Jason Bonnell
463126770a
## Motivation <!-- Explain the purpose of this PR and the goals it aims to achieve. --> - __Reduced Code Duplication__: Version parsing logic moved from individual Dockerfiles to the central build script - __Improved Edge Case Handling__: Better handling of ROCm versions with and without patch numbers (e.g., `6.2` vs `6.2.0`) - __Easier Maintenance__: Future version-related changes only need to be made in one place - __Cleaner Dockerfiles__: Simplified Dockerfiles focus on package installation rather than complex shell logic - __Updated Platform Support__: Refreshed container matrix to reflect current platform/ROCm version combinations - __Fix OpenSUSE Docker Generation__: OpenSUSE container generation fails due to a change to the `binutils-gold` package - __Error Handling__: Fix bug where errors in docker image build were being masked, allowing workflow to pass anyway. ## Technical Details <!-- Explain the changes along with any relevant GitHub links. --> - Updated `Dockerfile.opensuse` and `Dockerfile.opensuse.ci` docker files to remove `binutils-gold` - Not needed since we build `binutils` with systems anyways - Updated `rocprofiler-systems-containers.yml` to remove `pushd/popd` commands and just run the shell scripts - There was a silent failure observed here, which I verified in this PR before adding the fix for openSUSE - Refactor ROCm version parsing. Move this logic to the `build-docker.sh` script to reduce duplication. - Fix bug that caused ROCm 7.0 to fail installation. The trailing `.0` was being trimmed. - Fixed inconsistencies in `containers.yml` that lead to invalid ROCm-OS_VERSION combinations. - Formatting fixes - Removed trailing whitespace - Fix docker build warnings. Use an `=` rather than ` ` when assigning an environment variable.
375 baris
12 KiB
Bash
Executable File
375 baris
12 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
|
|
set-user-defaults()
|
|
{
|
|
: ${USER:=$(whoami)}
|
|
: ${ROCM_VERSIONS:="6.3"}
|
|
: ${DISTRO:=ubuntu}
|
|
: ${VERSIONS:=22.04}
|
|
: ${PYTHON_VERSIONS:="6 7 8 9 10 11 12 13"}
|
|
: ${BUILD_CI:=""}
|
|
: ${PUSH:=0}
|
|
: ${PULL:=--pull}
|
|
: ${RETRY:=3}
|
|
: ${SCRIPT_DIR=$(dirname "$(readlink -f "${BASH_SOURCE[0]:-$0}")")}
|
|
}
|
|
|
|
set-user-defaults
|
|
|
|
set -e
|
|
|
|
cd $(dirname ${SCRIPT_DIR})
|
|
|
|
declare -a MATRIX_DISTROS=()
|
|
declare -a MATRIX_VERSIONS=()
|
|
declare -a MATRIX_ROCM_VERSIONS=()
|
|
|
|
load-matrix()
|
|
{
|
|
local container_file="$(realpath "${SCRIPT_DIR}/containers.yml")"
|
|
if [ ! -f "${container_file}" ]; then
|
|
echo -e "\n Error: Cannot find ${container_file}"
|
|
exit 1
|
|
fi
|
|
|
|
# In form os-distro;os-version;rocm-version
|
|
local matrix_data=$(awk '
|
|
/matrix:/, /steps:/ {
|
|
if (/- os-distro:/) {
|
|
gsub(/[[:space:]]*- os-distro:[[:space:]]*"/, "")
|
|
gsub(/"/, "")
|
|
distro = $0
|
|
}
|
|
if (/os-version:/) {
|
|
gsub(/[[:space:]]*os-version:[[:space:]]*"/, "")
|
|
gsub(/"/, "")
|
|
version = $0
|
|
}
|
|
if (/rocm-version:/) {
|
|
gsub(/[[:space:]]*rocm-version:[[:space:]]*"/, "")
|
|
gsub(/"/, "")
|
|
rocm = $0
|
|
printf "%s;%s;%s\n", distro, version, rocm
|
|
}
|
|
}
|
|
' "${container_file}")
|
|
|
|
while IFS=';' read -r os_distro os_version rocm_version; do
|
|
MATRIX_DISTROS+=("$os_distro")
|
|
MATRIX_VERSIONS+=("$os_version")
|
|
MATRIX_ROCM_VERSIONS+=("$rocm_version")
|
|
done <<< "$matrix_data"
|
|
}
|
|
|
|
validate-distro()
|
|
{
|
|
local distro="${1}"
|
|
|
|
if [ -n "${distro}" ]; then
|
|
distro=$(tolower "${distro}")
|
|
case "${distro}" in
|
|
ubuntu|opensuse|rhel)
|
|
;;
|
|
*)
|
|
send-error "Unsupported distribution '${distro}'" "Supported distributions: ubuntu, opensuse, rhel"
|
|
;;
|
|
esac
|
|
fi
|
|
}
|
|
|
|
show-matrix()
|
|
{
|
|
local filter_distro="${1:-}"
|
|
|
|
if [ -n "${filter_distro}" ]; then
|
|
validate-distro "${filter_distro}"
|
|
filter_distro=$(tolower "${filter_distro}")
|
|
fi
|
|
|
|
echo ""
|
|
if [ -n "${filter_distro}" ]; then
|
|
echo " Supported ${filter_distro} + ROCm Combinations "
|
|
echo " =============================================="
|
|
else
|
|
echo " Supported OS + ROCm Combinations "
|
|
echo " =========================================="
|
|
fi
|
|
echo ""
|
|
echo " OS Distribution Version ROCm Version"
|
|
echo " ---------------- ------- ------------"
|
|
|
|
for i in "${!MATRIX_DISTROS[@]}"; do
|
|
if [[ -z "${filter_distro}" || "${MATRIX_DISTROS[i]}" == "${filter_distro}" ]]; then
|
|
printf " %-16s %-9s %s\n" "${MATRIX_DISTROS[i]}" "${MATRIX_VERSIONS[i]}" "${MATRIX_ROCM_VERSIONS[i]}"
|
|
fi
|
|
done
|
|
|
|
echo ""
|
|
echo "ROCm '0.0' means no ROCm installation (CPU-only build)"
|
|
echo ""
|
|
echo "Note: Patch versions are also supported (See: https://repo.radeon.com/amdgpu-install/)"
|
|
echo ""
|
|
}
|
|
|
|
# Cross checks arguments against compatibility matrix (ignores ROCm patch version)
|
|
validate-combinations()
|
|
{
|
|
# Check OS version combinations
|
|
for VERSION in ${VERSIONS}; do
|
|
VERSION_MAJOR=$(echo ${VERSION} | sed 's/\./ /g' | awk '{print $1}')
|
|
VERSION_MINOR=$(echo ${VERSION} | sed 's/\./ /g' | awk '{print $2}')
|
|
|
|
local os_version_valid=0
|
|
for i in "${!MATRIX_DISTROS[@]}"; do
|
|
if [[ "${MATRIX_DISTROS[i]}" == "${DISTRO}" && \
|
|
"${MATRIX_VERSIONS[i]}" == "${VERSION}" ]]; then
|
|
os_version_valid=1
|
|
break
|
|
fi
|
|
done
|
|
|
|
if [ ${os_version_valid} -eq 0 ]; then
|
|
send-error "Unsupported OS version :: ${VERSION}. See compatibility matrix for supported versions."
|
|
fi
|
|
done
|
|
|
|
# Check ROCm version combinations
|
|
for VERSION in ${VERSIONS}; do
|
|
for ROCM_VERSION in ${ROCM_VERSIONS}; do
|
|
local valid=0
|
|
ROCM_MAJOR=$(echo ${ROCM_VERSION} | sed 's/\./ /g' | awk '{print $1}')
|
|
ROCM_MINOR=$(echo ${ROCM_VERSION} | sed 's/\./ /g' | awk '{print $2}')
|
|
ROCM_MAJOR_MINOR="${ROCM_MAJOR}.${ROCM_MINOR}"
|
|
if ! ([ "${ROCM_MAJOR_MINOR}" == "0.0" ] && [ "${ROCM_VERSION}" != "0.0" ]); then
|
|
for i in "${!MATRIX_DISTROS[@]}"; do
|
|
if [[ "${MATRIX_DISTROS[i]}" == "${DISTRO}" && \
|
|
"${MATRIX_VERSIONS[i]}" == "${VERSION}" && \
|
|
"${MATRIX_ROCM_VERSIONS[i]}" == "${ROCM_MAJOR_MINOR}" ]]; then
|
|
valid=1
|
|
break
|
|
fi
|
|
done
|
|
fi
|
|
|
|
if [ ${valid} -eq 0 ]; then
|
|
send-error "Unsupported combination :: ${DISTRO}-${VERSION} + ROCm ${ROCM_VERSION}. See compatibility matrix for supported versions."
|
|
fi
|
|
done
|
|
done
|
|
}
|
|
|
|
tolower()
|
|
{
|
|
echo "$@" | awk -F '\\|~\\|' '{print tolower($1)}';
|
|
}
|
|
|
|
toupper()
|
|
{
|
|
echo "$@" | awk -F '\\|~\\|' '{print toupper($1)}';
|
|
}
|
|
|
|
usage()
|
|
{
|
|
set-user-defaults
|
|
print_option() { printf " --%-20s %-24s %s\n" "${1}" "${2}" "${3}"; }
|
|
echo "Options:"
|
|
print_option "help -h" "" "This message"
|
|
print_option "no-pull" "" "Do not pull down most recent base container"
|
|
print_option "matrix -m" "[ubuntu|opensuse|rhel]" "Shows compatibility matrix"
|
|
|
|
echo ""
|
|
print_default_option() { printf " --%-20s %-24s %s (default: %s)\n" "${1}" "${2}" "${3}" "$(tolower ${4})"; }
|
|
print_default_option distro "[ubuntu|opensuse|rhel]" "OS distribution" "${DISTRO}"
|
|
print_default_option versions "[VERSION] [VERSION...]" "Ubuntu, OpenSUSE, or RHEL release" "${VERSIONS}"
|
|
print_default_option rocm-versions "[VERSION] [VERSION...]" "ROCm versions (format: Major.Minor.Patch, patch defaults to 0 if not specified)" "${ROCM_VERSIONS}"
|
|
print_default_option python-versions "[VERSION] [VERSION...]" "Python 3 minor releases" "${PYTHON_VERSIONS}"
|
|
print_default_option "user -u" "[USERNAME]" "DockerHub username" "${USER}"
|
|
print_default_option "retry -r" "[N]" "Number of attempts to build (to account for network errors)" "${RETRY}"
|
|
print_default_option push "" "Push the image to Dockerhub" ""
|
|
#print_default_option lto "[on|off]" "Enable LTO" "${LTO}"
|
|
}
|
|
|
|
send-error()
|
|
{
|
|
usage
|
|
echo -e "\nError: ${@}"
|
|
exit 1
|
|
}
|
|
|
|
verbose-run()
|
|
{
|
|
echo -e "\n### Executing \"${@}\"... ###\n"
|
|
eval "${@}"
|
|
}
|
|
|
|
verbose-build()
|
|
{
|
|
echo -e "\n### Executing \"${@}\" a maximum of ${RETRY} times... ###\n"
|
|
for i in $(seq 1 1 ${RETRY})
|
|
do
|
|
set +e
|
|
eval "${@}"
|
|
local RETC=$?
|
|
set -e
|
|
if [ "${RETC}" -eq 0 ]; then
|
|
break
|
|
else
|
|
echo -en "\n### Command failed with error code ${RETC}... "
|
|
if [ "${i}" -ne "${RETRY}" ]; then
|
|
echo -e "Retrying... ###\n"
|
|
sleep 3
|
|
else
|
|
echo -e "Exiting... ###\n"
|
|
exit ${RETC}
|
|
fi
|
|
fi
|
|
done
|
|
}
|
|
|
|
reset-last()
|
|
{
|
|
last() { send-error "Unsupported argument :: ${1}"; }
|
|
}
|
|
|
|
reset-last
|
|
load-matrix
|
|
n=0
|
|
while [[ $# -gt 0 ]]
|
|
do
|
|
case "${1}" in
|
|
-h|--help)
|
|
usage
|
|
exit 0
|
|
;;
|
|
-m|--matrix)
|
|
shift
|
|
if [[ $# -gt 0 && ! "${1}" =~ ^-- ]]; then
|
|
show-matrix "${1}"
|
|
else
|
|
show-matrix
|
|
fi
|
|
exit 0
|
|
;;
|
|
"--distro")
|
|
shift
|
|
DISTRO=$(tolower ${1})
|
|
last() { DISTRO="${DISTRO} $(tolower${1})"; }
|
|
;;
|
|
"--versions")
|
|
shift
|
|
VERSIONS=${1}
|
|
last() { VERSIONS="${VERSIONS} ${1}"; }
|
|
;;
|
|
"--rocm-versions")
|
|
shift
|
|
ROCM_VERSIONS=${1}
|
|
last() { ROCM_VERSIONS="${ROCM_VERSIONS} ${1}"; }
|
|
;;
|
|
"--python-versions")
|
|
shift
|
|
PYTHON_VERSIONS=${1}
|
|
last() { PYTHON_VERSIONS="${PYTHON_VERSIONS} ${1}"; }
|
|
;;
|
|
--user|-u)
|
|
shift
|
|
USER=${1}
|
|
reset-last
|
|
;;
|
|
--push)
|
|
PUSH=1
|
|
reset-last
|
|
;;
|
|
--no-pull)
|
|
PULL=""
|
|
reset-last
|
|
;;
|
|
--retry|-r)
|
|
shift
|
|
RETRY=${1}
|
|
reset-last
|
|
;;
|
|
"--*")
|
|
send-error "Unsupported argument at position $((${n} + 1)) :: ${1}"
|
|
;;
|
|
*)
|
|
last ${1}
|
|
;;
|
|
esac
|
|
n=$((${n} + 1))
|
|
shift
|
|
done
|
|
|
|
# Validate input parameters for os-distros and rocm-versions
|
|
validate-distro
|
|
validate-combinations
|
|
|
|
DOCKER_FILE="Dockerfile.${DISTRO}"
|
|
|
|
if [ "${RETRY}" -lt 1 ]; then
|
|
RETRY=1
|
|
fi
|
|
|
|
if [ -n "${BUILD_CI}" ]; then DOCKER_FILE="${DOCKER_FILE}.ci"; fi
|
|
cd docker # Forced since PWD is parent dir
|
|
if [ ! -f ${DOCKER_FILE} ]; then send-error "File \"${DOCKER_FILE}\" not found"; fi
|
|
|
|
for VERSION in ${VERSIONS}
|
|
do
|
|
VERSION_MAJOR=$(echo ${VERSION} | sed 's/\./ /g' | awk '{print $1}')
|
|
VERSION_MINOR=$(echo ${VERSION} | sed 's/\./ /g' | awk '{print $2}')
|
|
VERSION_PATCH=$(echo ${VERSION} | sed 's/\./ /g' | awk '{print $3}')
|
|
for ROCM_VERSION in ${ROCM_VERSIONS}
|
|
do
|
|
# Parse ROCm version components
|
|
ROCM_MAJOR=$(echo ${ROCM_VERSION} | sed 's/\./ /g' | awk '{print $1}')
|
|
ROCM_MINOR=$(echo ${ROCM_VERSION} | sed 's/\./ /g' | awk '{print $2}')
|
|
ROCM_PATCH=$(echo ${ROCM_VERSION} | sed 's/\./ /g' | awk '{print $3}')
|
|
|
|
# Compute ROCM_VERSION_URL (for repo URLs) and normalize ROCM_PATCH
|
|
ROCM_VERSION_URL="${ROCM_VERSION}"
|
|
if [ -z "${ROCM_PATCH}" ]; then
|
|
ROCM_PATCH=0
|
|
elif [ "${ROCM_PATCH}" = "0" ]; then
|
|
ROCM_VERSION_URL="${ROCM_MAJOR}.${ROCM_MINOR}"
|
|
fi
|
|
|
|
# Compute ROCM_VERSN (numeric version for package names)
|
|
ROCM_VERSN=$(( ("${ROCM_MAJOR}"*10000)+("${ROCM_MINOR}"*100) + ("${ROCM_PATCH}")))
|
|
|
|
# Determine container tag
|
|
if [ "${ROCM_PATCH}" = "0" ]; then
|
|
CONTAINER=${USER}/rocprofiler-systems:release-base-${DISTRO}-${VERSION}-rocm-${ROCM_MAJOR}.${ROCM_MINOR}
|
|
else
|
|
CONTAINER=${USER}/rocprofiler-systems:release-base-${DISTRO}-${VERSION}-rocm-${ROCM_VERSION}
|
|
fi
|
|
if [ "${DISTRO}" = "ubuntu" ]; then
|
|
case "${VERSION}" in
|
|
24.04)
|
|
ROCM_REPO_DIST="noble"
|
|
;;
|
|
22.04)
|
|
ROCM_REPO_DIST="jammy"
|
|
;;
|
|
*)
|
|
;;
|
|
esac
|
|
verbose-build docker build . ${PULL} --progress plain -f ${DOCKER_FILE} --tag ${CONTAINER} --build-arg DISTRO=${DISTRO} --build-arg VERSION=${VERSION} --build-arg ROCM_VERSION=${ROCM_VERSION} --build-arg ROCM_MAJOR=${ROCM_MAJOR} --build-arg ROCM_MINOR=${ROCM_MINOR} --build-arg ROCM_PATCH=${ROCM_PATCH} --build-arg ROCM_VERSION_URL=${ROCM_VERSION_URL} --build-arg ROCM_VERSN=${ROCM_VERSN} --build-arg PYTHON_VERSIONS=\"${PYTHON_VERSIONS}\"
|
|
elif [ "${DISTRO}" = "rhel" ]; then
|
|
# use Rocky Linux as a base image for RHEL builds
|
|
DISTRO_BASE_IMAGE=rockylinux/rockylinux
|
|
verbose-build docker build . ${PULL} --progress plain -f ${DOCKER_FILE} --tag ${CONTAINER} --build-arg DISTRO=${DISTRO_BASE_IMAGE} --build-arg VERSION=${VERSION} --build-arg ROCM_VERSION=${ROCM_VERSION} --build-arg ROCM_MAJOR=${ROCM_MAJOR} --build-arg ROCM_MINOR=${ROCM_MINOR} --build-arg ROCM_PATCH=${ROCM_PATCH} --build-arg ROCM_VERSION_URL=${ROCM_VERSION_URL} --build-arg ROCM_VERSN=${ROCM_VERSN} --build-arg PYTHON_VERSIONS=\"${PYTHON_VERSIONS}\"
|
|
elif [ "${DISTRO}" = "opensuse" ]; then
|
|
DISTRO_IMAGE="opensuse/leap"
|
|
if [[ "${VERSION_MAJOR}" -le 15 && "${VERSION_MINOR}" -le 5 ]]; then
|
|
PERL_REPO="15.6"
|
|
else
|
|
PERL_REPO="${VERSION_MAJOR}.${VERSION_MINOR}"
|
|
fi
|
|
verbose-build docker build . ${PULL} --progress plain -f ${DOCKER_FILE} --tag ${CONTAINER} --build-arg DISTRO=${DISTRO_IMAGE} --build-arg VERSION=${VERSION} --build-arg ROCM_VERSION=${ROCM_VERSION} --build-arg ROCM_MAJOR=${ROCM_MAJOR} --build-arg ROCM_MINOR=${ROCM_MINOR} --build-arg ROCM_PATCH=${ROCM_PATCH} --build-arg ROCM_VERSION_URL=${ROCM_VERSION_URL} --build-arg ROCM_VERSN=${ROCM_VERSN} --build-arg PERL_REPO=${PERL_REPO} --build-arg PYTHON_VERSIONS=\"${PYTHON_VERSIONS}\"
|
|
fi
|
|
if [ "${PUSH}" -ne 0 ]; then
|
|
docker push ${CONTAINER}
|
|
fi
|
|
done
|
|
done
|