**** BEGIN LOGGING AT Wed Sep 16 02:59:57 2009
Sep 16 06:03:46 <russell_> hmm.  memory leaking somewhere on r16855 (wgt634u).  after about a week, i get a process that fails to allocate enough memory for a child process and it dies.  i can restart the process, but a little while later, it happens again.  on a reboot, it'll go about a week.
Sep 16 06:04:51 <russell_> kernel is 2.6.28.10
Sep 16 06:05:24 <russell_> any suggestions on how to debug?
Sep 16 06:10:12 <nbd> russell_: are you using madwifi?
Sep 16 06:10:38 <nbd> if so, please update to latest
Sep 16 06:10:43 <nbd> i fixed a rather nasty memory leak
Sep 16 06:12:01 <russell_> yes i am
Sep 16 06:12:32 <russell_> will rev up shortly and see if that fixes, thanks
Sep 16 06:18:01 <russell_> this one?  commit d7b870eaa3e8ef4245023415766fd607abd3f44a
Sep 16 06:19:54 <nbd> yup
Sep 16 06:26:56 <russell_> under what circumstances does it leak?  that is, what was the trigger?
Sep 16 06:27:41 <nbd> receiving beacons from other APs around
Sep 16 06:28:17 <nbd> so if you're in a wifi crowded area, you'll have a continuous leak independent of how much traffic you push through
Sep 16 06:28:28 <russell_> nice
Sep 16 06:29:08 <russell_> any idea how long it's been there?  forever or was it introduced recently-ish?
Sep 16 06:29:42 <nbd> no idea, i guess it's been there for a while
Sep 16 07:13:55 <dkostousov> Hi. How is it to add IE8 support like ie7 support?
Sep 16 08:57:24 <xMff> ping danage
Sep 16 09:15:42 <danage> hi xMff
Sep 16 09:16:12 <xMff> hey
Sep 16 09:16:20 <danage> sup
Sep 16 09:16:34 <xMff> do you remember what OpenWrt Overview in nuwiki was supposed to mean?
Sep 16 09:16:42 <danage> url?
Sep 16 09:16:44 <xMff> architectural overview?
Sep 16 09:16:55 <xMff> http://wiki.openwrt.org/start
Sep 16 09:18:08 <danage> xMff: i don't remember, no. i don't remember how it got in there, maybe it was on Bartman007's list. i guess it could be made anything that fits (i.e. page collection, description, whatever) or even deleted if it doesn't
Sep 16 09:18:13 <danage> do you know about TOH progress?
Sep 16 09:18:33 <xMff> no, but I requested a new meeting so we can finally sort it out
Sep 16 09:18:40 <danage> ok cool
Sep 16 09:18:54 <danage> there doesn't seem to have been a lot of progress in the last couple of weeks
Sep 16 09:19:24 <danage> when is the meeting?
Sep 16 09:19:27 <xMff> I tend to decide on variant "Alternative Layout"
Sep 16 09:19:35 <xMff> in http://wiki.openwrt.org/inbox/trialtoh
Sep 16 09:20:15 <xMff> not yet decided
Sep 16 09:20:43 <danage> me too
Sep 16 09:20:44 <xMff> I just started writing some extensive uci documentation
Sep 16 09:20:57 <danage> ah yes, i imagine that could talk a long time
Sep 16 09:20:58 <xMff> but might take a couple of days until it's somewhat usable
Sep 16 09:21:59 <xMff> I think we should invest some time to make stubs in the appropriate places so that users can edit away
Sep 16 09:22:22 <danage> agreed
Sep 16 09:22:31 <danage> i think we need some senior people to participate too
Sep 16 09:22:42 <danage> the framework is there, but it's not very good at the moment.
Sep 16 09:22:44 <danage> (imo)
Sep 16 09:23:12 <xMff> I'll help to evaulate information correctness
Sep 16 10:46:23 <CIA-36> nbd * r17591 /branches/8.09/package/madwifi/ (Config.in Makefile): madwifi: sync makefile/config changes from trunk to disable compression (fixes a crash bug reported on openwrt-devel@)
Sep 16 12:02:13 <CIA-36> nbd * r17592 /trunk/package/madwifi/patches/ (7 files): madwifi: fix iwconfig tx power display
Sep 16 12:24:24 <CIA-36> nbd * r17593 /trunk/target/linux/atheros/ (12 files in 2 dirs): atheros: remove 2.6.28 support
Sep 16 12:24:49 <CIA-36> nbd * r17594 /trunk/target/linux/generic-2.6/ (10 files in 4 dirs): move the updated version (for 2.6.30+) of ar8216/mvswitch to the generic files directory and remove patches for older kernel versions (no longer in use)
Sep 16 12:25:07 <CIA-36> nbd * r17595 /trunk/target/linux/atheros/config-2.6.30: atheros: disable CONFIG_COMPAT_NET_DEV_OPS, it is no longer necessary
Sep 16 13:56:58 <_trine> [florian], when I plug my ext4 pendrive into my WRT160NL it gives this message in logread:- Update your userspace programs to mount using ext4
Sep 16 14:15:27 <OutBackDingo> nbd: possible breakage with https://dev.openwrt.org/changeset/17594/trunk see http://pastebin.com/d260c5ba6
Sep 16 14:21:21 <CIA-36> jow * r17596 /trunk/package/base-files/ (Makefile files/lib/network/config.sh): [package] base-files: clear ip addresses on interfaces that are about to be added to a bridge
Sep 16 14:23:02 <CIA-36> jow * r17597 /trunk/package/swconfig/ (Makefile files/switch.sh): [package] swconfig: clear ip address on the switch interface, this prevents duplicate assignments of 192.168.1.1 on the DIR-300 at least
Sep 16 14:24:22 <acoul> rb532 goes 2.6.31: http://openwrt.pastebin.com/m16c3e312
Sep 16 14:24:55 <xMff> nice!
Sep 16 14:30:02 <[florian]> acoul: excellent :)
Sep 16 14:30:12 <danage> hi [florian]
Sep 16 14:30:13 <danage> sup
Sep 16 14:30:16 <[florian]> danage: hey
Sep 16 14:30:25 <[florian]> acoul: can you check that korina and via-rhine adapters work fine?
Sep 16 14:30:46 <[florian]> danage: oh well, finishing school stuff, hopefully it will end this week :)
Sep 16 14:30:47 <danage> i was gone for like 2 weeks, you play with your bcm6345 in the meantimeß
Sep 16 14:30:53 <danage> ah ok i see
Sep 16 14:30:57 <danage> i _should_ be doing the same
Sep 16 14:31:01 <[florian]> danage: not at all :p
Sep 16 14:57:05 <acoul> @florian: all rb532 lan ports work like a charm
Sep 16 14:57:18 <acoul> korina port is eth2
Sep 16 14:58:53 <[florian]> acoul: cool
Sep 16 14:58:59 <[florian]> acoul: thanks
Sep 16 14:59:00 <acoul> then the first via-rhine (from left to right after the korina port) is eth0 and the last one eth1
Sep 16 14:59:09 <acoul> this is little odd though
Sep 16 14:59:33 <acoul> as the port sitting in the middle of the three is eth0
Sep 16 15:00:30 <acoul> the cf is also working like a charm, you need this patch though: https://dev.openwrt.org/attachment/ticket/5216/066-block2mtd-probe.patch
Sep 16 15:01:08 <acoul> @florian: we should be thanking you ;-)
Sep 16 15:01:16 <[florian]> oki
Sep 16 15:01:31 <[florian]> acoul: well, in the beginning I managed to break more than I fixed :p
Sep 16 15:02:25 <acoul> well, openwrt is rock solid for me since day 1 ;-)
Sep 16 15:04:55 <acoul> @florian: check this: http://openwrt.pastebin.com/m5b604afb (still working like a charm pushing 20+ Mbit @ 24 x 365)
Sep 16 15:06:07 <[florian]> wgt634u ?
Sep 16 15:07:37 <OutBackDingo> acoul: what unit is this ?
Sep 16 15:07:40 <acoul>  asus wl500g  http://gallery.ozo.com/data/scaled/AWMN/Projects/Wrap-Asus-Box/wl500g.jpg
Sep 16 16:14:38 <nbd> OutBackDingo: about your error: try make target/linux/clean world
Sep 16 16:15:10 <OutBackDingo> nbd: yupp just updated, and did a fresh clean
Sep 16 16:15:16 <OutBackDingo> build is working now
Sep 16 16:15:18 <nbd> ok
Sep 16 16:22:25 <OutBackDingo> nbd: you have any knowledge about these NS5s and the third LED light always on ?
Sep 16 16:23:12 <nbd> nah, haven't played with the led stuff yet
Sep 16 16:23:34 <OutBackDingo> nbd: just curious because on NS2s its always off
Sep 16 16:23:42 <OutBackDingo> on NS5s its always ON
Sep 16 16:24:28 <OutBackDingo> but only when wireless is enabled
Sep 16 16:25:50 <OutBackDingo> nbd: ok, now thats odd, on latest trunk build its off
Sep 16 16:28:52 <nbd> i disabled GPIO LED registration at some point
Sep 16 16:28:57 <nbd> because it was causing trouble on various hw
Sep 16 16:29:01 <nbd> maybe that's what changed
Sep 16 16:29:10 <nbd> should be that way in recent 8.09 builds as well
Sep 16 16:31:50 <nbd> OutBackDingo_: did you get my messages?
Sep 16 16:33:39 <OutBackDingo_> nbd: nope
Sep 16 16:34:02 <OutBackDingo_> nbd: and scratch that last post i was logged into the wrong device
Sep 16 16:34:07 <OutBackDingo_> led is still on
Sep 16 16:34:40 <OutBackDingo_> as soon as wifi start / or wifi stop is run LED starts blinking
Sep 16 16:34:43 <nbd> s hw
Sep 16 16:34:43 <nbd> 18:32 < nbd> maybe that's what changed
Sep 16 16:34:43 <nbd> 18:32 < nbd> should be that way in recent 8.09 builds as well
Sep 16 16:34:43 <nbd> 18:33 -!- OutBackDingo_ [n=quassel@57-31.202-68.tampabay.res.rr.com] has joined #openwrt-devel
Sep 16 16:34:51 <nbd> oops
Sep 16 16:34:53 <nbd> wrong paste
Sep 16 16:35:04 <nbd> 18:31 < nbd> i disabled GPIO LED registration at some point
Sep 16 16:35:04 <nbd> 18:32 < nbd> because it was causing trouble on various hw
Sep 16 16:35:04 <nbd> 18:32 < nbd> maybe that's what changed
Sep 16 16:35:04 <nbd> 18:32 < nbd> should be that way in recent 8.09 builds as well
Sep 16 16:35:16 <OutBackDingo_> nbd: nope
Sep 16 16:35:23 <OutBackDingo_> lights sytill blinkin
Sep 16 16:39:58 <OutBackDingo_> nbd: with a ton pf phyerrs
Sep 16 16:39:58 <OutBackDingo_> according to athstat ath0
Sep 16 16:41:10 <nbd> that might be normal
Sep 16 16:41:22 <nbd> most of the time you can ignore those
Sep 16 16:41:26 <nbd> unless you see other problems as well
Sep 16 16:41:56 <OutBackDingo_> nbd: does loosinf connectivity qualify ???
Sep 16 16:42:00 <OutBackDingo_> LOL
Sep 16 16:42:25 <OutBackDingo_> nbd: its odd, some units, the lights on, they dont work, other units the lights off they work fine
Sep 16 16:42:35 <OutBackDingo_> sometimes the lights on theyll work occaionbally
Sep 16 16:43:16 <OutBackDingo_> yet you can reboot a unit that had a light on, itll go off and work, reboot another with light off, light will come on itll stop working
Sep 16 16:43:55 <OutBackDingo_> and the light is only on if wireless is enabled
Sep 16 16:44:24 <OutBackDingo_> and it flashes like its a traffic / activity led
Sep 16 16:44:57 <nbd> that's really odd
Sep 16 16:45:07 <OutBackDingo_> yupp
Sep 16 16:45:22 <OutBackDingo_> happens on wispstations also
Sep 16 16:45:27 <OutBackDingo_> seems all 5ghz gear
Sep 16 16:45:33 <OutBackDingo_> well ubiquiti gear
Sep 16 16:45:41 <OutBackDingo_> and im shicked nobodies seen this
Sep 16 16:46:06 <OutBackDingo_> i cant be the nly one running openwrt on ubnt ns5a
Sep 16 16:46:09 <nbd> i know quite a few people using ubnt 5 ghz gear
Sep 16 16:46:11 <OutBackDingo_> err ns5s
Sep 16 16:46:14 <nbd> none of those have reported such bugs
Sep 16 16:46:17 <OutBackDingo_> with OpenWRT
Sep 16 16:47:02 <OutBackDingo_> i mean ther is no logic or ryme/reason to this occuring
Sep 16 16:47:29 <OutBackDingo_> ive tested 8.09/8.09.1/8.09.2RC they all act the same way, even trunk
Sep 16 16:52:20 <OutBackDingo_> nbd: i mean it would be enough to think that something in wireless is turning the gpio on
Sep 16 17:04:12 <OutBackDingo_> nbd: ok gpioctl clear 7 turns off the light
Sep 16 17:04:44 <OutBackDingo_> im wondering since this is a NS5 Loco and doesnt have dual mode antennas its triggering something odd
Sep 16 17:05:44 <OutBackDingo_> but somrthing turns it back on even
Sep 16 17:06:29 <nbd> maybe the softled setting of the driver
Sep 16 17:09:17 <OutBackDingo_> nbd: however the ni would think it would do it on ALL units
Sep 16 17:09:40 <nbd> i haven't done any tests on the ns5 loco yet
Sep 16 17:09:49 <nbd> i only know people using the regular ns5
Sep 16 17:10:44 <OutBackDingo_> so it might be loco / wispstation specific
Sep 16 17:11:28 <OutBackDingo_> can you confirm that gpio 7 is in fact for antenna settings ?
Sep 16 17:12:08 <xMff> root@OpenWrt:/# grep 7 /lib/wifi/madwifi.sh
Sep 16 17:12:08 <xMff>                 NanoStation2) antgpio=7;;
Sep 16 17:12:24 <xMff> root@OpenWrt:/# grep NanoStation5 /lib/wifi/madwifi.sh
Sep 16 17:12:24 <xMff>                 NanoStation5) antgpio=1;;
Sep 16 17:15:33 <OutBackDingo_> heh so sysctl -w dev.wifi0.softled=0 and gpioctl clear 7 well disabled the blink blink :)
Sep 16 17:16:36 <OutBackDingo_> sooo to stop them from being ewnabled one needs to modify the madwifi.sh
Sep 16 17:17:13 <xMff> what's in /proc/sys/dev/wifi0/dev_name
Sep 16 17:17:16 <xMff> ?
Sep 16 17:17:58 <OutBackDingo_> xMff: NanoStation Loco5
Sep 16 17:18:29 <xMff> OutBackDingo_: and what's the right gpio for switching antennas there?
Sep 16 17:18:58 <OutBackDingo_> xMff: there is only one antenna on the unit and according to UBNT it cannot be switched
Sep 16 17:19:21 <OutBackDingo_> and there is no external connector on these either
Sep 16 17:19:35 <xMff> ah, well "NanoStation Loco5" matches none of the cases in madwifi.sh, so it's unlikely that "wifi" triggers the led
Sep 16 17:19:46 <xMff> can you recheck without gpioctl but softled off?
Sep 16 17:20:12 <OutBackDingo_> xMff: well according to wifi start, thats what triggers it when its enabled in /etc/config/wireless
Sep 16 17:21:16 <OutBackDingo_> so set softled off after boot, then enable wirless manually see what occurs
Sep 16 17:21:24 <xMff> OutBackDingo_: I meant, since "NanoStation Loco5" matches none of the exceptions in madwifi.sh, it will not execute any gpioctl command
Sep 16 17:22:00 <xMff> yes
Sep 16 17:22:21 <OutBackDingo_> xMff: testing
Sep 16 17:22:31 <OutBackDingo_> booting wifi disabled
Sep 16 17:24:42 <OutBackDingo_> ok booting wifi disabled there is no LED
Sep 16 17:26:05 <OutBackDingo_> though softled is still set to 1
Sep 16 17:26:16 <xMff> check dev.wifi0.ledpin
Sep 16 17:26:25 <xMff> iirc it defaults to 7
Sep 16 17:26:33 <xMff> which would trigger the red led on the loco5
Sep 16 17:26:56 <OutBackDingo_> yupp
Sep 16 17:27:01 <OutBackDingo_> set to 7
Sep 16 17:27:16 <xMff> so the issue is sofled + 7 as default
Sep 16 17:27:24 <OutBackDingo_> yupp
Sep 16 17:28:56 <xMff> so the only piece of info that's required now is what's the right one for the actual wifi led
Sep 16 17:29:10 <xMff> this needs to be added as expection to madwifi.sh or the driver itself then
Sep 16 17:29:27 <xMff> *exception
Sep 16 17:30:08 <OutBackDingo_> xMff: guess that info needs to come from Ubiquiti
Sep 16 17:30:15 <OutBackDingo_> ill speak to mike
Sep 16 17:31:03 <xMff> you can just prope them with gpioctl I think
Sep 16 17:31:16 <xMff> but if you can get first hand info, even better
Sep 16 17:31:56 <OutBackDingo_> if i run gpioctl get 0-7 7 is the only one set high
Sep 16 17:32:03 <OutBackDingo_> the rest are low
Sep 16 17:49:20 <OutBackDingo_> xMff: thanks thats provided alot of insight, though it doesnt answer why it only occurs on units in random methods, ill talk to ubnt and see if theres maybe different board revs also
Sep 16 18:14:32 <CIA-36> nbd * r17598 /trunk/target/linux/brcm-2.4/base-files/etc/init.d/netconfig:
Sep 16 18:14:32 <CIA-36> Fix Buffalo WBR-G54 and WBR-B11 support (updated patch)
Sep 16 18:14:32 <CIA-36> Signed-off-by: Roger Hardiman <roger@rjh.org.uk>
Sep 16 18:15:33 <CIA-36> nbd * r17599 /trunk/package/dnsmasq/files/ (dhcp.conf dnsmasq.init): (log message trimmed)
Sep 16 18:15:33 <CIA-36> dnsmasq startup script update: add notinterface parameter
Sep 16 18:15:33 <CIA-36> This tiny patch makes the dnsmasq init.d script listen to a parameter called "notinterface", which tells dnsmasq to not listen on a defined set of interfaces.
Sep 16 18:15:33 <CIA-36> By default dnsmasq *always* listens on the loopback interface, even if you tell
Sep 16 18:15:35 <CIA-36> it to specifically listen on a certain other interface. Sometimes this is not
Sep 16 18:15:37 <CIA-36> desirable behaviour, so you might want to turn it off. This is useful when, for
Sep 16 18:15:39 <CIA-36> example, you want to have your own DNS server listening on the loopback
Sep 16 19:12:32 <acoul> rb133 goes 2.6.31 with bfs-221 patches: http://openwrt.pastebin.com/m285e1efb
Sep 16 19:13:43 <nbd> how's performance with it? ;)
Sep 16 19:14:50 <acoul> haven't iperfed it yet, but multiple lan flodding with max packates 65k looks better
Sep 16 19:15:47 <acoul> it feels quite nice
Sep 16 19:17:33 <nbd> i'll forward port my bfs/cfs kconfig selection patch to .31 and bfs 221
Sep 16 19:17:38 <nbd> so we can commit it
Sep 16 19:18:02 <nbd> maybe bfs 221 will work on .30 as well
Sep 16 19:18:22 <nbd> i think i'll try .30 first
Sep 16 19:18:25 <acoul> it would be nice to get some real life fild tests on it
Sep 16 19:18:43 <acoul> =field
Sep 16 19:18:52 <nbd> yes
Sep 16 19:19:12 <nbd> i would like to add a toplevel menuconfig entry (build system) for bfs
Sep 16 19:19:18 <nbd> to make it easier to test
Sep 16 19:19:35 <nbd> as i expect testing to continue for a while until we know for sure which one is better
Sep 16 19:19:41 <nbd> considering that cfs keeps improving as well
Sep 16 19:20:14 <acoul> it looks like kolivas is working daily on it
Sep 16 19:20:29 <nbd> yeah
Sep 16 19:20:41 <nbd> i'm lurking in #ck
Sep 16 19:20:42 <nbd> ;)
Sep 16 19:21:13 <acoul> that will give a nice boot to performance
Sep 16 19:21:27 <acoul> so we are gracefull to kolovas no matter what :-)
Sep 16 19:21:33 <acoul> =boost
Sep 16 19:21:44 <acoul> it's late :-)
Sep 16 21:17:38 <nbd> acoul: managed to get bfs 221 to compile on 2.6.30, will test tomorrow
**** ENDING LOGGING AT Thu Sep 17 02:59:57 2009