Older blog entries for fscked (starting at number 2)

Thursday

Up at around 9:30 again. While having breakfast, a mistaken bite reminded me that wisdom teeth are around to give me a hard time. Fortunately it was nothing serious, not worth an emmergency visit to the dentist, anyway. I'll have to make an appointment sometime soon though, been a while since last time.

No classes in the morning so stayed in, checking email, news, websites, generally trying to stay on top of things. Gave yet another try to the test machine but eventually decided the disk is really terminally ill. Good news is that one of the in house mates has a small spare disk to lend so by monday, if all goes to plan, I can start testing kernels as planned. This is especially important considering the feature freeze is rapidly approaching, October 20 according to Linus himself, so the focus is then on bug chasing.

From what I could see, and mind you that my kernel knowledge is _nothing_ compared to the gurus that actually come up with new ideas and code them in, I think the kernel has seen some major improvements and got added some neat features. A few days ago there was quite a lot of discussion as to whether the next stable release should be called either 2.6 or 3.0. But I'm not even going there. In the end it's just a number. Of course there's always the marketing implications (3.0 is indeed much more appealing than 2.6) but it's useless to discuss. Linus seems to be more inclined to go for 3.0 and I don't argue. Nor should I :) In fact many things got at least partially done, like the rewrites of the block and IDE layers, lots of VM bits including most of -rmap, new O(1) scheduler and other important things from Ingo Molnar, preempt by Robert Love, you name it. Maybe the sum of the parts actually give way to a 3.0 release. We'll just see :)

On another note, decided to stop by the university's library to check if there are some books I'm willing to read. Came across Silberschatz's Operating Systems Concepts and it's indeed very nice reading. Planning to keep reading it. Together with Understanding the Linux Kernel (which could do with a reviewed and improved version) and Linux Device Drivers, 2nd Edition, I think it can be a good study base for me to learn more about this field of computer science. Afterwards, the first class of Computer Architecture after yesterday's introduction. Basically the professor gaves us the layout of the semester's subjects. Here's a short overview for anyone enough curious:

- Performance Analysis
- Architecture's Instruction Set (for MIPS)
- Building an Arithmetical Logical Unit
- Building a complete processor for MIPS arch.
- Performance improvement using pipelining
- Memory systems: caches and virtual memory
- I/0, network and other bits

I confess I'm especially attracted by the memory systems. Really looking forward to this.

Got home at around 18:00. Switched to non-productive mode. Oh wait, it was already turned on.

Wednesday

Got up at around 9:30. Wandered off to Uni to attend the first Computer Architecture class of the semester. Basically just talking about the program, the project work and setting up test/exam dates. Looking forward to this, it's a matter of particular interest to me. Oh yeah, for all of you interested in odd archs, let me tell you we're talking about MIPS here.

Today's highlight was a Linux documentary being aired on portuguese news channel. I don't know what's the standard in other countries, but Linux and Open Source get few to none attention from the mass media here in .pt. Even when it gets spotlight, it's usually crap. Today, however, was a delightful surprise. Insight was provided by very well known and respected hackers such as Linus Torvalds himself, David Miller, Eric S. Raymond, Alan Cox, Richard Stallman and a few others. Needless to say, I taped it. Very inspiring stuff.

Afterall, the test machine has really gone tits up. Well, the hard disk has anyway. Oopses quite a lot and gives out some crc errors randomly, especially upon booting. I'm very pissed off about it considering the machine ain't mine. I wouldn't be this pissed off if it was actually mine. This means I didn't get around to muck with contest afterall. Will probably do tomorrow anyway, using my workstation.

The new apt colleague has arrived. I hope you don't take me for a dictator but the poor fella only had Windows installed. Not good. Obviously that bit is already taken care of. He's now happily running Slackware 8.1. I hope he settles in just fine ;)

Tuesday

First entry. Or maybe not.

Woke up at 8:30. Not much to do on the outside, university is still before warm up phase, fooling around with freshmen and stuff so got home at around 17:00. Sunk in front of the tele watching a re-run of Fight Club.

Spent most of the night reading (lkml) mail which kept me from investigating further on Con Kolivas' contest. This is a benchmarking application designed to test different kernels under different workloads (no load, process load, memory load and i/o load). It seems senior kernel hackers are delighted with this. So worth checking out.

Had to juggle a bit with the router machinery which has to leave its headquarters since the new guy in the apt will be arriving soon and I don't think he'd be happy to sleep with a couple of fans and a couple of hard drivers buzzing on his ears. Considering there's not much (closed) space in the place, decided to put it in the living as much hidden as possible. However the noise is decidedly inconvenient, even if we get used to it in a while. I'll look out for silent fans anyway.

Set up a stale machine which had been lying around for a while (well, it's not really mine) to serve for kernel testing/benchmarking. Disaster strucks when rm'ing a couple of files, a nice oops ensued and a filesystem goes bye bye just like that. I think it's the first time I've thrashed something with such efficiency... I'll have to see what I can do/recover tomorrow, cuz honestly today I'm a bit too tired to investigate.