| 01:29.00 | *** join/#edev Sienna (n=Sienna@pool-71-246-71-7.bltmmd.east.verizon.net) |
| 01:29.39 | *** join/#edev agaffney (n=agaffney@gentoo/developer/pdpc.active.agaffney) |
| 01:49.04 | *** join/#edev Mesousa (n=Mohamdu@unaffiliated/hellomeow) |
| 01:59.31 | *** join/#edev agaffney (n=agaffney@gentoo/developer/pdpc.active.agaffney) |
| 02:08.29 | *** join/#edev T0mW (n=Tom@24.115.218.135.res-cmts.sth.ptd.net) |
| 02:24.48 | *** join/#edev geekosis (n=Xerox@h24-207-7-50.primesignal.com) |
| 02:44.44 | *** part/#edev geekosis (n=Xerox@h24-207-7-50.primesignal.com) |
| 03:09.00 | *** join/#edev agaffney (n=agaffney@gentoo/developer/pdpc.active.agaffney) [NETSPLIT VICTIM] |
| 03:09.26 | *** join/#edev chouimat|Zzzz (n=dieu@r2351064.cidc.net) |
| 03:11.42 | *** join/#edev Shingoshi (n=shingosh@c-67-169-195-234.hsd1.or.comcast.net) [NETSPLIT VICTIM] |
| 03:11.42 | *** join/#edev huaynac_ (n=huaynac@61.6.65.198) [NETSPLIT VICTIM] |
| 03:15.30 | *** join/#edev Shingoshi (n=shingosh@c-67-169-195-234.hsd1.or.comcast.net) |
| 03:22.55 | *** join/#edev RobotGuy (i=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 03:24.52 | RobotGuy | ~seen Sienna |
| 03:24.55 | purl | sienna <n=Sienna@pool-71-125-185-199.bltmmd.east.verizon.net> was last seen on IRC in channel #edev, 23h 40m 24s ago, saying: 'goodnight boys! I am going to go snuggle into the couch and continue reading'. |
| 03:58.38 | *** join/#edev DrZimmerman (n=theo@tbaumgartner.broker.freenet6.net) |
| 05:36.19 | *** join/#edev harpal (n=harpal@124.125.79.150) |
| 05:46.12 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 05:48.13 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 05:59.32 | *** join/#edev drath_ (i=vmaster@p5B07F10B.dip.t-dialin.net) |
| 06:46.19 | *** join/#edev howtoo (n=howtoo@195.159.62.66) |
| 07:20.06 | *** join/#edev chouimat|Zzzz (n=dieu@r2351064.cidc.net) |
| 07:28.28 | *** join/#edev [AD]Turbo (n=turbo@host131-128-dynamic.57-82-r.retail.telecomitalia.it) |
| 07:29.29 | [AD]Turbo | ciao all |
| 07:32.19 | RobotGuy | Greetings |
| 07:44.53 | *** join/#edev harpal (n=harpal@124.125.79.150) |
| 07:54.19 | *** join/#edev magnetron (n=magnetro@unaffiliated/magnetron) |
| 08:10.25 | *** join/#edev harpal (n=harpal@124.125.79.150) |
| 08:27.43 | *** join/#edev harpal_ (n=harpal@124.125.79.153) |
| 08:38.18 | *** join/#edev harpal__ (n=harpal@124.125.79.150) |
| 08:50.32 | *** join/#edev harpal_ (n=harpal@124.125.79.153) |
| 09:10.43 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 09:22.58 | *** join/#edev lyakh (n=lyakh@p57BD2BC0.dip0.t-ipconnect.de) |
| 10:00.23 | *** join/#edev samuel1 (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 10:41.12 | *** join/#edev samuelr (n=samuelr@dsl093-061-094.pit1.dsl.speakeasy.net) |
| 11:20.15 | *** join/#edev ali_as (n=as@ambix.plus.com) |
| 11:36.21 | *** join/#edev Appleman1234 (n=Appleman@CPE-124-186-229-83.qld.bigpond.net.au) |
| 11:54.57 | *** join/#edev hw_ (n=hw@p578b3905.dip0.t-ipconnect.de) |
| 12:00.24 | *** part/#edev Mesousa (n=Mohamdu@unaffiliated/hellomeow) |
| 13:01.44 | *** join/#edev GPSFan (n=kenm@12.10.255.246) |
| 13:01.44 | *** mode/#edev [+o GPSFan] by ChanServ |
| 13:04.14 | *** join/#edev chouimat|work (n=dieu@209.217.106.98) |
| 13:24.35 | chouimat|work | morning |
| 13:31.40 | *** join/#edev prpplague (n=dave@mail.americanmicrosystems.com) |
| 13:34.27 | *** mode/#edev [+o prpplague] by ChanServ |
| 13:49.26 | prpplague | ho ho hum |
| 13:49.36 | prpplague | time to nag bjdooks about the patches |
| 14:05.06 | *** join/#edev dijenerate (n=dijenera@205.244.151.65) |
| 14:06.46 | *** join/#edev adyer (n=adyer@adsl-69-223-13-78.dsl.chcgil.ameritech.net) |
| 14:17.22 | prpplague | interesting post - http://tech.slashdot.org/tech/08/04/07/0137208.shtml |
| 14:18.20 | prpplague | funny i consider myself very introverted, but i much pref irc over emial or forums |
| 14:18.45 | chouimat|work | prpplague: same for me but we are "functional introvert" ;) |
| 14:19.03 | prpplague | chouimat|work: hmm |
| 14:19.14 | chouimat|work | bad joke :) |
| 14:19.16 | prpplague | chouimat|work: and i don't care for social networking sites at all |
| 14:19.55 | chouimat|work | prpplague: http://labs.trolltech.com/blogs/2008/04/07/free-usb-hub-with-coffee-cup-holder/ |
| 14:20.06 | prpplague | looks |
| 14:21.09 | prpplague | chouimat|work: cute |
| 14:21.31 | chouimat|work | I need one :) |
| 14:21.34 | chouimat|work | maybe not |
| 14:22.24 | ali_as | prp, Ben has a 117 hour idle time, I'm hoping to catch him myself to ask about the ethernet driver he rewrote. |
| 14:23.07 | prpplague | ali_as: lets just say he tends to be very hard to contact |
| 14:24.26 | Mirell | It's too early in the morning to work... |
| 14:25.12 | ali_as | It took simtec support 24 days to get the first reply to my query, so hopefully it's all relative. |
| 14:26.37 | ali_as | I am just so tempted to jump to the ts-7800, but then I lose any chance I could make a working prototype of the final design myself. |
| 15:00.31 | *** join/#edev harpal (n=harpal@124.125.79.150) |
| 15:08.07 | *** join/#edev emil|k (i=emil@e179233074.adsl.alicedsl.de) |
| 15:09.08 | prpplague | ali_as: what is it you are doing? |
| 15:10.40 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 15:10.41 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 15:12.03 | prpplague | CosmicPenguin: greetings |
| 15:15.24 | CosmicPenguin | hi |
| 15:17.25 | prpplague | CosmicPenguin: hows biz? |
| 15:17.34 | CosmicPenguin | same ol |
| 15:17.44 | prpplague | ahh\ |
| 15:18.33 | ali_as | prpplague, jtag project, multiplatform. |
| 15:19.30 | prpplague | ali_as: you using one of simtec's cpu boards? |
| 15:20.13 | ali_as | Yes, the oki board. |
| 15:21.47 | ali_as | On paper it was perfect, 100Mbit ethernet, 60MHz ARM7 and a CPLD so no bitbanging. |
| 15:22.59 | ali_as | http://www.simtec.co.uk/products/EB675001DIP/ for anyone interested. |
| 15:25.20 | ali_as | But wget'ing a file to /dev/null maxes out at 880 kilobytes a second, so my goal of one megabyte a second through to and from the target device is starting to seem unreachable. |
| 15:27.06 | SpeedEvil | sure the thing is autobauding properly? |
| 15:27.21 | *** join/#edev T0mW (n=Tom@24.115.218.135.res-cmts.sth.ptd.net) |
| 15:27.41 | SpeedEvil | syslogs or lights on the far end may tell you. |
| 15:28.04 | ali_as | Actually, no, but both lights are on so I'm assuming 100Mbit. |
| 15:28.40 | ali_as | I've looked at the syslog and it doesn't tell me. |
| 15:30.05 | ali_as | But by the strangest coincidence to happen to me this year by a mile, bjdooks, who wrote the driver hangs out in #openjtag so when he's active I hope to ask him a few things. |
| 15:30.15 | SpeedEvil | :) |
| 15:33.40 | SpeedEvil | The 'gold' board? |
| 15:33.47 | SpeedEvil | Looks interesting |
| 15:34.33 | prpplague | gold board? |
| 15:35.02 | SpeedEvil | one of the variants of the above board |
| 15:35.10 | SpeedEvil | got 32M SDRAM, the others have nothing |
| 15:35.11 | ali_as | Gold, yes. |
| 15:35.50 | SpeedEvil | According to the FAQ, you should get 'DM9000: dm0: r1, 00:11:ac:00:06:00 int phy, link ok, 100Mbit full duplex' |
| 15:36.02 | SpeedEvil | or similar from the bootloader if it's 100mbit |
| 15:36.43 | ali_as | I think I get confirmation from the bootloader, but uclinux isn't telling me what it's using. |
| 15:37.24 | SpeedEvil | k |
| 15:38.32 | ali_as | eth0, gives me an IRQ, MAC address and the name of the driver. |
| 15:38.56 | ali_as | It's not even giving me link up/down messages. |
| 15:39.10 | SpeedEvil | I'd look at the link lights on the far end if poss. |
| 15:39.18 | ali_as | Both lit. |
| 15:39.25 | SpeedEvil | ah |
| 15:39.35 | SpeedEvil | what does it do the other way? |
| 15:39.38 | SpeedEvil | sending that is |
| 15:40.03 | ali_as | I don't know, the default kernel doesn't seem to have any program for sending. |
| 15:40.14 | ali_as | It's a bit... sparse. |
| 15:40.22 | SpeedEvil | is it stock wget? |
| 15:40.36 | ali_as | Its part of busybox I think. |
| 15:40.57 | SpeedEvil | sshs into phone. |
| 15:40.58 | ali_as | With wput not included. |
| 15:41.06 | prpplague | i really hate git |
| 15:43.00 | chouimat|work | prpplague: stop spreading the heresy :) |
| 15:49.18 | CosmicPenguin | don't hate what you don't understand |
| 15:49.48 | chouimat|work | CosmicPenguin: it's part of the human nature :) |
| 15:51.04 | CosmicPenguin | First understand it, then you can hate it |
| 15:52.09 | prpplague | CosmicPenguin: i understand how to use it, it does't perform the way i like |
| 15:52.17 | prpplague | CosmicPenguin: i much prefer cvs |
| 15:52.37 | prpplague | CosmicPenguin: but my opinion isn't worth much |
| 15:58.29 | Mirell | I could really deal without ever having to see Perforce again... |
| 16:16.36 | *** join/#edev ico2 (n=ico2@modem-151.cheetah.dialup.pol.co.uk) |
| 16:17.02 | Dr{Who} | prplague.. i have been tring last week to get gadget fs to work so far no luck. when i load the module and say file=/somefile.dat the driver faults.. I have not gotten a good stack trace yet because my kernel does not have enough debug data in it.. A few questions.. i have never debugged a kernel with a jtag port but would like to learn.. do we have docs / wiki data on that ? |
| 16:21.54 | *** join/#edev dijenerate (n=dijenera@205.244.151.65) |
| 16:22.21 | *** join/#edev lyakh (n=lyakh@p57BD2BC0.dip0.t-ipconnect.de) |
| 16:29.18 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 16:33.28 | prpplague | RobotGuy: hey, which pin did you break? |
| 16:33.40 | RobotGuy | prpplague: RXD0 |
| 16:34.52 | RobotGuy | I don't see how I could have danaged the module unless I got a tiny solder bridge when soldering the jumper wire. |
| 16:34.54 | prpplague | RobotGuy: console port uart |
| 16:34.58 | RobotGuy | prpplague: Yep |
| 16:35.37 | RobotGuy | I get no response from the module now, but JTAG flashing seems to work (no errors). |
| 16:36.33 | prpplague | RobotGuy: of course you won't since the console pin is broken |
| 16:36.59 | RobotGuy | I should get output through TXD0 though. |
| 16:38.18 | RobotGuy | I tried to fix it or at least make it usable again, but failed and the module doesn't work. |
| 16:39.04 | RobotGuy | Now I remember why I don't care much for hardware interfacing. |
| 16:39.21 | RobotGuy | I don't know if the module is recoverable or not. |
| 16:39.24 | prpplague | RobotGuy: are you using the bt module? or using the db9? |
| 16:39.41 | RobotGuy | Bluetooth tied to console. |
| 16:40.12 | RobotGuy | Hmmmm, wonder if Hammer is trying to talk to BT. |
| 16:40.55 | RobotGuy | I should still see output from TXD0 though, regardless. |
| 16:41.27 | prpplague | RobotGuy: not if you have the lines incorrect |
| 16:42.06 | RobotGuy | I double checked the connections for the BT module and they are OK - RX->TX TX-RX GND<>GND 3.3V<>3.3V |
| 16:43.02 | RobotGuy | warms up the soldering iron |
| 16:55.24 | RobotGuy | prpplague: I have output from the Hammer. |
| 16:55.32 | prpplague | RobotGuy: lovely |
| 16:55.38 | prpplague | RobotGuy: what was the problem? |
| 16:55.53 | RobotGuy | I think it has something to do with the BT, but not 100% sure yet. |
| 16:57.24 | RobotGuy | I can even enter stuff on the console now. |
| 16:59.43 | RobotGuy | Now that I know Hammer is operational, I can start figuring out why the BT on console port seems to be a problem. |
| 17:00.08 | prpplague | RobotGuy: which bt module are you using? |
| 17:00.34 | RobotGuy | prpplague: http://www.sparkfun.com/commerce/product_info.php?products_id=8469 |
| 17:01.17 | *** join/#edev Crofton (n=balister@66-207-66-26.black.dmt.ntelos.net) |
| 17:01.33 | prpplague | RobotGuy: have you tested the module first to see if you have communication? |
| 17:02.16 | RobotGuy | prpplague: I am not sure I would know how to do that. I don't think I have anything to test it with. |
| 17:02.32 | RobotGuy | Yeah, I know I should have a known good test for stuff like this. |
| 17:03.12 | prpplague | RobotGuy: how are you connecting the bt module to the carrier board? |
| 17:04.07 | RobotGuy | prpplague: Socket headers for the BT module to insert into. RX-->TX, TX-->RX, GND<>GND, 3.3V<>3.3V |
| 17:04.46 | RobotGuy | Hmmmm, I wonder if I have to cross the TX/RX lines with this BT module. |
| 17:04.52 | RobotGuy | looks |
| 17:05.12 | prpplague | RobotGuy: i mean which points are soldering to for the rs-232? |
| 17:05.51 | RobotGuy | prpplague: RXD0 -> BT-TX TXD0-> BT-RX |
| 17:05.59 | prpplague | hehehe |
| 17:06.06 | RobotGuy | Wrong? |
| 17:06.13 | prpplague | RobotGuy: do you have the carrier board schematic? |
| 17:07.18 | Mirell | Mmm, the constant battle between production software engineers and development software engineers |
| 17:07.31 | RobotGuy | prpplague: BTW, 4 threads on TCT forums need to go away. |
| 17:07.45 | prpplague | RobotGuy: ok i'll get them |
| 17:08.27 | prpplague | RobotGuy: you have the carrier board schematic? |
| 17:08.39 | RobotGuy | prpplague: Got schematic for carrier board up now. We should have that on the Wiki also. |
| 17:09.05 | prpplague | RobotGuy: tell me the location where you are soldering your tx/rx connections |
| 17:09.26 | prpplague | RobotGuy: for example you could say, J7 pins 1 and 2 |
| 17:10.08 | RobotGuy | prpplague: J9, pins 7 and 8 |
| 17:11.15 | prpplague | RobotGuy: ok that should be ok |
| 17:11.49 | prpplague | RobotGuy: i suspect you probably have something not correct with your bt module |
| 17:12.27 | RobotGuy | prpplague: Seems like I should just need RX and TX crossed and PWR/GND connected. Doesn't seem that difficult. |
| 17:12.55 | prpplague | RobotGuy: still gotta get the bt on your host side connected properly |
| 17:13.24 | RobotGuy | Ahhh, I think I need to jumper RTS and CTS on the BT module. |
| 17:14.04 | RobotGuy | There is also a reset line I many have to deal with. |
| 17:16.30 | *** join/#edev Cesc (n=ffileppo@151.61.154.104) |
| 17:21.03 | RobotGuy | prpplague: I think that was it - had to jumper RTS and CTS on the BT module. |
| 17:21.30 | RobotGuy | prpplague: Hammer is consoling with the BT module installed now. :) |
| 17:21.55 | prpplague | lovely |
| 17:22.58 | ali_as | Robotguy, it's back from the dead? |
| 17:23.02 | RobotGuy | I am not done yet though - have lost RX to the Hammer with the BT installed. |
| 17:23.08 | RobotGuy | ali_as: Hammer lives! |
| 17:23.15 | ali_as | Excellent! |
| 17:23.32 | RobotGuy | I have a jumper wire going from the Hammer module to RXD0 on the prototype area though. |
| 17:23.43 | RobotGuy | Because I broke that pin on the module. |
| 17:23.48 | ali_as | nudges a tiny coffin and a great deal of 'wake' grade alcohol into a corner. |
| 17:25.27 | RobotGuy | prpplague: I also installed socket headers into the strips next to the hammer socket and sd/mmc socket, so I have access to the pins for my solderless breadboard. |
| 17:26.30 | RobotGuy | I have a level shifter I could connect to the BT module. Hmmmm. |
| 17:27.39 | RobotGuy | thinks |
| 17:36.04 | *** join/#edev ico2 (n=ico2@modem-655.chimpanzee.dialup.pol.co.uk) |
| 17:37.48 | RobotGuy | Durn, I can't find my RS-232 level shifter. That's what I get for letting a friend help me reorganize things. I can't find anything! |
| 17:41.03 | Dr{Who} | prpplague.. sorry to ask again but any ideas on kernel faults on g_file_storage s3c2410_udc |
| 17:41.24 | prpplague | Dr{Who}: hmm, i don't recal you asking a first time |
| 17:41.36 | *** join/#edev ce_geek (n=ce_geek@roger-blanchette-4.umelst.maine.edu) |
| 17:41.46 | prpplague | Dr{Who}: got any error logs? |
| 17:41.55 | Dr{Who} | sorry you were afk ya i load the g_file_storage file=/somefile.dat |
| 17:41.57 | Dr{Who} | and it faults |
| 17:41.58 | *** join/#edev DrZimmerman (n=theo@tbaumgartner.broker.freenet6.net) |
| 17:42.16 | Dr{Who} | just the dump..but my call stack does not have any symbols |
| 17:42.33 | Dr{Who} | have to figure that out.. i rebuilt with debug data but it seems to not be showing.. |
| 17:42.48 | Dr{Who} | was going to do some learning on how to kdb with the flyswatter.. |
| 17:43.08 | Dr{Who} | figure its as good as any a time to learn this. never debugged with a jtag before. |
| 17:43.25 | prpplague | Dr{Who}: i assume you have all the other gadget drivers turned off and/or not loaded? |
| 17:44.14 | Dr{Who} | http://pastebin.com/m53fd7312 |
| 17:44.30 | Dr{Who} | let me dbl check. i know i have them all as modules.. |
| 17:44.33 | Dr{Who} | so i dont loaded any |
| 17:47.16 | RobotGuy | prpplague: The site for the microcom package needs to be updated to http://ftp.iasi.roedu.net/mirrors/openwrt.org/sources/ |
| 17:49.43 | prpplague | Dr{Who}: what kind of file are you using? |
| 17:49.44 | Dr{Who} | robot i did a small wiki update to a page you authored.. i adjusted the Hammer_How_to_build_Kernel page. the common size section.. can you verify it when you get a chance. |
| 17:49.57 | Dr{Who} | i just did a dd to create it |
| 17:50.23 | Dr{Who} | i also tried to build one that was a true fat but the last test i did i just dd and created a zero file |
| 17:50.42 | Dr{Who} | figure if it worked i woudl format with the linux box or windows box i plugged it into. |
| 17:50.50 | RobotGuy | Dr{Who}: Greetings. |
| 17:51.06 | Dr{Who} | http://docs.blackfin.uclinux.org/doku.php?id=gadget_file-backed_storage |
| 17:51.23 | Dr{Who} | howdy robotguy. |
| 17:51.47 | RobotGuy | Dr{Who}: Did you have your sale yet? I hope not. |
| 17:52.08 | Dr{Who} | nope : ( too busy U need to come over and take stuff.. : c ) |
| 17:52.28 | prpplague | Dr{Who}: hmm that should work |
| 17:52.43 | prpplague | Dr{Who}: i'll have some time this afternoon to test |
| 17:52.46 | RobotGuy | Dr{Who}: I will have a look at the Wiki. |
| 17:53.06 | Dr{Who} | ok cool. i will try to get debug data working.. i need to see a better backtrace now it just has address data but no symbols. |
| 17:53.52 | Dr{Who} | also if you have any kickstart pages or docs on flyswatter kernel debugging i would appreciate it. |
| 17:55.02 | RobotGuy | Dr{Who}: I think the sizes should actually be 0x200000 and 0x300000 for even megabyte boundaries, but I will have to check. I think I need to make some more updates to that page also. |
| 17:58.49 | RobotGuy | prpplague: Is it a 256K boundary that .size has to be set to? |
| 17:59.13 | prpplague | RobotGuy: its actually 128k |
| 17:59.24 | RobotGuy | Dr{Who}: Be sure you add your successes to the Wiki. :) |
| 17:59.27 | RobotGuy | prpplague: Ah, OK |
| 17:59.38 | prpplague | RobotGuy: but you need to make sure you account for the bootloader and kernel |
| 18:00.03 | RobotGuy | prpplague: I have included that in my update. |
| 18:00.34 | RobotGuy | bootloader + kernel for booting from flash and bootloader + kernel + rootfs if using an initrd. |
| 18:01.30 | Dr{Who} | im using 340000 |
| 18:01.50 | Dr{Who} | its a bit overkill but i plan on adding perl soon and that will take up some room. |
| 18:09.01 | *** join/#edev Madsy (n=madsy@ti500710a342-1506.bb.online.no) |
| 18:09.12 | ce_geek | Holy crap the Wiki has received some face lifts |
| 18:11.51 | RobotGuy | prpplague: When you get a chance, could you check my edits to the building a custom kernel page? |
| 18:12.05 | RobotGuy | ce_geek: We have been busy. :) |
| 18:12.16 | ce_geek | RobotGuy: Great1 |
| 18:12.18 | ce_geek | *Great! |
| 18:12.37 | RobotGuy | ce_geek: Whole new howto for booting from flash. :) |
| 18:13.21 | prpplague | RobotGuy: just a note, the hammer always boots from flash, you doc is about using flash based root filesystem |
| 18:13.44 | RobotGuy | prpplague: Ah, OK. |
| 18:13.46 | ce_geek | http://www.elinux.org/Hammer_Board_HOWTOS <--- is it listed there? |
| 18:13.57 | RobotGuy | ce_geek: Yes, of course |
| 18:14.36 | ce_geek | Which one is it? What's it called? |
| 18:14.56 | Madsy | Somehow, I find the ARM architecture a bit too difficult. I need a pretty simple evalutation board, so I can learn how to setup interrupts, GPIO, cache and write buffers, etc. Can someone here recommend me one? |
| 18:14.58 | prpplague | http://www.elinux.org/Hammer_How_to_Boot_JFFS2_Root_Filesystem |
| 18:14.59 | ce_geek | Oh the JFFS2 thing? |
| 18:15.02 | ce_geek | Ah okay |
| 18:15.15 | Madsy | I currently have a GP2X, which uses a MMSP2 SoC from magiceyes. |
| 18:15.16 | ce_geek | Needed the JFFS2 = flash connection |
| 18:15.27 | prpplague | Madsy: of couse i'm gonna recommend a hammer board, hehe |
| 18:15.39 | *** join/#edev landley (n=landley@adsl-68-93-84-18.dsl.hstntx.swbell.net) |
| 18:15.39 | *** mode/#edev [+o landley] by ChanServ |
| 18:15.58 | RobotGuy | prpplague: On the deletion of spam threads on the forums, I could handle that for you. I and another user did that on the Lynxmotion forums for awhile when spam was bad before they got a spam handler installed. |
| 18:16.05 | RobotGuy | waves at landley |
| 18:16.25 | Madsy | prpplague: If you honestly think that's a good choice, I'll keep that in mind. |
| 18:16.45 | RobotGuy | Madsy: Compare the Hammer specs with what you have. |
| 18:16.59 | Madsy | The current system I have is well-documented, but written in poor japanese-english. |
| 18:17.15 | Madsy | And the features are so many, that I don't even know where to start :-) |
| 18:18.09 | Dr{Who} | looking at the default hammer.ocd i think we have about 200k of wasted space between the boot loader and the kernel |
| 18:18.35 | RobotGuy | Madsy: Do you have a link to what you have now? |
| 18:18.36 | prpplague | Madsy: i;m biased |
| 18:18.50 | RobotGuy | prpplague: Me too, of course. :) |
| 18:18.57 | landley | waves at RobotGuy. |
| 18:19.28 | prpplague | Madsy: the hammer isn't as well documented as we would like, but there is a growing community of developers and open documentation |
| 18:19.31 | RobotGuy | Madsy: What features do you need? |
| 18:19.50 | Madsy | Robotguy: Not the entire spec of the GP2X console, but is uses a MMSP2 Soc which you can find here: http://www.mesdigital.com/english/Products/product_mmsp2.asp |
| 18:20.02 | Dr{Who} | i dont see the boot loader getting any bigger than the 32k it uses ever so maybe we can move the kernel closer to the boot loader.. |
| 18:20.30 | Madsy | RobotGuy: Basically the most basic evaluation board I can get, with an MMU and which runs linux. |
| 18:20.44 | Madsy | RS232 and JTAG is also most welcome. |
| 18:20.48 | Dr{Who} | is it worth it to you prpplague.. to do this on the iso you send out with a rebuilt bootloader etc. |
| 18:20.54 | ce_geek | So what's the advantage of booting from a JFFS2 root filesystem instead of using the initrd in RAM? Using memory somewhere else? |
| 18:21.09 | RobotGuy | Madsy: Hammer has 2x RS-232 + JTAG right on the module. |
| 18:21.18 | landley | DrWho: remember, the bootloader includes environment variables. |
| 18:21.22 | landley | Including your kernel command line. |
| 18:21.29 | prpplague | Dr{Who}: two issues |
| 18:21.46 | landley | Still, that shouldn't go much above 64k total. 256k is being a bit generous... |
| 18:21.46 | prpplague | Dr{Who}: primary one is that the flash has 128k erase blocks |
| 18:22.01 | prpplague | Dr{Who}: which means you can only erase in groups of 128k |
| 18:22.14 | Madsy | RobotGuy: Hmm.. looks like it also has an ARM 920, which is the current CPU I allready code for :-) |
| 18:22.20 | landley | shrugs. Reflash bootloader when you flash a new kernel is fine with me. My script does that anyway. |
| 18:22.20 | prpplague | Dr{Who}: so even though you are only using 32k for bootloader it takes up a 128k block |
| 18:22.24 | Dr{Who} | ya thats a good point forgot about that.. i need to wiki that |
| 18:22.29 | RobotGuy | Madsy: Hammer is also full Linux (kernel 2.6.22 or later) on a 40 pin DIP (32 MB RAM, 16 MB Flash, 200 Mhz Samsung S3C2410 ARM9). |
| 18:22.39 | *** part/#edev Cesc (n=ffileppo@151.61.154.104) |
| 18:22.41 | prpplague | Dr{Who}: second is that bootloader uses a storage area for variables |
| 18:22.42 | Madsy | Looking at the page now. |
| 18:22.43 | RobotGuy | Madsy: Hammer is ARM920T |
| 18:23.12 | prpplague | Dr{Who}: and currently it also ends up using 128k |
| 18:23.14 | RobotGuy | And I am reciting all of this from memory.. That is scary. |
| 18:23.14 | landley | prpplague: 32k of environment space should be plenty for all the variables it's ever likely to use. |
| 18:23.15 | prpplague | Dr{Who}: there are some things you can do to tweak this, but for now, just leaving it at 256k is probably a good idea |
| 18:23.16 | landley | Heck, 8k would probably be plenty... |
| 18:23.27 | landley | Actually 128k. |
| 18:23.30 | Dr{Who} | ok sounds good to me.. |
| 18:23.33 | prpplague | landley: it actually only uses about 256 bytes |
| 18:23.33 | landley | One erase block. Bootloader plus environment. |
| 18:23.42 | landley | prpplague: of environment? |
| 18:23.45 | prpplague | yea |
| 18:23.54 | Madsy | Uh.. the pin layout is the only documentation available? :) |
| 18:23.57 | landley | The linux kernel command line is in there, and users can stick all sorts of gunk into there. |
| 18:24.07 | Madsy | In addition to the ARM processor spec. |
| 18:24.09 | landley | Keep in mind you can set environment variables for PID 1 from the kernel command line. |
| 18:24.10 | RobotGuy | Madsy: There is another channel, #tincantools, if you would like to discuss this more and have a clear channel. |
| 18:24.11 | Dr{Who} | i need to modify my jffs2 creation script to include that erase block size.. |
| 18:24.23 | prpplague | Madsy: are you looking at the wiki page? |
| 18:24.23 | landley | So giving 'em a few k to play with is common courtesy. |
| 18:24.27 | Madsy | Thanks. |
| 18:24.40 | landley | dunno what the bootloader's limit for setting a kernel command line is... |
| 18:24.52 | landley | Vaguely remember a thread about that, long ago. |
| 18:25.00 | landley | Would probably remember more clearly if I wasn't operating on 1 hour of sleep. |
| 18:25.04 | prpplague | landley: the variables i am refering to are for the bootloader |
| 18:25.17 | prpplague | Madsy: the wiki page has alot more details |
| 18:25.17 | Madsy | prpplague: Nope, just tincantool.com |
| 18:25.26 | prpplague | Madsy: click on the documenation link |
| 18:25.53 | Madsy | If this board is proper documented, I'm seriously thinking about buying one. Thanks again. |
| 18:26.06 | RobotGuy | Madsy: http://www.elinux.org/Hammer_Board |
| 18:26.08 | landley | prpplague: but the kernel command line is taken from a bootloader environment variable. |
| 18:26.37 | landley | Madsy: the documentation for hammer and nail are all on elinux.org, not on tincantools.com. Don't ask me why. |
| 18:26.42 | prpplague | Madsy: again i'm biased |
| 18:26.50 | prpplague | landley: correct |
| 18:27.04 | prpplague | landley: not sure what you are driving at |
| 18:28.28 | RobotGuy | Madsy: We are documenting as fast as we can. I'm very active in Wiki work, but I am just a Hammer user, not employed by TinCanTools. |
| 18:28.48 | RobotGuy | I am working on adding bluetooth to the console port now. |
| 18:30.35 | RobotGuy | wishes there was access to the third serial port. |
| 18:31.57 | landley | prpplague: that browsing on the tincantools site doesn't mention things like the nail board. |
| 18:32.02 | prpplague | argh i can't figure out whats happend to this nand flash |
| 18:32.04 | landley | I understand it's not shipping yet... |
| 18:32.23 | landley | is operating on 1 hour sleep. |
| 18:32.31 | landley | Assuming I'm getting at anything may be unwarranted. |
| 18:32.32 | prpplague | landley: if you click on "documenation" it takes you to the hammer wiki pages |
| 18:32.39 | agaffney | landley: that's not operating :P |
| 18:33.08 | agaffney | 1 hour of sleep is walking around like a zombie with only a basic level of functioning |
| 18:33.32 | landley | agaffney: I have to stare at C++ code. |
| 18:33.44 | agaffney | I'm terribly sorry |
| 18:33.51 | agaffney | I wouldn't want to do that on a full night's sleep |
| 18:33.53 | landley | Last week there was a debate of the new corporate email retention policy, and an attempt to resolve it by calculating the cost of compliance. |
| 18:34.16 | landley | This job pays very well, and I'm wondering if I can last the full 5 months of the original contract. |
| 18:34.21 | agaffney | heh |
| 18:34.45 | RobotGuy | prpplague: Is there a problem with having two things connected to /dev/ttySAC0 at the same time? I'd like to be able to monitor bootup through a serial console and have everything go out bluetooth also. |
| 18:34.50 | landley | I'm currently trying to figure out why the exchange webmail logs me out between every click. |
| 18:35.47 | prpplague | RobotGuy: monitoring is fine |
| 18:36.07 | RobotGuy | prpplague: OK, so what I want to do is workable then. |
| 18:36.22 | landley | closes prpplague's LCD link from friday. |
| 18:36.43 | prpplague | landley: hehe which lcd link was that? |
| 18:36.46 | landley | Too expensive for a laptop-like hammer enclosure, by the way. Even if you got 5000 of them, it'd push the total price up over $100. |
| 18:36.47 | RobotGuy | prpplague: Right now Hammer is stopping at booting the kernel. No idea why at present. |
| 18:36.56 | landley | The tuxscreen one. |
| 18:36.59 | landley | http://store.earthlcd.com/EDMGRB8KJF?sc=7&category=148 |
| 18:37.04 | prpplague | landley: ahh no thats retail |
| 18:37.13 | prpplague | landley: you can get them in bulk for less than $25 |
| 18:37.22 | landley | Ooh, that's a lot more attractive. |
| 18:37.23 | adyer | RobotGuy: here's a perverse way to add a serial port - i2c to serial bridge: http://www.standardics.nxp.com/products/bridges/uart.i2c.master.gpio/ |
| 18:37.37 | RobotGuy | adyer: That is not what I am trying to do. |
| 18:37.46 | landley | If you can get the entire system price to total under $150, you've got something. (Under $100, you've got quite something.) |
| 18:38.06 | landley | Up around $200 it starts butting into the eeeeeeeeeeeeepc and olpc and similar. |
| 18:38.23 | RobotGuy | adyer: I am adding bluetooth to Hammer's console port. |
| 18:38.39 | landley | (Wasn't eeepc one of the characters in Voyage of the Dawn Treader? Swashbuckling anthrpomorphic rat that sailed over the edge of the world?) |
| 18:38.47 | landley | Reepicheep. Right. |
| 18:39.25 | RobotGuy | I know, for myself, I have to really justify spending $200.00 on a board or module. |
| 18:39.44 | landley | Of course, adding the world's cheapest 802.11b wireless blows the budget anyway... |
| 18:40.08 | RobotGuy | landley: It would have to be WiFi B/G for me. :) |
| 18:40.43 | *** join/#edev DrZimmerman (n=theo@tbaumgartner.broker.freenet6.net) |
| 18:40.58 | landley | RobotGuy: the entire device has 32 megs of ram and 16 megs of storage. |
| 18:41.14 | RobotGuy | landley: Sounds like Hammer |
| 18:41.14 | landley | 802.11b has the same range as G (last I checked), and 11 megabytes/second data transfer. |
| 18:41.23 | landley | RobotGuy: it is hammer. |
| 18:41.24 | RobotGuy | landley: Speed is what I want. |
| 18:41.43 | landley | On friday, prpplague and I were pondering a laptop-like enclosure for hammer with the world's cheapest keyboard, touchpad, and black and white display. |
| 18:41.51 | landley | And about 32 AA batteries for a week or so of battery life, if at all possible. |
| 18:41.52 | RobotGuy | landley: I have two USB WiFi B/G adapters. |
| 18:41.59 | landley | Ala the old TRS-80 model 100 of yore. |
| 18:42.30 | landley | (If you could install DRAM banks in groups of 4 back in the 386DX days, why not install AA batteries in groups of 4?) |
| 18:42.47 | RobotGuy | landley: I think one of my adapters is Broadcom based and the drivers are not ready for prime time yet, even in kernel 2.6/25-rc7. |
| 18:42.48 | ce_geek | robotguy; So what's the advantage of booting from a JFFS2 root filesystem instead of using the initrd in RAM? Using memory somewhere else? |
| 18:42.51 | landley | Not necessarily a real product, just "what might drive hammer unit volume and make it cheap?) |
| 18:43.10 | landley | ce_geek: you can drop cache pages if you have backing store. |
| 18:43.26 | landley | Memory pressure can't do anything with ramfs, the entire contents have to be in memory all the time. |
| 18:43.33 | RobotGuy | ce_geek: You get more RAM for applications. My Hammer gets 24 MB - 25 MB of RAM available now, booting from flash. |
| 18:43.38 | landley | jffs2 uses the flash as (compressed) backing store. |
| 18:43.51 | adyer | landley: as soon as your volumes get up to a few 1000's you might as well make your own board with hammer + other stuff on it |
| 18:43.59 | landley | Plus the data can persist when you write it. Initramfs won't save it across a reboot. |
| 18:44.16 | landley | adyer: sure. |
| 18:44.20 | landley | But it's a chicken and egg problem. |
| 18:44.24 | ce_geek | What are the disadvantages of this arrangement? |
| 18:44.30 | RobotGuy | landley: How so? |
| 18:44.35 | landley | ce_geek: more of a pain to set up. |
| 18:44.48 | ce_geek | landley: but once it's set up, it's just all around better? |
| 18:44.56 | landley | RobotGuy: if they can front the cash to manufacture 5000 or 10000 units of inventory at a time, the price goes way down. |
| 18:45.08 | landley | ce_geek: for large filesystems, yeah. |
| 18:45.16 | RobotGuy | landley: Yes, of course it would. |
| 18:45.21 | landley | There's a break even point, but it's probably about the 2 megabyte (uncompressed) mark. |
| 18:45.40 | RobotGuy | landley: My jffs2 image is 3.4 MB now. |
| 18:45.44 | landley | RobotGuy: but tying up that much cash in inventory you're not sure you can sell is a big risk for a small company to take. |
| 18:46.06 | landley | And if they make smaller production runs, the price is much higher and there's much less demand for them because they're more expensive. |
| 18:46.06 | RobotGuy | landley: Yes, I know. |
| 18:46.16 | *** join/#edev drath_ (i=vmaster@p5B07EFED.dip.t-dialin.net) |
| 18:46.29 | landley | If the suckers cost $49 a pop, everybody would want one. And if everybody wanted one, they could profitably make it that cheaply. |
| 18:46.36 | RobotGuy | There is one board I would REALLY like to have. |
| 18:46.40 | landley | For a value of "everybody" that's probably somewhere well into 6 figures of unit volume. |
| 18:46.45 | Dr{Who} | ok i added a wiki page on adding jffs2 to the hammer.. need to add more later but its a start. |
| 18:47.03 | RobotGuy | Dr{Who}: I have a howto for that. |
| 18:47.32 | Dr{Who} | ohh.. ok well you can nuke mine or snag any data from it then nuke.. |
| 18:47.39 | landley | prpplague: tincantools still isn't sending anyone to CELF? |
| 18:48.19 | prpplague | landley: no |
| 18:48.57 | RobotGuy | They could send ME! :D:D Just kidding, of course, but wishful thinking. |
| 18:49.06 | Dr{Who} | my data is not related to booting just adding a mountable partition |
| 18:49.06 | landley | prpplague: alas. |
| 18:49.21 | landley | I'll be showing my students my nail board during the cross compiling tutorial. |
| 18:49.55 | RobotGuy | Dr{Who}: Ah, OK. Just be sure you describe it properly so it is not confused with booting from flash. :) |
| 18:50.30 | Dr{Who} | will do. |
| 18:50.44 | RobotGuy | I don't know why my kernel does not boot now. It hangs at starting to boot, right after uncompressing the kernel. |
| 18:51.10 | ce_geek | :( |
| 18:51.54 | adyer | landley: nail boards and crosses? hmmmm |
| 18:53.00 | ce_geek | haha |
| 18:53.01 | prpplague | landley: sorry, i'm a bit busy at the momment, give a few to catch up |
| 18:53.37 | Dr{Who} | ok hows that robotguy.. |
| 18:54.16 | RobotGuy | Dr{Who}: Booting from the flash is different than just having the flash mountable as a partition. |
| 18:54.40 | ce_geek | Dr{Who}: Is it in the list of hammer board howtos? |
| 18:55.01 | prpplague | just a note again, you always boot from flash, using flash as the root filesystem is the difference |
| 18:55.09 | prpplague | ce_geek: it is now |
| 18:55.25 | RobotGuy | I meant using flash as a root partition. |
| 18:55.31 | Dr{Who} | i just added a link from the howto |
| 18:55.40 | RobotGuy | looks |
| 18:56.25 | ce_geek | Oh ya I see now. |
| 18:56.28 | ce_geek | kthx |
| 18:56.56 | Dr{Who} | presumptions are needed to be added like "you need mkfs.jffs2" etc |
| 18:57.08 | RobotGuy | I had my flash mounted from inittab. You have to be sure you have the JFFS2 filesystem in the kernel and not as a module also. |
| 18:57.09 | prpplague | why the two seperate howtos? |
| 18:57.21 | RobotGuy | prpplague: No idea |
| 18:57.43 | RobotGuy | prpplague: Using flash as a root filesystem versus using it for data. |
| 18:58.04 | RobotGuy | Dang, missed my free lunch today. :( |
| 18:59.07 | Dr{Who} | robot.. im presuming that buildroot is building your jffs2? |
| 18:59.20 | RobotGuy | Dr{Who}: Yes. |
| 19:00.03 | Dr{Who} | k. so my howto is a roll your own so thats good to just make a jffs2 etc. |
| 19:00.47 | RobotGuy | Dr{Who}: I think it's actually easier to do it from inittab. |
| 19:01.12 | Dr{Who} | ? |
| 19:01.17 | Dr{Who} | inittab? |
| 19:01.35 | agaffney | oh boy |
| 19:01.47 | RobotGuy | Dr{Who}: buildroot/target/device/Samsung/Hammer/target_skeleton/etc/inittab |
| 19:01.52 | agaffney | you're doing embedded stuff with linux and you don't know what inittab is? :P |
| 19:02.09 | Dr{Who} | i know exactly what it is : P |
| 19:02.25 | Dr{Who} | i did not under stand the context of his statemnt. |
| 19:02.28 | agaffney | it looked like you were questioning what it was :P |
| 19:02.30 | agaffney | ah |
| 19:03.29 | ds2 | morning |
| 19:03.42 | RobotGuy | Dr{Who}: I have a line to mount /dev/flash (/dev/mtdblock1) in my inittab (commented out now, of course). |
| 19:03.46 | ds2 | bah who needs inittab |
| 19:03.50 | ds2 | or for that matter init ;) |
| 19:04.12 | agaffney | linuxrc ftw? :P |
| 19:04.20 | ds2 | sure |
| 19:04.32 | ds2 | or just crash userspace and let all the interrupts in your tasks |
| 19:04.33 | landley | Actually the third thing you can do is boot your kernel _out_ of jffs2. |
| 19:04.35 | Dr{Who} | hmm i dont have that directory structure in my buildroot. |
| 19:04.45 | landley | In theory, the bootloady thing whose name I forget has a jffs2 reading driver... |
| 19:04.56 | ds2 | loadley: uboot |
| 19:04.58 | RobotGuy | I have this: "null::sysinit:/bin/mount -t jffs2 /dev/mtdblock1 /" now, but I am not sure it is valid for my config at this point. |
| 19:05.06 | landley | No, the asus or whatever it is bootloader has it too. abit. Something. |
| 19:05.19 | landley | (Heck, grub does this. If the FSF can manage it, by definition it's not that hard.) |
| 19:05.30 | ds2 | RobotGuy: you have root=/dev/mtdblock1 as the command line, right? |
| 19:05.37 | RobotGuy | Dr{Who}: You don't have the Samsung dir off buildroot/target/device ? |
| 19:05.53 | Dr{Who} | nope.. |
| 19:05.57 | RobotGuy | ds2: Yes, in APEX, I set that. |
| 19:06.13 | ds2 | RobotGuy: that line should no longer be needed |
| 19:06.17 | RobotGuy | Dr{Who}: Which version of buildroot are yiu using? |
| 19:06.21 | *** join/#edev LyosNorezel (n=Ralph@unaffiliated/lyosnorezel) |
| 19:06.34 | RobotGuy | ds2: Has to know where the root partition is. |
| 19:06.51 | ds2 | RobotGuy: that's what the root=/dev/mtdblock1 rootfstype=jffs2 does |
| 19:06.55 | Dr{Who} | probably svn |
| 19:07.20 | ds2 | not sure how that line is ever valid as by the time init runs, / better known or else bad things will happen |
| 19:07.22 | RobotGuy | Dr{Who}: Bad idea. |
| 19:07.42 | RobotGuy | Dr{Who}: Use the Jan 8th version available on the Wiki. |
| 19:07.51 | *** join/#edev ico2 (n=ico2@modem-655.chimpanzee.dialup.pol.co.uk) |
| 19:08.01 | RobotGuy | We can't support the SVN versions of buildroot. ;) |
| 19:08.21 | RobotGuy | OK, well, I have been tinkering with it, and it almost works now. |
| 19:10.01 | RobotGuy | Uh oh, I better be careful.. I might be starting to sound like a mini-prpplague. Hee hee |
| 19:10.37 | RobotGuy | hides |
| 19:12.37 | Dr{Who} | i have no issues with my snapshot so far. i did use the kernel provided on the iso as it is patched up specific to the hammer. I keep buildroot seperated as much as possible form my build process. |
| 19:13.16 | RobotGuy | Dr{Who}: You will still need the Samsung stuff from the Jan 8th buildroot on the Wiki. |
| 19:13.16 | Dr{Who} | a few things i have to patch are bash as it has a bug since bash-patch-011 |
| 19:14.17 | Dr{Who} | whats in the patch? |
| 19:14.38 | Dr{Who} | ill fetch it and look. |
| 19:15.22 | RobotGuy | Dr{Who}: All the stuff related to Hammer. device_table.txt, etc. |
| 19:15.37 | Dr{Who} | where is the patch at? |
| 19:15.42 | RobotGuy | Dr{Who}: You need to add that at buildroot/target/device |
| 19:16.15 | Dr{Who} | i dont use the device table portion of buildroot my distro has its own configs for that. |
| 19:16.15 | RobotGuy | Dr{Who}: Then edit buildroot/target/device/Config.in to add the selection for the Samsung processor. |
| 19:16.46 | RobotGuy | Dr{Who}: I guess you are on your own then. I don't know what you are using. |
| 19:16.58 | Dr{Who} | my init process builds all of that on the fly from a config inside of a package that can be modified after flashing. |
| 19:17.07 | RobotGuy | Dr{Who}: What are you using? |
| 19:18.56 | Dr{Who} | my distro is a branch of LRP i started 8 years ago. |
| 19:20.17 | chouimat|work | LRP? |
| 19:20.27 | Dr{Who} | linux router project |
| 19:20.49 | chouimat|work | been a long time since I used this ... |
| 19:21.28 | Dr{Who} | i moved it all over to buildroot this year.. originally i have a very a huge turd of files and packages i just maintain. now its all a wrapper aroudn buildroot. |
| 19:21.43 | Dr{Who} | originallly 2.2 kernel |
| 19:22.13 | Dr{Who} | basic idea is if i modify /etc/fstab i dont have to reflash i just backup my package etc.ufo and reboot and it has the saved settings. |
| 19:22.30 | Dr{Who} | ok lunch time.. bbiab |
| 19:24.03 | RobotGuy | prpplague: What kinds of things would hang Hammer at booting the kernel? I removed the BT module. |
| 19:25.42 | prpplague | RobotGuy: there are a ton of things |
| 19:25.43 | prpplague | RobotGuy: you could have turned of some of the serial console support |
| 19:25.43 | RobotGuy | Oh, wait, it booted! There was just a long pause after uncompressing the kernel. |
| 19:25.43 | prpplague | RobotGuy: you could have removed the floating point emulation |
| 19:25.44 | prpplague | RobotGuy: yea, thats the jffs2 being initialized |
| 19:25.46 | prpplague | RobotGuy: first boot after erase |
| 19:26.02 | RobotGuy | prpplague: I just didn't get all the usual syslog stuff on the console, which is weird. |
| 19:26.52 | RobotGuy | checks his inittab |
| 19:27.08 | prpplague | RobotGuy: sure you have "console=ttySAC0" for your kernel argument? |
| 19:27.52 | RobotGuy | prpplague: Whoops. I have "console=/dev/ttySAC0" |
| 19:28.14 | prpplague | RobotGuy: sure you have "console=ttySAC0" for your kernel argument? |
| 19:28.25 | prpplague | RobotGuy: just "console=ttySAC0" |
| 19:28.28 | prpplague | RobotGuy: no /dev/ |
| 19:28.32 | RobotGuy | prpplague: Changing it now. :) |
| 19:45.20 | RobotGuy | Now it says it can't open an initial console, but everything looks like it should be working. |
| 19:45.35 | RobotGuy | I don't remember what this error means. |
| 19:50.02 | prpplague | RobotGuy: probably means your jffs2 image is hosed |
| 19:51.21 | RobotGuy | prpplague: That's kind of vague. |
| 19:52.11 | RobotGuy | prpplague: I figured it has to do with my JFFS2 image. |
| 19:53.46 | RobotGuy | prpplague: http://www.pastebin.ca/975769 |
| 19:55.31 | *** join/#edev drath_ (i=vmaster@p5B07C4AA.dip.t-dialin.net) |
| 19:56.17 | landley | DrWho: there's a decent 2.6 patch for hammer kernels now. Prpplague posted it, and I snagged a copy here: http://landley.net/hg/firmware/file/713a53bce9a6/sources/patches/linux-2.6.25-rc7-hammer.patch |
| 19:58.28 | RobotGuy | landley: I believe it's also available on the Wiki now. |
| 19:58.47 | landley | Oh good. |
| 20:00.17 | prpplague | RobotGuy: you are either missing the ttySAC0 or your inittab is missing, or your busybox doesn't have init |
| 20:00.27 | prpplague | RobotGuy: there are sooo many things that this could be |
| 20:01.01 | RobotGuy | prpplague: I have busybox, and I have an inittab, and ttySAC0 is in the default command line. |
| 20:01.18 | ds2 | RobotGuy: is dev on your jffs2 image fully populated? |
| 20:02.22 | RobotGuy | ds2: Looks like it is. |
| 20:03.06 | RobotGuy | Everything looks fine so far. |
| 20:03.20 | ds2 | whoa i ought to get food |
| 20:11.11 | RobotGuy | I need to find an RP-SMA female to male cable. |
| 20:15.12 | RobotGuy | Yikes, $20.00 for a 12" RP-SMA M/F cable. |
| 20:26.38 | prpplague | RobotGuy: is your /dev/ populated? |
| 20:26.46 | prpplague | RobotGuy: did you copy the jffs2 to the right location |
| 20:26.48 | RobotGuy | prpplague: Yes |
| 20:26.58 | prpplague | RobotGuy: does the partition sizes match? |
| 20:27.04 | RobotGuy | prpplague: Yes, using the howto as reference. |
| 20:27.25 | RobotGuy | prpplague: I can't check partition size because it doesn't finish booting. |
| 20:27.39 | prpplague | RobotGuy: you have the partition size set in your kernel |
| 20:27.53 | prpplague | RobotGuy: i'd suggest starting by reloading your jffs2 image |
| 20:28.53 | RobotGuy | prpplague: .size is correct if that's what you are referring to. |
| 20:29.02 | RobotGuy | prpplague: Reloading now |
| 20:29.10 | prpplague | RobotGuy: what is the .size set for? |
| 20:29.30 | RobotGuy | prpplague: 0x140000 now. |
| 20:30.36 | prpplague | RobotGuy: ok, and you are coping your jffs2 image to nor:1280k+size of image ? |
| 20:30.55 | RobotGuy | prpplague: I'm building a new kernel, bloader, and jffs2 image now. |
| 20:31.04 | prpplague | uh ok |
| 20:31.31 | RobotGuy | prpplague: copy 0x30800000+<jffs2-size> nor:1280k |
| 20:31.51 | prpplague | yep that sounds correct |
| 20:31.54 | RobotGuy | prpplague: Using the exact byte count |
| 20:32.36 | RobotGuy | prpplague: It's scary how much of this I am retaining. |
| 20:35.23 | *** part/#edev LyosNorezel (n=Ralph@unaffiliated/lyosnorezel) |
| 20:39.22 | RobotGuy | Wow, 9 forum threads to go bye bye now. |
| 20:42.19 | prpplague | RobotGuy: yea, i really wish they'd fix that |
| 20:46.38 | RobotGuy | prpplague: Me too! Sheesh, it's getting way out of hand and I don't like having to keep logging in all the time. |
| 20:51.44 | chouimat|work | ? |
| 20:52.08 | prpplague | chouimat|work: pr0n spam on the forum |
| 20:52.22 | chouimat|work | ooooh good pr0n? |
| 20:53.12 | prpplague | hehe |
| 20:53.53 | RobotGuy | Most of those porn posts contain a virus. |
| 20:57.00 | landley | Just like real life! |
| 20:57.40 | RobotGuy | hides |
| 20:57.41 | landley | (Well, actually sex workers are generally adamant about getting regularly tested and their test results are an important part of their cirriculum vitae...) |
| 20:58.07 | landley | They're actually some of the safest non-virgins you can('t) have sex with, statistically speaking. |
| 20:58.15 | landley | But the line was just too obvious to pass up... |
| 20:59.09 | landley | Most forums eventually go to validation mode for anonymous users. |
| 20:59.11 | RobotGuy | landley: You've stood in line for them? |
| 20:59.24 | landley | Don't show the post until someone approves them, rather than playing whack-a-mole with live content. |
| 20:59.32 | landley | RobotGuy: in line...? |
| 20:59.42 | RobotGuy | landley: I don't know what software TCT is using, but I wish it were wider format. |
| 20:59.45 | adyer | chuckles |
| 20:59.52 | RobotGuy | landley: for the sex workers. :P |
| 21:00.08 | landley | RobotGuy: nope. |
| 21:00.21 | RobotGuy | But, enough of that. I don't think that is an appropriate topic to be discussing here. |
| 21:00.34 | landley | I haven't starred in a non-porn movie either, but following various aspects of the entertainment industry is a hobby of mine. |
| 21:01.30 | landley | P.S. http://www.fool.com/portfolios/rulemaker/2000/rulemaker000406.htm |
| 21:01.35 | Dr{Who} | robot.. not sure if it will help and it wont be fun to patch.. you can try adding kdb to your kernel.. then you can break into a debugger at exactly the point that fails.. |
| 21:01.38 | landley | I tend to cross-pollinate my hobbies. :) |
| 21:01.52 | landley | Often with the most tenuous excuses... |
| 21:02.15 | Crofton | you recommend any pr0n stocks? |
| 21:03.29 | T0mW | prpplague: did they fire you yet? |
| 21:03.42 | prpplague | T0mW: hehe not yet |
| 21:03.51 | prpplague | T0mW: i'm much more likely to quit first |
| 21:04.01 | T0mW | prpplague: well, you'll probably have to wait for payday for that anyway |
| 21:04.14 | prpplague | hehe |
| 21:04.19 | RobotGuy | prpplague: If ya go, tell me where so I can follow ya. |
| 21:04.34 | RobotGuy | Well, OK, he may not want me to follow him. |
| 21:04.39 | T0mW | prpplague: I've been fsck'ing around with debian on an IOP348 system |
| 21:05.11 | T0mW | prpplague: 1200MHz xscale, 8 SATA ports + 2 Gigabit LAN |
| 21:05.28 | ds2 | who is selling a 1.2GHz xcale? |
| 21:05.36 | T0mW | ds2: intel |
| 21:05.36 | landley | Crofton: yes actually, but they were in europe and I don't try the markets over there. |
| 21:05.40 | landley | Completely different accounting rules. |
| 21:06.25 | T0mW | ds2: 256Meg of DDR2 ram |
| 21:06.38 | ds2 | T0mW: I didn't know they actually sold those |
| 21:07.04 | T0mW | well, this is a development board, not the actual product |
| 21:07.08 | RobotGuy | prpplague: woooo hoooo, it boots AND I have WiFi (wlan0). Got to check and see if the WiFi is configured now. |
| 21:07.22 | prpplague | RobotGuy: what was the problem? |
| 21:07.26 | T0mW | ds2: fast |
| 21:08.12 | ds2 | I see. |
| 21:08.34 | ds2 | T0mW: have you done any benchmarks on it vs a nonXscale ARM? (say a 600MHz OMAP) |
| 21:08.52 | adyer | T0mW: how much $ for that IOP? |
| 21:09.07 | T0mW | ds2: naw, it does come up with 1192 bogoMIPs |
| 21:09.42 | T0mW | adyer: dunno, I just have the thing handed to me to put a kernel on it + iscsi |
| 21:10.13 | RobotGuy | prpplague: I don't remember. :( I hate memory lapses, or whatever it is I experience. |
| 21:10.20 | T0mW | not too bad, iscsi xfer rate (via hdparm -t) is about 60MB/s |
| 21:11.15 | RobotGuy | prpplague: This sucks. It wants a firmware file I apparently do not have. The other RT73 module never needed firmware. |
| 21:11.42 | ds2 | uh.. ALL the RT73 files needs firmware |
| 21:11.49 | adyer | T0mW: wow, lots of power - 8.4A @ 1.2V, probably not so mobile... |
| 21:11.58 | ds2 | just grab the tarball or do a CVS checkout from the codemonkey site...they have a firmware file in there |
| 21:12.55 | T0mW | adyer: lol, yeah, this is in a 1U chassis with redundant supplies |
| 21:13.32 | T0mW | PITA building the kernel + file system out. |
| 21:14.05 | T0mW | this is the first time I've built debian with debootstrap |
| 21:15.01 | CosmicPenguin | good times |
| 21:16.16 | RobotGuy | ds2: The module I used with kernel 2.6.22 did NOT require external firmware. |
| 21:16.37 | ds2 | RobotGuy: what is the name of the module? |
| 21:16.47 | RobotGuy | ds2: rt73usb |
| 21:17.00 | RobotGuy | ds2: But it pulls in other modules also |
| 21:17.35 | RobotGuy | ds2: rt73usb,rt2x00usb,usbnet |
| 21:18.08 | ds2 | the rt2x00 driver uses a firmware file, IIRC... maybe there is a way to compile it in |
| 21:18.53 | RobotGuy | ds2: Yeah, it's the rt2x00usb module that needs the firmware. |
| 21:19.19 | ds2 | okay, so the names got shuffled around. |
| 21:19.22 | RobotGuy | ds2: It is looking for rt73.bin now. This is not making sense to me. |
| 21:19.45 | ds2 | all the ones I have used looks for that... do you have udev/hotplug installed? |
| 21:19.53 | RobotGuy | ds2: No |
| 21:20.02 | ds2 | if so, I think all you need to do is drop it into /lib/firmware or /usr/lib/firmware |
| 21:20.20 | RobotGuy | ds2: I have to find the firmware module first. |
| 21:20.43 | ds2 | rt73.bin is not a module. it is part of the source tarball on codemonkey, the last time I looked |
| 21:23.11 | RobotGuy | ds2: I realize it is not a module. I just need to find it so I can see if I can make this work. |
| 21:23.59 | RobotGuy | ds2: I've already bought two different usb WiFi adapters - and the other one does not work because the broadcom drivers are not workable yet. |
| 21:25.41 | ds2 | okay |
| 21:25.53 | ds2 | http://rt2x00.serialmonkey.com/rt73-cvs-daily.tar.gz |
| 21:25.55 | ds2 | grab that tarball |
| 21:26.01 | ds2 | -rw-r--r-- rt2x00/rt2x00 2048 2006-10-06 07:22:12 rt73-cvs-2008040715/Module/rt73.bin |
| 21:26.01 | Dr{Who} | robot http://oss.sgi.com/projects/kdb/ but it wont patch well into 2.6.25 that kernel is soooooo changed.. i almost dont want to touch it at all till 2.6.26 when they get more of the drivers ported |
| 21:26.09 | ds2 | that's the firmware file, enjoy. |
| 21:26.11 | *** join/#edev ico2 (n=ico2@modem-2629.cougar.dialup.pol.co.uk) |
| 21:26.48 | prpplague | RobotGuy: you have to be careful, semantics can be a real gotcha when dealing with embedded |
| 21:27.28 | RobotGuy | prpplague: I know, and I know I haven't quite got there yet. I am trying though. |
| 21:27.56 | RobotGuy | WiFi is getting to be a major pita. |
| 21:28.12 | prpplague | RobotGuy: its always been a pita under linux |
| 21:28.35 | RobotGuy | prpplague: I can get my Ralink based adapter working under PC Linux just fine. |
| 21:30.36 | RobotGuy | W.A.L.T.E.R. will be pretty much not exciting or fun without WiFi. |
| 21:31.15 | prpplague | RobotGuy: you really should give it a girls name |
| 21:31.42 | RobotGuy | prpplague: Why? My BRAT is MsBRAT (pending a better name). |
| 21:32.06 | RobotGuy | Wheeled Autonomous Learning Terrain Exploring Rover |
| 21:32.27 | ds2 | you can always try to get the Marvel model working |
| 21:32.32 | ds2 | :) |
| 21:32.45 | RobotGuy | Marvell? There is a libertas driver. |
| 21:32.53 | ds2 | yeah |
| 21:33.06 | RobotGuy | I don't know if there are any Marvell based USB WiFi adapters though. |
| 21:33.17 | ds2 | use SPI |
| 21:33.35 | RobotGuy | I am getting way much embedded stuff in my head |
| 21:33.41 | ds2 | stares at the never ending pile of spanners }:-) |
| 21:33.49 | RobotGuy | ds2: I've never used SPI. |
| 21:33.56 | ds2 | it is just a serial bus |
| 21:34.01 | ds2 | nothing more then a fancy shift register |
| 21:35.00 | RobotGuy | ds2: I |
| 21:35.17 | RobotGuy | ds2: I've mainly used I2C for inter chip/device stuff. |
| 21:35.45 | ds2 | RobotGuy: yep, same idea except it doesn't have the start/stop or addressing but does use one other wire for a slave select |
| 21:36.16 | adyer | ds2: not quite - spi has two data lines, i2c shares one |
| 21:36.31 | RobotGuy | ds2: I know about the selects. I'm planning to use I2C for communication between Hammer and smart I/O handling modules |
| 21:36.59 | RobotGuy | MISO MOSI |
| 21:37.02 | RobotGuy | I think |
| 21:37.06 | ds2 | adyer: kinda... some devices will let you get away with a resistor to connect those 2 together |
| 21:37.27 | RobotGuy | doesn't care for shortcuts too much. |
| 21:37.38 | adyer | ds2: eww |
| 21:37.48 | ds2 | that's fine as long as the chip is not full duplex (clocking out data while it is being programmed) |
| 21:38.10 | prpplague | RobotGuy: you didn't get my email? |
| 21:38.42 | RobotGuy | prpplague: Nope |
| 21:38.58 | prpplague | RobotGuy: about the i2c level shifting? |
| 21:39.09 | RobotGuy | prpplague: Oh yeah, I got that one. |
| 21:40.19 | RobotGuy | prpplague: I'm just not ready to tackle it yet, and wondering if it's just not better to wait for RDP to be real. It's more circuitry than I might want to deal with or be able to fit on the proto area. |
| 21:40.57 | RobotGuy | It's going to be tight enough with two dsPIC4012's, the BT module, and connectors (if I can fit them all on). |
| 21:41.15 | RobotGuy | I can put one of the dsPICs under the BT module though. |
| 21:41.24 | prpplague | RobotGuy: got a question for you |
| 21:41.30 | RobotGuy | prpplague: OK |
| 21:41.48 | prpplague | RobotGuy: if you were doing your i2c bare metal, how would you write the code to talk via i2c? |
| 21:42.17 | Dr{Who} | robot.. not sure if this is what you want but it is where i am leaning.. I dont like 802.11x as it takes too much power and is a pita ie linux drivers etc. |
| 21:42.21 | Dr{Who} | this is what i am looking at |
| 21:42.23 | Dr{Who} | http://www.digi.com/products/wireless/point-multipoint/xbee-pro-series1-module.jsp |
| 21:42.53 | RobotGuy | prpplague: I do not know. I've not tried i2c bare-metal yet. I actually haven't done much bare-metal coding, except a bit for my LPC2148 boards using Crossworks. |
| 21:43.12 | prpplague | RobotGuy: how had you done i2c before? |
| 21:43.32 | adyer | prpplague: i2c slave or host? |
| 21:43.44 | prpplague | adyer: i2c host |
| 21:43.46 | RobotGuy | I've had FreeRTOS running on an LPC2148. The only I2C I have done before is with PICs and dsPICs. |
| 21:44.11 | RobotGuy | I've done master (PIC) to slave (dsPIC). |
| 21:44.12 | adyer | prpplague: if you don't care about speed, code it in C, you can steal the routines from u-boot (or linux I think) |
| 21:44.12 | prpplague | RobotGuy: and how did you code for the i2c on the pics? |
| 21:44.36 | RobotGuy | prpplague: I used the hardware i2c modules. C language. |
| 21:44.36 | prpplague | adyer: hehe, hang on, trying to do some research |
| 21:44.53 | prpplague | RobotGuy: ok, and what did the i2c hardware modules provide? |
| 21:44.54 | ds2 | just use a bluetooth PAN |
| 21:44.58 | ds2 | easier and lower power |
| 21:45.02 | ds2 | or DUN if you are lazy |
| 21:45.23 | RobotGuy | prpplague: Interrupt access for the slaves. |
| 21:45.27 | ds2 | $10-$15 for a USB Bluetooth adapter that should just work |
| 21:45.59 | prpplague | RobotGuy: ok, so did you have to configure some registers and such? |
| 21:46.06 | RobotGuy | prpplague: I have a whole bunch of I2C modules I want to write code for with Hammer. Sienna sent me three more to tinker with. |
| 21:46.24 | prpplague | RobotGuy: kernel modules? |
| 21:46.34 | RobotGuy | prpplague: Yep, had to configure registers properly for master and slave, all that low level stuff. |
| 21:46.38 | prpplague | RobotGuy: can you post them on the wiki ? |
| 21:47.02 | RobotGuy | prpplague: Post what? Ya lost me. |
| 21:47.16 | prpplague | <RobotGuy> prpplague: I have a whole bunch of I2C modules I want to write code for with Hammer. Sienna sent me three more to tinker with. |
| 21:47.35 | RobotGuy | I understand I2C pretty well now, and can work with it easily using PICs and dsPICs. |
| 21:47.36 | prpplague | RobotGuy: you said these are kernel modules |
| 21:47.47 | RobotGuy | prpplague: Oh, OK, you mean list the I2C modules I have? |
| 21:48.06 | prpplague | RobotGuy: post the i2c modules for the linux kernel that you have |
| 21:48.12 | RobotGuy | prpplague: No, I didn't say these are kernel modules. They are devices I want to write code for Hammer. |
| 21:48.24 | prpplague | <prpplague> RobotGuy: kernel modules? |
| 21:48.24 | prpplague | <RobotGuy> prpplague: Yep, had to configure registers properly for master and slave, all that low level stuff. |
| 21:48.40 | prpplague | ahh ok, i miss understood |
| 21:48.55 | RobotGuy | prpplague: I'm confused. I don't have a context. |
| 21:48.56 | prpplague | RobotGuy: ok, much clearer |
| 21:49.11 | RobotGuy | prpplague: Ah, good. |
| 21:49.12 | prpplague | RobotGuy: np, anyway, you had to configure all the registers and such, correct? |
| 21:49.20 | RobotGuy | prpplague: Yes, of course. |
| 21:50.16 | prpplague | RobotGuy: ok, now, _without_ using the i2c and spi driver support in the linux kernel, you code exactly the same as if you'd be doing it on a pic |
| 21:50.16 | RobotGuy | prpplague: I am wondering how hard it would be to write kernel modules for the I2C devices I have. It would probably be the proper way to do it. |
| 21:50.30 | prpplague | RobotGuy: you set up the registers, you configure the interrupts |
| 21:50.32 | prpplague | RobotGuy: all of that |
| 21:50.38 | *** join/#edev CosmicPenguin (n=nobody@163.181.251.103) |
| 21:50.38 | *** mode/#edev [+o CosmicPenguin] by ChanServ |
| 21:50.40 | prpplague | RobotGuy: but if you turn on the i2c and spi drivers |
| 21:50.41 | RobotGuy | prpplague: OK. Right. Same as bare-metal. |
| 21:50.52 | prpplague | RobotGuy: you have to use their api and have to use the kernel |
| 21:50.56 | prpplague | RobotGuy: to do your work |
| 21:51.23 | RobotGuy | prpplague: Well, I don't really have a problem with that, even if it is more work as long as the api is stable. |
| 21:51.54 | RobotGuy | Bare-metal coding would probably provide a faster responce, wouldn't it? |
| 21:52.03 | prpplague | RobotGuy: ok, just so you understand |
| 21:52.26 | RobotGuy | prpplague: I just downloaded "Linux Device Drivers, 3rd Edition" |
| 21:52.32 | prpplague | RobotGuy: a dedicated kernel module will provide the best response for your i2c and spi interfaces |
| 21:53.06 | RobotGuy | prpplague: OK. I want to do it the best way for best response. I think the kernel module route would be most beneficial in the end. |
| 21:53.48 | *** join/#edev nhoman (n=homann@89.100.83.14) |
| 21:53.53 | prpplague | RobotGuy: make sure you understand what i mean by dedicated driver |
| 21:54.13 | RobotGuy | prpplague: A driver just for a specific device. |
| 21:54.42 | adyer | RobotGuy: for prototyping in user mode, you might want to look at Documentation/i2c/dev-interface in the kernel tree |
| 21:54.47 | RobotGuy | prpplague: Like what is already in the I2C driver section. |
| 21:55.19 | RobotGuy | prpplague: I have been into that, which is how I found out what device files are needed. I added those to device_table.txt. |
| 21:56.19 | RobotGuy | prpplague: This is a major undertaking for me, but I will have loads of I2C stuff to post on the Wiki as I start to get things working. |
| 21:56.29 | prpplague | RobotGuy: a dedicated driver for your instance would not use the i2c or spi api |
| 21:56.43 | RobotGuy | prpplague: Now I am confused. |
| 21:57.04 | prpplague | RobotGuy: just because it is a kernel module doesn't mean it has to use the API |
| 21:57.10 | prpplague | RobotGuy: for either i2c or spi |
| 21:57.13 | RobotGuy | prpplague: I am thinking of a kernel module for each new I2C device. |
| 21:57.20 | prpplague | RobotGuy: i rarely use the i2c or spi intrastructure |
| 21:57.42 | RobotGuy | prpplague: OK, but that is you. :) I don't necessarily do things the same way you do. :) |
| 21:57.45 | prpplague | RobotGuy: if you have more than one device on the i2c bus, you'll need to use the i2c api |
| 21:57.59 | RobotGuy | prpplague: I will have many I2C devices on the bus. |
| 21:58.10 | prpplague | RobotGuy: just trying to make you understand all the options and the drawbacks of each |
| 21:58.39 | RobotGuy | prpplague: I figured I would be best off going the kernel module route. I think that's the way I need to go. |
| 21:58.49 | ds2 | regarding the dev-interface, has anyone written a generic 'i2c' command to send and recieve data on the i2c bus? think *BSD has something like that for basic testing |
| 21:59.02 | prpplague | RobotGuy: even if you go kernel module, there are several different ways to implement it |
| 21:59.17 | prpplague | ds2: we have one |
| 21:59.22 | prpplague | ds2: we also have a spi one |
| 21:59.29 | RobotGuy | ds2: Hmmmm, maybe that would be a good intro to embedded I2C to do before I get into writing kernel modules. |
| 21:59.42 | ds2 | prpplague: is that specific to the hammer or a generic one that will work on any i2c bus? |
| 22:00.00 | prpplague | ds2: its specific to the s3c24xx |
| 22:00.01 | RobotGuy | prpplague: I want the easiest way that provides the best response and access to device features. |
| 22:00.13 | ds2 | RobotGuy: it'll be fine for basic debugging but beyond that, the kernel side has different issues |
| 22:00.22 | ds2 | prplague: oh :( |
| 22:00.22 | prpplague | RobotGuy: hehe, loaded answer, as they are full of trade offs |
| 22:00.47 | RobotGuy | prpplague: OK, let's talk about what gives the best response first then. |
| 22:00.51 | ds2 | I questions the meaning of 'response' on a slow, non RT bus like I2C |
| 22:00.51 | prpplague | ds2: its not difficult, you make one for the i2c subsystem pretty easy |
| 22:01.10 | adyer | RobotGuy: response to what? do you have external non-i2c interrupts? |
| 22:01.15 | prpplague | RobotGuy: best response is going to be without the subsystem , if its coded properly |
| 22:01.28 | ds2 | prpplague: didn't say it was... just matter of time.. open/ioctls/r/w/close.. I think that's the Linux way instead of sockets like BSD |
| 22:01.33 | RobotGuy | prpplague: Without using the I2C subsystem? |
| 22:01.44 | prpplague | RobotGuy: correct |
| 22:01.57 | RobotGuy | prpplague: I want the best response using the I2C subsystem, and best access to device features. |
| 22:02.31 | prpplague | RobotGuy: there are no options when you are running i2c subsystem, you have to play by its rules |
| 22:02.56 | RobotGuy | prpplague: Are you saying I could do bare-metal, twiddling all the registers and such, without making the kernel unhappy? |
| 22:03.28 | prpplague | RobotGuy: you do have to do some things for the kernel, but you'd be writing a "dedicated driver" |
| 22:03.37 | prpplague | RobotGuy: which is what i was trying to explain to you |
| 22:03.50 | RobotGuy | prpplague: I'm OK with the rules. I want to go the kernel module route. I think (but might be wrong) that would be most beneficial to the community as a whole and get me where I want to be. |
| 22:04.26 | RobotGuy | One kernel module per device. |
| 22:05.41 | RobotGuy | prpplague: I think you don't see that I do understand. |
| 22:07.02 | prpplague | RobotGuy: but you keep saying "kernel module" |
| 22:07.08 | RobotGuy | Do I have to reupload my jffs2 image even if it does not change? |
| 22:07.11 | prpplague | RobotGuy: there are several different types of kernel module |
| 22:07.25 | prpplague | RobotGuy: only if you over write it or erase the flash |
| 22:07.51 | RobotGuy | <PROTECTED> |
| 22:08.14 | prpplague | RobotGuy: hehe |
| 22:08.18 | *** part/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 22:08.27 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 22:08.49 | prpplague | RobotGuy: you have a 1) kernel module with i2c API or 2) kernel module without i2c API |
| 22:08.54 | RobotGuy | Pressed the wrong button |
| 22:09.13 | prpplague | RobotGuy: 1) you have to use their API or 2) you get to do your own |
| 22:09.18 | RobotGuy | prpplague: I would think I want a kernel module with the I2C API. |
| 22:09.31 | RobotGuy | prpplague: I UNDERATAND that. |
| 22:10.09 | prpplague | RobotGuy: ok, i don't think you do, but ok |
| 22:10.20 | RobotGuy | What do I not understand???? |
| 22:10.58 | RobotGuy | I write a kernel module for each device that uses the I2C subsystem and the API. |
| 22:11.00 | *** part/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 22:11.12 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 22:11.30 | RobotGuy | What do I not understand?? |
| 22:12.12 | prpplague | RobotGuy: i think you are not understanding all the things that are involved in using the i2c API and how it isolates you from the hardware |
| 22:12.32 | CosmicPenguin | you don't need to write a kernel module for each |
| 22:12.34 | prpplague | RobotGuy: but give it a try |
| 22:12.38 | CosmicPenguin | in fact, you can do a bunch of it from userland |
| 22:13.09 | ds2 | nice thing about doing it in userland is having the potential option of making it nonGPL ;) |
| 22:13.10 | RobotGuy | Why would I not want to use the I2C subsystem and API???? |
| 22:13.28 | ds2 | RobotGuy: each subsystem you use adds to latency |
| 22:14.02 | RobotGuy | I am getting too confused!! I gotta go. Can't deal with this anymore right now. |
| 22:14.10 | prpplague | RobotGuy: np |
| 22:14.38 | CosmicPenguin | I don't understand - he is still using the API - its not like the API is all that anyway |
| 22:14.47 | prpplague | hmm, i don't think i'll bring that up again until he starts freaking out with stuff not working |
| 22:15.09 | prpplague | CosmicPenguin: his problem is that he is using custom pic devices with custom protocols |
| 22:15.10 | ds2 | the nice thing with the API is most of your driver will work if you drop the hardware onto another board |
| 22:15.18 | prpplague | CosmicPenguin: that need to pretty close real time |
| 22:15.52 | prpplague | CosmicPenguin: he thinks that by simply turning on the i2c driver in the kernel his custom interfaces will automagically work |
| 22:16.33 | CosmicPenguin | yeah - thats not really how it works |
| 22:17.11 | CosmicPenguin | its just i2c - not rocket science |
| 22:17.11 | adyer | prpplague: does the s3c2410 i2c driver sleep? |
| 22:17.31 | prpplague | adyer: i don't recall, we've always done our own since the s3c24xx i2c driver seemed to be pretty slow |
| 22:18.07 | CosmicPenguin | most i2c drivers will sleep |
| 22:18.16 | CosmicPenguin | in the LInux kernel that is |
| 22:18.30 | CosmicPenguin | just because of that lag time between response |
| 22:18.34 | CosmicPenguin | and kernel guys care not for real time |
| 22:18.37 | ds2 | Mmmm blocking on a 400KHz clock yummy |
| 22:18.58 | adyer | prpplague: I ran into that on i.MX - they had a polling loop, but if we ran through the scheduler it added a tick delay to every transaction |
| 22:19.02 | CosmicPenguin | I'm pretty sure our driver blocks - but its been a while since I've been in there |
| 22:19.11 | CosmicPenguin | s/blocks/sleeps/ |
| 22:19.19 | prpplague | for us, we have custom stuff that will never end up in the kernel, so there is no motivation to use the subsystem |
| 22:19.43 | prpplague | we can knock out a small custom kernel driver that talks directly to the hardware, and not worry abou tit |
| 22:20.58 | adyer | prpplague: I changed it to poll for a few thousand tries, then punt and run through the scheduler. that worked pretty well so far. |
| 22:21.57 | ds2 | prpplague: doesn't that make it painful if you had to switch processors? |
| 22:23.23 | prpplague | ds2: yes and no, 1) we usually design a cpu board that can be used in many projects 2) we generally have a life span for the product around 7 years 3) once a product is working, it is never touched unless a bug is found |
| 22:25.15 | ds2 | it still would cut into new product dev/code reuse if a new product uses a different proc? |
| 22:27.31 | prpplague | ds2: yes assuming that the same custom piece is used in the new device |
| 22:27.55 | prpplague | ds2: we generally do about 10-15 products based on the same processor and peripheral devices |
| 22:28.03 | prpplague | ds2: then ramp up for a new set |
| 22:28.09 | ds2 | I see |
| 22:28.28 | Crofton | http://techland.blogs.fortune.cnn.com/2008/04/07/amd-swings-the-ax-after-sales-shortfall/?cnn=yes |
| 22:28.49 | prpplague | ds2: its a weird business |
| 22:28.59 | prpplague | ds2: as CosmicPenguin and i have discussed in the past |
| 22:29.13 | ds2 | it is the nature of embedded. everyone is different |
| 22:33.58 | prpplague | heads home |
| 22:35.38 | adyer | Crofton: hmm, I'll have to see if AMD is more eager to sell me chips now, told me to buzz off a few weeks ago :-) |
| 23:06.24 | *** join/#edev prpplague (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:13.06 | prpplague | ho ho hum |
| 23:13.10 | ce_geek | ho ho hum? |
| 23:13.29 | ce_geek | http://datasheets.maxim-ic.com/en/ds/DS18S20.pdf <--- that's the temp sensor I wish to use. Reading up on it now. |
| 23:14.07 | ce_geek | I'm probably going to talk to it with an 8-bit device I'm familiar with and then see if I can be successful with the hammer. |
| 23:16.38 | *** join/#edev prpplague^2 (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:26.40 | *** join/#edev prpplague^2 (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:27.05 | prpplague | ho ho ho |
| 23:27.09 | ce_geek | Merry Christmas |
| 23:27.28 | prpplague | ce_geek no no, its: ho ho ho, merry freakin monday |
| 23:27.55 | ce_geek | Hehe |
| 23:27.56 | prpplague | it is interesting how the internet has influenced family life |
| 23:28.04 | ce_geek | :) |
| 23:28.06 | prpplague | my son tried to pull a fast one on my this evening |
| 23:28.13 | ce_geek | What family life? |
| 23:28.14 | ce_geek | :-/ |
| 23:28.21 | prpplague | "sorry dad, i forgot my homework at school" |
| 23:28.36 | prpplague | no problem, /dad downloads a copy from the school websitre |
| 23:28.45 | ce_geek | Hahahahaha. |
| 23:29.01 | prpplague | he's in his room doing it now and grumbling |
| 23:29.04 | prpplague | hehe |
| 23:30.02 | adyer | prpplague: make him back it up, so he can print a new copy when the dog eats it :-) |
| 23:30.16 | prpplague | ce_geek get your paper turned in? |
| 23:30.24 | ce_geek | prpplague: Yup!! |
| 23:30.49 | ce_geek | prpplague: Pumped. Actually in the lab this evening to get cranking with hammer stuff. |
| 23:31.00 | ce_geek | Reading the datasheet for the one-wire temperature sensor I'd like to play with. |
| 23:31.06 | ce_geek | http://datasheets.maxim-ic.com/en/ds/DS18S20.pdf |
| 23:31.27 | ce_geek | Thinking I should write code for an 8-bit device I'm familiar with and get that to work first, then make it work on hammer. |
| 23:32.05 | adyer | dallas.c if (qty_requested > 1) sleep(16 weeks); |
| 23:32.20 | ce_geek | Hah |
| 23:32.41 | ce_geek | I got two of them from sparkfun on 7/17/2006. |
| 23:32.47 | ce_geek | They've been waiting for me to play with 'em. |
| 23:33.25 | *** join/#edev prpplague (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:33.28 | prpplague | hmm |
| 23:33.34 | ce_geek | prpplague: Troubles? |
| 23:33.45 | prpplague | ce_geek yea cheap ass router |
| 23:33.52 | ce_geek | ah |
| 23:34.01 | prpplague | ce_geek i've been procrastinating about getting a much better one |
| 23:34.36 | prpplague | ce_geek when i moved back from overseas i was interested in getting up and running |
| 23:37.11 | prpplague | ho ho hum |
| 23:37.21 | prpplague | weird problem with a new board today |
| 23:37.25 | ce_geek | ? |
| 23:37.44 | prpplague | the nand flash seems to reset and when it does, it blanks the entire device and enables a soft write protect |
| 23:38.03 | ce_geek | nice |
| 23:39.16 | *** join/#edev RobotGuy (n=robotguy@dsl093-038-072.pdx1.dsl.speakeasy.net) |
| 23:42.47 | *** join/#edev prpplague (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:43.17 | *** join/#edev prpplague (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:44.57 | *** join/#edev prpplague (n=dave123_@ppp-70-244-174-210.dsl.rcsntx.swbell.net) |
| 23:45.16 | prpplague | wellllll @#$@#$@#$% |
| 23:45.54 | adyer | prpplague: your modem would work better without all the line noise :-) |
| 23:46.25 | Dr{Who} | ce_geek. i had excellent seccess with the 1wire code on this site.. http://www.ethernut.de/en/projects.html |
| 23:46.33 | ds2 | RING |
| 23:46.35 | Dr{Who} | very clean c code. |
| 23:46.39 | ds2 | RING |
| 23:47.02 | adyer | +++ATH0 |
| 23:47.16 | ds2 | NO CARRIER |
| 23:47.47 | adyer | ds2: lol, scary what you remember... |
| 23:48.48 | *** join/#edev ico2 (n=ico2@modem-2733.chimpanzee.dialup.pol.co.uk) |
| 23:50.39 | ds2 | adyer: semi recent work on faking out DUN to get a wince phone to talk to a board. board had to fake out a modem enough to keep wince happy |
| 23:51.13 | Dr{Who} | for 3 wire i had to write my own to talk to a ds1802 but its just a spi with a clock |
| 23:54.09 | Dr{Who} | i have some 1wire temp sensors if you cant find any i can send you a few via snail mail to play with. or you can get maxim to send you some free. |
| 23:56.16 | RobotGuy | was having chest pains earlier. :( |
| 23:56.33 | RobotGuy | Dr{Who}: I think I have some of those also. |
| 23:57.04 | RobotGuy | I think I got mine from Maxim. |
| 23:58.31 | Dr{Who} | yep.. they own them now. |
| 23:58.40 | Dr{Who} | and maxim is very good about samples.. : c ) and they are local for us : c ) |
| 23:58.59 | Dr{Who} | on ebay they are charging a lot for the 1w temp sensors. |