As the excitement of OLS continues onward…
Yesterday marked the beginning of the actual conference, so I managed to drag myself out of bed for the first talk of the morning which was Jonathan Corbet talking about thoughts on 2.7. Or thoughts on the Kernel Summit. The current idea of basically using 2.6 for most development except for “really radical things” kind of scares me as someone who would like 2.6 to be fairly stable. Then, attended a talk on NFSv4 and rpcsec_gss which was pretty interesting; that seems like it's coming along well and will be fairly useful eventually, even if it's likely to involve a flag day event. Next was a talk by Keith Packard on getting X out of the business of talking directly to the hardware — interesting ideas and supposedly largely bought into by kernel folks; it remains to be seen if it actually happens. It seems to make some degree of sense to me, although I can see both sides. Also went to Matt Domsch's talk on DKMS and started on my list of thoughts after looking at it for a bit — need to try to catch up with him still and sit down and go through the list. Although I can easily arrange talking with him and Gary at another time without any real difficulty.
Had dinner with a few folks and wandered around a little with before heading back for the evening reception. The cheese was pretty good. On the other hand, I wasn't such a big fan of the speaker. Didn't win the AMD64 raffle. Later, hung out with a whole huge pile of folks and played some Super Monkey Ball 2. Eventually departed back to my room and tried to catch up with my mail a little. Finally crashed out at around 4:30 or so.
Woke up again early and headed to more talks. First went and saw a little bit on FUSE (userspace filesystem stuff). I've looked at venus in years past when dealing with arla and looking at coda, so not super amounts of innovation in what I saw of it; code is supposed to be better, though. Then, another X talk — this time, Jim Gettys talked for a while about some of the problems with X right now and then Keith showed some demo-ware. It was quite pretty — the compositing stuff definitely looks cool. Acceleration of it on modern hardware promises to be a little bit dicey, though, just from getting vendors to care :-/ Had lunch with Matthew Miller and another guy from BU and mostly talked about OpenAFS for a bit. After lunch, it was the talk on virtualization on the Power5 stuff — especially interesting as I've been dealing with it a bit. For all of my gripes, it is kind of nice in a lot of ways, I'm just not entirely sold on the implementation details 🙂 Then, a talk about filesystem security which had lots of stuff covered — getting to a good, integrated solution here is looking more and more necessary. Then, hot plug memory which while kind of cool, doesn't seem to be all that baked yet (well, add seems okay-ish… but remove just isn't even close)
Dinner was with a crowd of us, the Specifix people (that is msw, ewt and mkj) and a couple of Dell folks. Had a good discussion about some things with mdomsch (again) and try to come up with some ideas. After dinner, went to the SELinux BOF where I was probably fairly troublesome… at the same time, the problems there are real. I think that the current working plan of using the “targeted” policy by default will let us start getting MAC used and understood by people. After dinner, more hanging out with much the same crowd although I did have a quick talk with rml about udev. And my hope is to get to bed earlier tonight to catch up a little.
So, it's been a busy few days and the next two don't really look to be any less so. But, it's been productive, so I'll live with the crazy business.