**** BEGIN LOGGING AT Fri Feb 19 02:59:57 2010 Feb 19 03:06:31 jow * r19727 /packages/utils/mc/Makefile: [packages] mc: fixup broken symlinks to host automake (#6699) Feb 19 04:10:14 lars * r19728 /packages/libs/aalib/Makefile: Feb 19 04:10:14 [packages] aalib: Set ncurses dir prefix. Otherwise configure is going find the Feb 19 04:10:14 hosts ncurses lib. Feb 19 04:11:49 lars * r19729 /packages/net/bahamut/ (Makefile patches/010-dont-use-host-zlib.patch): [packages] bahamut: Don't use hosts zlib Feb 19 04:19:47 lars * r19730 /packages/Xorg/lib/gtk1/Makefile: [packages] gta1: Properly fix paths in gtk-config Feb 19 04:21:04 lars * r19731 /packages/net/pure-ftpd/patches/ (. 010-dont-add-host-paths.patch): [packages] pure-ftpd: Don't add host lib and include paths. Feb 19 04:27:47 lars * r19732 /packages/Xorg/lib/libast/patches/010-dont-add-host-paths.patch: [packages] libast: Dont' add host lib and include dirs. Feb 19 04:37:18 lars * r19733 /packages/utils/anyremote/patches/ (. 010-dont-add-host-includes.patch): [packages] anyremote: Don't add host include dirs. Feb 19 04:38:13 lars * r19734 /packages/Xorg/app/eterm/patches/ (. 010-dont-add-host-paths.patch): [packages] eterm: Don't add host lib and include paths. Feb 19 04:46:28 swalker * r19735 /packages/utils/stress/Makefile: [packages] stress: update to 1.0.3 Feb 19 04:48:13 swalker * r19736 /packages/utils/pciutils/Makefile: [packages] pciutils: update to 3.1.7 Feb 19 04:55:43 lars * r19737 /packages/utils/rrdtool/patches/ (. 010-dont-add-host-include-paths.patch): [packages] rrdtool: Don't add host include paths. Feb 19 04:57:01 lars * r19738 /packages/net/ipcad/patches/010-no-host-includes.patch: [packages] ipcad: Don't add host include paths. Feb 19 04:57:37 lars * r19739 /packages/net/ndyndns/patches/010-dont-add-host-includes.patch: [packages] ndyndns: Don't add host include paths. Feb 19 05:44:25 rhk`: I take it you didn't see https://dev.openwrt.org/ticket/6370 Feb 19 08:35:11 swalker, you're right, didn't see it Feb 19 08:40:23 swalker, so why isn't it supported now if a patch has been provided 2 months ago? Feb 19 09:50:22 swalker, I've seen https://dev.openwrt.org/ticket/6700 also Feb 19 09:58:54 kaloz * r19740 /trunk/toolchain/uClibc/patches-0.9.30.1/500-avr32_add_varargs_handling_of_prctl_syscall.patch: [uClibc/avr32]: fix varargs handling of the prctl syscall Feb 19 10:13:18 hello Feb 19 10:13:24 i've read on the forum Feb 19 10:13:32 here https://forum.openwrt.org/viewtopic.php?id=23426 Feb 19 10:13:55 that current trunk should support FREESCALE mpc83xx Feb 19 10:14:04 but i can't find it among the available targets Feb 19 10:18:05 do you find it in your trunk? Feb 19 10:24:01 juhosg * r19741 /trunk/target/linux/adm5120/ (4 files in 2 dirs): adm5120: sync kernel configs Feb 19 10:24:05 juhosg * r19742 /trunk/target/linux/generic-2.6/config-2.6.30: kernel: add a bunch of missing symbols to the 2.6.30 config Feb 19 10:25:36 Lalloso: isn't that just the mpc83xx target? Feb 19 10:26:01 Lalloso: I'm pretty sure I've seen it Feb 19 10:26:39 Lalloso: maybe you need to enable broken targets (i.e. maybe it's not ready yet) Feb 19 10:26:59 ah okay let me try Feb 19 10:27:33 how do i enable broken targets? Feb 19 10:28:00 Developer options (advanced options?) Feb 19 10:28:20 thx a lot Feb 19 10:28:23 got it Feb 19 10:28:25 :-D Feb 19 10:28:58 it appears also infineon amazon Feb 19 10:29:01 nice! Feb 19 10:29:46 cshore are you still hiring at santa clara? Feb 19 10:30:01 no idea: I'm not the boss Feb 19 10:30:26 have you defined your target architecture? Feb 19 10:30:32 but I don't think so Feb 19 10:30:40 brcm63xx Feb 19 10:31:00 I've written an email to you Feb 19 10:31:04 about dsp Feb 19 10:31:06 ok Feb 19 10:44:18 rhk`: lack of manpower and what manpower is available is busy working on other areas? not exactly the first case of it happening though Feb 19 10:45:44 swalker: is there something like a buildbot? Feb 19 10:45:57 compiling for all the different architectures? Feb 19 10:46:19 that would allow to spot packaging bugs more easily Feb 19 10:46:46 there used to be a buildbot.openwrt.org years ago iirc but http://tksite.gotdns.org:8010/ is available now Feb 19 11:06:37 Collected errors: * Packages were found, but none compatible with the architectures configured Feb 19 11:06:44 is there a way to override this error? Feb 19 11:08:50 renaming the packages does not seem to suffice Feb 19 11:28:27 Lalloso: you need to override the arch in opkg.conf Feb 19 11:28:35 can somebody verify https://dev.openwrt.org/ticket/6701 ? Feb 19 11:31:43 swalker: while you're at the download methods (#6700), can you switch "svn checkout" to "svn export" ? Then the find-rm-.svn is not needed which should speed the whole thing up Feb 19 11:35:42 thx xMff Feb 19 11:36:50 Lalloso: you can specify multiple arches there and assign a priority to them, for example Feb 19 11:36:50 arch x86 1 Feb 19 11:36:50 arch i386 10 Feb 19 11:37:12 add whatever the package uses you try to install Feb 19 11:52:17 juhosg * r19743 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-wndr3700.c: Feb 19 11:52:17 ar71xx: fix polarity of GPIO buttons on the WNDR3700 board Feb 19 11:52:17 Signed-off-by: Scott Dudley Feb 19 11:52:21 juhosg * r19744 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-wndr3700.c: Feb 19 11:52:21 ar71xx: add missing GPIO LED for the WNDR3700 board Feb 19 11:52:21 Signed-off-by: Scott Dudley Feb 19 11:52:24 juhosg * r19745 /trunk/package/mac80211/ (Makefile patches/410-ath9k-wndr3700-led-pin-fix.patch): Feb 19 11:52:24 mac80211: fix GPIO pin assignment for the LEDs on the WNDR3700 board Feb 19 11:52:24 Signed-off-by: Scott Dudley Feb 19 11:52:27 juhosg * r19746 /trunk/target/linux/ar71xx/base-files/etc/uci-defaults/ (. wndr3700): Feb 19 11:52:27 ar71xx: add uci configuration for the WAN LED on the WNDR3700 Feb 19 11:52:27 Based on a patch by Scott Dudley Feb 19 11:52:29 juhosg * r19747 /trunk/target/linux/ar71xx/config-2.6.32: ar71xx: sync kernel config Feb 19 11:52:32 juhosg * r19748 /trunk/tools/firmware-utils/src/mktplinkfw.c: firmware-utils/mktplinkfw: add board definitions for TL-WR941ND v4 Feb 19 11:52:35 juhosg * r19749 /trunk/target/linux/ar71xx/ (image/Makefile profiles/tp-link.mk): ar71xx: create firmware images for the TL-WR941ND v4 board Feb 19 11:52:38 juhosg * r19750 /trunk/target/linux/ar71xx/ (7 files in 2 dirs): ar71xx: move AP91 specific ethernet initialization into a separate file Feb 19 11:52:41 juhosg * r19751 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-tl-wr741nd.c: ar71xx: fix wireless mac address on the TL-WR741ND Feb 19 11:52:43 juhosg * r19752 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/ (Kconfig mach-dir-600-a1.c): ar71xx: fix mac addresses on the DIR-600 rev. A1 board Feb 19 11:56:15 florian * r19753 /trunk/target/linux/brcm63xx/patches-2.6.30/070_bcm63xx_enet_vlan_incoming_fixed.patch: [brcm63xx] fix the receiving of packets with a 1500 mtu and vlan enabled (#6698) Feb 19 12:00:07 <[florian]> rtz2: ping? Feb 19 12:00:51 [florian]: pong Feb 19 12:01:10 <[florian]> rtz2: did you figure out why only 2/4 ports of the ic+ switch are enabled? Feb 19 12:01:30 [florian]: yes Feb 19 12:01:51 [florian]: I have a realtek switch, but it doesn Feb 19 12:01:53 't matter Feb 19 12:02:00 the phys aren'T initilized Feb 19 12:02:14 [florian]: you remeber the libphy stuff I did send you? Feb 19 12:02:36 <[florian]> rtz2: yes I do Feb 19 12:02:55 [florian]: it fixed the problem Feb 19 12:03:13 [florian]: but it is more or less a complete rewrite of the current libphy Feb 19 12:04:11 [florian]: I could get it in a state ready for merging, but nbd voiced some doubts, if it would be accepted into mainline Feb 19 12:05:21 [florian]: I would need somebody to take a look at it and tell me, if and how I should go on with it Feb 19 12:06:21 [florian]: problem is, the current libphy doesn't support multiple phys connected to one mac at all and it's really hard to extend it to do so Feb 19 12:07:10 <[florian]> rtz2: we could probably sequentially configure each and every single phy in the r6040 driver, could not we? Feb 19 12:08:25 [florian]: yes, but imho it wouldn't be a very nice solution Feb 19 12:09:23 [florian]: it's basically a problem nearly all drivers with switches connected to them face Feb 19 12:09:50 <[florian]> rtz2: no a problem with all drivers and switches, only a couple of dump switches Feb 19 12:10:12 <[florian]> rtz2: all other switches have a single interface using a dedicated phy address or a mdio/mdc lines Feb 19 12:11:57 [florian]: the other switches have also multiple phys Feb 19 12:12:03 <[florian]> which ones? Feb 19 12:12:19 <[florian]> they do have multiple phys, but only a single one is connected to the mac Feb 19 12:12:48 <[florian]> if you need to query the link status and characteristics, you do this using the "cpu" port of the switch Feb 19 12:14:15 [florian]: I didn't look at many datasheets but most realtek switches and an adm6996 an a TI switch I check work that way Feb 19 12:15:17 [florian]: the cpu port only shows the link status to the cpu, nothing else Feb 19 12:17:11 <[florian]> rtz2: ok, I might not be understanding something then Feb 19 12:17:16 [florian]: the network drivers usually attach only to one phy, because libphy doesn't support multiple phys Feb 19 12:17:52 <[florian]> also for historical reasons, most machines are using a single phy mac paradigm Feb 19 12:18:24 [florian]: a 5 port switch has 5 phys, 5 macs and a switching logic Feb 19 12:18:56 [florian]: all 5 phys are reachable over the MII bus Feb 19 12:19:05 or MDIO or how it's called Feb 19 12:19:56 [florian]: those switches used in routers have a mode, where one of the macs is connected to the mac in the cpu instead of it's phy Feb 19 12:20:43 [florian]: and to the unused phy, another mac in the cpu can be connected Feb 19 12:21:43 [florian]: the realtek switch actually reports 6 phys in that case (4 external phys, 1 mac connected to the cpu, one phy connected to the cpu) Feb 19 12:22:27 [florian]: i'm not sure, if this is the standard case, but i am sure that every switch has at least as many phys on the MDIO bus as ports Feb 19 12:24:08 [florian]: play around with the libphy code, it does probe for multiple phys, even if it can't use them Feb 19 12:24:32 [florian]: or use my code, it has some nice dump functions ;) Feb 19 12:33:38 kerneis * r19754 /packages/net/babel/ (4 files in 2 dirs): (log message trimmed) Feb 19 12:33:38 [package] babeld: update to 0.98, fix scripts Feb 19 12:33:39 - Full-featured UCI configuration file (no need for babeld.conf). Feb 19 12:33:39 - Do not check the pid file (useless) and ahcpd script (harmful). Feb 19 12:33:39 - Make /etc/config/babeld a config file. Feb 19 12:33:39 - Launch at most one instance of babeld. Feb 19 12:33:39 - Add status and reload options to init script. Feb 19 14:15:32 ping thepeople jow_laptop Feb 19 14:16:08 ping Kaloz Feb 19 14:16:08 pong Feb 19 14:16:12 jow_laptop hi Feb 19 14:16:22 jow_laptop could u please apply my nfs v4 patch Feb 19 14:16:52 jow_laptop another question is how do u tread nis with uclibc? its not supported Feb 19 14:17:01 will do it this evening Feb 19 14:17:10 nis? Feb 19 14:17:22 the share passwords over the network Feb 19 14:17:26 portmap is deprecated Feb 19 14:17:49 libtirpc Feb 19 14:17:59 #and another lib no replaces it transparently Feb 19 14:18:06 no=now Feb 19 14:18:38 but nis support is missing in uclibc Feb 19 14:19:27 well if you intend to deploy systems that need nis, you have to use another libc implementation Feb 19 14:19:27 libtirpc and rpcbind Feb 19 14:19:41 openwrt supports uclibc, eglibc, glibc Feb 19 14:19:51 hmm i wanted to update nfs-utils to the latest version but this needs them Feb 19 14:20:40 portmap is deprecated for nfs-utils >= 1.2 Feb 19 14:21:13 deprecated does not mean dropped Feb 19 14:21:14 on some mailing list there is a uclibc and lib discussing meaning this Feb 19 14:21:19 1. have libtirpc to use openssl for the crypto funcs Feb 19 14:21:24 2. have libtirpc to use gnutls for the crypto funcs Feb 19 14:21:36 3. implement key_encryptsession_pk, getnetname, cbc_crypt, ecb_crypt, Feb 19 14:21:36 getpublickey and key_gendes in uclibc. Feb 19 14:21:54 but all of this options suck to have openssl as a "dep" for nfs-utils Feb 19 14:21:57 a hard dependency on openssl for nfs would suck Feb 19 14:22:21 it would be hard to use it and a lot of things would have to be changed Feb 19 14:22:36 for uclibc at least Feb 19 14:22:56 maybe i can somehow wipe this unneed nis/crypto stuff out Feb 19 14:23:02 unneeded Feb 19 14:24:07 on gentoo therefor it seems portmap is still the standard because of this changes Feb 19 14:26:31 ping nbd - do u know where i can find an good explanation how to commit into svn from the git tree? Feb 19 14:26:52 well just they mark it as deprecated does not mean that you have to abandon it immediately, especially if the proposed alternative does not work at all on our platform Feb 19 14:27:04 yeah i know Feb 19 14:27:41 i will update for now to 1.1.6 and write in the commit that if someone has found a good soltuion for the 1.2.* update they should contact me Feb 19 14:27:54 is there somewhere a good place to say things like this? Feb 19 14:27:57 the ml? Feb 19 14:28:39 tripolar * r19755 /packages/net/nfs-kernel-server/Makefile: update nfs-kernel-server to 1.1.6 Feb 19 14:31:46 thepeople: would it be possible to add "htop" to my maintained packages? Feb 19 14:31:55 tripolar_: yes the mailing list would be a good place to ask Feb 19 14:32:05 tripolar_: I will get it added Feb 19 14:32:09 thanks Feb 19 14:32:52 thepeople i nearly ported one lib i will try to get further and when i got a little more i will post my findings on the mailing list Feb 19 14:42:16 tripolar_: i have a script called gsc - http://nbd.name/gsc - you can use that to clone a git tree and set up the tree config for committing back to svn with git svn dcommit Feb 19 14:42:36 you use it with gsc Feb 19 14:42:52 and only use git-svn rebase in the tree to update, do not git-pull anymore Feb 19 14:52:39 nbd okay thanks Feb 19 14:52:41 ill try it Feb 19 14:52:45 a little bit later Feb 19 15:03:13 ping xMff: I traced the iptables issue to libiptc.c, line 2584, the setsockopt fails Feb 19 15:04:00 Thats a kernel system call if I understand correctly, so this doesn't really help much Feb 19 15:04:18 netprince: so it is missing kernel support after all :-( Feb 19 15:04:58 jow_laptop: yes, I wanted to ask you if you think the problem could be that iptables expects the kernel interface to be in xtables format, not ipt format? Feb 19 15:05:16 unlikely, more likely some patch is missing Feb 19 15:07:16 There does seem to be some discrepancies between the 2.6 version of .checkentry and the 2.4 version, I'm going to explore that further Feb 19 15:10:11 ok Feb 19 15:38:38 netprince: can you paste the backtrace somewhere? Feb 19 15:41:22 jow_laptop: I didn't save it. :( It was not very interesting. I have a few notes though, will paste them Feb 19 15:41:58 hauke * r19756 /trunk/package/mac80211/ (4 files in 2 dirs): Feb 19 15:41:58 [mac80211] remove pcmcia and sdio Feb 19 15:41:58 This fixes #6701 Feb 19 15:41:58 Remove the pcmcia and sdio interface from b43 driver to not depend on these modules. Feb 19 15:42:20 jow_laptop: if you want, I could re-run the trace, but I've already re-flashed the router to test some things, so it would take some time Feb 19 15:42:58 netprince: do you know whether it is possible to reproduce the issue with x86? Feb 19 15:43:17 no, x86 works fine, I have tested on alix boards Feb 19 15:43:27 hm Feb 19 15:43:44 its only the 2.4 kernel, the multiport backport patch is incorrect somehow Feb 19 15:44:13 jow_laptop: http://openwrt.pastebin.com/m741c2f6 Feb 19 15:46:03 do you know whether it uses the _v1 variant in the iptables extension? Feb 19 15:47:15 yes, the v1 variant is the only implementation, the patch removes the v0 variant. (probably because it doesn't support ranges) Feb 19 15:48:56 I've gota go for a couple hours, bbl Feb 19 15:55:55 netprince: I believe I found it Feb 19 15:56:29 no, nvm Feb 19 16:51:45 netprince: I believe the 2.4 kernel does not support the xtables match revision facility, therfore the userspace part of iptables always chooses the v0 match which is incompatible to the v1 implementaion on the kernel side Feb 19 16:52:11 netprince: I think the easiest solution is to remove the registration of the v0 multiport matches entirely Feb 19 16:56:34 ... and the associated code Feb 19 16:56:51 since v0 is nether supported in v4 and v6 kernels, there's no need to carry that around Feb 19 17:22:39 nbd * r19757 /trunk/package/mac80211/patches/540-ath9k_beacon_timer_fix.patch: ath9k: fix beacon timer restarts after card resets - improves stability in AP mode Feb 19 17:33:48 nbd * r19758 /trunk/package/swconfig/src/cli.c: Feb 19 17:33:49 swconfig: Remove unused code Feb 19 17:33:49 Signed-off-by: Martin Mares Feb 19 18:25:25 jow * r19759 /trunk/target/linux/x86/ (config-default xen_domu/config-default): [x86] switch kernel compression from lzma to gzip and reorder symbols - fixes boot problems in VirtualBox Feb 19 18:52:29 pinbg johnyb Feb 19 18:52:31 pinbg jow_laptop Feb 19 18:52:43 pong tripolar_ Feb 19 18:53:00 can u now commit the nfsd patch Feb 19 18:53:17 I am still at work Feb 19 18:53:52 okay Feb 19 18:54:00 where are you from? Feb 19 18:54:14 germany Feb 19 18:54:32 jow_laptop okay - people work long there as it seems - nearly 8pm Feb 19 18:54:41 :P Feb 19 18:54:53 well, I started 2pm Feb 19 18:54:54 do you have a work ralated to openwrt? Feb 19 18:54:57 nope Feb 19 18:54:59 ohh thats nice Feb 19 18:55:14 getting paid to do openwrt stuff would be nice Feb 19 18:56:52 indeed. ok - I'm on the way home. bbl Feb 19 18:57:00 bye Feb 19 18:57:03 i have to go now Feb 19 18:57:10 having party tonight :) Feb 19 19:11:53 nbd it seems the git package co doesnt have a complete log some entries are missing .. is this know or a bug? Feb 19 19:12:44 what entries are missing? Feb 19 19:12:54 nbd git log messages Feb 19 19:13:04 which ones?? Feb 19 19:13:26 hauke * r19760 /trunk/package/mac80211/Makefile: Feb 19 19:13:26 [mac80211] rt2800lib depends on rt2x00-usb Feb 19 19:13:26 The rt2800lib module depends on rt2x00-usb if rt2x00-usb was build. Feb 19 19:13:47 nbd now it seems to work - tried it some hours ago and there where some log messages missing Feb 19 19:14:07 the repo only syncs every 10 minutes Feb 19 19:14:10 of commits - they changes where made but there where no log messages Feb 19 19:14:55 yeah my a temporary failure .. i update some minutes ago ... but i have seen this error befor Feb 19 19:15:01 some weeks ago Feb 19 19:15:07 my=may Feb 19 19:15:11 maybe Feb 19 19:15:32 maybe a pebkac Feb 19 19:15:34 :D Feb 19 19:15:46 <- Feb 19 19:17:00 nbd i sometimes use git log on a path to see only the commits that affected the path Feb 19 19:17:09 or a file Feb 19 19:21:44 nbd do you have a work related to openwrt? i mean do you get paid for some work on it? Feb 19 19:21:56 i get paid mostly for my wifi driver work Feb 19 19:22:01 so yes Feb 19 19:22:14 nbd okay and how did u get this job? Feb 19 19:22:34 by being good at what i do and submitting my stuff on public lists Feb 19 19:22:39 ;) Feb 19 19:23:03 yeah - it seems like this - you know what you are doing ;) Feb 19 19:23:22 free marketing ;-) Feb 19 19:23:56 exactly Feb 19 19:24:23 nbd so you didnt ask for the job - you where asked to do it Feb 19 19:24:34 right Feb 19 19:24:40 nice Feb 19 19:24:47 tripolar_: htop is added Feb 19 19:25:01 thepeople: thanks Feb 19 19:25:19 i have to go - have a nice day/night - bye Feb 19 20:31:25 xMff: interesting idea, I will try that Feb 19 20:32:51 netprince: the only reason that checkentry() could fail is a struct size mismatch, the _v1 structs from userspace and kernel have identical declarations so I think iptables is using the wrong version to the kernel Feb 19 20:33:11 s/using/sending/ Feb 19 20:34:08 ... because the v0 struct is shorter Feb 19 20:34:40 the strange thing is that iptables 1.3.8 worked fine with the patch in its current state Feb 19 20:35:10 1.3.8 didn't had the target/match revision facility iirc Feb 19 20:35:28 interesting Feb 19 20:36:26 maybe I'm totally wrong, I just read the code and patches... Feb 19 20:37:20 its worth a closer look Feb 19 20:38:40 I dont have time right now, will look closer later Feb 19 20:39:09 I'll give it a try on a wl500gp shortly Feb 19 20:39:20 cool Feb 19 21:20:21 netprince: indeed - that was it Feb 19 21:23:00 xMff: nice, I'd like to see your patch Feb 19 21:24:47 anyone know where to get the phydev iptables extension? kmod-ipt-extra and iptables-mod-extra claim to provide it but don't actually do so Feb 19 21:25:06 phydev=physdev Feb 19 21:25:37 cshore: iirc, that came with the ebtables patch Feb 19 21:25:43 jow * r19761 /trunk/package/iptables/patches/010-multiport-linux-2.4-compat.patch: Feb 19 21:25:43 [package] iptables: fix multiport match for brcm-2.4 Feb 19 21:25:43 Since the netfilter framework in Linux 2.4 lacks support for target and match revisioning, Feb 19 21:25:43 iptables falls back to the v0 implementation of multiport while only v1 is implemented on Feb 19 21:25:43 the kernel side, leading to an invalid argument error when trying to insert rules using Feb 19 21:25:44 multiport. This patch removes v0 support from the userspace match library to enforce the Feb 19 21:25:45 usage of the v1 match implementation. Feb 19 21:26:03 netprince: so I need ebtables? Feb 19 21:26:21 netprince: I think physdev is different from ebtables Feb 19 21:26:30 I was looking at it at one time Feb 19 21:26:51 thepeople: its been a while since I played with it, you could be right Feb 19 21:27:30 cshore: it is in iptables-mod-extra Feb 19 21:28:10 thepeople: it's not actually there...according to opkg: Feb 19 21:28:10 root@cadersedat:/# opkg files iptables-mod-extra Feb 19 21:28:10 Package iptables-mod-extra (1.4.6-1) is installed on root and has the following files: Feb 19 21:28:10 / Feb 19 21:28:10 /usr/ Feb 19 21:28:11 /usr/lib/ Feb 19 21:28:11 /usr/lib/iptables/ Feb 19 21:28:12 /usr/lib/iptables/libxt_pkttype.so Feb 19 21:28:12 /usr/lib/iptables/libxt_quota.so Feb 19 21:28:13 root@cadersedat:/# opkg files kmo Feb 19 21:28:28 cshore: trunk? Feb 19 21:28:32 yes Feb 19 21:29:16 cshore: what arch? Feb 19 21:29:21 brcm63xx Feb 19 21:29:28 netprince: https://dev.openwrt.org/browser/trunk/package/iptables/patches/010-multiport-linux-2.4-compat.patch Feb 19 21:29:58 cshore: according to the discription it should be there Feb 19 21:30:12 hauke * r19762 /trunk/ (8 files in 3 dirs): [brcm-2.4] update kernel to version 2.4.37.9 and refresh patches Feb 19 21:30:19 thepeople: I know, that's why I was surprised when it wasn't Feb 19 21:30:27 if not I would rack it up to a bug Feb 19 21:30:36 ok Feb 19 21:31:00 thepeople: I'll have see if I can see why Feb 19 21:31:31 cshore: check whether the corresponding symbols are enabled in the brcm63xx kernel config Feb 19 21:31:59 xMff: thats great! I've been missing a correct qos in 2.4 for some months, thanks for helping me. Feb 19 21:32:18 you're welcome Feb 19 21:32:33 xMff: cshore: they should be added via /includes/netfilter.mk which is called by package/kernel/modules/netfilter.mk Feb 19 21:34:06 do you know the kernel symbol name (.e.g CONFIG_?_PHYSDEV)? Feb 19 21:34:16 CONFIG_NETFILTER_XT_MATCH_PHYSDEV Feb 19 21:34:32 CONFIG_NETFILTER_XT_MATCH_PHYSDEV Feb 19 21:35:11 it's not present in the build_dir/linux.../linux.../.config Feb 19 21:35:13 I bet it was renamed Feb 19 21:36:08 I have it in r19111 on x86, came in kmod-ipt-extra.list Feb 19 21:37:37 netprince: where is that located? Feb 19 21:38:19 its the opkg info dir: /usr/lib/opkg/info Feb 19 21:38:57 netprince: is it actually in the file tho? Feb 19 21:39:20 hm I'm on r19631 Feb 19 21:39:31 yeah, I have it available to insmod Feb 19 21:40:11 xt_physdev.ko Feb 19 21:41:31 netfilter.mk is supposed to add the symbol...so I don't know why it's not in the final .config Feb 19 21:43:51 kernel doesn't have phsydev option anymore (make menuconfig in build_dir kernel) Feb 19 21:44:16 appears here for x86 in make kernel_menuconfig Feb 19 21:44:23 cshore: you should try enabling ebtables in kernel, I was pretty sure it was related somehow Feb 19 21:45:03 where would that be in the kernel config? Feb 19 21:46:20 just do it in menuconfig, CONFIG_PACKAGE_kmod-ebtables CONFIG_PACKAGE_kmod-ebtables-ipv4 CONFIG_PACKAGE_kmod-ebtables-ipv6 CONFIG_PACKAGE_kmod-ebtables-watchers Feb 19 21:46:44 cshore: I could be completely wrong though, its been a long time Feb 19 21:46:44 yep, that's it...enabling ebtables enables the physdev option Feb 19 21:47:03 so I'll try recompiling with ebtables Feb 19 21:47:17 looks like I found a dependency Feb 19 22:21:59 xMff: looks like cvs co -> cvs export, hg clone -> hg archive, bzr {co,branch} -> bzr export should be done as well Feb 19 22:22:51 at first I thought it was for a reason but since the whole metadata is deleted right after the checkout I assume we can just export Feb 19 22:24:26 hi all...I've done a build with ebtables and phsydev is now present Feb 19 22:24:45 should I report a bug on kmod-ipt-extra? and/or iptables-mod-extra? Feb 19 22:26:51 hm, ebtables is explicitely disabled in release builds, so physdev will never be available for release users Feb 19 22:27:50 xMff: why is it disabled? Feb 19 22:28:05 performance reasons if I recall correctly Feb 19 22:28:35 xMff: that only affected 2.4 kernels Feb 19 22:29:07 in 2.6 it can be en/disabled via sysctl, right? Feb 19 22:29:37 I'm not certain, I have never needed to disable it in 2.6 Feb 19 22:29:51 its def a problem in 2.4 Feb 19 22:30:27 so hg archive only works locally and github doesn't have the git archive capability enabled Feb 19 22:30:49 well I recently had to disable it in sysctl.conf because bridge firewalling is enabled by default in linux 2.6, this prevented lan<->wifi communication unless the firewall forward policy was set to accept Feb 19 22:31:14 swalker: would have been to easy Feb 19 22:32:07 netprince: I mean this keys: http://openwrt.pastebin.com/m3ac1c46b Feb 19 22:32:09 xMff: I actually need phsydev to allow bridge traffic using shorewall...I suspect it assumes bridge firewalling is on Feb 19 22:32:24 cshore: see the paste above Feb 19 22:32:39 cshore: ah shorewall... well another can of worms I suspect Feb 19 22:32:54 xMff: yes, I have done that before, but it doesn't prevent the perf problems in 2.4 Feb 19 22:33:11 netprince: yeah, I think it is not really separated in 2.4 Feb 19 22:33:19 xMff: so without ebtables there is no bridge firewalling? Feb 19 22:33:41 xMff: if so, then I don't need the bridge option in shorewall and hence don't need physdev Feb 19 22:34:32 cshore: I think you can hook into iptables on a bridge, but its limited Feb 19 22:35:20 netprince: I actually don't want bridge firewall....all I want to do is allow bridge traffic between bridged devices to act as if they were on a hub Feb 19 22:35:30 cshore: you just need to make sure that the bridge-nf-call-* caps are disabled if you don't want iptables to affect traffic passing a bridge Feb 19 22:35:39 xMff: cool Feb 19 22:45:51 jow * r19763 /trunk/package/mac80211/Makefile: [package] mac80211: fix typo introduced in r19760 (#6707) Feb 19 22:52:22 oh wow >2d uptime on my router, how is it possible :D Feb 19 22:52:29 time for a flash! Feb 19 23:01:40 stintel: I sometimes feel like that about the web server...there seem to be regular reboot required updates to Debian Feb 19 23:01:52 oh :/ Feb 19 23:03:05 did you read /usr/share/doc/linux-image-2.6.26-2-amd64/changelog.Debian.gz or similar to see if there was anything relevant for the system ? Feb 19 23:04:07 stintel: yes, but I apply them anyway...it's not a big deal and it's only my home server, so schedule maintenance doesn't apply Feb 19 23:04:27 stintel: just reboot and a couple of minutes later it's back up Feb 19 23:04:28 ah ic Feb 19 23:04:31 indeed Feb 19 23:05:07 ah well, I'm running Gentoo atm in a VM, for my screen+irssi+bitlbee+znc Feb 19 23:05:22 00:05:17 up 8 days, 15:29, 7 users, load average: 0.00, 0.00, 0.00 Feb 19 23:05:43 I once went for 40 days on my web server Feb 19 23:06:11 my desktop (Ubuntu) is regulary exhausting swap - it's usuing it even though memory is not full Feb 19 23:06:21 highest uptime on 1 of my machines was 700+ days, almost 2 years, then the PSU died Feb 19 23:06:57 highest uptime at home was somewhere near 300d Feb 19 23:07:14 stintel: that's pretty impressive Feb 19 23:07:22 you think ? Feb 19 23:07:48 stintel: for home workstation? I reboot at least once a week, usually every few days Feb 19 23:07:55 stintel: for one reason or another Feb 19 23:07:56 no no not home workstation ;-) Feb 19 23:08:04 I even do dualboot again with windows Feb 19 23:08:12 it hardly ever gets >2d Feb 19 23:09:01 stintel: I haven't used Windows except a few minutes in a VM, or because I was somewhere else and didn't have a choice, for several years Feb 19 23:10:43 I've had enough of Windows on 1/06/2008, switched my desktop from vista to Gentoo at that date, and now few weeks ago I decided to install Windows again, be it on another disk so I can dualboot Feb 19 23:11:32 stintel: what sorts of things do you need it for? Feb 19 23:11:52 I got tired of gvfsd-smb segfaulting when importing my music library on a samba share in rhythmbox Feb 19 23:12:06 xMff: hrm, was that firewall issue the same as #6253? Feb 19 23:12:38 cshore: my desktop? Feb 19 23:12:48 stintel: Windows Feb 19 23:12:55 swalker: you mean the bridge firewalling? the sysctl update should have solved this as well Feb 19 23:13:07 cshore: *sigh*, iTunes Feb 19 23:13:12 I have an iphone Feb 19 23:13:17 stintel: ah, yes, I understand Feb 19 23:13:27 and currently no spare money for a n900 :( Feb 19 23:13:29 swalker: r19214 Feb 19 23:14:16 right, so that ticket should get closed, no? Feb 19 23:14:24 correct Feb 19 23:14:50 can you do it? My pc is a bit laggy right now... Feb 19 23:14:53 cshore: if only I knew what kind of company apple was before I bought that shiny phone Feb 19 23:15:04 sure Feb 19 23:15:10 stintel: heh Feb 19 23:15:37 stintel: you mean closed source, or other things too? Feb 19 23:15:55 vendor lockin Feb 19 23:16:00 worse than microsoft Feb 19 23:16:02 imho Feb 19 23:16:08 stintel: I see Feb 19 23:16:17 stintel: that's good to know Feb 19 23:16:37 but then I crashed iTunes on an Mac Mini with Snow Leopard within 3 minutes Feb 19 23:16:47 so it doesn't even work if you have everything from Apple ^^ Feb 19 23:16:55 stintel: heh Feb 19 23:20:06 hmm did I mess something up with the patch to enable wep with hostapd, or am I just making other mistakes right now Feb 19 23:20:31 some one want to take a look at #4191? Feb 19 23:20:32 the patch looked okay Feb 19 23:20:41 yeah it's something different Feb 19 23:20:47 I enabled zabbix-agent Feb 19 23:21:13 http://openwrt.pastebin.ca/1803165 Feb 19 23:23:32 try to fix it or add a ticket to the batch of open issues Feb 19 23:24:11 already looking ;-) Feb 19 23:26:46 build #13 of ixp4xx is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/13 Feb 20 01:22:24 hello any one can helpme with openvpn client configuration? Feb 20 01:25:12 I did copy the configuration from other router but in new one with openvpn V2.10 do not startup tap device and openvpn 2.0.9 did up it Feb 20 01:36:54 openvpn --verb 5 --config /your/open/vpn.conf and read carefully **** ENDING LOGGING AT Sat Feb 20 02:59:57 2010