| 00:23:23 | vosque | is there anything odd I have to do to familiar to get pivot_root to work? The solution hasn't been obvious to me. |
| 00:24:44 | vosque | when I run it by hand I get: |
| 00:24:50 | vosque | # pivot_root /mnt /mnt/initrd |
| 00:24:50 | vosque | pivot_root: pivot_root: No such file or directory |
| 00:25:08 | vosque | I thought i tmight be a library/binary issue on the familiar side, but I coped all that over and it works |
| 00:40:53 | GoRK | PICOGUI LIVES!!!!! |
| 00:42:11 | GoRK | mmm need ppp |
| 00:42:43 | BZFlag | vosque: what fs are you trying to run pivot_root from? |
| 00:43:01 | GoRK | BZFlag: have you heard of anyone running ppp over the serial port? |
| 00:44:43 | BZFlag | not yet, no |
| 00:45:08 | GoRK | just wondering if it will behave with uclibc.. probably not, huh? |
| 00:45:52 | GoRK | i just got PicoGUI working but until i have a touchscreen driver written, I have to use a remote pointer -- which means i need to be able to talk to the phone with TCP/IP :-/ |
| 00:46:05 | GoRK | alternatively, i guess i could just preempt the touchscreen driver |
| 00:47:52 | GoRK | i guess i can yank the microwindows one pretty easily. brb |
| 00:52:43 | GoRK | is the ASSABET mouse driver the one i'm after from microwindows, anyone? |
| 00:56:15 | vosque | BZFlag: the 0.3 image on the website. |
| 01:11:53 | GoRK | what do the AA batteries in the phone back up exactly? |
| 01:13:55 | as2 | Phone numbers list maybe? |
| 01:14:03 | as2 | Real time clock? |
| 01:15:05 | GoRK | clock, sure... but do they maintain the state of any ram? |
| 01:18:08 | as2 | I dont actually know. |
| 01:18:34 | GPSFan | GoRK, on my TuxScreen under Inferno, I lose the time & date when I remove the wall power. Without schematics it's hard to know for sure. I don't lose the directory & option settings. |
| 01:19:42 | GoRK | woot i now have picogui server and clients running on tux |
| 01:52:34 | | 994 |
| 01:52:34 | BZFlag | ibot: 9 * 9 * 12 + 22 |
| 01:53:51 | BZFlag | batteries just keep the dsp alive. they don't backup anything. |
| 01:54:11 | BZFlag | phone dir etc are stored in the 512k serial flash. |
| 01:57:05 | vosque | it has serial flash too? |
| 01:57:06 | vosque | neat. |
| 02:00:10 | GPSFan | BZFlag, |
| 02:00:21 | BZFlag | linux is not talking to the serial flash yet either. |
| 02:00:44 | GPSFan | ah, that explains it, Is there any hope of getting schematics for the CPU section? |
| 02:01:32 | BZFlag | hope? yes. plans? no. but I'm still asking people. |
| 02:02:08 | BZFlag | though it appears that the serial flash is tied to the dsp and may get handled when we do the dsp code. |
| 02:02:12 | vosque | BZFlag: I'm using the v0.3 image to do the pivot_root.. is that broke or something? |
| 02:02:20 | BZFlag | no. That's what I use. |
| 02:02:30 | BZFlag | be sure to adjust your path. |
| 02:02:43 | BZFlag | you tried what I put up in PivotRoot ? |
| 02:03:00 | BZFlag | that is what is now in the build system. |
| 02:03:31 | vosque | yeah |
| 02:03:35 | vosque | I adjusted my path and such |
| 02:03:47 | BZFlag | finds 2 more phones he didn't count. |
| 02:03:47 | vosque | I'm using your script, in fact. |
| 02:03:57 | vosque | modified and stuff |
| 02:04:07 | vosque | it mounts on /mnt just fine, though |
| 02:04:09 | BZFlag | that exact script is working here for me. |
| 02:04:21 | BZFlag | got a /linuxrc on the new fs? |
| 02:04:59 | vosque | yeah |
| 02:05:01 | BZFlag | it is important to exec 2 things on the new fs. the second actually frees up the orignal fs. |
| 02:05:39 | BZFlag | so exec /linuxrc on real root, pivot_root, exec /linuxrc on new which execs /sbin/init |
| 02:05:40 | BZFlag | then you can unmount the original rootfs. |
| 02:06:04 | vosque | I just untarred task-familiar-complete into a directory and copied the dev entries from the flash image over. |
| 02:06:07 | vosque | <nod> |
| 02:06:24 | vosque | That's what it's doing, but it seems to have a problem with pivot_root |
| 02:06:34 | vosque | I posed what happens earlier when I run it by hand |
| 02:07:13 | BZFlag | try init=/bin/sh and then running the steps by hand? pivot_root complains? |
| 02:07:28 | vosque | hmm |
| 02:07:29 | vosque | hang on |
| 02:07:32 | vosque | I think I got it. |
| 02:08:00 | BZFlag | you can _not_ init=/linuxrc which spawns a shell and then expect to pivot_root. |
| 02:08:12 | BZFlag | linuxrc will have the rootfs locked. |
| 02:09:14 | vosque | ah |
| 02:09:19 | vosque | think I found the proble. |
| 02:09:20 | vosque | +m |
| 02:09:30 | BZFlag | afaict it must be process 1 that calls pivot_root. |
| 02:09:42 | vosque | This is going to be one of those 'bang my head against the wall until it starts working again' again things, I fear. |
| 02:09:44 | BZFlag | my bug or yours? ;-) |
| 02:10:16 | vosque | thought pivot_root worked like the chroot command, and you had to give the new target in relation to the new root |
| 02:10:28 | vosque | Cool! |
| 02:10:31 | vosque | now it's killing init! |
| 02:10:34 | vosque | this is a new behavior! |
| 02:14:15 | BZFlag | your linuxrc should not exit. it should exec /sbin/init |
| 02:14:30 | vosque | It should |
| 02:14:37 | vosque | but new behavior is better than nothing. |
| 02:18:41 | vosque | alright, that did it. |
| 02:50:37 | BZFlag | is hacking on pcmcia-cs for build system |
| 02:53:05 | andersee | cool. |
| 02:53:29 | andersee | is reading Stevens to try and figure out what I'm doing wrong in my code... |
| 04:06:55 | lfjewett | Anyone around? |
| 04:07:08 | andersee | nope |
| 04:07:14 | lfjewett | LOL Liar!! |
| 04:07:32 | lfjewett | Hey, how do you get a exectable file to your tuxscreen? |
| 04:07:46 | lfjewett | I have it hooked up via null modem. |
| 04:08:00 | lfjewett | and am trying to copy over a simple hello world app.. (first test) |
| 04:08:58 | andersee | Well, you could use rz/rz to send it via zmodem... |
| 04:10:25 | lfjewett | ok, thought I was missing something.. |
| 04:10:37 | lfjewett | I can get it there, with Zmodem. and the size is right but it doesn't run |
| 04:10:45 | Aelhaz | chmod +x? |
| 04:10:47 | lfjewett | comes back with no such file or directory |
| 04:10:52 | lfjewett | and Yeah did a chmod 755 on it. |
| 04:11:11 | Aelhaz | it is an armv4l binary, yes? |
| 04:11:37 | lfjewett | I used the arm-linux-gcc toolchain to compile it yes. |
| 04:12:00 | Aelhaz | ok, just asking the obvious questions |
| 04:12:04 | Aelhaz | how are you executing it? |
| 04:12:04 | lfjewett | a.out: ELF 32-bit LSB executable, Advanced RISC Machines ARM, version 1, dynamic |
| 04:12:04 | lfjewett | ally linked (uses shared libs), stripped |
| 04:12:12 | lfjewett | ./a.out |
| 04:12:31 | Aelhaz | linked against uClibc? |
| 04:13:10 | lfjewett | Yes. this is simply the default file from handhelds.org |
| 04:13:18 | Aelhaz | hmm |
| 04:13:32 | Aelhaz | wait, what? HH's stuff is glibc |
| 04:13:58 | lfjewett | going back to check the readme now.... |
| 04:14:56 | lfjewett | Yep your right glibc |
| 04:15:56 | lfjewett | <-- hasn't used linux and C enought to know where to go.. |
| 04:16:49 | Aelhaz | ok, well, the linux root fs that's distributed for the tuxscreen uses uClibc rather than glibc |
| 04:16:57 | Aelhaz | 'cause uClibc is a hell of a lot smaller than glibc |
| 04:17:12 | lfjewett | Ok makes sence. |
| 04:17:15 | Aelhaz | the only problem with that is uClibc isn't thread safe and doesn't support c++ |
| 04:17:21 | Aelhaz | or localization |
| 04:18:27 | lfjewett | Ok, then to compile a simple C program can you tell me what direction to go in.. After I get this right I will make a wiki page for others like myself :-) |
| 04:18:31 | Aelhaz | I'm no expert on this - I've never compiled anything against uClibc |
| 04:18:48 | Aelhaz | I just find trouble and shoot it |
| 04:19:33 | BZFlag | Russ|werk: what did you do to pcmcia-cs to get past it wanting inb and outb ? |
| 04:20:27 | lfjewett | Ahh well thanks anyway Aelhaz |
| 04:20:57 | lfjewett | Perhaps someone else listening in can give a hint when they get around here. |
| 04:21:02 | BZFlag | lfjewett: get the buildsystem out of cvs and build it there. |
| 04:21:12 | BZFlag | bbiab |
| 04:31:52 | Aelhaz | lfjewett: my tech experience is in high level tech support, network admin, and testing. I can't code worth shit . . . . but i have a knack for figuring out what's busted |
| 04:33:43 | lfjewett | I'm with you Aelhaz.. I'm a network analyst. |
| 04:34:07 | lfjewett | My codeing experiance was from 9th grade working with fortran 77 and a little C. |
| 04:35:00 | Aelhaz | lfjewett: network admin drove me nuts. I got sick of doing things the wrong way 'cause the boss said so. |
| 04:35:50 | Aelhaz | lfjewett: Maybe that was because i worked at a company where everybody either knew damn well how complex networks work, or has been around the first kind so long they think it's easy |
| 04:36:08 | lfjewett | I can understand that I work for a *large* ISP so we often do as we're told not as we do.. |
| 04:36:15 | Aelhaz | i'd do admin work again, for, say, an accounting firm |
| 04:36:53 | Aelhaz | but never again for an ISP |
| 04:37:27 | Aelhaz | I'd want an employer with wide profit margins, a stable market, and execs who just want it to work |
| 04:38:20 | lfjewett | LOL well the ISP I work for does have all 3 of those. |
| 04:38:31 | Aelhaz | well, it helps if they think it's magic |
| 04:39:10 | Aelhaz | the bane of my existance were execs who would read trade rags and then send out an edict based on what some dumbass part time pundit said |
| 04:39:36 | lfjewett | Haha, I can see that in my head. but haven't experianced it.. |
| 04:40:20 | lfjewett | Things around work are about as tight as they come.. 99% of our guys know there stuff inside/out. and the execs are most all very technical. |
| 04:40:56 | Aelhaz | There was a Dilbert . . . . . PHB walks up, says "We should build an SQL database" |
| 04:41:27 | Aelhaz | dilbert thinks . . "Uhoh. Does he know what he's talking about, or did he just finish reading a magazine article . . . . " "What color should it be?" |
| 04:41:36 | Aelhaz | . . . "I think Mauve has the most RAM" |
| 04:42:04 | lfjewett | LOL that reminds me to check dilbert.com tommarrow for new ones :-) |
| 04:44:34 | Aelhaz | my sister is a CPA at a well established accounting firm |
| 04:44:46 | Aelhaz | I admire their policy on computer hardware |
| 04:45:11 | Aelhaz | "if there's anything wrong with it, get rid of it immediately and replace it with the best price/performance you can find" |
| 04:45:27 | Aelhaz | my sister mentions one day, "my monitor flickers sometimes" |
| 04:45:38 | Aelhaz | she comes in the next morning and there's a brand new 19" viewsonic on her desk |
| 04:46:29 | Aelhaz | Seems some years ago they decided, "hey, we're accountants, lets run the numbers and see how much money we save by being stingy about equipment" |
| 04:46:44 | Aelhaz | and it turned out to be not much, and certianly not worth annoying the labor |
| 04:47:58 | lfjewett | At work, you ask for a new SUN E6500 and it'll be there in a day, ask for a new monitor for your desk and it could be months. |
| 04:48:41 | Aelhaz | heh |
| 04:49:47 | Aelhaz | at the isp we worked for, we'd ask for more server hardware, fill out a requisition form, and six weeks later we'd get, something that doesn't meet any of the specs |
| 04:50:33 | lfjewett | Ouch |
| 04:50:47 | Aelhaz | when we asked for monitors we got used 4 year old 15" compaqs |
| 04:50:57 | lfjewett | What ISP was this |
| 04:51:17 | Aelhaz | I once literally saw our purchasing manager buying refurbished TrippLite UPSs off the back of a truck |
| 04:51:21 | Aelhaz | AirSwitch |
| 04:52:04 | Aelhaz | the guy never figured out that there's a reason there's so many refurb tripplites out there |
| 04:52:34 | Aelhaz | (because they SUCK) |
| 04:53:36 | lfjewett | My work is a little big to get to play with stuff like that.. I almost seem to miss out on alot of good stuff. All our power is on building UPS and Generators. |
| 04:53:44 | Aelhaz | our purchasing manager was a bastard. |
| 04:54:01 | Aelhaz | he finally got fired, but it was long after I quit |
| 04:54:18 | lfjewett | I don't dare quit now, economy to bad, stock to low. |
| 04:54:37 | Aelhaz | yeah |
| 04:54:40 | lfjewett | BZ did you ever come back? |
| 04:54:44 | Aelhaz | well, airswitch is more or less gone anyhow |
| 04:54:55 | lfjewett | Yeah hadn't heard of them before.. |
| 04:54:56 | Aelhaz | changed their name to switchpoint, trying to sell infrastructure hardware now |
| 04:55:03 | Aelhaz | They have an interesting technology |
| 04:55:17 | Aelhaz | long haul ethernet, basically. |
| 04:55:44 | lfjewett | Hmm cheaper than Fiber |
| 04:55:59 | lfjewett | ? |
| 04:56:19 | lfjewett | Andersee are you available for a second? |
| 04:56:33 | andersee | sure. Whaddayou need? |
| 04:56:57 | lfjewett | After getting the ARM toolchain, I don't know what steps to take to compile a binary for the tuxscreen |
| 04:57:19 | lfjewett | I guess the toolchain is linked for glibc but the tuxscreen use's uClibc |
| 04:57:29 | Aelhaz | lfjewett: well, their value proposition is that it requires less expensive hardware than cable modems or dsl |
| 04:57:53 | andersee | So either use the uClibc wrapper, or compile with --static for a static glibc binary |
| 04:57:54 | Aelhaz | they say it's about $180 per house to roll out vs. about $500 per house for cable modem |
| 04:57:58 | lfjewett | Ahh.. |
| 04:58:23 | Aelhaz | i don't think they compete so well with DSL anymore |
| 04:58:31 | lfjewett | Thank you very much ander.. Can you pint me to a uClibc wrapper.. |
| 04:58:36 | Aelhaz | four years ago a DSLAM was an obscenely expensive chunk of equipment |
| 04:58:46 | lfjewett | Sorry fot the questions, I'm not a veteran at coding.. |
| 04:58:48 | Aelhaz | and that's much less the case now |
| 04:59:02 | andersee | lfjewett: grab the tuxscreen buildsystem. It has everything... |
| 04:59:18 | lfjewett | I just got that from CVS. |
| 04:59:21 | BZFlag | pcmcia-cs in cvs now. |
| 04:59:28 | lfjewett | Looking though it now. |
| 04:59:38 | andersee | You can use the uClibc gcc wrapper from that... |
| 04:59:44 | andersee | BZFlag: cool |
| 04:59:53 | Aelhaz | initial DSL rollout used to be in the millions of dollars, now it's in the hundreds of thousands |
| 05:00:04 | andersee | so that leave just udhcp and rz/sz? |
| 05:00:32 | lfjewett | that is the buildroot-tux from CVS correct.. I will read up and try and figure it out.. |
| 05:00:52 | lfjewett | Aelhaz: I just wish I could afford DSL to my place.. |
| 05:01:07 | BZFlag | perhaps. deploying now... |
| 05:01:43 | andersee | BZFlag: how complete are the set of pcmcia modules? |
| 05:02:32 | Aelhaz | lfjewett: out here in utah, qwest isn't unreasonable. 640k down / 320k up for about $30/mo on top of the regular phone line charges, $20/mo for isp w/ one static ip |
| 05:02:41 | BZFlag | I did not build the modules. using the kernel versions at present. |
| 05:02:43 | Aelhaz | I can get a block of 5 more ip's for $5/mo |
| 05:02:49 | Aelhaz | they cap it at 8 gigs each way per month tho |
| 05:02:57 | andersee | BTW, I got that select based disk copy working.. |
| 05:03:00 | Aelhaz | so it's not like i can serve up mp3's on gnutella 24/7 |
| 05:03:30 | BZFlag | and? perfomance? |
| 05:03:42 | andersee | Aelhaz: how do you check your xmission utilization? |
| 05:04:16 | Aelhaz | andersee: They supposedly have a way i can check that. I just know i get email from some daemon telling me to knock it off if i go over 2gb in a week |
| 05:04:22 | andersee | BZFlag: within 20 secs on me earlier threaded version.... |
| 05:04:51 | BZFlag | which is a % faster that dd with large block size? |
| 05:05:02 | andersee | Presumably my system activity over ~11 min could have caused that though |
| 05:05:41 | andersee | Well, the current version took 11 min 10 sec, while dd took 17 min 53 sec... |
| 05:05:55 | andersee | So I'm still way faster. :) |
| 05:06:13 | BZFlag | Child dir "." (ino #1) of dir ino #1 appears to be a hard link |
| 05:06:37 | andersee | This was copying 640 Meg of .ogg files from MO drive to MO drive. |
| 05:06:38 | BZFlag | what bs= on dd? |
| 05:06:48 | andersee | default |
| 05:07:07 | BZFlag | ahh. try like 1 meg out and see what that does. |
| 05:07:24 | andersee | BZFlag: that is a bug I left in mkfs.jffs2. Nothing too serious. |
| 05:07:38 | Aelhaz | andersee: I've been given the impression that i should be able to view a graph of my useage at http://www.xmission.com/stats/routers/ - I've never been able to log in to do it, tho |
| 05:07:46 | Aelhaz | might just be for corporate accounts |
| 05:08:01 | andersee | BZFlag: I think I gave / an inode, which it seems I shouldnt have done. |
| 05:09:53 | andersee | BZFlag: k. Trying using dd ... bs=1M |
| 05:10:23 | Aelhaz | one of these days I'll set up mrtg on my own router |
| 05:10:32 | Aelhaz | or better yet, ntop |
| 05:15:30 | andersee | BZFlag: I think the 'make install' for the cardmgr target needs a $(PREFIX) set so it doesnt try to install to the host |
| 05:15:49 | lfjewett | Thanks for the help, still don't understand 90% of this but I'll do some more reading tommarrow and try again. Least my box says "Hello World" now, even though it is a 900k file :-) |
| 05:16:33 | andersee | yipe. You added --static and compiled vs glibc, right? |
| 05:16:43 | BZFlag | andersee: I added the prefix in the ./Configure line. |
| 05:17:12 | lfjewett | Yes. |
| 05:17:19 | lfjewett | just to check it out. |
| 05:17:46 | lfjewett | I am also doing the make for the CVS stuff now but it's 21 megs or so.. so that'll be waiting for me in the AM. |
| 05:18:30 | andersee | Sorry about that. At least the buildroot itself is pretty lightweight |
| 05:19:22 | lfjewett | I'll learn this all eventually.. It's a far cry from fortran.. LOL |
| 05:20:27 | andersee | far :gt: indeed |
| 05:21:59 | andersee | BZFlag: upping the dd buffersize to 1M took dd into the same speed range as the other methods... |
| 05:22:34 | andersee | Lemme try upping the bufsize on the select based one to 1M now |
| 05:24:56 | BZFlag | heh, I installed all the pcmcia man pages... ;-) |
| 05:27:29 | GoRK | greetings |
| 05:27:55 | GoRK | BZFlag: you are working on pcmcia-cs for the build system? |
| 05:28:06 | BZFlag | GoRK: it's in there. tuning it now. |
| 05:28:25 | GoRK | ahh cool i am so ready for it! ... i got picogui up and running |
| 05:28:55 | GoRK | i need a nic so bad.... but they are very close now. the'll be here monday i think for sure |
| 05:30:42 | Aelhaz | <deleted> vendor from <deleted> upstate NY <deleted> wants a <deleted> signature for delivery of a <deleted> ten dollar heatsink . . . . |
| 05:31:03 | GoRK | also i got lrzsz working sort of with the build system. they don't get autobuilt but it'd probably be a matter of a quick 3 lines in the makefile. I'll do it within a copuple days and send a diff |
| 05:31:08 | Aelhaz | has to go to the <deleted> UPS dock monday to pick up the <deleted> thing |
| 05:31:27 | GoRK | dude call ups and get a dropoff notice put on file. you'll never have to sign for anything again |
| 05:31:57 | Aelhaz | GoRK: I didn't used to have any kind of problem with UPS, but I havea different driver now, and he's a weenie |
| 05:32:17 | Aelhaz | the old guy had no qualms about stacking up boxes next to my porch |
| 05:32:27 | Aelhaz | this guy won't do that |
| 05:32:39 | Aelhaz | guess I'll do that tho, getting tired of this BS |
| 05:33:26 | GoRK | well he'll have to if you have one of those things on file for your address :) they like it too because they can just ring the buzzer and bolt ... they dont have to wait around for you to mash on that plastic strip |
| 05:34:05 | Aelhaz | this new driver is so anal, has to leave the itty bittyest package with neighbors . . . |
| 05:34:29 | Aelhaz | drives me nuts, 'cause they're never home when i am |
| 05:35:23 | Aelhaz | I live in a neighborhood that's all condos. And i live on a corner. So he must think the porch is too visible, too dangerous |
| 05:36:12 | Aelhaz | hasn't clicked for him yet that the neighborhood is lousy with retired people who stay home all day and putter around their yards, and young couples with children too young to go to school, who stay home all day and let their kids play in *my yard |
| 05:36:36 | Aelhaz | Not exactly a prime neighborhood for swiping boxes off of people's porches. too many witnesses. |
| 05:37:46 | andersee | If you were in my neighborhood, my daughter would have every one of your packages in her bedroom as playtoys. :) |
| 05:37:52 | Aelhaz | hehe |
| 05:38:58 | Aelhaz | It's been a little bit of culture shock, actually. I grew up in suburbia, everyone was at least vaguely aware of property lines |
| 05:39:17 | Aelhaz | kids who live two blocks away come and play on my lawn and leave their bicycles on my porch |
| 05:39:25 | Aelhaz | 'cause it's the biggest chunk of grass around |
| 05:40:00 | Aelhaz | they don't have any concept of turf |
| 05:40:18 | Aelhaz | gets annoying when I'm tryin to sleep in on weekends . . . |
| 05:42:08 | Aelhaz | anyway, I'm goin to bed. g'nite all. |
| 05:42:15 | andersee | night |
| 05:44:00 | GoRK | do you have to do anything special to strip a library |
| 05:44:45 | GoRK | vs a regular binary |
| 05:47:11 | GoRK | -rwxr-xr-x 1 root 5764 162876 Sep 15 01:40 pgserver <-- PicoGUI server |
| 05:47:20 | BZFlag | andersee: looked at hard links again? |
| 05:55:35 | andersee | Lemme take a quick look. |
| 05:56:17 | andersee | BTW, you were going to add that PREFIX stuff so the cardmgr stuff would stop whining... |
| 05:57:26 | BZFlag | /dev/root 3072 1196 1876 39% /initrd |
| 05:57:38 | andersee | BZFlag: try commenting out line 928 in mkfs.jffs2 'output_pipe(ino, &sb);' |
| 05:57:54 | BZFlag | is now pivot_root on nfs from build tree! debian arm. ;-) |
| 05:58:02 | andersee | I think / should only have a dirent, not an inode... |
| 05:58:20 | andersee | :) |
| 05:58:32 | andersee | needs to order a nic so I can try that |
| 05:58:43 | GoRK | i can't find the ioctls that set brightness/contrast in the kernel source... i thought it was in ucb1200_ts.c |
| 05:58:43 | BZFlag | cardmgr is not complaining about prefix is it? I tell it path info on the ./configure line |
| 05:59:45 | andersee | BZFlag: It is for me... Tries to install to /etc/pcmcia and friends (and fails of course). |
| 05:59:56 | BZFlag | cvs up -PdA ? |
| 06:01:45 | BZFlag | hmm.... I got it now. I was not seeing that. |
| 06:01:47 | andersee | I still see nothing about PREFIX="$(TARGET_DIR)" under cardmgr... |
| 06:01:55 | andersee | It needs that |
| 06:02:01 | andersee | For the install target |
| 06:02:10 | andersee | make PREFIX="$(TARGET_DIR)" -i -C $(PCMCIA_DIR) install; \ |
| 06:04:48 | BZFlag | eep. |
| 06:05:10 | BZFlag | I ran a configure in another window and did not copy over the --target. added now. |
| 06:05:39 | BZFlag | cvs up -PdA ; rm -rf root pcmcia-cs* ; make |
| 06:06:21 | andersee | Tried that change I suggested for mkfs.jffs2? |
| 06:06:36 | GoRK | when i use pcmcia cards in this do i have to power off to insert them or can i just plug them in while running? |
| 06:08:14 | BZFlag | not yet... |
| 06:08:17 | andersee | BZFlag: I'm noticing -- looks like we arn't installing any kernel modules. |
| 06:08:29 | andersee | That will slow down pcmcia a bit |
| 06:08:35 | BZFlag | true at present. all compiled into the kernel now. |
| 06:08:46 | BZFlag | why? |
| 06:08:57 | andersee | just thought i'd mention it. |
| 06:09:12 | BZFlag | why do you say it will slow down pcmcia? |
| 06:09:12 | andersee | was looking for scsi support and noticed the lack |
| 06:09:43 | andersee | I was thinking of trying my adaptec 1460... |
| 06:10:10 | BZFlag | no problem. hack the kernel config to compile in your pcmcia-cs driver for the card. that's all that's needed. |
| 06:10:24 | BZFlag | GoRK: plug them in while running. |
| 06:10:39 | GoRK | ok |
| 06:10:57 | andersee | Hmpf. My threaded copy beats dd and select based copying consistantly by 20secs in my tests. grumble. |
| 06:11:38 | GoRK | im testing it now. my rootfs is 680556 bytes including picogui server and i think i have about 20 client apps and a couple themes in there too |
| 06:11:46 | GoRK | oh yeah and lrzsz too |
| 06:13:02 | andersee | BZFlag: is there a verion of blob yet that will boot the kernel direct from jffs2? |
| 06:13:29 | BZFlag | not yet. |
| 06:13:51 | BZFlag | when that exists, then I'll move the default config to all modules. |
| 06:14:18 | andersee | Who if anyone is working on that one? |
| 06:14:45 | BZFlag | Russ|werk has some patches that worked on an older blob. don't know where they are now. |
| 06:15:39 | BZFlag | then we will be 64k blob, 64k parameters, 3.9M rootfs |
| 06:15:50 | andersee | BZFlag: could you try out that mkfs.jffs2 hack for me to verify it removes the hardlink problem... |
| 06:16:00 | BZFlag | built it, flashing now |
| 06:16:04 | andersee | k, cool. |
| 06:16:44 | andersee | I would, but I need to rebuild and I don't want CPU use to mess up my timing tests... |
| 06:17:16 | BZFlag | mkfs.jffs2 did not complain and produced a filesystem. |
| 06:17:47 | andersee | good, good. I'd tested that much between runs... |
| 06:17:48 | BZFlag | thinks everyone should flood ping erik. oh.. he's behind a fw. darn. |
| 06:18:04 | andersee | :) |
| 06:18:42 | andersee | The real test is if the kernel whines when booting the phone... |
| 06:19:02 | BZFlag | Child dir "." (ino #1) of dir ino #1 appears to be a hard link <- still. |
| 06:19:20 | andersee | Gack! |
| 06:19:29 | andersee | mumbles |
| 06:20:44 | BZFlag | it did rebuild mkfs and the image. and I am on the new one. seems to have not fixed it. |
| 06:20:59 | GoRK | if i reboot do i have to reinsert the pccard? |
| 06:21:07 | BZFlag | GoRK: no. |
| 06:21:21 | andersee | mkfs.jffs2: writing '/' ino=1 |
| 06:21:21 | andersee | mkfs.jffs2: writing dir 'bin' ino=2 parent_ino=1 |
| 06:21:21 | andersee | mkfs.jffs2: writing dir 'etc' ino=3 parent_ino=1 |
| 06:21:21 | andersee | mkfs.jffs2: writing dir 'etc/init.d' ino=4 parent_ino=3 |
| 06:21:21 | andersee | mkfs.jffs2: writing dir 'etc/rc.d' ino=5 parent_ino=3 |
| 06:21:55 | andersee | I translate '/' to '.' so it still seems to be the same thing... |
| 06:22:26 | andersee | could you check the timestamp on mtd/util/mkfs.jffs2 just to be certain it got rebuilt? |
| 06:22:37 | BZFlag | GoRK: the thing that is broken is if you do a "cardctl eject ; cardctl insert" the insert will fail on many cards. "cardctl eject" is fine, but then remove an reinsert the card. |
| 06:22:49 | BZFlag | I checked the log. it did. |
| 06:23:13 | andersee | very wierd then. |
| 06:23:16 | GoRK | hmm when i build cardmgr and other pcmcia binaries don't get copied into root/ |
| 06:23:37 | BZFlag | GoRK: cvs up -PdA I just fixed that. |
| 06:23:41 | GoRK | ahh |
| 06:23:57 | GoRK | how groovy |
| 06:24:03 | BZFlag | cvs up -PdA ; rm -rf root pcmcia-cs* ; make |
| 06:25:12 | BZFlag | andersee: any reason to keep S05ramdisk1 in there? |
| 06:25:36 | GoRK | is the 512k data area available for use yet? |
| 06:25:42 | BZFlag | no |
| 06:26:24 | GoRK | i gather that that is where the telephony apps in inferno store user data? |
| 06:26:27 | andersee | BZFlag: that one is for a r/o root fs. not the case here, so sure. axe it |
| 06:26:48 | BZFlag | GoRK: correct |
| 06:27:10 | BZFlag | andersee: chopping now... |
| 06:27:24 | andersee | BZFlag: hmm. axing it will me when the kernel changes perms on pts etc they fo to flash though.... |
| 06:27:35 | andersee | s/ will me/ will mean/ |
| 06:28:16 | andersee | s/they fo/they go/g |
| 06:30:46 | GoRK | hmm this CF card doesnt seem to be working.. are some CF->PCMCIA adaptors 5V for some reason? |
| 06:31:10 | GoRK | lemme jam it into the laptop |
| 06:32:34 | BZFlag | andersee: symlinks in devices.txt yet? |
| 06:33:21 | BZFlag | andersee: perhaps dev -> /var/dev and /var as ramfs then? |
| 06:33:55 | BZFlag | GoRK: what card? if may just not have a config entry. |
| 06:34:09 | andersee | that would be best to avoid wearing out the flash... |
| 06:34:13 | BZFlag | oh... flash card? I didn't fix that yet |
| 06:34:24 | GoRK | shouldnt cardctl ident say it's there? |
| 06:34:55 | andersee | BZFlag: no, no symlink support yet for device tables |
| 06:35:33 | BZFlag | hmmm /me wonders how to create /var/dev |
| 06:35:46 | andersee | skel.tar |
| 06:35:57 | BZFlag | GoRK: is cardmgr running? then cardctl ident should show it, yes. |
| 06:35:58 | vosque | any reason you guys aren't using devfs? |
| 06:36:14 | BZFlag | is an old timer and never switched |
| 06:36:23 | vosque | I see. |
| 06:36:34 | andersee | devfs is not very lightweight |
| 06:36:37 | GoRK | yeah there it goes |
| 06:36:40 | GoRK | bottom socket |
| 06:36:51 | BZFlag | not sure that's a good reason. like andersee says it is large though. |
| 06:37:15 | vosque | I switched to devfs on all the machines I could when it finally settled down |
| 06:37:25 | andersee | And I really hate seeing 'df' show names that wrap over 80 columns... |
| 06:37:30 | vosque | size is a good reason, though |
| 06:37:55 | vosque | I don't mind it. |
| 06:38:20 | vosque | I actually like the descriptiveness of it. |
| 06:39:02 | andersee | It does make sense in a number of places. |
| 06:39:33 | andersee | But I doubt folks will be running fiber channel on the tuxscreen anytime soon.. |
| 06:40:38 | vosque | harsh |
| 06:40:48 | vosque | the qpe ipkgs's total just under 4 meg. |
| 06:41:12 | GoRK | can we stick 'ifconfig lo 127.0.0.1 up' into /linuxrc ? |
| 06:41:29 | vosque | then again, that's with a lot of internationalization. |
| 06:41:47 | GoRK | or you could use PicoGUI :) |
| 06:41:54 | vosque | heh |
| 06:42:05 | vosque | I *could*. |
| 06:42:13 | vosque | And I do plan on looking at it, eventually.. |
| 06:42:36 | GoRK | PicoGUI is crashing on the aqua theme. I dont think it works in 8bit... hmm. i'll have to see about modifying it.. all the other themes seem to work properly on the tux (the ones that dont use jpegs anyway -- not quite through with that) |
| 06:42:59 | vosque | I haven't thought about the 8bit thing.. hmm. |
| 06:43:28 | BZFlag | GoRK: flash cards are fixed now. |
| 06:43:33 | andersee | vosque: 4 meg total or compressed? |
| 06:43:39 | BZFlag | cvs up -PdA ; rm -rf root ; make |
| 06:43:43 | vosque | that would be compressed. |
| 06:43:53 | vosque | But that's everything and the kitchen sink. |
| 06:43:57 | vosque | It's also compiled against glibc. |
| 06:44:21 | andersee | Looks like qpe will have to run on CF or nfs then -- won't fit in flash. |
| 06:44:30 | andersee | The screenshots look pretty |
| 06:44:51 | vosque | It would have to be, right now. |
| 06:45:21 | andersee | how about if you compile w/o i18n? |
| 06:45:24 | BZFlag | andersee: they are all c++ too. |
| 06:45:26 | vosque | But there's a lot of stuff in QPE that just doesn't make much sense in a tuxscreen aspect, like the PPP control.. |
| 06:45:41 | GoRK | qpe won't compile against uclibc as of yet. C++ish |
| 06:45:43 | BZFlag | vosque: tux has a modem. ;-) |
| 06:45:53 | vosque | truethat. |
| 06:46:12 | vosque | Personally, it's going to sit on my desk and have an ethernet connection. |
| 06:46:13 | andersee | GoRK: I need to finish this project I'm on, so I can fix uClibc c++ support... |
| 06:46:17 | GoRK | yeah and i wish for ppp right now too so i could run it over this serial port |
| 06:46:37 | vosque | so ppp is kinda besides the point.. |
| 06:46:38 | vosque | hmm.. |
| 06:46:49 | vosque | lemme see if I can compile with out i18n |
| 06:46:50 | GoRK | i am kind of dead in the water without the ability to point and click in picogui... which i could do remotely if i had it on the net somehow |
| 06:47:46 | vosque | the X calibrate thing is wacky. |
| 06:48:00 | vosque | hmm |
| 06:48:06 | vosque | I should download the packages just to see if it works. |
| 06:50:17 | BZFlag | GoRK: lo starting now in linuxrc |
| 06:50:26 | GoRK | might also need a route add 127.0.0.1 lo in there also |
| 06:51:39 | GoRK | or really a route add -net 127.0.0.0/8 lo but dunno what syntax busybox uses for that .. my phone is flashing right now |
| 06:53:22 | GoRK | PicoGUI needs those.. it'll talk using EITHER unix sockets or TCP/IP.. kind of retarded but whatever |
| 06:54:07 | GoRK | hda: hda1 <-- whoot |
| 06:54:23 | BZFlag | GoRK: ok. added that. I'll commit in a bit... |
| 06:54:52 | GoRK | route add -net 127.0.0.0 netmask 255.0.0.0 lo |
| 06:56:48 | BZFlag | yep. that's what I did. |
| 07:01:35 | andersee | BZFlag: So udhcp and lrzsx still to go then... |
| 07:02:20 | BZFlag | and blob. |
| 07:03:35 | GoRK | i have lrzsz mostly done |
| 07:04:25 | GoRK | there is a magic configure line and a strip command and then you can probably integrate it in like 2 minutes |
| 07:04:47 | vosque | hmm |
| 07:06:48 | GoRK | CC=$(STAGING_DIR)/usr/bin/arm-uclibc-gcc LD=$(STAGING_DIR)/usr/bin/arm-uclibc-ld ./configure --host=arm-linux --build=`./config.guess` --disable-timesync --disable-nls --disable-syslog |
| 07:07:33 | andersee | BZFlag: I'm getting udhcp set up now. |
| 07:08:05 | GoRK | that configure built lrzsz against uclibc just fine then i stripped with arm-uclibc-strip --remove-section=.note --remove-section=.comment lsz lrz |
| 07:08:58 | GoRK | i didnt have to modify the source at all |
| 07:11:07 | vosque | ooo crap |
| 07:11:12 | vosque | have to build new packages |
| 07:11:18 | vosque | to compile for 8 bit |
| 07:11:22 | vosque | that actually might make things smaller |
| 07:13:53 | GoRK | wasnt someone talking about addressing the display as 12 bit? |
| 07:14:09 | vosque | Hmm |
| 07:14:10 | vosque | yeah |
| 07:14:12 | vosque | er |
| 07:14:18 | vosque | not 12-bit |
| 07:14:23 | BZFlag | hacks the ramfs setup... |
| 07:14:49 | vosque | We'd lose alpha-channel stuffs, which would make it a bit smaller. |
| 07:15:37 | BZFlag | it might be possible to address the tux screen in 12/16 bit. |
| 07:16:04 | BZFlag | the sa1100 and the panel both support it, it remains to be seen if the support hw does. |
| 07:16:23 | vosque | what does inferno use? 8-bit? |
| 07:17:08 | andersee | k. I just added udhcp. |
| 07:17:44 | andersee | is reminded he promised Russ he would make uClibc lockf work... |
| 07:19:39 | BZFlag | vosque: yes, 8 bit |
| 07:20:57 | BZFlag | dhcpc.o(.text+0x1e4): undefined reference to `lockf' |
| 07:21:15 | andersee | see my comments above... |
| 07:21:28 | BZFlag | so you broke cvs and are happy with that? |
| 07:21:29 | BZFlag | ;-) |
| 07:21:32 | andersee | is implementing lockf in uClibc |
| 07:21:37 | BZFlag | as we speak? |
| 07:21:43 | andersee | yup |
| 07:21:46 | BZFlag | cool. |
| 07:22:30 | BZFlag | I just commited the ramfs patch. |
| 07:22:50 | BZFlag | devices are now in /dev -> /var/dev -> /boot/var/dev |
| 07:23:21 | BZFlag | /boot/var gets copied to /var in /linuxrc |
| 07:23:47 | BZFlag | which hides the /var/dev -> /boot/var/dev symlink. |
| 07:25:58 | BZFlag | ramfs, tmpfs : prefs? |
| 07:26:13 | vosque | does jffs do any compression? |
| 07:27:39 | andersee | yep |
| 07:27:45 | andersee | err. jffs2 does |
| 07:27:55 | andersee | which is what we use |
| 07:28:06 | vosque | er, yeah jffs2 |
| 07:30:25 | BZFlag | tmpfs is ramfs with possible swap, yes? and we have no swap. |
| 07:32:09 | andersee | BZFlag: cvs up uClibc... |
| 07:35:00 | andersee | Looks like some pcmcia stuff wants inb, inw, outb, outw for arm |
| 07:35:33 | BZFlag | dhcpc.o(.text+0x1e4): undefined reference to `lockf' |
| 07:35:59 | BZFlag | andersee: pcmcia assumes it can build some test apps that are x86 specific. |
| 07:36:16 | BZFlag | even though it's been told it is building for arm |
| 07:36:49 | andersee | I just fixed the lockf thing. cvs up again. I'd forgotten to add the new stuff into a makefile |
| 07:38:44 | BZFlag | is rebuilding uClibs |
| 07:38:46 | BZFlag | is rebuilding uClibx |
| 07:38:48 | BZFlag | is rebuilding uClibc |
| 07:38:57 | BZFlag | eek. tired and can't type... |
| 07:39:06 | andersee | sorry. try another cvs up. |
| 07:39:28 | andersee | I forgot that the anon cvs pull is from oss. I just encouraged an early rsync... |
| 07:39:55 | BZFlag | still failing.... |
| 07:39:58 | BZFlag | ah. |
| 07:39:58 | andersee | Should see |
| 07:40:01 | andersee | U libc/misc/file/Makefile |
| 07:40:01 | andersee | U libc/misc/file/lockf.c |
| 07:41:22 | andersee | BTW, 'rm -rf uClibc/lib staging_dir' is sufficient... |
| 07:42:41 | andersee | A full uClibc rebuild is only needed when changing config options or struct definitions in the headers |
| 07:43:03 | andersee | Gack |
| 07:43:06 | andersee | /home/andersen/tuxscreen/buildroot-tux/staging_dir/lib/libc.so: undefined reference to `__getpid' |
| 07:43:27 | BZFlag | dhcpc.o: In function `pidfile_acquire': |
| 07:43:27 | BZFlag | dhcpc.o(.text+0x1e4): undefined reference to `lockf' |
| 07:44:18 | andersee | Odd. I don't see that one anymore. |
| 07:44:44 | BZFlag | /home/timr/work/shannon/buildroot-tux/staging_dir/usr/bin/arm-uclibc-gcc dhcpc.o clientpacket.o script.o options.o socket.o packet.o -o udhcpc |
| 07:44:44 | BZFlag | /home/timr/work/shannon/buildroot-tux/staging_dir/lib/libc.so: undefined reference to `__getpid' |
| 07:44:52 | andersee | right. |
| 07:44:57 | andersee | I'm digging on that one. |
| 07:45:13 | andersee | Looks like something odd happening in the headers... |
| 07:45:18 | BZFlag | I had not nuked stagingdir before. |
| 07:46:11 | andersee | While I fix this, want to encourage pcmcia to stop rebuilding every time? |
| 07:48:04 | andersee | k. Got it. A buglet in uClibc/misc/file/lockf.c |
| 07:57:02 | GoRK | ok i think i have the touchscreen almost working with PicoGUI |
| 07:58:25 | GoRK | like that is all too impressive to you guys :) |
| 08:03:46 | BZFlag | cardmgr used to build and does not now. |
| 08:04:33 | andersee | is getting sleepy |
| 08:05:26 | andersee | I'm off to bed I'm afraid. |
| 08:05:45 | andersee | night all |
| 08:06:56 | GoRK | lalala zmodem hurry up |
| 08:08:07 | GoRK | what exactly do the batteries in the phone back up? is that 512k area not really flash memory or something? |
| 08:09:33 | BZFlag | batteries keep the phone (dsp) alive when the power is out. they do nothing for the computer aspects |
| 08:10:02 | GoRK | ah |
| 08:11:51 | GoRK | is /dev/ts from the microwindows distro the same as /dev/ucb1200-ts from the buildsystem? |
| 08:16:01 | BZFlag | probably. |
| 08:17:04 | GoRK | nm found the problem. it is right |
| 08:17:23 | GoRK | i was not enumerating the driver properly |
| 08:26:49 | BZFlag | got udhcpc installed as well, but it does not get an address. |
| 08:27:15 | BZFlag | kick andersee when he gets up. ;-) |
| 08:27:17 | BZFlag | nite all |
| 08:28:04 | GoRK | kicks andersee while he's down |
| 08:28:07 | GoRK | muhahahahaha |
| 08:35:16 | vosque | oi |
| 08:35:30 | GoRK | whoot touchscreen works albeit upside down |
| 08:35:44 | vosque | no kidding? |
| 08:35:45 | vosque | hahaha |
| 08:35:50 | vosque | that's kinda cool. |
| 08:35:52 | GoRK | yeah lemme invert it |
| 08:36:00 | vosque | it's like a puzzle game. |
| 08:36:06 | GoRK | no i mean you touch upper left it draws the cursor in the lower right |
| 08:36:22 | GoRK | this toolkit actualy has a feature that will rotate the display 90 or 180 degrees |
| 08:36:58 | vosque | yeah |
| 08:36:59 | vosque | I got it |
| 08:37:05 | vosque | You have to guess where to put the stylus |
| 08:37:12 | vosque | Unintentional puzzle game |
| 08:37:24 | vosque | not a very complex one, mind you. |
| 08:37:45 | vosque | bothered his neck by trying to figure out how to make X go upright. |
| 08:40:47 | GoRK | i have to load up the calibration program to get the coffs then do a transform |
| 08:47:05 | vosque | is picogui what the agenda uses, or is that something else? |
| 08:54:56 | GoRK | umm i know picogui will run on the agenda |
| 08:55:19 | GoRK | i like it because it's so absolutely small and themeable |
| 08:56:20 | GoRK | plus it forces all your programs to look the same |
| 08:57:04 | vosque | <nod> |
| 08:57:11 | vosque | I'll go take a look at it.. |
| 08:57:43 | GoRK | hmm |
| 08:58:09 | GoRK | do you know offhand if the lcd supports any kind of hardware calibration or do you have to do it in software? |
| 08:59:37 | vosque | hmm.. |
| 08:59:39 | vosque | don't know offhand. |
| 08:59:54 | vosque | I got mine on wednesday, just learning how the thing works |
| 08:59:55 | GoRK | arent you working on X stuffs? |
| 09:00:03 | vosque | I got it working |
| 09:00:11 | vosque | I'm getting QPE to run right now |
| 09:00:22 | GoRK | did you port the Microwindows calibrator program? |
| 09:00:36 | GoRK | hmm i guess i should just look at that |
| 09:37:20 | vosque | hey |
| 09:37:23 | vosque | a perl interface |
| 09:37:23 | vosque | neat |
| 12:46:33 | vosque | hmm |
| 12:46:44 | vosque | got QPE to run, doesn't like the touchscreen much |
| 16:37:35 | BZFlag | anyone else try the buildroot out? |
| 17:17:53 | GoRK | i use it; but i guess you probably knew that already ... |
| 17:18:03 | GoRK | heh |
| 17:58:55 | BZFlag | did you get udhcpc to work yet? |
| 18:00:35 | GoRK | me? |
| 18:05:57 | andersee | BZFlag: me? |
| 18:08:35 | GoRK | i think it was compiling bt not getting an address |
| 18:09:55 | GoRK | this touchscreen driver is confusing me. nowhere in the kernel driver is it documented what raw_max_x actually means. ugh |
| 18:15:56 | GoRK | oh well at least its working |
| 18:26:28 | pattieja | is back (gone 19:51:13) |
| 18:29:59 | pattieja | is away: testing latency |
| 18:30:59 | pattieja | is back (gone 00:00:59) |
| 18:31:22 | pattieja | is away: testing latency |
| 18:31:58 | pattieja | is back (gone 00:00:36) |
| 18:33:59 | pattieja | is away: testing latency |
| 18:35:04 | pattieja | is back (gone 00:01:05) |
| 18:39:04 | GoRK | woot! got touchscreen calibration in picogui |
| 19:06:45 | GoRK | sweet picogui is almost ready now |
| 19:11:15 | mallum | ready for tuxscreen ? |
| 19:20:07 | GoRK | yeah i got the touchscreen driver pretty much finished except for being able to recalibrate on the fly |
| 19:20:46 | GoRK | all apps that i have tried so far work fine. all the themes except aqua work fine. There is not yet jpeg support (that is what im working on now) |
| 19:22:21 | BZFlag | well, udhcpc compiles and all, but I can't get it to do anything. |
| 19:23:18 | GoRK | BZFlag: do you have the kernel patch that lets usb1200-ts driver control the brightness and contrast? it must be used in the microwindows distribution but i dont find any kernel source for it |
| 19:24:34 | mallum | GoRK: did you build against uClibc ? |
| 19:24:39 | GoRK | yes |
| 19:24:41 | mallum | cool |
| 19:24:50 | mallum | thats great :) |
| 19:25:28 | GoRK | the mouse cursor leaves a little garbage on the display when moving around - i am also checking into how to fix that. i think i am polling to fast |
| 19:26:25 | Aelhaz | You could make a fine argument for getting rid of the pointer while the touchscreen is active |
| 19:26:34 | Aelhaz | since you're holding a pointer in your hand |
| 19:26:42 | GoRK | or the pointer is your hand |
| 19:26:48 | GoRK | ok that is a built in feature. i'll just use it |
| 19:27:15 | GoRK | but the pointer will probably still leave garbage when i use remote pointer or something |
| 19:27:25 | GoRK | i wont know that until i have a nic though |
| 19:27:37 | Aelhaz | *nod* |
| 19:34:48 | BZFlag | GoRK: I have a ts patch that hangs the kernel. I have not seen one that works yet. |
| 19:38:57 | GoRK | when i run the microwindows distribution, it is a lot brighter for some reason. is there something different that it does? |
| 19:40:05 | GoRK | it also seems to be able to turn the backlight on and off completely when it goes into the screen saver |
| 19:42:10 | GoRK | +#ifdef CONFIG_SA1100_SHANNON |
| 19:42:10 | GoRK | +if (machine_is_shannon()) |
| 19:42:10 | GoRK | +ucb1200_set_io(UCB_GPIO_BACKLIGHT, 1); |
| 19:42:11 | GoRK | +#endif |
| 19:42:43 | GoRK | +if (machine_is_shannon()) { |
| 19:42:43 | GoRK | +ucb1200_set_io((UCB_GPIO_BRIGHT_MASK & bl) << UCB_GPIO_BRIGHT, 1); |
| 19:42:45 | GoRK | +ucb1200_set_io((UCB_GPIO_BRIGHT_MASK & ~bl) << UCB_GPIO_BRIGHT, 0); |
| 19:43:08 | GoRK | +counter_bl = bl; |
| 19:43:09 | GoRK | +ucb1200_set_io((UCB_GPIO_CONTRAST_MASK & contrast) << UCB_GPIO_CONTRAST, 1); |
| 19:43:09 | GoRK | +ucb1200_set_io((UCB_GPIO_CONTRAST_MASK & ~contrast) << UCB_GPIO_CONTRAST, 0); |
| 19:43:15 | GoRK | +counter_contrast = contrast; |
| 19:43:22 | GoRK | those are in one of russ's diffs |
| 19:43:42 | GoRK | diff-tux1-tux2.gz |
| 19:50:02 | GoRK | hmm is that only for 2.4.6 |
| 19:55:11 | GoRK | the microwindows wm seems to be turning the backlight on and off with this: |
| 19:55:53 | GoRK | int fd = open("/dev/ts", 0); ioctl(fd, 64, 1); |
| 20:11:31 | pattieja | is away: I'm busy |
| 20:11:45 | pattieja | is back (gone 00:00:02) |
| 20:20:59 | GoRK | hmm now the cursor hides but still draws garbage |
| 20:58:50 | Russ | thanks andersee |
| 20:59:54 | GoRK | russ! |
| 21:00:44 | GoRK | how does microwindows turn off the backlight and set the brightness all the way up? in the source i see it using ioctl 64 on the /dev/ts but i dont see anything in your kernel patches |
| 21:01:02 | GoRK | the changelog mentions some patch you had.. |
| 21:01:05 | andersee | Russ: for what? |
| 21:01:15 | BZFlag | I told him about the build tree |
| 21:01:18 | Russ | its a seperate patch I sent cosmicpenguin |
| 21:01:22 | Russ | andersee: lockf |
| 21:01:29 | andersee | Ahh. no problem |
| 21:02:03 | Russ | I need to add then and update to the latest kernel |
| 21:02:05 | andersee | I stuff udhcp into the build system -- I don't have a nic so I don't know if I missed anything (config scripts for example) |
| 21:02:27 | BZFlag | Russ: it runs, but: Moreton Bay DHCP Client (v0.9.5) started |
| 21:02:27 | BZFlag | script /usr/share/udhcpc/default.script failed: No such file or directory |
| 21:02:27 | BZFlag | Sending discover... |
| 21:02:28 | Russ | yah, it needs scripts to do anything (other than talk to the server) |
| 21:02:30 | GoRK | ahh you're awesome. i am straining to see my widgets. thanks a bunch |
| 21:02:45 | BZFlag | I don't see anything on the wire. |
| 21:02:55 | Russ | BZFlag: is the interface up? |
| 21:03:00 | Russ | (ifconfig eth0 0.0.0.0 |
| 21:03:03 | Russ | ) |
| 21:03:36 | BZFlag | shows broad/multi but no up. |
| 21:03:44 | Russ | ? |
| 21:04:04 | BZFlag | that did it! |
| 21:04:26 | Russ | also, the options -q -f will probably be helpfull |
| 21:04:40 | Russ | (quit after obtaining lease, remain in foreground after first attempt) |
| 21:05:35 | BZFlag | it should bring the interface up first, no? |
| 21:05:39 | GoRK | ok the only thing left in picogui is getting rid of bits of the mouse pointer that are left behind and proper calibration (which is 90% done anyway) |
| 21:05:41 | Russ | the script does |
| 21:06:02 | GoRK | oh yeah and jpeg support |
| 21:06:36 | Russ | look in sample |
| 21:06:52 | Russ | its three scripts there, but it can be just one |
| 21:07:14 | Russ | er, samples/ |
| 21:07:20 | BZFlag | the scripts are called after getting a lease no? |
| 21:07:51 | Russ | at start, after getting a lease, after renewing a lease, and after losing a lease |
| 21:08:17 | Russ | so $1=deconfig means that the scripts should do something like: |
| 21:08:24 | BZFlag | ah, but they will not get a lease if eth0 is not brought up first. |
| 21:08:25 | Russ | ifconfig $interface 0.0.0.0 |
| 21:08:31 | Russ | ? |
| 21:09:14 | BZFlag | I tried running the client and did not get a lease. I then did a ifconfig eth0 0.0.0.0 up and then ran udhcpc and DID get a lease. |
| 21:09:40 | Russ | probably because you had it set to some IP that wasn't on the same network |
| 21:10:00 | BZFlag | it was down. no ip set at all. |
| 21:10:17 | Russ | oh, yah, it can't send anything when the interface is down |
| 21:10:36 | BZFlag | hence my comment above: <BZFlag> it should bring the interface up first, no? |
| 21:11:01 | Russ | yah, just take a look at samples/ |
| 21:11:10 | Russ | make install puts them in their proper places |
| 21:11:12 | BZFlag | we are in a loop...... |
| 21:11:19 | Russ | and they can easily be combined into one scrip |
| 21:11:20 | Russ | t |
| 21:11:25 | GoRK | break; |
| 21:12:11 | BZFlag | the scripts run _after_ getting a lease. we will _not_ get a lease cause the interface is down, you tell me to fix the scripts which will not be run cause the interface is down and we will not get a lease... |
| 21:12:46 | Russ | it runs a script at start, as well as after losing a lease ($1=deconfig) |
| 21:13:02 | BZFlag | ah! that's what I was missing. ok. |
| 21:13:18 | Russ | just play a bit in udhcp/samples/ |
| 21:13:23 | BZFlag | but, deconfig should down the interface no? |
| 21:13:30 | GoRK | <-- handy with break; |
| 21:13:33 | Russ | it deconfigures it |
| 21:13:34 | BZFlag | or is there a down action? |
| 21:13:47 | Russ | in that it takes away its IP address |
| 21:14:09 | Russ | so, ifconfig $interface 0.0.0.0, which has the side effect of bringing it up if its down |
| 21:14:19 | BZFlag | what happens when the client exits? it should down the interface then. |
| 21:14:31 | Russ | not if its for an initrd |
| 21:14:58 | BZFlag | true, but it supports other no? |
| 21:15:25 | Russ | it basicaly tries to implement as much mechanism as possible, with as little policy |
| 21:16:01 | Russ | so the choice of bringing down the interface when udhcp is killed is yours |
| 21:16:49 | Russ | usefull for odd behavior, like determining the default gw on a network, sending a release, and then giving yourself the default gw address |
| 21:17:26 | Russ | so if you have some router device, you can sets its IP by changing the GW on your DHCP server |
| 21:18:04 | Russ | or in the initrd example, you can use it for getting a lease, and using the IP for NFS, letting a dhcp client on the NFS take over later |
| 21:18:18 | BZFlag | is there anyway to generate a DHCPRELEASE message then? |
| 21:18:35 | Russ | SIGUSR...was it one or two... |
| 21:18:36 | Russ | one sec |
| 21:19:14 | Russ | udhcpc also responds to SIGUSR1 and SIGUSR2. SIGUSR1 will force a renew state, |
| 21:19:14 | Russ | and SIGUSR2 will force a release of the current lease, and cause udhcpc to |
| 21:19:14 | Russ | go into an inactive state (until it is killed, or receives a SIGUSR1). |
| 21:21:15 | BZFlag | where does all the lease info go? ie: does it save next-server or filepath? |
| 21:21:31 | Russ | its all in env vars available to the script |
| 21:21:47 | BZFlag | cool. thanx for the updates... /me goes to hack on it. |
| 21:21:58 | BZFlag | try out the buildroot when you get a chance. |
| 21:22:19 | Russ | right now, just dhcp options are given |
| 21:22:30 | BZFlag | those are dhcp options. |
| 21:22:35 | Russ | not bootp? |
| 21:22:42 | BZFlag | both |
| 21:22:51 | Russ | then its not a dhcp "option" |
| 21:23:03 | BZFlag | afaik all bootp options are dhcp options and then some. |
| 21:23:07 | Russ | dhcp options go after all the bootp stuff |
| 21:23:27 | BZFlag | but they are still valid dhcp options. |
| 21:23:27 | Russ | (and can also be stored in unused bootp fields) |
| 21:23:42 | Russ | "dhcp option" has a specific meaning |
| 21:23:50 | Russ | it has an option code, etc |
| 21:24:08 | Russ | the bootp stuff just sits at predetermined locations in the header |
| 21:24:45 | Russ | I can pass thet stuff if you use it for something though, but there are already relevant dhcp options for NFS root |
| 21:27:30 | Russ | dhcp: http://www.dhcp.org/rfc2131.html |
| 21:27:32 | GoRK | dhcpc.o: In function `pidfile_acquire': |
| 21:27:33 | GoRK | dhcpc.o(.text+0x1e4): undefined reference to `lockf' |
| 21:27:40 | BZFlag | ahh... sname and file can both be used for "dhcp options" ic. |
| 21:27:40 | Russ | option list: http://www.dhcp.org/rfc2132.html |
| 21:27:54 | GoRK | I thought that was fixed. i rm -rf uClibc and sources/uclibc* and remade... |
| 21:28:04 | Russ | right, if you need sname, file, and next-server, I can pass those |
| 21:28:14 | Russ | GoRK: did you do make install? |
| 21:28:22 | Russ | (on uclibc) |
| 21:28:30 | GoRK | this is in the build system |
| 21:28:30 | Russ | and is it getting CVS |
| 21:29:07 | Russ | right, but I would imagine stuff is linked against /usr/arm-uclibc-linux |
| 21:29:19 | BZFlag | I thought I'd setup /linuxrc to pivot_root based on dhcp reply. I was planning on using sname for the server and file for the path to chroot to. |
| 21:29:22 | GoRK | ok i think i got it |
| 21:29:24 | Russ | er, arm-linux-uclibc |
| 21:29:43 | GoRK | i didnt remove my staging_dir |
| 21:29:44 | BZFlag | GoRK: rm -rf root then make. |
| 21:29:49 | Russ | serverid- The IP of the server |
| 21:29:53 | BZFlag | oh, yeah. that too. ;-) |
| 21:29:58 | Russ | rootpath- The path name of the client's root disk |
| 21:30:13 | BZFlag | serverid is the id of the _dhcp_ server, no? |
| 21:30:16 | Russ | er |
| 21:30:56 | Russ | serverid, right is the ip of the server |
| 21:33:07 | Russ | what about temporily using the tftp server option until I pass nextserver in a ENV |
| 21:33:27 | GoRK | Russ: BZFlag said there was a brightness/contrast patch that he had that locked the system... Will this one do that or does it work like it should? |
| 21:34:02 | Russ | if it locks things, its not something I know about |
| 21:34:19 | GoRK | ok. cool. i am about to fire it up |
| 21:34:43 | Russ | anyway, I'm going to drive down to the mall and see if I can donate blood |
| 21:34:44 | Russ | bbiaw |
| 21:34:56 | GoRK | thanks for the patch |
| 21:35:51 | BZFlag | Russ: I'll find something. |
| 21:36:06 | BZFlag | I suppose I could add myself to the commit list and post the changes. ;-) |
| 21:40:59 | GoRK | aww yeah i can SEE THE SCREEN NOW |
| 21:43:56 | cory | woohoo, I just got 18 girls' email addys |
| 21:44:19 | cory | 18 ku freshmen on my floor |
| 21:45:57 | cory | I mean 17 |
| 22:10:22 | GoRK | is there a kernel commandline or can busybox do something to disable console blanking or do i have to fiddle with console.c in the kernel to turn it off? |
| 22:14:53 | Aelhaz | i think that's configurable in proc |
| 22:14:57 | Aelhaz | I forget *where |
| 22:17:40 | Russ | cory: umm...you could just go talk to them |
| 22:17:57 | cory | I talked to them to get their email addresses |
| 22:18:50 | Aelhaz | So what are you doing on IRC? |
| 22:18:59 | cory | I had the female ra with me for some legitimacy.. but that also means I'll have to follow ku privacy policies |
| 22:19:03 | cory | it'll be weird |
| 22:19:08 | cory | we're setting up a floor email list |
| 22:19:14 | cory | I'm writing the first email right now |
| 22:19:23 | Russ | just add them to your buddy list |
| 22:19:29 | cory | hehe |
| 22:19:39 | cory | there are a few aol emails in there |
| 22:23:44 | BZFlag | Russ: I think for every option udhcp receives that it does not understand we should set option<num>=contents if all characters are isprint(). thoughts? |
| 22:24:23 | Russ | you don't really know how to proccess it |
| 22:24:37 | Russ | is it a string of n characters? is it a 4 byte IP? |
| 22:24:43 | Russ | is it a pair of 4 byte IP's? |
| 22:24:51 | Russ | is it a list of n integers? |
| 22:24:57 | Russ | its just as easy to add them in |
| 22:25:26 | BZFlag | hence the isprint() ;-) |
| 22:25:32 | BZFlag | but ok. |
| 22:26:13 | Russ | isprint? |
| 22:26:58 | BZFlag | check the characters to determine if they are all printable. if so, then set then assume string. |
| 22:27:23 | Russ | a lot of IP's are printable strings |
| 22:27:48 | BZFlag | yeah, but having that in the env will not harm anything if they are printable. |
| 22:29:07 | Russ | all the options are documented, we can even #ifdef extension sets |
| 22:30:36 | BZFlag | true. |
| 22:35:40 | BZFlag | would be nice to handle "Static Route Option". thoughts on that for env? /me notices that masks are not specified, and hence broken by default. |
| 22:36:22 | Russ | lemme take a look |
| 22:38:47 | Russ | its a IP_PAIR |
| 22:38:53 | Russ | so it just needs to be added to the list |
| 22:38:58 | Russ | but yes, its kinda broken |
| 22:39:40 | GoRK | narf. i have to code keyboard support into the picogui input driver also |
| 22:40:17 | Russ | picogui doesn't work with keyboards? |
| 22:40:46 | GoRK | no it does just the input drivers are unified. |
| 22:41:17 | GoRK | i dont think it has ever been used on a machine with a keyboard that didn't have SDL or SVGAlib support |
| 22:42:05 | GoRK | so my tuxinput driver has to take over the keyboard as well... which is fine by me |
| 22:42:08 | Russ | it does it through the 2.4 input interface |
| 22:42:32 | Russ | so if you want to just link directly to that, instead of wait for it to come out of the keyboard driver, you can do that |
| 22:43:44 | GoRK | well first i have to see how it even pases key events from the input driver |
| 22:44:24 | Russ | what do you mean? |
| 22:44:39 | Russ | my keyboard driver sends all its events to the input driver |
| 22:44:57 | Russ | and the the keyboard input thingy picks them up and passes them along |
| 22:46:13 | GoRK | no i mean for picogui |
| 22:46:24 | cworth | ah, it's lovely to be on IRC with my tuxscreen. :) |
| 22:46:58 | cworth | now, if only I didn't get dropped charaters perioically. Anyone else have problms with that? |
| 22:48:35 | Aelhaz | yup |
| 22:48:38 | mallum | yup |
| 22:48:46 | Aelhaz | but i did under inferno, too |
| 22:49:00 | GoRK | the other little problem im having is giving picogui exclusive access to the framebuffer. i can still see a blinking cursor on top of my apps and the screen can (and will) fill with junk that the kernel prints |
| 22:49:06 | cworth | maybe I just need new batteries in the keyboard? I'll haveto experiment ater, (I've got another keybard at work I can trytoo) |
| 22:49:14 | GoRK | plus the console blanking |
| 22:49:26 | cworth | hrm... bummer. :( |
| 22:50:05 | Russ | running emacsOS I see |
| 22:50:12 | mallum | cworth: are you run IRC via X ? |
| 22:50:20 | mallum | s/run/running/ |
| 22:50:48 | cworth | mallum: yes IRC in zenirc in emacs, cross-displayed to my tux. |
| 22:52:16 | mallum | cworth: heh nice , I've never tried irc in emacs before |
| 22:52:39 | cworth | mallum: I cant do text outside of emacs I think. |
| 22:54:10 | mallum | cworth: did you end up building your own X for xstroke dev ? |
| 22:56:33 | cworth | mallum: I used to just native-compile with my iPAQ, (Debian/arm over NFS). But yes, I started cross-compiling X/xstroke after keithp started putting new tricks in Render for me. |
| 22:59:07 | BZFlag | sweet. udhcpc script is now in cvs. I need to update pcmcia to call it correctly. |
| 22:59:12 | BZFlag | bbiaw |
| 23:00:25 | cworth | more to the topic, xstroke seems quite coarse on the tuxscreen. How often do we get ts updates? |
| 23:00:49 | Russ | I just used the ucb ts driver |
| 23:00:54 | mallum | cworth: yes I need to figure out a dev enviroment to add some render stuff to xkbd |
| 23:02:28 | zmallum | discovers zenirc .... |
| 23:03:41 | cworth | mallum: it may be that ERC is actually the better emacs IRC client. |
| 23:04:34 | mallum | cworth: can zenirc do multiple channels ? |
| 23:05:23 | cworth | mallum: I do several, (all in the same buffer). Don |
| 23:05:28 | cworth | don |
| 23:05:50 | cworth | argh... stupid keyboard :) |
| 23:06:15 | mallum | cworth: but not in different buffers ? |
| 23:06:15 | cworth | I don't know if it can put separate channels in separate buffers or not. |
| 23:07:32 | cworth | mallum: I've never really needed it, but it shouldn't be too much elisp hacking if it's not there already. |
| 23:08:22 | mallum | shudders away from elisp ... |
| 23:19:30 | cworth | mandrake: greetings |
| 23:19:37 | cworth | oops |