Contributed by pitrh on from the fuzzing my clobberage dept.
Since some years now, I cannot attend a hackathon without having to deal with a USB problem. It's generally a bug that one of the hackers in the room can expose with his hardware. But in Cambridge it was different!
krw@ brought with him a couple of USB emulators/fuzzers so I started the hackathon by waiting, for the USB stack to panic. Even if fuzzing the stack is fun, because you start liking panics, that's something I can do alone & anywhere, so I moved on.After fixing a couple of bugs related to the routing table, I spent a lot of time discussing with claudio@ what should be the next step in this area. We discussed many times during the week about caching route entries, interaction with userland, SMP issues, memory consumption....
At this point I should realize that the good old hackathon saying "one discussion bring you to another" describes perfectly my week. I started discussing the routing table, again mostly for SMP purposes with bluhm@.
Then I jumped to input/touchpad issues with bru@, matthieu@ and jcs@.
In between I discussed CTF, Compact ANSI-C Type Format, integration in ddb(4) with jasper@. This even resulted in a couple of bug fixes in my ctfdump(1) implementation.
On the next day I had a discussion about DHCP and RTSOL, some crazy stuff is being prepared!
Then I managed to rewrite my Dynamic Profiling diff and to get it in with the help of guenther@. This should allow us to get the ddb(4) subsystem safe to be used as foundation for a tracing tool. Yes, this is related to jasper@'s work on CTF which is the format used by DTrace.
Finally I had nice talks with awolk@ who found some nice races in the hotplug logic of the Network Stack.
I'm certainly missing a lot of stuff, it was a really nice and busy hackathon for me. Thanks a lot to Gemma, Anil, the OpenBSD foundation and everybody else who helped organizing it!
Thanks for the report, Martin!
(Comments are closed)