Contributed by Paul 'WEiRD' de Weerd on from the low-level-lore dept.
As more and more developers arrive back home from France, more reports arrive to you to keep you informed of what happened in Nantes. This time, it's Philip Guenther (guenther@) who writes in with his report:
Non-OpenBSD life had led to me not being able to make a2k18, so Bob cajoled me into booking tickets for p2k18. I'm not a porter, but it's useful if they have another opportunity to
yell at mepolitely request fixes or missing functionality that would make porting easier.
Once there, I continued to work on making amd64's low-level trap handling simpler and clearer. The Meltdown work in Jan+Feb and the discussions then with developers from the other BSDs and Illumos had helped me understand some of the darker corners. A number of simplifications have now gotten in and there's one more in the works, though it causes hangs on one of visa@'s boxes that I still need to diagnose.
While doing that I served as sounding board for Theo and Bob's continued pledge enhancements, Mike's VMM work, and a dhclient glitch that Ken looked at for me.
Finally, on the penultimate evening at the bar, Robert asked about support for clock_nanosleep() and Landry asked about the timer_*() family of functions. I had done a proto-diff of clock_nanosleep() years ago but didn't get traction on getting it in then. Since then we've implemented more user-visible clocks (CLOCK_UPTIME, CLOCK_PROCESS_CPUTIME_ID, and per-thread CPU time) so it needs some love to get back into working order. So it goes: forward progress isn't free.
Nantes was beautiful, with mild weather, good food, and the crafting of Les Machines de L'ile was impressive. Thank you, Epitech, for the hosting, and Gilles for logistics!
Thank you, Philip!
(Comments are closed)