mirror of
https://github.com/flatcar/scripts.git
synced 2025-08-07 04:56:58 +02:00
834e0c11d7
When started by the Flatcar core user, the SDK failed to use UID 500 because inside the SDK there already is the core user from nss-altfiles with the same ID. This way, the SDK user was continuing with UID 1000 and had permission errors. Allow to reuse an existing ID for the SDK user. However, this only works when usermod doesn't find a process that uses this ID, and we had a race between the SDK entry points called by "docker start" and by "docker exec". The race is unwanted anyway because we don't want to execute the commands while setup_board is still running. Solve it by setting the entrypoint for "docker start" directly to "bash -l" in "docker create" (this is also what the entry point does as last step: sudo su -l).
74 lines
2.6 KiB
Bash
Executable File
74 lines
2.6 KiB
Bash
Executable File
#!/bin/bash
|
|
|
|
if [ -n "${SDK_USER_ID:-}" ] ; then
|
|
# If the "core" user from /usr/share/baselayout/passwd has the same ID, allow to take it instead
|
|
usermod --non-unique -u $SDK_USER_ID sdk
|
|
fi
|
|
if [ -n "${SDK_GROUP_ID:-}" ] ; then
|
|
groupmod --non-unique -g $SDK_GROUP_ID sdk
|
|
fi
|
|
|
|
chown -R sdk:sdk /home/sdk
|
|
|
|
# Check if the OS image version we're working on is newer than
|
|
# the SDK container version and if it is, update the boards
|
|
# chroot portage conf to point to the correct binhost.
|
|
(
|
|
source /etc/lsb-release # SDK version in DISTRIB_RELEASE
|
|
source /mnt/host/source/.repo/manifests/version.txt # OS image version in FLATCAR_VERSION_ID
|
|
version="${FLATCAR_VERSION_ID}"
|
|
|
|
# If this is a nightly build tag we can use pre-built binaries directly from the
|
|
# build cache.
|
|
if [[ "${FLATCAR_BUILD_ID}" =~ ^nightly-.*$ ]] ; then
|
|
version="${FLATCAR_VERSION_ID}+${FLATCAR_BUILD_ID}"
|
|
fi
|
|
|
|
if [ "${version}" != "${DISTRIB_RELEASE}" ] ; then
|
|
for target in amd64-usr arm64-usr; do
|
|
if [ ! -d "/build/$target" ] ; then
|
|
continue
|
|
fi
|
|
if [ -f "/build/$target/etc/target-version.txt" ] ; then
|
|
source "/build/$target/etc/target-version.txt"
|
|
if [ "${TARGET_FLATCAR_VERSION}" = "${version}" ] ; then
|
|
continue # already updated
|
|
fi
|
|
fi
|
|
|
|
echo
|
|
echo "Updating board support in '/build/${target}' to use package cache for version '${version}'"
|
|
echo "---"
|
|
sudo su sdk -l -c "/home/sdk/trunk/src/scripts/setup_board --board='$target' --regen_configs_only"
|
|
echo "TARGET_FLATCAR_VERSION='${version}'" | sudo tee "/build/$target/etc/target-version.txt" >/dev/null
|
|
done
|
|
fi
|
|
)
|
|
|
|
# This is ugly.
|
|
# We need to sudo su - sdk -c so the SDK user gets a fresh login.
|
|
# 'sdk' is member of multiple groups, and plain docker USER only
|
|
# allows specifying membership of a single group.
|
|
# When a command is passed to the container, we run, respectively:
|
|
# sudo su - sdk -c "<command>".
|
|
# Then, we need to preserve whitespaces in arguments of commands
|
|
# passed to the container, e.g.
|
|
# ./update_chroot --toolchain_boards="amd64-usr arm64-usr".
|
|
# This is done via a separate ".cmd" file since we have used up
|
|
# our quotes for su -c "<cmd>" already.
|
|
if [ $# -gt 0 ] ; then
|
|
cmd="/home/sdk/.cmd"
|
|
echo -n "exec bash -l -i -c '" >"$cmd"
|
|
for arg in "$@"; do
|
|
echo -n "\"$arg\" " >>"$cmd"
|
|
done
|
|
echo "'" >>"$cmd"
|
|
chmod 755 "$cmd"
|
|
sudo su sdk -c "$cmd"
|
|
rc=$?
|
|
rm -f "$cmd"
|
|
exit $rc
|
|
else
|
|
exec sudo su -l sdk
|
|
fi
|