| 00:01.19 | GoatBoy | well, someone attempted a PXA patch for cpufreq |
| 00:01.22 | kergoth | just thinkig from the linux distro perspective. i maintain openzaurus and openembedded, and want a useful OE built root filesystem for this device. our buildsystem generally yanks things from upstream and builds them, so making sure i'm covering my ass on this |
| 00:02.47 | kergoth | GoatBoy: cpufreq notifies you when the core frequency changes, not when your parent's clock changes. its conceivable that the core could change, but the clock thats driving you changes parents and doesnt change, or what have you. i'd rather see the clock tree reflected directly in sysfs, with a means of initiating a change of a given clock's rate via a sysfs attribute or two |
| 00:03.25 | dsaxena | kergoth: if you send me an email with a list of licensing questions/concerns, I can bring them up directly with Intel folk |
| 00:03.35 | dsaxena | anonymously if so desired |
| 00:03.52 | kergoth | okay, thanks |
| 00:03.58 | dsaxena | i've got a pretty good working relationship with their linux TME and he's really linux-friendly and savy |
| 00:07.22 | *** part/#nslu2-linux rwhitby-treo (upirc@203.202.55.130) |
| 00:09.07 | CIA-5 | 03joule360 * 10unslung/sources/sudo-1.6.8.patch: Add sudo patch. |
| 00:12.06 | ekpneo | would stuff concerning a user-space libusb matter if I compiled the library libusb with the 2.4.25 kernel headers? Even though we are using the 2.6.7 kernel? |
| 00:14.16 | ekpneo | er... I am anyway |
| 00:37.35 | *** join/#nslu2-linux CodeWhacker (~CodeWhack@nr24-216-68-196-228.fuse.net) |
| 00:39.19 | CodeWhacker | has anybody managed to get the linux-2.4.x-mod tree from Linksys cleaned up enough to compile a dupe of the distributed kernel? |
| 00:40.00 | kergoth | yes |
| 00:41.16 | CodeWhacker | don't s'pose you have a howto? :) |
| 00:42.38 | CodeWhacker | they didn't compile NFS into the kernel, so my NFS3 server won't run |
| 00:52.45 | joule360-away | A query. I am working on putting openssl into the cvs repository. Here's the problems: a) they don't use the standard configure, so I need to hand-patch the makefiles after config to use the correct build tools [easy fix, but kludgey], and: |
| 00:53.33 | joule360-away | b) The final libraries need to be linked with libdl. Unfortunately, not all platforms support libdl (OSX does not, for instance) so I need to provide a libdl location on the final link command line. |
| 00:55.05 | joule360-away | I can do this in one of several ways: I could copy libdl from the slug into the sources dir, so it is available for linking to, or b) I could require that somehow the config partition on the slug be mounted to the host system and the location of the mount be made available to the build script so it can be linked to, or 3) I can tell anyone compiling on the Mac that they cannot do this, and they should just install the ipkg. |
| 00:55.45 | joule360-away | What would be the preferred approach? (Frankly, I favor copying libdl into the cvs system in some manner, OR some method of copying it to the host system before building.... |
| 00:57.22 | joule360-away | I guess another choice is to have the openssl build d/l source for libdl and compile it into the staging dir first, but that seems like a lot of extra work when the library is already on the slug, and is actually only 8K is size, so it's not a HUGE addition to the cvs stuff.... |
| 00:57.42 | joule360-away | I'd appreciate it if someone smart would tell me a preferred way of handling this. |
| 00:58.08 | kergoth | taking a binary from the slug strikes me as an extremely hackish way of doing it |
| 00:59.46 | joule360-away | agreed, kergoth. Other suggestions? |
| 01:00.31 | kergoth | normally your crosscompiler toolchain would have an arm libdl |
| 01:00.42 | kergoth | shouldnt be a problem at all, regardless of build machine |
| 01:05.00 | joule360-away | yes, but |
| 01:05.20 | joule360-away | i don't have one with my crosstools. |
| 01:05.47 | kergoth | fix your crosstools :) |
| 01:07.11 | joule360-away | thanks. I am using the one built by McTails...guess I need a rebuild. |
| 01:10.06 | joule360-away | Is there a crosstool building howto out? |
| 01:10.28 | joule360-away | iIf anyone has a quicklink, thanks; o/w, time to google. |
| 01:12.48 | jacques | yay! I found my soldering iron! |
| 01:15.44 | joule360-away | Would someone please verify for me the crosstool version preferred for unslung development? The howto on the nslu2-linux website indicates 0.28-rc32, but I recall the logs indicating rc34. Please clarify and I will try a build for MacOSX.. |
| 01:33.40 | [g2] | joule360-away, I think either will work. r32 with the patch or rc34/35 |
| 01:34.07 | [g2] | I don't know about the MacOSX part though |
| 01:35.54 | [g2] | dsaxena greets, and where would the best place to look for ixp4xx patches be ? |
| 01:37.15 | [g2] | we've got some issue with 2.6.7 like serial hangups and clock drift (my nslu2 thinks it's up for 8 days, but its more like 8 hours) |
| 01:42.43 | jacques | that could be related to the other timing issues like why lmbench thinks it can only measure down to 1000000usecs |
| 01:45.53 | [g2] | jacques, is that with 2.6.7 ? |
| 01:51.00 | dsaxena | [g2]: you should pull 2.6-bk |
| 01:52.07 | dsaxena | it has a timer fix (not sure if related) and mtd compile fix |
| 01:53.17 | [g2] | dsaxena, thx. I'm kinda new to this, I've got to get bk literate :) |
| 01:53.46 | kergoth | speaking of which, i should throw up a tree to nslu2.bkbits.net |
| 01:55.05 | *** join/#nslu2-linux dyoung-web (~80ab9a96@cpe-24-31-90-123.hawaii.rr.com) |
| 01:56.48 | dyoung-web | Did somebody already answer dsaxena's query about building a workable CSR under 2.6? |
| 01:59.57 | joule360-away | Thanks [g2]. Is one or the other version preferred? |
| 02:00.40 | joule360-away | NB: does this version of crosstools build libdl? kergoth says I should have it with my current crosstools, so I'm confused as to why I don't...when McTails pops in, I'll have to ask him as he built this version. |
| 02:00.47 | *** join/#nslu2-linux jacques (~fontenot@c-67-171-198-193.client.comcast.net) |
| 02:01.50 | jacques | i gotta learn to use xchat better - half the time when my connection goes down it doesn't rejoin my channels |
| 02:01.57 | jacques | and I have no idea why |
| 02:04.01 | [g2] | joule360-away, right now they are mostly the same. gcc 3.4.2 is a candidate, but no real verification has been done yet. |
| 02:04.10 | joule360-away | ic. |
| 02:04.22 | kergoth | i'm using 3.4.2 for everything i'm doing |
| 02:04.23 | [g2] | I think I ran a chrooted buildroot for a little while |
| 02:04.29 | kergoth | oe built root filesystems, and kernel work |
| 02:04.49 | [g2] | kergoth: using or executing on the device ? |
| 02:04.56 | jacques | however, 3.4.2 doesn't build the current stuff in cvs - neither unslung nor anyu2 |
| 02:05.15 | kergoth | [g2]: both. |
| 02:05.33 | kergoth | everything -i'm- doing builds with it and works with it on the device |
| 02:06.09 | [g2] | that's *great* news |
| 02:06.30 | jacques | yeah I'd like to move to 3.4.2 but changes need to be made to unslung/anyu2 first (assuming a non-immediate move to oe) |
| 02:06.40 | [g2] | is that oe advertisement done yet ? |
| 02:06.50 | [g2] | I'd love to make use of that :) |
| 02:07.03 | kergoth | not yet, got distracted by real life earlier |
| 02:07.06 | kergoth | i'll finish it up shortly |
| 02:07.13 | [g2] | totally awesome |
| 02:07.24 | jacques | bah if I allowed myself to get distracted by real life I would have eaten by now |
| 02:07.32 | kergoth | hehe |
| 02:07.56 | [g2] | you mean this isn't real life ? |
| 02:09.03 | jacques | I mean, I spent six hours in the last two days searching for my soldering iron so I could add serial ports to my nslu2's |
| 02:19.06 | *** join/#nslu2-linux chris_ (~chris@pcp09941955pcs.hyatsv01.md.comcast.net) |
| 02:19.15 | chris_ | re |
| 02:19.18 | chris_ | anyone home? |
| 02:19.55 | chris_ | everyone must be out gettin drunt |
| 02:26.00 | [g2] | ok i've installed bk, can just pull the 2.6-bk repo ? or is there a little more to it than that ? |
| 02:29.46 | jacques | you need to clone it first |
| 02:30.00 | jacques | but there's not that much more to it |
| 02:35.41 | kas11 | I guess this buildroot got compiled with -Dno-freakin-network-scripts |
| 02:36.49 | [g2] | kergoth, so do I want to clone the oe bk repo at openlogging.org or is there another way of going about this ? |
| 02:37.10 | kergoth | hm? |
| 02:37.14 | kergoth | rtfm |
| 02:37.21 | kergoth | openembedded.org -> wiki -> GettingStarted |
| 02:37.39 | [g2] | thx |
| 02:53.12 | *** join/#nslu2-linux Chezdude (~chatzilla@space.ne.client2.attbi.com) |
| 03:15.30 | chris_ | fyi: i just posted build instructions for an nfs version 3 server on the mailing list |
| 03:34.15 | [g2] | chris_, thx |
| 03:34.24 | [g2] | nite all |
| 03:39.44 | kas11 | ok...what is the real eth interface name on this silly thing...it doesn't seem to like npe0 |
| 03:39.59 | chris_ | lol i read that as "nippy0" |
| 03:40.16 | kas11 | works for me *w* |
| 03:40.18 | chris_ | [g2]: glad to help |
| 03:40.38 | kergoth | kas11: can run ifconfig -a and look? |
| 03:40.41 | kergoth | s/can/cant/ |
| 03:41.53 | kas11 | duh...guess i need sleep |
| 03:46.36 | *** join/#nslu2-linux McTails (~mctails@ool-182cae46.dyn.optonline.net) |
| 03:47.21 | chris_ | i want a gige switch |
| 03:50.26 | *** join/#nslu2-linux gelloglue (~18153f45@CPE-144-136-196-72.sa.bigpond.net.au) |
| 03:50.42 | gelloglue | hey dudes |
| 03:50.54 | chris_ | hi |
| 03:51.29 | gelloglue | So hows the AnyU2 kernel going? |
| 03:52.02 | chris_ | search me! |
| 03:53.27 | dyoung-web | Hi Gelloglue. |
| 03:53.36 | McTails | dyoung-web: did you look at the Bug Tracker ? |
| 03:53.38 | dyoung-web | AnyU2 is moving along. |
| 03:53.39 | kas11 | I be making slow progress here...dont know where dyoung is...I think g2 is sorta off the reservation ;) |
| 03:54.16 | gelloglue | howdy all |
| 03:54.40 | kas11 | considering someone stole all my network scripts, I guess I am doing as wll as can be expected |
| 03:54.48 | McTails | kas11: Stole ? |
| 03:54.53 | kas11 | gelloglue: howdy |
| 03:54.56 | gelloglue | I'm just doing my weekly snoop |
| 03:55.33 | gelloglue | I noted the AnyU2 stuff today - looks like good start on 2.6 |
| 03:55.40 | kas11 | or something got run with -Dno_freaking_network_scripts_needed |
| 03:55.42 | dyoung-web | Basically where its at now, is 2.6.7 with CSR, up in the air on type of rootfs. |
| 03:56.21 | dyoung-web | I used your CSR patches for 2.6. They work good! |
| 03:57.31 | gelloglue | Actually they are from Farin and Winkler. |
| 03:57.57 | gelloglue | I just help to get them to post |
| 03:57.59 | kergoth | fyi, http://openembedded.org/oe_wiki/index.php/NSLU2AnyU2 |
| 03:58.15 | kergoth | walks through obtaining and setting up oe, and using it to produce a minimal uclibc root filesystem |
| 03:58.34 | kergoth | need to push some package changes to improve the resulting rootfs |
| 03:59.03 | gelloglue | Wow .. you guys move way to fast for me |
| 03:59.47 | dyoung-web | Ah, I was waiting for Kergoth's docs.. |
| 04:00.19 | kergoth | as things stand today in our package repo that resulting rootfs will not contain CSR, and the kernel built is 2.4 |
| 04:00.33 | kergoth | and emits the ext2.gz and zImage seperately |
| 04:00.44 | kergoth | need to push the changeset that does the flashable image bits with slugtool |
| 04:00.56 | kergoth | my tree's a bit out of sync with upstream, with sync up in a bit |
| 04:01.17 | gelloglue | Yep the CSR stuff.. Got to have them get it from Intel |
| 04:01.35 | kergoth | I need to enhance the way we populate the OE mirrors before i can push CSR |
| 04:01.50 | kergoth | otherwise itll get distributed on the mirrors, which wouldnt make intel happy with us |
| 04:02.53 | gelloglue | they have other stuff to do but the may ask the big guys to remove from the mirrors |
| 04:03.21 | kergoth | doesnt matter. |
| 04:03.36 | dyoung-web | One issue being battled right now in 2.6 is some time timing thing. I understand there have been some patches since, but I didnt get to it yet. |
| 04:03.43 | kergoth | i have enough shit to deal with without doing something grossly stupid that violates a license, thanks. |
| 04:19.26 | gelloglue | <PROTECTED> |
| 04:20.15 | kas11 | has been belly-up all day |
| 04:23.33 | gelloglue | ixp4xx-osdg cvs is ok. nslu is getting a proxy error |
| 04:25.45 | CIA-5 | 03dyoung8888 * 10unslung/sources/busybox.config: Added Large File Support; CONFIG_LFS=y |
| 04:25.57 | jacques | heh |
| 04:34.51 | dyoung-web | BBIAB |
| 04:34.54 | *** part/#nslu2-linux dyoung-web (~80ab9a96@cpe-24-31-90-123.hawaii.rr.com) |
| 05:09.39 | kas11 | well, I have my network scripts back anyhow |
| 05:16.51 | rwhitby-away | kergoth: I linked to your wiki page from the AnyU2 wiki page. |
| 05:17.59 | rwhitby-away | the sf cvs has known problems with cvs for projects starting with certain letters. todays outage is brought to you by the letter "n" :-( |
| 05:18.09 | kergoth | rwhitby-away: to the doc directly, or to the main index? keep in mind that i havent pushed CSR build in oe yet, so any images built with it have no ethernet, better hope people playing with it have serial |
| 05:18.37 | kergoth | need to deal with the licensing situation as it applies to our automatically updated mirror |
| 05:18.42 | rwhitby-away | gelloglue: if you need developer access, just let me or any of the other sf nslu admins know your sf.net id |
| 05:18.54 | rwhitby-away | kergoth: just to the doc - would you prefer to the main index? |
| 05:19.11 | rwhitby-away | anyone playing with AnyU2 know what they are getting themselves into |
| 05:19.24 | kergoth | the doc is fine, i just dont want increased support load on something so new |
| 05:19.38 | kergoth | okay, in that case, no worries. a link to the maini index in addition would be good |
| 05:19.43 | kergoth | for general OE documentation |
| 05:19.45 | kas11 | I haven't been able to get in with :ext yet...but I think that is just the normal delay for noobs |
| 05:20.00 | kergoth | since there's a big difference between a howto and reference material |
| 05:20.40 | rwhitby-away | kergoth: http://www.nslu2-linux.org/wiki/AnyU2/HomePage - how's that? |
| 05:21.35 | kergoth | questionable |
| 05:21.37 | kergoth | its unintuitive |
| 05:21.59 | kergoth | when i see a highlighted OpenEmbedded, i expect that would take me to an openembedded project link, not OE as it applies to the nslu2 |
| 05:22.08 | rwhitby-away | ah - right. |
| 05:22.21 | kergoth | i'd change that to just openembedded.org, and add another link to the howto, clearly marked as such |
| 05:22.22 | rwhitby-away | feel free to salt to your taste :-) |
| 05:22.45 | kergoth | nodt |
| 05:22.49 | kergoth | s/t$// |
| 05:22.57 | rwhitby-away | [I agree with you] |
| 05:23.08 | kergoth | atm i'm putting hte finishing touches on some of the deployment to a flashable image bits so i can push them |
| 05:23.41 | kergoth | then i'll start implementing a mechanism for doing source distribution with OE that obeys the LICENSE variable in .oe files |
| 05:23.52 | rwhitby-away | I personally really appreciate the work you are doing - it is exactly what I was after in setting a baseline for AnyU2 development. |
| 05:23.56 | kergoth | once that's implemented, i'll get treke to use it for the mirror, then I'll push the CSR build |
| 05:25.12 | rwhitby-away | [I'm just somehow unable to work with stuff unless it's in a build system - even if it's one as primitive as the Unslung build system that we hope to soon replace.] |
| 05:26.00 | kergoth | same. i refuse to hack anything together manually. i'm willing to put up with hacks, but not UNMAINTAINASBLE hacks. |
| 05:26.07 | kergoth | s/ASB/AB/; |
| 05:26.48 | rwhitby-away | I like the way you think, kergoth. |
| 05:28.00 | kergoth | the first openzaurus release was a modification of the embedix distro shipped on the device. not unlike some of the projects here |
| 05:28.06 | kergoth | but it was entirely manual |
| 05:28.09 | kergoth | nightmare |
| 05:29.08 | kergoth | got tired of that and constructed the new version from the ground up, writing my own buildroot implementation as i went. eventually our buildroot had like 500 packages in it. i was impressed at how well it scaled |
| 05:29.20 | kergoth | considering it was just makefiles and all |
| 06:03.14 | kergoth | kergoth@linux |
| 06:03.20 | kergoth | ~/code/oe/build-armeb$ /bin/ls -l tmp/deploy/images/ |
| 06:03.21 | kergoth | total 8208 |
| 06:03.32 | kergoth | -rw-r--r-- 1 kergoth users 8388608 2004-09-18 02:04 bootstrap-image-nslu2-20040918060402.img |
| 06:03.41 | kergoth | drwxr-xr-x 2 kergoth users 4096 2004-09-18 02:04 slug |
| 06:03.44 | kergoth | there, fixed |
| 06:16.26 | jacques | this is incredibly frustrating - I cant get the solder out of the holes in eht nslu2 |
| 06:17.39 | kergoth | rwhitby-away: bk pull; bk -r co -q; thatll get you the latest updates. you'll want to do that in both oe and packages. |
| 06:18.06 | kas11 | jacques: it isn't too hard on the 2 signal holes...pwr and gnd planes suck up the heat tho... |
| 06:18.07 | kergoth | should automatically emit an image suitable for flashing if you selected the correct distro now. |
| 06:18.39 | jacques | kas11, those are the ones i cant do |
| 06:19.52 | kas11 | I mounted my serail adaper on the back of the board so I used single pins on those...heated the pins and the pad at the same time and let the pin fall thru |
| 06:20.12 | kas11 | not too bad with a small pair of needlenose pliers |
| 06:22.22 | kas11 | on the Vcc pin I used a piece of hookup wire that I put in by itself...if you notice, 3 of the adapter pins match the nslu so you just have to run Vcc from the nslu2 around and over to Vcc on the adapter |
| 06:23.32 | kas11 | of course, I removed the DB-9 connector beforehand |
| 06:24.24 | jacques | i think this tip is too small |
| 06:25.08 | kas11 | and hardwired...I'm not gonna worry about ever removing a $20 adapter so putting it on a header wasn;t desirable here |
| 06:26.06 | kas11 | you prolly don't want to put more than 25 watts on the signal lines...but more heat is helpful on pwr and gnd |
| 06:27.05 | kas11 | if you are good, you will find that more solder is better than less...makes it easier to get heat where needed...but you have to be careful |
| 06:27.58 | jacques | careful yeah. I may alredy have destroyed this one |
| 06:28.16 | jacques | iron has slipped a few times trying til push the tip into the holes |
| 06:29.04 | jacques | i can hold it on the power and ground and push hard and it just wont melt |
| 06:29.10 | kas11 | I tinned the holds a bit and used a good solder sucker on the signal pins...I held my pins against the holes with plier and applied solder around the base until the pin went thru on Vcc and gnd |
| 06:29.13 | jacques | tip is too long and small |
| 06:30.28 | kas11 | but I have been doing this stuff for 40 years...I know it can be tricky if you don't know the technique |
| 06:30.57 | jacques | now they tell me. I dont have anywhere near that soldering experience |
| 06:31.14 | jacques | i built a hero 1 in high school |
| 06:31.19 | kas11 | have you got the 2 signal holes cleaned? |
| 06:31.45 | kas11 | those are easiest do start there if not |
| 06:31.53 | jacques | yeah they are not the problem |
| 06:32.08 | kas11 | are ytou using solder wick or a sucker? |
| 06:32.41 | jacques | i have both |
| 06:33.03 | jacques | since icant get solder to melt on two holes neither halps |
| 06:33.32 | jacques | trying a different tip |
| 06:33.52 | kas11 | tin your tip first |
| 06:34.34 | kas11 | make sure you have a little excess solder on the tip to help with heat transfer...keeps the heat distributed |
| 06:34.41 | jacques | other tip was like a needle and I dont think there was enough heat transfer |
| 06:36.38 | kas11 | I prefer a small ip...but sometimes it is hard to get enough heat to the work |
| 06:44.53 | jacques | btw how do you use wick? |
| 06:45.44 | kas11 | place the wick over the hole and press it against the work with your tip |
| 06:46.05 | jacques | ok thats what i thought |
| 06:46.11 | jacques | it's not working at all |
| 06:46.26 | kas11 | the solder wicks to the copper...keep moving the wick so yu have fresh untinned copper over the hole |
| 06:46.56 | jacques | yeah if any solder ever wocked to it |
| 06:46.58 | jacques | heh |
| 06:47.07 | kas11 | wi doubt you will have much luck with wick tho unless you have lots of heat |
| 06:47.08 | jacques | holes too small wick too big |
| 06:47.47 | kas11 | you would need to get the heat to the work fast...otherwise the wisk sucks up all your heat |
| 06:48.27 | kas11 | wick can overlap...you just need copper over the solder you want to remove |
| 06:52.03 | rwhitby-away | jacques: wick the signal holes, then just push through the other two. |
| 06:53.32 | jacques | rwhitby-away, ive been trying to push through for like an hour now |
| 06:53.45 | jacques | all I have done is caused the iron to slip across several traces several times |
| 06:54.06 | kas11 | rwhitby: hope you saw my post about the docs...I haven't forgotten...just can't get in yet |
| 06:55.51 | kas11 | and am still finding little gotchas anyhow...like 4 empty dirs needed in /etc/network that weren't in the rootfs |
| 06:58.42 | jacques | i need three hands |
| 06:59.18 | kas11 | a small vice is your friend |
| 06:59.44 | jacques | yeah i need one |
| 06:59.55 | jacques | ok now hsve all but gnd |
| 07:00.11 | kergoth | man, i read that and thought 'vice' as in sin, vices |
| 07:00.19 | *** join/#nslu2-linux rwhitby-treo (~upirc@203.202.55.177) |
| 07:00.21 | kergoth | and wondered why you'd want only a small one |
| 07:01.51 | rwhitby-treo | jacques: I used a four pin header, and pre- pushed the two outside pins back. then solder in the two inside pins. then push the outside pins back through, making sure the thing you are pushing the pin with is not heat conductive. |
| 07:01.58 | kas11 | Gnd is a real bad one...I would heat from the opposite side while pressing the installed header towards the board..should be easie since you have the 3 clean holes |
| 07:03.10 | jacques | ok i think i have four clean holes now - but seriously I am worried that i killed the board |
| 07:03.13 | kas11 | don't solder anything except the gnd pin until you have it all the way in the hole tho |
| 07:03.41 | kas11 | the board is pretty tough... |
| 07:04.14 | kas11 | if the signal traces didn't lift, you should be fine |
| 07:05.40 | jacques | at least I dont think i have done that :-) |
| 07:06.16 | tking-away | chris_ : could you post your nfs3 howto on the wiki? |
| 07:06.22 | kas11 | you would know if you had...trace would curl up |
| 07:08.36 | rwhitby-treo | jacques: now you've got the holes clear, check operation before soldering again. |
| 07:08.49 | jacques | rwhitby-treo, oops too late |
| 07:08.57 | jacques | i just soldered the header |
| 07:10.04 | jacques | i'll check now |
| 07:11.36 | tking | hi rwhitby-treo |
| 07:13.45 | rwhitby-treo | tk drvng |
| 07:13.50 | tking | k |
| 07:14.03 | tking | pay attention to driving... |
| 07:14.28 | jacques | well it beeped once |
| 07:14.30 | tking | (I've never mastered driving and typing. |
| 07:14.36 | jacques | after doing the normal flashing patterns |
| 07:14.58 | rwhitby-treo | only type at lights |
| 07:15.30 | jacques | dangit I need top change my ip address to try to talk to it and that will kill me connection to irc |
| 07:15.46 | jacques | ive tried netmask tricks and they havent worked for me in the past |
| 07:16.12 | kergoth | windows or linux? |
| 07:16.22 | kergoth | you dont need to change your ip |
| 07:17.00 | jacques | windows? please. |
| 07:17.22 | jacques | ooh virtual dev? |
| 07:17.29 | jacques | i dont thik i have that built into kernel |
| 07:17.42 | jacques | and as I said changing netmask didnt work when I tried it before |
| 07:18.07 | jacques | but I am open to suggestions :-) |
| 07:18.49 | kergoth | yes, if you can get the nslu2 into linux you can just bring up ixp0:0 |
| 07:18.54 | kergoth | thats what i did when i first got mine |
| 07:20.35 | jacques | you mean if I have serial on the nslu2 ? |
| 07:22.01 | jacques | or you mean on the host side? I don't think I built my kernel with that feature - let me see |
| 07:23.37 | jacques | ixp0 is the nslu2 2.4 device name - which is why I'm confused |
| 07:27.17 | jacques | right, no /proc/net/ip_alias |
| 07:27.26 | jacques | i dont build in stuff I dont use |
| 07:28.03 | kergoth | heh, yeah, can try adding a route to it manually |
| 07:28.32 | kergoth | route add -host 192.168.0.1/255.255.255.255 eth0 |
| 07:28.34 | kergoth | or whatever |
| 07:28.45 | kergoth | i can never remember route syntax, since its different on every single fucking unix variant |
| 07:29.29 | kergoth | ~emulate ash |
| 07:29.30 | jbot | Honey, you got real ugly. |
| 07:29.36 | kergoth | ~emulate ash |
| 07:29.37 | jbot | ACTION says "hey, what's that on your face?" and shovels dirt on your face. |
| 07:30.27 | rwhitby-treo | ~ash |
| 07:30.28 | jbot | it has been said that ash is a lightweight shell with many of the same capabilities as the bash shell or now included in busybox with command line editing a la bash in like 30k less than stand alone ash |
| 07:30.50 | rwhitby-treo | hmm |
| 07:30.55 | kergoth | rwhitby-treo: movie reference. Army of Darkness and Evil Dead 1 & 2 |
| 07:31.28 | rwhitby-treo | ahh |
| 07:31.53 | kergoth | i highly recommend watching army of darkness at a minimum if you get the chance |
| 07:32.00 | kergoth | ~emulate bender |
| 07:32.01 | jbot | Yeah, well I'm gonna build my own lunar space lander! With blackjack aaaaannd Hookers! Actually, forget the space lander, and the blackjack. Ahhhh forget the whole thing! |
| 07:33.12 | rwhitby-treo | know bender well |
| 07:33.28 | kergoth | :) |
| 07:33.39 | dyoung-away | Shop Smart. Shop S-Mart. |
| 07:33.43 | kergoth | we've been training the bots to entertain us when bored |
| 07:33.45 | kergoth | dyoung-away: indeed :) |
| 07:33.47 | kergoth | ~emulate ash |
| 07:33.48 | jbot | Gimme some sugar, baby! |
| 07:34.26 | kergoth | ~listkeys emulate |
| 07:34.37 | kergoth | hmm |
| 07:34.40 | kergoth | ~emulate windows |
| 07:34.41 | jbot | ACTION steals your wallet, then straps an anchor to your neck |
| 07:34.44 | kergoth | doh |
| 07:35.01 | dyoung-away | jacques, the 2.6 ixp425_eth patch changes ixp0 back to normal eth0. |
| 07:37.39 | dyoung-away | Who was bender now? |
| 07:37.45 | kergoth | futurama |
| 07:37.48 | rwhitby-treo | thank deity |
| 07:45.14 | kergoth | ~emulate stewie |
| 07:45.15 | jbot | Ha ha. Oh, this is so good it just HAS to be fattening. |
| 07:45.23 | kergoth | ~emulate stewie |
| 07:45.24 | jbot | Damn you, broccoli. |
| 07:53.59 | *** join/#nslu2-linux fontenot (~fontenot@c-67-171-198-193.client.comcast.net) |
| 07:55.09 | dyoung-away | ~emulate emulate3buttons |
| 07:55.19 | jacques | damn |
| 07:55.32 | jacques | when I downed the virtual interface it took my default route with it |
| 07:55.39 | kergoth | heh, oops |
| 07:55.42 | jacques | and I didnt notice til my connectiones dropped |
| 07:55.49 | jacques | so I missed anything anyone said |
| 07:56.02 | jacques | after <kergoth> thats what i did when i first got mine |
| 07:56.22 | dyoung-away | jacques, the 2.6 ixp425_eth patch changes ixp0 back to normal eth0. |
| 07:56.36 | kergoth | everything else was just drivel |
| 07:56.39 | dyoung-away | which may explain some confusions. |
| 07:56.53 | dyoung-away | Ash is my hero. |
| 07:56.57 | dyoung-away | Boom Stick. |
| 07:57.04 | kergoth | yeah :) |
| 07:57.19 | kergoth | i was trying to convince a friend of mine to name the next release of his project 'boomstick' |
| 07:57.50 | jacques | well I didn't kill the nslu2 that I can tell yet |
| 07:58.09 | dyoung-away | I had the same fear. I heated the hell out of it. |
| 07:58.20 | jacques | but I soldered a female header to it and I dont have any male headers (single row) |
| 07:58.33 | dyoung-away | I ended up putting quite a bit of solder on so I could wick the damn thing off. |
| 07:58.35 | jacques | dyoung-away, yeah I had to apply a lot of heat |
| 07:59.17 | jacques | and yeah that was probsably the most useful tip - kas11 told me to put solder on my tip - I had already tinned it but when I put a ball on the end it worked better |
| 08:00.13 | jacques | and when I heated one side and sucked from the otther it worked best - tho tricky to do |
| 08:01.30 | kergoth | heh, just noticed the oe uclibc builds its ldconfig static |
| 08:02.42 | jacques | how are ppl getting the serial lines through the case? |
| 08:03.46 | rwhitby-treo | flat cable next to usb socket - no mod to case |
| 08:04.25 | rwhitby-treo | picture in yahoo group pics area |
| 08:07.33 | jacques | rwhitby-treo, ah thanks looking |
| 08:10.13 | jacques | hmm those pics are awfully dark |
| 08:10.30 | jacques | and small |
| 08:10.34 | *** join/#nslu2-linux mjn3 (~mjn3@ip68-226-136-198.lf.br.cox.net) |
| 08:10.50 | jacques | how do you get it bigger? |
| 08:10.59 | jacques | dont tell me the 400x300 pic is 352k |
| 08:11.12 | rwhitby-treo | click on picture - sorry about lighting |
| 08:11.15 | jacques | Posted: 23-Aug-2004 - Resolution: 400 x 300 | Full Size: 352 KB |
| 08:11.32 | jacques | yeah clicking on pic was first thing I tried |
| 08:11.47 | jacques | probably uses some javascript bs that I have disabled |
| 08:12.00 | jacques | but it doesn't appear clickable at all |
| 08:12.01 | dyoung | I milled a slot above where the header is. |
| 08:14.00 | rwhitby-treo | just put cable through same hole as disk1 usb socket |
| 08:14.22 | jacques | `k sounds like a plan to me |
| 08:14.50 | rwhitby-treo | flat idc cable |
| 08:23.46 | rwhitby-treo | it's a nice snug fit, and no damage to case |
| 08:28.54 | dyoung | (at least you werent like me and broked the case) |
| 08:45.04 | jacques | do I need anything other than a straight-through serial cable ? |
| 08:45.34 | jacques | damn wires arent long enough to go out hole without blocking port |
| 08:45.34 | kergoth | the nslu2 uart is probably dte, may need a null |
| 08:45.44 | jacques | well what are you using? |
| 08:45.44 | kergoth | (crossover) |
| 08:46.09 | kergoth | yep, null modem cable |
| 08:47.11 | jacques | crap |
| 08:47.48 | kergoth | could always do the cross yourself at any point. make the connection coming out of the device wired like a modem (DCE) |
| 08:47.56 | kergoth | then a straight through serial cable from there would be fine |
| 08:48.22 | jacques | oh yeah - but would that work with the max232? |
| 08:48.31 | kergoth | what do you mean? |
| 08:48.46 | jacques | if all it does is level conversion then it should work |
| 08:48.52 | jacques | if it's "smarter" then who knows |
| 08:48.52 | kergoth | just run rx on the max232 to the tx pin on whatever cable you're connecting out to |
| 08:48.58 | kergoth | and tx to rx, and .. |
| 08:49.06 | jacques | ok I'll try both ways |
| 08:49.23 | rwhitby-treo | kergoth: I tried oe bk developer access as per wiki, but it wanted a password. do I need to tell bk how to use my ~/.ssh/identity rsa1 key? |
| 08:49.57 | kergoth | i havent added the key yet |
| 08:50.01 | kergoth | bk just calls ssh |
| 08:50.20 | rwhitby-treo | ah. after that it should work? |
| 08:50.25 | kergoth | if you're as new to bk as you seem, i'd rather you gain a bit of competence with it first |
| 08:50.28 | kergoth | yeah |
| 08:50.33 | rwhitby-treo | right |
| 08:50.47 | rwhitby-treo | no problem |
| 08:51.14 | kergoth | keys.. remember to bk edit all files before making changes, so its locked for editing.. |
| 08:51.19 | rwhitby-treo | experienced with clearcase, rcs, cvs - just not bk yet |
| 08:51.32 | kergoth | and for checking things in, either manually use bk ci for each file, then commit to create a changeset, or use the X frontend, bk citool |
| 08:51.55 | kergoth | i usually 'alias e="bk editor"', and set EDITOR appropriately |
| 08:52.03 | kergoth | then use 'e' to edit everything |
| 08:52.13 | rwhitby-treo | thx - will be cmd line - using CoLinux |
| 08:52.30 | kergoth | bk ci -y"Fixed xyz" blahfile |
| 08:52.31 | kergoth | bk ci -y"Fixed xyz" blahfile2 |
| 08:52.38 | kergoth | bk commit -y"Fixed xyz in blahfiles." |
| 08:53.32 | *** part/#nslu2-linux rwhitby-treo (~upirc@203.202.55.177) |
| 08:53.45 | jacques | ok I powered on the nslu2 and I see a + |
| 08:53.52 | jacques | thae mean anything to anyone? |
| 08:53.54 | kergoth | yeah |
| 08:53.59 | jacques | oh cool |
| 08:54.14 | jacques | I am in redboot - I hit ctrl-C but it didn't do anything immediately |
| 08:54.29 | jacques | and when I went back to that screen it's in redboot |
| 08:54.32 | kergoth | the rx on the unit not connected right? or just a fluke? |
| 08:54.51 | jacques | what were you gonna say abot the + ? |
| 08:55.00 | jacques | because it "appears" to be working |
| 08:55.04 | kergoth | cool |
| 08:56.42 | kergoth | Kernel panic: No init found. Try passing init= option to kernel. |
| 08:56.43 | kergoth | hmm |
| 08:56.48 | kergoth | guess i shrunk that rootfs too much :) |
| 08:57.20 | jacques | hmm |
| 08:57.29 | jacques | boot does some stuff but doesn't actually boot |
| 08:57.51 | jacques | how do I just boot it from redboot ? |
| 08:58.58 | jacques | anyone? |
| 08:59.03 | kergoth | "just boot it" |
| 08:59.13 | kergoth | you mean exec the kernel manually? |
| 08:59.18 | kergoth | or..? |
| 08:59.48 | jacques | well pretty much every bootloader I have used on arm - hh.org, blob, uboot, I can say "boot" and it boots the kernel from flash |
| 09:00.07 | jacques | this one does: |
| 09:00.11 | jacques | RedBoot> boot |
| 09:00.11 | jacques | have eRcOmM |
| 09:00.11 | jacques | copy kernel code from flash to RAM |
| 09:00.11 | jacques | copy ramdisk file from flash to RAM |
| 09:00.11 | jacques | run kernel |
| 09:00.12 | jacques | RedBoot> |
| 09:00.47 | jacques | apparently I was looking got "go" |
| 09:00.59 | kergoth | boot works here, dunno |
| 09:01.25 | kergoth | god late night tv sucks |
| 09:01.47 | jacques | not sure if go alone would be enough - but boot then go works |
| 09:03.10 | kergoth | ah right, the boot script mine's using calls boot then 'exec' at the appropriate addy |
| 09:03.16 | kergoth | so yeah, guess thats how it works |
| 09:03.55 | kergoth | ~/code/oe/build-armeb$ ls -l tmp/rootfs/sbin/ldconfig |
| 09:03.56 | kergoth | -rwxr-xr-x 1 kergoth users 14276 2004-09-18 05:04 tmp/rootfs/sbin/ldconfig' |
| 09:03.59 | kergoth | :) much better |
| 09:05.01 | dyoung | go 0x1d00000 |
| 09:05.09 | dyoung | (assuming you put your kernel there) |
| 09:05.21 | dyoung | or you can do move go |
| 09:05.28 | kergoth | move go? |
| 09:05.29 | dyoung | or move exec |
| 09:05.37 | dyoung | move then go |
| 09:05.40 | dyoung | or move then exec. |
| 09:05.45 | kergoth | ah. |
| 09:06.21 | jacques | cool I'm probably going to be asking a lot of stupid redboot questions in the near term - unless there is a nslu2-specific doc somewhere |
| 09:06.36 | *** join/#nslu2-linux tking_ (~tking@adsl-66-159-198-116.dslextreme.com) |
| 09:06.57 | dyoung | ip_address -l 192.168.1.77 -h 192.168.1.14 |
| 09:06.58 | dyoung | load -r -v -b 0x1d00000 /nslu2/vmlinuz.sda |
| 09:06.58 | dyoung | load -r -v -b 0x1000000 /nslu2/ramdisk.g2.rc.gz |
| 09:06.58 | dyoung | go 0x1d00000 |
| 09:07.32 | dyoung | is just about it, unless you make a brick, then you'll look at the wiki page and find the fis write stuff handy. |
| 09:07.45 | kergoth | heh, i should do that, instead of writing it into flash every time i need to test something |
| 09:07.50 | jacques | dyoung is this assuming I have a tftpd properly configured? |
| 09:07.57 | kergoth | yep |
| 09:08.09 | jacques | :-\ |
| 09:08.10 | dyoung | Yes, although someome said you can use -m http; but I havnt tried that method yet |
| 09:08.24 | dyoung | (in which case it uses http...) |
| 09:08.27 | jacques | i wonder if gentoo's tftpd works any better than its tftp client |
| 09:08.41 | kergoth | http for boot stuff.. silly. thats what tftp and bootp are -for- :) |
| 09:09.01 | dyoung | Yes, which is why I havnt tried it yet. |
| 09:09.10 | jacques | http would sure help if you wanna get stuff off a dirferent subnet tho |
| 09:09.11 | dyoung | Although you could do some really sick things.... |
| 09:09.19 | rwhitby-around | I use http, cause then I suck it down from the same place that I put the binaries for others |
| 09:09.30 | dyoung | "Hey Jacques, try boot my kernel at http://some_ip_address/zImage |
| 09:09.33 | dyoung | " |
| 09:09.42 | kergoth | hehe |
| 09:09.46 | kergoth | sick indeed |
| 09:09.46 | jacques | that's not sick. that's very cool |
| 09:09.57 | kergoth | implementing http in the bootloader is sick |
| 09:10.28 | rwhitby-around | boot, move and go are SerComm additions |
| 09:10.34 | rwhitby-around | they know the Linksys addresses |
| 09:10.37 | jacques | implementing a dvd player in the bootloader is sick, http is cool |
| 09:10.45 | rwhitby-around | load and exec are standard RedBoot |
| 09:11.03 | jacques | why do they always have to add crap |
| 09:11.16 | jacques | i cant wait til we have jtag |
| 09:11.18 | rwhitby-around | exec remembers the entry address from the last load, so always load your ramdisk first, then the kernel, and you'll save some typing on the exec |
| 09:11.31 | tking_ | I'm going to work on that this weekend. |
| 09:11.34 | rwhitby-around | cksum is also your friend, and it works on the last load |
| 09:11.34 | jacques | rwhitby-around, cool tip |
| 09:11.35 | tking_ | JTAG. |
| 09:12.06 | jacques | rwhitby-around, you recommend I read the regular redboot docs? |
| 09:12.16 | rwhitby-around | yeah |
| 09:12.29 | dyoung | Oh yes. cksum. cksum is your best friend/ |
| 09:12.35 | kergoth | jtag will be great\ |
| 09:12.35 | rwhitby-around | forget about reading the fis stuff though - none of it except write and erase are implemented |
| 09:12.42 | jacques | ok |
| 09:12.45 | jacques | good to know |
| 09:12.48 | rwhitby-around | alternatively, just read the redboot source code that we have |
| 09:13.13 | rwhitby-around | the recovery howto has all you need to know about fis write |
| 09:13.48 | dyoung | I got most of that stuff in a text file that I just copy/paste from these days. |
| 09:13.55 | rwhitby-around | read the standard doco for dump and memcmp and stuff like that |
| 09:14.28 | jacques | when I was working on the gumstix I made kermit macros |
| 09:15.02 | kergoth | kermit's handy for that |
| 09:16.04 | jacques | yep. i dunno why i still use minicom |
| 09:16.25 | jacques | habit from the ipaq I guess |
| 09:16.38 | kergoth | i'd like to know why nearly every terminal emulator sucks |
| 09:16.40 | kergoth | i dont get it |
| 09:16.49 | kergoth | minicom, kermit, teraterm, hyperterminal, etc |
| 09:16.58 | kergoth | i miss the old nice DOS terminal emulators from the bbs days |
| 09:17.06 | kergoth | telemate and telix and shit |
| 09:17.09 | kergoth | heh |
| 09:17.15 | dyoung | ProComm Plus. |
| 09:17.20 | jacques | yeah I used shit all the time :-) |
| 09:17.24 | kergoth | yeah, procomm was great. powerful as hell |
| 09:17.29 | kergoth | it was a beast though |
| 09:17.31 | tking_ | procomm was awesome |
| 09:17.34 | dyoung | Heck I STILL use it. |
| 09:17.38 | rwhitby-around | anything better than tera term pro for windows? |
| 09:17.42 | kergoth | i wonder where i put my procomm floppies |
| 09:17.44 | kergoth | thatd be the way to go |
| 09:17.52 | rwhitby-around | [tera term pro is crap] |
| 09:18.02 | kergoth | tera term is small and simple |
| 09:18.04 | kergoth | which is good |
| 09:18.10 | kergoth | but its way tooooo simple |
| 09:18.11 | kergoth | heh |
| 09:18.13 | rwhitby-around | can't see macros though |
| 09:18.23 | dyoung | It makes me color-ls impared. |
| 09:18.24 | kergoth | yeah, it only supports a couple of the vt terminal types |
| 09:18.36 | kergoth | shit for file transfer protocol support as well |
| 09:18.49 | kergoth | i vote we get off our asses and add serial support to putty |
| 09:19.05 | tking_ | they have it already... |
| 09:19.07 | tking_ | puttytel |
| 09:19.09 | jacques | this damn linksys nslu2 doesn't like any of my linksys switches/routers - what the heck? |
| 09:19.15 | kergoth | puttytel can do serial? |
| 09:19.17 | kergoth | cool |
| 09:19.21 | tking_ | oops..I'mm wrong... |
| 09:19.22 | kergoth | does it suck? |
| 09:19.24 | kergoth | heh |
| 09:19.31 | kergoth | didnt think so. its open source, windows and linux.. |
| 09:19.35 | tking_ | lemme check taht again. |
| 09:19.42 | kergoth | definately a good project to add it to |
| 09:21.06 | kergoth | root@nslu2:~# df|grep -v Filesystem|head -n 1 |
| 09:21.07 | kergoth | /dev/ram0 8059 3403 4247 44% / |
| 09:21.09 | kergoth | hmm, better |
| 09:23.30 | kergoth | pfft, 400k /lost+found |
| 09:23.33 | kergoth | silly ext2 |
| 09:23.36 | rwhitby-around | kergoth: will be in a position to try your kernel and rootfs from oe in half to one hour (after dinner) |
| 09:23.42 | kergoth | down to 2991 |
| 09:23.46 | kergoth | rwhitby-around: cool |
| 09:25.23 | rwhitby-around | kergoth: what crosstool do you recommend, or does OE build it for me? |
| 09:25.29 | kergoth | oe builds it for you |
| 09:25.33 | kergoth | just oemake bootstrap-image and wait |
| 09:26.04 | dyoung | I've been waiting to type oemake bootstrap-image for a while. I keep screwing up the installation of linux on my new shiny buildbox. |
| 09:26.23 | dyoung | because the old crusty box is COMPLETELY full. |
| 09:26.33 | dyoung | I think I might have 100m free. |
| 09:26.42 | rwhitby-around | 2Gb free for oemake now |
| 09:26.42 | kergoth | not much left for me to trim in the image. looks like mkfs and fsck from e2fsprogs isnt going into bootstrap atm, need to correct that for nslu2 |
| 09:27.55 | kergoth | looks like around 150k in lrzsz, which is primarily useful in a debug environment, when we're having problems with the networking |
| 09:28.01 | kergoth | i could see removing that for a production image |
| 09:29.43 | kergoth | interesting, looks like the OE portmap package doesnt actually include a startup script to run the portmapper on start. does mount -t nfs start it automatically when necessary, or do we need a startup script? |
| 09:29.58 | dyoung | need startup script, I believe. |
| 09:30.31 | kergoth | hmm, seems strange if thats the case, i'd have thought pb_ would have tested this, since he's the one who added it to the stock task-bootstrap package set |
| 09:30.50 | dyoung | I'm not expert though, I may be wrong. |
| 09:31.31 | kergoth | i'll ask him later today |
| 09:33.55 | rwhitby-around | kergoth: |
| 09:33.58 | rwhitby-around | ERROR: Nothing provides nslu2-firmware-binary |
| 09:33.59 | rwhitby-around | ERROR: dependency nslu2-firmware-binary (for bootstrap-image) not satisfied |
| 09:33.59 | rwhitby-around | NOTE: no buildable providers for bootstrap-image |
| 09:34.11 | rwhitby-around | just after slugtool-native completes |
| 09:34.50 | rwhitby-around | Nothing to pull in oe and packages |
| 09:35.23 | rwhitby-around | removing tmp and trying again |
| 09:35.43 | kergoth | you didnt run bk -r co -q in packages after doing a bk pull. |
| 09:35.50 | rwhitby-around | did |
| 09:36.02 | kergoth | dunno what to tell you then. your environment is broken. |
| 09:36.12 | kergoth | ls packages/nslu2-binary-only/ |
| 09:36.19 | rwhitby-around | ok, I'll debug further |
| 09:36.38 | kergoth | oh, no, my environment is broken |
| 09:36.41 | rwhitby-around | no packages/nslu2-binary-only |
| 09:36.45 | kergoth | it helps if i'd actually push the .oe |
| 09:36.50 | kergoth | oops :\ |
| 09:36.57 | rwhitby-around | np |
| 09:37.25 | rwhitby-around | [that's the purpose of my testing :-)] |
| 09:37.29 | kergoth | since the oe CIA messages dont come here, i recommend idling in #oe |
| 09:37.34 | kergoth | there, bk pull and bk -r co -q in packages again |
| 09:38.52 | rwhitby-around | thx - got it |