[DanielAdolfsson]

Stacked on #96 — review #95 and #96 first.

README

The rewrite that came with #93 is the base and most of it is kept. This says the things it could not, because it was written for a fork.

Status, and credit where it is due

For six years this project's public answer to "which version should I use" was that the current one was in maintenance, its replacement was not ready for production, and there was no estimate. That is why the reporters in #71 concluded ndppd was abandoned and moved to ndpresponder.

The honest answer now is different, and the README says it: 1.0 is in beta, is running in production, and is being worked on again.

It is being worked on because of @totoCZ. The rewrite sat untouched from June 2020 until they took it into production behind a Juniper switch, worked out everything that was wrong with it, and sent the fixes back — #92 and #93. Several are the kind of bug only a real network finds:

• a solicited-node multicast MAC wrong by one byte, so every EUI-64 target silently never received a probe;
• a socket membership dropped by the parent process at fork, leaving the surviving daemon unable to see the packets it exists to answer;
• an NA correct in every respect except the one flag that would have let a neighbour cache believe it.

That is in the Status section and again under Credits, by name.

What ndppd does not do

A new Limitations section: ndppd does not route, does not enable forwarding, does not send RAs, is not a substitute for the kernel's neighbour table on its own host, is Ethernet-only, is defeated by MLD snooping, and needs its interfaces to exist at startup.

That list is most of the open tracker — #55, #69, #72, #77, #79 — and #60 has been open since June 2020 asking for exactly this. Each item links the issue it corresponds to.

The rest

• How to choose a rule mode, the decision people most often get wrong, including what static really commits you to: a static /64 claims eighteen quintillion addresses on your provider's link.
• gmake on BSD, which answers How to make under FreeBSD #66 — the Makefile opens with ifdef and BSD make replies "Need an operator".
• A worked rewrite example, which is what ndppd with unique local IPv6 addresses on the internal interface #77 is really asking for.

Man pages

ndppd.conf(5) documented invalid-ttl as defaulting to 5000 when it is 10000, used route eth0 where it meant proxy eth0, and was missing keepalive, router, target and rewrite entirely.

Bounds are now stated for every numeric option, because the parser rejects out-of-range values rather than clamping them, and the error does not say what the range was.

ndppd(8) gained --netns, what the verbosity levels actually are, and a note on promiscuous mode and on log buffering.

The shipped systemd unit could not work

ndppd(8) documents /etc/ndppd.conf as the default configuration path. The compiled-in default was the relative path "../ndppd.conf", and ndppd.service passes no -c at all. systemd runs a service with its working directory at /, so the shipped unit looked for /ndppd.conf:

info  | Loading configuration "../ndppd.conf"...
error | Failed to load configuration: No such file or directory

The default is now built from SYSCONFDIR, so a packager who moves it moves the documentation's claim with it, and 110-config-path.sh asserts the binary and its own man page name the same path — whatever that path is.

---

Relevance to the open tracker

• Documentation is missing #60 (documentation is missing) — largely addressed for 1.0. The Limitations section is the part that was actually missing.
• How to make under FreeBSD #66 (building on FreeBSD) — answered: gmake.
• ndppd with unique local IPv6 addresses on the internal interface #77 (ULAs on the internal interface) — answered: ndppd does not route, and rewrite is what this user wants.
• MLD Snooping will cause the NDP Proxy to fail. We need the MLD Proxy. #55, ndppd does receive locally generated neighbor discovery #69, ndppd and tun/wg interfaces #72, ndppd does not wait for docker interfaces #79 — documented as known limitations with links, not fixed. They remain open.

These issues were all filed against 0.2.x, and the README changes here are on 1.0-devel only. One loose end for master: its README still tells every visitor that 1.0 "is not yet stable enough to be used in production, and I currently have no estimate when it is". That has been the project's public answer since 2020 and is now wrong in the other direction. A paragraph on master would stop the two branches contradicting each other.

Verification

make check: 228 unit + 55 integration assertions, exit 0. Man pages render with asciidoctor -b manpage.

🤖 Generated with Claude Code