| 00:14:38 | SpaceCoaster | GoRK: back yet? |
| 00:20:14 | Lenolium | *pling* |
| 00:29:35 | Lenolium | Mornin |
| 00:29:49 | SpaceCoaster | mornin? |
| 00:30:24 | SpaceCoaster | Lenolium: I have a new driver for you to try. |
| 00:33:56 | Lenolium | SpaceCoaster: Cool. wanna mail it my way: rawb@rawb.org? |
| 00:34:44 | SpaceCoaster | Lenolium: sure, it supports polling and uses interrupts |
| 00:35:32 | Lenolium | Great! |
| 00:35:55 | Lenolium | I have to work on sunday, but I hope to be able to get something out before I have to go in (6-7pm tommorrow) |
| 00:36:19 | SpaceCoaster | It no longer flashes and beeps at startup, in fact it does nothing until the open call when it resets the DSP. |
| 00:37:30 | SpaceCoaster | GoRK discovered that it needs a sanity cmd sent every few seconds, I checked it and 4 seconds is the most, so every two is about right. |
| 00:38:16 | SpaceCoaster | The buzby.c has an example buzby -poll. If the DSP gets broken, close+open will reset it. |
| 00:39:11 | Lenolium | SpaceCoaster: Cool. |
| 00:39:31 | Lenolium | SpaceCoaster: And the idle handling is handled by the daemon, and not the driver, right? |
| 00:40:16 | SpaceCoaster | Correct, as we have polling this seems sensible, when the driver is closed it doesn't do anything, so why keep the DSP alive? |
| 00:41:15 | SpaceCoaster | I has implements a circular buffer and has upto 64 responses from the DSP, after that older ones are thrown away. |
| 00:41:32 | SpaceCoaster | s/I has/It/ |
| 00:50:18 | Lenolium | SpaceCoaster: Sounds good. |
| 00:50:52 | SpaceCoaster | just going to run a test and then mail it to you and GoRK. |
| 00:51:17 | Lenolium | Cool. |
| 00:55:54 | SpaceCoaster | Lenolium: sent |
| 00:58:48 | Lenolium | SpaceCoaster: THX |
| 01:02:05 | GPSFan | SpaceCoaster: hi, you still awake? |
| 01:02:16 | SpaceCoaster | yes |
| 01:02:48 | GPSFan | Would you be interested in a schematic of the Intel dev bord connections for the UCB? |
| 01:03:30 | SpaceCoaster | indeed! and GoRK would be even more interested than me |
| 01:04:12 | GPSFan | OK, I'll put it up in the schematics section, the UCB is a UCB1100, but is the same pinout as the UCB1200. |
| 01:05:52 | GPSFan | The file is codecs.pdf |
| 01:07:33 | SpaceCoaster | GPSFan: got it thanks |
| 01:07:36 | GPSFan | I was looking at the spi interface and wondered if it was the same as on Brutus, it is close, Brutus uses GPIO22 for the interrupt wheras Tux uses 16. |
| 01:58:29 | SpaceCoaster | root |
| 02:21:52 | Lenolium | is away: Dancing. |
| 04:32:22 | hypaStyle | hey anyone around ? |
| 04:32:26 | Alehaz | yep |
| 04:33:18 | hypaStyle | do you know of any phone related software that runs on the tuxscreen |
| 04:33:26 | hypaStyle | at the website it has a lot of suggested features |
| 04:33:31 | hypaStyle | but no actual software |
| 04:36:33 | Alehaz | yes |
| 04:36:38 | Alehaz | go look at the wiki |
| 04:36:48 | Alehaz | the buildroot page should be interesting |
| 04:36:54 | Alehaz | also the microwindows page |
| 04:37:26 | Alehaz | we're still working on supporting all the hardware. the telephony stuff is just starting to work. |
| 04:40:02 | hypaStyle | i also couldnt find a whole lot of docs on the apis to talk to the phone hardware |
| 04:40:07 | hypaStyle | or is that still the works ? |
| 04:42:40 | Alehaz | we've got a spec, it's brand new |
| 04:43:07 | hypaStyle | linkage |
| 04:43:28 | Alehaz | http://cosmic.censoft.com/dspspec.html |
| 04:43:38 | Alehaz | http://www.tuxscreen.net/wiki/view//LucentDSP |
| 04:43:55 | Alehaz | it's just hooked up to a bunch of gpio pins |
| 04:44:06 | Alehaz | communicating with the DSP is slow and painful |
| 04:44:11 | hypaStyle | straight dsp commands |
| 04:44:13 | hypaStyle | ouch |
| 04:44:20 | hypaStyle | no easy interface hehe |
| 04:44:22 | Alehaz | 500ms udelay |
| 04:44:27 | hypaStyle | 500ms |
| 04:44:31 | hypaStyle | JESUS |
| 04:44:32 | Alehaz | nah, we have a simple interface now |
| 04:44:50 | hypaStyle | talking straight to the dsp is a simple interface ? |
| 04:45:15 | Alehaz | well, we have a simple device driver for pushing bits at the dsp, and a simple c program for sending raw commands |
| 04:45:18 | Alehaz | and reading responses |
| 04:45:23 | hypaStyle | ah |
| 04:45:26 | Alehaz | and we can turn stuff on and off and whatnot |
| 04:45:45 | hypaStyle | can you get all the caller id info and everything in that first 500ms |
| 04:45:56 | Alehaz | yeah |
| 04:46:22 | hypaStyle | can you actually capture the data from the phone and make an answering machine program ? |
| 04:46:43 | Alehaz | It turns out that the DSP audio is not physically connected to the SA1100 audio |
| 04:46:58 | Alehaz | but there are like seven relays on the dsp board with nothing hooked up |
| 04:47:12 | hypaStyle | ouch :( |
| 04:47:27 | Alehaz | so it's possible there could be a simple modification to hook up the dsp to the ucb1200 |
| 04:47:30 | hypaStyle | so how can you do like the call filtering thing if you cant talk to it |
| 04:47:36 | Alehaz | which would allow recording and playing back audio |
| 04:47:55 | Alehaz | what do you mean? |
| 04:48:12 | hypaStyle | one of the things they recommended in the faqs |
| 04:48:18 | hypaStyle | was to make a call filtering program |
| 04:48:24 | Alehaz | hmm |
| 04:48:26 | hypaStyle | that if telemarketers call |
| 04:48:33 | hypaStyle | they hit 1 if its spam |
| 04:48:39 | hypaStyle | 2 if there not selling something |
| 04:48:46 | Alehaz | oh, yeah, probably based on not knowing the audio wasn't connected |
| 04:49:08 | hypaStyle | so if you did those hookups that would all be possible ? |
| 04:49:13 | Alehaz | people at philips & lucent tell us that they've modified some of theirs to be VoIP gateways, so it's definately possible |
| 04:49:20 | Alehaz | just require some soldering |
| 04:49:27 | Alehaz | and maybe some minor bits & pieces |
| 04:49:27 | hypaStyle | ah ok |
| 04:49:37 | hypaStyle | so its not a really useful device till you do that |
| 04:49:48 | hypaStyle | all you can get from the phone is caller id info until you do the mods right |
| 04:49:57 | Alehaz | well yeah |
| 04:50:07 | Alehaz | and you can controll the telephone with software |
| 04:50:18 | hypaStyle | to dial ? |
| 04:50:43 | Alehaz | we haven't worked out the precise mods yet - the major stuff getting worked on is auto-vcc for the pcmcia slots, and figuring out how to address more than 16 megs of memory |
| 04:50:56 | Alehaz | Well, without a daemon talking to the dsp, it only pulse dials |
| 04:51:12 | Alehaz | and it needs a watchdog on a 2 second timer or it resets itself |
| 04:51:36 | hypaStyle | hrmm |
| 04:51:41 | hypaStyle | so the project is in its infancy ? |
| 04:51:52 | Alehaz | nah, it's definately adolescent at this point |
| 04:52:05 | Alehaz | you weren't here when we were just figuring out how to calibrate the touchscreen |
| 04:52:19 | hypaStyle | so is inferno or linux where most of the development being done on the device ? |
| 04:52:31 | Alehaz | inferno is basically done and forgotten |
| 04:52:37 | Alehaz | in 1998 |
| 04:52:53 | Alehaz | has web, email, address book, with dialer and caller id |
| 04:52:55 | hypaStyle | what is actually on the inferno version any software ? |
| 04:52:57 | hypaStyle | oh ok |
| 04:53:02 | hypaStyle | does that come with the device ? |
| 04:53:04 | Alehaz | sortof a cheezy interface |
| 04:53:08 | Alehaz | yeah, comes on the flash chips in it |
| 04:53:18 | Alehaz | oh, and the browser is useless |
| 04:53:25 | Alehaz | inferno won't use the pcmcia slots at all |
| 04:53:32 | hypaStyle | ah |
| 04:53:48 | hypaStyle | does the linux version have the callerid and dialer software ? |
| 04:53:59 | Alehaz | linux uses them for anything that's 3.3v at the moment - nics, scsi, microdrive, etc |
| 04:54:09 | hypaStyle | ya i saw that |
| 04:54:10 | Alehaz | not yet, but i imagine it'll materialize very soon |
| 04:54:31 | hypaStyle | so you have to reinsert the pcmcia devices when you reboot it still ? |
| 04:54:48 | Alehaz | nah |
| 04:54:51 | hypaStyle | i saw that at the website |
| 04:54:56 | hypaStyle | k cool |
| 04:54:59 | Alehaz | well, hmm, maybe you do, Idunno |
| 04:55:07 | Alehaz | there's no reset line hooked up to the pcmcia slots |
| 04:55:08 | hypaStyle | you use one at your house ? |
| 04:55:12 | hypaStyle | ah |
| 04:55:24 | Alehaz | but i think when you reset the whole phone the cards get reset, not sure |
| 04:55:29 | Alehaz | I don't recall |
| 04:55:32 | Alehaz | yeah, I have one here |
| 04:55:44 | hypaStyle | ya what software do you use on it ? |
| 04:56:16 | Alehaz | Right now it's running the Familiar distribution ment for ipaqs, it's got xchat displayed on xwindows from another box |
| 04:56:53 | hypaStyle | none of the phone functions though eh |
| 04:56:54 | hypaStyle | ? |
| 04:57:18 | Alehaz | well, the phone rings, and you can talk on it, and you can make calls if you don't mind pulse dialing |
| 04:57:28 | Alehaz | but the speakerphone doesn't work or anything |
| 04:57:45 | Alehaz | that stuff will be easy to implement now that we have a programming spec |
| 04:57:46 | hypaStyle | speakerphone requires software on the pc |
| 04:57:49 | Alehaz | we've only had the spec for a few days |
| 04:57:53 | hypaStyle | ah cool |
| 05:00:14 | hypaStyle | are there anyother websites or chatrooms that talk about development for the phone |
| 05:00:56 | Alehaz | there's some QNX people working on some of them |
| 05:01:00 | Alehaz | I forget where they're at |
| 05:01:39 | Alehaz | if you come around here during the day during the week, someone will prolly know |
| 05:02:57 | Alehaz | this irc channel and the wiki are pretty much it for linux on the phone |
| 05:03:07 | Alehaz | there's a mailing list too |
| 05:20:55 | hypaStyle | mailing list |
| 05:20:59 | hypaStyle | where do i subscribe |
| 05:25:09 | Alehaz | http://lists.sourceforge.net/lists/listinfo/tuxscreen-devel |
| 05:29:37 | collord | ...grrr pcmcia sucks... and i don't even have a tuxscreen yet |
| 05:29:56 | Alehaz | collord: why do you say that? |
| 05:30:25 | collord | i'm just fighting with a dlink 802.11 cardbus and an old toshiba... |
| 05:30:43 | Alehaz | I didn't think there were any cardbus wireless cards |
| 05:30:47 | Alehaz | you mean CF? |
| 05:31:19 | collord | the dlink 650 claims to require 32-bit cardbus, and i can't get it running under PCIC |
| 05:32:03 | Alehaz | dwl-650? |
| 05:32:25 | Alehaz | is the business end of the card gold plated & bumpy? |
| 05:32:35 | collord | yes, that's the one. |
| 05:33:11 | collord | err, bumpy but only gray plastic. (Both sides are business ends) |
| 05:33:23 | Alehaz | I mean the pcmcia connector end |
| 05:33:36 | Alehaz | cardbus has a row of bumps, and is gold plated |
| 05:33:40 | Alehaz | for better grounding |
| 05:33:50 | Alehaz | it also physically doesn't insert into 16 bit slots |
| 05:34:20 | Alehaz | http://www.dlink.com/products/wireless/dwl650/ <-- this says it's 16 bit pcmcia type II |
| 05:34:30 | Alehaz | which toshiba are you using? |
| 05:35:02 | Alehaz | my notebook is a relatively old toshiba |
| 05:36:42 | collord | tecra 740CDT. kernel tree pcmcia, 2.4.12-ac3. |
| 05:37:12 | Alehaz | eu, don't use kernel pcmcia unless you need to |
| 05:37:20 | Alehaz | it barely works |
| 05:37:26 | Alehaz | arm needs it because nothing else works |
| 05:37:50 | Alehaz | but on x86 you're far better off using david hinds' pcmcia-cs |
| 05:38:20 | other | i've used pcmcia-cs with arm |
| 05:38:25 | other | not the networking stuff tho.. |
| 05:38:40 | Alehaz | other: yeah but the actual socket driver is in the kernel |
| 05:38:46 | other | yeah |
| 05:39:01 | Alehaz | because the i82365 driver from pcmcia-cs can't work on arm |
| 05:39:06 | other | im talking the card spec modules |
| 05:39:19 | collord | that's the route i'm pursuing after seeing the sizable diff between the two i82365.c |
| 05:39:24 | other | ide-cs in the kernel was[is?] severely b0rken |
| 05:39:44 | collord | it just torques me that linux can't get stuff like this in one place and easy |
| 05:40:15 | Alehaz | I think it'd be nice if the pcmcia-cs tree could be integrated into the kernel but it seems like that's looking like a lot of work, and dave hinds keeps a pretty good lid on things as it is |
| 05:42:06 | Alehaz | *Shrug* |
| 05:42:39 | Alehaz | other: pcmcia-cs's ide_cs has issues too, with respect to not being able to share interrupts |
| 05:42:49 | Alehaz | other: dahinds says it |
| 05:42:55 | Alehaz | says it's a bug in the kernel ide driver |
| 05:43:04 | collord | Alehaz: I'm pretty sure this card is cardbus... it's in reqs elsewhere. my linksys prism2 worked perfectly with this config though. |
| 05:44:32 | Alehaz | collord: prism2 is a 16 bit chipset |
| 05:44:44 | Alehaz | collord: I'm connected to the net right now with a prism2 nic |
| 05:44:49 | Alehaz | it's definately not cardbus |
| 05:45:43 | collord | Alehaz: well this card won't work in my tosh but will in my newer micron and a Ricoh PCI adapter *shrug* |
| 05:45:56 | Alehaz | I'm on a toshiba portege 3015 |
| 05:46:36 | Alehaz | collord: Look at the end of the card - if it's pcmcia type II, one side of the end of the card will have thin tabs top and bottom, and the other will have a thin tab just on the bottom |
| 05:46:53 | collord | all running kernel tree pcmcia... if not cardbus probs, then general support for ToPIC95 lacking in the kernel tree pcmcia |
| 05:46:56 | Alehaz | if it's cardbus it will have a thick tab that goes almost the whole thickness of the card, to prevent it from being inserted into 16 bit slots |
| 05:47:26 | collord | you're right, the physical interface is the 16 bit pcmcia |
| 05:47:57 | Alehaz | ToPIC95 here, uses i82365.o |
| 05:48:08 | Alehaz | and it's a cardbus controller |
| 05:48:24 | collord | you have PCIC or Cardbus/16-bit option in bios? |
| 05:48:30 | Alehaz | I don |
| 05:48:37 | Alehaz | don't remember - I *think pcic |
| 05:48:46 | Alehaz | but it doesn't actually matter, so long as it's not "auto" |
| 05:51:25 | Lenolium | is back (gone 03:29:30) |
| 05:51:38 | Alehaz | i remember it had to be pcic to use a lame dos app to upgrade the firmware on some lame wireless nics i ended up not using |
| 05:53:36 | collord | fyi the kernel tree crap won't even start if not in PCIC mode |
| 05:54:01 | Lenolium | hypaStyle: I'm working on a better interface to the DSP tonight. Should have something to show by Sunday night (hopefully) |
| 05:55:48 | collord | Alehaz: anyway, thanks. this is OT so I'll shut up now. |
| 05:56:25 | Alehaz | collord: s'alright |
| 05:56:47 | Alehaz | collord: just wait until you have to start arguing with resource allocation, then you'll *really hate pcmcia |
| 05:57:14 | Alehaz | if you want to get the full effect, go find a Actiontec 750 pci pcmcia kit and an older Athlon and try to make it work |
| 05:57:23 | Lenolium | is going to start really hating PCMCIA soon, he will be attempting to put a 16-bit ISA PCMCIA adapter kit in an Alpha running Linux |
| 05:58:02 | Lenolium | Alehaz: I have an actiontec 750 card, just no frontend drive. (stupid ignoring big red letters on ebay auctions) |
| 05:58:05 | Alehaz | Lenolium: I've had pretty good luck with my isa adapter kits, but they've all been in peecees |
| 05:58:30 | Alehaz | Lenolium: I have one w/ the front end drive at lineo which is a paperweight |
| 05:58:54 | Alehaz | Lenolium: the Ti cardbus bridge on those things requires that the bios be smart enough to resource allocate before boot time |
| 05:59:05 | Alehaz | or that the cardbus driver be smart enough to initialize the pci bus |
| 05:59:27 | Alehaz | spent hour after hour trying to make 'em work in AMD 751 chipset mobos |
| 05:59:30 | Alehaz | got nowhere |
| 05:59:37 | Alehaz | complete waste of time & money |
| 06:00:10 | Alehaz | dahinds said adding support should be "easy" but "tricky" |
| 06:00:27 | Alehaz | by "easy" he must mean "if you write pci bridge drivers in your sleep this is easy" |
| 06:00:38 | Alehaz | ANYway . . . |
| 06:01:29 | Lenolium | Alehaz: So, I take it they are not supported under Linux? |
| 06:03:27 | Alehaz | Lenolium: they are, but if the bios doesn't know to allocate resources for both slots, and if the pci fixup code in the cardbus driver doesn't know how to map resources on your pci chipset, you're up shit creek |
| 06:04:07 | Alehaz | I forget which c file it's in in pcmcia-cs, there's a pci fixup program that has specific instructions for a whole list of pci bridges |
| 06:04:25 | Alehaz | not much code per pci chipset, but each one looks like chicken scratches . . |
| 06:10:33 | collord | cha-cha-cha *kernel panic* cha-cha-cha |
| 06:11:05 | AlHaz | (yes, I have more than one computer online) |
| 06:14:51 | AlHaz | I spent the day trying to get a dvd decoder board working under linux . . . man, what a pain in the ass . . . |
| 06:16:28 | hypaStyle | Lenolium : can you tell me a little more about your interface ? |
| 06:17:38 | collord | AlHaz: success? |
| 06:19:07 | AlHaz | collord: Nope |
| 06:19:21 | Lenolium | hypaStyle: Well, I'm just starting on it, but it will be a daemon that sits around and handles the dsp, resets it when it needs it and generally keeps it happy. What my daemon will do is to let multiple things connect to the DSP at one time, and they can feed it commands, or can tell it what events they are interested in. |
| 06:19:24 | AlHaz | I've got a Hollywood+ board and an ide dvd drive i picked up in like |
| 06:19:26 | AlHaz | '98 |
| 06:19:43 | AlHaz | allegedly the HW+ can be used under linux with an experimental driver |
| 06:20:13 | Lenolium | hypaStyle: All of the kernel work has been done by <SpaceCoaster>, so I'm just making a cleaner way to get at it. |
| 06:20:20 | AlHaz | there are at least 60 or so permutations of configuration options, all marked "this might solve some problem or other for some people, or not" |
| 06:20:38 | Lenolium | AlHaz: Now that's a quality driver. ;) |
| 06:21:04 | AlHaz | Lenolium: Well, I think they haven't figured out a reliable way to determine the hardware revisions |
| 06:21:24 | AlHaz | if they could accurately detect your hardware rev they could automagically select which options you need |
| 06:21:37 | AlHaz | or at the very least they could do like bttv does, and just have card=foo |
| 06:21:42 | Lenolium | AlHaz: Ahh.. so they are doing stuff like linksys did with their ethernet cards, four cards, one PCI-ID. |
| 06:21:47 | AlHaz | Lenolium: Yeah |
| 06:23:16 | AlHaz | Bus 0, device 17, function 0: |
| 06:23:17 | AlHaz | Multimedia controller: Unknown vendor Unknown device (rev 1). |
| 06:23:17 | AlHaz | Vendor id=1105. Device id=8300. |
| 06:23:37 | AlHaz | "8300" means Sigma EM8300 |
| 06:23:47 | AlHaz | which is only a third of the chips on the board . . . |
| 06:23:52 | AlHaz | they all say they're rev 1 |
| 06:24:02 | AlHaz | and they're all id 1105 . . . |
| 06:24:07 | AlHaz | vendor, even |
| 06:24:33 | AlHaz | there's an eeprom on the board, on the i2c bus, but that doesn't say much either |
| 06:25:13 | AlHaz | so there's various pal and ntsc versions of the card |
| 06:25:25 | AlHaz | and they all look basically more or less the same on both pci and i2c |
| 06:25:25 | Lenolium | AlHaz: Oh, yay. |
| 06:25:28 | AlHaz | or at least really close |
| 06:25:50 | AlHaz | there must be subtle differences since the windows drivers seem to tell them apart |
| 06:26:06 | AlHaz | So then you take the windows vxd, and you run this perl script against it that lifts out the microcode |
| 06:26:11 | Lenolium | Why do bizarre multimedia devices seem to think that they don't need to follow any of the rules. |
| 06:26:13 | AlHaz | and there are three versions of the microcode in it |
| 06:26:27 | AlHaz | and only one is likely to work |
| 06:26:38 | Lenolium | Oh, now that sounds like fun. |
| 06:26:58 | AlHaz | Lenolium: ESR has insinuated that this is the reason it's so hard to get programming specs for a lot of them |
| 06:27:12 | AlHaz | Lenolium: it's not because they're protecting their IP, it's because they're EMBARRASED |
| 06:27:35 | AlHaz | makes sense to me . . . |
| 06:29:03 | AlHaz | but the way a disturbing number of suits think, they probably think they're protecting precious secrets that their entire business depends on |
| 06:29:55 | Lenolium | AlHaz: I think that a lot of hardware is probably designed that way. "Hey guys, did you see the new schedule, we're supposed to have this done, next week. Quick! Hack something together!" |
| 06:30:26 | AlHaz | which is dumb . . . if engineers at competing companies want to know how something works, they'll damn well find out one way or another. short of putting all required componentry on one board and encasing it in epoxy and titanium, you really can't protect your IP |
| 06:31:03 | AlHaz | and even then, there are a dozen companies in the world that literally shave the packaging off your chips and put 'em under microscopes |
| 06:31:34 | hypaStyle | arent there some anti reverse engineering laws in the usa |
| 06:31:46 | hypaStyle | relating to dmca |
| 06:32:14 | AlHaz | hypaStyle: the DMCA implies a lot of that, but US law doesn't apply outside of the US |
| 06:32:19 | hypaStyle | tru |
| 06:32:25 | AlHaz | no matter how much legislators want it to |
| 06:32:43 | hypaStyle | ya |
| 06:32:53 | AlHaz | so these companies are in canada, hong kong, etc |
| 06:32:58 | hypaStyle | ya |
| 06:33:14 | AlHaz | there are european countries where reverse engineering is a right explicitly granted to the product holder |
| 06:34:08 | Lenolium | If you are running in a highly competitive hardware business, you shouldn't depend on others not being able to copy what you have done, you should focus on getting something better out the door before they get a chance. Example: Nvidia |
| 06:34:34 | AlHaz | Yeah, NVidia kicked ass by having the best product |
| 06:34:57 | AlHaz | their engineering process is interesting, too |
| 06:35:03 | Lenolium | Yup, and releasing a new product every six months (even if they are only marginally better than the previous ones) |
| 06:35:03 | AlHaz | "obsolete our product before anyone else can" |
| 06:35:23 | AlHaz | Lenolium: I hear they have two engineering teams working in competition with eachother |
| 06:35:26 | AlHaz | at all times |
| 06:35:42 | Lenolium | After a certian level of complexity, it becomes easier to figure it out yourself than use somebody else's stuff. |
| 06:35:54 | collord | ...and they are binary-only driver people... |
| 06:36:16 | Lenolium | AlHaz: Oh boy, with the downturn of the economy, I bet the competeion is REALLY fierce right about now. |
| 06:36:33 | AlHaz | collord: nvidia's got suits that are convinced that keeping things "secret" is vital, probably |
| 06:36:44 | AlHaz | they seem to waffle |
| 06:36:52 | Lenolium | collord: Of course, no hardware hacker could ever understand assembly... *wink* |
| 06:37:16 | AlHaz | I mean they released some stuff in the past that was open, and then they obfuscated the code, and then they deobfuscated it, and then they started releasing binaries . . . |
| 06:37:19 | collord | AlHaz: of course. plus, DRI stinks and nvidia rules. |
| 06:37:23 | AlHaz | they need a swift kick in the pants |
| 06:38:04 | AlHaz | I don't use 3d so much. If i want to play games, the duron downstairs has a win98 partition and a geforce2-mx |
| 06:38:13 | collord | (in the language of #loki, sorry) |
| 06:38:15 | AlHaz | the big linux box has a Matrox G200 |
| 06:38:45 | AlHaz | I need to get the matrox stuff working right, I don't even have Xv at this point |
| 06:39:08 | AlHaz | I could do 3d on it, but aside from playing some BZFlag a g200 is probably more or less useless for 3d |
| 06:39:48 | AlHaz | it gets worse in the business world, you know, the "everything is secret" suits aren't hardly the worst of 'em |
| 06:39:50 | Lenolium | The wheaties talks to the ucb1200? I wonder what it says. |
| 06:39:55 | collord | AlHaz: two generations crusty. But nice framebuffer drivers for 2D |
| 06:40:32 | AlHaz | there's an entire school of thought in business management and especially sales & marketing that you must "win", and you don't "win" unless the other guy gets the shaft |
| 06:40:56 | AlHaz | there are actually business schools teaching people that a good product at a fair price is the wrong thing to do |
| 06:41:06 | Lenolium | AlHaz: My ATI Rage/Mobility Pro on this laptop (Dell 5000e, exactly like bzflag's), works great, you might want to look into one of these boards, they are way cheap. |
| 06:41:34 | AlHaz | Lenolium: Nah |
| 06:42:00 | AlHaz | I just picked up a G200 Marvel to replace the plainjane G200 |
| 06:42:13 | AlHaz | integrated tv tuner and mjpeg codec |
| 06:42:23 | collord | mmmm... mjpeg codec |
| 06:42:58 | collord | ... for when bt chips just won't cut it |
| 06:43:04 | Lenolium | AlHaz: Does the integrated mjpeg codec work in Linux? |
| 06:43:29 | AlHaz | Lenolium: Yup. Zoran publishes specs for their chips :) |
| 06:44:08 | Lenolium | AlHaz: You're kidding me? A hardware codec supported in Linux by something other than reverse engineering??? I gotta get me one of these. |
| 06:44:29 | AlHaz | so it does video overlay just like a Bt8x8, and hardware mjpeg compress / decompress |
| 06:44:42 | AlHaz | and saves me a pci slot |
| 06:44:53 | collord | and won't drop frames |
| 06:45:03 | AlHaz | yeah, no more dropped frames |
| 06:46:15 | AlHaz | Lenolium: they did have to reverse engineer some of the matrox controll registers . . . but the codec itself had a driver already, it's more or less the same thing as on the iomega buz |
| 06:46:58 | collord | and some of it came from teh LM33 project, right? |
| 06:47:06 | AlHaz | collord: yeah |
| 06:47:16 | AlHaz | all three are zoran mjpeg |
| 06:47:32 | AlHaz | the buz just has an awful pci-pci bridge chip on it |
| 06:48:06 | AlHaz | anyway, I'm gonna go to bed |
| 06:48:38 | collord | 'night |
| 07:53:40 | GoRK- | anyone still up? |
| 07:53:52 | GoRK- | hrm i should use the user that I'm already logged on with :) brb |
| 07:54:34 | Lenolium | GoRK: I'm up |
| 07:54:34 | GoRK | heh |
| 07:54:35 | GoRK | ok |
| 07:54:44 | Lenolium | I am playing around with buzby |
| 07:54:59 | GoRK | Lenolium: cool i wanted to talk to you actually... i was going to start writing a daemon to make the phone work as a phone.. i didnt know if you had started |
| 07:55:02 | Lenolium | Wanted to get my tuxscreen all ready to go, so tommorrow I could start on the daemon. |
| 07:55:24 | Lenolium | Actually, I was going to write the daemon, and have a client listener do the phone-enabling. |
| 07:55:33 | GoRK | Lenolium: i already have written a .h file for inclusion in the kernel patch and was going to start work on the major functionality |
| 07:55:55 | GoRK | Lenolium: ahh... i was going to write a daemon that did the whole thing but allowed behavior modifications from clients |
| 07:56:12 | GoRK | the whole thing being phone functionality |
| 07:56:41 | Lenolium | GoRK: Remember, there is much more to that DSP than the basic phone functionality. Us VoIP people are hoping for much more. :) |
| 07:57:10 | GoRK | clients could attach and ask for callbacks or clients could attach and take responsibility for controlling the operation or actions of events.. eg. what happens when a digit is pressed on the keypad |
| 07:57:20 | GoRK | Lenolium: i am voip people :) |
| 07:57:47 | Lenolium | GoRK: Geez. That sounds basically identical to what I was planning to do, but I was trying to decide if I should build in default functions or not. |
| 07:58:24 | GoRK | Lenolium: I figured that one of those 7 extra digital swtiches present in the LED controls can control some relays taht connect the audio interfaces (ie. handset speaker and mic) to the ucb1200 audio (or maybe even telecom) interfaces |
| 07:58:54 | GoRK | Lenolium: basically we'd use the extra LED control lines to switch relays to do audio routing to and for the audio interfaces. |
| 07:59:38 | GoRK | so in theory we could use the dsp LED interface to switch ucb1200 audio output from the internal speaker to an external line-out jack or somethign like that |
| 08:00:50 | GoRK | my girlfriend's roommate has a lucent answering machine that I would bet lots and lots of money on having a wheates IC in it.... |
| 08:01:08 | GoRK | and it has a LED panel that displays the number of messages on the machine |
| 08:01:40 | GoRK | last night i could not account for so many unused LED control lines, but if you are trying to control a single-digit numeric led display, it makes perfect sense |
| 08:02:38 | Lenolium | Yup, that sounds about correct. Do we have a diagram of that area of the board yet? do we know if the ucb1200 is even remotely connected to the DSP? |
| 08:03:35 | GoRK | i may have gotten gpsfan psyched up enough about it to do a ucb1200<->mongoose<->telephone hardware |
| 08:03:39 | GoRK | schematic |
| 08:04:06 | GoRK | i am 99.999999% sure the ucb1200 is not currently connected to the dsp |
| 08:04:21 | GoRK | at least as far as audio goes |
| 08:04:30 | GoRK | i would ****love**** to be proven wrong |
| 08:05:31 | GoRK | the good news is that i think i can wire it up :) |
| 08:05:57 | GoRK | ugh i need to get home.. im still over at a friends house and it's 5am |
| 08:06:05 | Lenolium | Hehe. |
| 08:06:06 | GoRK | err 4am anyway :) read my watch wrong |
| 08:06:23 | GoRK | im partied out :) |
| 08:06:33 | GoRK | anyway bbiab.. bout 5 minutes |
| 08:06:35 | Lenolium | COMMAND: DSP_LOAD_MSG_DATA_CMD 0x17 |
| 08:06:39 | GoRK | oh yeah |
| 08:06:45 | Lenolium | Description: Block of encoded speech to ble laoded into DSP RCELP coder |
| 08:07:07 | GoRK | the dsp has the capabilities to record and playback audio to/from the 512kb flash that is onboard the mongoose chip |
| 08:07:20 | GoRK | i am 99.99999% positive of that now, too |
| 08:07:41 | GoRK | it is the only sensible thing |
| 08:07:45 | Lenolium | I wonder if we can load random data also... |
| 08:09:07 | GoRK | yes. you can write to the flash from the current dsp interface. it is *horribly* slow but totally possible. It should be possibel to download custom audio files to the dsp's flash and make the dsp play them... custom rings, answering machine, etc |
| 08:09:48 | GoRK | the 512k flash doesnt seem to hold anything pertinent to the operation of the dsp itself either (at least any address that is not reserved) |
| 08:10:09 | GoRK | so it seems to be used totally for answering system, callerid, etc data |
| 08:11:20 | GoRK | it would make sense since lucent could use the same IC in devices such as answering machiens that were designed with a PIC or something rather than a big fat strongarm processor as a controller |
| 08:11:37 | GoRK | ok well ill talk to you again in a bit. i am heading home now |
| 08:11:39 | Lenolium | GoRK: I'm gonna write my daemon tommorrow, if yours is better than mine, well pfffft, you win. :) |
| 08:12:01 | GoRK | Lenolium: well let's at least use the same header file.... ill send it to you when i get home |
| 08:12:16 | GoRK | Lenolium: talk to you in a few minutes... |
| 08:12:22 | Lenolium | GoRK: rawb@rawb.org |
| 08:12:28 | Lenolium | GoRK: I'm going to bed. :) |
| 08:12:36 | GoRK | Lenolium: ok. well catch you tomorrow, probably |
| 08:12:40 | GoRK | later |
| 08:18:36 | Lenolium | *plong* |
| 08:28:28 | GoRK | ok |
| 10:57:48 | scanline | is away: sleep |
| 13:39:19 | SpaceCoaster | GPSFan: you awake today? |
| 13:44:49 | GPSFan | SpaceCoaster: good morning! |
| 13:45:08 | GPSFan | Just put up a wiki page. |
| 13:45:25 | SpaceCoaster | how are you this fine and sunny day (in NH at least) |
| 13:45:50 | GPSFan | It's cloudy and cool on the Western Slope of Colorado. |
| 13:46:38 | SpaceCoaster | It 60 degrees here, Colorado another nice part of the US |
| 13:47:20 | GPSFan | Yep, 42 this am, not too much to do outside today but feed the critters, and build some fences. |
| 13:49:23 | SpaceCoaster | sounds good, I have a hardware type question are you ready? |
| 13:49:38 | GPSFan | Sure, fire away! |
| 13:50:26 | SpaceCoaster | ok, it's to do with resetting the wheaties module, I think this it is really just the DSP reset that matters. |
| 13:50:40 | SpaceCoaster | ^this^that |
| 13:50:58 | SpaceCoaster | The code looks like this... |
| 13:51:22 | SpaceCoaster | GPCR = DSP_RESET; |
| 13:51:22 | SpaceCoaster | udelay(200); |
| 13:51:22 | SpaceCoaster | GPSR = DSP_RESET; |
| 13:51:34 | SpaceCoaster | with better indents :-) |
| 13:51:58 | GPSFan | SpaceCoaster: that would be GPIO17 I believe. |
| 13:52:28 | SpaceCoaster | indeed GPIO17, now the spec says nothing about the delay, but take it out and no dice |
| 13:52:47 | SpaceCoaster | thats dsp1650.pdf |
| 13:53:40 | GPSFan | from the Mongoose schematic, it looks like the only place that GPIO17 goes is the DSP chip, so I'd look at that spec and see if there are any good wors regarding reset timing (pulse width) or how lon it takes for the dsp to init after a reset. |
| 13:54:10 | GPSFan | I'll take a peek and see if I see anything obvious. |
| 13:54:29 | SpaceCoaster | well, if we send it anything within 800ms, after reset, it pukes |
| 13:54:43 | SpaceCoaster | that is milliseconds! |
| 13:55:44 | GPSFan | Wow, i wouldn't think it would take that long to init, what about reset pulse width? |
| 13:56:21 | SpaceCoaster | errr, isn't that the udelay(200), i.e. 200 microseconds? |
| 13:58:13 | GPSFan | could be, I don't have the code to look at, and that's prpbably a call to a subroutine thet twiddles the bits. |
| 13:58:36 | GPSFan | but 200us seems a bit long for a reset pulse. |
| 13:59:15 | SpaceCoaster | it was a guess, the original code used 500000us so it's better :-) |
| 13:59:45 | SpaceCoaster | do you speak linux device drivers? |
| 13:59:46 | GPSFan | wati 1 while I look at the dsp .pdf.... |
| 13:59:55 | GPSFan | no, |
| 14:00:05 | SpaceCoaster | wanna learn? |
| 14:01:30 | SpaceCoaster | ok page 105 of dsp1650.pdf has a diagram of reset waveforms |
| 14:02:21 | GPSFan | that's where I' |
| 14:02:28 | GPSFan | m looking |
| 14:03:36 | GPSFan | looks like 150ns minimum, no max given so 200us should be ok. |
| 14:04:19 | GPSFan | again, that's just to reset the chip, the DSP then starts its program, which may do lots of stuff. |
| 14:04:42 | GPSFan | memory test, program checksums, stack init, etc. |
| 14:05:02 | SpaceCoaster | cool, so 200us it is! |
| 14:05:25 | GPSFan | there is also a watchdog timer that may have to be tickled every so often. |
| 14:06:16 | SpaceCoaster | yeah I think the 800ms is a program init thing, I sleep() instead of delay, which lets the kernel do other things, so it's not an issue. |
| 14:07:10 | GPSFan | I saw you & GoRK talking about watchdogs last night, what were you sending the DSP? |
| 14:07:20 | SpaceCoaster | hmm, GoRK determined experimentally that he needed to send it a command every so often, I found it to be 4 seconds max. |
| 14:08:11 | SpaceCoaster | it was an 0xB7, a sanity command, returns some bytes saying "please reset me" or "i'm ok" |
| 14:09:26 | GPSFan | that sounds like it tickles the watchdog, where are you getting the api for this? |
| 14:10:41 | SpaceCoaster | have you seen the additional docs under http://tuxscreen.net/download/??? |
| 14:10:52 | GPSFan | Page 62 of the dsp spec says there are 3 watchdog reset periods, 2^12, 2^14, 2^16 clock tics. |
| 14:11:29 | SpaceCoaster | i see |
| 14:11:50 | SpaceCoaster | do you know what our clock is? |
| 14:11:52 | GPSFan | Hmmmm, let me look, I've been busy with the CPU schematics, and only a passing interest in DSP has kept me following what you are doing, let me look... |
| 14:12:05 | SpaceCoaster | good man! |
| 14:13:04 | GPSFan | I see the dsp spec, which I have, but what else is there re the dsp?? |
| 14:13:09 | SpaceCoaster | hhm those are watchdog clock cycles, which is CLKIN/128 |
| 14:13:38 | SpaceCoaster | there are some docs in howdidthisgethere |
| 14:14:44 | GPSFan | Ah! redily evident, I'll look them over, and yes those are CLKIN/128 cycles. |
| 14:15:07 | SpaceCoaster | what would be your ballpark estimate for the clock? |
| 14:18:49 | SpaceCoaster | 2^16 divides it by 8M, 2^14 divides it by 2M, 2^12 divides it by 500k |
| 14:19:07 | SpaceCoaster | s/divides/multiplies/ |
| 14:19:30 | GPSFan | the Mongoose schematic says 4.??? MHz for the xtal, but there is an intternal PLL that multiplies it up. But the block diagram says that CLKIN is the same freq as the xtal. |
| 14:19:58 | GPSFan | so 4Mhz?? |
| 14:20:28 | SpaceCoaster | hmm my ALU has turned off, help:-) |
| 14:20:57 | SpaceCoaster | 8M cycles at 4MHz is 2 seconds? |
| 14:21:20 | GPSFan | yep, |
| 14:22:31 | SpaceCoaster | the DSP did seem to be happy with 4seconds between data tho' |
| 14:24:20 | GPSFan | SpaceCoaster: I'd say stick with the sanity message at no less than 4 seconds, untill things become clearer. I've got to bail out for a while, hungry critters to feed, and work to do before winter comes, |
| 14:24:56 | SpaceCoaster | thanks, this is all starting to make sense |
| 14:25:02 | GPSFan_away | good! |
| 14:38:54 | SpaceCoaster | GPSFan_away: Page 107 talks about a Powerup Reset Pulse of 120ms, I assume that this isn't for us, we reset about 10seconds after power up. |
| 14:42:47 | SpaceCoaster | is away feeding the chipmunks |
| 15:24:34 | kram | has returned. |
| 15:52:29 | SpaceCoaster | has returned |
| 16:12:56 | mallum | yo |
| 16:14:20 | AlHaz | y0 |
| 16:16:49 | mallum | Alehaz: got the Cats to boot |
| 16:17:11 | SpaceCoaster | congrats mallum |
| 16:17:11 | AlHaz | Cool |
| 16:17:16 | AlHaz | how are you booting it? |
| 16:17:24 | ctr | ok |
| 16:17:25 | ctr | bbiab |
| 16:17:26 | ctr | :) |
| 16:17:31 | mallum | Alehaz: I managed to eventually get it to boot off the tftp image from the hardisk |
| 16:17:46 | mallum | Alehaz: after buying another network card that it failed to recognise |
| 16:18:04 | ctr | is away: offLicence->get( new Wine( Wine.RED ) ); room->tidy(); gf->arrive(); |
| 16:18:17 | AlHaz | mallum: ugh, that's disappointing about the nic |
| 16:18:25 | mallum | gf->mount() |
| 16:18:33 | ctr | ehehe |
| 16:18:34 | ctr | :) |
| 16:18:47 | mallum | Alehaz: yeah |
| 16:19:41 | mallum | Alehaz: the firmware also doesn't seem to like links, it wouldn't boot /vmlinuz ( which was linked to /boot/vmlinuz.... ) not boot it from a full path |
| 16:20:05 | mallum | Alehaz: I had to copy it the root over vmlinuz and that seemed ok |
| 16:20:29 | mallum | Alehaz: hopefully I can upgrade to woody and get a 2.4 kernel on it |
| 16:20:46 | mallum | happy now, but what a royal pain in the arse |
| 16:23:30 | AlHaz | bringing up a strange board the first time is always a pain |
| 16:23:46 | AlHaz | at least it wasn't a motorola product, in which case it would have arrived with a blank flash simm |
| 16:23:55 | AlHaz | a $700 debugger |
| 16:23:58 | AlHaz | and a copy of CodeWarrior |
| 16:25:38 | AlHaz | for Windows |
| 16:27:54 | ctr | is back (gone 00:09:49) |
| 16:33:13 | mallum | Alehaz: heh, a phone ? |
| 16:34:00 | AlHaz | nah, nearly all motorola development boards are shipped that way |
| 16:34:05 | AlHaz | no code running on them at all |
| 16:34:23 | AlHaz | so you absolutly positively must buy their annointed jtag or cop or bdm product |
| 16:34:39 | AlHaz | and use their provided flash program, which happens to be a code warrior project |
| 16:34:56 | AlHaz | Some manager at motorola thinks that forcing every customer to install CW is a Good Thing |
| 16:43:19 | mallum | AlHaz: hmm, not good |
| 16:43:49 | AlHaz | and more often than not the protocol the hardware debugger speaks is proprietary and secret |
| 16:44:30 | mallum | AlHaz: is this at lineo, you still there ? |
| 16:44:43 | AlHaz | mallum: Yeah, I'm still there. It's Motorola that's doing that, tho. |
| 16:45:05 | AlHaz | Motorola owns a great deal of lineo stock |
| 16:45:23 | AlHaz | and we try to support a lot of motorola products |
| 16:45:29 | AlHaz | and that means i get to test them |
| 16:46:08 | AlHaz | it's difficult to market linux development kits for motorola powerpc boards because they arrive dead to the world . . . and there's often no existing way to get a boot loader onto the board from anything but windows |
| 16:46:43 | AlHaz | code warrior just adds insult to injury |
| 16:48:20 | wej | mornin |
| 16:48:22 | wej | evenin |
| 16:50:09 | wej | has anyone here tried any 802.11 cf cards? |
| 16:50:35 | wej | i've tried a standard linksys pcmcia card, but it crashes the tuxscreen.. i've been working on it for a while but to no avail... |
| 16:50:43 | wej | i'm on my way to fry's to be the guinea pig. |
| 16:50:53 | Alehaz | wej: see if you can compile wlan-ng for arm |
| 16:51:08 | wej | i tried, alehaz. |
| 16:51:11 | wej | should i revisit that? |
| 16:51:11 | Alehaz | wej: it's orinoco_cs that crashes, and it'll crash with everything |
| 16:51:17 | wej | orinoco crashes.. |
| 16:51:21 | wej | i've been using wvlan_cs tho |
| 16:51:24 | wej | and it crashes as well. |
| 16:51:25 | Alehaz | hmm |
| 16:51:57 | Alehaz | same sort of error message? |
| 16:53:46 | wej | yeah |
| 16:53:48 | wej | it's an oops. |
| 16:54:00 | Alehaz | there's arm-specific code in wlanng - you could join the mailing list and ask how it's done |
| 16:54:11 | Alehaz | I've come close to compiling it but not succeded |
| 16:54:24 | Alehaz | but I lack motivation because there's a hub three feet from my tux |
| 16:56:30 | mallum | ok cinema for me now |
| 16:56:33 | mallum | bye all |
| 16:59:23 | wej | heh |
| 16:59:29 | wej | i'm working omcpiling it now |
| 16:59:31 | wej | man. it's a mess. |
| 16:59:53 | RobFlynn | is away: Shower... |
| 17:11:44 | sb-- | i just read about this thing on slashdot.. what all can it do? |
| 17:13:43 | SpaceCoaster | sb--: recent slashdot or a month ago? |
| 17:15:17 | sb-- | i just read a recent comment on /. that led me to the old article, so both |
| 17:16:02 | SpaceCoaster | cool, at the moment it is really a developers toy |
| 17:16:24 | collord | Alehaz: i got that bloody dwl-650 up-- needed the tosh cardbus option and pcmcia-cs drivers |
| 17:17:01 | sb-- | spacecoaster: well what can it run? |
| 17:17:18 | sb-- | i'm assuming you can squeeze linux on the 4mb flash rom, right? |
| 17:18:17 | Alehaz | collord: good to hear it's working |
| 17:18:38 | Alehaz | sb--: we don't just squeeze, we've got a few megs free |
| 17:18:47 | SpaceCoaster | sb--: indeed, it fits easy |
| 17:18:51 | sb-- | alehaz: okay :) well can you tell me what all it can do? |
| 17:18:53 | sb-- | i'm debating getting one |
| 17:18:57 | sb-- | i'd like a new toy to play w/ |
| 17:19:02 | sb-- | but i wanna know what it's capable of |
| 17:19:10 | collord | wej: are you sure the linksys will work with 3.3V? |
| 17:19:17 | Alehaz | well, what fits on the flash, you can have a simple microwindows gui, etc |
| 17:19:55 | SpaceCoaster | sb--: it has a 640x480x8 display with a touchscreen so you can develop very pretty apps. |
| 17:20:07 | Alehaz | collord: we're not sure at all. When you plug them into the phone, "cardctl ident" shows the card is sane enough to say who it is |
| 17:20:30 | sb-- | any networking capabilities? |
| 17:20:31 | Alehaz | collord: so we think it will probably work |
| 17:20:38 | Alehaz | collord: at this point we lack a working driver |
| 17:20:49 | Alehaz | sb--: sure, pcmcia slots w/ network cards in 'em works fine |
| 17:20:52 | sb-- | is the touchscreen active or passive? |
| 17:21:05 | sb-- | can it run X? |
| 17:21:11 | sb-- | so can you network it and run X apps on ti? |
| 17:21:20 | SpaceCoaster | sb--: it has pcmcia so it can be networked, only 3v cards work which means you have to be careful. |
| 17:21:21 | RobFlynn | is back (gone 00:21:28) |
| 17:21:35 | Alehaz | I've never understood what the difference is supposed to be wrt active and passive touchscreens. It's a normal 4 wire resistive touchscreen |
| 17:21:47 | sb-- | i meant active or passive matrix :) |
| 17:21:57 | Alehaz | the screen is DSTN |
| 17:22:06 | sb-- | ahh, okay |
| 17:22:07 | Alehaz | lcd is, i mean |
| 17:22:08 | SpaceCoaster | X works using the Xfbdev from the familiar distribution, very usable, quick |
| 17:22:09 | sb-- | not active |
| 17:22:58 | SpaceCoaster | the screen is fine, this is for use indoors, so laptop, pda issues don't apply |
| 17:23:24 | SpaceCoaster | it wouldn't look good in bright sunshine |
| 17:24:05 | sb-- | how exactly do you install linux on it? i mean, how do you move files onto it? |
| 17:24:30 | SpaceCoaster | downloaded as images over the serial port, see the wiki for details |
| 17:24:46 | sb-- | hrm.. sounds cool |
| 17:24:49 | SpaceCoaster | http://tuxscreen.net/ |
| 17:25:38 | SpaceCoaster | the phone doesn't integrate well with Linux yet but this is rapidly approaching and we are making progress daily |
| 17:25:43 | sb-- | digital answering machine software? :) |
| 17:26:13 | Alehaz | not yet |
| 17:26:25 | Alehaz | and we're not 100% sure we can do that |
| 17:26:29 | SpaceCoaster | hmmm, maybe Lucent talked of having to modify it to do VOIP and answer machine is similar functionality |
| 17:26:31 | GPSFan | SpaceCoaster: The reset data on page 107 (120ms) refers to the built-in powerfail reset timing set by the RC on RSTB pin 99. |
| 17:27:05 | SpaceCoaster | GPSFan: ooops, quick context switch, is that good? |
| 17:27:08 | sb-- | from the picture it looks like the screen can swivel all around? or maybe fold down? |
| 17:27:48 | SpaceCoaster | it folds flat and turns about 30 degrees either side of center, not much |
| 17:27:50 | Alehaz | it folds flat against the phone and almost vertical, and it swivles a few degrees from side to side |
| 17:28:18 | SpaceCoaster | Alehaz: we are sounding like echoes, I'm gonna eat lunch... |
| 17:28:19 | sb-- | so when it's flat, is this thing pretty small? |
| 17:28:31 | sb-- | anyone got a picture of theirs in a room so i can get some idea of the size? |
| 17:28:47 | GPSFan | SpaceCoaster: It isn't good or bad, if the powerfail ckts are enabled that's the length reset you will get on powerfail recovery. And the Mongoose doesn't have the recomended RC on pin 99. |
| 17:29:13 | GPSFan | SpaceCoaster: BTW the RSTB pin on the DSP is bi-directional. |
| 17:31:08 | SpaceCoaster | GPSFan: sorry I don't understand what this means. |
| 17:31:28 | Alehaz | damn, I was gonna go eat lunch . . . |
| 17:31:28 | SpaceCoaster | is a software hacker |
| 17:32:03 | sb-- | anyone got a pic of their tuxscreen online? |
| 17:32:18 | Alehaz | there are many pics at www.tuxscreen.net |
| 17:34:31 | sb-- | i found pictures, but none of it in a room where i can see the size :) |
| 17:34:42 | sb-- | tons of pictures of boxes in garages and its insides |
| 17:34:57 | collord | sb--: imagine that the size of the phone receiver is about the distance from mouth to ear... |
| 17:35:20 | SpaceCoaster | it's 12x9x3.5 with the screen flat |
| 17:35:43 | SpaceCoaster | 12x9x9 with the screen raised |
| 17:36:01 | sb-- | cool, thanks |
| 17:36:14 | sb-- | does the infrared keyboard have good range? |
| 17:36:50 | SpaceCoaster | never tried more that 6 inches |
| 17:37:06 | SpaceCoaster | goes to eat |
| 17:48:42 | SpaceCoaster | sb--: ok, I have a 16ft tape measure it works at that distance |
| 17:49:23 | SpaceCoaster | sb--: though you can't see the screen, except with binoculars :-) |
| 17:49:36 | Disconnect | question.. if I got a tuxscreen next week or so, and the crucial 32M sodimm referenced from the wiki, what are the chances of ending up with a 40M linux tuxscreen (after some amount of non-soldering-type work ;) ..) |
| 17:50:38 | SpaceCoaster | inferno works with 40mb so soldering shouldn't be neccesary, just programming |
| 17:52:05 | Disconnect | ok. was looking at http://tuxscreen.net/wiki/view/Main/SODIMM and there was some confusion (mainly cuz I haven't really looked at the tuxscreen before now, I suspect) |
| 17:52:35 | Disconnect | thanks :) |
| 17:52:55 | Disconnect | is off to read more about the linux-on-tuxscreen before asking more questions (since they are prolly already in the wiki) |
| 17:53:04 | SpaceCoaster | but GPSFlag has discovered some aspects of the way the memory is wired which may impact the software fix |
| 17:53:43 | SpaceCoaster | hmm, do i detect another slashdotting... |
| 17:56:02 | Disconnect | hrm. that doesn't sound promising. eventual goal was 128M or so (as ram prices drop) but for our short-term needs 40m is great.. (16M isn't - for that, we've got ipaqs ;) ..) |
| 17:56:10 | sb-- | woa, you can add a 32M sodimm? |
| 17:56:18 | sb-- | how much does that cost? |
| 17:56:39 | SpaceCoaster | yup but linux doesn't see it all at the moment, just 8mb |
| 17:56:45 | Disconnect | yah. :( |
| 17:57:13 | Disconnect | sb--: see the wiki link above - its $43.19 unless you are a new customer ($36.71 for first order, if I'm reading the promo correctly) |
| 17:57:17 | SpaceCoaster | but a bright dude with a soldering iron and a schematic can fix anything, right! |
| 17:57:33 | sb-- | heh, well i got a soldering iron :) |
| 17:57:37 | Disconnect | SpaceCoaster: anyone figure out why? |
| 17:57:54 | Disconnect | has a soldering iron and knowledge enough to know its best if it stays unplugged ;) ..) |
| 17:58:37 | prpplague | SpaceCoaster: whats this about soldering? someone got a hardwired ram solution? |
| 17:58:43 | SpaceCoaster | GPSFan has figured out the wiring and others have hypothesised on a boot loader issue but no firm solutions |
| 17:59:14 | SpaceCoaster | prpplague: not a solution, GPSFan has the details |
| 17:59:37 | prpplague | ohh |
| 18:00:27 | GPSFan | SpaceCoaster: I see you are back from lunch, to answer your last question, the RSTB pin on the dsp will accept a reset from the CPU OR it will pull itself low if the powerfail ckts are enabled AND a powerfail happens. However, we don't know what is between the RSTB pin on the DSP and the GPIO17 |
| 18:00:50 | SpaceCoaster | ahh I understand |
| 18:01:00 | Disconnect | more general question.. what does a tuxscreen do "out of the box" other than being a rather large non-cordless phone? ;) |
| 18:01:12 | prpplague | can't seem to escape the x86 arch fot the arm arch |
| 18:02:55 | SpaceCoaster | the inferno software is not too bad, address book, notepad, i think it has a web browser, via an isp, that sort of thing |
| 18:03:10 | Disconnect | prpplague: depends on what you are doing. (which is why it takes me 2 suitcases to haul my haulables ;) .. ipaq, libretto, 2 laptops..) |
| 18:03:41 | Disconnect | its got pcmcia - can I slap a wireless card into it? (and what about mp3 or even just speakerphone?) |
| 18:03:48 | sb-- | libretto's are awesome |
| 18:04:28 | Disconnect | really wants one. But I gotta have at least some vague reason for the fiance or it'll just sit up here. If it does something -she- wants, it goes somewhere useful like the living room, where we want an mp3 player/webpad thing ;) ..) |
| 18:05:17 | Disconnect | sb--: sorta. 20G drive makes a big difference. the keyboard sucks for real work (great for the don't-haul-a-laptop-can't-use-an-ipaq case tho) and the headphone jack is mono :( |
| 18:05:40 | Disconnect | sb--: biggest flaw in this one tho is that it doesn't charge while running. :( |
| 18:06:02 | sb-- | sounds like you're even more of a hardware junkie than i am :) |
| 18:06:27 | Disconnect | heh. |
| 18:06:27 | prpplague | Disconnect: i'm in R&D and we have a ton x86 based stuff i'd love to dump and work on arm solutions, but everytime i start working on arm they pull me off on some x86 based stuff |
| 18:07:17 | Lenolium | *pling* |
| 18:07:50 | Disconnect | collect stuff at auction (ebay mostly) .. one of the 2 laptops belongs to work (and I'm actually taking 4 on this trip tomorrow, cuz 2 of them have the new win2k images I'll need.. upgrading rental laptops to a new image.. the image is on the server already, but I feel better having it w/ me as well ;) ..) |
| 18:08:49 | sb-- | disconnect: yah, i frequent ebay as well. i've been eyeing the libretto/picturebook mini-laptop type computer, but haven't had the need to justify the $$$ yet, but boy do i want one :) |
| 18:08:51 | Disconnect | prpplague: although the 2 U5s my work was throwing away a few jobs ago. and the mac (7100/80) was $100 w/ 21" monitor .. (the problem is I then throw lots of money into this stuff, which kinda defeats the purpose of getting it cheap - 20g drive for libretto, G3 upgrade for mac, etc) |
| 18:08:55 | sb-- | or a loox |
| 18:09:40 | Disconnect | wouldn't spend $600 on a used g3, but he spent $100 on a used 7100, then $500-$600 upgrading it - new drive, g3 board, ram, etc. go figure. |
| 18:10:41 | Disconnect | whats killing me now (and its good I'm on an embedded kick ;) ..) is power. bought a house, and 3/4 of it is on one 20A circuit. so only one desktop is running, even with kvm I can't use the others w/o popping the circuit (hell, it goes on its own sometimes even with the lights out etc) |
| 18:10:53 | prpplague | if it has an arm core, i've been buying it |
| 18:10:58 | Disconnect | heh |
| 18:11:27 | prpplague | i've got about 8k left in the budget time jan1 |
| 18:11:55 | Disconnect | prpplague: lemme highly recommend ipaq 3100 for embedded arm. right now we're debating between that (w/ $99 64M ram upgrade) and the tuxscreen ($30 40meg ram upgrade) .. but this is all personal stuff, no budget except what I can afford ;) |
| 18:12:23 | SpaceCoaster | prpplague: wow a man with a budget, will you be my friend? |
| 18:12:23 | Tangent | Evening all |
| 18:12:43 | Disconnect | prpplague: the current project is running the debian 'buildd' on an armv4l box, so that intimate (debian on ipaq) can get armv4l stuff instead of the current debian-arm (armv3l) .. from what I understand, its a big jump. |
| 18:12:49 | prpplague | SpaceCoaster: only if your willing to trade info/knowledge |
| 18:12:52 | Disconnect | ..speaking of intimate... |
| 18:12:54 | Disconnect | hiya Tangent |
| 18:13:09 | Disconnect | prpplague: heh. you hiring? ;) |
| 18:13:26 | SpaceCoaster | prpplague: wanna buy a working DSP? |
| 18:13:26 | Disconnect | Tangent: btw gave ibot some messages for you. |
| 18:13:39 | Tangent | Disconnect : Yeah.. just read them |
| 18:13:42 | prpplague | Disconnect: ya but probably not for what you guys are used to making... |
| 18:14:02 | Disconnect | prpplague: heh. depends. most of what I do IRL is sys admin and/or sys admin manager. (I can herd cats. ;) ..) |
| 18:14:13 | Tangent | Disconnect : If we can make crosscompiling buildd work, then I'm all for it |
| 18:14:52 | SpaceCoaster | prpplague: are the other arm cores as neat as the StrongArm? |
| 18:14:56 | prpplague | Disconnect: our big thing is we just installed cisco ip phones, and our "winblows sysadmin" has no idea how to manage anything |
| 18:15:38 | Disconnect | prpplague: I grok cisco. (not at the CCIE level, but I manage networks as well as *nix boxen ;) ..) |
| 18:16:02 | Alehaz | cisco ip phones aren't really cisco |
| 18:16:18 | prpplague | SpaceCoaster: i'm not a guru, but from what i've seen, strongarm seems to gear more towards a real os, rather than dedicated oprations |
| 18:16:19 | Alehaz | they bought a company so they could slap their name on the product |
| 18:16:22 | Disconnect | Alehaz: half of cisco isn't really cisco ;) |
| 18:16:39 | Alehaz | Disconnect: true |
| 18:16:42 | prpplague | Alehaz: vovida? |
| 18:17:03 | Alehaz | I had the misfortune of working at a company that was a cisco ip telephony beta site |
| 18:17:12 | Alehaz | prpplague: I forget, I think it started with an X |
| 18:17:14 | Alehaz | er, S |
| 18:18:12 | prpplague | i like the phones, but the winblows back end sucks |
| 18:18:20 | Alehaz | yup |
| 18:18:44 | Alehaz | are they still shipping a rack mount proliant with a cisco faceplate and a DT24X card shoved in it? |
| 18:18:53 | prpplague | yep |
| 18:18:55 | Alehaz | running NT, of course, or have they switched to 2k |
| 18:19:14 | prpplague | 2k professional i believe |
| 18:19:36 | Alehaz | prpplague: the early DT24x cards were made out of far-too-thin circuit board, they kept popping out of the pci slot |
| 18:20:41 | prpplague | Alehaz: we've been down twice due to network viruses, and have about 10 replicatable problems that cisco can't seem to fix |
| 18:20:45 | Alehaz | the early phones were these stupid looking white things with a character lcd display on 'em, with a built in two port 10baseT switch |
| 18:21:02 | Alehaz | prpplague: Sounds like they haven't progressed much since '99 |
| 18:21:43 | Alehaz | prpplague: we found that the only way to make 'em work was to put them on their own fully switched ethernet with their own ip segment |
| 18:21:58 | SpaceCoaster | Alehaz: do you have any pointers for public domain VOIP? |
| 18:21:58 | Alehaz | separated from the rest of the network with an extra port in the router |
| 18:22:24 | Alehaz | SpaceCoaster: I'm not a good source of info on that subject. fooling around with dialpad and then using the cisco ip phones has convinced me that ip telephony is a sow's ear |
| 18:22:39 | prpplague | same here |
| 18:22:48 | Alehaz | prpplague: And I've heard of other installations configured that way, makes it seem sorta pointless |
| 18:23:20 | Alehaz | At that point just use some other digital pbx that doesn't have a seven layer network model to worry about |
| 18:24:17 | prpplague | Alehaz: we've tried both ways and we still have problems |
| 18:24:34 | Lenolium | SpaceCoaster: Starting on the daemon. |
| 18:24:38 | Alehaz | find an industrial equipment auction and buy an old panasonic or toshiba system that someone's ripped out in an upgrade, if it's a small enough company |
| 18:25:07 | SpaceCoaster | Lenolium: you are the man! |
| 18:25:11 | sb- | is there a speaker on this baby? |
| 18:25:21 | sb- | or audio out? |
| 18:25:21 | Alehaz | sb-: hooked up to the phone, yes |
| 18:25:24 | sb- | haha |
| 18:25:38 | Alehaz | sb-: there's built in audio in the sa1100 but it doesn't appear to be hooked up to anything at this point |
| 18:25:58 | sb- | gotcha |
| 18:26:17 | collord | Alehaz: the dlink 650 won't work with toshiba in win98 unless in cardbus mode |
| 18:26:38 | SpaceCoaster | Alehaz: Gork can make it do fancy things using the wheaties driver. |
| 18:26:49 | Alehaz | collord: weird |
| 18:27:00 | Alehaz | SpaceCoaster: the speaker? |
| 18:27:13 | Alehaz | SpaceCoaster: but you can't just push PCM digital audio at it |
| 18:27:16 | collord | Alehaz: yeah. it's academic at some point, but i think it's prob. a voltage problem |
| 18:27:42 | collord | Alehaz: so i *don't* expect this to work in tuxscreen... |
| 18:27:57 | SpaceCoaster | Alehaz: no it connects to the line though. |
| 18:27:58 | Alehaz | collord: it sounds like a bios and/or driver issue |
| 18:28:18 | Alehaz | collord: I've used my linksys card in an aincent ISA pcmcia adapter, and it's based on the same chips |
| 18:28:28 | sb- | will a wireless card work in it? |
| 18:28:32 | Lenolium | Alehaz: Hopefully, by the end of tonight, we will be able to dial from the pad, and from the command line. |
| 18:29:01 | Alehaz | sb-: theoretically but no reports of success yet |
| 18:29:14 | collord | Alehaz: yeah my linksys worked just fine in PCIC mode, both win (same cw10.sys driver!) and lin. |
| 18:29:50 | Alehaz | collord: weird |
| 18:30:54 | collord | Alehaz: perhaps that PCIC won't do 3.3v whereas cardbus will... so will be interesting to see tuxscreen's 3.3v pcmcia |
| 18:31:11 | Alehaz | collord: sounds odd, since my portege has the same cardbus bridge . . . |
| 18:31:34 | collord | Alehaz: cat /proc/pci... do you see the ToPIC95? |
| 18:31:54 | Alehaz | yep |
| 18:32:22 | Alehaz | Bus 0, device 19, function 0: |
| 18:32:22 | Alehaz | CardBus bridge: Toshiba ToPIC95 (rev 7). |
| 18:32:22 | Alehaz | Slow devsel. Fast back-to-back capable. Master Capable. No bursts. Min Gnt=128.Max Lat=4. |
| 18:32:22 | Alehaz | Bus 0, device 19, function 1: |
| 18:32:22 | Alehaz | CardBus bridge: Toshiba ToPIC95 (rev 7). |
| 18:32:23 | Alehaz | Slow devsel. Fast back-to-back capable. Master Capable. No bursts. Min Gnt=128.Max Lat=4. |
| 18:32:46 | collord | Alehaz: then you're in Cardbus/16 mode in the bios... and the pcmcia-cs will support that, but kernel-tree barfs |
| 18:33:04 | Alehaz | collord: I see it in proc/pci regardless of that setting in bios |
| 18:33:26 | collord | Alehaz: you sure? mine went away (perhaps resource conflict then?) |
| 18:33:36 | Alehaz | collord: pretty sure |
| 18:34:14 | Alehaz | I can check next time i reboot |
| 18:34:20 | Alehaz | but that happens rarely at best |
| 18:34:48 | prpplague | has anyone experience "Illegal instruction" errors on the tux-build? |
| 18:35:10 | collord | Alehaz: it's academic-- I'm resolved that the dlink is wierd. i just needed three boxes to help sniff out problems with the prism2 drivers... |
| 18:35:15 | Alehaz | prpplague: don't recall, where at? |
| 18:35:34 | prpplague | with the rz, sz and df commands? |
| 18:36:11 | Alehaz | hmm, no, they worked ok |
| 18:36:34 | prpplague | ok thanks |
| 19:22:26 | sb- | hrm.. |
| 19:22:31 | sb- | cant install openbsd on my 486 :/ |
| 19:22:34 | sb- | piece of sheyat |
| 19:25:10 | prpplague | does anyone else find that you really need around 36 hours per day to get everything done? |
| 19:26:55 | other | i was thinking more. |
| 19:27:07 | collord | any particular deadline? |
| 19:28:35 | prpplague | collord: just seems little hard to get everything in - computers,beer,kids,wife,bills,computers.... |
| 19:31:06 | Tangent | prpplague : For the last couple of months, I've not gone to work, and only slept for around 4-6 hours a night... I'm still busy all the time, and don't think I'll be able to finish my current project before I get a job |
| 19:32:59 | Disconnect | Tangent: ditto here, except I have had the job.. even doing some of it @ work there isn't time. |
| 19:33:17 | prpplague | Tangent: i'm beginning to think my job is interfering with my work |
| 19:33:28 | Disconnect | prpplague: hehehe |
| 19:33:33 | Tangent | prpplague : Yeah.. I know that feeling... |
| 19:33:49 | Tangent | prpplague : Last job I had, both me and my boss worked from home all the time |
| 19:34:08 | Tangent | We used to say that we didn't like going into the office because we couldn't get any work done there |
| 19:34:17 | Tangent | It was totally true |
| 19:36:30 | collord | prpplague: and then you'll get your ceiva... |
| 19:38:11 | collord | prpplague: sorry, i'm obsessed with those. |
| 19:38:15 | prpplague | collord: that too |
| 19:38:16 | as_ | Evnin all. |
| 19:38:29 | mallum | Tangent: got my Cats board up, booting and networked :) |
| 19:38:30 | Tangent | as_ : Evening |
| 19:38:46 | as_ | Your on early Tangent ;] |
| 19:38:49 | prpplague | collord: me too, i'm in search of a cool arm hack |
| 19:38:51 | Tangent | mallum : Fantastic :).. How cool is it? |
| 19:38:56 | Tangent | as_ : I was going to say the same to you |
| 19:39:01 | as_ | Haha. |
| 19:39:26 | mallum | Tangent: well it was a pain in the arse to get working, but I now have a 256meg, 200mhz arm compile box |
| 19:39:50 | Tangent | mallum : Very handy thing to have |
| 19:40:01 | Tangent | mallum : Once I get a job, I may have to get one |
| 19:40:19 | Tangent | Tho' I may force myself to wait until I've played more with some other toys |
| 19:40:24 | mallum | Tangent: well it only cost me just over 300 quid |
| 19:40:36 | mallum | Tangent: I had most of the parts lying around |
| 19:40:46 | mallum | Tangent: not sure it will run a 2.4 kernel though |
| 19:41:16 | Tangent | mallum : I have plenty parts here to plug into one.. I decided on a spending freeze until I know where I'm gonna get more money from tho' |
| 19:42:35 | wej | anyone tried the Soho 10baseT ethernet card? |
| 19:42:38 | wej | it seems to identify it.. |
| 19:42:41 | wej | and i don't get the Vcc message.. |
| 19:42:46 | wej | so that tells me it's running ok at 3.3v |
| 19:43:04 | wej | ident comes up weird on it tho.. say sit's an NDC something |
| 19:43:33 | wej | SWEET |
| 19:43:36 | wej | it looks like it's working. |
| 19:43:39 | wej | 29 bucks at Fry's |
| 19:44:35 | collord | wej: iirc, sohoware is an ndc brand |
| 19:44:38 | mallum | wej: cool |
| 19:44:46 | wej | very nice. |
| 19:44:51 | collord | wej: so that's expected |
| 19:44:56 | wej | it's hung, but it found the card, ercognized the mac, and brought up the int. |
| 19:44:57 | wej | heheh. |
| 19:45:01 | wej | collord: nice. |
| 19:46:29 | wej | also says it's NE2000 compatible... |
| 19:46:33 | wej | dunno whyu it's hanging tho. |
| 19:47:27 | Alehaz | hung? |
| 19:50:47 | prpplague | Tangent: spending freeze? just do like congress and spend the money, worry about where you're going to get later... |
| 19:51:43 | Tangent | prpplague : Hrm... |
| 19:52:02 | Tangent | prpplague : If only I could print the stuff too ;) |
| 19:52:27 | prpplague | Tangent: thats what ink jet printers are for |
| 19:59:03 | sb- | where can i get a list of 3.3V cards? |
| 20:04:24 | Alehaz | sb-: go look at the wiki |
| 20:06:12 | sb- | at the wiki? |
| 20:09:09 | Alehaz | yes |
| 20:09:17 | Alehaz | go to www.tuxscreen.net and click "wiki" |
| 20:09:20 | Alehaz | it's all there. |
| 20:14:48 | GoRK | hello folks |
| 20:15:45 | prpplague | do you guys normally use two seperate boxes one as a stable box, the other as a dev box? |
| 20:16:00 | GoRK | prpplague: you mean desktop machines? |
| 20:16:11 | prpplague | GoRK: ya |
| 20:17:10 | GoRK | i sort of do... I keep a linux and a windows box side by side.. depends on what im doing with them. mainly i use the linux machine |
| 20:17:28 | GoRK | for both development and regular things |
| 20:18:25 | prpplague | GoRK: i've been doing dev work on my main desktop, but as of late, because i've been messing around so much, i've created problems on my box |
| 20:18:45 | GoRK | prpplague: why, are you not running debian? |
| 20:18:53 | prpplague | i think i'm going to break down and seperate my work onto 3 different boxes |
| 20:19:06 | GoRK | :) |
| 20:19:27 | prpplague | GoRK: i'm running rhat :-| |
| 20:20:24 | sb- | how good is the refresh rate on the screen? |
| 20:20:52 | GoRK | sb-: it's a DSTN screen. it sucks |
| 20:29:41 | wej | anyone compiled ssh against uclibc? |
| 20:30:06 | GoRK | afaik you can't compile ssh against uclibc for some reason but you can compile lsh with some patches |
| 20:30:41 | prpplague | GoRK: doesn't ssh require some c++ support that uclibc doesn't have yet? |
| 20:32:39 | wej | updated the wiki with his soho card info |
| 20:32:51 | wej | prpplague: dunno. sounds likely tho |
| 20:32:58 | GoRK | prpplague: dunno. i just heard it doesnt compile from the ucblibc mailing list |
| 20:40:49 | SpaceCoaster | GoRK: hi ya, taking a break from buzby hacking? |
| 20:40:51 | sb- | GoRK: so i guess the screen isn't good enough to play live video on huh? :) |
| 20:40:55 | wej | how goes it, SC? |
| 20:41:08 | GoRK | sb-: you could... It'd probably be a little blurry |
| 20:41:28 | SpaceCoaster | good, I don't want to look at soho card info, is it good or bad news? |
| 20:41:31 | wej | is building microwindows with one-hand-on-the-keyboard |
| 20:41:42 | wej | SC: soho card works well |
| 20:41:52 | GoRK | sb-: but if you're talking video phone or something... talking head type video should be ok. |
| 20:41:54 | wej | 29 bucks at fry's |
| 20:41:59 | GoRK | wej: did you get a sound card working in the tux? |
| 20:42:08 | wej | GoRK: haven't tried... |
| 20:42:27 | GoRK | wej: ah. what sound card is this then? a pcmcia card? |
| 20:43:26 | wej | the Soho ethernet card |
| 20:43:28 | wej | is what i was talkin about. |
| 20:43:53 | GoRK | hahahahaha i am stupid |
| 20:44:07 | GoRK | i kept seeing soho and reading sound |
| 20:44:58 | wej | heh. |
| 20:45:04 | wej | no worries. |
| 20:45:28 | wej | last night i looked at my wife and said 'pass the salt' and what i really meant to say was 'you ruined my life you wench.' |
| 20:45:49 | GoRK | lol |
| 20:46:25 | Lenolium | SpaceCoaster: Question, do you still reset on every open()? Or, how do I reset it if the sanity check comes back bad. |
| 20:46:47 | GoRK | Lenolium: there is a 2 byte software reset. |
| 20:46:51 | prpplague | wej: rofl |
| 20:47:14 | Lenolium | GoRK: I thought I remembered something like that, but what is it? |
| 20:47:18 | GoRK | SpaceCoaster: hardware reset should probably be implemented in an ioctl. it would alleviate the problems of *having* to reset on open |
| 20:47:30 | SpaceCoaster | Lenolium: yeah every, close and open, I don't flush the read queue so you won't lose anything. |
| 20:48:17 | SpaceCoaster | GoRK: but if it wasn't open the odds are pretty good that you will have to reset. |
| 20:48:59 | GoRK | Lenolium: i didnt include the command in the wheaties.h because on arm it is better to do reset with the io pins... but the software reset command is B9 BB |
| 20:49:24 | GoRK | SpaceCoaster: yeah i guess that's right... still probably need a reset ioctl though .. :) |
| 20:49:45 | SpaceCoaster | GoRK: will do. |
| 20:50:06 | SpaceCoaster | GoRK: have you tried the new driver yet? |
| 20:50:26 | GoRK | SpaceCoaster: nope not yet, though i'm going to compile it up here in a bit |
| 20:50:52 | SpaceCoaster | GoRK: I included wheaties.h, was that ok? |
| 20:51:13 | GoRK | Lenolium: on software reset, my dsp responsds with 0xb6 0x02... power state indicators |
| 20:51:33 | Lenolium | GoRK: I think I'll just do a hardware reset for now. |
| 20:52:19 | GoRK | SpaceCoaster: yeah that's awesome, though it's not 100% complete.. if anyone finds any comands or responses that are necessary (not just ones that are part of the dsp's 'spec' then please let me or spacecoaster know so we can include them in the .h file |
| 20:52:41 | GoRK | Lenolium: are you writing your daemon? |
| 20:52:56 | GoRK | i havent even gotten to start yet |
| 20:53:37 | Lenolium | GoRK: I'm through most of the generic basic stuff. If you want a copy, let me know and I'll DCC it to you. |
| 20:54:21 | GoRK | Lenolium: yeah i'd like it.. but if you're going to be around in about an hour, I'd rather get it then, since i'll be back at home where my phone is |
| 20:54:51 | Lenolium | Ahh.. I'll send it to you before I head off to lunch then. |
| 20:55:18 | GoRK | hrm.. can you just email it to me? |
| 20:55:50 | GoRK | johnl@blurbco.com |
| 20:55:57 | GoRK | whenever |
| 20:59:01 | prpplague | hmm just a few more weeks till the first lord of the rings movie... |
| 20:59:18 | GoRK | yep. wired ran a good article about it last month |
| 20:59:58 | GoRK | i never was too into lotr so i didnt even know a movie was coming out :) hehheh |
| 21:01:45 | prpplague | GoRK: i was a big fan when i was younger, don't have much time for that stuff anymore, maybe when my son gets older i can brainwash him into liking it too |
| 21:02:08 | GoRK | i need to get to my laser printer so i can print these schematics :) |
| 21:04:36 | SpaceCoaster | GoRK: just re-reading you mail, nice idea about the 512KB flash |
| 21:04:53 | GoRK | SpaceCoaster: it's the only thing that makes sense |
| 21:05:28 | SpaceCoaster | BZFlag felt that it was used for user data, phonelist, config etc. |
| 21:05:33 | GoRK | otherwise there's no purpose for it at all... i think inferno stores the user data on there simply because it's half a megabyte that's not used for anything else... |
| 21:06:01 | GoRK | in inferno, I think it is... Everytime you access call lists and whatnot inferno bogs down bigtime... which makes sense considering all the dsp access that's going on |
| 21:06:47 | GoRK | and since the shannon doesnt include a digital answering system.... no need to waste that much flash, eh? |
| 21:07:19 | SpaceCoaster | true, do you think the 512KB is a standard part on the wheaties module? |
| 21:07:31 | GoRK | from the inferno source, i don think so |
| 21:07:45 | GoRK | but it's nto part of wheaties, it's on the mongoose board |
| 21:07:53 | wej | man they put coffee in their coffee in brasil |
| 21:08:14 | SpaceCoaster | is wheaties just the DSP |
| 21:08:39 | GoRK | SpaceCoaster: there are ifdef's in it for small flash (128 sectors of 256 bytes ea) and large (2048 sectors of 256 bytes ea) |
| 21:09:28 | GoRK | wheaties is a chip that includes the dsp (dsp1650) I think there is other circutry on there besides the dsp, but the flash chips are not on wheaties |
| 21:09:44 | Proteus | *blink* So wheaties is an ASIC? |
| 21:10:39 | wej | yeah gork did all the ic for it.. laid the chip out with Magic and sent it off to a fab in afghanistan for some used pinball parts he told them were devices with which to build atomic weapons. |
| 21:10:54 | SpaceCoaster | the dsp1650 manual talks about 24kwords (maximum) of internal flash rom is this different flash |
| 21:10:56 | GoRK | haha |
| 21:11:00 | SpaceCoaster | hehe |
| 21:11:02 | GPSFan | SpaceCoaster: the 512K flash is an Atmel AT45DB041A and is on the CPU baord, it is accessed via the SPI. |
| 21:11:37 | wej | has anybody built strace against uClibc? |
| 21:11:41 | wej | or wlan_ng? |
| 21:11:44 | wej | i'm gonna work on that next today. |
| 21:11:48 | wej | i'm going to pit every brain cell ih ave left against it. |
| 21:11:54 | GoRK | i want to know if someone has built gdbserver :) |
| 21:11:56 | wej | i'll let you know how badly i'm beaten. |
| 21:13:00 | GoRK | SpaceCoaster: i think that is the acutal dsp program, and i dont think it's able to be rewritten except maybe through jtag ... i am just speculating though |
| 21:13:32 | SpaceCoaster | wej: Echelon will have your name/number and the feds will be knocking on the door, you used ALL the keywords, it was you last birthday man. |
| 21:14:27 | GoRK | SpaceCoaster: or that may just hold dsp code to get the dsp running off of an external rom.. 24kb doesnt seem like a lot to do what the dsp does |
| 21:15:07 | SpaceCoaster | GoRK: the specs say it can be read protected in the factory, I guess that could mean no off chip access. |
| 21:15:24 | wej | SpaceCoaster: Dammit. I better try all those drugs I was scared to do. |
| 21:15:28 | GoRK | it'd make sense if it was just bootstrap stuff |
| 21:17:51 | SpaceCoaster | true, I haven't perused the rest of the code, just focused on the small area we needed, in a couple of days I see what else we've got. |
| 21:18:37 | SpaceCoaster | I saw some code for using the SSP, do you know what that is? |
| 21:19:36 | SpaceCoaster | GoRK: did you see GPSFan came up with a rational explanation of the 2secs or so timeout, it's a DSP watchdog timer |
| 21:20:04 | Lenolium | well, I got wheatiesd running, and I think it's resetting the dsp properly, every 1.5s |
| 21:20:34 | SpaceCoaster | cool, can I "mee too" on an email copy. |
| 21:21:28 | Lenolium | Sure, let me send it out, all it does is sit there right now, but it's pretty ready to be hooked up to actually be able to do reading and writing. |
| 21:21:54 | GoRK | Lenolium: cool |
| 21:22:09 | GoRK | SpaceCoaster: no i missed it but i read your note. i figured it was |
| 21:27:04 | SpaceCoaster | GoRK: it looks like the SSP is used to transfer to the serial flash |
| 21:27:15 | GoRK | hmm |
| 21:28:57 | SpaceCoaster | We need to select the device and configure the SSP bus, I guess the flash is an SSP device and supports one of the SSP protocols? |
| 21:30:24 | SpaceCoaster | hmmm I do a search on AT45DB041A and top of the list is TuxScreen :-) |
| 21:31:29 | scanline | is back (gone 10:33:42) |
| 21:32:31 | GPSFan | SpaceCoaster: http://www.atmel.com/atmel/acrobat/doc1432.pdf |
| 21:32:43 | SpaceCoaster | it appears to be 4mbit with 264 byte pages, 264 = 256 + 8, weird, thanks for the URL |
| 21:33:22 | GPSFan | SpaceCoaster: I just finished ringing out that circuit. ;>) |
| 21:35:30 | SpaceCoaster | Think of a use for it, then I'll have the motivation for making it work in the driver! |
| 21:35:55 | SpaceCoaster | We should be able to something with 512KB of memory. |
| 21:36:27 | SpaceCoaster | Could it run a super slow filesystem? |
| 21:36:34 | GoRK | SpaceCoaster: user state data. That way the user can flash the OS without worrying abou their telephone's settings being destroyed |
| 21:38:08 | Lenolium | Alright. Now it annoyingly clicks every time it sends an idle message to the DSP. I'll call this done for now. |
| 21:38:26 | SpaceCoaster | GoRK: yeah but would that be best as just another mounted filesystem? |
| 21:38:28 | Lenolium | And it gives a HAPPY_TONE on startup. |
| 21:39:10 | SpaceCoaster | Lenolium: excellent I removed that feature from the driver as it was tricky to implement with interrupts. |
| 21:40:11 | GoRK | SpaceCoaster: maybe. if it's fast enough. Otherwise, it might best be used as some sort of persistent storage for a filesystem that runs entierly in ram.. eg. a ramdisk taht does writes in the backgroudn when it would not impact system performance |
| 21:40:24 | prpplague | argh, compiling on a 200mhz box takes forever.... |
| 21:40:40 | Lenolium | Alright, I'm sending out this thing. |
| 21:40:46 | GoRK | the device could be implemented as an mtd device suitable for a jffs2 filesystem or something but would actually be in RAM and only write to flash when the system is ilde |
| 21:42:24 | SpaceCoaster | GoRK: sounds good, I'll put it in on the list of things to investigate |
| 21:45:19 | Lenolium | Alright, sent. |
| 21:45:26 | Lenolium | Could one of you guys check if you got it alright? |
| 21:45:40 | GoRK | did you send it to the list or to individuals? |
| 21:46:03 | SpaceCoaster | got it |
| 21:46:05 | Lenolium | Just to you and spacecoaster, It's not exactly useful yet, unless you want to tourture yourself. |
| 21:46:09 | GoRK | heh |
| 21:46:12 | GoRK | ok yeah i got it |
| 21:46:35 | Lenolium | The little clicky,clicky is not too annoying, but is annoying enough. |
| 21:47:55 | GoRK | CMD_TONE TONE_DEST_SPK TONE_HAPPY :) |
| 21:48:14 | SpaceCoaster | GoRK: you beat me |
| 21:50:01 | GoRK | read and ignore :) |
| 21:50:49 | GoRK | bbiab |
| 21:51:25 | SpaceCoaster | Lenolium: what causes the clicky clicky? |
| 21:52:55 | SpaceCoaster | Lenolium: you use an int for sanity, shouldn't that be unsigned char? |
| 21:56:05 | Lenolium | SpaceCoaster: It does the click every time it does a sanity check that passes. |
| 21:57:16 | Lenolium | Alright, I'm heading off to Client #2, I'll be back... tonight... maybe. |
| 21:57:29 | Lenolium | *plong* |
| 21:58:54 | prpplague | http://jobsearch.dice.com/jobsearch/jobresults.cgi?sr=1&hp=25&cf=07.56c8ad80&brief=0&banner=1 |
| 22:01:45 | scanline | is away: food |
| 22:15:58 | sb- | so if i get a tuxscreen + ethernet card and install linux + X, can i stream video onto my X display? :) |
| 22:17:26 | mallum | but in 8bit color |
| 22:17:58 | sb- | doh!Q |
| 22:18:22 | sb- | almost a full-proof plan.. foiled by a shitty video card |
| 22:18:39 | sb- | ill just unsoder the video chip, find a better one, soder it on, then write some drivers for it |
| 22:19:01 | mallum | sb-: and maybe solder in a new screen too ;) |
| 22:19:27 | SpaceCoaster | sb-: lets try realplay onto the tuxscreen! |
| 22:20:10 | sb- | and decss! |
| 22:20:20 | SpaceCoaster | sb-: hmm looks pretty pukey |
| 22:20:25 | sb- | heh |
| 22:20:37 | sb- | im debating buying one of these bad boys |
| 22:20:42 | sb- | is the screen good enough for web surfing? |
| 22:21:06 | SpaceCoaster | sb-: yes, at $99 bucks it has to be a no brainer |
| 22:22:20 | sb- | how do i hook it up to my pc? do i gotta buy a special cable? |
| 22:22:58 | SpaceCoaster | serial cable for downloading stuff or ppp networking, ethernet using pcmcia card |
| 22:23:53 | sb- | so a M-M serial cable? |
| 22:24:04 | sb- | and there's software for linux or windows to do this? |
| 22:26:37 | scanline | is back (gone 00:24:52) |
| 22:27:50 | SpaceCoaster | sb-: the cable is F-F crossover (null modem), like you would use for connecting two PCs, the software is described on the Wiki, development happens on Linux and these are really development boxes |
| 22:31:11 | sb- | what do you mean by they 'are really development boxes' ? |
| 22:34:01 | SpaceCoaster | well, the software that is on them is a bit clunky, usable but clunky. They are real bargains for incorporating them in another project, home automation, weather station, GPS display, kitchen recipes etc. So a lot of people are buying them to develop stuff for them. The linux software doesn't come with any GUI apps. |
| 22:35:41 | Alehaz | SpaceCoaster: well, the microwindows gui is optional, and picogui is on the horizon, but aside from a pretty good game of snake they don't do much at this point |
| 22:36:23 | sb- | heh, all the development ive done has been with vi, emacs, gcc, and javac :) |
| 22:36:28 | sb- | so i dont need a gui :) |
| 22:36:42 | SpaceCoaster | Alehaz: sorry didn't mean to ignore those fine works. but no apps. |
| 22:37:03 | sb- | but can you irc off them? :) |
| 22:37:10 | sb- | that'll be a big step, when you can irc off your phone |
| 22:37:15 | Alehaz | SpaceCoaster: Yeah, basically no apps |
| 22:38:07 | SpaceCoaster | contradicting myself, viewml is a basic web browser with microwindows |
| 22:39:01 | SpaceCoaster | sb-: it runs linux so, anything you can compile for the arm processor will probably work fine. |
| 22:40:38 | sb- | no one has tried to irc off it yet? |
| 22:41:28 | Alehaz | sb-: Well, I irc off mine all the time, but it's running xchat remotely |
| 22:41:38 | Alehaz | alhaztux is the client on my tuxscreen |
| 22:42:12 | sb- | heh, cool |
| 22:43:53 | mallum | Alehaz: did you build X statically against libc ? |
| 22:47:54 | Alehaz | mallum: Nope, I pivot_root to familiar |
| 22:48:15 | mallum | Alehaz: uh ok |
| 22:49:05 | Alehaz | I have a Familiar root fs on a CF card |
| 22:49:25 | mallum | Alehaz: nice. |
| 22:49:37 | Alehaz | SpaceCoaster: but viewml doesn't actually work on the tuxscreen yet |
| 22:49:42 | mallum | ok bedtime for me |
| 22:49:45 | mallum | night all |
| 22:50:24 | SpaceCoaster | Alehaz: it loaded pages and worked a bit for me, i think |
| 22:50:45 | Alehaz | hmm, i'll try again |
| 22:50:48 | Alehaz | mine cranked and got nowhere |
| 22:51:26 | SpaceCoaster | Alehaz: didn't have a working touchscreen at the time so my testing was limited to specifying URLs on the command line, the touchscreen is better now though :-) |
| 22:51:27 | other | is the pcmcia supposed to be hot swappable? |
| 22:52:01 | Alehaz | other: of course |
| 22:53:41 | Alehaz | why do you ask? |
| 22:56:16 | SpaceCoaster | Alehaz: where did sb- go, we spend 20minute answering questions and he doesn't say thanks, ungrateful sod. |
| 22:56:50 | Alehaz | SpaceCoaster: not the first time either |
| 22:56:59 | Alehaz | I think from now on it's "go read the wiki" |
| 22:59:12 | SpaceCoaster | Alehaz: yeah, tho' I amazed at the mix of skills on this channel, it's the most fun I had developing stuff in years. |
| 22:59:34 | Alehaz | yeah, this has been a great hack |
| 23:02:35 | SpaceCoaster | true and I think it will get even better as the hardware opens up. |
| 23:03:33 | Alehaz | yeah |
| 23:03:43 | Alehaz | I've been involved in some disappointing projects in the past, this is a nice change |
| 23:03:44 | SpaceCoaster | I have a weather station, security cameras and a home automation system which I will control from it. |
| 23:04:11 | SpaceCoaster | Alehaz: commercial or open source projects? |
| 23:05:41 | Alehaz | SpaceCoaster: "lets hack this old weird hardware" projects |
| 23:06:05 | Alehaz | that were riddled with lots of very anxious but basically ignorant people |
| 23:06:17 | Alehaz | fumbling around in the dark, etc |
| 23:07:17 | SpaceCoaster | aahh, I understand, this one is different, mainly because the TuxScreen is quite mainstream, it's ahead of the bunch and not behind in most respects. |
| 23:09:11 | SpaceCoaster | what will happen when they are all sold will be interesting. |
| 23:15:51 | Alehaz | yeah |
| 23:16:30 | GoRK | all i do is ask a bunch of questions and dissapear :) |
| 23:16:42 | Alehaz | :P |
| 23:17:07 | other | Alehaz: i ask because i stuck this pcmcia card in and my unit rebooted. |
| 23:17:29 | SpaceCoaster | GoRK: yeah, that's you alright :-) |
| 23:18:08 | GoRK | yeah you can irc off me too |
| 23:19:02 | GoRK | im an intuitive autonomous irc agent. |
| 23:19:10 | GoRK | that was a big step |
| 23:19:39 | SpaceCoaster | GoRK: seen mallum |
| 23:19:53 | GoRK | mallum went to bed |
| 23:20:43 | SpaceCoaster | good, bit slow, obviously slower processor, but good |
| 23:21:15 | SpaceCoaster | GoRK: what next with the DSP? |
| 23:21:16 | GoRK | it's all aobut the complex logic. I have to calculate if you really want to know if ive seen mallum |
| 23:22:17 | GoRK | SpaceCoaster: well i think that the driver is robust enough that i can get a program working that makes it work like an acutal phone.. perhaps first w/o callerid or speakerpone, but nonetheless functional |
| 23:23:08 | OrngBug | hello |
| 23:23:08 | GoRK | then... maybe once gpsfan gets to the ucb1200 and mongoose board's header i can do some sort of hardware modification to get audio in/out working |
| 23:23:51 | GoRK | hrm what's the best program for printing C source code? like on paper... w/line numbers and maybe a function index |
| 23:24:06 | scanline | doxygen? |
| 23:24:19 | NickB | a2ps? |
| 23:24:27 | GoRK | hmmmmm |
| 23:26:30 | GoRK | thanks |
| 23:29:03 | SpaceCoaster | GoRK: the only thing I might add to the driver is a reset ioctl as was suggested, Lanoliums close/open code comment made me sad. |
| 23:32:45 | GoRK | close/open comment? |
| 23:32:55 | GoRK | oh.. the poor man's reset? |
| 23:33:33 | GoRK | he could have used the software reset also. b2bb or whateve it was |
| 23:34:18 | SpaceCoaster | yeah, so the ioctl is in! It wouldn't be so bad if he didn't check all the return codes :-) |
| 23:35:18 | SpaceCoaster | The device driver book, on the net, that Russ suggested is good for this stuff, makes it easy. |
| 23:37:42 | GoRK | cool |
| 23:47:17 | SpaceCoaster | GoRK: are you a hitch hikers guide fan? |
| 23:47:46 | GoRK | i just like to say share and enjoy |
| 23:48:36 | SpaceCoaster | me too |
| 23:49:05 | GoRK | printing large pdfs over a slow vpn sucks |
| 23:49:33 | GPSFan | GoRK: SpaceCoaster: the latest schematics have just been uploaded. |
| 23:49:55 | SpaceCoaster | k, i'll turn the printer on! |
| 23:50:00 | GoRK | GPSFan: sweet. i am printing them now :) |
| 23:50:09 | GPSFan | There is now a wiki page, http://tuxscreen.net/wiki/view/Main/KenMcguire |
| 23:50:25 | GPSFan | which shows rev's |
| 23:51:02 | GoRK | is Pcmcia.pdf different from pcmcia.pdf ? |
| 23:52:38 | GPSFan | uh.... looks like I munged the file names a bit, check the date. I just uploaded them, will fix momentarily. |
| 23:52:53 | GoRK | ok well i gotta go get my printouts. ill bbiab |
| 23:53:02 | GoRK | must drive across town |
| 23:55:41 | GPSFan | fixed, somehow I typed incorrectly! can't imagine how?? |