OpenWrt

[00:01] <joebush> Interesting, dnsmasq doesn't get the parameter to not listen on LAN port
[00:07] <joebush> nevermind, my bad.
[00:22] <joebush> is there a problem with ipcalc? END doesn't seem to be calculated correctly
[00:23] <[olli]> joebush, dnsmasq?
[00:23] <joebush> yes, the call in the startup script
[00:23] <joebush> was fine in rc4
[00:24] <[olli]> joebush, openwrt uses the variables dhcp_start and dhcp_num, to what are they set?
[00:24] <joebush> 100 and 100
[00:24] <joebush> it's the calculation for end that gets screwed up
[00:24] <joebush> I get 236
[00:25] <joebush> testing ipcalc yields some interesting results
[00:26] <joebush> is it possible to browse the source on the dev wiki?
[00:26] <joebush> for ipcalc, that is
[00:27] <[olli]> joebush, https://dev.openwrt.org/file/tags/whiterussian_rc5/openwrt/package/base-files/default/bin/ipcalc
[00:28] <joebush> oh, it's a script hehe. Could have viewed it locally
[00:35] <spectra-> i've been using openwrt with wds for a while now on my wrts with antennas on the roof, to connect a few households
[00:36] <spectra-> i guess since rc4, problems started appearing.. connections got slower and eventually didn't work at all
[00:37] <spectra-> speaking in numbers.. when we put up the wrts, on the first day, we had like 2,5mb/s via wifi.. using hyperwrt with wep
[00:37] <spectra-> no routing, all households in one network
[00:38] <spectra-> when we switched to openwrt, speed dropped to 1-1,5mb/s, which still was acceptable (i thought it was due to wpa)
[00:38] <spectra-> anyhow, the speed kept dropping and lately, nothing worked at all
[00:39] <spectra-> now, just to give it a try, we switched two of the access points to hyperwrt again... with wpa, we get almost 1mb/s again
[00:39] <spectra-> with openwrt running only one hour ago, the connection was far from being solid
[00:41] <EnolaGay> i am looking for someone able to mount a cramfs image and tar.gz me it. plz contact me. ( cramfs image is "big endian with holes" format)
[00:41] <spectra-> the only actual change now is that the 'main' ap (equipped with an omni antenna) only has to handle one wds connection, whereby under openwrt nvram was set to three wds connections
[00:42] <spectra-> so now.. any ideas?
[00:43] <K1> hmm, are there updated lighttpd packages? Not 1.3.13 but something like 1.4.11?
[00:53] <akrot_> jop
[00:53] <joebush> olli: looks like the script's logic is faulty
[00:53] <joebush> do you have a link to ipcal from rc4?
[00:53] <joebush> ipcalc
[00:54] <[olli]> joebush, yeah, just thought about the same
[00:54] <[olli]> joebush, https://dev.openwrt.org/file/tags/whiterussian_rcr/openwrt/package/base-files/default/bin/ipcalc
[00:54] <[olli]> joebush, https://dev.openwrt.org/file/tags/whiterussian_rc4/openwrt/package/base-files/default/bin/ipcalc
[00:54] <joebush> file not found :o
[00:54] <joebush> tried that before
[00:57] <joebush> ipcalc is binary on my other boxes
[00:58] <[mbm]> ?
[00:59] <joebush> ipcalc isn't calculating end properly for dnsmasq
[00:59] <[mbm]> specific example?
[01:00] <joebush> ipcalc 10.0.0.1 255.255.255.0 10.0.0.100 5
[01:00] <joebush> gives an end address for dhcp of 10.0.0.101
[01:00] <joebush> seems the logic is faulty
[01:01] <joebush> have the source for rc4 ipcalc? It's a binary, so I can't view it
[01:01] <frop> http://www.rafb.net/paste/results/lj5U0z14.html
[01:03] <[olli]> [mbm], http://pastebin.com/624053 <-- another ipcalc example
[01:03] <joebush> This fixes it end=start+ip2int(ARGV[4])
[01:03] <joebush> at least for my simple examples
[01:05] Action: [mbm] hates awk's broken math
[01:05] <joebush> hehe
[01:05] <[mbm]> it can only add 31 bits
[01:05] <[mbm]> not 32
[01:06] <joebush> I did a manual calculation and had problems with that algorithm though
[01:06] <[mbm]> the routines are correct, it's just awk's ability to do math
[01:07] <joebush> end=or(start,and(ip2int(ARGV[4]),compl(netmask)))
[01:07] <joebush> Is that supposed to take a number of addresses and return the end address?
[01:08] <[mbm]> ack ..
[01:08] <[mbm]> end=or(network,and(ip2int(ARGV[4])+start,compl(netmask)))
[01:09] <joebush> that one I believe :)
[01:09] <joebush> thanks, much better
[01:10] <[mbm]> hmm.. awk still isn't parsing it right
[01:10] <joebush> works for me
[01:10] <[mbm]> yeha, for that case.. not for all
[01:11] <joebush> limit doesn't seem to work hehe
[01:11] <[mbm]> no, it's the end
[01:11] <joebush> you're right
[01:11] <joebush> too tired :+
[01:13] <joebush> well, that's good enough for now at least
[01:39] <Quozl> On desktop Linux, starting pppd with MPPE causes modutils to load the ppp_mppe_mppc module automatically, yet this does not happen on OpenWrt White Russian RC4. Is there a module alias list?
[01:39] <[mbm]> Quozl: there's no modprobe; if you want a module loaded, you have to use insmod
[01:40] <Quozl> [mbm]: thanks. So adding to /etc/modules is the way ... pptp ipkg doesn't add /etc/modules.d entry, yet depends on kmod-mppe, where shall i make the suggestion?
[01:41] <[mbm]> right, you add to /etc/modules
[01:41] <Quozl> or /etc/modules.d if it's done by a package?
[01:41] <[mbm]> you can file a ticket on dev.openwrt.org
[01:41] <[mbm]> right
[01:41] <Quozl> thanks.
[02:08] <Omnic> hrm.
[02:08] <Omnic> now, something like firehol (that defines a simple firewall 'language') ported to the openwrt environment would be cool..
[02:09] <Omnic> Or perhaps written from scratch would be better.
[02:10] <Quozl> [mbm]: https://dev.openwrt.org/ticket/412
[02:13] <Quozl> wiki.openwrt.org ... "500 Internal Server Error (sorry - we're updating)" on submit of a change.
[02:16] <[mbm]> Quozl: try the wiki again
[02:17] <Quozl> [mbm]: same message ... (refresh of a POST)
[02:18] <[mbm]> hmm ok.. I'll look into it later .. somewhat busy
[02:18] <Quozl> [mbm]: no worries.
[02:22] <Omnic> [mbm]: I successfully patched my ppp (rp-pppoe.so) and kernel (pppoe.o) on Friday to support my 3com DSL modem (has non-standard pppoe frame types).. I'm going to rework the patches to make changing the frame types into options, would you accept such a patch?
[02:23] <Quozl> Omnic: shouldn't it go upstream? has it already (in 2.6.16?)
[02:23] <Omnic> Quozl: yes, but openwrt doesn't use that kernel.. (given, I haven't checked kamikaze)
[02:23] <[mbm]> busy.
[02:24] <Omnic> Quozl: as for has it been done upstream? highly likely not.
[02:24] <moonwatcher> nbd: ?
[02:25] <nbd> yep
[02:26] <moonwatcher> emule is really happy :)
[02:26] <moonwatcher> everythign is great :)
[02:26] <moonwatcher> what now?
[02:27] <moonwatcher> ui?
[02:27] <nbd> sounds great :)
[02:27] <moonwatcher> god i just spent 72 hours strech writign a HUGE xml document
[02:28] <Quozl> Omnic: PPTP for instance has DSL modem and provider specific quirks in it that I've accepted ... often because the RFC is vague ... so I'd expect upstream kernel team to welcome well presented suggestions on PPPoE. Just wondered, that's all.
[02:29] <Omnic> Quozl: yeah, highly likely, and if I get around to doing this, I'll submit upstream too.. but in this instance, I'm also interested in OpenWRT getting the changes, since that's what I use it on.
[02:53] <moonwatcher> oh wow!
[02:53] <moonwatcher> i havent noticed
[02:53] <moonwatcher> RC5 is out...
[02:54] Action: moonwatcher looks at [mbm] excitingly... look its out...
[02:55] <[mbm]> they wouldn't let me call it rc6 :/
[02:55] <moonwatcher> shoot
[02:56] <moonwatcher> [mbm]: never mind, they will let you name the next one RC6...
[02:56] <[mbm]> I'm sure there's more that I forgot to list on the changelog
[02:56] <Quozl> ;-)
[02:56] <moonwatcher> well thats what happnes when you only write the changelog before the release...
[02:56] <Quozl> i always do a recursive diff to verify my changelogs ... but admittedly i'm dealing with one or two packages at a time. ;-}
[02:57] <Quozl> pity the web interface doesn't have english in it for rc5. ;-)
[02:57] <[mbm]> well, there's all the crap on dev.openwrt.org/timeline
[02:57] <[mbm]> but most of that is just 'fixed bug #xxx'
[02:58] <[mbm]> and I'm not even going to begin to list all the bugfixes
[02:59] <[mbm]> Quozl: yeah, somehow I think l33t will be added before English
[02:59] <Quozl> [mbm]: well done on the release ... rsync'ing it now.
[03:16] <Omnic> [mbm]: in the readme:
[03:16] <Omnic> What the heck is with all these directories?
[03:16] <Omnic> You *legend*
[03:17] <J4k3> hahaha wtf
[03:17] <J4k3> webif doesn't have english?!
[03:17] <J4k3> haha
[04:16] <[hC]> Anyone know how i may try to debug this? I'm using openwrt to join another wireless network in client mode, a network scan sees the network, but trying to join it does nothing at all. Ive also seen this happen when connecting from windows xp it will 'attempt' to join the other wireless network but cant (although it sees it with full signal)
[04:16] <[hC]> Is there a way i might debug why the connection wont establish? its a netgear wireless router im attaching to.
[04:17] <J4k3> you're not in its mac filter?
[04:17] <[hC]> its open
[04:17] <[hC]> no restrictions. if i leave it and try to connect again in like 5 hours, it'll probably let me in.
[04:17] <J4k3> is it your AP?
[04:18] <[hC]> yep
[04:18] <[hC]> upstairs
[04:18] <[hC]> when it connects, its great. sometimes it simply wont let me on. an iwconfig shows an empty SSID even though i specified it
[04:18] <[hC]> and it worked previously.
[04:18] <[hC]> tried rebooting my wrt54g... nothin.
[04:18] <[hC]> joined another open AP nearby (non netgear) and it works fine
[04:20] <Quozl> [hC]: wish i knew how to debug that kind of thing too, but sorry, i don't know.
[04:21] <[hC]> No problem, thanks though.
[04:21] <[hC]> Ive tried connecting with various wireless cards, all do the same thing. I think its a netgear specific thing.
[04:22] <Quozl> It certainly sounds like a problem with the AP.
[04:22] <[mbm]> the ssid shown in iwconfig is that of the last connection
[04:22] <[mbm]> so even though you've specified it, unless it actually connects, iwconfig won't show it
[04:22] <[hC]> iwlist eth1 scanning shows it, at a -45dBm signal level
[04:23] <[hC]> and iwconfig infact does not show it, cause it doesnt connect.
[04:23] <[mbm]> and unfortunately most of the wireless is provided in binary only form with no debug info
[04:23] <[hC]> im trying to determine why it would show up with a strong signal, yet not let me on.
[04:23] <[hC]> ill try to dick with the wireless settings on it later.
[04:23] <[mbm]> I'm guessing that if the problem shows up with both openwrt and xp the problem is on the netgear side
[04:24] <[hC]> ESSID:"NETGEAR"
[04:24] <[hC]> Channel:6
[04:24] <[hC]> Quality:0/0 Signal level:-53 dBm Noise level:-92 dBm
[04:24] <[hC]> thats what iwlist shows
[04:31] <Eggplant> [hC], how far away is that router? ive got a -50 on a ap that was 2 miles away (its running HOT)
[04:32] <[hC]> you know,
[04:32] <[hC]> i just found out
[04:32] <[hC]> WEP was enabled on it by my roomate
[04:32] <[hC]> and he didnt feel like it was important to mention.
[04:32] <[hC]> :)
[04:32] Action: Quozl chuckles
[04:32] <[hC]> Always stupid things
[04:32] <[hC]> the XP connection thing though, happened all the time.. with no wep key.. that i was not imagining.
[04:33] <Quozl> although you could wish there was something that would say "can't connect, no WEP key."
[04:33] <[hC]> yeah
[04:33] <[hC]> linux is great for that sort of stuff in regards to its wireless tools
[04:33] <[hC]> it either works, or it seems like theres simply no radios in sight.
[04:34] <[hC]> or even if iwlist showed you that it was wep enabled :)
[05:34] <roh> muh
[05:47] <Quozl> Successful upgrade of two WOAP54G's to RC5, using mtd of squashfs trx.
[05:48] <ejka> Omnic: svn up
[05:48] <Omnic> ejka: -ECHAN :)
[05:48] <Omnic> and how'd you know it was an openwrt tree? :P
[05:49] <ejka> Omnic: :)
[05:49] <Omnic> wow, there's not a whole lot that's changed in RC5
[05:49] <Quozl> Omnic: that's good, isn't it?
[05:50] <Omnic> not really good or bad, just an observation.
[05:50] <Quozl> Omnic: what has the rate of change been like?
[05:52] <Omnic> god knows why you're asking me.
[05:52] <Quozl> Omnic: well, you said there's not a lot that has been changed, so i was wondering why you thought that ... had you measured what was previously changed.
[05:53] <Omnic> Quozl: 'svn up'. lists the files added/changed between revisions.
[05:53] <Omnic> there was less than 30.
[05:53] <Omnic> I lie. there was 30.
[05:54] <Quozl> Omnic: oh, so you're saying the *number of files* changed is low ... what was the complexity of the changes? How many lines changed?
[05:54] <r0tu> Hrrm.. I"ve been trying and trying.. and don't want to bother anyone.. but I'm almost out of hair! I can't get my IPtables to block local netbios broadcasts withing owrt... I've tried to block it on each table and am blocking udp/tcp on 137:139
[05:54] <Omnic> Quozl: buggered if I know, I'm not an svn user.
[05:55] <r0tu> by nbd on Sun Mar 26 20:39:56 2006
[05:55] <r0tu> Users on #openwrt: r0tu burtonez roh Revenger amonkey Sinbios alp aurel32 frop [hC] _Perun thepeople mrverbos1 Eggplant webmind |dennis| Pervert sunjon wizardyesterday Quozl admin__ inh LeeJunFan funky Weedy joebush Kevin` MaxHorvath J4k3 StefanG Kryojenik brosioz Denmark gerdi_ EnolaGay Rawplayer daniel_bergamini akrot_ iDaywalker kempo_ r0nny p-crat_ russell_ ptitlouis atari_ Jah_ netmax^OFF obi Hydra tojoe Blorb cybe
[05:55] <r0tu> Users on #openwrt: Plnt RedT0mt0m noxlux ixs mXr _ivc reinhold|bot eXi KarlNapf Bac9 tris Bold z3ro Ge0rG gr8w11ne LuX Doughbwoy koollman Vince|| HackerMaster 42 zoobab piiE NAiLzzZ crodler sky jnc fish Th- radi0head redondos catastrop toktok crashe
[05:55] <r0tu> Whooa!!! Sorry about the post
[05:56] <zinx> be thankful i'm not opped
[05:56] <zinx> or you'd have been automagically permbanned
[05:56] <r0tu> iptables -A FORWARD -p UDP --dport 137:139 -j DROP
[05:56] <r0tu> ? why ... it was a mistake
[05:56] <zinx> because my script doesn't care if it was a mistake :DD
[05:56] <r0tu> lol, ahhh
[05:57] <r0tu> zinx: do you have any ideas about what I may be doing wrong.. or where / what I have to add to my script to accomplish this seemingly simple task
[05:57] <mohr> maybe I'm mistaken, but it appears that there isn't a squashfs 8mb rc5 image
[05:58] <J4k3> mohr: squashfs doesn't have the 4mb/8mb issue
[05:58] <mohr> oh
[05:58] <Quozl> mohr: nor does rc4 ;-)
[05:59] <mohr> so I can flash my new wrtsl54gs with openwrt-brcm-2.4-squashfs.trx from rc5?
[06:02] <mohr> thanks for making openwrt, you're the best :P
[06:04] <J4k3> wow, thats beautiful
[06:04] <J4k3> my patch still hasn't been committed.
[06:05] <J4k3> nor has the workaround been made... wbx committed against the makefile, but obviously didn't bother testing it either.
[06:14] <J4k3> hmm wbx changed a couple things, but missed the root of the problem entirely.
[06:26] <dacb> hey
[06:26] <dacb> I am trying to figure out which of the pads (image bottom, board center http://24.18.210.117/~dacb/530g.JPG ) are the uart's GND, RX, and TX for the WL-530g. Has anyone here cracked open one of those devices?
[06:27] <Quozl> dacb: no, i haven't got one ... but GND will usually match the chip ground and power supply ground, TX will usually source or sink a few milliamps, and RX will float.
[06:28] <Quozl> dacb: on that basis it's usually not difficult to figure it out, if done carefully.
[06:29] <dacb> Quozl: you think using a DVM on those pads would be better than looking for RX by attaching a TTL serial and reboting the device a mess of times?
[06:30] <Quozl> dacb: i'd use a DVM in volts mode to find GND first, that should be easy, if there's any ambiguity about it then keep trying until GND is proven.
[06:31] <dacb> rgr
[06:31] <Quozl> dacb: then i'd use a DVM in volts AC mode while sending large amount of text out /dev/ttyS0, so as to find TX ... or use a CRO or sampler. RX should be silent.
[06:31] <Quozl> dacb: then i'd tie TX and RX together on the board temporarily and verify that "cat /dev/ttyS0" shows whatever I send to /dev/ttyS0 with an echo.
[06:32] <Quozl> short cuts lead to short circuits. ;-}
[06:33] <dacb> I like your idea of testing /dev/ttyS0 via linux, but I don't have a shell via 'net w/ the wl-530g and the 1.9.4.4 firmware :(
[06:33] <dacb> So I will have to dig about as you suggest
[06:33] <dacb> w/ DMM
[06:42] <[mbm]> .
[06:43] <Quozl> dacb: perhaps the firmware emits out TX during pre-boot and boot. That may be sufficient for you.
[06:47] <dacb> I'm almost sure of it - the boot flags from the firmware source show a serial console
[06:49] <dacb> not really going to help me find RX :)
[07:00] <Quozl> dacb: finding RX ... it should float to either 0V or 5V given a 10k resistor. That's a good indication that it is an input, TTL or CMOS. If it doesn't float, then it's not an input, try next pin.
[07:10] <dacb> Quozl: can you clarify: I will connect the pin in question (probably to gnd first) via 10k resister and if the V at the pin floats, it is a good sign it is an in?
[07:13] <Quozl> dacb: measure the voltage on the pin relative to ground. then connect a 10k resistor from the pin to ground, and measure the voltage again. for an input, it will be pulled to ground by the 10k resistor ...
[07:14] <Quozl> dacb: then remove the resistor, connect a 10k resistor from the pin to the supply (maximum 5v), and measure the voltage again. for an input, it will be pulled to 5v by the 10k resistor.
[07:14] <Quozl> dacb: an output on the other hand will simply continue to reflect whatever it has been told to show.
[07:16] <dacb> rgr
[07:26] Action: Quozl finishes rsync of RC5 ... coolies.
[07:48] <SplinterFL> anyone know how to set default 404 in the httpd.conf ? (default httpd in openwrt)
[07:49] <[mbm]> the error is hardcoded into httpd
[07:49] <SplinterFL> anyway to redirect to a cgi?
[07:49] <SplinterFL> captive portal
[07:50] <[mbm]> nbd had some code for that, can't remember if it made it into rc5
[07:50] <SplinterFL> is rc5 downloadable yet?
[07:51] <EnolaGay> i am looking for someone able to mount a cramfs image and tar.gz me it. plz contact me. ( cramfs image is "big endian with holes" format)
[07:51] <[mbm]> SplinterFL: see topic, see forum post, see release announcement on openwrt.org .. etc, etc ...
[07:52] <SplinterFL> ugg.. must have downloaded v4 just before 5 was posted yestarday...
[07:52] <[mbm]> EnolaGay: mount it yourself, nobody is going to take an interest in a random channel spam that doesn't offer any reason or incentive
[07:53] <EnolaGay> i cant, the cramfsck i have do not mount "cramfs image with holes", it seem i need a big endian cpu/environement
[07:54] <SplinterFL> try an emulator?
[07:55] <EnolaGay> do not think it will work. and btw i do not read a lot about this mysterious cramfs with holes format. any know more about it ?
[07:55] Action: Quozl finds a 12V to 5V DC/DC converter module in Australia and orders a couple
[07:56] <SplinterFL> dl's v5 to see whats new..
[07:58] <fish_> re
[07:58] <EnolaGay> mkcramfs can make image with holes (there is an option for this) but i didn't find any tools to extract this kind of cramfs
[07:59] <[mbm]> ...
[08:01] <AlbertoSt> any plans on supporting siemens se515? (broadcom 6345)
[08:01] <[mbm]> none of the developers have such a device
[08:02] <AlbertoSt> ok
[08:03] <J4k3> [mbm]: please commit the madwifi-old patch I just sent. People keep asking me about it
[08:04] <J4k3> nbd said he was going to add something last time around, but never got around to doing it... I don't think committing this patch is going to prevent him from doing whatever he needs to do either way.
[08:05] <Denmark> Its already released, I am afraid..
[08:06] <[mbm]> J4k3: ? you mean ticket #414 or did you actually sign your name?
[08:08] <J4k3> 414
[08:08] <J4k3> I forgot to sign it
[08:09] <[mbm]> ok, I'll take a look later
[08:10] <AlbertoSt> I think the developers have a se515 already ;) http://forum.openwrt.org/viewtopic.php?id=2341 ... I guess it will just take some time..
[08:18] <Denmark> how bad is it to delete nvram, (mtd erase nvram), apart from loosing current configuration?
[08:21] <[mbm]> Denmark: depends on the device, it's generally considered a bad idea
[08:23] <Denmark> [mbm]: Yeah... its an ASUS WL-500gd. I guess the first variable you want back, is boot wait.
[08:24] <[mbm]> probably better to just learn what the variables do
[08:24] <Denmark> ok .. flashing from wlan ...
[09:28] <mohr_> I've set up wds per the openwrt wiki, but I just can't seem to make it work
[09:28] <mohr_> can someone give me some tips?
[09:28] <[mbm]> mohr_: gimme 10 min
[09:28] <mohr_> I've read the Configuration howto and the WDS howto
[09:28] <mohr_> ok thanks
[09:41] <[mbm]> mohr_: that post answer your question?
[09:41] <mohr_> I've read it I think
[09:41] <mohr_> lemme double check
[09:42] <[mbm]> you read it? I just posted it .. *points at wrt(
[09:43] <mohr_> well something very similair anyway, posted to the openwrt forums
[09:43] <mohr_> quite right, not that post, but another like it
[09:43] <mohr_> basically I don't want a bridge, I want a repeater with wds
[09:45] <[mbm]> it's all the same
[09:47] <mohr_> mbm, how can I tell if the wds interfaces are "added" to br0?
[09:48] <[mbm]> brctl show
[09:48] <mohr_> I have wds0.4915 in my interface list
[09:48] <mohr_> ok
[09:48] <[mbm]> if you have it, it'll be added to the bridge
[09:48] <[mbm]> word of warning, having the interface only means a wds connection attempt, not always an established connection
[09:48] <mohr_> ok, the interface is listed under the "interfaces" column in brctl show
[09:50] <mohr_> the interface name is listed on both devices, and has the same number
[09:50] <mohr_> the firewall on the secondary unit has been disabled
[09:50] <mohr_> the lazywds and wl0_wds nvram settings are correct
[09:51] <mohr_> (triple checked)
[09:51] <mohr_> all encryption has been disabled per the wiki
[09:51] <mohr_> but I cannot ping the gateway from the "client" wrt
[09:52] <mohr_> do you have any suggestions for troubleshooting?
[09:53] <srcquench> moroning
[09:55] <[mbm]> mohr_: I actually like to use lazywds for troubleshooting
[09:56] <mohr_> how so?
[09:56] <[mbm]> if I turn off wl0_wds then it doesn't create new wds interfaces automatically
[09:57] <[mbm]> so if a new connection is attempted with wl0_lazywds then I know that it saw the connection
[09:57] <[mbm]> because there will be a new wds* interface
[09:58] <[mbm]> .. anyywas, most common error is forgetting that all wds devices need to be on the same channel, or setting the wl0_wds to the wrong mac address
[09:58] <mohr_> ah, so set lazywds to 1 and clean wds?
[09:58] <mohr_> clear*
[09:59] <[mbm]> yeah, although you might need to install wl and do 'wl wds none' to actually clear out the old wds* interfaces (or just reboot)
[10:01] <mohr_> mbm, ok thank you
[10:01] <mohr_> I'll keep trying
[10:01] <mohr_> gotta go for now, rebooting the gateway
[10:07] <mohr> [mbm], you still there?
[10:08] <mohr> [mbm]: you were correct, my tired eyes had missed a colon on one of the routers
[10:08] <[mbm]> cool
[10:08] <[mbm]> so everything works now?
[10:08] <mohr> thank you for your help :)
[10:09] <mohr> yes, it seems so
[10:09] Action: Revenger hunts for a changelog rc4->rc5
[10:09] <[mbm]> Revenger: dev.openwrt.org/timeline
[10:09] <mohr> now I must sleep :)
[10:09] <Revenger> gracias.
[10:09] <mohr> good night
[10:09] <mohr> and thank you again :)
[10:10] <[mbm]> that will show you every damned commit from the start of the project up until the latest commit
[10:10] <[mbm]> and is rather impossible to actually read
[10:14] <Revenger> [mbm]: true. unless you are used to that stuff :P
[10:16] <[mbm]> Revenger: well, imho most of the commit messages are lower than worthless in terms of describing what the commit was actually for
[10:18] <Revenger> *nod*
[10:18] <Revenger> but better this than nothing
[10:19] <[mbm]> Revenger: is there a specific change you're looking for? it might be quicker to look through the tickets and see the bug reports that got fixed
[10:20] <Revenger> [mbm]: nbd made some fixes and created a test-image for me (leds on a buffalo AP) a while ago.
[10:20] <Revenger> wondering if that one made it into rc5
[10:20] <[mbm]> should.. rc5 patches every bug that was reported for rc4
[10:21] <Revenger> i guess its quicker to just try it (tm) than searching thu those commits :P
[10:38] <Quozl> wiki.openwrt.org "500 Internal Server Error (sorry - we're updating)" ... on submit of a wiki page update.
[10:39] <Quozl> ... worked second time.
[10:39] <Quozl> PPTPDHowto and PPTPClientHowto updated for RC5 with lab testing ... t'was fun creating a routing loop and tangling the poor thing.
[10:56] <Quozl> Does anyone know how to reduce transmit power on WRT54G running OpenWRT 1.0 RC5?
[10:59] <Pooh22> Quozl: the normal response to that question is probably "don't", but you can try looking at wl or iwconfig ;-)
[11:00] <Pooh22> I've seen some people take off the antenna's to reduce the output (which also doesn't work very well to reduce the power)
[11:00] <Quozl> Pooh22: yeah, i know the usual response is don't, and i understand the risks of warping the spectral response ... but my antenna dish is too good and i want to reduce the power a tad.
[11:01] <Pooh22> then it probably has a different "spectral response" anyway?
[11:02] <[mbm]> Quozl: wl0_txpwr
[11:02] <Eggplant> Whats wrong with just turning down the TX power?
[11:02] <Quozl> [mbm]: thanks.
[11:02] <[mbm]> if you want to really lower the output, set that to 1
[11:02] <[mbm]> (aka 1mW)
[11:02] <[mbm]> oh and wl0_txpwr was ignored before rc5
[11:03] <Eggplant> [mbm], any known issues with wl0_antdiv?
[11:03] <Quozl> Eggplant: same reason why turning it up isn't always good ... the output stage is being operated out of spec, may make things worse.
[11:04] <Eggplant> Quozl, ive never heard that before, and ive been doing RF for 10 years
[11:04] <Eggplant> I dont claim to be a guru by any means, only that that is news to me after a long time
[11:05] <Quozl> Eggplant: okay, i've only done it for 23 ... but i've no spectrum analyser. http://wl500g.info/showthread.php?t=12&page=2 shows an example of the spread into nearby frequencies ... if you can cope with that garbage, then it should be okay.
[11:07] <ads> morning
[11:07] <Pooh22> Are there any rc5 upgrade notes for "heavily" modified rc4 systems?
[11:08] <Eggplant> Quozl, looks like they are running hotter then spec, not colder
[11:08] <Quozl> Pooh22: one of mine was heavily modified ... you lose all changes bar what is in nvram, so take a copy of what you need first.
[11:08] <Pooh22> ok, tnx Quozl
[11:09] <Quozl> Pooh22: and it's in the README. ;-}
[11:09] <Pooh22> duh! ;-)
[11:09] Action: Pooh22 flogs self
[11:11] <Pooh22> I may need to recompile some packages for rc5 (if they've not been upgraded already)
[11:17] <joebush> mbm: from the ipcalc problem we discusses yesterday...when exactly does 'end=or(network,and(ip2int(ARGV[4])+start,compl(netmask)))' not work?
[11:17] <joebush> that is the line you copied in
[11:17] <joebush> and why not simplify it too 'end=start+ip2int(ARGV[4])'?
[11:18] <[mbm]> joebush: already patched
[11:18] <joebush> since start is already calculated
[11:18] <[mbm]> it should be end=start+ARGV[4]
[11:19] <[mbm]> and it was changed in rc5
[11:19] <joebush> cool, that's what I was thinking
[11:19] <joebush> yeah, guess ip2int is overkill for an int :)
[11:19] <[mbm]> (we updated the rc5 binaries after we noticed, since the release announcement hadn't been posted)
[11:20] <jnc> oh for crap's sake, another rc
[11:20] Action: jnc grins
[11:21] <jnc> realise i don't complain about hard work in making the rc5 possible
[11:21] <jnc> just my own irritation when something new comes out, feeling like i have to update to it asap
[11:23] <joebush> mbm: don't suppose you've had any revelations about my strange dnsmasq problem? No change with rc5
[11:24] <[mbm]> joebush: not really
[11:25] <joebush> yeah, nobody knows why it happens. It isn't crucial since the only query I need on my 500gx is downloads.openwrt.org which i stuck in hosts
[11:25] <joebush> But would be nice to figure it out. It's bugging the hell out of me not to know why it happens :)
[11:30] <Quozl> how do i add text to a ticket?
[11:34] <wbx> joebush: describe your problem to me.
[11:34] <wbx> Quozl: not possible at the moment.
[11:47] <joebush> wbx: DNS queries work fine for clients after my 500gx is booted up, but from the shell I can't do queries. If I kill dnsmasq and restart it from the script, the 500gx can do queries, but the clients can't.
[11:47] <joebush> I did a fresh install of rc5 last night, but it didn't help. Wiped nvram and everything
[11:48] <joebush> To update ipkg I had to put downloads.openwrt.org in the hosts file.
[11:48] <[mbm]> joebush: dnsmasq has nothing to do with queries from the router itself
[11:48] <[mbm]> those are done directly using the nameservers in /etc/resolv.conf
[11:48] <[mbm]> which should be your isp's nameservers
[11:49] <[mbm]> dnsmasq will read /etc/resolv.conf and create a new dns server for the lan
[11:49] <joebush> Right, which works correctly on boot.
[11:49] <[mbm]> .. have you been playing with resolv?
[11:49] <joebush> except something is then stopping local queries from the router
[11:50] <joebush> nope
[11:50] <[mbm]> firewall changes?
[11:50] <joebush> nothing. fresh install
[11:50] <[mbm]> hmm..
[11:50] <joebush> exactly hehe
[11:51] <[mbm]> let me reflash one of mine and se if I can reproduce it
[11:51] <joebush> The irritating part is how the problem gets reversed after killing and restarting dnsmasq
[11:51] <joebush> My wl-hdd didn't have this issue
[11:52] <joebush> was using it as a router while I waited for the 500gx
[11:53] <joebush> brb...
[12:04] <joebush> back
[12:06] <joebush> I just confirmed that this happens the first time dnsmasq is started regardless of when it is started. I had it not run at start and ran it manually after the router was booted and it worked exactly the same. After killing and restarted was also the same as before.
[12:07] <joebush> Was thinking that start order was somehow effecting it, but that isn't the case apparently.
[12:14] <[mbm]> joebush: hmm .. just out of curiosity, change the dns servers listed in /etc/resolv.conf to something other than your isp's servers
[12:14] <[mbm]> (I'm wondering if your isp is doing something weird)
[12:15] <joebush> can't hurt to try.
[12:15] <joebush> wan_dns?
[12:16] <[mbm]> what's wan_proto?
[12:16] <joebush> dhcp
[12:16] <joebush> the address is obtained automatically
[12:16] <[mbm]> then it will ignore wan_dns
[12:17] <joebush> should be possible to manually configure dns even if using dhcp for address
[12:17] <[mbm]> it is, you have to remove the symlink for /etc/resolv.conf
[12:17] <[mbm]> and manually put the file there
[12:18] <joebush> hehe, ok
[12:18] <joebush> Was taking a closer look at udhcpc to see if there was a parameter there
[12:18] <[mbm]> which reminds me.. the other option is to run dnsmasq with /tmp/resolv.conf and then have /etc/resolv.conf use 127.0.0.1
[12:20] <joebush> I thought /etc/resolv.conf was consistenly used. what looks for nameservers in /tmp/resolv.conf?
[12:20] <joebush> forget it
[12:20] <joebush> read wrong
[12:21] <[mbm]> antyhing that creates resolv.conf manually uses /tmp/resolv.conf
[12:21] <[mbm]> er .. s/manually/automatically/
[12:21] <[mbm]> antyhing that creates resolv.conf automatically uses /tmp/resolv.conf
[12:21] <[mbm]> which /etc/resolv.conf points to
[12:21] <[mbm]> you can tell dnsmasq to use /tmp/resolv.conf directly
[12:22] <[mbm]> and then use /etc/resolv.conf to point to dnsmasq for local queries
[12:22] <joebush> that should work
[12:22] <joebush> hopefully...will let you know soon
[12:24] <joebush> actually, that is what the script did when rc4 was installed on my wl-hdd
[12:24] <joebush> that is why it worked right
[12:24] <joebush> didn't notice that difference until you mentioned it
[12:24] <joebush> why would the script be different when installed on the 500gx?
[12:25] <[mbm]> it never did, you must have changed it yourself
[12:25] <joebush> nevermind, that was dhcp.leases
[12:26] <joebush> got a bit overexcited for a sec.
[12:30] <joebush> which parameter is used to specify the resolv.conf file?
[12:30] <joebush> forget it hehe
[12:30] <joebush> skipped passed it. I need to start waiting 3 extra seconds before hitting enter
[12:33] <joebush> giving it a try, brb
[12:40] <joebush> actually, that parameter was for dnsmasq.conf, not resolv.conf
[12:41] <joebush> is it possible to force dnsmasq to use another resolv.conf?
[12:41] <joebush> shit, there it is
[12:41] <[mbm]> -r, --resolv-file=path Specify path to resolv.conf (defaults to /etc/resolv.conf).
[12:41] <joebush> again, 3 seconds!!
[12:41] <[mbm]> you can probably put it in dnsmasq.conf
[12:41] <joebush> have to start to listen to myself
[12:41] <Quozl> ;-)
[12:41] <joebush> just couldn't find the parameter :|
[12:42] <[mbm]> syntax would probably be "resolv-file /tmp/resolv.conf"
[12:42] Action: Quozl used strings on it
[12:42] <Mitu> anybody has conected an wrt to an ipcop VPN ?
[12:43] <joebush> that fixed it
[12:44] <joebush> thanks mbm
[12:44] <joebush> were you able to duplicate it, btw?
[12:45] <joebush> (reboot to confirm it works on fresh start)
[12:46] <[mbm]> joebush: nope
[12:46] <[mbm]> works fine for me
[12:46] <joebush> your 500gx didn't havve that issue?
[12:46] <[mbm]> isn't a 500gz
[12:46] <joebush> oh
[12:47] <joebush> brb, then I can ask you about ethers....doesn't appear to be working in rc5
[12:49] <joebush> nevermind, that is working now too
[12:49] <joebush> well, my day suddenly looks brighter :D
[12:59] <ixs> Sonic: generally, that is no problem. You might need to write a custom startup script however, as I do not know, if openwrt supports that out of the box.
[12:59] <[mbm]> Sonic: if you want multiple ip addresses, use vlan1:0 for the second ip address
[13:00] <[mbm]> (changing vlan1 will probably block your internet access)
[13:00] <Sonic> in fact vlan1 hasn't any ip because my wan ip is on ppp0
[13:01] <ixs> [mbm]: might even be easier to implement. But in generall the pppoe stuff is independent of the tcp-configuration on the interface
[13:01] <[mbm]> ah ok I get it
[13:01] <[mbm]> Sonic: foo_ifname=vlan1 foo_proto=static foo_ipaddr=x.x.x.x
[13:02] <[mbm]> set those vars and then 'ifup foo'
[13:02] <Sonic> thanks I will try
[13:03] <[mbm]> (you can change 'foo' to anything)
[13:07] <Sonic> finally I have just type : ifconfig vlan1 192.168.1.2 netmask 255.255.255.0
[13:07] <[mbm]> Sonic: that's essentially what the above does
[13:07] <Sonic> okay
[13:13] <nikolaj_a> can anyone help me with this: iftop: can't load library 'libpcap.so.0.8'
[13:13] <ixs> nikolaj_a: you need to install libpcap.
[13:13] <[mbm]> ipkg install libpcap
[13:13] <ixs> ipkg should do the job
[13:15] <nikolaj_a> it installed automatically, but its version 0.9
[13:15] <joebush> do I need to install kmod-usb2 AND kmod-usb-uhci?
[13:16] <[mbm]> usb2 is for usb2.0 support only and udhci is for 1.0/1.1
[13:16] <joebush> only...ok...guess my printer isn't usb2 afterall :|
[13:23] <gile> how can i implement ip_conntrack_sip in the new RC5
[13:26] <Sonic9_> okay I have set a ip static on vlan1, how can I restore dhcp on vlan1 ? or no ip ? thx
[13:26] <florian_> Sonic9_: nvram set wan_proto=dhcp or none
[13:27] <[mbm]> Sonic: huh? you should still have wan_ifname=vlan1 set
[13:27] <Sonic9_> but I must reboot to apply .. how can I restore without reboot router, with ifconfig it's possible ?
[13:27] <[mbm]> you don't reboot
[13:28] <[mbm]> any setting like wan_ifname/wan_proto can be applied with "ifup wan", which is why I gave you that "ifup foo" earlier
[13:28] <Sonic9_> okay
[13:30] <gile> in the sorce of the RC5 i did not find ip_conntrack_sip , i think i read in the forum that conntrack wil be added in RC5 for sip and rtsp
[13:30] <gile> and now i can only find rtsp
[13:32] <[mbm]> gile: it didn't get added in time for the rc5 release
[13:32] <gile> so can i include it somehow?
[13:33] <[mbm]> gile: sure, go read about cross compiling and creating packages and then create yourself a package
[13:33] <[olli]> seems i killed my d-link dsl-360t by just pressing the reset button - http://pastebin.com/624786
[13:34] <[mbm]> [olli]: huh? that looks like more than just a reset button
[13:34] <[olli]> [mbm], really, just the reset button
[13:34] <[mbm]> wonder what it's trying to write
[13:35] <Revenger> during flash, huh?
[13:35] <[olli]> Revenger, no. only reset
[13:35] <Revenger> blyme.
[13:35] <[mbm]> [olli]: so it's bricked? or did it only crash once
[13:35] Action: Revenger crosses d-link out of the to-buy list.
[13:35] <gile> mbm: can you point me in the right direction, link or something
[13:36] <[mbm]> gile: wiki.openwrt.org has a search
[13:36] <[olli]> [mbm], i guess it's a brick. it crashes all the time now. btw. that was with the original firmware, never tried openwrt
[13:37] <[mbm]> !?!? ..
[13:37] <gile> mbm: thanks :)
[13:40] <Quozl> [olli]: that's amazing and sad ... so this is the built-in loader firmware that did this, right?
[13:40] <gile> mbm: is there a way to put netfilter patch in the source for RC5 and make a new build
[13:40] <[olli]> Quozl, right
[13:42] <[olli]> google tells me that the fritzbox which is ar7 as well have the same problems
[13:42] <[mbm]> gile: http://wiki.openwrt.org/BuildingPackagesHowTo http://downloads.openwrt.org/docs/buildroot-documentation.html
[13:44] <[olli]> and i can't even reach the bootloader console
[13:45] <Quozl> [olli]: could you be looking at a failure caused by something else, and the reset button being coincidence?
[13:45] <[mbm]> [olli]: it's trying to write something to flash but can't .. I don't understand why (and their error handling sucks if it crashed that bad)
[13:53] <[olli]> have no time currently to debug this thing.. theirfor i will not buy d-link in the future if it's such a crap
[13:54] <Quozl> [olli]: they could easily have fixed it in a later revision of boot loader or PCB ... it's the sort of problem that plagues manufacturing for a while until it gets fixed, but doesn't hinder their ability to ship product.
[13:55] <Quozl> [olli]: when i worked in a similar situation, we had 30% failure rate during reflashing for a few weeks on the manufacturing floor before we figured out how to fix it ... all it meant was passing the 30% fails back up the line to have their eeprom erased at a special workbench.
[13:56] <Quozl> [olli]: while i think that d-link is crap, i can imagine it might not always be crap. ;-}
[13:57] <Quozl> WL-HDD RC4 ... with USB modules present, it reboots every few minutes ... removing the modules prevents the reboots. No serial port so I haven't tried to find anything more.
[13:59] <[olli]> Quozl, yeah. i know what you mean. but i bricked it that bad with the original firmware
[14:00] <Quozl> [olli]: static damage perhaps? i don't know the model, but did it have a console already?
[14:01] <joebush> had a problem with dnsmasq again after installing the usb modules
[14:01] <[olli]> Quozl, static damage.. could be possible, but i don't think so. it has headers for a serial console .. a ttl converter is required
[14:02] <joebush> removed and reinstalled, and reconfigured the resolv.conf files again and now it's ok
[14:02] <joebush> wierd stuff
[14:07] <Quozl> what, there's new stuff on a web interface? darn, missed it. where's my shell? ;-}
[14:08] <joebush> was list_installed replaced by something?
[14:08] <joebush> in ipkg
[14:08] <wbx> joebush: are you using rc5?
[14:08] <joebush> ye
[14:08] <joebush> s
[14:08] <Quozl> joebush: ipkg list_installed works for me on rc5.
[14:08] <wbx> joebush: have you done ipkg update?
[14:08] <Quozl> ah, i did an ipkg update.
[14:08] <joebush> yes
[14:09] <joebush> it isn't even listed as an option in the help
[14:09] <Quozl> joebush: does it say ipkg version 0.99.149?
[14:10] <joebush> hmm, where do I get the version?
[14:10] <Quozl> ipkg --version
[14:10] <joebush> that doesn't work
[14:10] <joebush> hehe
[14:10] <joebush> option isn't listed
[14:11] <Quozl> you've got one seriously dodgy ipkg there. try "ipkg invalidcommand", does it list a version at the beginning?
[14:11] <joebush> no, no version anyway
[14:11] <Quozl> strange. are you sure you are on rc5?
[14:12] <joebush> installed ipkg...looks better now
[14:12] <joebush> yes
[14:12] <joebush> installed it last night
[14:12] <joebush> maybe I got an incomplete binary
[14:13] <joebush> using the micro trx
[14:13] <joebush> 4mb
[14:14] <joebush> which binary did you flash?
[14:15] <joebush> no change in the fie I used. Maybe it's a faulty compile
[14:16] <joebush> s/fie/file
[14:17] <zviratko> hi, does someone have experience with Asus WL-300g? When I try to upload the image via web it says "Invalid firmware image", and TFTP timeouts :/
[14:19] <wbx> zviratko: which image you are using?
[14:19] <joebush> The WL-HDD has that problem too. You need to use the update utility that came with the unit. You should notice that light stops blinking. Let it timeout. Then run the ftp command
[14:19] <joebush> Works for me on the WL-HDD
[14:19] <wbx> zviratko: can you try the flash.sh script?
[14:19] <wbx> zviratko: i think wl500g, wl-hdd and wl300g need special tftp commands.. ASUSSPACELINK foobar..
[14:23] <lut1> after the release of RC5 I wonder if I can use wpa_supplicant for wpa instead of nas, is it possible?
[14:26] <wbx> lut1: wpa_supplicant has support for madwifi and other cards, but not for broadcom. so it depends on your device and wifi card.
[14:26] <joebush> wbx: but how do you run that as a single command?
[14:26] <wbx> joebush: the script? sh flash.sh openwrt-2.4-jffs2-4MB.trx asus
[14:27] <joebush> i meant the ASUSSPACELINK hack
[14:27] <joebush> windows tftp is single command
[14:28] <joebush> never got that wo work, but I guess most people here use Linux, so it isn't a big problem
[14:28] <joebush> s/wo/to
[14:28] <lut1> wbx: hmm openwrt has no lspci command, asus wl-500g deluxe uses broadcom or?
[14:28] <joebush> yes
[14:28] <wbx> joebush: does windows tftp accept commandline?
[14:28] <wbx> lut1: yes
[14:29] <joebush> yes, but it is a single command operation
[14:29] <joebush> not sure how the hack can be implemented
[14:29] <Wonka> lut1: openwrt has an lspci ipkg
[14:29] <joebush> that's why I use the utility the way I do
[14:29] <wbx> joebush: can you help me?
[14:30] <joebush> with what?
[14:30] <wbx> docs
[14:30] <lut1> thx, do you know if wpa_supplicant will ever support broadcom? or will there be open broadcom drivers?
[14:30] <joebush> I'm no expert, so it depends what you'd like me to do.
[14:30] <Pooh22> did something change with dnsmasq from rc4 to rc5 (dns stopped working with my old config file)
[14:30] <wbx> there are open reverse engineered broadcom drivers. the just have no Master mode yet.
[14:31] <lut1> so these drivers can't be used in aps?
[14:32] <wbx> joebush: look at the page here: http://wiki.openwrt.org/OpenWrtDocs/Hardware/Asus/WL500GD
[14:32] <wbx> Send image with Firmware Restoration technique
[14:33] <wbx> is the description there okay? can you fix documentation issues? and if they are fine. can you make a extra wiki page and create links from all Asus specific models to that page?
[14:33] <wbx> be part of an open source project and help.
[14:33] <wbx> lut1: can be used, but at the moment only as client
[14:35] <lut1> damn, I'm looking forward to have my computer with completely free software. atm I'm stuck to ati, java and broadcom :(
[14:35] <joebush> Looking now...I think I am going to reflash later since the micro trx is obviously a bit sick
[14:35] <zviratko> wbx: what flash.sh script? I have the original ASUS firmware inside...
[14:35] <[olli]> lut1, good luck
[14:35] <wbx> zviratko: then just use webupgrade with the trx file
[14:35] <zviratko> wbx: i tried the newest WR release, 4MB 8MB and SquashFS versions
[14:35] <zviratko> wbx: i tried, it says invalid image
[14:35] <zviratko> wbx: md5sum is right
[14:36] <wbx> zviratko: never use 8mb jffs on asus wl* 4 mb flash devices
[14:36] <zviratko> wbx: also, i did use the GET ASUSSSPACELINK... TFTP command, it just causes the PWR light to go off and times out
[14:36] <zviratko> wbx: i just tried :)
[14:36] <wbx> zviratko: hmm. so webupgrade does not work, tftp does not work. strange
[14:37] <wbx> zviratko: can you ping the ip in Failover Mode?
[14:37] <zviratko> wbx: you mean the one when i power on while holding reset? (and PWR is slowly blinking?)
[14:37] <zviratko> wbx: yes i can
[14:37] <wbx> yes.
[14:38] <zviratko> wbx: yes i can
[14:38] <wbx> okay, then try again the flash.sh script
[14:38] <zviratko> what flash.sh script? :)
[14:38] <wbx> zviratko: which operating system you use?
[14:38] <zviratko> OSX
[14:38] <wbx> http://openwrt.org/downloads/utils/flash.sh
[14:38] <wbx> hmm. don't know if osx has a usable tftp client installed
[14:39] <zviratko> wbx: did just does the same thing
[14:39] <joebush> I assume the firmware recovery application on the cd doesn't work on OSX
[14:39] <wbx> http://wiki.openwrt.org/OpenWrtViaTftp
[14:39] <zviratko> wbx: i type the same thing :)
[14:39] <wbx> zviratko: may be you forgot binary ;p
[14:39] <joebush> does OSX have windows emulation?
[14:39] <wbx> zviratko: or it is called octet ..
[14:40] <wbx> zviratko: read the MacOS X section, did you connect via cross over cable?
[14:40] <zviratko> wbx: ah
[14:40] <zviratko> MacTFTP Client commands: (For use with OS X. tftp in terminal did not work for me)
[14:40] <zviratko> i just used the command line one
[14:43] <joebush> wbx: I'll take a closer look at the page later after I reflash. But if this page is generic for flashing Asus products, it can be separated completely from the 500gx page
[14:44] <joebush> Then I can examine the Asus products pages and remove similiar info and put in a link
[14:45] <joebush> wbx: after rereading what you wrote before I realized that that is what you wanted me to do hehe
[14:48] <joebush> wbx: are you running rc4 or 5 on a 500gx?
[14:53] <wbx> joebush: right now? on my system beside me i am running kamikaze 2.4
[14:54] <wbx> joebush: okay, just say "ready" if i should look over your changes :}
[15:20] <wbx> zviratko: any news?
[15:44] <joebush> wbx: http://wiki.openwrt.org/OpenWrtDocs/Hardware/Asus/Flashing
[15:45] <joebush> Just threw that together to start gathering flashing info generic to Asus products
[15:46] <joebush> will try to take a closer look later after reflashing
[15:48] <joebush> Does anyone know how a WDS enables AP would react to being wired physically to the net in addtion? Will it survive or cause a loop?
[15:49] <joebush> That is assuming it is the second AP on the wired lan
[15:50] <[mbm]> your question really doesn't have anything to do with wds, only creating routing loops
[15:50] <[mbm]> and if it's actually a loop, then yes it will cause problems
[15:51] <KeefR> hm, does spanning tree work over wlan? that might solve that problem
[15:52] <[mbm]> spanning tree really doesn't help much
[15:57] <wbx> joebush: okay, good starting point. but wl500g should not be reset to factory defaults. wl500gx is fine. and wl500g needs special tftp get command, wl500gx needs not.
[15:57] <KeefR> why not, stp puts one of the links in blocking mode
[15:57] <KeefR> so no traffic is passed: no loop
[15:58] <joebush> mbm: does that mean it isn't smart to link as many ap's as possible with wds? they create unwanted lops?
[15:58] <wbx> joebush: you can create 5 ap and use wds.
[15:58] <wbx> joebush: just one have the uplink..
[15:58] <wbx> joebush: the rest is one big wireless
[15:59] <[mbm]> KeefR: that's the ideal version, in reality you run into stp errors
[15:59] <joebush> wbx: I'm aware that gx doesn't need them. Thought I mentioned the models that do
[15:59] <KeefR> [mbm]: most of the time it works, but normal stp is simply to slow
[16:00] <joebush> wbx: I have 4, and they are all linked together. meaning each of them is linked to the remaining 3
[16:00] <joebush> although I have no way of checking which actually are able to establish a connection to the others
[16:01] <joebush> But based on mbm's comment, I probably a loop (or several)
[16:01] <KeefR> probably yes ;-)
[16:01] <joebush> i.e I set it up thinking like a mesh
[16:01] <joebush> guess wds isn't quite a mesh :D
[16:02] <joebush> btw, doesn't anyone know if it is possible to check which ap's are connected in the wds?
[16:05] <wbx> joebush: check brctl show
[16:05] <[mbm]> joebush: you can always take KeefR's comment and turn on stp and hope for the best.. just that I wouldn't be the least bit surprised if that turns out to be problematic
[16:05] <wbx> joebush: and look at ifconfig -a
[16:06] <[mbm]> wbx: the wds interfaces represent wds connection attempts .. in other words I'll have a wds interface for every mac in wl0_wds regardless of if they're connected or not
[16:06] <KeefR> [mbm]: don't know how it works over wlan, but in a fully meshed environment stp isn't a good choice
[16:06] <KeefR> with 2 devices connected via lan and wlan it might work
[16:06] <[mbm]> KeefR: well, bridging is never really a good choice
[16:07] <KeefR> depending on the wlan part (never testet that stuff)
[16:08] <KeefR> we used to have atm in the backbone, never had a problem with the fully meshed design...with 10GE it's getting ugly now...
[16:09] <Ge0rG> hey KeefR :)
[16:10] <KeefR> Ge0rG: got a new flat, so finally I can use all the aps *G*
[16:10] <frop> someone know ImageBuilder?
[16:11] <frop> if i put an files/etc/readme.txt
[16:11] <frop> in squash, it goes to /rom/etc/readme.txt?
[16:11] Action: [mbm] probably should increase the stp forwarding delay .. pretty sure I left it at 0 (completely disabled)
[16:11] <KeefR> [mbm]: but that might be the best default config
[16:11] <srcquench> hei Ge0rG
[16:12] <joebush> mbm: so other than using one mac at a time to see if there is a connection, there is no possible way to know which AP's are connected in a wds?
[16:14] <[mbm]> frop: yes.
[16:14] <frop> [mbm] ok...and for packages that add users to the system? like quagga?
[16:15] <[mbm]> frop: have them use the 'nobody' account
[16:15] <frop> is "new" image with new user in /rom/etc/{passwd,group}
[16:15] <[mbm]> you can separate them if you want
[16:15] <joebush> Oh, yes. Using lazywds you can see which once actually connect :D
[16:16] <[mbm]> frop: but after you've removed root priveledge there's not much point in further separation
[16:16] <frop> mmmh
[16:16] <[mbm]> joebush: well, you can also get status with wl .. just depends on how you want to use the information
[16:16] <joebush> Interesting. I show the same AP connected twice with wds
[16:17] <frop> i'd like privilege separation possibilty...i'll try to build something with that
[16:17] <joebush> Just to check the links. It would be nice to be able to check signal strength too, but I doubt that's possible
[16:17] <[mbm]> joebush: easily possible .. ipkg install wl; wl sta_info <mac>
[16:18] <[mbm]> and wl rssi <mac>
[16:18] <joebush> wasn't aware of that. thanks
[16:18] <joebush> can it list all connected clients?
[16:18] <[mbm]> works for any connection, not just wds
[16:18] <[mbm]> wl assoclist
[16:19] <joebush> iwlist ethx ap doesn't work here. So I'll get wl and check it out
[16:19] <[mbm]> that's because the driver doesn't really speak the wireless-extensions language that the iw* utils use
[16:20] <[mbm]> so we've written a (buggy) translation layer to remap the requests
[16:20] <Ge0rG> srcquench: :)
[16:20] <[mbm]> if you notice any errors with iwconfig or any of the iw* utils file a bug report or ask nbd
[16:22] <joebush> are you sure wds ap's will show up in assoclist?
[16:22] <joebush> doesn't seem to
[16:22] <[mbm]> oh.. misunderstood your question
[16:22] <[mbm]> assoclist is clients
[16:23] <[mbm]> you want 'wl wds'
[16:23] <joebush> ahh
[16:23] <joebush> nice
[16:23] <[mbm]> unfortunately that will just be wl0_wds
[16:23] <joebush> No, it is listing what I am connected to since this AP is in lazy mode :D
[16:24] <joebush> connected to all 3
[16:24] <joebush> wl0_wds is empty
[16:24] <[mbm]> just remember you can't have them all as lazy and that while in lazywds mode all you have is your encryption security
[16:25] <joebush> hmm, or does it gather the macs of all the wds ap's on the net via the one it connects to??
[16:25] <joebush> Not worried. I live in the middle of nowhere
[16:25] <[mbm]> lazywds should be whatever aps attempted to do a wds connection to the device
[16:26] <[mbm]> (all connections are accepted, no connection attempts are made)
[16:27] <joebush> should all 3 references to wds in ifconfig be identical?
[16:27] <[mbm]> ?
[16:27] <[mbm]> you eman the interface names?
[16:27] <joebush> all are called wds0.4915 thought they whould eb different
[16:27] <joebush> yes
[16:27] <[mbm]> the name is actually longer, just that it gets cut off
[16:28] <joebush> ahh
[16:28] <[mbm]> (and technically it should be wds0.0, wds0.1 .. just that there's a bug somewhere in the binary only wireless driver)
[16:28] <joebush> as long as it works :\
[16:29] <[mbm]> I still find the bug to be annoying, I just haven't figured out any way that I can fix it without seeing the code
[16:30] <joebush> oh, binary only driver
[16:31] <[mbm]> brctl will show you the full name
[16:31] <joebush> how can I get the full (buggy) if name up?
[16:31] <joebush> hehe
[16:31] <joebush> jeez, 3 seconds
[16:31] <joebush> knew I'd seen them before. But couldn't remember where
[16:32] <joebush> so, they should be wds0.0.* wds0.1.* etc?
[16:33] <[mbm]> no, it should be wds0.0 wds0.1 wds0.2 .. instead it's wds0.49153 wds0.49154
[16:33] <joebush> oh
[16:34] <joebush> well, that wouldn't certainly be neater
[16:34] <[mbm]> in other words, the number didn't get initialized to 0
[16:34] <joebush> is it consistent on every boot, or random?
[16:34] <[mbm]> so it pulled in the value 49153 from some other area of mem
[16:34] <joebush> or incremental
[16:34] <[mbm]> always seems to start at 49153
[16:34] <joebush> substract 49153? :D
[16:35] <[mbm]> well give me the damned source and I'll fix it :P
[16:35] <joebush> hehe
[16:36] <joebush> is it possible to remap the interface name?
[16:36] <joebush> so it at least looks like it should
[16:36] <[mbm]> no point in adding more code to hide a bug in the existing code
[16:37] <joebush> to make it look pretty ;)
[16:37] <[mbm]> hmm damned bot
[16:37] <joebush> hehe
[16:37] Action: [mbm] fixes
[16:37] <DeltaF> haha
[16:37] <[olli]> oh, the bot is broken
[16:38] <[mbm]> yeah I broke it .. just a sec
[16:38] <[mbm]> fixed :P
[16:48] <[olli]> nbd, good work with the firewall webif page!
[16:49] <[mbm]> [olli]: better than nothing, still needs work
[16:50] <[olli]> true, but for basic things it works really good
[16:53] <joebush> yeah, that isn't bad
[16:56] <joebush> what's the difference between Destination IP and Forward to?
[16:56] <kellinaidan> joebush: context?
[16:56] <joebush> I am looking to forward to the destination IP
[16:56] <LostFrog> Destination IP is what matches, Forward to is what the dest IP is mangled to.
[16:56] <joebush> oh, in the firewall setting in webif. I'm testing it
[16:56] <joebush> oh, of course. thanks
[16:57] <kellinaidan> joebush: LF is right.
[16:58] <joebush> is it possible to enter a port range?
[16:59] <[mbm]> yeah although I forgot if the web interface used 123-456 or 123:456
[16:59] <joebush> must be :
[16:59] <joebush> I tried -)
[16:59] <joebush> or not
[17:00] <[mbm]> I know on the commandline it's -m multiport --dports 123:456
[17:00] <joebush> must be the control script that doesn't allow it
[17:01] <Pooh22> shouldn't the webif for the firewall show some schematic of the interfaces and their ip-nrs?
[17:01] <[mbm]> Pooh22: that's under the status page
[17:02] <Pooh22> well, by default (no fw) nothing shows, I guess
[17:03] <Pooh22> Having no firewall active is also something that could be shown on the status page
[17:03] <[mbm]> huh? you misunderstood
[17:04] <[mbm]> there's always a firewall enabled
[17:04] <[mbm]> the webpage just lets you override it
[17:04] <Pooh22> [mbm]: of course, my mistake
[17:05] <Pooh22> so the stuff shown on the status page could be extended
[17:05] <[mbm]> name one thing in webif that couldn't be extended :P
[17:06] <Pooh22> to include the available subnets per interface and forwarding rules
[17:06] <Pooh22> of course, I understand that it doesn't get there by itself ;-)
[17:06] <joebush> why don't I have to enter a password?
[17:06] <[mbm]> </captain obvious>
[17:06] <joebush> :O
[17:06] <[mbm]> joebush: you will
[17:07] <joebush> when?
[17:07] <[mbm]> joebush: you're allowed to visit the about and status page without a password, after that it'll prompt you
[17:07] <[mbm]> (your browser may have already cached your password_
[17:08] <joebush> lots a long caching time. Was in yesterday afternoon hehe
[17:08] <joebush> No logout?
[17:09] <LostFrog> Your box recognized you. :) It's the new facial recognition software.
[17:09] <joebush> hehe
[17:09] <[mbm]> joebush: blame your browser
[17:09] <joebush> that must be it
[17:09] <joebush> I am. It's just ignoring me though
[17:09] <[mbm]> threaten to install IE
[17:10] <joebush> :D
[17:10] <joebush> found it
[17:11] <joebush> It's in the meny in Firefox apparently
[17:11] <joebush> menu
[17:11] <joebush> was looking in options
[17:14] <joebush> ipkg search iwconfig
[17:14] <joebush> got nothing in the shell. Hoping I have better luck in this window
[17:14] <Pooh22> joebush: you have to run ipkg update before finding it
[17:14] <joebush> it has been run many times
[17:14] <[mbm]> iwconfig is already included :P
[17:14] <Pooh22> hmm
[17:15] <Pooh22> is that new in rc5?
[17:15] <joebush> yes
[17:15] <[mbm]> Pooh22: you mean iwconfig? no, it's been included in the past several revisions
[17:15] <joebush> I'm running micro right now
[17:16] <Pooh22> I'm probably having a serious monday condition ;-)
[17:16] <[mbm]> joebush: oh .. ipkg install wireless-tools
[17:16] <joebush> tx
[17:16] <Pooh22> is it possible to post several dns servers in the dhcp-options of dnsmasq?
[17:17] <[mbm]> yes, exactly as shown in the dnsmasq.conf example
[17:18] <Pooh22> weird, I have that and I only get the first one
[17:18] <joebush> reflashing. back in a bit
[17:18] <Pooh22> I'll refresh the if on the client side
[17:19] <[mbm]> Pooh22: did you remember to restart dnsmasq after editing the configs?
[17:19] <Pooh22> [mbm]: yes
[17:19] <Pooh22> running dhclient on the interface doesn't update my resolv.conf on the client side
[17:20] <Pooh22> it looks like: dhcp-option=lan,6,10.0.0.254,217.114.97.2,217.114.97.3
[17:20] <Pooh22> I only get 10.0.0.254
[17:20] <[mbm]> why dhcp-option=lan ?
[17:20] <Pooh22> on the client (running mandriva 2006)
[17:21] <Pooh22> I also have wifi,.....
[17:21] <Pooh22> my wifi is eth1 and lan is vlan0
[17:22] <Pooh22> I'd like to add the ipv6 address of the wrt in the list as well, but not if it will be the only address that comes through
[17:24] <Pooh22> my wlan if line looks like dhcp-option=wifi,6,192.168.2.1,217.114.97.2,217.114.97.3
[17:26] <joebush> btw, rc5 micro trx 4mb has at least an incomplete ipkg
[17:27] <joebush> many of the options were missing
[17:27] <[mbm]> yeah, it uses ipkg-sh
[17:27] <joebush> should I report it, or has it been recompiled since last night?
[17:27] <[mbm]> not ipkg
[17:27] <joebush> oh
[17:27] <[mbm]> it's not a bug, you selected 'micro'
[17:27] <joebush> didn't realize.
[17:27] <[mbm]> so it's 'micro' :P
[17:28] <joebush> list_installed isn't even there?
[17:28] <joebush> you should call it nano :p
[17:28] Action: [mbm] can't remember .. haven't used ipkg-sh in ages
[17:28] <[mbm]> it's a shell script, you can try to read it
[17:29] <joebush> that was my first time trying it
[17:29] <joebush> ipkg was pissing me off, but it was ok after I 'reinstalled' it hehe
[17:32] <joebush> oh, have to change resolv.conf again.
[17:32] <joebush> I wish other people had this problem
[17:33] <joebush> So everyone doesn't think I'm making it up to get attention
[17:33] <joebush> :D
[17:39] <Pooh22> [mbm]: my n770 does have all three ips in the resolv.conf, so it's not the dnsmasq config on the wrt
[17:43] <[olli]> yeah, rc5 runs great so far (router part) .. don't know the wireless stuff
[17:53] <grandrew> :-)
[17:56] <grandrew> I've got a asus wl300g (doesn't have a WAN) with WDS to another asus. I've configured it to have lan and wifi splitted (to be a router). WDS stopped to work :( Are there any special issues with WDS and bridge split?
[17:58] <grandrew> Everything was perfect until I did wan_ifname=eth0 and lan_ifnames=eth2 (I also did wifi_ifname=eth2 and set the ipaddr, for sure)
[17:59] <wbx> grandrew: how you installed openwrt? which version you use?
[17:59] <joebush> is it eth0 and eth2? My HDD uses eth1 for lan port and eth2 for wlan
[18:00] <joebush> you still need to have the wireless (eth2) on the bridge (br0)
[18:00] <joebush> the wds interfaces will join it there
[18:03] <grandrew> hmm I see the ping replys from my second asus in the old subnet attached to asus-router lan port. It means that wds link is still at bridge :( this is exactly not what i want
[18:04] <grandrew> is there any way to have a routed access to the wds link?
[18:04] <joebush> I'm not sure it is very clear exactly what your goal is
[18:05] <joebush> about you state you are using wds, but you accessed the other asus through the lan port
[18:06] <grandrew> I had comp - asus1 ~~WDS~~ asus2 ~ comp2(all in subnet1). I want comp --(subnet1)- asus1 ~(WDS,subnet2)~ asus2 ~(subnet2)~ comp2
[18:06] <grandrew> -- means wire link; ~~ means wireless :-)
[18:07] <joebush> the computer is attached to the port?
[18:08] <grandrew> yes. comp1 is on lan of asus1
[18:09] <joebush> you say that wds stopped working?
[18:09] <joebush> is there no communication between the ap's?
[18:09] <grandrew> it hasn't really. just it does not work the way I want it
[18:10] <joebush> what 'is' happening?
[18:10] <grandrew> i can not ping asus2 form asus1 but can form comp1 if it is in the old subnet
[18:10] <grandrew> and i cant ping asus2 if comp1 is in the new subnet
[18:11] <grandrew> and it should be reversed: comp1 should ping asus2 only using asus1 as the router
[18:12] <grandrew> and it just pings it as if it was in the 'lan' subnet
[18:13] <joebush> are you sure the ip addresses are correct for your configuration? Remember that a wds configuration isn't layer 3 (someone stop me if this is incorrect). IP addresses aren't necessary for communication between the ap's
[18:13] <joebush> so it is possible you are able to communicate between nodes, but not access the ap's if the addresses aren't properly configured
[18:14] <grandrew> I guess what i've got means that wds0 is bridged to br0 instead of eth2 (i want eth2(wifi) and wds0 with asus2 and all the wifi stuff on the subnet2)
[18:14] <joebush> wds 'must' be part of the bridge
[18:15] <grandrew> hm
[18:15] <joebush> eth2 should be as well
[18:15] <joebush> wds becomes part of thebridge automagically
[18:15] <grandrew> eth2 is now definitely not :-)
[18:15] <wbx> root@delaware:~# brctl show
[18:15] <wbx> bridge name bridge id STP enabled interfaces
[18:15] <wbx> br0 8000.0014bf1bada3 no eth1
[18:15] <wbx> wds0.49153
[18:15] <wbx> root@delaware:~# nvram get wifi_ifnames
[18:15] <wbx> eth1
[18:15] <wbx> root@delaware:~# nvram get wifi_ifname
[18:15] <wbx> br0
[18:15] <joebush> use pastebin.com
[18:15] <wbx> just split the bridge and make a new one with only the wifi device
[18:15] <[olli]> nbd, ping
[18:16] <wbx> joebush: to late. ;=)
[18:16] <joebush> oh, thought it was grandrew copying in his config :D
[18:16] <grandrew> br0 8000.0013d459fd91 yes eth0
[18:16] <grandrew> wds0.49153
[18:16] <grandrew> =)
[18:17] <grandrew> so how can I make a bridge with eth2 and wds0?
[18:17] <joebush> gransrew, from what I can tell it may just be the way you want
[18:17] <joebush> wds0 automatically becomes part of the bridge
[18:17] <joebush> just make sure eth2 is in the bridge
[18:18] <grandrew> i want eth2(the wifi) and wds0 to be bridged in the subnet2. I want eth0 to be subnet1
[18:18] <grandrew> hm
[18:18] <grandrew> how do I do that?
[18:18] <joebush> right, wan_ifname=eth0
[18:18] <joebush> lan_ifnames=eth2
[18:18] <joebush> lan_ifname=br0
[18:18] <grandrew> right =) thanks!
[18:18] <joebush> it is probably already corect
[18:19] <grandrew> wan is eth2 at my conf
[18:19] <joebush> Oh!!
[18:19] <joebush> hehe
[18:20] <joebush> but remember, you will automatically behind behind the firelass on asus1
[18:20] <joebush> and all clients that connect to either asus1 or asus2 will be outside
[18:21] <joebush> Anyone, is it possible to have the wan_ifname as part of br0?
[18:22] <joebush> hmm, wan_ifname=br0 maybe :o
[18:22] <joebush> Grandrew, you can try the above setting if you would rather secure the wireless side
[18:23] <joebush> and lan_ifname=eth0
[18:23] <joebush> depends what you are looking to do
[18:23] <joebush> good luck
[18:24] <wbx> samot: hi
[18:24] <wbx> samot: can you give me a short pmon crash course?
[18:25] <grandrew> okay, I've understood my problem. I wanted to have lan port as LAN while it is only possible to have it 'WAN'
[18:25] <joebush> I don't want to say it isn't possible
[18:25] <joebush> Just not how it was configured apparently
[18:25] <EnolaGay> is someone with a big endian system can mount a cramfs for me and send me the content in a tar.gz ?? contact me thx.
[18:25] <joebush> I have had my wl0 as WAN before
[18:26] <grandrew> hmm. wan_ifname=br0 will not help as wan_ifnames= is just ignored (as was in RC3)
[18:26] <grandrew> or it wasn't?
[18:26] <joebush> I used wan_ifname=eth2, but since wds joins the bridge with eth2 I'm unsure
[18:27] <wbx> make new devices and change /etc/init.d/S40network
[18:27] <wbx> grandrew: like wds_ifnames
[18:27] <wbx> or wifi_ifnames
[18:27] <wbx> grandrew: you need to set prefix_ifname, prefix_ifnames, prefix_ipaddr, prefix_proto and prefix_netmask
[18:28] <grandrew> okay, I'll just swap wan and lan and rewrite firewall rules.
[18:28] <grandrew> it's easier than rewrite all the init scripts :-)
[18:28] <[mbm]> grandrew: why change the firewall? jsut change the variables
[18:28] <grandrew> hm
[18:30] <wbx> [mbm]: do you know how to reflash via pmon on a v1 device?
[18:31] <[mbm]> the command? no I don't remember any pmon commands
[18:32] <[olli]> hmm. forward:proto=tcp dport=1234:192.168.1.2:22 works but (nmap show open) but forward:proto=tcp dport=12345:192.168.1.2:80 is still closed
[18:32] <samot_> wbx: hi. wap crash course: let's begin
[18:33] <grandrew> so can I have WDS separated from both wifi sides (no bridge at all, just another P2P link)?
[18:33] <grandrew> that would be even better :-)
[18:34] <[mbm]> grandrew: the way it's setup it'll always try to add it to br0, so just reserve the br0 for whatever you want as a wds network .. use br1 if you need another bridge for lan/wifi
[18:35] <Hybernaut> yay for RC5
[18:35] <Hybernaut> thanks to everyone
[18:35] <grandrew> [mbm], thx
[18:37] <wbx> samot: ah, cool
[18:37] <joebush> can wds be bridged separately from wl0??
[18:37] Action: [mbm] must have missed all of samot's comments
[18:37] <[mbm]> joebush: wl0? get your terms right
[18:37] <samot_> wbx: sorry i have some problems with inet connection
[18:37] <joebush> eth2 :p
[18:38] <joebush> I know what I mean
[18:38] <samot_> wbx: have you seen http://wiki.openwrt.org/OpenWrtDocs/Hardware/Linksys/WAP54Gv10
[18:38] <[mbm]> joebush: you can have the wds in br0 and eth2 in br1 .. doesn't matter
[18:39] <joebush> I'm just curious what happens to the traffic since wds goes over eth2
[18:39] <grandrew> does wds use wifi traffic?
[18:39] <grandrew> =)
[18:39] <joebush> that's why I assumed they had to be in the same bridge
[18:39] <wbx> samot_: i get the power cable and read the link, after that i ask if questions remain
[18:39] <Kaloz> [18:38:17] < samot_> wbx: have you seen http://wiki.openwrt.org/OpenWrtDocs/Hardware/Linksys/WAP54Gv10
[18:39] <[mbm]> grandrew: wds is just a point to point link
[18:40] Action: Kaloz is tempted to kill everyone who adds a new page for each hardware revision
[18:40] <[mbm]> Kaloz: so merge the page with the original and then kill samot :P
[18:40] <samot_> kaloz: dont kill me. I belive the reason I did that is explained
[18:40] Action: Stereo does the happy rc5 flashing dance
[18:40] <samot_> kaloz: in WAP54G page
[18:41] <[olli]> ah, busybox isn't listening on the wan port
[18:41] <[olli]> busybox httpd*
[18:41] <[mbm]> samot_: so how badly does rc5 break on the wap?
[18:41] <grandrew> right. but (if you know) does WDS share the same CTS as normal wifi links? I mean, will it double-dropdown the speed if used?
[18:41] <samot_> mbm: dont know
[18:42] <Stereo> samot_: it could be a section on the wap54g page
[18:42] <[mbm]> grandrew: hmm good question
[18:42] <samot_> mbm: the version i run is cool and i dont have time break it again
[18:44] <grandrew> I'll test this :-) Anyways, Kismet shows WDS link as 'noise'
[18:45] <[mbm]> think that qualifies as the most pointless use of /nick
[18:47] <Stereo> [mbm]: no, that'd have to be all the people who /nick foo-afk or foo-zzz
[18:47] <grandrew> but WDS uses the same transmiter as WIFI I guess -> speed dropdown is surely provided
[18:48] <Stereo> thank goodness irssi lets you /ignore *away* NICKS
[18:49] <[mbm]> Stereo: well, then there's those people that are in 30 channels and change their nick for a quick joke in one of the channels
[18:49] <Eagle_Fire> and then nbd kicks them from this one
[18:49] <[mbm]> * foo is now known as it_wasnt_me; * it_wasnt_me is now known as foo
[18:49] <Stereo> while having a serious conversation somewhere else
[18:50] <[mbm]> I hate irc spam from people that aren't active in the channel
[18:51] <Eagle_Fire> well at least you have ban power
[18:51] <[mbm]> If you've been talking and suddenly have to go I don't mind an away notice .. I just get annoyed when out of the blue there's an away message from someone I didn't even know was here to begin with
[18:59] <Stereo> there's always /ignore * ACTIONS -regexp -pattern is (away|gone|back)
[19:00] <MioZ> I want to access to my openwrt whiterussian firmare by ssh from the wan. What iptables rules must I add ? Thanks a lot for your help
[19:00] <grandrew> hm. Can I do 'WDSap~~WDSap~~WDSap~~WDSap' ?
[19:00] <Eagle_Fire> grandrew, yes
[19:00] <[mbm]> MioZ: see /etc/firewall.user
[19:01] <grandrew> did anybody test how many 'uplinks' can broadcom bridge iface handle then?
[19:01] <[mbm]> "broadcom bridge interface" ? don't know why you're calling it broadcom's
[19:02] <grandrew> br0 isn't just broadcom cpu(or whatever chipset) feature?
[19:02] <[mbm]> br0 = bridge 0
[19:02] <[mbm]> it's software.
[19:02] <[mbm]> standard linux component
[19:03] <grandrew> wow
[19:03] <MioZ> mbm : thx ;)
[19:04] <grandrew> so if I use, for example, 4 WDS links in 4 bridges, how will it eat mac addresses and all that stuff? It is the work of some switching mechanism to choose whether to send the packet further or to broadcast it 'here' i guess
[19:06] <grandrew> or is it just broadcasted through the whole set of bridges?
[19:06] <grandrew> maybe that's a question to the linux bridge driver developers =)
[19:07] Action: grandrew went reading the sources
[19:07] <[mbm]> grandrew: the bridge itself acts as an ethernet switch between interfaces, unicast messages are sent to the correct itnerface, broadcast messages are spammed
[19:08] <grandrew> cool 8-) thanks
[19:30] <dacb> Anyone here had luck putting a serial console on an ASUS WL-530G?
[19:44] <samot> mbm: wap 1.0 + rc5 = a good brick
[19:44] <[mbm]> samot: not surprised, I'm just wondering where
[19:45] <samot> mbm: this looks bad diag boardtype: 0000000
[19:45] <[mbm]> hmm
[19:47] <[mbm]> samot: that shouldn't do anything other than ignore the leds and reset button
[19:48] <samot> mbm: probably unconected phy workaround is not activated and eth do not work
[19:49] <md```> anyone knows how i can find out more about the hardware specifics of my netgear he 102? it runs vxworks and i'm connected through a serial console...
[19:49] <[mbm]> samot: when you say it's bricked, do you mean you can't even get to the shell via serial?
[19:50] <[mbm]> md```: there's usually a help function, although I can't think of any functions that would tell you much about the hardware
[19:51] <spectra-> talking about bricks... i've got two wrt54g v2 units here, both with constantly blinking power led and randomly (at intervals between 1 and 10 seconds) lighting _all_ switch leds (though only one port is actually connected). both don't boot up at all, of course. nope haven't tried serial yet.
[19:51] <samot> mbm: serial works. I just investigating it further
[19:52] <[mbm]> samot: there was a switch from et.o to b44 for the ethernet, but I thought we already had the workarounds in b44
[19:52] <md```> [mbm]: yeah i found the help function already :)
[19:52] <md```> but that only gives me stuff like which hardware and software revision it is
[19:53] <[mbm]> spectra-: all of the switch leds lighting means that the switch chip is being reset .. happens on powerup or when the ethernet driver loads .. the dmz led will be the more interesting led to look at
[19:53] <samot> mbm: eureka. just sill \r in default nvram. if i reset boardtype and boardnum, it shows diag boardtype: 0000041a and eth works
[19:53] <[olli]> md```, well. the vxworks os sucks. take the router back
[19:53] <samot> s/sill/silly/
[19:53] <spectra-> [mbm]: dmz never lights up at all
[19:53] <md```> [olli]: heh, i got it for free
[19:54] <md```> and the router doesnt seem so bad so far, only it runs vxworks :/
[19:54] <[mbm]> samot: so the new firstboot/jffs2 stuff works?
[19:54] <Stepan> hm.
[19:54] <Stepan> http://www.netgear.co.kr/Support/Product/FileInfo.asp?IDXNo=155 the zip file in the bottom contains a thing called "telnetenable.exe"
[19:54] <Stepan> i wonder whether that works on all vxworks based routers
[19:54] <spectra-> [mbm]: so something is causing the device to constantly reboot, that's what i had figured before :) any quick suggestions what i could do _without_ having to pack out my soldering iron? :)
[19:55] <samot> let me see
[19:55] <J4k3> I managed to finally get the dwl-2100ap's to stop crashing once a week
[19:55] <[mbm]> spectra-: if it was flashed with openwrt and you aren't seeing the dmz led then something is crashing openwrt in the kenrel bootup
[19:55] <J4k3> (another shitty vxworks device)
[19:55] <J4k3> all the default settings were crapl
[19:55] <spectra-> kind of remarkable that both are v2 units... maybe there was a bug in the ethernet driver or something, only concerning those things
[19:55] <J4k3> damned AP default configured to blast CTS packets.
[19:56] <[mbm]> J4k3: oems probably sabotoge the devices to sell more
[19:56] <spectra-> [mbm]: yeah, openwrt was on them before they started acting weird and eventually giving up
[19:56] <J4k3> [mbm]: no suprise... that and dlink is a 'lowest bidder' operation it appears.
[19:56] <spectra-> actually it's still on 'em, i guess ;)
[19:56] <[mbm]> spectra-: hmm think to figure out what's happening you'll need serial
[19:57] <[olli]> J4k3, bricked a d-link today by hitting the reset button (with default firmware on it)
[19:57] <spectra-> mmmm kay.. i was anyways thinking about putting lumberg connectors in my wrts, so i can simply connect them with siemens mobile phone cables
[19:58] <spectra-> guess that would be the cheapest solution... maybe i should write a wiki page about that
[19:58] <[mbm]> [olli]: that's got to be some great QA if you can take a device out of the box and brick it with only the reset button
[19:58] <spectra-> anyone gets an old siemens mobile phone to hack those connectors out :)
[19:58] <[olli]> yeah
[20:00] <samot> mbm: jffs seems to be ok if it has enough space to be writeable. Did you debug firstboot to cope with jffs failure correctly (sth like i sent you)?
[20:02] <[mbm]> samot: yeah there were a few patches
[20:03] <samot> mbm: i try to flash normal image
[20:03] <md```> Stepan: what advantages does telnet have over serial interface?
[20:06] <spectra-> [mbm]: another thing... i tried to install openwrt on two wrt54gs (running hyperwrt) yesterday. both don't respond anymore now, although led status seems okay (dmz lights up showing boot process, dmz goes off and everything looks fine..).
[20:06] <[mbm]> spectra-: use failsafe
[20:07] <spectra-> couldn't do any further research on that yet, anyways thats strange..
[20:08] <spectra-> yeah, still gotta do that ;)
[20:08] <spectra-> it sits on the roof.. i love it when those things give up ~:~
[20:08] <spectra-> err, :~
[20:09] <[mbm]> well, we decided that a network accessable failsafe mode would be a bad idea so failsafe requires physical access to the device
[20:12] <spectra-> i sure do agree :) anyhow it's always a bit annoying having to climb up there, hrmz
[20:12] <spectra-> at least it's on 100ft tower or something :)
[20:13] <spectra-> at least it's _NOT_ on a 100ft tower..
[20:13] <spectra-> bah, today sucks ;)
[20:13] <Ge0rG> it sucks really really bad
[20:19] <[mbm]> Ge0rG: the day after a release announcement always sucks
[20:19] <Ge0rG> [mbm]: if it only was the only reason
[20:45] <Jonek01> Cool - RC5 :-)
[20:53] <estel> hi; i've tried to update all packages on my wrt54gs (firmware: whiterussian), but ipkg refuses to install /bin/ip-calc, because it is already provided by busybox
[20:53] <estel> any ideas?
[20:54] <[mbm]> don't try to use ipkg upgrade to install rc5
[20:54] <estel> i've just wanted to get more up2date
[20:54] <estel> any ideas for fixes?
[20:55] <[mbm]> yes. download rc5 and install it properly.
[20:56] <estel> but that means that all my configuration will go down the drain, right?
[20:56] <[mbm]> nvram will be preserved
[20:57] <estel> nvram, but all the other config-files will be overwritten
[20:57] <[mbm]> right.
[20:57] <Eagle_Fire> however, you can save them if you wish
[20:58] <[mbm]> well, don't do something stupid like install rc5 and then revert it back to rc4 by restoring individual files
[20:58] <Eagle_Fire> of course, no one would be that stupid
[20:59] <estel> i don't understand what's the problem between rc4 and rc5
[20:59] <[mbm]> ...
[20:59] <Eagle_Fire> it's a new version
[20:59] <estel> why is an upgrade from rc4 to rc5 a problem?
[20:59] <Eagle_Fire> it's only a problem if you make it a problem
[20:59] <[mbm]> estel: we've never supported ipkg upgrade .. go read that 'common mistakes' thread in the topic
[20:59] <Eagle_Fire> just remember reflashing is the only way to upgrade your kernel
[21:00] <friiman> [mbm] are you there?
[21:00] <[mbm]> no
[21:00] <[mbm]> go away
[21:00] <[olli]> oO
[21:00] <estel> but just pakages with kernel modules depend on a certain kernel version ;-)
[21:00] <friiman> :'(
[21:00] <Eagle_Fire> if your configuration works fine as it is you don't have to upgrade. upgrades are only when you want to do new things
[21:01] <[mbm]> openwrt gets installed via a reflash, not via ipkg.
[21:01] <Eagle_Fire> estel: it appears you have correct information about openwrt. we don't have to argue semantics - you know what will break and ostensibly how to fix it
[21:01] <morcego> After running ipkg upgrade today (running RC4), I started getting the following messages. Any idas ? -> jffs2_reserve_space_gc of 196 bytes for garbage_collect_dnode failed: -4 || Waiting for chip to read, status = 7
[21:02] <[mbm]> *smack*
[21:02] <friiman> anyone tell me, what mean "child mac adress" and "parent mac adress" ? i want put other ap no-suported by openWRT but, configured to work on repeater mode.
[21:02] <[olli]> ohoh
[21:02] <macsat> youre on smack MBM ?
[21:02] <Eagle_Fire> friiman, as far as WDS goes that doesn't make sense
[21:03] <Eagle_Fire> wds is a peer protocol
[21:03] <Eagle_Fire> if they have some protocol that has parent and child then it is probably not compatible with wds
[21:03] <friiman> Eagle_Fire hummm. so, i need to put openWRT aP like AP an the other, repeater, right?
[21:03] <Eagle_Fire> that would help, yes
[21:03] <friiman> hum. ok
[21:03] <friiman> let me try
[21:03] <Eagle_Fire> hot damn, the rc5 changelog is large
[21:04] <[mbm]> only about 1000 entries
[21:04] <Eagle_Fire> only, sure
[21:04] <friiman> nice work for openWRT team.
[21:05] Action: grandrew still thinks that even RC3 was already perfect :-)
[21:06] <[mbm]> then rc4 broke it and now rc5 fixed it again .. it's a vicious cycle
[21:06] Action: J4k3 still thinks openwrt is getting slow-walked.
[21:06] <friiman> Eagle_Fire i use one ap with WDS, is necessary disable "wl0_lazywds" and "wl0_wds" to work on AP mode?
[21:07] <[mbm]> J4k3: if you didn't complain 24x7 about everything I might actually listen when you complain
[21:07] <blop> is there any docs about the /proc/switch yet?
[21:08] <Eagle_Fire> yeah, remove lazywds and wds to work in no-wds mode
[21:08] <[olli]> blop, i don't think so. the docs say fixme
[21:09] <[mbm]> blop: not really .. uses the same syntax as the vlan*ports variables .. you can see an example in the /etc/preinit script
[21:09] <friiman> Eagle_Fire "nvram unset wl0_lazywds & nvram unset wl0_wds" ?
[21:09] <Eagle_Fire> not sure if it's single ampersand but otherwise looks fine
[21:10] <blop> ok thx :)
[21:10] <[mbm]> single & is for unning commands in the background, && is a logical AND
[21:10] <[mbm]> (so that should be a &&)
[21:11] <Eagle_Fire> super
[21:13] <wbx> gorec: depends.
[21:13] <friiman> is necessary after nvram commit reboot the ap?
[21:13] <[mbm]> gorec: you're SOL if you expect wifi
[21:13] <Eagle_Fire> friiman, not necessary
[21:13] <friiman> hum ok
[21:13] <Eagle_Fire> but it usually helps when you want to apply the settings
[21:13] <Eagle_Fire> to apply the settings for wifi use "/sbin/wifi" command
[21:14] <[mbm]> friiman: a reboot is never required, and 'nvram commit' is for copying the settings to flash (running config to stored config)
[21:14] <gorec> i'm does'nt expert, but i try to build openwrt+madwifi under mips4600 (SoC Atheros)
[21:14] <[mbm]> gorec: madwifi doesn't support the ar531x
[21:14] <friiman> oh tks both
[21:14] <friiman> Eagle_Fire or reboot , right?
[21:15] <Eagle_Fire> reboot will apply the settings, yes
[21:15] <[mbm]> damned windows users ...
[21:15] <friiman> ok
[21:15] <gorec> really?!?!? - but in source many strings like "AR531x"...
[21:15] <[mbm]> gorec: only works if you can replace the hal binary .. the one used by madwifi won't work
[21:16] <gorec> where i can get that hal?
[21:16] <[mbm]> gorec: the simple answer is you can't.. go look at the madwifi site .. there's tons of similar complaints
[21:17] <gorec> anybody do this work succes before?
[21:17] <[mbm]> hell, atheros is even giving contracted oems a hard time
[21:17] <friiman> Eagle_Fire can i put some 10 lines in your pvt?
[21:17] <[mbm]> gorec: you're not listening
[21:17] <gorec> sorry - it's my emoution :-((
[21:17] <Eagle_Fire> friiman, sure, but it's better to use pastebin
[21:17] <Eagle_Fire> that way everyone can see
[21:18] <friiman> yep
[21:18] <friiman> http://pastebin.com/625533
[21:18] <wbx> but there are leaked hals ;)
[21:18] <friiman> right config for ap mode?
[21:19] <Eagle_Fire> friiman, sounds right - does it work?
[21:19] <friiman> root@OpenWrt:~# /sbin/wifi
[21:19] <friiman> Usage: wlconf <ifname> up|down
[21:20] <friiman> :/
[21:20] <Eagle_Fire> "wifi up" then
[21:20] <gorec> [mbm]: thank you for you help
[21:21] <svu> is there a way to update dyndns from behind NAT?
[21:21] <friiman> Eagle_Fire now i want config other ap (non-supported by openWRT) to repeat the traffic
[21:21] <svu> using openwrt
[21:21] <Eagle_Fire> svu, yes
[21:21] <friiman> he ask me child and parent mac address
[21:21] <friiman> what i put on that?
[21:21] <svu> Eagle_Fire, how? ez-ipupdate does not deal with nat nicely
[21:21] <Eagle_Fire> friiman, try having the parent be the openwrt ap and the child be the non-openwrt
[21:21] <frop> fucking...i'm the only one that use a DNS not supported by dyndns
[21:21] <frop> :P
[21:21] <frop> i've to "make" my client with nc
[21:22] <Eagle_Fire> tee hee
[21:22] <Eagle_Fire> svu, can you tell ez-ipupdate to ask the internet what its ip is?
[21:22] <fishhead> anyone know any known issues with the befw11s4
[21:22] <friiman> Eagle_Fire how i see mac add from ap supported ?
[21:22] <fishhead> yeah I know it doesn't run openwrt but you guys know the ins and outs better than most
[21:22] <svu> Eagle_Fire, AFAIK the answer is no. Everyone is recommending ddclient for such situations - but there is no ddclient built for openwrt
[21:22] <Eagle_Fire> friiman, ifconfig eth1
[21:23] <fishhead> or an easy way to find out if 2 wireless networks in an area are butting heads?
[21:23] <Eagle_Fire> fishhead, "butting heads" ?
[21:23] <fishhead> like could know knock the other one out
[21:23] <fishhead> RF wise
[21:23] <Eagle_Fire> svu, two options then - build ddclient for openwrt, or add support in ez-ipupdate
[21:24] <blop> =)
[21:24] <Eagle_Fire> fishhead, wi-viz might help you in that respect
[21:24] <svu> Eagle_Fire, thaaaanks ;)
[21:24] <blop> or run ddclient on another box behind the wrt
[21:24] <Eagle_Fire> svu: you're welcomd
[21:24] <fishhead> thx sir
[21:24] <svu> blop, that's what I am doing now - I just thought about moving the service to wrt...
[21:24] <blop> =)
[21:25] <fishhead> oh won't help me
[21:25] <fishhead> requires openwrt :/
[21:25] <blop> shouldnt be too complicated to get a static version of ddclient running on wrt :)
[21:25] <Eagle_Fire> people in this channel generally run openwrt
[21:25] <svu> another, even more stupid question - could I have WINS server on wrt?
[21:25] <fishhead> yes I know :)
[21:25] <fishhead> waiting for a ppc port of openwrt myself :P
[21:25] <Eagle_Fire> svu: can't rule it out
[21:25] <svu> Eagle_Fire, :)) you are being really nice:)
[21:25] <Eagle_Fire> wi-viz doesn't support ppc
[21:25] <blop> svu : samba does that , isnt there a pkg yet?
[21:26] <fishhead> wait
[21:26] <fishhead> there is a powerpc port of openwrt?
[21:26] <svu> blop, there is no samba on my list of available packages...
[21:26] <[olli]> svu, look at the backports repository
[21:26] <svu> [olli], ok, I'll check, thanks
[21:27] <[olli]> svu, http://downloads.openwrt.org/backports/rc5/
[21:28] <spectra-> [mbm]: one of the v2 just booted again (don't ask me why, i just left it sitting alone without power a while, then it worked again..). i flashed rc5 using mtd. now it again keeps resetting the switch, although the power and wifi (not dmz) are continously lighted
[21:28] <xFallenAngel> yay im getting another asus
[21:28] <Eagle_Fire> yAy!
[21:28] <jimp> wbx: hi
[21:28] <svu> [olli], ghm... 371K ... not enough space on my wrt I'm afraid...
[21:29] <[mbm]> spectra-: it's normal to see the switch leds flash with activity, it's only when all of them (including ones not connected) light up that you know it's a reset
[21:29] <spectra-> hrmz, if course that's what i mean
[21:29] Action: svu does not want full samba, only wins... but probably there is no such a thing...
[21:29] <blop> :)
[21:30] <wbx> jimp: hi
[21:30] <jimp> wbx: just saw your message about the belkin patches
[21:31] <spectra-> [mbm]: do you know of any hardware faults that could occur (especially on those v2 things)?
[21:31] <[mbm]> spectra-: only if nvram got reset
[21:31] <xFallenAngel> svu: i dont know if it was suggestged, but isnt wins only one of the daemons? cant you just install all but smbd? (i think nmbd was wins)
[21:32] <spectra-> hmm, i thought wrts had a bootloader which rewrites missing nvram variables, so mtd erase nvram would do no harm?
[21:32] <roh> heh.. a lot of people use still samba 1.9 or so because its _much_ less binarysize
[21:32] <[mbm]> spectra-: it writes some really horrible defaults
[21:32] <svu> xFallenAngel, yes, probably just wins should be ok... if there would be such package...
[21:33] <spectra-> hmmm this gives me an idea what could have gone wrong :)
[21:33] <svu> just nmbd
[21:33] <jimp> wbx: the router I was working with is now installed and being used, so I'll have to add a serial port to another one
[21:33] <fish_> ah hi roh
[21:33] <jimp> before I can try the new patches
[21:34] <xFallenAngel> svu: download the ipkg, its just a tar.gz with a different extension and some packaging, then you can scp it over
[21:34] <[mbm]> spectra-: loading either a recent linksys firmware or a recent openwrt will try to correct the nvram settings .. assuming it stays stable long enough for that to happen
[21:34] <jimp> I've also been meaning to try brcm-2.6, since the 2.4 kernel won't really ever support the flash correctly
[21:34] <spectra-> [mbm]: thanks, good to know. too bad it doesn't seem like it'd run stable.. i'll check out that linksys tool for flashing
[21:35] <wbx> jimp: why that?
[21:35] <svu> xFallenAngel, ok, I'll try. Thanks.
[21:35] <roh> wrt recent samba-builds are too fat for a lot of embedded-apps. so a lot of appliances still use samba 1.9 (which has less features, is buggy and propably exploitable) but can be compiled to binarys which still fit into the images
[21:35] <[mbm]> spectra-: it's a problem with the sdram refresh settings
[21:35] <jimp> wbx: it's a DataFlash part, the support for that wasn't ever added to mtd before they stopped developing the 2.4 branch
[21:35] <[mbm]> spectra-: so you're getting random corruption
[21:36] <jimp> wbx: right now it sort of works by pretending it's a NOR flash with a tiny erase size,
[21:36] <wbx> jimp: but i thought the serial flash driver works with the device?
[21:36] <jimp> wbx: which is fine for most cases, but jffs2 won't work on that.
[21:36] <jimp> wbx: squashfs works ok, which is what i'm using
[21:36] <spectra-> [mbm]: sounds plausible :/
[21:37] <jimp> wbx: but it means I have no writable partition
[21:37] <wbx> jimp: ah, okay. and backporting mtd changes from 2.6 is not an option?
[21:37] <jimp> wbx: might be, I haven't looked into it much
[21:39] <jimp> wbx: any particular suggestions about the branch delay slot / bogomips problems?
[21:39] <jimp> wbx: other than just trying a newer gcc?
[21:39] <[mbm]> jimp: the delay slot is the nop in 'jump addr; nop'
[21:40] <jimp> [mbm]: yeah... i know...
[21:40] <svu> ghm.... nmbd wants libsmb. which is 500K. Which I don't have:)
[21:40] <[mbm]> jimp: instead of turning them off globally you should just identify the missing nop and manually insert it
[21:41] <jimp> [mbm]: it's got to be in gcc-generated code, though, digging for that is going to be a pain
[21:41] <[mbm]> jimp: otherwise you'll make the kernel one instruction bigger for every single branch and call
[21:41] <wbx> jimp: mbm is your contact for the CFLAGS option ;)
[21:41] <fish_> is there no other smb/nmbd implementation beside samba?
[21:41] <jimp> [mbm]: I know the current solution is not optimal, but tracking this down isn't trivial
[21:41] <wbx> fish_: no.
[21:41] <roh> svu you could try compiling it statically and strip everything out you don't need
[21:41] <[mbm]> jimp: nobody ever said it would be ;)
[21:42] <svu> roh, ghm... well, it seems it is cheaper to run it outside of wrt;)
[21:42] <fish_> wbx: i know that it is on the "wishlist" of gatling (http://www.fefe.de/gatling/)
[21:42] <jimp> [mbm]: tracking it down and fixing the one case seems like a hack, too
[21:42] <jimp> [mbm]: what about other places where gcc is making mistakes and it's not causing such an obvious effect?
[21:42] <Kaah> crap i used ipkg upgrade
[21:42] <Kaah> :(
[21:42] <[olli]> lol
[21:43] <Kaah> was in the wrong channel
[21:43] <Kaah> crap crap crap
[21:43] <Kaah> so is there a solution to this problem on the forums? :P
[21:43] <[mbm]> Kaah: yeah, install rc5 properly :P
[21:44] <Kaah> aha :P thx
[21:44] <Kaah> does the same firewall settings work for rc5?
[21:44] <Kaah> firewall.user
[21:44] <[olli]> Kaah, yes, they should work
[21:45] <Kaah> thx
[21:45] <[mbm]> Kaah: rc5 is rc4 with bugfixes for everything that was ever complained about in rc4
[21:45] <Jonek01> Question about backports for RC5: Shouldn't the packages that are in the main package repository not be removed from the backports/rc5 dir?
[21:45] <Eggplant> [mbm], everything?
[21:45] <Kaah> okidoki
[21:45] <[mbm]> Eggplant: pretty much
[21:46] <Eggplant> im sure i complained rc3-rc4 about upgrade methods =)
[21:46] <Kaah> thx for the help, i have alot to do then :D
[21:46] <Eggplant> wish it was "safer" over the wireless, but that my fault for putting units where ya can not get ot them via eth
[21:47] <[mbm]> Eggplant: doesn't really matter how you upgrade it.. just make sure you validate teh md5sum before using the mtd command
[21:48] <svu> is it worth the risk upgrading to RC5 from RC4?
[21:48] <[mbm]> risk?
[21:48] Action: [mbm] wonders what the hell svu is smoking
[21:48] <zinx> [mbm]: there is a small chance you might open a blackhole by upgrading
[21:48] <svu> [mbm], well. risk of losing the config (or hardware) by incorrect procedure...;)
[21:49] <fish_> stupid question: is RC5 kamikaze?
[21:49] <zinx> svu: your device doesn't have boot_wait?
[21:49] <svu> zinx, AFAIK it does
[21:49] <Jonek01> svu got the birdflu and chickens out :-)
[21:50] <[mbm]> svu: upgrading is pretty robust .. if you abort in the middle of it or send a corrupt firmware all that happens is that it waits for a new one to be tftp'd
[21:50] Action: zinx idly wonders if mini-fo has been added to whiterussian yet
[21:50] <svu> [mbm], ok, I see. I can just trust you on that:)
[21:50] <[mbm]> zinx: why? symlinks work
[21:50] <zinx> [mbm]: they're ugly
[21:50] <zinx> [mbm]: and annoying to work with
[21:51] <Eagle_Fire> you get used to them
[21:51] <fish_> or is 1.0 whiterussian and kamikaze is the current development?
[21:51] <zinx> Eagle_Fire: I never get used to anything. I modify the software to fit me :D
[21:51] <wbx> fish_: no.
[21:51] <[mbm]> zinx: the overlay stuff is bloat, plus it doesn't deal well with editing a file and reverting it back to it's original state (generlaly leaves two identical copies)
[21:52] <wbx> fish_: whiterussian will be OpenWrt 1.0, kamikaze is 2.0 when ready
[21:52] <spectra-> anyone with a wrt54g v2 unit here who could paste me his nvram variables? to make sure everything's fine now here
[21:52] <zinx> [mbm]: not a big problem for me :/
[21:52] <spectra-> [mbm]: the router is up with rc5 now... now i want to check wether the system-critical nvram settings are correct or not..
[21:53] <[mbm]> spectra-: it's the sdram_* variables you need to verify
[21:53] <spectra-> i don't have a stock (or even just working) v2 unit here
[21:54] <spectra-> so if anyone would be so nice to show me his... sdram_* variables :)
[21:54] <svu> [mbm], don't you know by any chance whether milkfish (SIP router) available for RC5? will the old one (from RC4) work with RC5?
[21:54] <[mbm]> spectra-: what's boardtype on that thing?
[21:54] <spectra-> boardtype=0x0101
[21:54] <[mbm]> svu: all the official packages have been updated for rc5
[21:56] <svu> [mbm], the milkfish package was listed separately, in [milkfish] section - that is why I am asking. I do not know whether it is considered as official or not
[21:57] <fish_> wbx: ah, ok thx
[21:58] <[mbm]> spectra-: hmm that one shouldn't have the error I'm thinking of
[21:58] <spectra-> uh :)
[21:58] <spectra-> yet it seems to be running fine
[21:58] <warthawg> what's up with sveasoft?
[21:58] <spectra-> suggestions how to bring load on it?
[21:59] <[mbm]> warthawg: the tons of links on openwrt.org to answer that
[21:59] <warthawg> mbm: i leave you guys alone for 2 weeks, this happens, and then you release RC 5. it's just not fair.
[22:00] <[mbm]> warthawg: bah, it was impossible to find a gap in your schedule so we could pull that off
[22:01] <warthawg> heh
[22:02] <svu> [mbm], milkfish packages are not on your side for both rc4 and rc5 - so how my openwrt knows about their availability?
[22:02] <Stepan> hm. i found out how to calculate the wgt624 config file checksum
[22:02] <Stepan> but it's only one of them :/
[22:03] <[mbm]> svu: ipkg only knows about items you've added to ipkg.conf .. we take no responsibility and provide no support for packages that didn't come fro openwrt.org
[22:03] <Stereo> [mbm]: you take responsability for those who did? :)
[22:03] <svu> [mbm], I see. may be it was bundled by lads on milkfish.org. Thanks.
[22:04] <[mbm]> Stereo: a look at how many tickets were patched suggests we might
[22:04] <[olli]> [mbm], just got another 500 error in the wiki
[22:05] <Stereo> thus the :)
[22:05] <spectra-> [mbm]: hmm, it seems to run smoothly now... could you give me a hint how to effectively produce some load and ram i/o?
[22:05] <[mbm]> [olli]: hrm .. damned python bug
[22:05] Action: [mbm] goes to fix .. again ..
[22:09] <xFallenAngel> last time i upgraded i read something about some magic way to make it work when i have no access to the router, that method included setting some no_swap_root or whatnot nvram var. That was kinda misleading, how do i upgrade without having to plug in the cable?
[22:11] <morcego> What happened to tracker.openwrt.org ? Any replacements ?
[22:14] <Kaah> is it safe for me to run the trx upgrade?
[22:14] <Kaah> i did run the ipkg upgrade :P
[22:16] <Jonek01> Kaah, procedure for going to RC5 is: backup every change you made -> upgrade with the .trx (which by default resets the jffs filesystem) and then reload your packages and settings.
[22:16] <Kaah> yay thx
[22:16] <Kaah> :)
[22:17] <Jonek01> remember the part with jffs filesystem, any changes to config files will be lost.
[22:17] <Kaah> yepp.. i have made backups
[22:26] <Kaah_> hmm the router did not reboot with the -r
[22:27] <spectra-> [mbm]: well, no real change. the router ran fine for about 15 minutes, i tried a reboot, the same weird behaviour again...
[22:27] <wbx> Kaah_: never seen it. which mtd version you use?
[22:28] <[mbm]> spectra-: hmm I'd like to see the console output if you ever hook up a serial
[22:28] <spectra-> sure, me too :)
[22:28] <Kaah_> wbx dont know :( rc4?
[22:28] <Kaah_> hmm now im getting worried cant login to in via telnet or ssh after using passwd
[22:29] <Kaah_> now it works
[22:29] <Kaah_> my fault! :P
[22:29] <Kaah_> puhh
[22:37] <Kaah> whoho everything works! :D
[22:38] <Kaah> where is the webinterface firewall config stored?
[22:45] <samot> mbm: i tried to flash bigger (than micro) image to wap. after jffs creation failure root stays r/o squashfs and tmpfs prototype of root is at /mnt. Hmm... i already solved all that and not accepting my changes is your bussiness...
[22:48] <gorec> anybody used ar5k linux port (openHAL) for atheros ?
[22:49] <Mister_X> why, gorec?
[22:50] <gorec> Mister_X: ?
[22:50] <Mister_X> why do you want to know who use ar5k?
[22:50] <gorec> :-)) - i want use it too
[22:51] <Mister_X> use it
[22:51] <Mister_X> with madwifi-old
[22:51] <Mister_X> not ng
[22:52] <gorec> i have dwl7100ap - and want use linux on it - but all my work in this way break closed hal in madwifi :-(
[22:53] <gorec> it's work on madwifi-old?
[22:55] <gorec> Mister_X: may be you know - what is it "ahb" - i think it's a bus in a board but don't know use it under linux...
[22:59] <nightwatcher> d
[23:00] <nightwatcher> ping
[23:01] <nightwatcher> nvram setting correctly
[23:03] <nightwatcher> hallo jemand wach
[23:07] <[olli]> nbd, ping
[23:14] <[olli]> .
[23:15] <Vega27> woot whats up folks im back again
[23:16] <splinterfl> nbd, you in?
[23:16] <nbd> yep
[23:16] <Vega27> nbd is there a list of changes on this release of RC5
[23:17] <nbd> not yet
[23:17] Action: nbd can't remember
[23:17] <Vega27> aight.
[23:17] <[olli]> nbd, is the following possible with /etc/config/firewall? - http://pastebin.com/625798
[23:17] <spectra-> alright people.. don't set clkfreq to 216 on wrt54g v2 :) just keeps 'em rebooting and rebooting...
[23:17] <Vega27> is it the same release as pre-RC5 version you had me check out or should i change to RC5 ;o?
[23:18] <[olli]> spectra-, who said you should overclock your router?
[23:18] <splinterfl> spectra- I could have told you that
[23:18] <nbd> [olli]: should be possible, yes
[23:18] <Vega27> spectra- is that version 2.2 also
[23:18] <Vega27> ?
[23:18] <splinterfl> I tried that last night on an old thing, but was smart enough to know how to fix it
[23:18] <nbd> Vega27: change to rc5 :)
[23:18] <[olli]> nbd, do you have an example?
[23:18] <nbd> forward:proto=tcp dport=1234:192.168.1.1:22
[23:18] <nbd> i think
[23:18] <spectra-> [olli]: well i had a later version which ran kind of unstable, and with 216mhz it ran stable
[23:18] <[olli]> nbd, tried that already. doesn't work
[23:19] <spectra-> i recall that being in the wiki once, too?
[23:19] <spectra-> dunno, anyways... not a good idea on v2 :)
[23:19] <spectra-> Vega27: no idea... in doubt, just don't
[23:19] <splinterfl> nbd is there a way to make a default 404 page in the httpd.conf (captive portal)
[23:19] <nbd> splinterfl: no
[23:19] <Vega27> nbd is my hero :O
[23:19] Action: Vega27 hands nbd a case of bawls.
[23:21] Action: splinterfl tries to avoid putting a redundant httpd server on the router..
[23:22] <Vega27> ;o
[23:22] <[mbm]> splinterfl: you could always just change the httpd code
[23:22] <splinterfl> domain show login, filename urls give 404...
[23:22] <[olli]> nbd, does not work for me. nmap still shows closed. in firewall.user it works fine
[23:23] <Vega27> well dam changed my screen rez and the change log link showed up lol >.>
[23:23] <splinterfl> how hard? Wish there was a way to just do if -e /www/404.htm show page, else default
[23:23] <[olli]> nbd, the firewall is restarted with S45firewall
[23:23] <nbd> i will look into it
[23:24] Action: [mbm] scrolls up -- [olli] what about the firewall?
[23:24] <[olli]> [mbm], /etc/config/firewall does not work for me
[23:25] <[mbm]> oh.
[23:26] <[mbm]> someone told me that webif's firewall settings don't work from IE .. I'll have to check that one later
[23:27] <moonwatcher> what is the advantage of using the trx with the mtd -r write?
[23:27] <moonwatcher> as opose to using the webif?
[23:27] <[mbm]> moonwatcher: the web interface just runs the mtd commands
[23:28] <Vega27> i friggin love how openwrt has such an well implemented dchp server >.>
[23:29] <Vega27> i was really getting fed up w/ constant ip conflicks w/ linksys firmware :X
[23:29] <moonwatcher> [mbm]: oh in that case forget i asked, i just use the webif :)
[23:29] <moonwatcher> humm... does that mean i can upgrade remotely?
[23:29] <moonwatcher> from the internet?
[23:29] <[mbm]> moonwatcher: yep, you always could
[23:31] <moonwatcher> hum...
[23:31] <moonwatcher> no shit...
[23:31] <moonwatcher> and it will boot and reestablish the pptp connection and i come back?
[23:32] <moonwatcher> oh but the dyndns script will be missing so i wont know the new ip...
[23:32] <[mbm]> well, it's slightly harder to get into failsafe from several miles away .. so you'd better hope it boots
[23:32] <moonwatcher> i get online right after an upgrade but no dyndns...
[23:32] <[mbm]> moonwatcher: well, you could always build a firmware image preconfigured for dyndns
[23:33] <moonwatcher> yeah...
[23:33] <moonwatcher> i knew it would end in that answer...
[23:33] <moonwatcher> ill do it someday
[23:33] Action: moonwatcher is still slowly digesting the linux world
[23:34] <[mbm]> no, I'm going to end it with this answer: use the original linksys firmware
[23:34] <[mbm]> ;)
[23:34] <samot> mbm: firstboot@2mb
[23:34] <samot> mbm: good news: partition too small for jffs does not get erased each boot
[23:35] <samot> mbm
[23:35] <[mbm]> samot: but otherwise it'll create the jffs2 partition, attempt to copy data to it and crash
[23:36] <samot> mbm: and leave r/o squasfs as root
[23:36] <[mbm]> because when I rewrote the firstboot stuff I forgot to change the 'cp;jffs2root --clean' to 'cp && jffs2root --clean'
[23:37] <samot> mbm: i seem you planned use tmpfs root instead. it sits at /mnt after jffs failure
[23:37] <tty56> hi, i installed whrn rc5 today, everything works fine, and finally imq is included... cool
[23:37] <[mbm]> samot: not for failure
[23:37] <tty56> but i cant get openntpd working, when i run it it breaks with some error messages:
[23:37] <samot> for what?
[23:37] <[mbm]> samot: because creating a jffs2 partition at bootup just takes too damned long and there's startup scripts which need to write to the filesystem
[23:38] <tty56> ntpd[6499]: fatal: stat: No such file or directory
[23:38] <[mbm]> samot: so tmpfs is our temporary jffs2 filesystem
[23:38] Action: moonwatcher is going to flash RC5... brb
[23:38] <tty56> ntpd[6498]: dispatch_imsg in main: pipe closed
[23:38] <samot> should i find your words in logs... layer :-)
[23:38] <tty56> does anyone know that this means?
[23:39] <samot> side target was use tmpfs is jffs init fails
[23:39] <samot> s/is/if/
[23:39] <[mbm]> samot: rc4 and earlier would nuke the flash and setup a jffs2 partition before even turning on the network .. increased the bootup time after an install a good minute
[23:40] <Jonek01> Maybe webif should depend on ipkg, since running with the micro image only have ipkg-sh installed, which does not support ipkg list_installed.
[23:40] <[mbm]> samot: yeah, it really should fall back to tmpfs .. I'll have to go and rethink that .. should n't be too hard
[23:40] <samot> much easier if you had accepted the diff i sent
[23:41] <[mbm]> samot: your patch had a hardcoded value
[23:41] <[mbm]> and it was made before the failsafe changes
[23:42] <samot> don't fool me
[23:42] <samot> i just finish your changes
[23:42] <samot> and the hardcoded value? you mean number of free blocks?
[23:43] <samot> it is "hardcoded" in jffs2 code as well
[23:43] <[mbm]> seems like it would be more portable to pick a condition like cp failure
[23:47] <samot> this is the second mechanism i used. The first is check free space to avoid jffs init if there is no space.
[23:47] <[olli]> nbd, any idea what could be wrong?
[23:47] <samot> i think that it should work even without checking number of free blocks
[23:48] <nbd> [olli]: not yet. will look into it tomorrow
[23:48] <samot> if you dislike that so much
[23:48] <[olli]> nbd, ok
[23:49] <Eggplants> anyone ever seen the TP-Link hardware?
[00:00] --- Tue Mar 28 2006