From the ISC web site [LWN.net]

From the ISC web site

Posted May 30, 2025 10:08 UTC (Fri) by joib (subscriber, #8541)
In reply to: From the ISC web site by rschroev
Parent article: Local vulnerabilities in Kea DHCP

There's dhcpcd (used in Ubuntu initramfs, presumably Debian and other Debian-based distros as well). Not sure what other distros are using for initramfs needs. And NetworkManager as well as systemd-networkd have their own internal clients, which I guess is what most distros are using by default.

I'm not aware of any maintained open source relay agent. One particular brand of switches I'm aware of hasn't migrated off isc dhcp relay, presumably the vendor is supporting it themselves.

to post comments

From the ISC web site

Posted May 30, 2025 16:46 UTC (Fri) by auerswal (subscriber, #119876) [Link] (1 responses)

Dnsmasq (https://dnsmasq.org/doc.html) contains DHCP relay functionality, is available under the GPL, version 2 or version 3, and maintained.

From the ISC web site

Posted May 30, 2025 18:38 UTC (Fri) by joib (subscriber, #8541) [Link]

Oh! I'm familiar with dnsmasq, but I didn't realize it provides dhcp relay functionality as well. Thanks!

From the ISC web site

Posted Jun 1, 2025 13:18 UTC (Sun) by bferrell (subscriber, #624) [Link] (1 responses)

from my previous post:

I spent a full 36 hours troubleshooting an odd behavior... If ISC dhcp is set to send extra routes, the last extra route ends up the default route for dhcpcd. A lot of fiddling with ISC dhcpd will get dhcpcd to work right... But then it breaks dhcp in iphones, android devices, "smart" plugs and Windows.

The machine that prompted me to investigate (a Raspberry Pi) had dhcpcd removed and the ISC DHCP client installed.

Yes, I DID discuss it with the dev and gave up. It does faithfully support an obscure RFC. But the RFC is broken (as far as I'm concerned).

From the ISC web site

Posted Jun 2, 2025 15:18 UTC (Mon) by hmh (subscriber, #3838) [Link]

This is interesting... can you post some details? a link to a bug report, or the reference to the "obscure RFC", etc ?