linux.conf.au report

linux.conf.au report

By Greg Lehey

Thursday, 15 January 2004

In to town relatively late this morning, even later than intended, and missed the beginning of Jeremy Malcolm's talk about SCO [ http://www.lemis.com/grog/sco.html ]. After that listened to Chris Yeoh talking about the Linux Standards Base [ http://www.linuxbase.org/ ]. It sounds very well thought out. It's a pity that the BSDs don't want to know about it.

We had far too little time for lunch, only an hour, including the time it took us to get to Rundle St after people had finally got their act together, so we were quite late getting back, and I missed about half of Peter Chubb's talk about user-mode drivers. Pity, but I suppose I'll get it all on the conference CD. Then Sean Burford on reverse engineering from binaries. A good introduction, but no techniques that I didn't already know. He did point to some useful tools I hadn't heard of, though, and that should be useful.

After that I had invited a number of people out to my place, but timing was tight, and in the end only Peter and Lucy Chubb came out, so we spent most of the time talking about bassoon fingering systems. Pleasant evening.

Friday, 16 January 2004

First up this morning was my Vinum presentation [ http://www.vinumvm.org/papers/LCA2004/slides.pdf ], which went off without much trouble. Was rather flattered that Ian Gilfillan came in to listen to the start, certainly not because of interest in Vinum.

After that to Paul McKenney's [ http://www.rdrop.com/users/paulmck/ ] talk about Read-Copy-Update. I had heard the term before, and I thought I had basically understood the concept, but it turns out that I was completely mistaken. The idea sounds a lot better now, and the figures that Paul showed spoke for themselves.

To lunch at Cafe Michael 2, which turned out to be yet another Thai restaurant, not as good as the place across the road yesterday, and with terrible service. As a result missed the first session after lunch, and then to hear Jonathan Corbet talk about the Linux kernel 2.6, which provided a good overview.

Then back home before returning for the conference dinner, at which we took a number of photos [ http://www.lemis.com/grog/dinner-20040116.html ].

Saturday, 17 January 2004

Into town late today with an intention to listen to Jeremy Allison's “security soup” talk, but Arjen Lenz waylaid me and got me to attend James Cameron's talk about PIC microcontrollers, which I didn't regret. Amusingly, it ties in with a device I had been thinking of for controlling my temperature-controlled fridge: Paul Sorenson had pointed me to a computer controlled device, and it eventuated that James had designed it. He also spent some time talking about it and demonstrated it, showing that it's probably exactly what I'm looking for.

At lunchtime they had arranged a dunking session, and I was able to keep myself from being involved. To the Red Rock noodle bar, where I had “kimchi noodles” apparently without any kimchi. Back to hear Tridge [ http://samba.org/tridge/ ] talking about his junk code directory. It seems that he's finally tidied up his remote mail scripts, which I must investigate.

After that, conference close, and then tried to organize the people who wanted to come to my place. The challenges were many: Martin Pool needed to hire a car, Bdale had to go to the hospital to visit Pia Smith, who had come down with what proved to be flu and not SARS yesterday (she had just got back from China on Tuesday), and others wanted to get back to the hotel. Finally found the last person (Jeremy Allison) talking to Linus, and it proved that Linus didn't have anything else to do, so he came along as well. They went ahead with Rasmus while I took Martin to the airport to pick up his car.

Back home, Tridge was going on about some anomalies between the behaviour of multiple threads and multiple processes on multiple processors on a loop through readdir(), while Linus observed that nobody else cared. Well, not quite. I wondered how this would work on FreeBSD, but it turned out that the processes blocked on Giant, so that was a non-starter. Score one for Linux.

Apart from that, spent a lot of time talking about all sorts of things. Took still more photos [ http://www.lemis.com/grog/barbecue-20040117.html ], and read out the fairings [ http://www.lemis.com/grog/fairings.html ] saga, which the Linux people didn't know, but they found it good enough for a spontaneous round of applause. Pleasant evening.

Sunday, 18 January 2004

And still the conference isn't over! Into town today to pick up Janis Johnson, Paul McKenney and Tridge and to Geoffrey Bennett's speakers' barbecue. We had barely got there, however, when we got a phone call with the news that El Mago had torn down a fence round one of the eucalyptus trees in the paddock, not for the first time, but this time he was lame, so off without even having anything to eat and back home, by which time he was no longer lame, and there wasn't a scratch on him.

Sat down for an improvised lunch, and was barely through that when I got a call from Michael Davies telling me that Martin Pool had had a car accident and was at the Strathalbyn hospital. Fortunately he wasn't badly hurt, though the car was a writeoff. As I had half expected, it happened on the Paris Creek Road, the one that Yana drove to school on for years. Brought him home, and Yana took him to the airport.

Didn't do much else. After a week, the party^Wconference is over. It has really been one of the most interesting ones I've been to.