check: run tests in a systemd scope for mandatory test cleanup

TLDR: If systemd is available, run each test in its own temporary
systemd scope.  This enables the test harness to forcibly clean up all
of the test's child processes (if it does not do so itself) so that we
can move into the post-test unmount and check cleanly.

I frequently run fstests in "low" memory situations (2GB!) to force the
kernel to do interesting things.  There are a few tests like generic/224
and generic/561 that put processes in the background and occasionally
trigger the OOM killer.  Most of the time the OOM killer correctly
shoots down fsstress or duperemove, but once in a while it's stupid
enough to shoot down the test control process (i.e. tests/generic/224)
instead.  fsstress is still running in the background, and the one
process that knew about that is dead.

When the control process dies, ./check moves on to the post-test fsck,
which fails because fsstress is still running and we can't unmount.
After fsck fails, ./check moves on to the next test, which fails because
fsstress is /still/ writing to the filesystem and we can't unmount or
format.

The end result is that that one OOM kill causes cascading test failures,
and I have to re-start fstests to see if I get a clean(er) run.

So, the solution I present in this patch is to teach ./check to try to
run the test script in a systemd scope.  If that succeeds, ./check will
tell systemd to kill the scope when the test script exits and returns
control to ./check.  Concretely, this means that systemd creates a new
cgroup, stuffs the processes in that cgroup, and when we kill the scope,
systemd kills all the processes in that cgroup and deletes the cgroup.

The end result is that fstests now has an easy way to ensure that /all/
child processes of a test are dead before we try to unmount the test and
scratch devices.  I've designed this to be optional, because not
everyone does or wants or likes to run systemd, but it makes QA easier.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Eryu Guan <guaneryu@gmail.com>
Signed-off-by: Eryu Guan <guaneryu@gmail.com>

diff --git a/check b/check
index 5072dd82c1853f86593131a6673bf52d4cf3b4c3..83f6fc8bdf3ef24f6af0cd5b2570f5bd2ae030dc 100755 (executable)
--- a/check
+++ b/check
@@ -521,6 +521,12 @@ _expunge_test()
        return 0
 }
 
+# Can we run systemd scopes?
+HAVE_SYSTEMD_SCOPES=
+systemctl reset-failed "fstests-check" &>/dev/null
+systemd-run --quiet --unit "fstests-check" --scope bash -c "exit 77" &> /dev/null
+test $? -eq 77 && HAVE_SYSTEMD_SCOPES=yes
+
 # Make the check script unattractive to the OOM killer...
 OOM_SCORE_ADJ="/proc/self/oom_score_adj"
 test -w ${OOM_SCORE_ADJ} && echo -1000 > ${OOM_SCORE_ADJ}
@@ -528,8 +534,26 @@ test -w ${OOM_SCORE_ADJ} && echo -1000 > ${OOM_SCORE_ADJ}
 # ...and make the tests themselves somewhat more attractive to it, so that if
 # the system runs out of memory it'll be the test that gets killed and not the
 # test framework.
+#
+# If systemd is available, run the entire test script in a scope so that we can
+# kill all subprocesses of the test if it fails to clean up after itself.  This
+# is essential for ensuring that the post-test unmount succeeds.  Note that
+# systemd doesn't automatically remove transient scopes that fail to terminate
+# when systemd tells them to terminate (e.g. programs stuck in D state when
+# systemd sends SIGKILL), so we use reset-failed to tear down the scope.
 _run_seq() {
-       bash -c "test -w ${OOM_SCORE_ADJ} && echo 250 > ${OOM_SCORE_ADJ}; exec ./$seq"
+       local cmd=(bash -c "test -w ${OOM_SCORE_ADJ} && echo 250 > ${OOM_SCORE_ADJ}; exec ./$seq")
+
+       if [ -n "${HAVE_SYSTEMD_SCOPES}" ]; then
+               local unit="$(systemd-escape "fs$seq").scope"
+               systemctl reset-failed "${unit}" &> /dev/null
+               systemd-run --quiet --unit "${unit}" --scope "${cmd[@]}"
+               res=$?
+               systemctl stop "${unit}" &> /dev/null
+               return "${res}"
+       else
+               "${cmd[@]}"
+       fi
 }
 
 _detect_kmemleak