Delan Azabani 2014, day 2: a kernel fit for a colonel

An excellent start to the second day of LCA 2014 was unexpectedly seeing the one and only Linus Torvalds seated on the opposite side of the Octagon Theatre during the keynote. To counter this, while I'm wearing the official shirt today, I've forgotten my lanyard pass, so I really hope I'm not rejected from any talks.

Topics on the Linux kernel track included filesystem integrity, advice for new contributors to the kernel, testing device drivers, further details of block device replication with MARS, minimising idle power consumption without fully suspending, and features of the new POWER8 CPU, which includes an impressive twelve cores and eight hardware threads per core. Linus occasionally offered his wisdom, concluding to applause, "I don't do storage."

For lunch, we headed into the city for Café Biscotti, apparently the only coffee shop in Perth that accepts Bitcoin. The very friendly barista said that there were about ten Bitcoin customers before us in the week since accepting Bitcoin, and even offered us a free muffin afterwards. In the new Bankwest building; definitely would recommend.

Briefly switching among the Astronomy and Open Programming miniconf tracks yielded some insights into high-performance computing and large-scale data processing with the Square Kilometre Array, plus some lightning talks including a very interesting one about IPython Notebook, positioned as a web-based alternative to Mathematica.

Moving back to the kernel miniconf, Linus and three other core developers gave an open-ended panel discussion, taking intriguing questions like "What were your most tricky bugs to fix?" as well as some obvious questions such as "In this post-Snowden world, do you see more companies switching to Linux?"

Linus' encouraging response? "I'm a cynic and a pessimist, so I think nothing will change."

Out on the grass as a cool sunset loomed, the key signing BoF session was surprisingly stringent. All equipped with photo identification, we formed a line and verified each other, then we checked for matching fingerprints on the printed key list and each key owner's gpg --fingerprint output. I've already received one signature, which is a great start.

Half price waffles make for a guilty but fulfilling dinner. I have great expectations for many sessions tomorrow including a history of Linux filesystems, reverse engineering firmware drivers, and a massive live upgrade jumping ten years from Red Hat to Debian across thousands of servers. Exciting stuff.