cgroup_fj_common.sh: Do not disable systemd related cgroup subsystems

The cpu, io, memory, and pids subsystems under the root cgroup cannot be disabled
because they are used for systemd. Currently, the test of cpu and memory subsystems
in cgroup_fj_function.sh and cgroup_fj_stress.sh will report on same
systems (Fedora 37, Debian trixie) an error:

    echo: write error: Device or resource busy
    cgroup_fj_function 10 TWARN: echo -cpu > /sys/fs/cgroup/cgroup.subtree_control failed

Link: https://lore.kernel.org/ltp/20240809073010.793324-1-maxj.fnst@fujitsu.com/
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Ma Xinjian <maxj.fnst@fujitsu.com>

testcases/kernel/controllers/cgroup_fj/cgroup_fj_common.sh

@@ -77,7 +77,10 @@ common_cleanup()
 
     cgroup_cleanup
 
-    [ "$cgroup_version" = "2" ] && ROD echo "-$subsystem" \> "/sys/fs/cgroup/cgroup.subtree_control"
+    if [ "$cgroup_version" = "2" ] && [ "$subsystem" != "cpu" ] && [ "$subsystem" != "io" ] \
+            && [ "$subsystem" != "memory" ] && [ "$subsystem" != "pids" ]; then
+            ROD echo "-$subsystem" \> "/sys/fs/cgroup/cgroup.subtree_control"
+    fi
 }
 
 . cgroup_lib.sh