Did you know stuff happened today while you weren't watching?

This month is really starting to piss me off. Sad. It was really such a good color too.

I feel like sleeping but I can't. Of course I can't. I slept all day. I'm tired of doing work for no gain. I'm too busy doing work for no gain that I don't have time to do work for personal gain. Why can't I leave society yet? Aparently I'm barely old enough to declare myself socially unfit and leave society (you have to be 16 to do that).

And why don't I just leave? Leave school, everything behind and go hide for myself somewhere? Why else but the fear of committment. If I do that now, it'll be comitting myself to thinking the way I think now for the rest of my life. There'd be no going back. And years later when I no longer thought this way (as I'm guarenteed to change -- I change hour by hour, its obvious that I'm going to have a drastically different mindset in a few years time), I'll look back and curse myself for being so idiotic and not thinking ahead...for committing. And its one of the worst possible feelings for me to know that I did something wrong out of ignorance. Ignorance is evil.

And thats the irony of my life. Yours too. Ignorance is the worse thing. Yet we live in a universe where we can only reduce ignorance over time, and that time only gives us the opprotunity to look back and wonder about it. Our nature requires us to think hypothetically. To question what we could've accomplished better or faster if we'd known one bit of information earlier in our lives.

If you'd learned to read one day earlier, would you still be in the position you are now?

Everything comes back to the above. Why do I bother doing homework? Going to school? Because not doing and not going would be too major of a change. An irreversable change. If I don't get a diploma from high school, there's really not many options later. I could just dump it all right now, go get a GED quite easily. But thats major. Colleges don't like GEDs. And not being able to get into college is a major problem.

I don't fear change. I fear the irreversability of it.

However, over the past while, it has become increasingly aparent that there's something stronger in me than fear. Apathy. And as you have probably guessed, there's a point where apathy crosses fear and nothing gets done. Not because the fear has decreased, but because the effort needed to fear is too great. Apathy is winning.

And there you have it. The summary of my life for the past few months: Apathy is winning.

Went to school this morning. Complained to Janeal for 90 minutes. And came home. I think they did room/bus assignments for the SoCal trip. Er, I slept through that. I should talk to her. Anyway, I have aparently contracted this lovely cold-like thing. I feel just wonderful, be assured.

Slept all day.

"part of politics as an enlightened individual is knowing how to vote for someone that has a chance of winning" (josh) Am I the only one thats rather disturbed by such a statement?

Hrmph. I had a lot to cover my opinion on that. I removed it. Not because I didn't like it. But because if I wrote that much, I'd have to write that much more again to explain it. And then most likely have to write more explaining that bit too. I'm just going to spare myself the effort and leave it. [I will mention this, just for the sake of putting my mind at rest: It is not overidealism that is hurting America. It's overrealism.]

There, I saved you the dissertation. Happy?

Hrmph. Overuse of AIM over the past year has quite drastically debilitated by ability to write. Write write, not just write. Blahhh... Its taken away the incentive as well.

I'm going to school tomorrow, despite the fact that I still feel utterly horrible (in fact worse than I did this morning). I need to ask C where the absenteeism counter gets reset (methinks its on the 9wk boundry, which means I can not only stay home a couple more days in the next few weeks, but also get another 10 days before the end of the year...if not, then I'll probably fail all my classes simply on a technicality (certainly not a GPA-related technicalilty, that would make too much sense) because I've been absent too much. (Its actually a financial technicality, despite the fact that they won't admit to that being the reason they're going to fail you with an A in the class...if you don't go to school 90% of the class time, they lose the state funding for you...less money for the school means they have to find some why to release their anger towards you...hence, they fail you. Such a nice and logic school system they are too.) Anyway, its a shorter day. Not too bad.

Friday looks to be incredibly boring as well. Aparently most everyone (except me, I was not invited nor told what was going on) is going to NAU for some reason or another. Other than a calc test, that should be a sleep day.

I'm not going to get all that government crap done. Someone want to do all this homework for me? I'll pay you...

Aparently there was some bit of confusion over the last entry. "Bah. Bye." is rather different than, say, "Babye" (one of those lovely popular slangy thingies that we all love to hate). You stand corrected. Yes. That means you Stacy.

The purpose of that statement was to convey my sentiments concerning Friday. It was a fairly blah day. In a bad way. It included getting up earlier than usual, and then sitting in one spot (a very, very unformatable chair) from 0900 to 1830, save an hour where we were allowed to sit somewhere else. I was thoroughly unamused by the time 1830 came around. And was happy to come home. In total, my waking day lasted over twenty hours. I prefer to have naps throughout the day, even short ones. I cannot tolerate the ridiculous sleep-eight-hours-and-stay-up-for-sixteen crap. I need breaktimes in the middle of the day. Anyway, the twenty hours was not the troubling bit. The troubling bit was me not getting my nap that day.

Saturday was far better. Still had to get up early (had to be in Mesa by 0900). But the rehersals were minor, and an hour shorter than scheduled. The only particularly annoying bit was the stench of an electrical fire on stage. Hah. It was disgusting. Had two and half hours for lunch. Concert at 1400. Other than the sitting around waiting to perform part, the concert was quite enjoyable. They didn't let us listen to the choir (which was really good) or the orchestra (which I didn't think sounded great in rehersal, but...eh). Hopefully the CD is decently mastered. The band's performance was amazing, however. Thankfully. That Humburg guy is really annoying for the first five or six hours, but he is a good director. And his jokes do get funnier as you become less conscious with boredom.

I spent nearly all of Sunday sleeping. Yay. Same for Monday, mostly. I did do a bit of homework then too. But nothing substantial. (I need to cram a months worth of government homework into the next two days...wheee...)

Didn't really get any wireless work done. Did make headway in things relating to doing more work, however.

After deciding that the ODI driver was probably not written to be the best driver available for the card, I started looking for other stuff. I came back to GNU binutils' support for the NLM format. Got really weird results when disassembling. Sent mails off to the addresses I could find that were related to the NLM support. Got a reply late Monday evening. Answered both of my questions. Firstly, the extraneous numbers following a properly decoded line of assembly are not in fact extraneous. Aparently objdump only likes to print five bytes of opcode per line. On six and seven byte instructions (which are allowed on ia32), it prints the sixth and seventh bytes on the following line. Very confusing. The other thing was that it would generate calls to impossible addresses (like, at offset 0x17, there'd be a 'call 0x18', which is not possible, as the call instruction must be at least two bytes...). The answer here was that its a result of relocation mangling. Yee. The proper way to get usable information is to use the -dr options to objdump. Not pretty, but it does work. That makes it print the real address of the call. And at least in this NLM, it gives me real names to symbols. Even better.

So far the NLM code looks closer to what I want in Linux. It uses multiple recieve/transmit buffers, for one thing. And I haven't yet seen any stupid DMA controller manipulation yet either. So either NetWare Server handles that for the driver, or the card can actually do busmaster DMA. I haven't gotten that far into the code yet.

I only got an hour or so to really look at it yesterday. The NLM MLID HSM layout is really an improvement on the old ODI MLID interface. At least if you're trying to disassemble them. All the major callbacks from the OS are listed at compile-time (and therefore static at disassemble time) in a big table. The NLM driver is actually quite a bit smaller than the ODI driver (though by how much, I can't tell because the ODI driver also has the full MSM and TSM linked in; the NLM is the driver code only, the MSM and TSM are all dynalinked at run time -- as I said, NetWare is a Real Operating System, DOS is not, therefore the ODI driver has to fill in all the pieces that DOS is incompetent in while the NLM can just make assumptions about it).

I'm basically stuck in two modes are far as examining the NLM code goes: 1) identify the major parts of the driver by looking at it from the MLID/Netware interface perspective, and 2) finding the routines that match the routines in the ODI driver (I have a lot of knowledge about the ODI driver, much of which should be applicable to the NLM -- they do drive the same hardware after all!). I haven't had the chance to make a lot of progress in either mode.

I haven't had a chance to do much of anything. I'm to the point where I have so much crap to get done that I spend only a small amount of time on each thing, and get nothing done overall. This is bad. I have lots of for-a-grade stuff to get done Soon, and I'm not prepared to do it. I'm just glad I only have two graduation requirements this semester. Wheee... I'm going to fail something, and I don't get to make the conscious decision which thing it is...

Oh, oh, and IT RAINED!!! All those days of no weather. They were an omen. It rained. Did I mention it rained? First non-negligable drop occured around 0200 Monday, followed by various 15min spurts following. Last night there was aparently quite a downpour, judging from the size of the puddles.

Some good things might be happening as far as my involvement with AIM goes (ie, money). I may have a decent job this summer in The Valley as well. We'll see.

Did I mention it rained?

Bah. Bye.

It's getting there. Tonight was the fun part: DMA. Did I say fun? Oh, wait.

Was in a bad mood the first part of the morning. Not helped by the fact that I had to do a lab by myself in physics because everyone in that class is incompetant, including the author of the lab sheet. Blah. Went to get an AP form from Denton. Talked with her for a bit. Late to calc. Big loss. Came home at noon. Slept.

Fidgeted with the IBM card for my usual three hour allotment. Got DMA working. In both directions. (I think. I assume that my writes to the card work if the subsequent read from the card yeilds expected results.) This got me a 0x30byte config block, which includes various bits of information such as the microcode version string and the MAC address. Its a helluvalot of work just to get a MAC address, but at least its not a SEEPROM. And it means I got DMA working. The two last things I need to figure out before I can move on to actual transcieving: 1) figuring out all the weird indirections they use to setup the ISR's, and 2) doing the actual cell insertion. I'll have the debugger during the cell insertion, but from then on, I have to competely rely on running the code on paper. And that means I actually have to know what I'm doing by then.

I have to go to regional band rehersal tomorrow and saturday. 0900-1830 tomorrow. 0900-1500 or so on saturday. And I volunteered for this?

Looks like I'm going to be taking at least three AP tests this year. Calc AB just to say I did (not that any school I go to for a CE major will take it anyway). CompSci AB, just to spite the situation that happened last year. And Bio, because I think it will be easy, and its an excuse to do more bio stuff. I'm still thinking about English. I'm going to have to sit through all the preparation rubbish in compandlithellofnoremorse, so I might as well take the test. Even if I get a 3, its better than nothing. If I get a 1 or a 2, I can always cancel the score.

For my future reference, the AP schedule is here. That's going to be a hellish week if I take the English test. CS on the 9th, Eng on the 10th, and Calc on the 11th. And then Bio the day before graduation. Wheee.

Hrmph. I said that.

I want rain. Now. Yay. The world must be ending again. We have no weather.

Yet another boring morning. And that timedwriting did come this afternoon in compandlithellofnoremorse. [I'm just going to add another word onto "compandlit" every week until class is over.] I finished in 20minutes. It doesn't take very long to get a C. And since I have no idea how to get an A, thats the best I can even try for. We had to come up with our own topic today. We'll see if thats a good thing or not. It doesnt really matter. I'm going to ramble and ramble on about nothing anyway, whether they give me something to ramble about or not. (But I got to implicitly insult people in this one. Yay.)

Stared at assembly all evening. Or tried. I kept getting distracted and eventually gave up. Read some amusing bits in the newspaper. Haha. Fish amuse me.

Oh well. G'night.

"Because we all know Adam Sightings are Bad."

Hrmph. Really really boring day.

All morning was irritatingly inactive. Just all sat around morning. Slept all afternoon.

Traced the IBM code all the way through a couple DMA transfers. Aparently it doesn't support bus-master DMA, only polled DMA. Yick. Polled DMA features all the disadvantages of DMA (complexity) and all the disadvantages of PIO (wasted cycles). Oh well. I'm still a bit confused, and I'll trace through it at least once more before I actually try to implement it. I did figure out the ordering and framing of Intel enter/exit stack frames, however. The blocked functions are actually quite a bit easier to read than the standard asm subroutines, as the paramters and variables are easier to see (they're all at 2byte-aligned offsets from bp, but of course the stack grows upside down...). Very nice. Just wish more of it was that way.

Did compandlithell and calc homework later this evening. I'm bored again now, so I guess I'll go read for a bit. Maybe even go to bed...

Yay. It can read its DIP switches. And more importantly, it can be reset. Interesting stuff. Only a few more years and I should have it all the way through the init process...

It would appear that lots-o-intelligence is on the card. Resetting is a matter for toggling a bit in one of its eight registers, and then waiting for another bit to come up. Thats it. (You can also watch the LED on the board go off for a bit while its resetting. But thats not very exciting, and is aparently the only use of that LED that I've found yet. Boo. I want blinkenlights damnit.]

It has the signs of a simple request-response flow. I think you write a command byte to a register, flag a "command waiting" bit in another register, wait for the ack bit to come up (the same bit as in reset, actually), and then read out the response. This is all done using measly 8bit registers, btw.

Anyway, that was most of my evening. The rest of the day was rather average. Got the highest grade in the class on that calc test from friday. Unskewed score of 75%. Hah. Stupid mistakes. At least I remembered the procedures for simple arithmetic this time. Corrections took a bit. Too much distraction...bah. Didn't end up doing a timed writing in compandlithell after all. Oh, but its coming. No doubt.

Sigh. Time for bed. Soon. Maybe.

Er, oops. Forgot to write...yes..thats it.

Thursday I was actually rather ill and slept for 20 hours or so. That could've been better.

Friday I did have to go to school. That could've been better.

Saturday I went to the VNSA booksale. Picked up several dozen books for a grand total of 24$. Did nothing much of the rest of the day.

Today I did less than yesterday, with exception that today included watching a lot more television (Not the real kind...the Simpsons kind).

Worked a bit on friday night trying to get PDMA working. I've determined that Madge has hopelessly rewired the TMS380 PDMA/SWHRQ pins on their cards to the point of me having no idea what the hell is going on. So, ISA Client Plus PnP support may be impossible without using Madge's firmware. I also came upon some information from Olicom, plenty to write ISA and PCI support for. Time and interest is the issue at this point. I still need to backport to 2.2, but it looks like I'm too late for 2.2.15, which means I'm going to have to test everything as modules and try to get a user-compilable tarball out instead. Ugh.

Quickly read through an x86 book yesterday (that I got at the sale for .90$). I now know a lot more about the IBM assembly code I'm looking at. I've actually started writing the kernel module (yikes). Detection is easy. Right now I'm stuck in the hard reset routine. I'm not getting the ack back from the microcode like I should. But I don't want to go into it tonight. I want to take a look at the HPC microcontroller datasheet and a meter, see what's connected where. The more of the HPC thats connected to the host bus the better. If the HPC is allowed to directly DMA into host memory, this could be interesting.

"Do not use this appliance on pain of explosion." Haha.

Anyway, the past few days have mostly been me sleeping a lot, and trying to avoid thinking about school. Blah.

Zrmph.

Slept through what I would've had of school today. Oops. Damn alarm.

Worked with the wireless thing all day and all night. I have a lot of code traced, but I'm going to have to do it over again just to make sure. Having the ODI source really simplifies determining what routines are driver-specific and which aren't. Very nice. But its still assembly, and its still time-consuming to read (and reread (and reread (and reread))).

Inbox is back up to 50+ messages again. Full of reminders and crap. Got rid of three of this evening by applying Christoph's two patches, as well as updating tms380tr to work with 2.3.43-7 (though I havent even upgraded my tree to that yet, so it doesn't compile yet...).

You don't watch MTM enough. There were in fact two different theme song lyric sets, however, neither said "can never tell" (probably because those lyrics wouldn't make any sense at all...). As I said, there was another set. From 1971-72, the opening line was "How will you make it on your own...", instead of "Who can turn the world on with your smile..." as was used in 1970 and 1973-77. You just happened to dispute me during an MTM marathon on nick at night, and i just happened to have clrec+bladeenc ready for it. Listen to it for yourself if you don't believe me. [Never dispute the Great One. Haha.] [It's 800kb too. Bladeenc doesn't do well at less than 128kb. It hates my standard recording rate of 48khz as well. Bah.]

Lalala.

Do I really have to go to school again... sheesh. This going-everyday-thing is really boring you know...

Why am I a "moron"? Does this have something to do with the end of yesterday's entry?

Today would've been much better if I hadn't had to stop doing what I'm doing to go to sleep because I have to go to school. But. Oh well.

Nothing day (haha, reference to yesterday...) really. All morning was just boring. Compandlit less painful than usual. She's throwing out the surprise timed writings as well as letting us bring up the grade on the Othello paper (and I'm certainly going to take the opprotunity to bring my B+ up to an A-. hah.). Came home and slept (I say that everyday; I should find a new way to say it). Homework in the early evening.

Started pouring over IBMWL.COM ODI/LSL driver disassemblies. Weee. 64 pages of 4pt, four column landscape pages. Its fun. Intel assembly is hell (have I mentioned that?). However, I've discovered lots of interesting things this evening/morning. And now I am certainly glad that I decided to use the ODI driver instead of the NDIS driver. With this, I get all of Novells documentation. And for anyone who's dealt with novell's developer relations department, you know what I mean as far as magnitude goes: lots of API documentation, lots of information, and most importantly, lots and lots of verbosely commented source code.

ODI drivers are built by writing a few required routines and then linking it with the main MSM module. Therefore, the actual driver-specific (HSM) portion of the resultant object is very small percentage-wise (like 10-30% I'd say). And Novell gives out the source for the MSM code. So I just have to compare the source, figure out what's added in my disassemblies, and go from there. The Novell MSM source is very well commented, and matches perfectly to the code in the disassembled output.

This project was looking grim for a bit, as I didn't know how long I could survive going over this much asm. But it just got a whole lot easier (finding the MSM source), and its looking better.

I have an especially short day tomorrow. Hopefully I'll have time to finally make some real progress here (maybe actually analyse a routine with a real outb/inb in it!).

For those who care, or more importantly for my own reminding, all the required Novell info is aviable in a 20mb lancomp_all.exe file from developer.novell.com. "Free registration required."

It's 0300. I guess I should go to bed now.

Ahg. My clock skipped from midnight to 0130. Oops.

Nothing particularly exciting at school. Had to stay a bit late to make the physics teacher happy. Came home and found a rather large box full of wireless cards. Yay.

The antenna is a bit smaller than I expected, but still larger than the Entry antennas. The boards are rather long. I don't like the way they did the MCA/ISA on the same card, but whatever (the ISA connector is raised about a quarter-inch off the board, just waiting to be borked off). It comes with no rear plate. You have to screw on either an ISA plate or an MCA plate depending on what you want.

It's got two large EPROMs, and another 1mb or so of EEPROM/flash. There's three major chips on the board. The most entertaining is a National HPC46400. It's a 20mhz microcontroller. This may indicate that the majority of the code runs on the card, meaning that the host driver is minimal. We'll hope thats the case. Another chip of interest is an IBM22-ALDC1005S00. It's a hardware line-speed 3:1 compressor. Neat. Datasheet is available at chips.ibm.com. The third chip is the most important, as well as the least documented. IBM part number 54G1943, Toshiba part number TC1600GA8AF. This is the chip my driver is going to have to deal with. Glad I know so much about it...

Spent the evening trying to set up a NetWare server. It all appeared to go well the first install, until I found out that all the directory services files were corrupt. Attributed all this strange behavior to an overclocked 486, put in a regular 486sx-33 and reinstalled. No weird stuff since. Other than that I can't see the server from the client. Erm. And I haven't even put the wireless card in yet.

If someone would like to donate some OS/2 disk images, I could avoid this netware thing altogether. Anybody have some?

I'll have to continue that tomorrow. Or finish up the Madge stuff. Weee.

"Who can turn the world on with your smile? Who can take a nothing day and suddenly make it all seem worthwhile? Well it's you girl and you should know it. With each glance and every little movement you show it. Love is all around no need to waste it. You can have the town, why don't you take it. You're gonna make it after all...you're gonna make it after alll..." [I felt the need to include it in its entirety (according to me, I'm sure its correct), just to enhance your annoyance with it.]

It can't possibly be monday. No.

Slept til 1530. Very enjoyable. Finished up some calc stuff that I'd been neglecting (related rates are much easier to comprehend the second time around...). Did nothing for quite some time. Did some Simpsons watching.

Worked on ATULA stuff the rest of the evening. Aparently some things that got touched in the final sixth of yesterday broke things. Fixed those. Fixed numerous bugs. Figured out the required behavior during interrupts. The Madge Smart 16/4 AT Plus (ISA) should be working fine now (doing real DMA and everything). The only other Madge ISA card I have is the ISA Client PnP. Its not a DMA card. It only supports PIO/PDMA, but it does support that weird 32bit PIO, which requires the Madge firmware. So, with my driver you'll be stuck with 8bit PIO. Which is going to be really really slow. But hell, it'll be supported. And slow support is better than none (right?). I will have to write all that PDMA stuff someday. That card also uses an AT24 SEEPROM, but luckily I already have the routines for that. Just have to figure out which bits to twiddle. Madge says the Olicom OC-3117/3118 is the same as the AT Plus. I'm having a difficult time believe this, as the Olicom is PnP and the AT Plus is not. But we'll see. (The Olicom also has an HC46 SEEPROM on it, which increases my doubts further.)

Probably get the PDMA stuff working for the ISA Client PnP next. That's going to take quite a bit of effort, but the PDMA stuff will all be general tms380tr stuff and will be applicable to other cards. I've been meaning to write it for quite a while now (ever since I was told there were Alpha's with borked DMA controllers that needed it). I'm not going to put too much effort into the Olicom's until I get more programming info. I mean, if they're not going to support me once again, I can't justify my efforts. I have better things to do. People who need Olicom support can probably trade in their cards for Madge's. Or better yet, get a SysKonnect.

Speaking of syskonnect, Christoph mailed me a bug patch today. Aparently somebody had a card that was trying to recieve and send simultaneously, and was getting the skbuf pointers mixed up between the two. That must've been amusing.

Anyway, aparently I have to go to school tomorrow. Boo.

Weee. I'm bored. But bored on purpose.

Slept in til 1030. Finished up Heart of Darkness. It was enjoyable, not just because I finished, but because it was actually enjoyable (gasp). The last twenty pages or so are just bizare. But entertainingly bizare. I should read the cliffs/monarch notes just to make sure I hate it by Tuesday. (Gods forbid I enjoy something I am forced to read by compandlit...) Took a shower, did nothing for quite a while.

Worked on ISA Madge (ATULA) support all evening. First third of that was spent writing the actual initialization/probe routine. Very complicated. Another third was spent getting DMA to work. Various problems were showing up there. Forgot that you have to enable the host DMA controller, as well as parsing the ROM DMA channel value wrong. This was causing a DMA timeout for the test transfer (but I didn't know that, because the backend code lumped all the errors into one error, and I had no idea what was going on -- so I copied yet another error code table from the Yellow Book into printk's). That all got figured out, which got it all the way into the light on the MAU turning on. Woohoo. Except it won't trasmit or recieve. The following sixth of the evening was spent trying to figure that out. Aparently interrupts get squashed sometime after BUD. Oops. My attempt to locate it caused Weird Problems on the machine. (Like, RAM errors during POST. Very Bad.) I used the final sixth of the lot to cleanup some things and add some comments. All in all, I HATE ISA.

Oh well. I'm done.

Hrmph. It's hard to believe that the rest of the world thinks January is over. It seems like it lasted forever for you people. Luckily I'm still in December and I could care less.

Today lasted forever. Blah. At least I slept a lot. Yay. Most of the morning was just dreadfully boring. Was quite surprised to see Dadiv's father wandering around my physics class. Talking to him proved a good distraction from the mind-numbingly idiotic "activity" I was supposed to be doing. Came home as soon as possible, read a few more pages of Darkness and went to sleep.

Been staring at assembly code all evening. Intel should be put out of business for desiging a processor with this many instructions. Any processor with more than 550pages worth of instruction reference is just bad for the mind. (Can I have my MIPS back pleeeasse?) Oh well. My Wireless cards are coming sometime next week. I'm not sure how much I can learn without the cards in hand. When I get them, I'll of course need to set up the cell in working order with working drivers and all that. Just to see how its all suppose to work. I'd like to get a physical look at the cards too. There's some references in the drivers to SEIKO/EPSON that are frightening me a bit (what could they of all companies have to do with this?). Found some Bad Behavior Problems (aka Bugs) in the disassembler I am/was using. I really don't feel like fixing that code, so I'll probably just look for something better. Microsoft CodeView for DOS is looking pretty good right now. Though I have no real idea how to use it, and it has no help (yay for microsoft). Anyone have a copy of Turbo Debugger for DOS laying around? Mine seems to have gotten corrupt. (I could regrab it off the original disks, but, that would require finding a 5.25in drive.)

Sigh. I'm tired of being awake for now. I should do that ISA Madge driver tomorrow.

Its December 66th, despite the fact that for the past two days I've said January, and today I say Frerbruraryr. It's still last year. Really.

I didn't go to school today. Yay. Though I probably should've. I don't care.

Made it to page 41 of Heart of Darkness. You know, I liked this book for the first few pages. Then I read more. And I hated it. Now I've read even more and I'm starting to like it again. Amazing. I'll hate it by the end I'm sure. Only 31 pages left (of course that ends up being several hours worth of reading...).

I haven't really done anything today. Tomorrow's friday. Yay. Three more days and we start all over again...

"your color looks like puke" I knew it was coming. Yay. Tara is so very predictable.

Average daaayy.... Boring morning. Compandlit was just damned frightening. Teacher was not there, and in addition to that, we had a surprise timed writing on Heart of Darkness...weeee.... Luckily it wasn't terribly difficult. I don't think. I expect yet another C. Actually, I expect worse. But, oh well. Timed writings are evil. Compandlit would be a decent (not great, not good, but decent) class without the timed writings.

Isn't there a Brockway or somesuch name in the Flintstones? I really thought there was. Seeing the name in Invisible Man made me think of it. And its been bugging me ever since. I think he's in the upper management of the quarry, but I can't remember for sure. Probably one of those one/two-episode characters.

I won Mel's little trivia thing yesterday. Haha. And I made the list (see "Satan").

Jay's Linux/SNA is back. Uhm...Wow. I guess that year without public mention did the project some good. Or possibly it was TurboLinux hiring him and paying him to work on it... [Whine. I want a job at a Big Linux Place.]

After getting a message from TI saying that the RAM testing during BUD/OPEN is unavoidable on the TMS380, I decided to go see if there was something else causing the driver to take so long to open. I ended up going line-by-line through the initialization procedure and matching everything with the Yellow Book. Was able to eliminate several extra delays. Also got rid of having to have a byte-width read/write callback. That makes everything a little cleaner and smaller. (I'm not sure why byte reads/writes were getting used at all anyway; perhaps it was just a slip.) The driver was also enabling interrupts at places it didn't need them, which has weird-race-problem written all over it. Not that I ever had any problems, but still. I fixed that as well. Upon testing, I find that the adapter open delay has decreased by about 30% on the PCI cards, and nearly halved on the MCA cards. I can't really explain the latter, except maybe in that its a slower machine and I killed some loops that wouldn't have affected the PCI box. Shrug. Also added some extra error handling (with messages) for added robustness. I stopped at the DMA test, but I will probably look over that eventually as well. There's probably some more tweaks to do. I would like to add PDMA support soon as well, so I can maybe support more ISA cards without having Real Support for them, though thats probably doubtful (chances are if they have special DMA handling, they also have special IRQ handling, and it /requires/ interrupts).

A new friend from comp.sys.ibm.ps2.hardware dug up a couple old non-Entry IBM Wireless ISA/MCA cards that he's going to send me. Yay. That saves me a bit of money (though if I want PCMCIA, I'll actually have to buy it myself...but this will at least let me see for sure if I can write the driver before I spend money on it). Hopefully the boards aren't too complicated. The NDIS driver is 40kb. Eeek.

RS/6000 came today. Boots up to error 213 on the front LED panel. I have no idea what that means (hell, it took me 25minutes to figure out how to get the damned lid off without breaking anything). Adding in PS/2 RAM from the mod80 doesn't get any different response. The only differing response I did get out of it was when I put reall ECC SIMMs in it (from the Alpha). That made it real mad: flashed 888 instead. I hope error 213 isn't fatal. And I hope I don't have to buy RAM.

Since today seems to be a big catchup day on small comments (probably because I'm avoiding homework), I suppose I'll add another one I forgot. I mentioned the regional band audition last Saturday. I didn't mention that it went fairly horribly. I made 6th (aka last) chair band. Which means I had the very lowest score that they accepted. They took nine people. I was the ninth... Oh well. The music is quite easy. I'm not sure what the overall level of the pieces is, but my part looks to be a about a level threeandahalf/four. Which is fairly sad for a regional band (for what should be the most skilled region in the state). Our district honor band music was a hellofalot more difficult than this stuff. I'm rather confused.

I guess I can't avoid that homework much longer...

I like this color. Its one of the best colors yet. Its so...perplexing.

Better day than yesterday. Failed the physics test. Hah. Someone relating to that class is a complete dolt. I'm not allowed to say who it is. (But there's no way in hell I should fail a test if I gave the correct answers. Even if I didn't follow directions that I wasn't given.) [Nooo ranting...] Managed to get calc homework done while it was being taught (weeee...). And then took a government test in less than ten minutes. Came home. [Aparently the new rule is that in order to get all A's, I have to fail exactly one class per semeseter. Last semester 'twas humanities (tho I can get that to an A whenever I get around to writing that paper). This semester its aparently going to be physics. At least I'm not failing graduation requirements...(this is my fifth science class for gode's sake, and I only needed three to graduate)...not that it would matter...they let us graduate with defeciencies anyway...just to get rid of us...]

Olicom has binary-only developement kits for their token ring cards on their denmark ftp site. Maybe I can get the source for them.

My contact at Madge-Germany says he has "major customers" asking for Smart 16/4 AT Plus Ringnode ISA support for linux 2.2. That sounded important enough. Spent the evening backporting the 2.3.41 version of my driver to 2.2.15pre5. I had to write a bit of code around the probing routines to work with Space.c, since I wrote all my code to use the new-style setup routines from the start (unlike most drivers, which start out as old-style and get hacked into new-style, I had to go in reverse, which isnt pretty). Several hours were spent tracking a single bug, which was of the wait-three-hours-and-say-doh variety. (Uninitilized pointers are one thing. Uninitialized pointers to uninitizalied pointers which should've even existed are another.) Anyway, abyss, madgemc, and tmspci are all working until 2.2.15pre5 now (madgemc has some problems, but I doubt anyone will notice). I'll get the ISA driver working with the two ISA cards I have (as well as the Olicom OC-3118 which is aparently the same as the AT Plus), and then release major patches for both 2.2.15pre5 and 2.3.42. I've also got someone bugging me for PCMCIA support (in 2.2). We'll see.

During my Oops tracing, I got to play with serial console on x86 (dual-faulting oops that was scrolling off the console then Aiee'ing the interrupt handler, bad). Quite fun. Now all it needs is a way to do a hard-reset of the machine from the serial console. Then it would be just like a real unix box.

Hrm. Enough of this. (I have to go to compandlit tomorrow. Pout. And I'm only on page five of heart of darkness, which we were suppose to have done tomorrow...)

Speedchoice says someone is (illegally) transmitting on my transmit channel frequency. Damnit. They're still going to rot in hell. Even if its not their fault. (At least they were honest. There's probably a large notice on my customer record that says not to give me dumb answers or else I'll call out all the techs...)

This color reminds me of some wonderfully mystical candy out to provide me limitlessly tasty delight.... Or its just annoyingly bright. Either way.

Sluggish day. Only fell asleep in one class though. Physics was just dumb. Calc was slow, but provided much amusement in mocking Stacy (more than usual). Slept in compandlit, after taking a multiple choice test over invisible man. Came home and slept some more.

Madge/Olicom sent me another non-TMS380 Olicom card. Only 16/4. I don't know what I'm going to do with them.

I may be getting some non-Entry IBM Wireless cards. They should be easier to inverse engineer. I think they may even have a PIO/non-DMA mode. They don't have Win95 drivers however. But the ISA/MCA card (yes, two architectures supported on one physical card) seems quite nice, and appears to do on-board wireless-to-tokenring bridging. Looks interesting. Hopefully its got some standard chips on it too.

Okay, thats enough.