**** BEGIN LOGGING AT Mon Jun 21 02:59:57 2010 Jun 21 04:20:02 build #55 of rdc is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/rdc/builds/55 Jun 21 06:49:43 hello Jun 21 07:48:14 <_trine> this does not look good and it keeps happening OpenWrt Kamikaze bleeding edge, r21771 Jun 21 07:48:16 <_trine> http://dpaste.com/209694/ Jun 21 07:50:18 well, its just too many connections Jun 21 07:50:39 <_trine> but I'm not doing much with the router Jun 21 07:51:01 <_trine> it's just basically on irc Jun 21 07:51:02 wc -l /proc/net/ip_conntrack Jun 21 07:51:51 <_trine> root@OpenWrt:~# wc -l /proc/net/ip_conntrack Jun 21 07:51:51 <_trine> 72 /proc/net/ip_conntrack Jun 21 07:52:14 do it again if you see the dropping packet errors Jun 21 07:52:40 <_trine> 51 /proc/net/ip_conntrack Jun 21 07:53:19 <_trine> I don't know about this Jun 21 07:53:32 <_trine> it's beyond my knowledge Jun 21 07:54:08 looks like asterisk is eating too much memory Jun 21 07:55:07 <_trine> well normally asterisk sits there running and the memory is normally stable at about 70% Jun 21 07:55:28 <_trine> then all of a sudden recently I have been getting this Jun 21 07:55:41 <_trine> It didn't used to happen Jun 21 07:55:56 <_trine> well not that I noticed anyway Jun 21 07:56:25 last time I've seen such a behaviour was when a windows spam zombie was in my network Jun 21 07:56:38 <_trine> oh Jun 21 07:56:45 <_trine> I am in windows now Jun 21 07:56:47 outgoing smtp killed the router Jun 21 07:56:52 too much connects Jun 21 07:56:57 3-5 per second Jun 21 07:57:07 <_trine> how can I check for that Jun 21 07:57:16 I used tcpdump Jun 21 07:58:00 iftop is okay too Jun 21 07:58:53 <_trine> I'm looking now with iftop Jun 21 07:59:28 <_trine> maybe I will have to keep looking in case it's intermittent Jun 21 08:00:23 <_trine> what should I be watching for Jun 21 08:00:40 many simultanous connections from a specific host Jun 21 08:01:07 if your asterisk is doing many connections you can try to disable conntrack for sip Jun 21 08:01:27 <_trine> how Jun 21 08:01:42 <_trine> in filters Jun 21 08:01:44 <_trine> ? Jun 21 08:02:06 there is a "NOTRACK" target for iptables Jun 21 08:02:37 <_trine> ah again out of my depth of knowledge Jun 21 08:03:35 iptables -t raw -I INPUT -p udp --dport 5060 -j NOTRACK Jun 21 08:03:50 same for tcp Jun 21 08:04:02 port 5061 too Jun 21 08:04:08 <_trine> thank you I will give it a try Jun 21 08:04:22 and then add it to OUTPUT as well Jun 21 08:04:48 also watch your ipv6 Jun 21 08:04:57 it does conntrack too Jun 21 08:08:20 <_trine> back in a bit I have a small family crisis Jun 21 08:22:37 <_trine> 'ptables v1.4.8: Invalid target name `NOTRACK Jun 21 08:22:52 <_trine> seems to be a typo there too Jun 21 08:23:06 <_trine> generated by the router Jun 21 08:27:40 maybe its not included in your build Jun 21 08:28:54 <_trine> what do I need including Jun 21 08:29:23 <_trine> there does seem to be a router generated typo though Jun 21 08:29:46 <_trine> no 'i' at the start of the word iptables Jun 21 08:30:20 <_trine> it's a ' instead Jun 21 08:37:38 <_trine> I sent you a link to my ip tables Jun 21 08:40:24 jow * r21857 /trunk/package/ (broadcom-wl/Makefile broadcom-wl-old/Makefile): [package] broadcom-wl, broadcom-wl-old: binary 2.6 wl restricts the whole menuconfig submenu to brcm47xx, make menu depend on brcm47xx or brcm-2.4 in old and new driver package while brcm-2.4 target is still around Jun 21 08:47:40 <_trine> I entered it manually before I did it in firewall.user Jun 21 08:47:44 <_trine> root@OpenWrt:~# iptables -t raw -I INPUT -p udp --dport 5060 -j NOTRACK Jun 21 08:47:45 <_trine> iptables: No chain/target/match by that name. Jun 21 08:47:45 <_trine> root@OpenWrt:~# Jun 21 08:48:35 _trine: you may need an extra module (package) for the NOTRACK target Jun 21 08:49:23 <_trine> do you know what it's called Jun 21 08:52:47 no, but i'd take a guess at kmod-ipt- Jun 21 08:55:39 kmod-ipt-conntrack-extra Jun 21 08:59:58 install the userspace part Jun 21 09:00:06 iptables-mod-conntrack-extra Jun 21 09:00:18 that will pull the appropriate kmod along Jun 21 09:00:46 if you only install the kmod, you cannot use it since the libraries are missing Jun 21 10:04:35 <_trine> sorry got called away from the computer Jun 21 10:11:22 <_trine> root@OpenWrt:/mnt/root# iptables -t raw -I INPUT -p udp --dport 5060 -j NOTRACK Jun 21 10:11:22 <_trine> iptables: No chain/target/match by that name. Jun 21 10:11:32 <_trine> it still gives me that reply Jun 21 10:20:33 then its probably in iptables-mod-nat-extra Jun 21 10:59:25 <_trine> anyone know where rtz has got to? Jun 21 11:26:56 hi guys... everytime i do "/etc/init.d/network restart" the number if the wlan interface is increased in "/var/run/hostapd-phy0.conf" (interface=wlan0,1,2,3,...) and hostapd cannot start because of this ("Could not set interface 'wlan3' UP"). i'm using the latest trunk... some months ago it was working fine Jun 21 11:27:43 i think there may be a problem in the script which generates the "hostapd-phy0.conf"... but i don't know where i can find it Jun 21 11:28:34 the interface "lan" is a bridged interface ("eth1 wlan0") Jun 21 11:30:17 this is normal Jun 21 11:30:34 you should not reference wireless ifnames in the network config Jun 21 11:31:40 what should i reference instead? Jun 21 11:31:56 you should use "option network" in the wireless config Jun 21 11:34:50 this is confusing (quote luci): "You can bridge several interfaces by ticking the "bridge interfaces" field and enter the names of several network interfaces separated by spaces." Jun 21 11:35:16 yes I know Jun 21 11:36:49 hm doen't help... "interface=wlan4" :( Jun 21 11:37:19 whats the issue? Jun 21 11:37:23 hostapd not coming up? Jun 21 11:38:16 yes... everytime when i restart networking "iwconfig" shows a new wlan interface wlan1,2,3,4,... Jun 21 11:38:40 and "/var/run/hostapd-phy0.conf" contains the number of the new interface Jun 21 11:38:57 when i adjust this number manually (wlan0) and run hostapd it starts fine Jun 21 11:39:29 sounds like an hostapd bug then Jun 21 11:46:55 jow_laptop: do you know where the "hostapd-phy0.conf" is generated? Jun 21 11:50:11 jow_laptop: is openwrt-hackers@ the right place for questions/proposals of slightly unsure legality (I'd like to have /some/ confidentiality, i.e. not indexed by google, until I can assume my actions are most likely legal)? Jun 21 11:52:44 KanjiMonster: yes Jun 21 11:52:53 qecko: /lib/wifi/hostapd.sh Jun 21 11:54:41 good Jun 21 12:00:01 hcg * r21858 /trunk/ (21 files in 13 dirs): Refactor omap35xx sub-targets to cater for beagleboard and gumstix overo Jun 21 12:16:51 i try to integrate the tulip module in trunk withous success. i get a warning WARNING: kmod-tulip is not available in the kernel config. Jun 21 12:17:41 ive addes the definition for this kernel modul to package/kernel/modules/netdevices.mk Jun 21 12:17:57 it look like this define KernelPackage/tulip Jun 21 12:17:57 TITLE:=TULIP Network Support Jun 21 12:17:57 SUBMENU:=$(NETWORK_DEVICES_MENU) Jun 21 12:17:57 DEPENDS:=@LINUX_2_6 Jun 21 12:17:57 KCONFIG:= \ Jun 21 12:17:58 CONFIG_NET_TULIP=y \ Jun 21 12:17:58 CONFIG_DE2104X=m \ Jun 21 12:17:59 CONFIG_DE2104X_DSL=0 \ Jun 21 12:17:59 CONFIG_TULIP=m \ Jun 21 12:18:00 CONFIG_TULIP_MWI=y \ Jun 21 12:18:00 CONFIG_TULIP_MMIO=y \ Jun 21 12:18:01 CONFIG_TULIP_NAPI=y \ Jun 21 12:18:01 CONFIG_TULIP_NAPI_HW_MITIGATION=y \ Jun 21 12:18:02 CONFIG_DE4X5=m \ Jun 21 12:18:09 use pastebin... Jun 21 12:18:20 sry Jun 21 12:18:24 np Jun 21 12:19:40 check whether those symbols are overridden in target/linux/generic-2.6/config-2.6.* Jun 21 12:21:04 here the suitable format for irc @ http://pastebin.com/QHZ5msWz Jun 21 12:25:13 and try adding this to the generic config: http://openwrt.pastebin.com/vZvKBYGr Jun 21 12:36:45 jow_laptop: okay, i found out that mac80211_hostapd_setup_bss() in "/lib/wifi/mac80211.sh" contains "config_get ifname "$vif" ifname" which returns the wrong ifname... (vif=cfg035040 ifname=wlan1,2,3,...) Jun 21 12:39:56 qecko: well try uci -P /var/state revert wireless and repear Jun 21 12:39:58 repeat Jun 21 12:40:17 the count logic is okay, sounds more like state vars going out of sync Jun 21 12:42:25 <_trine> I still cant get this to work iptables -t raw -I INPUT -p udp --dport 5060 -j NOTRACK Jun 21 12:42:49 <_trine> I have recompiled the firmware to include the extra mods Jun 21 12:42:58 <_trine> but it still refuses Jun 21 12:43:28 <[Fate]> in some ar71xx setup code, I see ethX's PHY interface mode set to RMII, but with a speed of 1000M. how does that work? shouldn't it be RGMII? Jun 21 12:43:34 jow_laptop: no luck, wlan16 -> /sbin/wifi -> wlan17, ... Jun 21 12:43:54 qecko: cat /tmp/.uci/wireless Jun 21 12:44:37 /tmp/.uci/ is empty Jun 21 12:45:10 <_trine> When I try to enter it I get this :- Cannot use -I with -I Jun 21 12:45:37 _trine: on ssh? Jun 21 12:45:44 <_trine> yes Jun 21 12:45:54 try to enter it manually Jun 21 12:46:01 sounds like you have weird chars in there Jun 21 12:46:08 <_trine> that's what I'm doing in ssh Jun 21 12:46:20 <_trine> do you mean to type it in Jun 21 12:46:27 yes... Jun 21 12:46:30 qecko: weird Jun 21 12:46:32 <_trine> and not cutting and pasteing Jun 21 12:46:34 qecko: check /var/state Jun 21 12:46:39 _trine: yes Jun 21 12:47:25 <_trine> and I get :- iptables: No chain/target/match by that name. Jun 21 12:47:32 jow_laptop: /var/state/ contains an empty file "wireless" Jun 21 12:48:21 <_trine> I off out for a bit Jun 21 12:48:27 <_trine> see you later Jun 21 12:48:37 <_trine> I'm ** Jun 21 12:50:25 qecko: well, search for config_set "$foo" ifname "$bar" Jun 21 12:50:33 qecko: and find out how they're counted Jun 21 12:54:42 build #61 of goldfish is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/goldfish/builds/61 Jun 21 13:01:35 jow_laptop: https://dev.openwrt.org/browser/trunk/package/mac80211/files/lib/wifi/mac80211.sh in function "enable_mac80211()", line 229 "/sys/class/net/wlan$i" gets counted... there are indeed 19 wlan directories right now Jun 21 13:02:28 jow_laptop: the question is: why does restarting wifi lead to a new directory? Jun 21 13:19:22 qecko: maybe the old vifs are not properly destroyed Jun 21 13:29:12 jow_laptop: in disable_mac80211(), line 184 the for-loop isn't executed cz: "ls: /sys/class/ieee80211/phy0/device/net: No such file or directory", instead the directories .../device/net:wlan[0,1,2,3] do exist Jun 21 13:29:49 qecko: that could be the explaination for your issues Jun 21 13:30:08 maybe the structure in sysfs changed recently Jun 21 13:41:26 jow_laptop: changed line 184 to "for wdev in $(ls /sys/class/ieee80211/${phy}/device/ | grep net: | cut -d':' -f2 2>/dev/null); do" Jun 21 13:41:29 works now :) Jun 21 13:42:57 my eyes bleed ;) Jun 21 13:43:25 can you post the output of ls -1 /sys/class/ieee80211/${phy}/device/ ? Jun 21 13:43:30 *paste Jun 21 13:44:27 jow_laptop: http://pastebin.com/RmQkR08F Jun 21 13:45:20 before this i had net:wlan[0,1,2,...22] Jun 21 13:51:20 for wdev in $(ls /sys/class/ieee80211/${phy}/device/ | sed -ne 's/^net://p'); do Jun 21 13:53:22 jow_laptop: sure, works too ;) but i thought that grep and cut is more "low-level" than sed... sed may not be available in some environments Jun 21 13:54:14 it slow, ugly and inefficient Jun 21 13:54:23 and needs more forks Jun 21 13:54:34 sed is guaranteed to be there Jun 21 13:54:37 iirc Jun 21 13:55:00 jow_laptop: you're absolutely right :) on openwrt-devices no problem, but i know devices which have grep+cut but no sed (regexps) Jun 21 13:55:23 why would such device use openwrt scripts? ;) Jun 21 14:07:20 jow_laptop: whatever... thanks for your help :) before commiting this to the svn it would be better to check the format of the sysfs directory structure first... otherwise other ppl may have the same error again ;) Jun 21 14:13:09 yes Jun 21 14:13:14 what kernel do you run? Jun 21 14:19:19 jow_laptop: 2.6.33.2 Jun 21 14:25:18 okay Jun 21 14:27:11 jow_laptop: finally got a serial port installed on my whr-g54s: http://openwrt.pastebin.com/z4E4BEw6 Jun 21 14:28:16 netprince: this looks innocnetz Jun 21 14:28:18 -z Jun 21 14:28:39 *innocent Jun 21 14:29:49 here is what happens when entering failsafe by pressing a button: http://openwrt.pastebin.com/s7AQLZEz Jun 21 14:31:36 hmm Jun 21 14:32:22 can you create a ticket regarding this switch issue? Jun 21 14:32:48 sure Jun 21 14:34:24 I'm going to revert the apsta mode and see what it looks like Jun 21 15:14:12 jow_laptop: is this the patch I linked last week? https://dev.openwrt.org/changeset/21857. Did you get it to compile? I had some trouble but haven't sorted it out yet... Jun 21 15:19:58 netprince: what do you mean? :) Jun 21 15:20:20 the patch above fixes kmod-brcm-wl-old for brcm-2.4 Jun 21 15:20:37 since the k2.6 package restricted the whole menu to brcm47xx Jun 21 15:21:33 but I committed all the profile stuff I gave you in pastebin previously Jun 21 15:35:30 jow_laptop: I couldn't get it to compile, and I had to reduce the patch down to http://openwrt.pastebin.com/ubEHxW1u Jun 21 15:35:51 I can try building the latest rev in a few mins and let you know if you like Jun 21 15:36:08 thats exactly the issue fixed by the patch you linked above Jun 21 15:36:35 https://dev.openwrt.org/changeset/21857 Jun 21 15:36:43 no need to rebuild, it'd have the same effect Jun 21 15:36:50 as your diff Jun 21 15:37:08 ok Jun 21 16:17:30 build #53 of ps3 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/53 Jun 21 16:22:02 strange, the new wl module will sometimes load on my whr-g54s after about 5 mins, and sometimes it just hangs Jun 21 16:25:10 it is disabled in /etc/config/wireless Jun 21 16:37:15 same for your switch as it seems Jun 21 16:37:27 maybe they share some critical resource Jun 21 16:37:32 gpio lines or something Jun 21 16:37:54 that would make sense Jun 21 16:39:19 and the hang produces no oops or something on the serial console? Jun 21 16:39:39 nope, just sits, I can actually press and get newlines Jun 21 16:39:53 sometimes it will eventually continue Jun 21 16:40:01 hmm Jun 21 16:46:38 there is the usually kernel taint warnings, and Disabling lock debugging due to kernel taint Jun 21 16:48:04 I'll try removing it from modules.d and loading it manually Jun 21 16:48:16 ok Jun 21 16:53:33 same, just hangs Jun 21 16:53:50 had top running via ssh, it hangs too Jun 21 17:01:20 strace: http://openwrt.pastebin.com/d6hSPdDb Jun 21 17:03:22 updated strace: http://openwrt.pastebin.com/gnKFCta0 Jun 21 17:36:46 build #64 of cobalt is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/64 Jun 21 18:20:39 nbd: why isn't 'uci rename network.wan=wan-3g' working? Jun 21 18:23:28 because section names must not contain '-' Jun 21 18:23:33 ok. Jun 21 18:24:42 let's test multiwan script now :) Jun 21 18:25:15 nbd: what is the best way to bring up wan connection at boot time? Jun 21 18:27:43 dogge10: did 3g work? Jun 21 18:28:16 Lechucks: with vodafone websessions, yes Jun 21 18:28:19 :) Jun 21 18:28:26 tell me Jun 21 18:29:13 what? Jun 21 18:29:32 how did you do? jeje Jun 21 18:29:50 i have same modem and router but i cant do it Jun 21 18:30:34 yesterday i deactivate autorun Jun 21 18:31:11 well just installed kmod-usb-serial, kmod-usb-serial-option, kmod-usb2, kmod-usb-ohci, kmod-usb-uhci & comgt Jun 21 18:31:26 then setup /etc/config/network Jun 21 18:31:30 that's it Jun 21 18:32:54 are you insert anything in /dev/modules.d/60-usb-serial ? when do you install this modules reboot or insert the modem? Jun 21 18:33:35 cat /etc/modules.d/60-usb-serial Jun 21 18:33:37 usbserial vendor=0x19d2 product=0x0063 Jun 21 18:34:25 lsusb command? Jun 21 18:35:23 lsusb: http://openwrt.pastebin.com/mXchrtMP Jun 21 18:37:14 lechucks: lsusb: http://openwrt.pastebin.com/mXchrtMP Jun 21 18:37:35 what is genesys logic? Jun 21 18:38:02 i think the sdcard on board Jun 21 18:38:11 sdcard is connected via usb Jun 21 18:38:18 my lsusb http://openwrt.pastebin.com/VrANMjJN Jun 21 18:38:26 dogge10: ok do it Jun 21 18:39:00 what next? Jun 21 18:39:21 i have fail in pincode timeout Jun 21 18:39:21 configure wan section in /etc/config/network Jun 21 18:39:35 can you pastebin this section? Jun 21 18:39:47 i had this too. just power cycle the board Jun 21 18:40:09 i dont understand Jun 21 18:40:27 do you say that all i need is reboot? Jun 21 18:40:53 i think so Jun 21 18:40:57 uci export network: http://openwrt.pastebin.com/kXGdpS4u Jun 21 18:41:25 ttyUSB3 Jun 21 18:41:29 i have USB0 Jun 21 18:41:57 try with network.wan3g.device=/dev/ttyUSB3 Jun 21 18:43:43 * dogge10 is playing now with the multiwan script :) Jun 21 18:44:55 ok ill try this Jun 21 18:45:36 dogge10: not found Jun 21 18:53:15 dogge10: not command network Jun 21 19:07:25 dogge10: are you create chatscripts or something? Jun 21 19:10:49 build #51 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/51 Jun 21 19:32:25 dogge10: try this Jun 21 19:33:06 pppd noaccomp nopcomp novj nobsdcomp noauth lock crtscts 115200 /dev/ttyUSB3 persist usepeerdns defaultroute replacedefaultroute unit 0 linkname wan ipparam wan connect USE_APN=event.vodafone.de Jun 21 19:33:18 and see what tell us logread Jun 21 20:12:55 build #62 of gemini is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/62 Jun 21 21:40:30 jow * r21859 /trunk/package/kernel/modules/netdevices.mk: [package] kernel: package tulip family modules into kmod-tulip Jun 21 21:51:06 jow * r21860 /packages/sound/mpd/ (Makefile files/mpd.init): [packages] mtd: make grep for playlist config option more explicit to discard comments (#7499) Jun 21 21:55:07 jow * r21861 /packages/utils/ (4 files in 4 dirs): [packages] minicom, microcom, picocom, setserial: s/terminal/Terminal/ for submenu entries (#7494) Jun 21 21:59:01 jow * r21862 /packages/utils/pwcrypt/ (. Makefile): [packages] add "pwcrypt", a cli wrapper around the crypt() call (#7490) Jun 21 22:04:54 build #52 of xburst is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/xburst/builds/52 Jun 21 22:10:52 jow * r21863 /packages/utils/debootstrap/ (6 files in 3 dirs): [packages] add "debootstrap", allows bootstrapping Debianish systems from within OpenWrt (#7500) Jun 21 22:14:53 jow * r21864 /packages/net/unbound/Makefile: [packages] unbound: update to v1.4.5 (#7498) Jun 21 22:23:23 jow * r21865 /packages/admin/zabbix/Makefile: [packages] zabbix: don't chown on host system (#7497) Jun 21 22:26:35 jow * r21866 /packages/utils/ntfsprogs/ (. Makefile): [packages] add "ntfsprogs", utilities to manage ntfs volumes from within OpenWrt (#7501) Jun 21 22:27:32 <[Fate]> jow: on a packaging spree? =) Jun 21 22:28:10 just getting some tickets resolved Jun 21 22:29:52 jow * r21867 /packages/net/haproxy/Makefile: [packages] haproxy: update to v1.4.7 (#7475) Jun 21 22:34:43 jow * r21868 /packages/net/olsrd/patches/900-force_copy_olsrd_conf_file.patch: [packages] olsrd: force non-interactive install (#7470) Jun 21 22:39:58 phew Jun 21 22:40:00 done Jun 21 22:40:05 \o/ Jun 21 22:42:24 its funny to fix stuff I'd never use Jun 21 22:51:01 <[Fate]> yeah debootstrap from within openwrt looks weird Jun 21 22:51:50 actually it could make sense for those hybrid owrt kernel / debian systems Jun 21 22:52:01 or for embedding in a bootable openwrt iso Jun 22 00:27:11 want to move it to the admin subdir or fix the section/category Makefile vars? Jun 22 00:27:37 sentence cut off? Jun 22 00:30:35 no Jun 22 00:31:14 then I don't understand the question Jun 22 00:31:41 ah you mean debootstrap? Jun 22 00:32:37 aye Jun 22 00:32:44 sure Jun 22 00:34:41 jow * r21869 /packages/ (admin/debootstrap/ utils/debootstrap/): [packages] debootstrap: move package to the admin directory **** ENDING LOGGING AT Tue Jun 22 02:59:56 2010