NetworkManager segfaults when starting and an interface is up

[Walter]

This is the kernel output I get when restarting NetworkManager with an interface being up. This is reproducible across all of my machines. I am running Funtoo with gentoo sources 4.13.9 and have been able to reproduce this problem for over 6 months.

[Thu Oct 26 22:40:36 2017] NetworkManager[484]: segfault at 8 ip 0000000000479ef7 sp 00007fffb903eb60 error 4 in NetworkManager[400000+23b000]
[Thu Oct 26 22:46:40 2017] NetworkManager[9577]: segfault at 8 ip 0000000000479ef7 sp 00007fffefab0200 error 4 in NetworkManager[400000+23b000]
[Thu Oct 26 22:48:37 2017] NetworkManager[9936]: segfault at 8 ip 0000000000479ef7 sp 00007ffd46445950 error 4 in NetworkManager[400000+23b000]
[Thu Oct 26 23:16:19 2017] NetworkManager[4029]: segfault at 8 ip 0000000000479ef7 sp 00007fff07455bc0 error 4 in NetworkManager[400000+23b000]
[Thu Oct 26 23:17:23 2017] NetworkManager[4474]: segfault at 8 ip 0000000000479ef7 sp 00007ffd73cca230 error 4 in NetworkManager[400000+23b000]

[Walter (1os843jpl0fa6sl1lwb3gyvu73nyetct)]

I am also running NetworkManager 1.4.4-r1.

[Walter (1os843jpl0fa6sl1lwb3gyvu73nyetct)]

I see this repeatedly in the logs. Just prior to the segfault, NetworkManager is clearing out DNS information.

Oct 26 22:40:37 localhost NetworkManager[484]: <info> [1509072037.0735] dns-mgr: Removing DNS information from /sbin/resolvconf
Oct 26 22:40:37 localhost kernel: NetworkManager[484]: segfault at 8 ip 0000000000479ef7 sp 00007fffb903eb60 error 4 in NetworkManager[400000+23b000]

[Walter (1os843jpl0fa6sl1lwb3gyvu73nyetct)]

I think I found the issue:

#cat /etc/resolvconf.conf
# Use the local name server
name_servers=127.0.0.1

# added for dnscrypt + dnssec
options edns0

The options edns0 is clearly invalid, so, once I commented it, it now works. I can restart NetworkManager as often as I like and it appears to come up just fine.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.