Bug#781210: systemd asserts on function cg_is_empty_recursive, crashes

Martin Pitt mpitt at debian.org
Sat Mar 28 07:25:56 GMT 2015


Hello all,

I ran the bisect today, and it seems this commit fixes it:

  http://cgit.freedesktop.org/systemd/systemd/commit/?id=7b3fd6313c4

I'll backport it to 215 ASAP and verify that it still works.

For the record, this is my autopkgtest:

====== debian/tests/control =============
Tests: t
Depends:
Restrictions: needs-root

====== debian/tests/t =============
#!/bin/sh -e

fail() {
	echo "ERROR: $1" >&2
	exit 1
}

# install local systemd replacement
if [ -e /tmp/systemd ]; then
    rm -f /lib/systemd/systemd
    cp /tmp/systemd /lib/systemd/systemd
    rm /tmp/systemd
    echo "Rebooting with new systemd..."
    /tmp/autopkgtest-reboot boot1
fi

DEBIAN_FRONTEND=noninteractive apt-get install -y strongswan

ps aux|grep ipsec
PS=`pidof starter` || fail "strongswan starter not running"
PC=`pidof charon` || fail "charon not running"
#systemctl stop strongswan
kill $PS $PC 2>&1|| true
sleep 1
[ ! -d /proc/$PS ] || fail "starter still running"
[ ! -d /proc/$PC ] || fail "charon still running"
if journalctl -b | grep 'dumped core' >&2; then
    fail "found crash"
fi
set +e
systemctl status strongswan
RC=$?
set -e
[ $RC -eq 3 ] || fail "unexpected status code $RC"
====================================


Martin

-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)




More information about the Pkg-systemd-maintainers mailing list