Support cgroup v2 `Unified Mode`

[Pijukatel]

The system should be able to correctly report the maximum available memory when restricted through cgroups.

Scenario examples:

1 Example invocation without a container:
systemd-run --user --scope --quiet --no-ask-password -p MemoryMax=50M -p MemorySwapMax=0 python3 sensor.py

Should report 50M

2 Example invocation with container:
sudo docker run --rm --memory=50m --memory-swap=50m -v "$(pwd)/sensor.py:/sensor.py:ro" python:3.12-slim python3 /sensor.py

Should report 50M

3 Restricted container, but even more restricted script:

  sudo docker run --rm \
    --memory=50m --memory-swap=50m \
    --cgroupns=private --privileged \
    -v "$(pwd)/sensor.py:/sensor.py:ro" \
    python:3.12-slim \
    bash -c '
      mount -o remount,rw /sys/fs/cgroup 2>/dev/null || true
      mkdir -p /sys/fs/cgroup/init /sys/fs/cgroup/inner
      echo $$ > /sys/fs/cgroup/init/cgroup.procs          # vacate the namespace-root cgroup
      echo +memory > /sys/fs/cgroup/cgroup.subtree_control # now delegation is allowed
      echo 30M > /sys/fs/cgroup/inner/memory.max
      echo 0   > /sys/fs/cgroup/inner/memory.swap.max
      echo $$ > /sys/fs/cgroup/inner/cgroup.procs          # move shell (→python) into the tight cgroup
      exec python3 /sensor.py
    '

Should report 30M