheck
+++ b/tools/canonical-certification-precheck
@@ -839,9 +839,9 @@ if lspci | grep "3D controller [0302]"; then
     gpu="GPGPU Detected"
     echo " $gpu"
     echo " In order to test the discovered GPGPU device(s)"
-    echo " You will need to run the gpu-setup.sh script "
-    echo " included with the Server Certification suite to"
-    echo " install the necessary drivers and toolkit."
+    echo " You will need to install the checkbox-provider-gpgpu"
+    echo " package and run the gpu-setup script to install"
+    echo " the necessary drivers and toolkit."
     eval info${i}=\"$gpu\"
     info
 else
diff --git a/tools/gpu-setup.sh b/tools/gpu-setup.sh
 deleted file mode 100755
index 22a8430..0000000
--- a/tools/gpu-setup.sh
+++ /dev/null
@@ -1,79 +0,0 @@
-#!/bin/bash
-
-if [[ $EUID -ne 0 ]]; then
-   echo "ERROR: This script must be run as root"
-   exit 1
-fi
-
-echo "Configuring system for GPU Testing"
-echo "**********************************"
-echo "*"
-
-# For now we need internet access for this to work.  Future versions will
-# remove this necessity
-echo "*  Testing network connectivity"
-[[ ! `ping -c 1 www.ubuntu.com` ]] \
-    && echo "ERROR: This script requires internet access to function correctly" \
-    && exit 1
-echo "*"
-echo "**********************************"
-echo "*"
-echo "*  Adding nVidia package repository"
-## Leave the following bits for now, it's useful code to template if we sort
-## out local hosting in the future.
-#file_pattern="cuda-repo-ubuntu$(lsb_release -r | cut -f 2 | sed -es/\\.//)*.deb"
-#gw_ip=`ip route | awk '/default/ { print $3 }'`
-
-# attempt a download from our MAAS server (presume the SUT's gw IP is MAAS)
-#wget -r -l1 -nd --no-parent -A "$file_pattern" http://$gw_ip/nvidia/
-#nvidia_pkg=`find / -name $file_pattern`
-
-#if [ -z "$nvidia_pkg" ]; then
-#    echo "*  No local package found, installing remote installer"
-#    wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
-#    dpkg -i cuda-repo-ubuntu1804_10.0.130-1_amd64.deb
-#    apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
-#else
-#    echo "*  Local package detected at $nvidia_pkg... installing"
-#    key_cmd=`dpkg -i $nvidia_pkg |awk '/To install the key, run this command/{getline; print}'`
-#    $key_cmd
-#fi    
-
-## For now, require internet access and installing directly from nVidia
-# SAUCE: https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=debnetwork
-
-OSRELEASE=ubuntu`lsb_release -r | cut -f2 |sed -e 's/\.//'`
-
-wget https://developer.download.nvidia.com/compute/cuda/repos/$OSRELEASE/x86_64/cuda-$OSRELEASE.pin
-mv cuda-$OSRELEASE.pin /etc/apt/preferences.d/cuda-repository-pin-600
-# Ran into a case where the 16.04 repo uses http while 18.04 uses https :/
-KEYSERVER_URL_TAIL="://developer.download.nvidia.com/compute/cuda/repos/$OSRELEASE/x86_64/7fa2af80.pub"
-SUCCESS=1
-for KEYSERVER_URL_HEAD in https http; do
-    KEYSERVER_URL=$KEYSERVER_URL_HEAD$KEYSERVER_URL_TAIL
-    apt-key adv --fetch-keys $KEYSERVER_URL
-    SUCCESS=$?
-    [ $SUCCESS == 0 ] && break
-done
-
-add-apt-repository "deb http://developer.download.nvidia.com/compute/cuda/repos/$OSRELEASE/x86_64/ /"
-
-# Install necessary files
-apt update
-echo "*  Installing necessary pacakges"
-apt install -y build-essential git cuda
-
-#fix the path to get nvcc from the cuda package
-CUDA_PATH=$(find /usr/local -maxdepth 1 -type d -iname "cuda*")/bin
-export PATH=$PATH:$CUDA_PATH
-
-# get the gpu-burn repo and build it
-echo "*  Cloning gpu-burn repo"
-GPU_BURN_DIR=/opt/gpu-burn
-git clone https://github.com/wilicc/gpu-burn.git $GPU_BURN_DIR
-cd $GPU_BURN_DIR
-echo "*  Building gpu-burn"
-make && echo "*  Build completed..."
-echo "*"
-echo "*  Completed installation. Please reboot the machine now"
-echo "*  to load the nVidia proprietary drivers"
diff --git a/units/gpgpu-only.pxu b/units/gpgpu-only.pxu
 deleted file mode 100644
index b92f279..0000000
--- a/units/gpgpu-only.pxu
+++ /dev/null
@@ -1,15 +0,0 @@
-id: gpgpu-only
-unit: test plan
-_name: GPGPU Compute Testing
-_description:
- Tests for GPGPU Computations (non-graphical)
-nested_part:
-    com.canonical.certification::server-info-attachment-automated
-    com.canonical.certification::server-firmware
-    com.canonical.certification::server-miscellaneous
-    com.canonical.certification::gpgpu-tests
-include:
-bootstrap_include:
-    device
-    fwts
-