**** BEGIN LOGGING AT Tue Feb 15 23:59:56 2005 Feb 16 00:18:53 hm Feb 16 00:18:55 ~hail siddy Feb 16 00:18:58 * jbot bows down to siddy and chants, "I'M NOT WORTHY!!" Feb 16 00:21:30 morning Feb 16 00:23:58 someone tried to crosscompile perl modules? (Like PerlMagick?) Feb 16 00:24:54 I haven't any luck trying it to build PerlMagick native Feb 16 00:40:37 I was able to install CPAN modules that compiled stuff. Hadn't tried anything outside of the CPAN shell environment. Feb 16 00:42:51 okay, i've figured out the initrd / switchbox / real-root-dev / pivot_root stuff now... i guess i'll write a wiki page on it soon Feb 16 00:44:15 giel: if you can figure out the weird pivot_root reboot problem that we have in switchbox 4 that would be great ... Feb 16 00:45:33 tell me about that problem? 'cause i guess i've solved it Feb 16 00:46:31 i know for sure the combination of init in openbox and linuxrc in switchbox 4 and the current command line parameters doesn't work Feb 16 00:46:34 and i know why Feb 16 00:46:39 so is that the problem? Feb 16 00:46:48 that the slug reboots after pivot_rooting? Feb 16 00:47:02 that is, the moment it tries to exec chroot init? Feb 16 00:47:40 no, we can pivot_root fine. Feb 16 00:48:20 the problem occurs when unslung is running after pivot_rooting to it, and then you type "reboot" and get a kernel panic instead of an orderly reboot Feb 16 00:48:31 oh that's the same issue yeah Feb 16 00:48:32 dyoung and glc know all about it Feb 16 00:48:36 i think Feb 16 00:48:49 kernel panic: init not syncing? something like that? Feb 16 00:49:29 the pivot_root works fine, and the slug is perfectly operational (running Unslung with web interface and everything) whilst pivoted. the problem only occurs when you try and reboot from that perfectly good running system Feb 16 00:49:50 well, i'll talk with dyoung and/or glc about it, but the solution probably lays in the command line parameters Feb 16 00:50:12 could be Feb 16 00:50:14 bbiab Feb 16 00:50:59 to be sure i need someone using switchbox 4 and unslung for testing\ Feb 16 01:44:26 jbot: seen christopher Feb 16 01:44:27 christopher is currently on #nslu2-linux. Has said a total of 7 messages. Is idling for 1d 55m 25s Feb 16 01:49:36 hmm Feb 16 01:49:51 I'm not in a position to do unslung-4/switchbox-4 testing right now. Feb 16 01:50:27 my DevSlug is busy doing bootloader stuff. So I either need to serial enable another one, or pay the 40min penalty. Feb 16 01:58:00 it can be done without serial :P Feb 16 01:58:12 anyway, let's wait if glc has the opportunity for some testing Feb 16 01:59:01 Whats the issue with the cmdline? Feb 16 01:59:29 well, the problem rwhitby described could well be because of init not having PID 1 Feb 16 01:59:55 using the pivot_root method, the command line should include root=/dev/ram0 init=/linuxrc Feb 16 02:00:15 if it doesn't, you're using a combination of the old real-root-dev and the new pivot_root, which imposes problems Feb 16 02:02:06 ofcourse, with a bootloader which mounts the root filesystem the whole switchbox/initrd stuff isn't necessary anymore Feb 16 02:05:33 Okay so you suggest that if I comment out the one line in switchbox-4 /linuxrc it will just work? Feb 16 02:06:07 I'll try paying the 20min penalty right now. Feb 16 02:06:15 (dunno if its gonna work though) Feb 16 02:23:05 uh? Feb 16 02:23:10 what one line? Feb 16 02:23:34 you should edit the kernel command line Feb 16 02:24:14 But you just said there was a conflict between the real-root-dev and the pivot_root. So eliminating the real-root-dev once and for all should have the same net effect. Feb 16 02:24:33 uhm, well, first of all the real-root-dev is misplaced in switchbox 4 Feb 16 02:24:47 but i was not referring to the actual real-root-dev command as well as to the method Feb 16 02:24:49 Its there for historical reasons, and I dont think we need it. Feb 16 02:25:06 and the pivot_root method includes the two command line parameters Feb 16 02:25:14 I'm just suggesting that eliminating the real-root-dev will eliminate any conflict. Feb 16 02:25:26 if those aren't there, the kernel is expecting the linuxrc to terminate Feb 16 02:25:32 (that is the easiest for me to test right now) Feb 16 02:25:39 i don't think that'll help Feb 16 02:25:47 having looked at the kernel source that is Feb 16 02:26:42 Drat. I was hoping to avoid having to recompile the whole stuff to test your suggestion. ;-) Feb 16 02:26:54 simpler put: if there's a root=/dev/ram0 init=/linuxrc on the commandline, the kernel knows you're planning to pivot_root and doesn't expect the linuxrc to end but to spawn init itself Feb 16 02:27:13 if you don't have those, the kernel is expecting linuxrc to set real-root-dev and to end, so the kernel can spawn init Feb 16 02:27:37 I'm just being lazy. Feb 16 02:28:13 ofcourse, we all are :) Feb 16 02:28:41 I'm into the JTAG verify; so I can fool with it shortly. Feb 16 02:28:47 (if it worked) Feb 16 02:28:56 k, don't hurry Feb 16 02:29:04 i'm at work, so i should be doing other stuff anyway :) Feb 16 02:29:16 I still have 0x10000 bytes to go. Feb 16 02:32:42 dead as a doornail. Feb 16 02:32:53 Guess i need to pay the full price. ;-) Feb 16 02:34:05 damnit Feb 16 02:35:23 at least you have jtag :) Feb 16 02:37:15 I'm gonna have to serial enable another one eventually. Feb 16 02:37:56 yeah, me too Feb 16 02:37:59 but being lazy, I was hoping to do that at the same time I fatten it. Feb 16 02:38:25 altho, actually when i'm running a stable openslug i'd like to actually use it Feb 16 02:38:32 so i won't need serial then Feb 16 02:39:38 I was thinking of having 2 for dev work. Feb 16 02:40:27 well, at first i wasn't planning any dev work at all Feb 16 02:40:39 but then i really wanted openslug once i found out about it :) Feb 16 02:48:31 There goes the first block. one more to go. Feb 16 02:52:55 gotta lunch Feb 16 02:52:57 good luck Feb 16 03:22:54 Ooh this is a nasty bug I didnt consider. Feb 16 03:23:39 maybe I shouldnt worry about it.... Feb 16 03:23:56 I think there are only a couple people that are going to be blowing away their sysconf intentionally. Feb 16 03:33:28 giel, okay confirmed that in unslung-4.x init is NOT pid 1 Feb 16 03:33:33 for me its pid 9 Feb 16 03:33:46 building a image based on your suggestion. Feb 16 03:34:52 giel: if this works, then we can mainstream switchbox 4 ... Feb 16 03:35:21 (after glc and myself have a chance to torture it a bit...) Feb 16 03:35:35 of course :-) Feb 16 03:35:59 I'm hopefuly that the LED issue will be resolved too. Feb 16 03:36:16 but my understanding is that and the LEDs are the last issues between us and an alpha of unslung 4.x Feb 16 03:36:29 yes. Feb 16 03:48:39 nice. Feb 16 03:48:49 It beeps now and the LED WORKS. Feb 16 03:49:46 nice Feb 16 03:50:04 check it in danno Feb 16 03:50:28 and I'll turn devslug back on to test it out later Feb 16 03:52:44 pushed. Feb 16 03:53:43 that was weighing heavily on my mind for the past weeks. Feb 16 03:54:01 now that its cleared up, we can plow ahead with impunity. ;-) Feb 16 03:54:21 Thanks Giel! Feb 16 03:54:30 same here - it had brought unslung firmware development to a standstill Feb 16 03:54:36 ~worship giel Feb 16 03:54:39 * jbot prostrates himself before giel. "I'm not worthy!" Feb 16 03:55:21 I'm also secretly happy that /dev/slug can finally go away. Feb 16 03:56:47 yep - a source of endless questions from people who know what they are doing :-) Feb 16 03:57:37 i take from this it worked? Feb 16 03:58:07 so says dyoung :-) Feb 16 03:58:18 oh, one note, if you're building a kernel with devfs enabled and automaticaly mounted, /dev/ram0 should be replaced by /dev/rd/0 Feb 16 03:58:38 or at least, so says werner Feb 16 03:58:55 (= the author of the pivot_root stuff Feb 16 03:58:55 ) Feb 16 04:02:19 Well, I'll be more enthusiastic after more involved testing, but I *can* say this now. Feb 16 04:02:24 1) init is pid 1 Feb 16 04:02:33 2) the LED obeys my command(s) Feb 16 04:02:36 3) It beeps Feb 16 04:02:47 which is way more than I could have said 2 hours ago. Feb 16 04:02:50 and what happens on reboot? Feb 16 04:03:02 It reboots. Feb 16 04:03:51 The next step is to comment out some of those ugly /proc/foo/panic lines and see if it still does the right thing. Feb 16 04:04:11 but that will have to wait, I'm getting sleepy already. Feb 16 04:11:50 I've rebooted through several permutations of (stuff) and have not seen the "panic" message yet. that tells me its probably working the way its supposed to. Feb 16 04:12:04 excellent! Feb 16 04:12:12 gastro is up and building ... Feb 16 04:12:41 Hmm... Recovery mode seems to have issues.. Feb 16 04:15:00 Heh Feb 16 04:15:02 Thats my fault Feb 16 04:15:13 When I rearranged linuxrc, I seem to have forgotten some stuff. Feb 16 04:18:11 Ah... Feb 16 04:18:18 Okay thats a unforseen issue. Feb 16 04:18:21 We can fix that. Feb 16 04:18:29 the ramdisk comes up as read-only Feb 16 04:18:40 so we cant cheat and symlink... Feb 16 04:18:41 oh, add rw on the kernel command line Feb 16 04:18:53 root=/dev/ram0 rw Feb 16 04:19:15 and in theory you don't need anything which says real-root-dev or panic Feb 16 04:21:14 Those are switchbox changes which I dont want to deal with today. I'll look at it tommrow. Feb 16 04:21:24 sleep time now. Feb 16 04:21:52 night dyoung Feb 16 04:22:16 gastro should have finished building the native packages by then ... Feb 16 04:22:58 night dyoung Feb 16 04:23:13 but you don't need switchbox changes to have the ramdisk mounted rw Feb 16 04:23:52 I think he was talking about removing all trace of real-root-dev .. Feb 16 04:24:01 .. and panic. Feb 16 04:24:16 ah ok Feb 16 04:24:26 i could do that tonight when i'm home Feb 16 04:24:39 where's the switchbox 4 repo? Feb 16 04:24:40 on sf? Feb 16 04:24:47 bk Feb 16 04:24:55 ah Feb 16 04:25:03 there's no hurry. Feb 16 04:25:03 bk://nslu2-linux.bkbits.net/unslung-4.x Feb 16 04:25:07 hm, i don't want to mess with bk yet, absolutely no idea how it works Feb 16 04:25:36 Its been stagnating so long, I really dont think one more day is going to make a difference. ;-) Feb 16 04:25:47 exactly what I was thinking. Feb 16 04:25:59 indeedy Feb 16 04:26:07 I can fiddle it in 10hours or so. Feb 16 04:26:09 well, have a good night's rest dyoung, you deserved it ;) Feb 16 04:26:22 Feb 3 is the last time I build unslung firmware ... Feb 16 04:26:24 Thanks for the input, its appreciated. Feb 16 04:26:37 hey, anything i learn i tell here Feb 16 04:26:44 that's the power of a community Feb 16 04:26:47 it seems so obvious now you've pointed it out :-) Feb 16 04:27:00 well, actually werner pointed it out Feb 16 04:27:04 so i had the same feeling Feb 16 04:27:12 I know that stuff, but just didn't connect it with the errors we were seeing ... Feb 16 04:27:23 although i almost had it by looking through the kernel source Feb 16 04:27:37 i knew exactly what went wrong but didn't yet knew how to correct it Feb 16 04:28:12 rwhitby: i've had those panic-on-reboot issues with a embedded platform at work, so i knew that one Feb 16 04:28:18 an* Feb 16 04:28:36 ah - the power of seeing something before ... :-) Feb 16 04:29:06 and ofcourse i didn't see them on my slug due to lack of serial, but i'm sure they were there :) Feb 16 04:29:55 next step i want to take is removing all mtd references from the nslu2 patches Feb 16 04:29:59 so i don't need mtd in my kernel Feb 16 04:30:17 but that's just for my specific case Feb 16 04:31:48 why? Feb 16 04:33:06 i just takes up precious memory Feb 16 04:33:07 :) Feb 16 04:33:14 i don't use the jffs2 at all Feb 16 04:39:15 rwhitby, geez youre too fast. Feb 16 04:39:38 ? Feb 16 04:40:03 I was gonna mod switchbox, change the cmdline to rw and push THAT as 4.1 ;-) Feb 16 04:40:16 so that'll be 4.2 Feb 16 04:40:35 no, push it as 4.1 too - we haven't announced 4.1 yet :-) Feb 16 04:40:45 4.1.1 :) Feb 16 04:40:53 whats that line.... Feb 16 04:41:01 "Commit early, commit often...." ?? Feb 16 04:41:20 yeah, and always take every opportunity to bump the most significant version number Feb 16 04:41:39 so it will be 4.2 :-) Feb 16 04:41:53 I have numerous "oops, forgot to bump the PR" commits. Feb 16 04:42:25 I dunno, if its not already 4.2 by tommrow we'll see. -) Feb 16 04:48:32 anyway, i've got 24mb free mem atm Feb 16 04:48:57 and i think i can have a system with hdd and ssh access and syslogging with 25mb free Feb 16 04:50:23 dyoung-zzzz: I'm going to fix the ?= in unslung.conf, and pull that into unslung-4.x ... Feb 16 05:12:56 03rwhitby 07 * r1.2988.1.1 10openembedded/conf/distro/unslung.conf: Updated unslung.conf to match new rules on overriding TARGET_FPU and TARGET_OS. Feb 16 07:03:57 03rwhitby * 10unslung/make/xemacs.mk: Bumped to 21.4.17 cause 21.4.16 has been withdrawn. Feb 16 07:04:32 03rwhitby * 10unslung/sources/xemacs/control: Bumped to 21.4.17 cause 21.4.16 has been withdrawn. Feb 16 08:27:30 hey folks! I did the next hack of that cool ixp425_eth driver: I added support for netconsole, anybody needing that? Feb 16 08:28:34 what is it? Feb 16 08:28:42 (so the obvious answer is: no) Feb 16 08:29:14 you can dump the console-messages via udp to your network. very useful for debugging without serial! ;) Feb 16 08:30:06 read $LINUX_SRC/Documentation/network/netconsole.txt Feb 16 08:30:10 ok Feb 16 08:30:37 ey, that could be nice Feb 16 08:32:13 i think i like it Feb 16 08:32:30 hehe, thought so. Feb 16 08:33:06 <[g2]> siddy, rock'n the house ! Feb 16 08:33:36 <[g2]> siddy, iirc that logs the kernel boot on the net right ? Feb 16 08:34:15 yes, if you load the ethernet-driver and the netconsole-module early enough you can see nearly the whole boot-messages. Feb 16 08:34:51 <[g2]> cool, can you paste a log in the wiki ? Feb 16 08:35:00 morning Feb 16 08:35:06 <[g2]> morning sunshine ! Feb 16 08:35:22 i really need one of those destroymornings.com tshirts Feb 16 08:35:32 <[g2]> how's that k'man this morning Feb 16 08:35:54 struggling toward consciousness Feb 16 08:35:56 you? Feb 16 08:36:13 [g2], I get the logs starting from the usb-initialization, so not the whole thing. I could change the module-loading a bit to get more I guess. Feb 16 08:37:29 that's what coffee is for Feb 16 08:37:36 coffee! Feb 16 08:37:43 speaking of which, where the heck is my coffee cup? Feb 16 08:38:21 <[g2]> fine, thx. [g2] is pretty much a 24-hours person -- I just sleep now and then and to sleep-camel every so often Feb 16 08:39:20 hehe Feb 16 08:39:25 * [g2] doesn't drink coffee but gets plenty of caffeine from soda Feb 16 08:39:32 and sugar Feb 16 08:39:35 <[g2]> nod Feb 16 08:39:48 [g2], but at first I'm waiting for my slug to crash once more and pray to the lord that this netconsole gives me some info on that... Feb 16 08:39:57 Well, I do have a 12 pack of Mt. Dew underneath the desk... Feb 16 08:40:03 blegh Feb 16 08:40:07 but still need the coffee hit in the morning Feb 16 08:40:24 ahhhhh Feb 16 08:40:28 same. mt dew doesnt pack quite enough of a punch early on Feb 16 08:40:37 Jolt Feb 16 08:41:03 and Mr. Pibb is like #2 after Mt. Dew for caffiene Feb 16 08:41:05 <[g2]> siddy, so you crash every now and then ? Feb 16 08:41:11 experimentally I've found that 4 mt. dews in a short period of time (15 minutes) = 2 cups of coffee Feb 16 08:41:52 [g2], yes, but it doesn't seem to be network-related this time. The box is completely dead from time to time, not even responding to the power-button. Feb 16 08:42:34 <[g2]> siddy, your network is pretty busy right ? Feb 16 08:43:27 * [g2] wonders aloud if there's a memory leak in the driver somewhere or one of the apps your are running Feb 16 08:43:44 [g2], not really. an uptream of 110kbit suffices for this crash. Feb 16 08:43:53 s/uptream/upstream/ Feb 16 08:44:21 <[g2]> you mean out of the box Feb 16 08:44:28 yes Feb 16 08:44:37 <[g2]> ftp ? Feb 16 08:44:48 seeding two torrents... ;) Feb 16 08:44:59 <[g2]> ctorrent ? Feb 16 08:45:02 yes Feb 16 08:45:31 <[g2]> strace ctorrent Feb 16 08:45:45 hm... a bug in userspace shouldn't crash the whole kernel. Feb 16 08:46:24 <[g2]> nod. it *shouldn't* Feb 16 08:46:59 <[g2]> does ctorrent run fine on Unslung ? Feb 16 08:47:14 I'll just wait and see what netconsole says on the next crash, will come in some hours I guess. Feb 16 08:47:24 <[g2]> GL Feb 16 08:47:58 <[g2]> siddy I liked the last wiki page ! Feb 16 08:48:56 <[g2]> siddy, what do you have to do to setup the netconsole ? Feb 16 08:49:33 [g2], you have to enable some kernel-options and patch the ixp425-driver to support it. Feb 16 08:50:14 * siddy is waiting for someone to shout "wiki"... Feb 16 08:50:17 ;) Feb 16 08:50:37 wiki! Feb 16 08:50:43 * [g2] whispers could please add that to the wiki ? Feb 16 08:51:19 hehe, will do so in the near future, busy atm. Feb 16 08:51:46 <[g2]> do you have the patch and a defconfig ? Feb 16 08:52:34 not yet. Feb 16 08:53:33 hm... my defconfig will probably be too weird for others, added and removed lots of stuff... I'll just post the neccessary config-changes. Feb 16 08:55:58 <[g2]> ok. It's nice to be able to build from a working set Feb 16 09:09:34 http://www.nslu2-linux.org/wiki/OpenSlug/EnableNetconsole Feb 16 09:10:09 [g2]: Feb 16 09:10:10 330 10/02 Greg KH ( 96) [ANNOUNCE] hotplug-ng 001 release Feb 16 09:10:14 * kergoth is catching up on email Feb 16 09:10:22 "I'd like to announce, yet-another-hotplug based userspace project: Feb 16 09:10:22 linux-ng. This collection of code replaces the existing linux-hotplug Feb 16 09:10:22 package with very tiny, compiled executable programs, instead of the Feb 16 09:10:22 existing bash scripts. Feb 16 09:10:23 " Feb 16 09:10:51 <[g2]> kergoth, THAT looks COOL! Feb 16 09:11:02 there used to be a project like that Feb 16 09:11:12 greg k-h and erik andersen had done work on it. diethotplug Feb 16 09:11:17 but it got bitrotted and stuff Feb 16 09:11:19 this looks cool Feb 16 09:11:24 Those are all static binaries, linked with klibc (which is included in Feb 16 09:11:24 the hotplug-ng package, just like udev.) Feb 16 09:17:18 kergoth: that's just what we need! Feb 16 09:17:30 'cause the hotplug scripts we use now are b-l-o-a-t-e-d Feb 16 09:17:33 i want it on my desktop too! Feb 16 09:17:36 hehe Feb 16 09:17:40 http://marc.theaimsgroup.com/?l=linux-hotplug-devel&m=110808263031219&w=2 Feb 16 09:24:08 thnxie Feb 16 09:24:44 <[g2]> kergoth, looks great, I'll have to try building that running it soon Feb 16 09:25:09 i'll add it to oe in a few minutes. Feb 16 09:25:10 <[g2]> I'm sure VoodooZ_work will really be interested to Feb 16 09:25:17 with a couple patches from the thread on the mailing list Feb 16 09:25:27 <[g2]> ~praise kergoth Feb 16 09:25:30 All hail kergoth! Feb 16 09:26:46 <[g2]> kergoth, I don't know if you've heard, but dyoung and I have run bewoolie's APEX bootloader and booted Unslung/Openslug off it Feb 16 09:27:53 <[g2]> it's teeny tiny .... <16K stripped Feb 16 09:28:32 cool Feb 16 09:28:35 glad to hear it Feb 16 09:28:47 * kergoth wants to try APEX on one of his omap boards one of these days Feb 16 09:29:05 <[g2]> I think you'd like it Feb 16 09:29:30 does apex need initrd or can it mount a rootfs directly? Feb 16 09:29:47 ? Feb 16 09:30:11 apex is a bootloader. it initializes hardware, sets up the tag list, etc, and launches a kernel Feb 16 09:34:37 i know what a bootloader is Feb 16 09:34:59 but redboot uses the initrd mechanism, while lilo and grub can do without Feb 16 09:36:28 what are you talking about? Feb 16 09:37:05 a bootloader doesnt need to mount the rootfs at all. if you're using an initrd, the bootloader usually has a mechanism to obtain it (flash, network, whatever) and load it into ram where the kernel can find it. its the kernel thats making use of it from there Feb 16 09:37:57 that is true Feb 16 09:38:07 The NSLU2 redboot doesn't use initrd. It's just been hacked to automatically load it into a certain part of RAM Feb 16 09:38:20 that's what i meant Tiersten Feb 16 09:38:20 It doesn't care or know what it's loading. Just that it's of length and goes to in memory Feb 16 09:38:25 right Feb 16 09:38:28 right Feb 16 09:38:43 hm, come to think of it... Feb 16 09:39:09 your initial question of "does apex need initrd or can it mount a rootfs directly?" doesn't really make sense Feb 16 09:39:15 i know Feb 16 09:39:25 it's completely not what i meant :) Feb 16 09:39:32 heh Feb 16 09:39:38 but i'm tired and hungry and making up excuses for sounding stupid :) Feb 16 09:39:44 go eat :) Feb 16 09:39:51 ignore this last line of remarks, please Feb 16 09:39:52 :) Feb 16 09:39:54 heh Feb 16 09:41:03 what i really meant was: is apex configurable to load or not load the initrd from flash to memory? Feb 16 09:42:31 don't see why it can't be Feb 16 09:42:44 redboot should have been configurable but they hardcoded it to do it Feb 16 09:43:44 loading the initrd from flash to memory is a feature. if it can do it at all, it will almost certainly be optional Feb 16 09:51:30 03jp30 * 10unslung/ (make/gconv-modules.mk sources/gconv-modules/control): add iconv binary to gconv-modules package Feb 16 09:59:02 hi, all, anyone like to help me test gconv-modules? Feb 16 10:01:48 ...not that there's much to test, it just packages up some bits of the toolchain that are needed for iconv() to be functional... and installs an iconv binary Feb 16 10:03:04 it would be helpful to hear from people using encodings whether the iconv binary is functional. Feb 16 10:19:31 kergoth: Good show! (re: hotplug-ng) Feb 16 10:19:56 I'm looking forward to trying that! Feb 16 10:26:02 03jp30 * 10unslung/Makefile: gconv-modules is ready for testing Feb 16 10:32:58 heh Feb 16 10:33:05 my sister has bought a slug Feb 16 10:33:07 bleh Feb 16 10:33:49 ~smack greg k-h for making some of the most common buildsystem mistakes in hotplug-ng (the same ones he makes in udev) Feb 16 10:33:52 * jbot smacks greg k-h for making some of the most common buildsystem mistakes in hotplug-ng (the same ones he makes in udev) upside the head. Feb 16 10:34:16 ~botsnack Feb 16 10:34:17 :), ByronT Feb 16 10:34:51 whatdiddidoo? Feb 16 10:34:55 i -really- need to write that list of the most common buildsystem mistakes Feb 16 10:35:00 someday.. Feb 16 10:35:08 wiki it! Feb 16 10:35:08 we need more hours in the day Feb 16 10:35:12 yeah, someday, we all need to do stuff! Feb 16 10:35:15 yeah! Feb 16 10:35:16 stuff Feb 16 10:35:17 I vote for a 28 hour day Feb 16 10:35:31 hmmmm stuff Feb 16 10:35:33 i like stuff Feb 16 10:35:47 preferably good stuff. Feb 16 10:35:51 i dont like bad stuff Feb 16 10:35:55 oh! gimme the good stuff babeh! Feb 16 10:35:59 hehe Feb 16 10:36:11 i'm gonna buy a car next week! that's good stuff! Feb 16 10:36:17 (or the week after) Feb 16 10:36:21 what car? Feb 16 10:36:28 presumably an audi tt Feb 16 10:36:32 but i might change my mind Feb 16 10:36:46 <[g2]> save your money an get an extra slug :) Feb 16 10:37:02 my friend, for that kind of money i could buy 1700 slugs :P Feb 16 10:37:19 WITH serial :) Feb 16 10:37:22 probably more that just 1700 slugs Feb 16 10:37:34 well, yeah, i ain't buyin' a new one Feb 16 10:37:39 <[g2]> giel, distcc's builds in 5 seconds ! Feb 16 10:37:42 i have money to spend, but not THAT mcuh Feb 16 10:37:52 [g2]: the problem will be the linking, then :) Feb 16 10:38:12 ah, audi. cool Feb 16 10:38:29 i'm going to see if i can afford a tt quattro Feb 16 10:38:33 4-wheel drive rocks Feb 16 10:39:31 anyway, this isn't #car-linux :P Feb 16 10:39:41 it was just some of the 'good stuff' Feb 16 10:40:05 i drive a chrysler crossfire. '04. got it relatively recently Feb 16 10:40:07 definately good stuff Feb 16 10:40:19 crossfire! good stuff! Feb 16 10:41:08 bought it yourself or lease? Feb 16 10:41:48 bought Feb 16 10:41:58 and got it into a car accident 2 months after i got it Feb 16 10:42:00 isnt life great? Feb 16 10:42:01 hehe Feb 16 10:42:05 ~stab murphy Feb 16 10:42:07 car accidents are cool Feb 16 10:42:08 * jbot runs at murphy with an origami Swiss Army knife, and inflicts a nasty paper cut. Feb 16 10:42:13 i mean, the rush of it! Feb 16 10:42:14 take that, murphy! Feb 16 10:42:24 adreneline is wonderful. much better than caffeine Feb 16 10:42:28 uhu Feb 16 10:43:12 oh, how i'm going to drive along small belgium mountainroads at unreasonably high speeds! Feb 16 10:43:20 good plan Feb 16 10:43:38 my other option is something completely different: a smart roadster Feb 16 10:44:18 it weighs next to nothing and you can chip it up to like 125bhp Feb 16 10:44:28 and it really sticks to the road Feb 16 10:45:25 * [g2] generated lots of adreneline when roadracing motocycles a while ago Feb 16 10:45:43 * [g2] seconds that it's better than caffine Feb 16 10:46:28 <[g2]> and less painful than accidents Feb 16 10:46:31 that's why i love to ride on my mountainbike through the city centre of amsterdam Feb 16 10:46:43 also, at unreasonably high speeds Feb 16 10:46:55 yay, unreasonably high speeds Feb 16 10:47:00 that's the keyword Feb 16 10:47:21 all three of those words are good, but together they're even better Feb 16 10:47:21 <[g2]> that's why biking is the most dangerous form of travel per passenger mile Feb 16 10:48:25 <[g2]> I definitely feel less comfortable on my Mountain bike at 40MPH+ than the motorcycle at 100MPH+ Feb 16 10:48:45 true Feb 16 10:48:52 what's your land speed record? Feb 16 10:49:18 <[g2]> I used to hit 45-46 on a hill on the way to work Feb 16 10:49:26 <[g2]> on the mountain bike Feb 16 10:49:47 i mean all time record Feb 16 10:49:52 any form of transportation Feb 16 10:50:11 while in control of the vehicle Feb 16 10:50:13 <[g2]> well planes fly at 500+ but they don't count Feb 16 10:50:16 well, sort of Feb 16 10:50:31 <[g2]> I've flown plane and that doesn't count either Feb 16 10:50:33 that's what's the 'land' is doing in 'land speed record' :) Feb 16 10:51:21 <[g2]> the 600cc bikes from when I was racing did 130MPH about the same as the old RX-7 Feb 16 10:51:59 155mph Feb 16 10:52:05 <[g2]> superbikes now go 180+ Feb 16 10:52:15 hm, all those mph, my head overheats Feb 16 10:52:24 i'm used to kmh Feb 16 10:52:27 93 Honda Prelude VTEC.. that's all the speedo would show and it was limited Feb 16 10:52:33 <[g2]> 62.5mph = 100 Kph Feb 16 10:52:41 i know the ratio, i'm just lazy Feb 16 10:52:43 that's with me in control... Feb 16 10:52:56 with you *out* of control? Feb 16 10:53:02 driving Feb 16 10:53:47 well, i was doing 230km/h (my guess is that's 150-155mph) in one of my previous cars Feb 16 10:53:49 <[g2]> Well I was dragging a knee at 130MPH or 400+KPH so there :) Feb 16 10:54:04 in a friend's super moded s2000 (dual turbos, 700+ horse at flywheel, 487 hp at wheels) the speedo wouldn't show over 155mph Feb 16 10:54:05 which was http://www.safekimono.nl/veenendaal.jpg Feb 16 10:54:24 <[g2]> s/400/200/ Feb 16 10:54:30 I almost needed new shorts... Feb 16 10:54:52 what is that kind of crappy speedo in such a car :P Feb 16 10:55:49 http://www.toddcomp.com/S2000.html <- a stealth supercar Feb 16 10:55:53 http://www.safekimono.nl/zandvoort.jpg Feb 16 10:56:00 who needs a speedo? just judge the speed by the screams of your passenger Feb 16 10:56:36 4 cylinder that out produces the Lotuses of years ago Feb 16 10:56:37 who needs passangers? only extra weight! Feb 16 10:57:19 heh Feb 16 10:57:34 passenger, victim and ballast are all interchangable Feb 16 10:58:02 anyway, i'll let you know when i've bought a car, 'cause i guess i won't be touching the slug much for some time :) Feb 16 10:58:13 that s2000 needs the extra weight... at full throttle it feels like it's about to get airborne Feb 16 10:58:27 ByronT: that's what spoilers are for Feb 16 10:58:40 <[g2]> Ok since we are already *way* off topic -- who's got a CAN to usb 2.0 adapter so the slug can capture all the realtime data ? Feb 16 10:58:42 but that would take away the stealthiness Feb 16 10:58:59 Passenger == Ballast... Feb 16 10:59:07 ehm... a retractable spoiler! Feb 16 11:00:34 i hope this is going to be my next car http://static.marktplaats.nl/fotos/autos/audi/2/45668.jpg Feb 16 11:04:22 [g2]: that's what the GBA is for! realtime engine management readings :) Feb 16 11:05:10 heh, get that gba cart with the fpga, attach a CAN controller to it.. Feb 16 11:05:14 bleh Feb 16 11:06:16 " 700+ horse at flywheel, 487 hp at wheels" ? Feb 16 11:06:34 dyno readings Feb 16 11:06:46 it loses 213HP in the driveline? Feb 16 11:06:59 I cannot conceive of such a thing Feb 16 11:07:20 30-40HP driveline loss is typical in a car like that Feb 16 11:07:28 guesstimates from what I understand... American car makers always give flywheel hp measurements Feb 16 11:07:33 where does all that energy go? Feb 16 11:07:48 converted into heat it would melt the vehicle Feb 16 11:07:50 have no earthly.. Feb 16 11:08:08 I revise my statement then to 487hp at wheels Feb 16 11:08:14 :) Feb 16 11:08:23 with racing fuel of course Feb 16 11:08:41 with just premium fuel the wheel hp drops to 410 or so Feb 16 11:08:43 you know what's a good thing? i live 50km of the german border Feb 16 11:08:53 and you know what the speed limit is on the german highways? Feb 16 11:09:49 *sigh* not another mad dutch on our pretty autobahns... ;) Feb 16 11:10:07 hey, i think i'm entitled to it when driving a proper german car! Feb 16 11:11:12 hm... you desecrate that audi with your yellow licence plate! :-p Feb 16 11:11:36 hey, did you recognize the two earlier pics? Feb 16 11:11:44 that's a german car, too Feb 16 11:11:46 which ones? Feb 16 11:11:58 http://www.safekimono.nl/veenendaal.jpg http://www.safekimono.nl/zandvoort.jpg Feb 16 11:12:15 i modified it somewhat, but it still is german :) Feb 16 11:12:37 err... opel? Feb 16 11:12:48 yeah Feb 16 11:12:53 a manta Feb 16 11:12:53 manta? Feb 16 11:12:56 correct :) Feb 16 11:13:05 lol, perfect! Feb 16 11:13:09 you've just identified yourself as a real german :) Feb 16 11:14:16 it's missing the "fuchsschwanz"! Feb 16 11:14:32 it does have the 180bhp :) Feb 16 11:14:36 and the '76 corvette hood Feb 16 11:17:30 I can't wait for april to get here (canada) so I can take my baby (Kawa Ninja ZX6) out!! Feb 16 11:18:41 i can't wait for july, so i can get to canada Feb 16 11:18:42 L) Feb 16 11:19:19 giel, going to OLS? Feb 16 11:19:58 yep Feb 16 11:20:00 that's the plan Feb 16 11:20:02 cool Feb 16 11:20:05 my plan too Feb 16 11:20:10 have to write my paper / get it accepted etc Feb 16 11:20:27 cool, let's meet up Feb 16 11:20:37 Cool. perhaps we'll see each other there. Feb 16 11:20:50 I live in Ottawa. Feb 16 11:20:58 yeah, let's drink to that Feb 16 11:21:23 i'm sure i'll be around. after going last year, i just need to go again Feb 16 11:21:23 VoodooZ_work, you going to OLS? Feb 16 11:21:32 how much does OLS cost though? I know the guys organizing it so perhaps I'll try to get in for free. Feb 16 11:22:00 jacques: I'd love to go but we'll see if I can get my work to pay for me. Feb 16 11:22:19 we should have a BOFS Feb 16 11:22:25 yeah! we should! Feb 16 11:22:38 i mean, we are ;) Feb 16 11:22:45 ~BOFS Feb 16 11:22:52 ~BOF Feb 16 11:23:34 cosmicpenguin (another oe user, geode), pb_, and others will be around. should do an OE bof or something Feb 16 11:23:49 I remember going to the Black Thorn party a few years back and meeting guys like Alan Cox. Pretty cool. Feb 16 11:24:19 cool I'd like to see cosmicpenguin Feb 16 11:24:21 well, i have some talke to do with andrew Feb 16 11:24:25 talking* Feb 16 11:24:38 Andrew Hutton you mean? Feb 16 11:24:42 morton Feb 16 11:24:45 I'd like to see pb but not sure he's going this year since he's going to australia in april Feb 16 11:24:53 ahh right Feb 16 11:24:55 forgot about that Feb 16 11:25:31 * kergoth thinks he should probably do something to increase his visibility in the community Feb 16 11:25:43 what about submitting a paper :) Feb 16 11:25:54 or wear a bright yellow suit Feb 16 11:25:56 hehe Feb 16 11:26:22 and then end mails to th LKML with "kergoth, the guy in the bright yellow suit" Feb 16 11:26:30 proposal deadline already passed :-( Feb 16 11:26:36 yep Feb 16 11:26:42 my proposal is still 'pending' Feb 16 11:26:51 but i'm not really worried about getting accepted Feb 16 11:27:42 embedded seems to be fairly low profile, all things considered Feb 16 11:27:51 lots of big iron and desktop mentality Feb 16 11:29:11 i'm off Feb 16 11:29:24 gotta relax for a bit Feb 16 11:29:39 want to turn in early, presentation tomorrow morning Feb 16 11:30:39 later giel Feb 16 11:31:02 yeah, ttyl Feb 16 11:39:24 <[g2]> giel, GL with the presentation Feb 16 11:42:59 Hi people, I'm stuck my foot in the libusb problem, and I wonder if anyone have some advice Feb 16 11:57:46 was that a freudian slip? Feb 16 11:59:52 Janne: I don't think libusb works Feb 16 12:01:01 [g2], any success with eall or still waiting? Feb 16 12:01:35 Janne: peteru used libusb initially for the Puppy app but found it to be flakey. I think he wrote his own routines Feb 16 12:05:26 <[g2]> christopher, I think ka6sox is shipping the FatSlug today Feb 16 12:12:34 :) Feb 16 12:12:38 bbs Feb 16 12:21:12 ok, where do I start with writing my own usb-routine? Feb 16 12:35:45 Janne: *shrug* Talk to peteru Feb 16 12:44:36 ok Feb 16 14:20:23 make/libusb.mk has my comments on the root cause of libusb issues. Given my limited amount of available time back then, I declared libusb a lost cause for the time being and put it on the back burner. Feb 16 14:21:16 I don't have any immediate plans to return to it - perhaps there's more hope for libusb2, which I believe is in development. Feb 16 14:33:01 oooooh... HitchHiker's Guide to the Galaxy preview at amazon.com Feb 16 16:59:11 03djf * 10unslung/sources/mutt/control: added deps: cyrus-sasl, libdb Feb 16 17:32:16 * rwhitby-web joins nslu2-linux-hot-cars .... Feb 16 17:33:19 lol Feb 16 17:33:50 so will netconsole go into unslung-able-4.x ? Feb 16 17:34:15 ? Feb 16 17:36:33 before all the land speed record talk, siddy announced that he had netconsole working Feb 16 17:36:42 Oh Feb 16 17:38:16 Unfortunatley, I didnt setup my irc logs by date, so I dont know what day im looking at anymore Feb 16 17:38:19 heh Feb 16 17:38:50 anyone tried the gdp ipk? Feb 16 17:39:00 Oops should be gdb... Feb 16 17:41:03 My first attempt messed up if ncurses was present but I think that's fixed now. Feb 16 17:43:37 shenson: gdb hasn't been added to one of the READY_FOR_TESTING Makefile variables yet, so assume it's not ready for testing.... Feb 16 17:43:50 ... so we assume .... Feb 16 17:45:06 Erm its in NATIVE_AND_CROSS_PACKAGES_READY_FOR_TESTING Feb 16 17:45:35 oh - was it added in the last 12 hours? Feb 16 17:46:04 * rwhitby-web didn't have a chance to cvs up before leaving for work this morning Feb 16 17:47:36 Log says date: 2005/02/15 18:00:30. Feb 16 17:47:56 I'm new to this, just wondering what the procedure was. Feb 16 17:48:53 Yep, I just checked and it is there. Weird that CIA didn't report it in the logs (which I did get a chance to check this morning, and gdb wasn Feb 16 17:48:59 't mentioned in them) Feb 16 17:49:14 you've followed the procedure correctly :-) My mistake ... Feb 16 17:52:13 Ah good, thanks. Please email me if there's any problems. Feb 16 17:52:31 building now if you can wait ... Feb 16 17:52:52 Yes sure. Native build takes a while :-) Feb 16 17:53:00 building cross :-) Feb 16 17:55:05 I don't get the chance to use IRC much. My "day job" is OpenSSL. Feb 16 17:56:16 :P Feb 16 17:59:33 so did we package openssl.ipk correctly ;-) Feb 16 17:59:38 newbie bb q: i think i have everything set up, but i'm getting "ERROR: Nothing provides openslug-image" (used openslug instead of unslug when mentioned) Feb 16 18:00:37 Yeah looks good. I've added some fixes for 0.9.7f which should make the build easier. Feb 16 18:00:48 cool Feb 16 18:01:25 cpage_, huh? Feb 16 18:01:32 cpage_: probably an error in your local.conf file - check the variables there carefully. Feb 16 18:01:35 ok Feb 16 18:01:38 Is it right that the crypto accelerator stuff doesn't work on the slug? Feb 16 18:01:48 it's a 420, not a 425 Feb 16 18:02:01 (even though the kernel reports a 425) Feb 16 18:02:08 building gdb cross I get: utils.c:61:20: curses.h: No such file or directory Feb 16 18:02:41 my build is still going ... Feb 16 18:02:45 cpage_, I don't really understand what you said, but I suspect you may need DISTRO = "openslug" Feb 16 18:03:06 rwhitby, did you have a chance to bk heave-ho that unslung.conf change? Feb 16 18:03:23 jacques - i had that... trying the reference local.conf in the wiki Feb 16 18:03:25 dyoung-web: yep Feb 16 18:03:34 hmm Feb 16 18:03:41 you pu shed it to unslung-4.x repo? Feb 16 18:03:45 dyoung-web: seemed that it was already fixed in unslung-4.x ... Feb 16 18:03:50 Its telling me theres nothing to heave ho. Feb 16 18:03:53 oh.. okay gotcha. Feb 16 18:04:17 I've been periodically heave-ho-ing from the main repo. Feb 16 18:04:21 it built for me at least, but I didn't get a chance to test it (Elise woke up) Feb 16 18:04:39 I built tested 14 hours ago, it was working then. Feb 16 18:05:27 There was a curses.h problem which I think I've fixed. Feb 16 18:06:05 my gdb cross build finished successfully. Feb 16 18:06:30 Ah good. It save me hours tracing a bug in ctorrent. Feb 16 18:06:50 garpinc: I think there is a version problem with gift-opennap - it is continually repacking the ipk - can you fix ? Feb 16 18:08:18 shenson: gdb ipk will be uploaded to the cross feed in the next half hour or so (it's doing a full upload of all packages, so it takes a while) Feb 16 18:10:05 jacques: this will include the new crosstool-native version 5 which has the syslimits.h added to it. We can test it with a full rebuild of native and cross packages on gastro .... Feb 16 18:10:22 OK, thanks. Feb 16 18:11:05 rwhitby-web, check private comm Feb 16 18:11:40 jacques - thanks - seems to be building now Feb 16 18:14:07 I'd better get off. Its 2am here and I've lots of "SHA1 hacked" panic to look forward to tomorrow. Feb 16 18:15:07 is this something we've heard here first ? Feb 16 18:16:38 back later Feb 16 18:17:09 It was announced at the RSA conference but its nowhere near as serious as some Feb 16 18:17:12 are making out. Feb 16 18:20:35 cpage_, what did you change to make it work? Feb 16 18:25:50 jacques - i replaced the modified local.conf from bitbake with the one on StepByStepSetup Feb 16 18:25:58 cpage_, ah ok Feb 16 22:28:28 hey, anyone know how to enable objc in "crosstool", I am going to try GCC_LANGUAGES="c,c++,objc" in nslu2-cross335.sh. Feb 16 22:30:24 the goal is to have a crosstool-native gcc compiler with objective-C enabled. Feb 16 22:58:45 just booted openslug... ssh'd in Feb 16 22:59:23 add yourself to the openslug group on groups.yahoo.com/nslu2-linux Feb 16 22:59:30 there is a database section. Feb 16 22:59:43 i'm already in the group Feb 16 22:59:50 cool Feb 16 22:59:59 I mean add your name to the group of opensluggers Feb 16 23:00:03 ya Feb 16 23:02:35 added Feb 16 23:02:46 cool Feb 16 23:02:54 I keep forgetting to add my name. Feb 16 23:03:21 man... ixp425's ethernet driver throws interrupts like crazy Feb 16 23:03:29 *ethernet portion? Feb 16 23:03:42 yes...about 4k/sec Feb 16 23:03:46 wow Feb 16 23:03:48 or maybe 5k/sec Feb 16 23:04:00 what'd happen if the interrupt was turned off and we simply polled the thing? Feb 16 23:04:20 heya tjyangaway Feb 16 23:04:29 cya soon! Feb 16 23:05:11 ka6sox, I am back, almost ready for bed. Feb 16 23:05:37 I should be headed that way too. Feb 16 23:05:51 but I'm working on a problem. Feb 16 23:06:31 me too, I forgot how to trigger the compilation of crosstool ie toolchain. Feb 16 23:06:49 is it already built? and you want to rebuild it? Feb 16 23:07:06 yes,sir, with objective-C language enabled. Feb 16 23:08:01 okay you will need to go to your builds directory and throw away the .ipk and the directory for the crosstool build. Feb 16 23:08:08 then you can make crosstool Feb 16 23:08:30 tried that, I ecen move build to build.old Feb 16 23:08:38 hmmm Feb 16 23:08:39 ecen->even. Feb 16 23:09:07 I need to look at the current makefile...brb Feb 16 23:09:30 thanks, that is ok. may be I am tired should go to bed. Feb 16 23:10:06 strugling with sleep head is no use ;) Feb 16 23:10:07 rm builds/crosstool-native* packages/crosstool-native* Feb 16 23:10:19 rm -rf even.... Feb 16 23:10:24 make crosstool-native Feb 16 23:10:53 not make crosstool-naive yet unstil make crosstool is ok. Feb 16 23:11:15 oh, okay then s/-native//g Feb 16 23:11:30 actually n t Feb 16 23:11:32 not Feb 16 23:11:38 you need to remove it from the toolchain directory. Feb 16 23:12:26 or something along those lines. Feb 16 23:12:30 I did this armv5b-softfloat-linux.old in toolchain even. Feb 16 23:13:29 ok I will give it another try since there are two helping hands here. Feb 16 23:13:32 ah, but crosstool directory thinks its built Feb 16 23:13:43 because toolchain/crosstool probably has a .built file in it Feb 16 23:14:34 so you will need to trick it by removing the .built and/or .configured; or just mv crosstool crosstool.old Feb 16 23:15:59 you know what ? move unslung unslung.old;cvs -d:ext:tjyang@cvs.sourceforge.net:/cvsroot/nslu co unslung is quicker. Feb 16 23:20:30 dyoung, rm toolchain/crosstool/.built did the trick, thanks Feb 16 23:21:11 Cool. Feb 16 23:21:13 ;-) Feb 16 23:21:46 You might have to remove the .configured too since you are trying to compile it differntly. Feb 16 23:23:49 hmm, I can't find .configured under toolchain/crosstool/* Feb 16 23:24:20 sorry, found it. Feb 16 23:25:08 [tjyang@dual crosstool]$ grep objc *.sh Feb 16 23:25:08 nslu2-cross335.sh:GCC_LANGUAGES="c,c++,objc" Feb 16 23:25:08 [tjyang@dual crosstool]$ Feb 16 23:25:39 what is objc anyways? Feb 16 23:25:40 am I modifing the right .sh file ? Feb 16 23:25:56 Objetive-C Feb 16 23:26:52 Yes, I believe thats the right file Feb 16 23:27:13 gnustep.org community like to have objc support.so they can port gnustep over to nslu2 ;) Feb 16 23:27:27 Anyone know what the story is with svn.berios.de tonight? Feb 16 23:27:39 berlios.de rather. Feb 16 23:28:53 make toolchain is compiling, time for bed. night all. Feb 16 23:29:02 Okay good luck! **** ENDING LOGGING AT Wed Feb 16 23:59:56 2005