| 00:00:41 | _as | Or to give the impression all your products are below normal price. |
| 00:00:44 | prpplague | TomW: ohh, we definetly don't do below cost, we actually make a profit but very little |
| 00:01:21 | TomW | prpplague: for example. "we" spent about $10K writing a software package to simplify choosing the config operations of a device that only sold for $200. We then gave that software away as a free download. We sold a lot of those $200 units. |
| 00:01:32 | prpplague | TomW: our deal is that a comparable product (i.e. a thin client ) is twice as expensive if we buy it from somewhere else, and it doesn't do exactly what we want |
| 00:02:43 | prpplague | TomW: we save a ton of money of tech support and lost customer by making the product easier to use and easier to fix when there is a problem |
| 00:03:07 | TomW | prpplague: before we had the software package, people would have to use a hex programmer (special unit) to put the options into the $200 unit. The also had to use a worksheet to figure out the values to use. |
| 00:03:53 | TomW | prpplague: yes, exactly my point. the benefit in "free" software can be recouped in various ways. |
| 00:03:57 | prpplague | TomW: since we started our embedded approach our daily tech phone support went from +2000 to around 100 |
| 00:04:13 | TomW | prpplague: cool. |
| 00:05:44 | _as | Hiya cb. |
| 00:06:25 | mallum | "John is also one of those long-hair hippie freaks who likes Bach, Buxtehude, and Beethoven -- and some Beatles thrown in for good "measure". |
| 00:06:25 | mallum | " ... hmmm |
| 00:07:03 | TomW | prpplague: I have a customer that choose to build a proprietary software solution on Windoze 98 / NT /2000. I could not convince him of the value of linux. Well, his tech support costs have skyrocketed as he now has customers calling in to say "I installed your software an Quatro Pro started crashing". heh |
| 00:07:38 | TomW | mallum: nothing wrong with Classical, I enjoy it all the time. ;-) |
| 00:08:57 | TomW | prpplague: With linux, he could have tightly controlled the system. BTW, he was shipping the PC to his customers with the software preloaded. |
| 00:09:20 | TomW | prpplague: they then installed "other usefull software" on that same PC. heh |
| 00:09:47 | cbiffle | _as: Howdy. |
| 00:10:46 | cbiffle | So am I the only one that thinks being able to 'md5 kernel' before rebooting from within blob would be a useful thing? :-) |
| 00:10:47 | mallum | TomW: I wasn't critisising, just thought it was funny sounding ... |
| 00:10:54 | TomW | mallum: np |
| 00:11:21 | cbiffle | Since, after all, we're having to give blob 128k aaaanyway...I've fit MD5 in under 9k... |
| 00:11:57 | TomW | mallum: burning flash now. I wanted to test your other apps (minivol, etc) before I moved into revamping the matchbox.mk. |
| 00:12:10 | cbiffle | If anyone else thinks this might be a pleasant and tasty thing to have, I'll go hammer out a patch. (My phone is bricked anyway, so if I'm going to have to reinstall blob, I'm going to do it right, dammit! :-) |
| 00:12:23 | _as | You could fit CRC-32 in a few hundred bytes ID think ;] |
| 00:12:42 | cbiffle | _as: Probably true, probably true. |
| 00:13:13 | cbiffle | _as: I could probably get MD5 in less than that with some creativity. But no promises. :-) |
| 00:13:20 | cbiffle | _as: (Less than 9k, not less than 100 bytes.) |
| 00:14:00 | _as | I think the main problem.. |
| 00:14:23 | _as | Is that by the time youve installed the blob with md5, you dont need to verify it anymore. |
| 00:14:25 | prpplague | TomW: ya thats why we've made a strong push towards linux based thin clients and X |
| 00:15:02 | cbiffle | _as: The reason I'm suddenly interested is that I've installed one bad ramdisk and one bad blob within the past three days; I'd have liked to have known that blob didn't transfer correctly before I rebooted and lost the phone. :-) |
| 00:15:32 | vosque | could borrow the ymodem code from the ipaq bootloader.. |
| 00:15:39 | cbiffle | vosque: Does ymodem do CRC? |
| 00:15:45 | vosque | that has som crc checking in it, I think. |
| 00:15:52 | heather | slaps russ around with a large trout |
| 00:15:55 | heather | teehee |
| 00:15:58 | cbiffle | LOL |
| 00:16:08 | vosque | hahha |
| 00:16:13 | cbiffle | vosque: Yup, that might be good enough. |
| 00:16:41 | cbiffle | has remembered that `sum -o 3` in BSD does standard CRC32, and is amenable to the idea. |
| 00:18:11 | cbiffle | _as: So, do you think some sort of checksumming would come in handy? |
| 00:18:54 | _as | I think if you can run it without a reboot on an unmodified tux that would be very useful. |
| 00:19:01 | prpplague | a friend of russ? |
| 00:19:18 | cbiffle | prpplague: This is a definition of the word 'friend' with which I was unfamiliar, from the looks of that trout! |
| 00:19:31 | cbiffle | _as: You mean without reinstalling blob, or...? |
| 00:19:49 | prpplague | cbiffle: lol |
| 00:20:41 | _as | As a stand alone program for example. |
| 00:21:14 | cbiffle | _as: From within blob? |
| 00:21:23 | _as | From within sboot. |
| 00:21:43 | cbiffle | _as: Hm. |
| 00:21:51 | cbiffle | _as: I don't find myself using sboot very often. :-) |
| 00:22:14 | prpplague | heh, my 4 year son just ask "dad, why does microsoft suck?" |
| 00:22:20 | _as | If you can install blob sucessfully, you dont needa checksum for anything else as you can redo. |
| 00:23:00 | cbiffle | _as: Wrong. :-) Two situations come to mind: (1) upgrading blob, and (2) moving back to Inferno (which seems to be the agreed-upon diagnostic trick for touchscreen/DSP issues around here). |
| 00:23:10 | cbiffle | _as: It was the latter that bricked my phone last night. |
| 00:23:57 | _as | I think its the blob install that people will be most worried about. |
| 00:24:44 | cbiffle | _as: I just find it odd that there are MD5sums of the various components on the website, but no way to check them once they're on the phone. |
| 00:25:40 | _as | I dont like MD5, I think its a sledgehammer to crack a nut. |
| 00:26:08 | cbiffle | CRC32 then. I don't have a preference. I'd just like to ensure that this doesn't happen to anyone else. |
| 00:27:02 | _as | I dont think it will, most people wont upgrade blob or reinstall sboot. It would be a useful feature though. |
| 00:27:40 | _as | We have JTAG to fall back on, but installing blob bothers me the most. |
| 00:27:53 | TomW | mallum: I screwed something up. bbl |
| 00:28:25 | cbiffle | _as: Unless we bundle a JTAG with the phone, I consider a CRC in blob to be considerably more accessible. |
| 00:29:16 | _as | If you need to get rid of blob. |
| 00:29:45 | cbiffle | Well, as soon as I started having hardware issues, the first thing I was told was 'Reinstall Inferno.' So getting rid of blob seems to be a standard step in the diagnostic process. |
| 00:31:27 | _as | Its problem becuase your in the unusual position of wanting to use the phone as a phone. ;] |
| 00:31:35 | _as | Err, probably. |
| 00:31:54 | cbiffle | _as: Silly me. I'll return it to its more traditional use as a bookend shortly. |
| 00:32:19 | _as | I cant use the DSP at all until I make a PFI faker, it doesnt bother me apart from not being able to get an inferno prompt. |
| 00:33:01 | cbiffle | I'm not sure I got an answer before from anyone -- is the LCD driven by the mainboard or the DSP board? |
| 00:33:21 | prpplague | _as: ohh well, who needs the DSP anyway! |
| 00:33:36 | prpplague | cbiffle: the lcd is driven by the main board |
| 00:33:48 | prpplague | cbiffle: but the connections go through the DSP board |
| 00:34:13 | _as | That reminds me. Do we have a LCD datasheet, all the links on the site were dead, as were everything I found on google? |
| 00:34:15 | cbiffle | prpplague: That's what it looked like from the top-layer traces, but I'm hesitant to crack the two board apart to look beneath. |
| 00:34:50 | _as | Its a b*tch to get the DSP board back in again. |
| 00:34:54 | prpplague | _as: i think i have a copy too, let me check |
| 00:35:11 | _as | I had to do some impromptu case remoddling to make it easier with a dremel. |
| 00:36:09 | prpplague | http://www.abcsinc.com/small-linux/ceiva/lm8v31.txt |
| 00:36:22 | _as | Tnx, looking. |
| 00:36:25 | prpplague | _as: there are some specs, but i don't have the pin outs up |
| 00:36:31 | prpplague | _as: they are at the office |
| 00:36:41 | prpplague | _as: remind me monday and i'll post |
| 00:36:57 | _as | Ok, will do, I'm after as much tech info as possible. |
| 00:37:20 | file | returns from his windows-fixing job |
| 00:37:26 | file | stupid Windows |
| 00:37:31 | file | whenever it started, it restarted the computer |
| 00:38:07 | cbiffle | Yaaaaay windows! |
| 00:38:12 | cbiffle | returns to hiding in BSD. |
| 00:38:30 | file | it was evil |
| 00:38:41 | file | all I could say was: use your restore CD, if not - your doomed |
| 00:39:14 | prpplague | _as: hey i found some specs on my machine at the office, give a few to post |
| 00:39:53 | _as | Would you mind emailing them? |
| 00:40:28 | _as | Or whatever is easiest. |
| 00:41:37 | file | waves to prpplague |
| 00:42:26 | prpplague | file: howdy |
| 00:42:36 | prpplague | _as: np |
| 00:42:57 | _as | alchemy@midasalchemy.freeserve.co.uk :] |
| 00:47:34 | prpplague | _as: ahh new place to send spam |
| 00:48:57 | _as | Doh. |
| 00:54:43 | file | curses! |
| 00:54:49 | file | NCURSES! |
| 00:55:25 | file | NCURSES! DANG IT! |
| 00:56:22 | _as | Unexpected reboot? |
| 00:57:03 | file | well |
| 00:57:12 | file | my Windows is unstable |
| 00:57:17 | file | it refuses input from keyboard/mouse randomly |
| 00:57:26 | file | and then a minute later it totally freezes |
| 00:58:03 | _as | Hmm, hardware problem? |
| 00:58:15 | file | never happened with Windows 95'a' |
| 00:58:26 | file | only started when I switched to 'b' |
| 00:58:42 | _as | The reinstall might have allocated different hardware resources. |
| 00:59:10 | _as | Check for problem IRQs, multiple instances of hardware drivers... |
| 00:59:15 | file | doubt it |
| 01:00:12 | file | I did, everything is right |
| 01:00:49 | as_ | Re. |
| 01:01:00 | as_ | Windows used to install multiple monitor drivers. |
| 01:01:21 | file | only mine is on |
| 01:01:47 | as_ | Which used to cause windows protection errors on bootup. |
| 01:02:08 | file | that's never happened to me |
| 01:02:38 | as_ | It happened a lot on my machine, depends on what combination of hardware etc. |
| 01:02:52 | as_ | What 98 is currently doing is forgetting drivers. ITs going senile. |
| 01:03:09 | file | lol |
| 01:03:45 | as_ | I booted up in 16 colour low rez today. |
| 01:03:54 | file | LOL |
| 01:04:09 | Russ | prpplague: she lives downstairs, fellow CSE major |
| 01:04:12 | as_ | It forgot the monitor driver, and the graphics card driver. |
| 01:04:25 | file | hahahaha |
| 01:06:09 | as2 | Re. |
| 01:08:37 | prpplague | Russ: good looking? lol |
| 01:13:08 | prpplague | wow, that was fun! first time i've ever had a linux box kernel panic(without me tinkering) |
| 01:14:02 | file | what did it say? |
| 01:14:12 | prpplague | _as: the doc is on the way |
| 01:14:26 | as2 | Cheers. |
| 01:14:35 | prpplague | _as: look like mozilla cored and then x and then the system |
| 01:15:02 | as2 | ?!? |
| 01:15:21 | prpplague | as2: ya that what i saud |
| 01:15:26 | prpplague | s/saud/said |
| 01:15:49 | as2 | I barely understood 2 consecutive words. |
| 01:16:22 | prpplague | as2: are you refering to my kernel panic? |
| 01:16:48 | as2 | Oh I see, file asked. |
| 01:17:02 | prpplague | as2: oops sorry |
| 01:17:04 | as2 | Taken down by your browser. |
| 01:17:51 | as2 | I missed that line totally. |
| 01:18:06 | prpplague | as2: i really need to re-document those pinouts but i've been swamped |
| 01:18:55 | Russ | prpplague: nvidia driver? |
| 01:20:22 | as2 | How is it wired, does the CPU continuesly strobe data to the screen, or does it talk to a screen controller? |
| 01:20:33 | prpplague | Russ: naw, sis |
| 01:21:18 | prpplague | as2: what are you asking? |
| 01:21:51 | as2 | I dont have a clue how its working. Is there a screen buffer in main RAM I can write directly to for example? |
| 01:22:14 | prpplague | as2: are you talking about the lcd interface? |
| 01:22:50 | as2 | HAng on, line clock pixel clock, this makes a little sense. |
| 01:22:52 | as2 | Yes. |
| 01:23:02 | prpplague | as2: if you are i'm still learn about lcd interfacing, TomW or scanline might be a better source of info |
| 01:23:29 | prpplague | as2: the sa-1100 has a lcd drive mode builtin |
| 01:23:42 | as2 | Ahh, interesting. |
| 01:23:46 | prpplague | as2: it seems to be well documented in the strong arm ref |
| 01:23:54 | as2 | I'll read that then. |
| 01:24:20 | prpplague | as2: the only thing that doesn't seem uniform is the control of the lcd bias voltage and the contrast |
| 01:24:48 | as2 | I'm wondering if sboot has what I need, Ive the feeling linux will want to isolate me from the hardware as much as possible. |
| 01:27:40 | prpplague | as2: if you are a hard core hardware guru, you may find that running a more rich environment like linux on embedded system will tend to make you jump through more hoops to get your work done |
| 01:27:45 | as2 | Well, linux on the PC would do that but I'm unsure if this is setup on the tux. Russ will probably know how the kernal handles user/supervisor mode. |
| 01:28:07 | prpplague | as2: however, IMHO, the work is worth it when you get more flexiblity on the final product |
| 01:28:26 | as2 | Remove the word guru and thats pretty close. |
| 01:28:33 | Russ | the sa1100 is making constant reads to memory |
| 01:28:51 | Russ | (for the lcd) |
| 01:28:53 | prpplague | as2: sorry i flench at the word guru too |
| 01:28:55 | as2 | Is the screen buffer protected memory in linux? |
| 01:29:09 | Russ | everything is by default |
| 01:29:20 | Russ | memory is added to processes, not taken away |
| 01:29:41 | cbiffle | Russ: Have you built a JTAG dongle, btw? |
| 01:29:54 | Russ | ya |
| 01:29:56 | as2 | Does the core run in supervisor mode and the applications in user mode? |
| 01:30:02 | Russ | core? |
| 01:30:06 | vosque | sboot? |
| 01:30:14 | as2 | The OS, linux. |
| 01:30:26 | Russ | I don't know if there is even a special mode |
| 01:30:42 | Russ | just page tables maybe |
| 01:30:50 | cbiffle | Russ: Is it possible to do so without using surface-mount? I notice the PCBs on the 'net are SMD. |
| 01:30:54 | as2 | Theres 4 CPU modes in older ARMs, I assume that will also be true on SA. |
| 01:31:02 | Russ | ya, you can get AHC logic in a DIL |
| 01:31:23 | cbiffle | Russ: I'll have to do that, then. I suspect it'd be a lot cheaper to build one myself. :-) |
| 01:31:26 | prpplague | cbiffle: trust me, unless your are really good soldering, its worth the $30 to get it from BZFlag |
| 01:31:35 | vosque | notes the topic. |
| 01:31:45 | Russ | prpplague: a DIL jtag connector is really easy to make |
| 01:31:50 | vosque | "BZFlag is out of jtag dongles at present" |
| 01:32:02 | cbiffle | prpplague: See vosque's comment. :-) |
| 01:32:02 | as2 | Let me put it another way, if I try writing a dword into the screen buffer from an application process will it be caught and produce an error? |
| 01:32:22 | Russ | you will segfault |
| 01:32:25 | cbiffle | prpplague: I'm good at soldering, but more importantly, I bought the phone using loaned money and need to have it boot to justify the expense. |
| 01:32:36 | Russ | if you open up /dev/fb, and use that, you'll be fine |
| 01:32:38 | prpplague | ohh |
| 01:32:44 | as2 | Ok, I might think seriosly about avoiding linux. |
| 01:33:06 | Russ | as2: if you want to write to it directly, map /dev/mem |
| 01:33:25 | as2 | I dont know enough about linux to know what that means. |
| 01:33:32 | Russ | as2: its a unix thing |
| 01:33:35 | cbiffle | Russ: I know in BSD there's a device, /dev/io I believe, that root-privileged processes can open that causes memory protection to temporarily turn off for memory-mapped devices; is there a similar hack in Linux? |
| 01:33:42 | Russ | as2: you mmap the fb region of /dev/mem |
| 01:33:54 | Russ | as2: but using /dev/fb would probably be a better solution |
| 01:33:55 | prpplague | cbiffle: ohh didn't notice |
| 01:34:04 | Russ | as2: why do you want to write dircetoly |
| 01:34:12 | as2 | I want to vomit data to the screen buffer as fast as possible. |
| 01:34:23 | Russ | as2: then mmap /dev/mem |
| 01:34:28 | prpplague | Russ; got some refernce on making a DIL jtag? or are we talking about the ones on the LART site? |
| 01:34:54 | Russ | just look at the schematic |
| 01:35:21 | Russ | basically, a chip with buffers, and a decoupling cap for it |
| 01:35:39 | as2 | The JTAG interfaces are very simple, its just a buffer to drop the pport lines to 3.3. |
| 01:35:54 | as2 | The SA isnt 5v tolerent. |
| 01:35:57 | cbiffle | Russ: Ah, so mmaping /dev/mem in Linux is a hack around the memory protection? |
| 01:36:12 | Russ | cbiffle: not a hack, its the proper way of doing it |
| 01:36:39 | cbiffle | Russ: Forgive my usage of 'hack.' Perhaps I should have said 'a way to circumvent the memory protection intentionally.' I didn't mean to imply it was hackish. :-) |
| 01:37:00 | Russ | its not a circumention, you are telling the kernel what to do with the page tables |
| 01:37:03 | cbiffle | Russ: It seems like quite a logical way of doing it. |
| 01:37:08 | prpplague | cbiffle: http://www.abcsinc.com/small-linux/mmap.txt |
| 01:37:42 | cbiffle | prpplague: I'm familiar with mmap. :-) |
| 01:38:09 | cbiffle | Russ: Do you think as2'd see the speed increase he's looking for over /dev/fb? It wouldn't surprise me. |
| 01:38:13 | prpplague | cbiffle: ohh sorry, i guess i'm not with it this evening |
| 01:38:20 | cbiffle | prpplague: That's okay. :-) |
| 01:38:28 | Russ | he will have direct access |
| 01:38:50 | Russ | he might want to watch how he does writes and reads to optimize caching |
| 01:38:55 | cbiffle | Russ: Well, yes. |
| 01:39:48 | Russ | q3a |
| 01:40:13 | cbiffle | realizes where he originally knows Russ from: Quake tourneys in the McC lab. |
| 01:41:47 | prpplague | as2: you get your doc? |
| 01:42:08 | as2 | Yep prp. |
| 01:42:23 | prpplague | as2: just checking |
| 01:42:33 | as2 | Thanks :] |
| 01:42:42 | prpplague | np |
| 01:42:46 | prpplague | anything to help |
| 01:43:42 | as2 | I'm a bit bothered by mention of pafge sizes. |
| 01:44:51 | cbiffle | as2: How so? |
| 01:45:17 | as2 | Last ARM I used was linear mapped. |
| 01:45:18 | cbiffle | as2: Bad x86 flashbacks? |
| 01:46:14 | cbiffle | I was rather surprised when I saw discussion of ARM paged memory. I haven't read up on it enough to know if it's just the organizational technique, or if it really does an x86-esque segmentation. |
| 01:47:05 | prpplague | cbiffle: where did you see the discussion? i'd like to add that to my notes |
| 01:47:14 | as2 | x86 segmentation is a legacy problem. |
| 01:47:35 | as2 | ARM doesnt have any history of segmentation or massacaring jews. |
| 01:48:55 | as2 | I wonder how much of linux had to be altered. |
| 01:49:15 | prpplague | you'd have to ask the real guru rmk |
| 01:49:23 | cbiffle | as2: .........riiiight. ARM: It wasn't involved in the holocaust! |
| 01:49:57 | as2 | Well, if it was it was only following orders. |
| 01:50:14 | cbiffle | vetos this entire thread. :-) |
| 01:50:42 | as2 | Moving swiftly onward. |
| 01:51:07 | as2 | The MEMC used to be a seperate chip and RISCOS did everything in usermode anyway. |
| 01:53:17 | prpplague | ahh, so much to learn so little time |
| 01:59:04 | | wish i knew, cbiffle |
| 01:59:04 | cbiffle | ibot: time? |
| 01:59:13 | as2 | 3 am here. |
| 01:59:31 | prpplague | 9pm here |
| 01:59:36 | cbiffle | So yes, having finally had the bright idea of looking at the JTAG schematics, it could be knocked out in discrete 7400-series logic without a hitch. |
| 01:59:45 | prpplague | as2: where in the uk are you? |
| 01:59:50 | cbiffle | (Well, okay, probably not stock 7400s for the 3.3v-ness.) |
| 01:59:55 | cbiffle | 7pm here. |
| 02:00:03 | prpplague | cbiffle: which schematic are you looking at? |
| 02:00:10 | cbiffle | prpplague: jtag4.pdf. |
| 02:00:22 | as2 | Midlands. |
| 02:00:24 | prpplague | cbiffle: on the LART page? |
| 02:00:27 | as2 | Middle england. |
| 02:00:37 | prpplague | as2: ohh ya i asked already |
| 02:00:50 | cbiffle | prpplague: On tuxscreen.net. Now that I've zoomed in enough, the schematic suggests 74VLX-series parts. |
| 02:01:08 | cbiffle | (kghostview gets slower and slower as level of zoom increases, which is the exact opposite of what I'd think would be happening.) |
| 02:01:11 | prpplague | needs to go get a checkup for alzheimiers |
| 02:01:24 | cbiffle | prpplague: Lay off the aluminum canned beverages. :-) |
| 02:01:45 | prpplague | cbiffle: i'll have you know, i only drink beer from bottles! |
| 02:01:51 | cbiffle | prpplague: Good man. |
| 02:02:00 | cbiffle | prpplague: Ahem. Good person of presently indeterminate gender. |
| 02:02:03 | as2 | Every interface IVe seen uses a different subfamily. |
| 02:02:42 | as2 | The quality of your LPT port makes or brakes it. |
| 02:02:53 | cbiffle | as2: If I'm not mistaken, we've just got three bits going to the phone, one coming back, and on each line a buffer to protect from voltage differentials. Duck soup. |
| 02:03:06 | cbiffle | as2: This is true -- that, and whether or not the BSD lpt ports can look like Linux lpt ports. :-) |
| 02:03:27 | as2 | They should support SPP mode. |
| 02:03:36 | as2 | Thats all it needs. |
| 02:03:44 | cbiffle | as2: I just meant in terms of API. |
| 02:04:02 | cbiffle | as2: They moved to a new LPT port device model in 4.x, and while I think the change was for -greater- compatibility with Linux, I don't remember exactly. |
| 02:04:32 | cbiffle | as2: (This is FreeBSD specifically; I can't speak for the others. :-) |
| 02:04:45 | as2 | Hmm, it would be stupid not to have a SPP software interface though. |
| 02:05:12 | cbiffle | SPP is certainly available. What I'm saying is that chances are excellent that the device names and ioctls are hardcoded into the source, and may require tweaking. :-) |
| 02:05:25 | as2 | Fair enough. |
| 02:05:29 | cbiffle | We can't exactly just write bytes into the 0x300 range after all. |
| 02:05:49 | as2 | My motherboard port is glitchy so JTAG wont work, even though normal devices like printers will. |
| 02:05:55 | cbiffle | blinks, realizes that the -serial- port is at 0x3f8, and can't remember where the LPT port is. Ah, well. |
| 02:06:20 | cbiffle | as2: Some latching circuitry in the dongle might help that, actually; if you're getting odd voltage fluctuations, these buffers won't stop it. |
| 02:06:50 | as2 | Theres about 3 locations for it depending on how the hardware has it wired. |
| 02:06:56 | prpplague | cbiffle: http://128.242.116.79/ViewUser.asp?UserID=1210 |
| 02:07:05 | cbiffle | I am curious as to why this schematic uses three-state buffers and simply wires the control lines to ground. |
| 02:07:13 | as2 | I tried latching, doesnt help. |
| 02:07:32 | cbiffle | prpplague: LOL |
| 02:08:27 | as2 | Some interfaces use the HighZ state. And compatability is usually non existant. |
| 02:08:33 | cbiffle | prpplague: I'm at the psych department at Arizona State, and some grad students and I were discussing the fact that higher levels of aluminum in the brain have been used as a diagnostic for Alzheimers. I pointed out that all of them were drinking acidic soda out of aluminum cans, and that I hoped the connection wasn't causal. The next day, every one of them came in with bottles. |
| 02:08:59 | cbiffle | as2: Reeeeally. The great thing about standards is that there are so many to choose from! |
| 02:09:19 | as2 | Some interfaces include a nottapreset line, which is not requred by the standard and rather pointless. |
| 02:09:51 | prpplague | cbiffle: well a lot of underarm deodearants contain large amounts of aluminum componets that are ion unstable |
| 02:09:55 | as2 | As you can reset the tap to a known state from anywhere with 5 TMS=1 pulses. |
| 02:10:01 | cbiffle | prpplague: *nod* It's been implicated in breast cancer, I hear. |
| 02:10:41 | cbiffle | as2: Eeeek. Yuck. So basically we need a PIC or something hanging off a serial port serving as an intermediary. |
| 02:10:46 | cbiffle | (Something with predictable hardware) |
| 02:10:55 | as2 | Aluminium cans have a plastic coating in them. |
| 02:11:11 | cbiffle | as2: I know, I just like to frighten people unnecessarily. :-) |
| 02:11:22 | as2 | Haha. Evil. |
| 02:11:28 | cbiffle | is evil, yes. |
| 02:12:30 | as2 | I want to make a serial port based JTAG dongle, but its even less likley to be compatable with anything anyone else is doing. |
| 02:12:51 | cbiffle | as2: If JFlash is sufficiently modular (haven't looked at it), it could be adapted. |
| 02:13:33 | as2 | I'm largly C illiterate. Which proves to be a major problem, I can read it more or less now. |
| 02:13:42 | cbiffle | as2: One of the nice things about the lpt interface on FreeBSD (Linux may well do this too, I just haven't tried) is that you can get to the control lines of the port as a separate device, so you can toggle/read them just like any other device. |
| 02:14:17 | cbiffle | as2: It's just a portable assembler, imho. The syntax can be a bitch at first, though, particularly with regards to multiple-indirection and data structures. :-) |
| 02:14:21 | as2 | Thats interesting, and err, rather pointless when I think about it. |
| 02:14:25 | cbiffle | LOL |
| 02:14:40 | cbiffle | as2: If you're in SPP mode, you -have- to use the control lines to receive input. |
| 02:14:57 | cbiffle | as2: This just lets you do it with byte writes instead of ioctl calls. |
| 02:15:25 | as2 | It only seems useful if you were going to allocate them seperatly to programs. |
| 02:15:43 | as2 | Which would require people using devices that only used a few lines. |
| 02:16:28 | cbiffle | as2: Scratch what I said about it being a separate device. What in fact happens (it's coming back to me, slowly) is that you can request a separate file-descriptor for the control lines from an API call. So the end effect is the same, but it's not visible to other processes. |
| 02:16:54 | cbiffle | has confused himself. |
| 02:17:07 | as2 | Hmm. |
| 02:17:25 | cbiffle | I'll read up on this and get back to you on it. I haven't used it for a year or so, and that was only to drive a bargraph. :-) |
| 02:18:23 | as2 | I'm not that interested, dont invoke anything like work. |
| 02:18:41 | cbiffle | *grin* |
| 02:20:57 | prpplague | retires to read arm ref manual and drink beer |
| 02:21:08 | prpplague | night all |
| 03:51:55 | TomW | something badly broken here in the latest build. looking... |
| 05:37:53 | TomW | andersee: early to tell. I have to rerun compiles to verify what I just did, but it looks like busybux-0.60.3 breaks things on the Tux. |
| 05:39:36 | TomW | andersee: cardmgr messages for the pcmcia card detection scan report (on boot) is missing returns on lines. tinyx.mk complains something about running out of Socket resouces(?). |
| 05:40:30 | TomW | andersee: I reverted back to busybox-0.60.2 and everything works okay. |
| 05:40:59 | TomW | andersee: I will retrace my steps with a clean copy of the buildroot-tux dl. |
| 06:17:10 | andersee | TomW: Hmmm. |
| 06:17:17 | andersee | TomW: that sounds odd. |
| 06:17:42 | andersee | TomW: tinyx ran out of flash space for me... |
| 06:17:59 | TomW | Almost done building latest buildroot-tux using 0.60.2 busybox. I'll test using that, then if all works, I'll update to 0.60.3 |
| 06:18:07 | andersee | k |
| 06:18:17 | TomW | later. |
| 06:32:37 | TomW | andersee: k, 0.60.2 works with latest build. now to try 0.60.3... |
| 06:45:34 | TomW | andersee: it is something about busybox-0.60.3 |
| 06:46:20 | andersee | :( |
| 06:46:39 | TomW | might not be your fault though. |
| 06:46:46 | TomW | _XSERVTransMakeAllCOTSServerListeners: failed to create listener for local |
| 06:46:46 | TomW | |
| 06:46:46 | TomW | Fatal server error: |
| 06:46:47 | TomW | Failed to establish all listening sockets |
| 06:46:49 | TomW | |
| 06:46:51 | TomW | [1] 63 Exit 1 ./Xfbdev -ac |
| 06:46:55 | TomW | oops, try again... |
| 06:47:00 | andersee | What does df show? |
| 06:47:24 | TomW | /usr/X11R6/bin # ./Xfbdev -ac & |
| 06:47:24 | TomW | /usr/X11R6/bin # _XSERVTransSocketUNIXCreateListener: mkdir(/tmp/.X11-unix) failed, errno = 2 |
| 06:47:24 | TomW | _XSERVTransMakeAllCOTSServerListeners: failed to create listener for local |
| 06:47:24 | TomW | |
| 06:47:24 | TomW | Fatal server error: |
| 06:47:25 | TomW | Failed to establish all listening sockets |
| 06:47:27 | TomW | |
| 06:47:29 | TomW | [1] 63 Exit 1 ./Xfbdev -ac |
| 06:47:44 | TomW | andersee: there is no 127.0.0.1 local in ifconfig. |
| 06:48:06 | andersee | Hmm. |
| 06:48:22 | TomW | sounds like a local socket listener? eh? |
| 06:48:24 | andersee | \/linuxrc problems? |
| 06:48:33 | TomW | could be. looking.. |
| 06:48:39 | andersee | \/linuxrc is supposed to fire up lo |
| 06:49:02 | andersee | I wonder if I accidentally changed the busybox /linuxrc support setting? |
| 06:50:13 | TomW | nope, I added a 127.0.0.1 device (lo) and route, no luck. |
| 06:50:16 | TomW | ifconfig lo 127.0.0.1 up |
| 06:50:16 | TomW | route add -net 127.0.0.0 netmask 255.0.0.0 lo |
| 06:50:38 | TomW | You usihg blob-0.4 yet? |
| 06:50:43 | andersee | nope |
| 06:50:50 | TomW | 2.0.5-pre3? |
| 06:51:43 | andersee | 2.0.4, as hacked by me to do 115200 many months ago |
| 06:52:52 | TomW | 2.0.4 was before they moved the allocations of the flash memory? we now have 3 meg, where the ramdisk was smaller? |
| 06:53:44 | andersee | dunno. I flashed in this version of blob back in like september |
| 06:53:55 | andersee | And havn't touched it since |
| 06:54:06 | TomW | current ramdisk is: /usr/X11R6/bin # df -h |
| 06:54:06 | TomW | Filesystem Size Used Available Use% Mounted on |
| 06:54:06 | TomW | /dev/mtdblock2 3.0M 2.8M 204.0k 93% / |
| 06:54:20 | TomW | yes, tinyx is big. |
| 06:55:17 | andersee | sigh. |
| 06:55:28 | andersee | I wonder what is breaking things for you? |
| 06:55:46 | TomW | I'll copy the 0.60.2 busybox image over the one I currently have, then reboot. maybe it'll give more info? |
| 07:02:31 | TomW | andersee: same problem. |
| 07:03:08 | TomW | doesn't appear to be in the binary itself. |
| 07:04:47 | andersee | Edit build/busybox-0.60.3/Config.h, disable BB_FEATURE_LINUXRC, rm the busybox binary, then recompile... |
| 07:04:50 | TomW | another hint: during the boot of the Tux, the messages print out as expected. Except: after "starting internet superserver". Then the messages get printed out without carriage returns. The text stair-steps down the LCD. |
| 07:05:22 | TomW | k. |
| 07:05:24 | andersee | I think I accidentally enabled BB_FEATURE_LINUXRC,, which is overwriting the default (correct) one.... |
| 07:10:20 | TomW | I uploaded via Zmodem. rebooted, same problem. Do you want a full upload of the ramdisk image? |
| 07:11:05 | andersee | You will need to either reflash, or manually remove /linuxrc and then replace it with the original one. |
| 07:11:20 | andersee | Right now, I bet you $10000000 /linuxrc is a symlink to busybox |
| 07:11:30 | andersee | It should be a shell script |
| 07:11:50 | TomW | I am not going to pay you! |
| 07:11:57 | TomW | it is symlinked. |
| 07:12:13 | andersee | rm it |
| 07:12:14 | TomW | let me upload an older linuxrc. |
| 07:12:19 | TomW | rm it? |
| 07:12:26 | andersee | replace with the original /linuxrc shell script |
| 07:12:34 | TomW | K |
| 07:12:53 | andersee | rm /linuxrc |
| 07:13:02 | andersee | Then replace it with the original /linuxrc shell script |
| 07:13:30 | andersee | Don't just use cp onto the symlink, or you will hose busybox |
| 07:15:18 | TomW | andersee: Give that man his winnings! |
| 07:16:07 | TomW | building new ramdisk image with 0.60.3, target_skeleton/linuxrc will be a text file. |
| 07:16:09 | TomW | ... |
| 07:16:29 | andersee | goes and fixes the busybox patch |
| 07:19:52 | andersee | k. sources/busybox.patch has been fixed |
| 07:21:21 | TomW | k. I will test this (should work ok), then I will do another fressh buildroot-tux checkout and a full build. Let you know abt that one in half hour. k? |
| 07:22:26 | TomW | ... never assume (the full build). |
| 07:23:49 | TomW | BTW, AMD has 2Meg X 16bit Flash that looks to be a drop-in replacement for the 1Meg X 16bit we now have. 8Meg Flash memory in the Tux, anyone? |
| 07:24:19 | andersee | :) |
| 07:24:27 | TomW | heh |
| 07:25:26 | TomW | the "quickie" worked. Now to do the full build... |
| 07:39:41 | TomW | andersee: BTW, AMD also has a 29LV641D device which may boost a Tux to 16Meg of Flash memory... :-) |
| 07:45:46 | fontenot | what are you guys up to? |
| 07:45:49 | fontenot | breaking things? |
| 07:46:25 | TomW | fontenot: nope, fixing stuff. slight prob with busybox. looks like it is fixed now. |
| 07:48:00 | TomW | fontenot: andersee has shaved 50K off the busybox binary. |
| 07:48:35 | fontenot | nice! |
| 07:49:09 | TomW | ... needed room! BTW, AMD also has a 29LV641D device which may boost a Tux to 16Meg of Flash memory... |
| 07:49:37 | andersee | :-) |
| 07:49:47 | andersee | And added features at the same time... |
| 07:50:30 | TomW | but, I want that IDE drive working. then I won't need the flash. I think I'll see if they'll be crazy enough to sample me some parts, then have them on hand. You never know... |
| 07:51:23 | TomW | What do we need libXt for? anything important, like xhost, etc.? |
| 07:51:50 | fontenot | no |
| 07:52:01 | fontenot | i checked (relatively sure) |
| 07:52:04 | TomW | would be nice to reclaim almost a meg by getting rid of that thing. |
| 07:52:09 | fontenot | exactly |
| 07:52:55 | TomW | > 1 meg of free drive space would make this tinyx more palatable to most people. I only have 200K free right now (no matchbox loaded). |
| 07:54:10 | fontenot | remember that since jffs2 is compressed, the 900kB libXt takes about 500kB of flash |
| 07:54:23 | TomW | k |
| 07:54:44 | TomW | drat, thought it would be more. :-( |
| 07:55:50 | fontenot | this is strange - |
| 07:55:50 | TomW | Is D. Jeff Dionne still active at Lineo? |
| 07:55:59 | fontenot | -rwxr-xr-x 1 fontenot fontenot 348344 Apr 26 23:58 libXt.so.6.0* |
| 07:56:05 | fontenot | and that's not even stripped |
| 07:56:17 | fontenot | where did we get the 900kB number ? |
| 07:56:22 | TomW | ? |
| 07:56:46 | fontenot | stripped and compressed with gzip -3 it's: -rwxr-xr-x 1 fontenot fontenot 141352 Apr 28 00:58 libXt.gz* |
| 07:56:56 | fontenot | am i going insane? |
| 07:57:00 | TomW | -rwxr-xr-x 1 root root 829140 Apr 28 2002 libX11.so.6.2 |
| 07:57:06 | TomW | that is from my Tux. |
| 07:57:12 | fontenot | libX11 is required |
| 07:57:20 | fontenot | we are discussing libXt |
| 07:57:36 | TomW | shake head |
| 07:57:39 | TomW | tired.. |
| 07:57:54 | fontenot | that came out a little harsh |
| 07:57:58 | fontenot | didnt mean it that way |
| 07:58:06 | TomW | np, none taken |
| 07:58:15 | fontenot | -rwxr-xr-x 1 fontenot fontenot 816152 Apr 28 01:00 libX11.so.6.2* |
| 07:58:23 | fontenot | that's stripped libX11 for me |
| 07:58:36 | fontenot | -rwxr-xr-x 1 fontenot fontenot 463594 Apr 28 01:00 libX11.gz* |
| 07:58:54 | fontenot | where did I get that 900kB libXt figure? I cannot remember |
| 07:59:10 | fontenot | checking something |
| 07:59:34 | TomW | I am not stripping the Xlibs, can they be stripped? |
| 08:00:05 | fontenot | TomW: yes they can be safely stripped with this command: |
| 08:00:22 | fontenot | arm-uclibc-strip -x --strip-unneeded --remove-section=.note --remove-section=.comment foo.so |
| 08:00:35 | fontenot | if you strip more they may become nonfunctional |
| 08:00:50 | TomW | that is the same command that is in the Makefile as the var $(STRIP). |
| 08:00:54 | fontenot | cool |
| 08:01:02 | fontenot | then you can use it on libs if you are no talready |
| 08:01:02 | TomW | checking... |
| 08:01:07 | fontenot | it will save space |
| 08:01:41 | TomW | no -x in $(STRIP), |
| 08:01:53 | TomW | so I will $(STRIP) -x foo.so |
| 08:02:23 | TomW | How do you strip Xfbdev? Same way? |
| 08:02:50 | fontenot | yes |
| 08:03:00 | fontenot | i got that strip commnad from andersee originally I think |
| 08:03:21 | TomW | That is why your binaries, etc, are smaller. I thought it was due to the compiler versions. |
| 08:04:54 | fontenot | TomW: I think there is a toolchain effect too - I think those numbers I was giving you were unstripped |
| 08:05:01 | TomW | tinyx still compiling.... |
| 08:05:07 | TomW | drums fingers. |
| 08:05:09 | fontenot | when we were comparing binary size |
| 08:05:39 | fontenot | TomW: did you see this: http://intimate.handhelds.org/jacques/matchbox14.png |
| 08:09:40 | TomW | -rwxr-xr-x 1 tom tom 876806 Apr 28 04:12 root/usr/lib/libX11.so.6.2* |
| 08:10:53 | fontenot | stripped? |
| 08:12:07 | TomW | yup, they won't strip. |
| 08:12:13 | TomW | checking.. |
| 08:12:28 | fontenot | wont strip? |
| 08:13:43 | TomW | -rwxr-xr-x 1 tom tom 829140 Apr 28 04:16 libX11.so.6.2* |
| 08:14:11 | TomW | did it manually and DID not use arm-uclibc-strip, used the arm-linux-strip. |
| 08:15:29 | fontenot | ok that's close to my value |
| 08:15:59 | fontenot | both of those strips should work - naturally use the one in the buildroot |
| 08:16:07 | fontenot | i always forget which that is |
| 08:18:17 | TomW | got it. had to use arm-linux-strip to get it done. odd... |
| 08:19:47 | andersee | TomW: just use $(STRIP) -- it is set in the toplevel makefile for you |
| 08:20:37 | andersee | Just add --strip-unneeded for libs |
| 08:20:42 | andersee | i.e. |
| 08:20:51 | andersee | $(STRIP) --strip-unneeded libfoo.so |
| 08:21:16 | TomW | andersee: yup, next compile I will. Thought about the $(STRiP) -x $$file may have been a problem with the -x in wrong position of command. |
| 08:24:11 | TomW | drums fingers... |
| 08:25:53 | fontenot | TomW: did you look at my screenshot? |
| 08:27:08 | TomW | fontenot: ah! you rotated the screen. |
| 08:27:26 | TomW | also have some color there. |
| 08:27:38 | fontenot | yep, xrandr is fun |
| 08:28:36 | TomW | andersee: Works as advertised. No problems, thanks! |
| 08:28:54 | TomW | fontenot: now, to test matchbox again... |
| 08:29:14 | andersee | np |
| 08:30:53 | fontenot | checking out latest buildroot cvs |
| 08:31:09 | fontenot | andersee: any reson for me to dl new uclibc snap since last night? |
| 08:34:56 | andersee | nah |
| 08:35:00 | fontenot | cool |
| 08:35:21 | andersee | I've was busy with busybox last night, and took the day off today |
| 08:35:28 | fontenot | what??? |
| 08:35:32 | fontenot | :-D |
| 08:35:41 | fontenot | took a saturday off - for shame |
| 08:35:44 | andersee | shocking I know... |
| 08:35:59 | andersee | is going straight to hell for taking a day off |
| 08:37:22 | fontenot | lol |
| 08:40:11 | fontenot | so what exciting new features can I expect in busybox 0.60.3 ? |
| 08:40:19 | TomW | no color, but matchbox looks ok. I have 112K left on flash. |
| 08:40:30 | andersee | fontenot: color ls... |
| 08:41:19 | andersee | fontenot: tab completion now looks at your current PATH setting. Previously you only got tab completion for /bin:/usr/bin |
| 08:41:24 | fontenot | 112k wow what are you putting in there? |
| 08:41:27 | andersee | tons of bug fixes |
| 08:41:36 | fontenot | cool |
| 08:45:50 | TomW | andersee: Erik, that color ls looks great in rxvt under matchbox. You da' man! |
| 08:46:12 | andersee | :) |
| 08:46:23 | TomW | pats andersee on the back |
| 08:46:52 | TomW | andersee: ok , you got your "atta boy", back to work with you.. |
| 08:46:59 | TomW | ;-) |
| 08:47:55 | TomW | fontenot: ok, mallum has a tarball that to be used instead of the cvs. I will start on that next, though not tonite. |
| 08:48:13 | TomW | s/that/that is/ |
| 08:49:23 | fontenot | TomW: cool that will protect us from him when he starts rewriting the theme engine |
| 08:54:37 | TomW | fontenot: how did that import command go? import -window root -display 192.168.211.7:0 foo.png ? |
| 08:55:22 | TomW | fontenot: yes, he was pretty adamant about using the matchbox-release-latest.tar.gz |
| 08:56:45 | TomW | fontenot: ok, figured out the import cmd. |
| 08:59:30 | TomW | fontenot: oooh! this looks nice! |
| 09:01:07 | fontenot | i wanna see! |
| 09:01:21 | fontenot | import -window root -display 192.168.1.101:0 matchbox16.png |
| 09:09:50 | TomW | http://tuxscreen.net/wiki/view/Main/BuildingTinyX |
| 09:10:20 | fontenot | coooool nive theme - which one is that? |
| 09:10:23 | fontenot | nice |
| 09:10:35 | TomW | default, :-/ |
| 09:10:43 | fontenot | what's that after the minitime ? |
| 09:10:54 | TomW | miniwave |
| 09:11:05 | TomW | it runs, but plays dead. |
| 09:11:18 | fontenot | what's it do ? |
| 09:11:49 | TomW | dunno, just needed something on the docking bar to clutter it up. ;-) |
| 09:12:57 | fontenot | heh :-) |
| 09:13:06 | TomW | I suspect that the tarball mallum left for me *may* produce somewhat smaller binaries. maybe. |
| 09:13:58 | fontenot | i need to go ahead and unlock another of my tuxes |
| 09:32:23 | fontenot | TomW: what's your site again? |
| 09:36:29 | TomW | http://www.openhardware.net |
| 09:36:54 | TomW | final changes to the files were just commited. |
| 09:44:53 | fontenot | TomW: to which files? |
| 10:05:24 | fontenot | ARG |
| 10:05:37 | fontenot | inconsistent blob flashing instructions! |
| 10:05:52 | fontenot | D!blob F!all______ ??? |
| 10:06:04 | fontenot | c D!tuxscreen-blob F!all ??? |
| 10:06:08 | fontenot | which to believe |
| 10:07:01 | fontenot | anyone awake? |
| 10:08:49 | fontenot | anyone? |
| 10:11:14 | fontenot | dammit |
| 10:12:14 | TomW | fontenot: ?? |
| 10:12:38 | TomW | c D!tuxscreen-blob F!all |
| 10:13:07 | TomW | fontenot: http://www.openhardware.net/TuxScreen/FlashUnlock.html |
| 10:14:20 | TomW | fontenot: changed files were make/tinyx.mk & source/tinyx-011010.patch |
| 10:15:47 | fontenot | TomW: thanks I already did it - the doc on wiki says D!blob F!all______ and I think that is dangerously wrong |
| 10:16:14 | fontenot | TomW: what changes did you make to the tinyx patchfile ? |
| 10:16:20 | fontenot | just curious |
| 10:16:40 | TomW | look at my page. It should be plainly laid out. |
| 10:17:10 | TomW | probably lots of changes today. Best to go get it. |
| 10:17:11 | fontenot | Consider yourself LARTed! |
| 10:17:19 | TomW | heh, |
| 10:17:35 | TomW | picks fontenot off ceiling. |
| 10:17:43 | fontenot | :-D |
| 10:18:13 | TomW | today is Sunday? |
| 10:18:19 | fontenot | yep |
| 10:19:41 | TomW | another week gone.. |
| 10:19:58 | fontenot | yeah :-\ |
| 10:20:08 | fontenot | but it feels like we got a lot done :-) |
| 10:20:25 | TomW | well, we did get a lot accomplished this week. Monday I didn't have a GUI / Window Manager. |
| 10:21:13 | TomW | I keep admiring the screenshot on the Wiki! heh |
| 10:21:51 | fontenot | arg that's right - the x stuff was > week ago |
| 10:22:21 | TomW | ? no, it was this past week, wasn't it? |
| 10:22:23 | fontenot | downloading kernel |
| 10:22:31 | fontenot | TomW: hmm maybe you are right |
| 10:22:38 | fontenot | i am getting old - time is going by too quickly |
| 10:23:12 | TomW | Revision 1.1 / (view) - annotate - [select for diffs] , Tue Apr 23 07:59:11 2002 UTC (5 days, 2 hours ago) by tomwalsh |
| 10:23:20 | TomW | whacks fontenot |
| 10:23:33 | fontenot | cool - so all the tinyx buildroot stuff was in the last week! |
| 10:23:38 | fontenot | that makes me feel good |
| 10:24:30 | fontenot | heh kernel does not like a blank root image region |
| 10:24:31 | TomW | yup, yup, yup. We got to finish it by Monday evening at the latest. No project can extend for more than a week, at least in my shop! |
| 10:24:46 | fontenot | ok we will define it to be done |
| 10:24:47 | TomW | heh, fill up that flash! |
| 10:24:52 | fontenot | and start on another project |
| 10:25:08 | fontenot | yep I am sbout to dl a 2.7MB root |
| 10:25:41 | TomW | ... like getting the IDE interface running. I have that 10G drive mounted inside my Tux, all ready for the electronics (PCB) to be wired to J602. |
| 10:26:34 | TomW | ... board is fabricated, chips installed, "all dressed up and no where to go". |
| 10:26:37 | fontenot | how are you handling the power requirements of the drive? |
| 10:26:53 | TomW | External +5 supply (Wall Wart). |
| 10:27:22 | fontenot | cool |
| 10:28:12 | TomW | That supply may be the main supply for the Tux. I would just need a POR (Power On Reset) signal and a 3.3v supply. I can whip up a 1"x1" board to do that in a jiffy. |
| 10:28:59 | fontenot | Very cool |
| 10:29:17 | TomW | hmmm, maybe a MAX728 is needed. The tux supply is 1.25 Amps at 3.3v. |
| 10:30:34 | TomW | The IDE board is 2.95" X 2.2". It has a 34 pin header on it (0.100" centers) and a 44 pin header as well (2MM). |
| 10:31:00 | TomW | It sits behind the IR diode. |
| 10:31:16 | fontenot | i would Love to see pictures of the internals when you mount these parts |
| 10:32:18 | TomW | Finding some way to attach a ribbon cable to J602 is a bitch. I finally decided to strip the cable wires and solder them to the connector. |
| 10:34:26 | TomW | You will see it, I will document what I did. There is some precise work involved in clearing a space for the board (some standoffs need to be removed). The drive is mounted on the shield over the PCMCIA sockets, a bit tricky locating where to drill the mounting holes in the shield to attach the drive. |
| 10:35:32 | fontenot | ARG |
| 10:35:33 | TomW | I'll feel better once I start getting some data from the drive. I had to use a PLD + 2 latches to be able to talk to the drive via the 8 bit port. |
| 10:35:38 | fontenot | /dev/mtdblock2 3072 3072 0 100% / |
| 10:35:45 | fontenot | my image is too big |
| 10:35:46 | fontenot | ??? |
| 10:35:52 | fontenot | -rw-r--r-- 1 root root 2752512 Apr 28 03:21 tuxscreen-image.jffs2 |
| 10:35:53 | fontenot | ???? |
| 10:35:55 | fontenot | WTF? |
| 10:36:43 | TomW | That is because you need to use blob-2.0.5pre3. 2.0.4 has only 2.5 (?) meg for ramdisk. 2.0.5pre3 has a 3meg ramdisk area. |
| 10:36:59 | fontenot | GRRRRR |
| 10:37:06 | fontenot | hey, why do we have libXaw ? |
| 10:37:09 | fontenot | what's that for ? |
| 10:38:20 | fontenot | TomW: do you know the size in bytes my root image needs to be under ? |
| 10:39:55 | TomW | under? |
| 10:40:04 | TomW | for 2.0.4? |
| 10:40:08 | fontenot | yep |
| 10:40:18 | TomW | Sorry, I don't have that source. |
| 10:40:36 | TomW | You have JTAG? |
| 10:41:28 | fontenot | nope |
| 10:41:51 | fontenot | I mean how many bytes in size should I make sure my root image is under to know it will fit? |
| 10:41:58 | TomW | Maybe you can reblob by having blob put 2.0.5 in, it can do that. |
| 10:42:32 | TomW | I don't know. All my tux's are running 2.0.5pre3 |
| 10:42:54 | TomW | you will have to reblob. |
| 10:42:58 | fontenot | TomW: oK I removed libXaw and libXt - trying again |
| 10:43:14 | fontenot | i am afraid of using newer blob until I know it is safe |
| 10:43:34 | fontenot | why does newer blob allow more space ? |
| 10:44:15 | TomW | because they assumed something like 1-2 meg for the kernel. |
| 10:45:11 | fontenot | ah |
| 10:45:12 | TomW | in the newer blob, they allow 3 meg for ramdisk & 1 meg for kernel (minus 128K for blob + params). |
| 10:45:31 | fontenot | when are they gonna get the new blob blessed? |
| 10:45:38 | fontenot | i mean the one with jffs2 support |
| 10:47:06 | TomW | not sure. BZFlag mumbled something about us needing that jffs kernel soon. |
| 10:47:16 | fontenot | heh we needed it months ago |
| 10:47:23 | fontenot | but now we REALLY need it |
| 10:47:29 | fontenot | since we are running out of space |
| 10:48:15 | fontenot | wow did I just properly receive a file from you? |
| 10:48:22 | TomW | All the files that you will need are in /usr/X11R6 tree. Why not copy that tree to CF or hard drive, then symlink to that image? |
| 10:48:35 | fontenot | BTW, can you give me an md5sum for that blob you sent me? |
| 10:48:48 | TomW | k, ... |
| 10:51:29 | fontenot | BTW, I dont suppose anyone figured out why the blue screen thing with X ? |
| 10:51:38 | TomW | 26a3afe5ef4a680efec43b86912101d8 tuxscreen-blob |
| 10:51:57 | TomW | heh, who has had time? |
| 10:52:19 | TomW | wait! |
| 10:52:42 | TomW | I have a JTAG, let me test that blob first. It is a fresh compile. |
| 10:53:12 | fontenot | lol |
| 10:53:17 | TomW | ?? |
| 10:53:38 | fontenot | just that you sent me an untested blob |
| 10:54:04 | fontenot | do you know the proper way to redirect stderr in the busybox shell ? |
| 10:54:20 | fontenot | is it >& or >2 or ... ? |
| 10:57:00 | TomW | dunno, haven't done much with ash. |
| 10:59:03 | fontenot | it's ./foo > /dev/null 2>&1 |
| 10:59:20 | fontenot | anyway that does not help th blus screen situation like I thought it might |
| 11:00:36 | fontenot | where did you put the mathcbox themes? |
| 11:00:59 | fontenot | found them |
| 11:03:04 | TomW | fontenot: good, it is the same syntax as bash. |
| 11:04:02 | fontenot | hmm calibration is rather off on this tux |
| 11:04:07 | fontenot | not extreme |
| 11:04:13 | fontenot | but annoying |
| 11:04:45 | TomW | fontenot: you have a stock tux? No settings needed for > 16Meg DRAM? |
| 11:05:26 | fontenot | TomW: sorry did not understand the question |
| 11:05:46 | TomW | You have an 8 meg SODIMM in that Tux? |
| 11:06:14 | fontenot | yes |
| 11:06:18 | TomW | total of 16meg DRAM (normal tux as shipped). |
| 11:06:25 | fontenot | yep all mine are like that |
| 11:06:55 | TomW | mine aren't, some 16Meg others 64Meg. |
| 11:07:07 | fontenot | :-) |
| 11:08:04 | TomW | hmmm. for some reason, I cannot build a blob. Good thing I hollered. Sorry abt that. |
| 11:09:46 | fontenot | np |
| 11:11:07 | fontenot | hmm xset wants libXmuu.so.1 |
| 11:11:12 | fontenot | which luckily is tiny |
| 11:11:28 | fontenot | ./xdpyinfo: can't load library 'libXtst.so.6' |
| 11:11:39 | fontenot | ./xhost: can't load library 'libXmuu.so.1' |
| 11:11:47 | fontenot | ./xauth: can't load library 'libXmuu.so.1' |
| 11:11:58 | fontenot | ./xrandr: can't load library 'libXrandr.so.1' |
| 11:11:59 | fontenot | :-D |
| 11:12:03 | fontenot | all tiny |
| 11:44:23 | TomW | fontenot: I have a blob. But I broke my 64Meg Tux. |
| 11:44:26 | TomW | sobs |
| 11:45:36 | TomW | I am uploading the kernel into it now. I changed the version number so that I could be sure that it was the blob I put in it. |
| 11:47:14 | fontenot | TomW: I am not going to try a new blob tonight - so don't know yourself out |
| 11:47:22 | fontenot | knock |
| 11:47:42 | fontenot | TomW: can you fix your 64MB tux with JTAG ? |
| 11:48:18 | TomW | I think it is a ram location of where blob puts itself into place in a 64Meg unit. Too tired to deal with it now. |
| 11:49:26 | fontenot | yeah it's getting really late :-) |
| 11:49:27 | TomW | did you get tinyx running with matchmbox? |
| 11:50:00 | fontenot | on the newly flashed tux? yes |
| 11:50:51 | TomW | cool |
| 11:52:17 | fontenot | :-) |
| 11:52:24 | fontenot | think it's time to make some food now |
| 11:55:41 | TomW | ok, cul |
| 11:55:49 | fontenot | later |
| 12:46:48 | file | hey GPSFan |
| 12:47:22 | GPSFan | morning file: how's Canada today? |
| 12:47:31 | file | sunny |
| 12:47:44 | GPSFan | still snow? |
| 12:47:44 | file | unlike yesterday, thank the lord |
| 12:47:53 | file | a bit of snow on the ground |
| 12:48:04 | file | it was snowing and falling ice pellets yesterday |
| 12:49:04 | file | GPSFan: guess what I'm doing today :) |
| 12:49:30 | GPSFan | haven't the foggiest. |
| 12:49:43 | file | airbrushing my Cybiko |
| 12:50:23 | GPSFan | don't forget to post pics. |
| 12:50:39 | file | I will |
| 19:41:04 | as_ | Evnin all. |
| 20:14:29 | PuckCh | hi. I've problems using the last CVS buildroot : busybox 0.60.3 has link proiblem : nslookup.o: In function `nslookup_main': |
| 20:14:29 | PuckCh | nslookup.o(.text+0x284): undefined reference to `__res_state' |
| 20:14:45 | PuckCh | Is someone else having the same problem ? |
| 20:20:03 | [trebb] | Yep, same problem... |
| 20:20:26 | PuckCh | I've put back 0.60.2, and all is OK. |
| 20:20:45 | [trebb] | commenting out nslookup in Config.h solves this... |
| 20:21:10 | [trebb] | 0.60.3 is nice with the color support. |
| 20:21:37 | PuckCh | ok. Will try this now. |
| 20:26:39 | PuckCh | the compilation is OK without nslookup. Thanks. flashing now... |
| 20:27:00 | [trebb] | fine! :) |
| 20:27:55 | PuckCh | I will make an email on the mailing list about this. |
| 20:30:11 | root__ | is playing with BitchX on his TuxScreen |
| 20:30:47 | root__ | should create a user... |
| 20:33:02 | [trebb] | Did you get TinyX to compile? Thats where I'm struggling right now.. |
| 20:33:40 | [trebb] | No Xfbdev is build... :( |
| 20:34:06 | mallum | hi |
| 20:34:32 | [trebb] | Hi |
| 20:35:36 | PuckCh | I'm only using PicoGUI, not TinyX. Sorry. ;) |
| 20:50:12 | PuckCh | is someone using a CompactFlash ? I've problem with mine. |
| 20:51:48 | [trebb] | Yes, me :) Whats the problem? |
| 20:52:17 | PuckCh | don't see the CF. |
| 20:52:33 | PuckCh | One time, it was OK. But just one time. |
| 20:53:18 | [trebb] | hmm, I just do a mount -text2 /dev/hda /mnt/cf and it is there... |
| 20:54:03 | PuckCh | at this time, it's not an ext2 fs, but a vfat (it's a new CF) |
| 20:54:41 | [trebb] | Well, then mount -tmsdos /dev/hda /mnt/cf should do it. It did for me... |
| 20:54:49 | PuckCh | but I don't think it's the problem. I think the problem is with pcmcia. |
| 20:54:59 | PuckCh | dump_cis don't see the card. |
| 20:57:10 | PuckCh | I've also an ethernet card. And for it, this is OK (after adding the axnet_cs in the kernel) |
| 20:58:50 | PuckCh | I have see onetime some ide_cs cards has problems with timing. maybe this is my problem ? |
| 20:59:17 | PuckCh | but don't know how to change this parameter (a more slow reset) |
| 21:05:39 | TomW | fontenot: we have a color pallete problem. |
| 21:07:05 | [trebb] | PuckCh: hmm, dunno... when I insert the card cardmgr recognises it as hda. Just checked it again. Have to mount it as vfat though. |
| 21:07:25 | [trebb] | Don't know what's he prob with yours, sorry. |
| 21:08:30 | PuckCh | [trebb]: yep. I have only one time the card recognised as hda. Other time, not recognised at all. :( |
| 21:09:23 | PuckCh | But I only have do some reboot, not a physical remove/insert of the card. Will try now... (will take some time, as the Tux is not in the same room ...) |
| 21:11:08 | [trebb] | have to go... cul |
| 21:21:39 | bhima | yawns. |
| 21:21:46 | bhima | haven't been around here in awhile. anybody awake? |
| 21:47:42 | TomW | mallum: you around? |
| 22:36:40 | prpplague | howdy howdy |
| 22:37:41 | TomW | and a very wet good afternoon to you! |
| 22:40:00 | prpplague | TomW: lol, its awfull hot and dry here |
| 22:42:27 | TomW | prpplague: thankfully, it is wet today (and last night!) |
| 22:42:59 | mallum | TomW: hi, around now ... been watching 'the hitcher' |
| 22:43:24 | prpplague | TomW: well i've got most of my traces done for the lcd, i'll have to wait to finish till next weekend |
| 22:47:19 | TomW | mallum: ok. there might be some confusion generated with matchbox-release-latest.tar.gz as it unpacks into a versioned directory: matchbox-0.2.0. |
| 22:48:14 | TomW | mallum: maybe not, as the makefile would fail as it couldn't find the versioned directory to build in. Your thoughts? |
| 22:48:48 | TomW | prpplague: are you laying out a board, or hand wiring it? |
| 22:49:21 | mallum | TomW: hmm, yeah good point |
| 22:50:09 | TomW | mallum: I don't think it will be an issue until you change versions. What would make the Tux build transparent is: |
| 22:50:21 | TomW | mallum: 1. make a --without-miniapm |
| 22:50:40 | TomW | mallum: 2. then we can keep the name of the dir to simply: 'matchbox' |
| 22:51:23 | TomW | mallum: the only thing I have to do to your stuff is patch util/Makefile.{in,am} to remove references to miniapm. |
| 22:51:29 | mallum | TomW: yep |
| 22:51:45 | mallum | TomW: that should be done automatically by the configure |
| 22:52:00 | mallum | TomW: it does the detection and sets a var |
| 22:52:19 | mallum | TomW: ie checks for apm.h |
| 22:52:35 | mallum | TomW: I couldn't get the logic right in Makefile.am for it to work well |
| 22:52:36 | TomW | mallum: aside from getting rid of miniapm, the only other extra-special stuff I have to do is an arm-linux-strip in src/ and util/ for "find $dir -type f -perm +111'. |
| 22:53:11 | TomW | mallum: ... and, of course, rm -rf build/man |
| 22:53:35 | mallum | TomW: probably easiest to make a patch and apply that to fix miniapm etc |
| 22:53:47 | TomW | mallum: --without-miniapm and --without-manpages would be ideal |
| 22:54:05 | prpplague | TomW: i'm hand wiring right now, i need the experience to understand where everything should go |
| 22:54:20 | mallum | TomW: do a DESTDIR=./build make install-strip |
| 22:54:24 | TomW | mallum: yup, that is what I do now, I run a patch file against the source to clean out the Makefile references to miniapm. |
| 22:54:27 | mallum | TomW: Im open to patches :) |
| 22:54:51 | TomW | mallum: k, tnx on the hint to the install-strip. |
| 22:55:21 | TomW | mallum: heh, yeah, I have never fscked with the configure system myself! |
| 22:55:38 | TomW | mallum: no, wait, I have done some work with it.. |
| 22:55:59 | TomW | mallum: tnx, maybe I'll patch you later. :-) |
| 22:56:06 | mallum | TomW: its very wierd to get your head round |
| 22:56:16 | TomW | heh |
| 22:56:39 | TomW | remembers last configure nightmare |
| 22:57:40 | TomW | mallum: I noticed that you stripped out most of the pixmaps in this distro, is that wise? |
| 22:57:51 | TomW | mallum: I mean, they won't be needed. |
| 22:57:57 | TomW | mallum: ? |
| 23:00:35 | TomW | prpplague: with hand wiring, two rules: #1 keep the wires short. #2 keep the wires away from each other, you need to keep crosstalk down to a minimum. So don't give in to bundling the wires together. |
| 23:00:45 | mallum | TomW: stripped out pixmaps ? |
| 23:01:03 | mallum | TomW: there may be some extra unused ones in cvs I havn't removed ? |
| 23:01:36 | prpplague | TomW: ok thanks, what about when doing bundles have everyother wire a ground? |
| 23:02:34 | TomW | mallum: like minitimestrip.xpm? |
| 23:02:55 | TomW | mallum: or is that only because it is incorporated into the binary? |
| 23:03:15 | TomW | Xfbdev minitime oclock xeyes xrandr |
| 23:03:15 | TomW | matchbox minivol rxvt xhost xrdb |
| 23:03:15 | TomW | mbcontrol miniwave xauth xlsfonts xset |
| 23:03:15 | TomW | minisys monolaunch xdpyinfo xmodmap xsetroot |
| 23:03:36 | mallum | TomW: oh yeah its incorperated into the binary - xpms are just c header files |
| 23:03:38 | TomW | That is my target dir for the matchbox stuff (with tinyx + Xapps). |
| 23:04:07 | TomW | mallum: just wanted to check, no xpms needed, except when theme-ing. |
| 23:04:18 | mallum | TomW: yeah, just for theming |
| 23:04:32 | mallum | TomW: autoconf is a nightmare :( |
| 23:04:36 | TomW | prpplague: interlacing GNDs is not needed. |
| 23:04:54 | TomW | mallum: I know, let me think on it, I had done that before. |
| 23:04:56 | prpplague | TomW: ok |
| 23:05:17 | prpplague | TomW: i really appreciate the advice, never had anyone around to ask such things |
| 23:05:25 | TomW | mallum: I do so much shit that I cannot remember, sometimes, just what it is that I *do* know. :-/ |
| 23:06:55 | mallum | TomW: it check if apm.h is present and sets USE_APMLIB, but getting that info over to the actual bin_Programs in ( util/Makefile.am ) just seems impossible |
| 23:07:20 | TomW | mallum: it looks as if this whole package (tinyx, rxvt, matchbox, and buildroot normal apps) is comming in at well under 2 meg of flash. |
| 23:07:58 | mallum | wow |
| 23:08:14 | TomW | mallum: that should make the people with blob-2.0.4 happy, I think they only have a 2 or 2.5 meg Flash area for the ramdisk image. |
| 23:08:35 | TomW | mallum: with blob-2.0.5 you get 3 meg for ramdisk. |
| 23:09:40 | TomW | prpplague: np, always willing to work with those that are willing & able to learn. When I have the time. |
| 23:10:48 | prpplague | TomW: thanks |
| 23:10:50 | TomW | prpplague: thing that drives me nuts is when people see me on a list-server someplace and start sending me personal email to get their questions answered! Grrrr! |
| 23:10:52 | file | no Cybiko coding this week |
| 23:12:08 | prpplague | TomW: would you prefer these questions go to the mailing list? |
| 23:12:45 | prpplague | sjm: howdy |
| 23:12:52 | sjm | morning all |
| 23:12:54 | TomW | prpplague: the question is, how to let these people know that that is considered "consulting" and that I spend some of my time, donating it, to help others when I can via list-server. Some people just haven't a clue what the 'net is about. |
| 23:13:06 | TomW | sjm: hi, we have RAIN! |
| 23:13:24 | TomW | sjm: NorthEastern Pennsylvania. |
| 23:13:54 | TomW | does raindance |
| 23:14:27 | prpplague | TomW: i understand |
| 23:15:25 | prpplague | TomW: thats why i try to keep my questions down to a min and to the point, anyway thanks again, think i need to do some reading |
| 23:15:28 | TomW | prpplague: when I come on an IRC channel, I also expect that. So, your questions are "on-topic". |
| 23:15:29 | prpplague | later folks |
| 23:16:38 | mallum | TomW: I've put a new matchbox 0.2 tarball in my downloadables dir with a small change to the configure stuff |
| 23:16:42 | fontenot | TomW: yeah mallum and I know there is a colormap problem :-\ |
| 23:16:49 | fontenot | TomW: what are you seeing? |
| 23:17:05 | mallum | TomW: it should now only build miniapm if it finds apm.h |
| 23:19:21 | TomW | fontenot: almost a random value thing, like something is not being initialized properly. Somethings the docking bar & rxvt have one color, then next time it will be differnt. Keep running startx from the keyboard and the colors will settle down (after 2-3 invokations) to a steady (but unpredictable) values. |
| 23:19:42 | TomW | s/Somethings/Sometimes/ |
| 23:19:46 | fontenot | TomW: ah that's interesting new information |
| 23:19:58 | fontenot | shower time |
| 23:20:37 | TomW | fontenot: aside from that, the apps behave okay. no oeprational problems that I can see, only that the colors are screwed & unpredictable. |
| 23:20:51 | TomW | fontenot: we are now below 2Meg in image size!!! |
| 23:22:07 | TomW | fontenot: this is thanks to andersee & mallum's optimizing, plus our stripping of the Xlibs. |
| 23:29:51 | TomW | drat! I still don't have this miniapm thing licked in the patches. Ok, on to the config.in... |
| 23:35:26 | mallum | TomW, fontenot: X was apparently developed on a vax with 2 megs of storage |
| 23:35:37 | TomW | HA! |
| 23:35:48 | mallum | TomW: so we've come full circle :) |
| 23:37:07 | mallum | TomW: it didn't have color then of course ... |
| 23:37:49 | mallum | TomW, fontenot: http://zap.crl.dec.com/lxr/http/source/X/ |
| 23:40:37 | mallum | ok bedtime for me, night all |
| 23:41:01 | mallum | TomW: remember to give the new 0.2 tarball a go if you get a chance ... |
| 23:42:04 | fontenot | TomW: COOL how did you get image size down?? |
| 23:43:03 | fontenot | ok just read it |
| 23:43:11 | fontenot | that's great |
| 23:45:52 | TomW | fontenot: -rw-r--r-- 1 tom tom 2097152 Apr 28 19:48 tuxscreen-image.jffs2 |
| 23:51:07 | TomW | erasing d????????????[xmodem upload - Press CTRL-C to quit]????????????? |
| 23:51:08 | TomW | erasing d?Sending tuxscreen-image.jffs2, 16384 blocks: Give your local |
| 23:51:08 | TomW | erasing d?XMODEM receive command now. |
| 23:51:08 | TomW | erasing d?Bytes Sent:2097152 BPS:10239 |
| 23:51:08 | TomW | erasing d |
| 23:51:08 | TomW | erasing d?Transfer complete |
| 23:51:10 | TomW | erasing d |
| 23:51:12 | TomW | erasing d? READY: press any key to continue... |
| 23:51:14 | TomW | erasing d??????????????????????????????????????????????????????????????? |
| 23:51:26 | TomW | |
| 23:51:26 | TomW | CReceived 2097152 (0x00200000) bytes |
| 23:51:26 | TomW | blob> |
| 23:51:36 | TomW | 2 MEG! |
| 23:53:18 | TomW | fontenot: Filesystem Size Used Available Use% Mounted on |
| 23:53:46 | as_ | IE really doesnt like the tuxscreen site. |
| 23:53:53 | TomW | fontenot: Filesystem Size Used Available Use% Mounted on |
| 23:53:53 | TomW | fontenot: /dev/mtdblock2 3.0M 2.5M 524.0k 83% / |
| 23:54:07 | as_ | Twice Ive downloaded files and IE hasnt stopped. |
| 23:54:20 | TomW | as_: heh |
