Right ... the slides for my presentation are done but the writeup of the report is still not quite there yet, I hope to finish it soon. Mental note to self: I am hopeless at giving talks in Cantonese. So don't even try next time to do it in Cantonese.

Assignments ... there's one outstanding one but it seems quite easy to do, but it's dragging and I really would like to get it done by the end of this week.

Some more people are starting to have a vested interest in the work that I am doing so I should receive a bit more help these one or two weeks.

很久没有写网络日志了，想想大概是因为这一个月实在是太忙了，若不是前两天Roy Chan的提醒，估计现在还没想呢。

备课

从这个月开始，我就正式开始自己的第二份工作了。从小到大，从来都没有真正想过要在学校里面长时间工作，而到了要找工作的时候，忽然觉得学校真的是挺好的，最重要的就是有自己的时间，能够做自己想做的事情。

以前当学生的时候总觉得老师们讲课水平有待提高，现在自己终于可以正式站到讲台上，忽然发现原来自己的水平也不怎么高，而这种感觉真是越讲越强烈。

I finally found out why it wasn't initializing the page tables. Combination of several factors, but basically, exceedingly dodgy code (thankfully, not written by me ...) and a bad merge of L4 microkernel kernel code that made it have amnesia during a page mapping interprocess communication (IPC) message. Also: the newer L4 kernel (the one which I had been using) boots but it is missing some features which have been added to the included L4 kernel, and the included L4 kernel apparently has some problems with my architecture. I need to look into that later on.

I've also put in an evil hack to make the bootloader do the right thing, I think that I need to speak to someone on how to get this fixed properly. The build system that is used for this particular software system is Scons. Scons is good, but it doesn't always do the right thing. When it doesn't, it's very painful (like now, the thing with the bootloader).

Good news is that thesis is on track, I just got the stuff I needed booting, which is great news.

It seems that there is some problem with the page table initialization though and I need to take a look into that. This shouldn't be hard.

Something harder would be to fix the kernel that I'm using. It needs some features but for my architecture port it's missing those.

Assignments ... distributed memory assignment is out of the way. It was fun, but I got tortured in that.

Still have networks assignment ...

Ethics (!!) essay ...

Test for networks on Friday and I have not even studied ... maybe I should ...

Debian GNU/Linux 参考卡片
Submitted by tony on 2005, March 30, 1:52 PM. GNU/Linux
http://211.92.88.40/~tony/Blog/show.php?id=95
http://people.debian.org/~debacle/refcard/

1. 获取帮助
man page 或 man bash:
读取每个命令的联机帮助 或 多数配置文件.

command [--help, -h]:
大多数命令的简要帮助.

/usr/share/doc/[package-name/]:
在这里查找所有文档, 以及包含具体选项的 README.Debian 文件

Web documentation
参考, 手册, FAQ, HOWTO, 等. http://www.debian.org/doc/

http://lists.debian.org 处的邮件列表
社区通常是非常用于的, 用于用户查询.

記錄生活中的芝麻綠豆事，喜怒哀樂，情緒宣泄。

都差唔多一年喇
但係仲未放得低...
可能因為得唔到既野始終係最好既
越得唔到就越想得到
人就係咁架喇...

The hard disk in my new IBM X40 started misbehaving around 2am last night. Some time earlier it was making rather loud noises. And then two commands (ps and smartctl) locked up the machine and I have to used sysrq magic keys to reboot it. It's too soon for the harddisk to fail because the notebook was bought in October last year, so it's about 5 months only. My original plan was to buy a new harddisk, transfer data from the failing one to the new one as much as possible, and then take the failed one to an IBM repair center for a replacement. But the sad news is I could not find a 40GB 1.8" harddisk (which the IBM X40 can only use, it can't use the 2.5" ones) in Beijing's Zhongguancun! Zhongguancun to Beijing is what the Golden computer center to Hong Kong, you go there for any kind of computer hardware and parts. For all the shops I asked which sold 1.8" harddisk, they only had 20GB. Once again, it proves that if you want to play with the newest electronic toys, your best bet is to try your luck in Hong Kong :|

Got a PowerPC64 toolchain running. I had to use powerpc64-linux, because powerpc64-elf didn't work. Oh well. That's OK. Linux uses ELF anyway.

GCC 3.4.x didn't seem to like the Pistachio code very much, and I had to switch back to GCC 3.2.3. I could probably have used GCC 3.3.x, but couldn't be bothered grabbing it off the mirror. I now have PowerPC64 L4 kernel sitting on my hard drive, I've to try and get it to boot soon. Anyway I'm not entirely convinced that the PowerPC64 port code is entirely correct if GCC 3.4.x barfs on it ... GCC 3.2.x is just probably a bit more loose with its checking.