18:58:40 | seletz | evening |
19:45:05 | prpplague | howdy seletz |
19:45:13 | prpplague | seletz: is it monday already for you? |
19:45:28 | seletz | prpplague: no, not yet. |
19:45:37 | seletz | prpplague: have to finish CF reading |
19:46:06 | seletz | prpplague: tomorrow and tuesday i'll be in munuch, so i've to do it now |
19:46:15 | seletz | s/munuch/munich/ |
19:46:47 | seletz | prpplague: hi, btw :) |
19:47:23 | prpplague | seletz: ahh |
19:48:09 | prpplague | is taking a break on dev work |
19:49:01 | seletz | damned blocks get not read completely. grr. |
19:49:15 | prpplague | i've getting into the mindset of "can't see the trees because of the forest" |
19:49:35 | prpplague | seletz: still working on the psuedo tar scheme? |
19:49:39 | seletz | prpplague: i know that feeling |
19:50:04 | seletz | prpplague: hmm, yup. But there's no "pseudo". Its real posix tar |
19:50:20 | seletz | prpplague: and its only the first step |
19:50:27 | seletz | prpplague: cramfs is next |
19:51:04 | prpplague | seletz: ahh, i didn't know all of the details |
19:51:17 | prpplague | seletz: so is the fat16 option been dropped? |
19:51:29 | seletz | prpplague: no. _I_ need fat16 |
19:52:14 | prpplague | seletz: mind filling me in, if you got a sec? |
19:52:15 | seletz | prpplague: But i need time to get fat16 working. Have to decode the linux fs driver first and do a bit reverse-öengeneering. Have some docs, tho. |
19:52:39 | prpplague | seletz: ohh |
19:52:41 | seletz | prpplague: np, whats the prob? |
19:53:15 | prpplague | seletz: nothing, i was jsut curious what you plans were on the CF stuff, i get it now |
19:53:58 | seletz | prpplague: my final goal ist to have a simple way for system updates (system update=blob, kernel, initrd, cramfs) |
19:54:29 | seletz | prpplague: now peopla disagree on how to store the update file (i.e. fat, cramfs, raw tar). |
19:54:43 | prpplague | seletz: ya that is a sticky area |
19:55:08 | seletz | prpplague: thus i thought it would be best if the CF code, the FS code, the Flash code the partition code know nothing of eachother. |
19:55:55 | prpplague | seletz: i did play with some laptop memory yesterday, they used a smaller type of the connector you suggested on the sa-1110 we talked about |
19:55:59 | prpplague | seletz: nice idea |
19:56:02 | seletz | prpplague: It came to me that it would be best if we had some sort of chainable "io drivers". You caould define read and write chains. |
19:56:26 | prpplague | seletz: i like the sound of that |
19:56:41 | seletz | prpplague: read chain: CF->DOSpart->FAT->FILES |
19:57:01 | seletz | prpplague: write chain: ->BLOBPART->FLASH |
19:57:15 | seletz | erm, sorry |
19:57:38 | seletz | prpplague: read: CF->DOSPART->FAT16->RAM |
19:57:53 | seletz | prpplague: write: RAM->BLOBPART->FLASH |
19:58:15 | prpplague | smooth |
19:58:26 | seletz | prpplague: each of these blocks know nothing of echother. FAT16 is exchangeable with TAR,CRAMFs for ex. |
19:58:58 | seletz | prpplague: or leave it out completely. My test chains (which _do_ work) are: |
19:59:53 | seletz | prpplague: read: RAM0->TAR-> write: RAM1 |
20:00:15 | seletz | prpplague: all i do is a io_copy( src, dest) and the selected file gets copied to ram1 |
20:00:30 | prpplague | seletz: smooth |
20:00:47 | prpplague | seletz: sounds like you doing some nice work |
20:01:08 | seletz | prpplague: CF access is easy too: io_read( dest, src, amount, ptr_to_driver ); |
20:01:35 | seletz | prpplague: like memcpy(). no blocking, no sectors, no heads/cyls. All managed by the driver. |
20:01:37 | prpplague | seletz: what is "amount" blocks,bytes,records? |
20:01:48 | seletz | prpplague: memcpy() |
20:01:51 | seletz | prpplague: bytes |
20:02:10 | seletz | prpplague: the driver does the blocking |
20:02:26 | seletz | prpplague: so it looks like linear RAM |
20:02:30 | prpplague | seletz: how are you determining how many bytes to read? |
20:03:36 | seletz | prpplague: well, you say src="0x420" as src address. i do a src/512. now i know it's block 2. |
20:04:08 | seletz | prpplague: there's some nasty shifting to do to get the offsets right. |
20:04:21 | seletz | prpplague: but _tat_ part works already. |
20:04:29 | seletz | s/tat/that/ |
20:04:37 | seletz | prpplague: got me? |
20:06:34 | prpplague | seletz: ya sorta, FS stuff was never my strong point |
20:06:50 | prpplague | seletz: i guess i need to do some more reading |
20:07:25 | prpplague | seletz: how long will the sa-1100 still be available? |
20:07:27 | seletz | prpplague: it's just like UNIX: do a open("/dev/hda") and do read()s. Don't worry about CHS and blocking. |
20:07:38 | prpplague | seletz: ok gotcha |
20:07:41 | seletz | prpplague: IMHO quite long |
20:07:54 | seletz | prpplague: think all those mobile gadgets |
20:08:09 | seletz | prpplague: oh, sa-1100. Hmm. |
20:08:56 | seletz | prpplague: still i think they'll be available quite long. Maybe INTEL will licence them to others too. |
20:09:43 | seletz | prpplague: btw, something new on the we-want-a-small-board front? |
20:10:29 | prpplague | seletz: nothing new, just struggling with budget issues |
20:11:03 | seletz | prpplague: well, we'll see i guess. |
20:11:28 | prpplague | seletz: ya, i'm seriously considering funding the project myself |
20:13:02 | seletz | prpplague: well. _that_ would be great. I did that too. I have the strong feeling that one could make quite a bit of money with WLAN access points and such. |
20:13:46 | seletz | prpplague: or routers and similar small gadgets. |
20:14:43 | prpplague | seletz: ya i bet, but the main thing i want is something to be able to learn on and incorporate into products with ease |
20:19:29 | seletz | prpplague: What products do you have in mind? |
20:21:29 | prpplague | seletz: what i envision for lack of a better description is a "hack kit" |
20:22:11 | seletz | prpplague: lol |
20:22:44 | seletz | prpplague: well, i guess you'd not need those kits in numbers? Sys 1000? |
20:22:53 | seletz | s/sys/say/ |
20:22:55 | seletz | :) |
20:23:05 | prpplague | seletz: good docs with examples, a board that is easy to develop i/o;s to, and some sort of generic box thats easy to work modifiy |
20:23:24 | prpplague | seletz: ya qty1k would be nice |
20:23:45 | seletz | prpplague: you're kidding, are you? |
20:23:55 | prpplague | seletz: no |
20:24:37 | prpplague | seletz: i get 2 or 3 calls a week from universities wanting to buy my tcsx-1 to teach embedded systems |
20:24:56 | seletz | prpplague: Ahhh, edu merket. Sweet. |
20:25:29 | seletz | prpplague: One would have to make linux geeks want to have it. |
20:25:57 | seletz | prpplague: Like, say, show how to build a mp3 player for a car HiFi with the hack kit. |
20:26:05 | seletz | prpplague: or similar. |
20:26:06 | prpplague | seletz: exactly |
20:26:22 | prpplague | seletz: even promote a "hack contest" we freebies |
20:26:48 | seletz | prpplague: Sounds good. |
20:27:08 | prpplague | seletz: it all comes down to funding |
20:27:12 | seletz | prpplague: well, why not just make those hack kits? |
20:27:58 | seletz | prpplague: well. funding is ok if we get at last some of the kits sold. |
20:28:37 | seletz | prpplague: for me this was a good year so far. Last year was ok too. |
20:29:13 | seletz | prpplague: Those kits could take the same concept fro cpu modules we talked about. |
20:29:34 | prpplague | seletz: IMHO we could 1k units easy of the cost is less than $600 |
20:30:47 | seletz | prpplague: boxed. It's a hack kit, so no real need of a case. No display needed too. |
20:31:42 | seletz | prpplague: One could produce that for say $400 IMHO. |
20:32:26 | prpplague | seletz: hmm that would be nice, and if we picked a standard pactec box and made sure that the mounting holes match that would be perfect |
20:33:00 | seletz | prpplague: yeah. |
20:34:54 | prpplague | drools with anticipation |
20:35:02 | seletz | prpplague: the big part of the funding is to get the terms right with the manufactorer. You always have some initial costs. The initial cost are moderate. But once the production runs, you have to sell constantly. |
20:36:09 | seletz | prpplague: constantly, that is: You order say, 500 units. Then those units need to be sold. |
20:36:42 | prpplague | seletz: what kind of lead time between orders and finished production? |
20:37:53 | seletz | prpplague: well, i guess it's all in 2-3 months range (with beta boards and so on. Tests etc.) |
20:38:11 | seletz | prpplague: maybe even quicker, as the design is already done. |
20:38:26 | seletz | prpplague: schematics i mean |
20:39:21 | seletz | prpplague: how many units could be sold quickly? |
20:40:34 | prpplague | seletz: initialy i think we could sell around 100-200 per month |
20:41:01 | prpplague | seletz: with more press coverage i think that it might go as high as 200-400 |
20:41:34 | seletz | prpplague: you really think that so many geeks are out there? |
20:41:51 | seletz | prpplague: well, EDU market. Hmm. |
20:42:11 | prpplague | seletz: in the US, ya |
20:42:25 | seletz | prpplague: Embedded processing is still hyped in the US? |
20:43:37 | seletz | prpplague: i mean, do universities give courses in embedded computing? |
20:44:16 | seletz | prpplague: if so, they'd need many units. 10-20 for ech course. |
20:44:55 | prpplague | seletz: ya its a hot topic |
20:45:12 | seletz | oh man. |
20:47:35 | prpplague | seletz: so what do think? can we crank out something pretty easy? |
20:47:47 | seletz | prpplague: well, yes. |
20:48:08 | seletz | prpplague: i really mean it. That would be quite easy. |
20:48:42 | prpplague | seletz: are we still looking at purchasing 1000 units at once to make this work? |
20:48:55 | seletz | prpplague: no display, no case. Take paktec if you want a case. take serial LCD module if you want a display. cool. |
20:49:14 | seletz | prpplague: no, that's a different story. |
20:49:43 | seletz | prpplague: That would be better done in "xxx units a month" or so. |
20:49:52 | seletz | prpplague: open ended, so to say. |
20:50:18 | seletz | phone. brb |
20:50:21 | prpplague | np |
20:51:45 | seletz | mam.... |
20:57:19 | prpplague | seletz: why not sent me an email with everything you think can squeeze into the "hack kit" |
20:57:36 | prpplague | seletz: i'm think like a 5x7 board |
20:58:50 | seletz | prpplague: inches or cm :) |
21:01:36 | seletz | prpplague: How do you think the cooperation would work between us? |
21:02:14 | seletz | prpplague: your compyny (absinc i think) would have nothing to do with it, would it? |
21:03:44 | prpplague | seletz: i probably would get them to "sponsor" the project and provide web space and phone services |
21:05:07 | seletz | prpplague: if i could get some sort of "buisness plan" done, i could raise some money to get some backup. |
21:05:50 | seletz | prpplague: i hate those buzzwords, btw. |
21:06:03 | prpplague | seletz: lol, np |
21:06:12 | prpplague | seletz: btw, inchs |
21:11:05 | seletz | grrr, only 0x100 bytes read. rest is junk. Why is this |
21:11:26 | prpplague | seletz: bad i/o day? |
21:12:30 | seletz | looks as if yes |
21:12:59 | seletz | prpplague: i'll keep the hack kit in mind, that's for sure. |
21:16:27 | prpplague | seletz: we can talk more about it at OLS |
21:17:00 | seletz | sure |
22:25:16 | seletz | grr |
22:25:26 | seletz | doesnt work & im tired |
22:25:43 | seletz | needs sleep |
22:25:51 | seletz | see ya |