Allright. I was going to post in-depth investigation results to sum up what stepping through the code taught me (pretty much nothing, it all goes well until I lose the stepping thread due to callback/async stuff), but right before sending the text I let GDB finish the program execution, and lo and behold, some DHCP6 traffic appeared in my tcpdump terminal, and the tests finally pass. I'm thus assuming that some timeout somewhere expired while I was stepping through the code (I of course dialed up all the test timeouts).
I thus suspect that the bug lies within this call and its side-effects : https://git.launchpad.net/ubuntu/+source/systemd/tree/src/libsystemd- network/sd-dhcp6-client.c?h=applied/ubuntu/impish-proposed#n1648 -- You received this bug notification because you are a member of Ubuntu Touch seeded packages, which is subscribed to systemd in Ubuntu. https://bugs.launchpad.net/bugs/1940635 Title: systemd-networkd failing to acquire a DHCP6 lease from dnsmasq on armhf Status in glibc package in Ubuntu: New Status in netplan.io package in Ubuntu: New Status in systemd package in Ubuntu: New Bug description: systemd-networkd is failing to acquire a DCHP6 lease from dnsmasq on armhf since glibc 2.34-0ubuntu1, failing systemd (tests-name=networkd- test.py) and netplan.io (test-name=ethernets) tests on armhf. Reproducer: * Setup an armhf container, e.g. via: autopkgtest systemd --test-name=networkd-test.py --shell -U --apt-pocket=proposed=src:systemd,src:glibc -s -- lxd autopkgtest/ubuntu/impish/armhf * It will fail and drop you into the shell * cd test/ && apt install python3-nose * nosetests3 -v -s -m "test_.*_dhcp_ip6" networkd-test.py This is unrelated to the recent dnsmasq changes (LP: #1894619), as that would fail on all architectures, not just armhf. It still passes inside an amd64 LXD container, so it is also not related to armhf being tested inside a container. But it shows the difference between armhf vs amd64 container, that on armhf we're missing the DHCPSOLICIT (and therefore DHCPREPLY) messages, as can be seen from the dnsmasq log: dnsmasq-dhcp[]: DHCPSOLICIT(router_eth42) 00:02:00:00:ab:11:57:1e:20:2f:9e:56:5f:34 dnsmasq-dhcp[]: DHCPREPLY(router_eth42) 2600::1f 00:02:00:00:ab:11:57:1e:20:2f:9e:56:5f:34 autopkgtest-lxd-rtypaf The issue is most probably the same that causes the currently failing netplan.io/ethernets autopkgtest on armhf, if tested against glibc 2.34-0ubuntu1. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/glibc/+bug/1940635/+subscriptions -- Mailing list: https://launchpad.net/~touch-packages Post to : touch-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~touch-packages More help : https://help.launchpad.net/ListHelp
