OpenBSD Journal

a2k17 hackathon report: Kenneth Westerback on the hidden wonders of the build system, the network stack and more

Contributed by pitrh on from the make make make things right dept.

Still fresh from the just completed hackathon down under Ken Westerback writes,
The trip to Brisbane was much nicer on the new direct flight from Vancouver, which saved me one airport traversal. And the high tech automated border control process was very cool. Whereas Brisbane itself was very hot.
Retreating from the heat into the hackroom I quickly cleaned out a collection of M's that were in my tree as a result of spending some time reading the warning messages generated during a system build. No actual bugs were found but some surprising code was accidentally uncovered and exposed to the light of day. e.g. an enterprising attempt in make(1) to remove files using 'sudo' if they resisted a first try. This was forcibly expunged by Theo.

I then trolled through most of the directories I had touched and removed a raft of superfluous whitespace. Thus saving the eyesight of future spelunkers.

By the time this was complete the paperwork had been processed for allowing beer in the hackroom and serious coding could begin.

My mission this time was to move forward on a new experiment in improving the management of the system network configuration. The approach was defined in Cambridge in September but I spent the intervening months figuring out how to write a proper OpenBSD daemon. About which I turned out to be wrong on every count until I went back and re-read old papers and new code.

That done, I used a2k17 to poke and prod the network stack into accepting a new routing socket message, called RTM_PROPOSAL for the moment. I updated 'route monitor' so it could display the contents of this message and I hacked dhclient so it issued the new message when it wanted to configure its interface. Which I then could see appearing in route monitor.

I put up a "Mission Accomplished!" email, updated my tree to pull in the latest work others were committing, booted a new kernel and was staring at a ddb prompt. A quick trip to the next table and claudio@ pointed out that his latest change to the routing socket code meant I had to rework my changes a bit. I did so and sent out an updated "Mission Accomplished!" email. Unfortunately I was not quick enough to save florian@ from visiting the same ddb prompt.

A skeletal daemon to receive and consider the RTM_PROPOSAL message was next. And then florian@ started to make progress on getting his slaacd daemon to use RTM_PROPOSAL messages. If all continues to go well this all could be hitting the tree soon.

And then a2k17 was over.

Thanks to dlg@ and jmatthew@ and their helpers for another great hackathon at UQ!

(Comments are closed)


Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]