**** BEGIN LOGGING AT Wed Jan 17 02:59:57 2007 Jan 17 05:25:49 ~seen cdoban Jan 17 05:26:05 i haven't seen 'cdoban', rwhitby Jan 17 05:52:26 03bzhou * r5208 10optware/trunk/make/multitail.mk: multitail: 4.2.0 -> 4.3.0 Jan 17 05:55:59 03bzhou * r5209 10optware/trunk/make/mpg123.mk: 0.63 -> 0.64 Jan 17 08:26:51 Hello, i'm trying to compile the 2.4 kernel with USB support for an IXP425 processor. First off, is this the right place for the question?(hint: bewildered newbie here :) Jan 17 09:40:11 ravehanker: Why a 2.4 kernel? Jan 17 09:52:05 NAiL:- My 2,6 isn't compiling itself Jan 17 09:52:23 NAiL:- I'm trying to Compile Snapgear for Intel IXP425 processor Jan 17 09:54:23 NAiL:- Assembler messages: Jan 17 09:54:24 Error: unrecognized option `-marmv5te' Jan 17 09:54:39 This is the error i get when i try to compile the 2.6 kernwl Jan 17 09:54:43 *kernel Jan 17 09:54:51 ravehanker: which toolchain you use? Jan 17 09:55:19 hrw|work:- The one i got from the vendor. Hold on i'll find out what it's called Jan 17 09:56:53 ravehanker: use normal one instead of vendor one Jan 17 09:57:11 hrw|work:- Ah, the one in the snapgear site? Jan 17 09:57:31 hrw|work:- arm-linux-tools-20030927.tar.gz --> This is what i'm using right now Jan 17 09:58:47 hrw|work:- The site shows a uclinux tool chain and a Linux tool chain which one should i use? Jan 17 10:01:30 ravehanker: you want ixp425 support which is armv5te. 20030927 means gcc 2.95 which do not know about any ARM core other then armv4 Jan 17 10:01:46 ravehanker: use gcc 3.4 or 4.1 for 2.6 kernel Jan 17 10:02:18 hrw|work:- I do have gcc4.1 running on my system, which of these two tool chains display on the main site should i use? Jan 17 10:02:55 *displayed Jan 17 10:11:43 ravehanker: the problem is that I do not track nslu2-linux development Jan 17 10:12:39 hrw|work:- Well,I'm happy to tell you that what i'm going is not exactly nslu2-linux dev. It has got more to do with compiling linux kernel on snapgear with USB support :) Jan 17 10:12:56 Which is why i'm still wondering if i'm at the right place :-/ Jan 17 10:13:05 * ravehanker doesn't want to get flamed for ignorance Jan 17 10:15:38 ravehanker: I think you have two options: Jan 17 10:15:54 1) Ask Snapgear for help with their vendor kernel and toolchain that you are using Jan 17 10:16:09 2) Use a stock linux kernel and a stock linux toolchain. Jan 17 10:16:19 3) Use OpenEmbedded Jan 17 10:16:48 there's probably other options, but none other than #1 would involve using the vendor kernel or toolchain. Jan 17 10:17:14 rwhitby:- what's a stock linux kernal? paid support like Monta Vista? Jan 17 10:17:24 no, the one you download from www.kernel.org Jan 17 10:17:47 rwhitby:- That's seems like a far shot. What is OpenEmbedded btw? Jan 17 10:17:57 ~openembedded Jan 17 10:18:07 i guess openembedded is a flexible build tool based on Gentoo's package manager (portage), and a large upstream package repository and build classes. It facilitates building embedded Linux distributions and packages for any arbitrary target platform. http://www.openembedded.org Jan 17 10:18:47 rwhitby:- Thanks! Jan 17 10:18:55 * rwhitby always thought OpenEmbedded was based on Debian packaging, not Gentoo ... Jan 17 10:19:19 (well, ipkg based on dpkg ....) Jan 17 10:19:32 rwhitby:- Debian would be more likey it. I grew up in that neighbouhood :) Jan 17 10:19:55 rwhitby: the build system is quite similar to portage. The package system is somewhat similar to dpkg Jan 17 10:20:33 rwhitby: oecore^Wbitbake was modeled on portage Jan 17 10:23:51 rwhitby, could you please rebuild oleg and ddwrt repository Jan 17 10:24:29 oleo: sure. is the new contents incompatible with the old contents, and is there a migration plan in place with the users? Jan 17 11:51:34 rwhitby : Oleg suggested rebuild a while ago. I also noticed that some packages segfaulted (svn). There were also problems with linking libs produced by gcc 3.4.6. Users are notified in Optware-uClibcBuild page. Migration plan for wl500g->oleg is aranged with Oleg. Jan 17 11:51:59 Maybe this will happen with 7g release of his firmware. Jan 17 11:59:38 Hmmmm ... busybox-1.3.1-r0 refuses to build while building debianslug Jan 17 11:59:53 NOTE: make -e MAKEFLAGS= CROSS=arm-linux- oldconfig Jan 17 11:59:53 make[2]: Entering directory `/home/pascal/slug/debianslug/tmp/work/busybox-1.3.1-r0/busybox-1.3.1' Jan 17 11:59:53 make[2]: *** No rule to make target `oldconfig'. Stop. Jan 17 11:59:53 make[2]: Leaving directory `/home/pascal/slug/debianslug/tmp/work/busybox-1.3.1-r0/busybox-1.3.1' Jan 17 11:59:53 FATAL: oe_runmake failed Jan 17 12:14:59 It shouldn't be trying to build busybox 1.3.1 at all Jan 17 12:43:52 03operron * r5210 10optware/trunk/make/ (busybox.mk ds101j-kernel-modules.mk smartmontools.mk): ds101j-kernel-modules: first checkin Jan 17 12:47:10 NAiL: any clue what's going on there in that case ? Jan 17 12:55:11 03operron * r5211 10optware/trunk/make/busybox.mk: busybox: Undoing change committed inadvertently in r5210 Jan 17 12:56:52 03operron * r5212 10optware/trunk/make/smartmontools.mk: smartmontools: Undoing change committed inadvertently in r5210 Jan 17 13:02:52 03operron * r5213 10optware/trunk/sources/ds101j-kernel-modules/ (. armeb-config): ds101j-kernel-modules: first checkin Jan 17 13:13:29 03operron * r5214 10optware/trunk/Makefile: ds101j-kernel-module: added to CROSS_PACKAGES_READY_FOR_TESTING Jan 17 14:37:45 mwester: ping Jan 17 15:23:52 Caelian|work: Yeah. I committed the .bb with DEFAULT_PREFERENCE_slugos = "1", which I didn't mean to do :P Jan 17 15:27:25 Caelian|work: I've committed the fix now. Should be fixed in a couple of minutes Jan 17 15:29:52 Caelian|work: Try syncing now ;) Jan 17 16:50:17 03bzhou * r5215 10optware/trunk/make/adduser.mk: adduser: up IPK_VERSION to trigger rebuild (reverse -r5210 impact) Jan 17 19:18:33 03bzhou * r5216 10optware/trunk/make/libcurl.mk: libcurl: rm staged libcurl.la Jan 17 21:09:40 03oleo * r5217 10optware/trunk/ (make/transmission.mk sources/transmission/transmissiond.c): transmission: r1320->1378 Jan 17 21:09:40 03bzhou * r5218 10optware/trunk/ (Makefile make/libnxml.mk): added and promoted libnxml Jan 17 21:11:00 03bzhou * r5219 10optware/trunk/ (Makefile make/libmrss.mk): added and promoted libmrss Jan 17 21:12:04 03bzhou * r5220 10optware/trunk/ (Makefile make/rsstail.mk): added and promoted rsstail Jan 17 21:30:29 need to rename some ipks, rpm knows a directive "obsoletes ", does ipkg do something like this? Jan 17 21:40:48 gda_: i'm not aware of ipkg built-in feature for that Jan 17 21:41:17 we add the old package to optware/Makefile PACKAGES_OBSOLETED Jan 17 21:41:27 eno: thank, I will risk the complains Jan 17 21:42:17 the problem is that the makefile name doesn't change but the sub packages Jan 17 21:43:08 eno: ds101g-kernel-modules makes now kernel-modules-* Jan 17 21:43:16 then you don't even need to add to PACKAGES_OBSOLETED Jan 17 21:43:28 It should be kernel-module-* Jan 17 21:43:57 autoclean.pl can handle that right now Jan 17 21:43:58 * rwhitby installs a 2.6.20-rc5 Debian kernel on the nas100d Jan 17 21:44:48 subpackages no longer need to be named with 'primary-package-...' Jan 17 21:45:30 people with kernel-modules-nfsd will never get noticed that kernel-module-nfsd was updated Jan 17 21:46:03 eno: what do you mean with "subpackages no longer need to be named with 'primary-package-...'" Jan 17 21:46:12 that's unfortunely true Jan 17 21:46:14 the chance of a kernel module getting updated, without changing the kernel version, is small. Jan 17 21:46:38 CONFLICTS is what you want, I think. or there might even be a SUPERCEDES Jan 17 21:47:04 rwithby: this is right, but maybe, some day, I will add a script that loads the module with the right options Jan 17 21:47:24 (best to ask in #oe for the right syntax to obsolete with ipkg) Jan 17 21:47:25 gda_: it was a limitation of scripts/optware-autoclean.pl scripts, see -r4570 Jan 17 21:50:40 ixp4xx_mac driver 0.3.1: eth0 on NPE-B with PHY[0] initialized Jan 17 21:52:59 nice Jan 17 21:55:36 * NAiL compiles Jan 17 21:56:15 gda_: there is a Replaces: header Jan 17 21:58:15 eno: this is what it asked just some moments before in #oe :) Jan 17 21:58:27 2.6.20-rc5 debian kernel works Jan 17 21:59:33 03rwhitby * r83 10debian/trunk/ (5 files in 3 dirs): Debian 2.6.20-rc5 experimental kernel Jan 17 22:00:13 svn co the "debian" repo from nslu2-linux, and type make to build the debian experimental kernel. then just dpkg -i that on the slug Jan 17 22:01:18 (note that most people should just wait until the next 2.6.18 kernel is uploaded by tbm, which contains the same stuff back-ported to 2.6.18) Jan 17 22:01:35 You only need 2.6.20 for the nas100d support. Jan 17 22:02:55 03oleo * r5221 10optware/trunk/ (make/transmission.mk sources/transmission/transmissiond.c): transmission: sync with CLI Jan 17 22:04:29 eno: doesn't work :( Jan 17 22:04:30 * NAiL wonders how long it'll take to compile apache 2.2.4 and php 5.2.0 on the TS101 Jan 17 22:04:48 > ipkg info kernel-module-tun Jan 17 22:04:48 Package: kernel-module-tun Jan 17 22:04:48 Version: 2.4.22-4 Jan 17 22:04:48 Replaces: kernel-modules-tun Jan 17 22:04:48 Status: install user installed Jan 17 22:04:49 Architecture: powerpc Jan 17 22:04:53 Installed-Time: 1169101993 Jan 17 22:05:30 ipkg list_installed |grep kernel-modul Jan 17 22:05:30 kernel-module-nfsd - 2.4.22-4 - Jan 17 22:05:30 kernel-module-tun - 2.4.22-4 - Jan 17 22:05:30 kernel-modules-tun - 2.4.22-3 - Jan 17 22:05:40 oleo: starting oleg and ddwrt rebuild now Jan 17 22:05:48 (including toolchain rebuild) Jan 17 22:08:09 eno: which clean target removes the 'host' directory? Jan 17 22:08:18 nm - it's a symlink Jan 17 22:10:03 rwhitby: right now there's no such target, individual package cleans host/builds/ Jan 17 22:10:20 eno: no probs Jan 17 22:10:21 very much like staging Jan 17 22:10:50 staging is a hell after a short period of time Jan 17 22:12:36 yeah, you can see ugly hack like 'rm -rf $(STAGING_INCLUDE_DIR)/apache2' in -stage Jan 17 22:13:30 osas: did you fix the asterisk staging? Jan 17 22:13:56 yay. apache2 compiled :) Jan 17 22:14:02 but OTOH, to solve it cleanly almost requires each package to have its own staging area and build all its dependencies Jan 17 22:15:09 eno: only this would work, or to do it like "mach" in a chroot Jan 17 22:15:55 in #oe they ignore me :( Jan 17 22:16:53 there're examples under builds/ipkg-opt/test Jan 17 22:20:02 gda_: ds101j-kernel-modules also a DS101G_SPECIFIC_PACKAGES, right? Jan 17 22:20:33 eno: no, ds101j specific Jan 17 22:20:58 ds101j is arm, ds101g is ppc Jan 17 22:21:18 i don't think we have a ds101j feed Jan 17 22:21:34 ds101j is nearly the same like ds101 Jan 17 22:21:52 ok, so maybe DS101_SPECIFIC_PACKAGES Jan 17 22:22:01 optware shouldn't make a difference there Jan 17 22:22:04 just half the ram and 1/8th the flash :-+ Jan 17 22:22:12 nite all Jan 17 22:22:37 NAiL: and some probs with glibc during bootstrapping? Jan 17 22:23:18 eno: after looking on the examples, I can't see what I made wrong Jan 17 22:23:32 ipkg problem? Jan 17 22:24:47 and Replaces not working? Jan 17 22:25:19 look what I have posted at 23:34 Jan 17 22:27:03 really depends on what you expect Replaces to do Jan 17 22:27:14 ? Jan 17 22:27:17 ideally it should remove kernel-modules-tun Jan 17 22:27:51 of course I expect this Jan 17 22:28:07 at least rpm is doing it Jan 17 22:36:42 rwhitby : tnx for rebuild Jan 17 22:49:53 03gda * r5222 10optware/trunk/make/ds101g-kernel-modules.mk: ds101g-kernel-modules: sub packages now named kernel-module-* instead of kernel-modules-* Jan 17 22:55:00 03gda * r5223 10optware/trunk/make/ds101g-kernel-modules-fuse.mk: ds101g-kernel-modules-fuse: renamed ipk Jan 17 22:55:08 n8 Jan 18 01:41:13 morning Jan 18 01:43:40 morning rwhitby Jan 18 01:44:20 hey VoodooZ **** ENDING LOGGING AT Thu Jan 18 02:59:57 2007