today's blog

After I've done the last post, I had a quick glance at things I wrote earlier. I think I enjoyed reading what I wrote previously, it's nice to see what I've been up to, since one tends to forget things quite quickly. So here we go: one other thing I did today was to get my microphone to work. The bloody thing wasn't working, I think I believed that it didn't work, and so it did not. I come to this conclusion because I was really sceptical about the new cable I had received before christmas. I tried it and it wasn't giving any sound. The guitar was just fine, and I recorded some stuff with that. Then today I took the cable, mic and the sound card to digital village. They tried the mic, and it was just fine. Then I got back home, I turned the phantom power on (don't know why they have that buzzy name for that) and it just worked. and not just worked, aber es war fantastisch! the sounds (vocal + guitar) is so detailed and clear, that you love to record almost the sound of anything. Your breathing, typing on the keyboard, everything sounds good. So I did some recordings. But unfortunately when you play back, it doesn't sound as good as it sounded during the recording. It actually sounds quite dull, than what it sounds real-time. I asked this on cakewalk mailing list, hopefully didn't sound too much of a newbie and will get a reply. One more thing, today I cooked bulgur rice, first time in my life, and it tasted quite good I must say. Man, I'm good at cooking.

L4 pb926 port ready for testing

I have just built an L4 image based on pb926 platform. It merely supports the irq controller, system controller, timers. It took a bit longer than my expectation due to various reasons. First, the code is a mixture of c and c++ and it's the first time I'm dealing with c++ in a close-up manner. It wasn't much of a problem but the linking stuff had some suprising outcomes. c++ inline functions seem not to be exported from an object file at all, whereas in c surely you can do that. I asked this in comp.lang.c++ but it seems to be overwhelmed with questions and nobody gave a satisfactory answer.

Apart from that, I spend considerable time trying to figure out how it would be the cleanest and fanciest way to support irqs on multiple controllers. I will need to improve on that still, I think the new linux way (as of russel king's implementation) is the best way of doing it.

Now list of things to do:
1) Fix code that may not give the expected outcome. (There's a lot of it I'm sure)
2) Add version 5 architecture support. (This may may be the most major effort as I haven't done any cache/tlb debugging).
3) Add UART support for console messages.
4) Add LED support for further debugging opportunities.
5) Fix irq implementation and make it better.
6) Fix c/c++ code so that they're clearer.
7) Fix comments.
8) Test everything.