**** BEGIN LOGGING AT Tue Jun 01 02:59:57 2010
Jun 01 07:09:56 <acoul> good morning
Jun 01 08:49:59 <CIA-38> acoul * r21645 /trunk/target/linux/ar71xx/generic/config-default: ar71xx: fix rootfs mount issue (closes #7319)
Jun 01 09:09:00 <CIA-38> nico * r21646 /packages/net/vpzone/ (8 files in 3 dirs): [packages] add vpzone - A flexible multipurpose layer 3 VPN daemon
Jun 01 10:32:55 <Gottox> I got a problem about backfire/packages. I'm trying to compile libopenh323 which results in an error complaining about speex.
Jun 01 10:33:10 <Gottox> http://pastebin.com/r7sUJaV2
Jun 01 13:39:42 <aroscha> hi
Jun 01 13:39:51 <aroscha> question for backfire:
Jun 01 13:40:04 <aroscha> does anyone here compile it with glibc enabled?
Jun 01 13:40:15 <aroscha> I am running into build breaks
Jun 01 13:41:17 <aroscha> is there some magic combination glibc version X + gcc version Y (I am talking about the toolchain now)
Jun 01 14:01:50 <{Nico}> aroscha: it usually depends on the arch, what combination are you triying to build ?
Jun 01 14:01:58 <aroscha> x86
Jun 01 14:01:59 <aroscha> alix
Jun 01 14:02:06 <aroscha> hi Nico :)
Jun 01 14:02:26 <{Nico}> aroscha: you should also probably try eglibc, which seems to support cross-compilation better
Jun 01 14:02:29 <{Nico}> hi aroscha
Jun 01 14:02:36 <aroscha> ok , will try :)
Jun 01 14:02:38 <aroscha> thanks!
Jun 01 14:02:50 <{Nico}> where does it break, do you have a build.log ?
Jun 01 14:02:57 <aroscha> will re-create it
Jun 01 14:03:02 <aroscha> not anymore
Jun 01 14:05:01 <{Nico}> iirr, sticking with the defaults (gcc-4.1.2 + glibc-2.6.1) used to work :)
Jun 01 14:11:58 <aroscha> ok
Jun 01 14:12:10 <aroscha> will re-build it and post the error message
Jun 01 14:12:24 <aroscha> merci
Jun 01 14:28:24 <aroscha> {Nico}: this is the error message: http://pastebin.com/624qdk9v
Jun 01 14:28:55 <aroscha> have not found a build.log there
Jun 01 14:41:39 <CIA-38> nbd * r21647 /trunk/target/linux/orion/base-files/etc/uci-defaults/hardware: orion: fix lan and wan mac address for wrt530n v2 (patch by maddes from #7113)
Jun 01 14:42:51 <CIA-38> nbd * r21648 /branches/backfire/target/linux/orion/base-files/etc/uci-defaults/hardware: orion: add wrt350nv2 mac address fix from r21647
Jun 01 14:48:38 <builder> build #44 of uml is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/uml/builds/44
Jun 01 15:12:36 <aroscha> {Nico}: same issue with eglibc
Jun 01 15:12:57 <aroscha> http://pastebin.com/ZrCa9rYN
Jun 01 15:13:23 <aroscha> where should I be able to find a more detailed build.log or similar log file?
Jun 01 15:33:52 <CIA-38> nbd * r21649 /branches/backfire/package/mac80211/ (28 files in 2 dirs): backport the latest version of the mac80211 package to backfire. includes fixes for wpa key handling, throughput issues, etc.
Jun 01 16:28:48 <Gottox> I got an error when I try to compile openh323. Seems like something is wrong with speex/openh323
Jun 01 16:31:14 <Gottox> seems like -L is missing.
Jun 01 16:32:27 <Gottox> also I can't find libspeex.so in staging_dir/target-i386_uClibc-0.9.30.1/root-x86/usr/lib only libspeexdsp.so.1@ don't know if that's a mistake.
Jun 01 17:38:03 <CIA-38> nbd * r21650 /trunk/package/mac80211/patches/530-ath9k_queue_fill.patch: ath9k: fix queue stopping/starting logic, should slightly reduce RAM usage under load and make throughput more smooth
Jun 01 18:47:33 <stintel> whoa saved by git reflog :>
Jun 01 19:03:32 <aroscha> re
Jun 01 19:30:34 <CIA-38> nbd * r21651 /trunk/package/mac80211/patches/530-ath9k_queue_fill.patch: ath9k: one more queue stop/start fix
Jun 01 20:21:13 <_trine> can the makefile for ngircd be updated please http://dpaste.com/202011/
Jun 01 20:42:17 <aroscha> i don't get it
Jun 01 20:42:25 <aroscha> I always get build stops at busybox
Jun 01 20:42:50 <nbd> show me
Jun 01 20:43:18 <aroscha> take a look at pastebin.com/rrsuhaP7
Jun 01 20:43:28 <aroscha> I did a make --debug -j 9 V=99
Jun 01 20:43:40 <aroscha> and before that a make clean
Jun 01 20:44:04 <nbd> that log contains no usable info
Jun 01 20:44:09 <aroscha> but there is no other error message where
Jun 01 20:44:10 <aroscha> exactly!
Jun 01 20:44:20 <rtz2> aroscha: leave out the -j
Jun 01 20:44:29 <aroscha> ok
Jun 01 20:44:31 <aroscha> will try
Jun 01 20:44:39 <nbd> or use BUILD_LOG=1
Jun 01 20:44:44 <aroscha> rtz2: same thing
Jun 01 20:44:45 <aroscha> ahhh
Jun 01 20:44:47 <aroscha> ok, that might help
Jun 01 20:45:04 <nbd> BUILD_LOG=1 will create per-directory log files
Jun 01 20:45:05 <rtz2> aroscha: well, it should show you the actual error without -j
Jun 01 20:45:36 <nbd> aroscha: when using make -j, the actual error messages may be hundreds or thousands of lines above the exit
Jun 01 20:45:38 <aroscha> rtz2: nope
Jun 01 20:45:46 <aroscha> got it
Jun 01 20:45:54 <rtz2> aroscha: paste what output you get
Jun 01 20:45:55 <nbd> because make lets each individual build job finish before bailing out
Jun 01 20:45:57 <aroscha> crap
Jun 01 20:45:58 <aroscha> libssp
Jun 01 20:46:15 <aroscha> libssp.so: no such file or directory
Jun 01 20:46:37 <aroscha> BTW: that is something I noticed some time ago already in 8.09
Jun 01 20:47:22 <aroscha> if you select stack smashing protection in Advanced Configuration->toolchain options then it runs into this error
Jun 01 20:47:23 <aroscha> ok.
Jun 01 20:47:26 <aroscha> ticket, merci
Jun 01 20:48:01 <aroscha> https://dev.openwrt.org/ticket/3924
Jun 01 20:48:20 <aroscha> ok, slightly different
Jun 01 20:48:26 <aroscha> but ... same thing somewhat
Jun 01 20:50:07 <aroscha> ok, got it
Jun 01 20:52:51 <aroscha> so the bug in this case is that if you select ssp then it does not put the .so into the staging_dir//toolchain-i386_gcc-4.1.2_uClibc-0.9.30.1/lib/
Jun 01 20:53:44 <Weedy> speaking of compilation related bugs
Jun 01 20:54:00 <Weedy> if you use sstrip on libs dnsmasq breaks
Jun 01 21:43:55 <CIA-38> jow * r21652 /trunk/package/base-files/ (3 files in 3 dirs): [package] base-file: move alias setup to route hotplug, this fixes alias sections for non-static interfaces like ppp, dhcp, 6in4 etc.
Jun 01 21:46:56 <Weedy> jow_laptop: oi
Jun 01 21:47:56 <Weedy> is alias supposed to allow things like using a dsl modem in bridge mode but still having a static ip that allows you to reach the modems admin page?
Jun 01 21:48:04 <xMff> yes
Jun 01 21:48:10 <Weedy> :O
Jun 01 21:48:16 <Weedy> so very fucking want
Jun 01 21:48:33 <xMff> you probably still need a manual masq
Jun 01 21:49:07 <Weedy> if i could do that i could write a munin plugin to get my modem stats
Jun 01 21:49:17 <stintel> or add the route to your internal subnet into your modem
Jun 01 21:49:26 <xMff> stintel: ah right
Jun 01 21:49:40 <stintel> if the crappy device supports it ;-)
Jun 01 21:49:59 <stintel> but I guess they mostly do
Jun 01 21:50:06 <stintel> gn
Jun 01 21:50:09 <Weedy> i don't know enough about routing to understand that
Jun 01 21:50:46 <xMff> suppose your modems status page is on 10.0.0.1 netmask 255.255.255.0 in bridge mode
Jun 01 21:51:14 <xMff> you'd add an alias to wan with ipaddr 10.0.0.2, netmask 255.255.255.0 on your openwrt box
Jun 01 21:51:23 <Weedy> that i get
Jun 01 21:51:32 <xMff> suppose your owrt lan is 192.168.1.1 netmask 255.255.255.0
Jun 01 21:51:52 <xMff> then you'd setup a static route on the modem with target net 192.168.1.0 / 255.255.255.0 via gateway 10.0.0.2
Jun 01 21:52:08 <Weedy> i see
Jun 01 21:52:25 <xMff> otherwise the modem has no clue on which iface to answer requests made from 192.168.1.x since it does not have this subnet by itself
Jun 01 21:52:32 <Weedy> and if I can't do that?
Jun 01 21:52:39 <xMff> then you need a masquerade
Jun 01 21:52:48 <Weedy> hmm
Jun 01 21:53:23 <xMff> iptables -I POSTROUTING -o $wanif -d 10.0.0.1 -j MASQUERADE
Jun 01 21:53:36 <xMff> where $wanif is the real ethernet interface, not ppp0 or so
Jun 01 21:53:53 <xMff> oh  iptables -t nat ...
Jun 01 21:55:20 <aroscha> what is the right init when I start openwrt from within a kvm / qemu environment (ok... I should say the kernel complains that it does not find init. And it does not accept -append init=/bin/busybox on the command line)
Jun 01 21:55:35 <xMff> init=/etc/preinit
Jun 01 21:55:52 <xMff> there was an issue with kvm
Jun 01 21:56:00 <xMff> and squashfs images
Jun 01 21:56:23 <xMff> the block2mtd layer is initialized after the virtio stuff
Jun 01 21:56:31 <xMff> so you should use ext2
Jun 01 21:57:21 <aroscha> xMff: merci!
Jun 01 21:57:23 <aroscha> will try
Jun 01 21:58:50 <CIA-38> jow * r21653 /trunk/package/firewall/ (Makefile files/lib/core_interface.sh files/lib/core_rule.sh): [package] firewall: Initial alias interface support. This allows to define zones covering alias interfaces and associated entries like rules and forwardings.
Jun 01 21:59:34 <aroscha> xMff: nope... somehow it does not find the preinit. Must be something with the rootfs
Jun 01 21:59:35 <aroscha> strange
Jun 01 21:59:48 <aroscha> it can mount it
Jun 01 22:00:15 <aroscha> what is your kvm line ?
Jun 01 22:00:33 <xMff> I don't use kvm, xen here
Jun 01 22:00:43 <xMff> but I helped stintel to implement a kvm guest subtarget
Jun 01 22:00:55 <xMff> check https://lists.openwrt.org/pipermail/openwrt-devel/2010-March/006526.html and following
Jun 01 22:00:59 <xMff> maybe they contain some clue
Jun 01 22:01:50 <aroscha> ok
Jun 01 22:02:11 <xMff> you need kernel_menuconfig
Jun 01 22:02:19 <xMff> and enable virtio stuff iirc
Jun 01 22:02:30 <xMff> CONFIG_VIRTIO_BLK=y  is the most important
Jun 01 22:02:54 <xMff> I don't know kvm well
Jun 01 22:03:05 <xMff> but maybe you can also change the storage emulation to something else
Jun 01 22:03:14 <xMff> non-virtio but faked ide or so
Jun 01 22:17:18 <aroscha> will try
Jun 01 22:23:09 <aroscha> xMff: But you don't mean  having kvm inside backfire?
Jun 01 22:23:14 <xMff> no
Jun 01 22:23:16 <aroscha> ok
Jun 01 22:23:19 <aroscha> ouff
Jun 01 22:30:05 <soma> i have openwrt in kvm running and i simply activated kvm guest support in the kernel and start it quite regular (-hda <combined-jffs.img>)
Jun 01 22:30:44 <xMff> soma: it works with squash/jffs2 images for you?
Jun 01 22:30:53 <aroscha> soma: nice
Jun 01 22:31:10 <soma> jffs for sure, squashfs too i think
Jun 01 22:32:18 <aroscha> soma: interesint...
Jun 01 22:32:22 <soma> but im not using those virtio devices afaik
Jun 01 22:32:28 <xMff> ah ok
Jun 01 22:32:32 <xMff> this explains it then
Jun 01 22:32:41 <aroscha> all I get is "Decompressing Linux... Parsing ELF... done"
Jun 01 22:33:04 <aroscha> soma: so your command line is kvm -hda mysquashfs.img
Jun 01 22:33:05 <aroscha> ?
Jun 01 22:33:15 <xMff> aroscha: if the console is virtio too, that one won't work either
Jun 01 22:33:28 <aroscha> ah, yes. ok
Jun 01 22:33:38 <aroscha> ok, I will try with X11 or SDL then
Jun 01 22:34:19 <soma> aroscha: http://pastebin.com/Z6MzR4mC
Jun 01 22:34:56 <aroscha> soma: would you be able to share the .config file for this?
Jun 01 22:35:01 <aroscha> and the kernel config ?
Jun 01 22:35:43 <soma> sure. and if you like here are the images: http://www.wgaugsburg.de/openwrt/x86-grsec/21426/
Jun 01 22:36:24 <soma> aroscha:  where can i send them?
Jun 01 22:36:33 <aroscha> aaron@lo-res.org
Jun 01 22:39:24 <aroscha> i guess it must be something in my config. Works with the standard download from openwrt.org (x86)
Jun 01 22:42:37 <aroscha> soma: danke :)
Jun 01 22:47:20 <soma> hope it helps you to find out whats wrong with your config then
Jun 01 22:48:14 <aroscha> yup
Jun 01 22:48:24 <aroscha> though there might be some differences since I am on backfire
Jun 01 22:48:29 <aroscha> anyway, thanks!
Jun 01 22:48:43 * aroscha needs to continue tomorrow. Got up at 7:00 a.m. ;-)
Jun 01 22:48:48 <aroscha> bye!
Jun 01 23:11:44 <CIA-38> nbd * r21654 /trunk/package/mac80211/ (Makefile patches/016-remove_pid_algo.patch): mac80211: remove the pid rate control algorithm to save a few kbytes
Jun 01 23:19:29 <builder> build #44 of x86 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/x86/builds/44
Jun 02 00:07:04 <cshore> ping nbd
Jun 02 00:08:21 <cshore> Anyone know if nbd has both trunk and 0.9 luci in git?  I he has a git repo of luci, but I'm not sure what branches
Jun 02 00:08:31 <cshore> *know
Jun 02 00:09:14 <nbd> yes, both trunk and 0.9 are in the same repo
Jun 02 00:09:26 <cshore> ah ok, is 0.9 a branch?
Jun 02 00:09:45 <nbd> yes
Jun 02 00:10:00 <cshore> nbd: thanks (and thanks for hosting the repos)
Jun 02 00:10:20 <nbd> no problem
Jun 02 00:10:54 <cshore> nbd: btw have you tried stgit at all...I'm finding it useful to keep track of patches on top of OpenWRT
Jun 02 00:11:17 <nbd> nope, never tried it
Jun 02 00:11:32 <nbd> i don't usually keep patches on top of openwrt
Jun 02 00:11:43 <nbd> i commit them
Jun 02 00:12:21 <cshore> nbd: I'm keeping my device-specific configs there, and for the company I work for anything that is specific to them
Jun 02 00:12:43 <nbd> and those have to be patched on top of openwrt?
Jun 02 00:12:47 <nbd> can't you keep them in an environment?
Jun 02 00:12:57 <nbd> (using scripts/env)
Jun 02 00:13:02 <cshore> nbd: what's that?
Jun 02 00:13:13 <nbd> ./scripts/env manages multiple sets of .config and files/
Jun 02 00:13:18 <nbd> and lets you switch between thme
Jun 02 00:13:28 <nbd> (using git internally)
Jun 02 00:13:32 <cshore> nbd: I see, I wasn't aware of that script
Jun 02 00:13:44 <nbd> i use it all the time
Jun 02 00:13:46 <cshore> nbd: any docs?
Jun 02 00:13:53 <nbd> it's documented in docs/
Jun 02 00:13:58 <nbd> however the command line help should be enough
Jun 02 00:13:59 <nbd> it's very simple
Jun 02 00:14:14 <cshore> nbd: always something new to learn
Jun 02 00:15:01 <cshore> nbd: how to I sync the envs with a remote git repository?
Jun 02 00:16:13 <nbd> you can go to env/ and use regular git commands
Jun 02 00:16:17 <nbd> each environment is represented as a git branch
Jun 02 00:16:24 <cshore> nbd: ok, great, thanks
Jun 02 00:16:33 <nbd> and each environment save creates a new commit
Jun 02 00:16:38 <nbd> so you have a history of your changes as well
Jun 02 00:16:45 <cshore> nbd: sweet
Jun 02 00:16:54 <nbd> maybe at some point i'll add a menuconfig hook that auto-commits in the environment
Jun 02 00:17:17 <nbd> right now it only saves when you switch or when you tell it to save
Jun 02 00:17:43 <cshore> nbd: ok, that's not that bad...and sometimes what you want anyway
Jun 02 00:18:05 <cshore> nbd: besides doing it without env you have to save manually anywya
Jun 02 00:18:20 <nbd> i could also add a ./scripts/env upload command where you can configure the remote url and don't have to add a remote entry for each branch
Jun 02 00:18:42 <cshore> nbd: that'd be nice
Jun 02 00:19:05 <cshore> nbd: in your oh so copious spare time ;-)
Jun 02 00:19:32 <nbd> wouldn't take long
Jun 02 00:19:49 <nbd> scripts/env is just a trivial shell script
Jun 02 00:19:57 <nbd> easy to read and edit
Jun 02 00:20:29 <cshore> nbd: ok, that's good to know
Jun 02 00:21:00 <cshore> nbd: what was that env2 script that was posted to the dev list?
Jun 02 00:21:58 <nbd> that does a little more and also interacts with the repository itself
Jun 02 00:22:08 <nbd> i've never tried that since scripts/env does all that i need
Jun 02 00:22:22 <CIA-38> jow * r21655 /trunk/package/base-files/files/ (etc/hotplug.d/iface/10-routes lib/network/config.sh):
Jun 02 00:22:22 <CIA-38> [package] base-files:
Jun 02 00:22:22 <CIA-38> - add aliases to device by default (not ifname)
Jun 02 00:22:22 <CIA-38> - introduce option "layer" to select the target ifname to attach the alias to:
Jun 02 00:22:22 <CIA-38> - 3 use tun device (tun over bridge over device) fallback to bridge or device
Jun 02 00:56:34 <builder> build #46 of xburst is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/xburst/builds/46
Jun 02 00:59:38 <CIA-38> jow * r21656 /trunk/package/firewall/files/lib/core_interface.sh:
Jun 02 00:59:38 <CIA-38> [package] firewall:
Jun 02 00:59:38 <CIA-38> - support alias ifnames different from parent ifname
Jun 02 00:59:38 <CIA-38> - properly handle multiple subnets per alias (v4+v6)
Jun 02 01:11:09 <swalker> thepeople: can you accept openwrt.org
Jun 02 01:11:19 <swalker> er
Jun 02 01:11:27 <swalker> 's hostkey for the buildbots?
Jun 02 01:12:18 <CIA-38> nbd * r21657 /trunk/package/mac80211/patches/300-mac80211_debugfs_debloat.patch: mac80211: reduce the size of the debugfs code
Jun 02 01:13:42 <swalker> geoff-: were you going to take care of ps3's GELIC_WIRELESS_OLD_PSK_INTERFACE missing kernel symbol?
Jun 02 01:15:54 <cshore> ping nbd:
Jun 02 01:16:06 <cshore> or anyone who knows how to use ./scripts/env
Jun 02 01:16:23 <cshore> files/ is now empty: is this normal?
Jun 02 01:16:47 <cshore> and env has dirs for the files that ere in files, but the symlink points to an empty dir
Jun 02 01:17:03 <cshore> *were
Jun 02 01:17:17 <xMff> this happens if you did not start with env from the beginning
Jun 02 01:17:42 <xMff> it basically moves all customizations away and replaces them with symlinks to per-env files and folders
Jun 02 01:17:58 <cshore> xMff: how do I import?
Jun 02 01:18:04 <xMff> just redo
Jun 02 01:18:07 <nbd> just throw something into the empty files/
Jun 02 01:18:11 <xMff> fill files again etc.
Jun 02 01:18:16 <cshore> ok
Jun 02 01:19:19 <xMff> once you're done, do ./scripts/env save
Jun 02 01:19:54 <xMff> you usually do not need to do that manually since it will happen anyway when you create a new env / switch to another one
Jun 02 01:21:10 <cshore> xMff: let's say I have a patch series with the changes that I want to end up in env.  How to I initialize env and get the changes into is so I can save them?
Jun 02 01:21:36 <xMff> I don't think you can use it for managing patch series
Jun 02 01:21:43 <xMff> at least I never did that
Jun 02 01:22:05 <cshore> xMff: I just want to get the results of the series into env (i.e. once they're all applied)
Jun 02 01:22:05 <xMff> it is for separating multiple files/, kernel and global .configs
Jun 02 01:22:41 <cshore> xMff: Just trying to not have to redo everything
Jun 02 01:22:41 <rtz2> can't you somehow ?mport it into the git repository
Jun 02 01:22:47 <xMff> if your patch modifies non-files/ or .config it will apply to all arches
Jun 02 01:22:47 <rtz2> i
Jun 02 01:22:59 <xMff> regardless of env
Jun 02 01:23:06 <nbd> cshore: you could create a new directory in env/
Jun 02 01:23:16 <nbd> cshore: and symlink patches/ to that directory
Jun 02 01:23:23 <cshore> xMff: I only have .config and files
Jun 02 01:23:28 <nbd> scripts/env will pick up anything you throw into env/
Jun 02 01:23:36 <nbd> files/ is for stuff that gets added to the rootfs only
Jun 02 01:23:39 <xMff> then aptch them in
Jun 02 01:23:44 <xMff> and do env save
Jun 02 01:23:48 <cshore> ok
Jun 02 01:24:14 <cshore> xMff: Can I do commit messages with env?
Jun 02 01:24:22 <xMff> no
Jun 02 01:35:26 <cshore> hmmm....I'm not sure what's the better option then
Jun 02 02:52:49 <builder> build #55 of mpc52xx is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/55
**** ENDING LOGGING AT Wed Jun 02 02:59:57 2010