OpenWrt

[00:05] <synth> ugh i always brick my wrt when i upgrade it
[00:05] Action: synth runs to get coffee
[00:06] <[mbm]> hmm just noticed a wrt54gc on the linksys site
[00:31] <[mbm]> ..
[00:33] <synth> heh
[00:33] <synth> mbm, which piece of hardware do you recommend that openwrt supports?
[00:33] <synth> this asus one has usb it appears
[00:33] <nbd> yeah, get the asus deluxe
[00:34] <synth> i'm just sittin ehre wonderin if i want to proceed with adding SD to the WRT, or if i should give this to my mom and get something beefier
[00:34] <nbd> well, the sd mod will always be slow
[00:34] <synth> gpio is limited to what speed?
[00:34] <nbd> with an asus deluxe, you can get several mb/s transfer rate
[00:34] <synth> yeah and i could attach IDE disk in a usb caddy at that point
[00:34] <synth> which would be k-rad
[00:35] <synth> squid :D
[00:35] <nbd> :)
[00:35] <synth> so how do these diff brands end up supported? they all use the same parts?
[00:36] <Death_INC> synth: yea same board
[00:36] <synth> ahh. OEM reference designs? like nvidia
[00:37] <synth> thats good then.. i should look into the asus. i wont buy another mobo from them but if this is made by another mfg'r :D
[00:37] <Death_INC> lol
[00:37] <synth> i got burned, i bought their products for the better part of 10 years and the last 3 boards have been shit
[00:38] <Death_INC> mine's fine :b
[00:38] <synth> i had problems with the a7v8x (i blame via, mostly) a7n8x-x and -e
[00:39] <nbd> well... i'd say the asus deluxe is the most stable openwrt-supported unit there is
[00:42] <nbd> i know i asked this before... but is anybody with crashing v1 hardware here right now? (linksys wrt54g v1.x, asus wl-500g)
[00:43] <[mbm]> nbd: inh did get the speed to a somewhat impressive 300KB/s
[00:43] <nbd> i know
[00:44] <nbd> with massive overclocking :)
[00:44] <[mbm]> hehe ...
[00:44] <[mbm]> yeah we spent awhile looking at the clkfreq supporting functions
[00:45] <Dabian> What about the radio?
[00:45] <[mbm]> what about it?
[00:45] <Dabian> Have the frequency spectrum been studied?
[00:46] <Dabian> (Sorry for my English)
[00:46] <Dabian> I mean .. what frequencies it can listen to and/or recieve.
[00:46] <Dabian> receive, even.
[00:46] <[mbm]> it's your standard wifi gear, 14 channels across the 2.4Ghz spectrum
[00:47] <Dabian> I wonder for instance, if I would be able to stream sound, and listen to it in the next room.
[00:47] <Dabian> [mbm]: So .. its not possible to change this?
[00:47] <[mbm]> sure
[00:47] <Dabian> I mean .. FM for instance. .. thats around 87.5MHz to 108.0MHz .. or something.
[00:48] <Dabian> Thats very far from 2.4GHz ..
[00:48] <synth> no way ;)
[00:48] <synth> google gnu radio and see what it takes to capture aribtrary frequencies
[00:48] <Dabian> gnu radio :)
[00:48] <synth> theres a radio card they have, and it grabs just about anything.. custom board they made
[00:48] <[mbm]> no, this is 2.4Ghz only, atleast until someone publishes detailed technical manuals (even then I doubt the radio could get outside the 2.4Ghz range)
[00:49] <synth> this kitten is gonna DIE if it keeps it up
[00:49] <[mbm]> several tv tuner cards also feature AM/FM radios if that's what you want
[00:49] Action: synth is being figurative of course
[00:50] <Dabian> [mbm]: I guess you're right .. just wondering if people had played with it .. I read that sveasoft was able to send with more power than the wrt54g-standard-firmware.
[00:50] <Dabian> I assume the same is possible with OpenWRT
[00:50] <nbd> Dabian: of course
[00:51] <dogs> can someone help assist me in setting up VPN etc on a WAG54 ?
[00:51] <Dabian> So thats why I wondered if anything else had been tried with the radio :)
[00:51] <nbd> dogs: this is the openwrt channel
[00:51] <synth> Dabian: antennas have to be physically designed for the spectrum they're capturing
[00:51] <nbd> dogs: and we have no hardware support for the wag54g yet
[00:51] <dogs> yeah.. but since its linksys firmware..
[00:52] <dogs> I would thnk you guys would have some idea on VPN setup
[00:52] <synth> Dabian: so you can up the energy, yeah, but the physical characteristics still stop you afaik
[00:52] <dogs> just common terminology
[00:52] <nbd> we don't use linksys firmware
[00:52] <synth> heh
[00:52] <dogs> nbd.. I mean VPN terminology
[00:52] <synth> i didnt want to futz with openswan or openvpn so yesterday i setup ppp in a ssh tunnel
[00:52] <synth> works really well
[00:53] <synth> im sure i could tune it..
[00:53] <dogs> Local Secure Group:
[00:53] <nbd> do you know if there's any way to make it work with windows?
[00:53] <nbd> ppp over ssh, i mean
[00:53] <dogs> What would I put for the Local Secure Group:
[00:53] <synth> sure
[00:53] <Dabian> dogs: There is no VPN in the wrt54g
[00:53] <[mbm]> synth: there's some nice papers written on why encapsulating tcp over tcp like that sucks
[00:53] <Dabian> dogs: If you were on OpenWRT, I'd say look in the swan project.
[00:54] <nbd> Dabian: better openvpn
[00:54] <Dabian> ok
[00:54] <synth> mbm: there's extra layers there but it functionally its fine
[00:54] <nbd> Dabian: it's a lot easier to set up and it's just as secure
[00:54] <synth> mbm: i'll look, i'd like to see
[00:54] <Dabian> I had problems setting up swan .. then again, I didn't have any endpoints ..
[00:54] <synth> i hear openvpn is very easy
[00:55] <synth> vs swna
[00:55] <synth> swan
[00:55] <nbd> yeah
[00:55] <Dabian> Right now I have so many projects, I cannot structure them, and get nothing done.
[00:55] <Dabian> Which seems suboptimal to me.
[00:55] Action: nbd needs many projects to not get bored
[00:56] <[mbm]> synth: been awhile since I read them but I think the basics where that tcp added overhead to deal with retransmission which was redundant and that you had to set the mtu lower to properly handle a few aspects of the transmission
[00:56] <Dabian> I could need a friend to help me on my projects .. but I'm too crazy to find anyone share my visions. :-)
[00:56] <ralf> z3ro: hi
[00:56] <ralf> z3ro: what time zone do you have?
[00:56] <nbd> Dabian: what visions?
[00:56] <synth> mbm: i found it at http://sites.inka.de/sites/bigred/devel/tcp-tcp.html
[00:57] Action: [mbm] always has tons of projects going on .. I'm easily distracted
[00:57] <synth> i have too many projects, thats my problem. i can count 3 embedded linux toys on my desk
[00:57] <z3ro> ralf: I'm busy at the moment, back in about an hour.
[00:57] <synth> handheld, empeg, wrt :D
[00:57] <z3ro> ralf: I'm in nz, but I use utc.
[00:57] <ralf> bye
[00:57] Action: nbd likes to do heavy multi-tasking with completely different projects.
[00:58] <Dabian> nbd: Well .. my OpenWRT project is almost running .. I have to debug the shaping thing, and I have to move my mail-server from the old server, and then I want a two-ring network. Optimally without public ip on the wrt.
[00:58] <Dabian> nbd: Its kinda hard to explain without drawings.
[00:58] <[mbm]> synth: fyi the cipe project listed there has been attacked for numerous security issues
[00:58] <Dabian> nbd: If you're interested, I can make an ASCII representation?
[00:59] <nbd> Dabian: i don't see a problem in that description yet :)
[00:59] <[mbm]> Dabian: you'll want atleast an ip for administrative tasks
[00:59] <Dabian> nbd: I know its possible :)
[00:59] <Dabian> mbm: I would want a private IP for that.
[00:59] <Dabian> if possible.
[00:59] <[mbm]> ah ok, misunderstood
[00:59] <Dabian> I have no need to access stuff from the outside.
[00:59] <synth> mbm: ok, that makes sense ;) that sucks..
[01:00] <synth> some meltdown issue after a while
[01:01] <synth> i can ssh out of work but not much else, so i dunno what i'll do
[01:01] <synth> i hate as/400 :-\
[01:01] <synth> ssh/ppp worked welll because i could initiate the session from within the terminal i'm in
[01:01] <Dabian> Why hate AS/400? Its merely a database platform, no?
[01:02] <synth> no its a little-bro of the mainframe era, its db-like for sure
[01:02] <synth> the whole thing is an OO experience
[01:02] <synth> but more than anything its for number crunching, finances
[01:02] <Dabian> k
[01:03] <synth> a motherfuck to work with if you dont know it already. commands like wrkoutq strprtwtr
[01:03] Action: ralf is away ( http://www.autistici.org.nyud.net:8090/shockraver/ph0t0.htm)
[01:03] <synth> but its terminal emulation sucks it eats high characters
[01:04] <synth> i cant even get it let me use anything but the keyboard in its telnet
[01:04] <synth> and it owns the escape, ctrl-c et cetera
[01:04] <synth> so you sit down on it and it feels like a relic
[01:04] <synth> funny, now that i think about it, AIX was like this
[01:05] <synth> IBM. buncha stuffy bastards.
[01:08] <[mbm]> in my experience it's not so much the computers but how well they're configured and administered .. most people have this idea of running the latest greatest software on older machines and all the overhead of the new object oriented designs in today's software grinds it to a halt
[01:08] <synth> eh, i dont do that on production servers :D
[01:09] Action: synth breaks his shit all the time at home! no problem
[01:09] <synth> "honey, net is goin down gotta reboot the wrt"
[01:09] <synth> *7 hours later i'm asleep on the couch and the wrt is rebooting itself perpetually*
[01:10] <synth> upgrade gone bad... natch..
[01:10] <[mbm]> well the other thing that annoys me is how friendly machines are .. I get really annoyed if I'm on a unix machine that can't compile and doesn't have the standard set of utils
[01:10] <[mbm]> which is also a function of how much the admins actually use the system
[01:10] <t0mt0m> good morning
[01:14] <synth> yeah.. actually, that reminds me, i wanted to setup something CVS-ish for myself so i can pull down all my tools on any machine
[01:14] <synth> i work primarily on Winders at work so i need my cygwin tools, sysinternals pstools et cetera
[01:14] <synth> and of course, my .REG's so i can actually work and not curse at Winders
[01:16] <[mbm]> synth: I'm not sure you'd need a revision system for that
[01:17] <synth> well, when i make a change to a file, i'd like for it to propegate rather than update by hand?
[01:17] <synth> i have enough BS, trust me
[01:17] <[mbm]> sure, but do you need a log of when the file changed and why?
[01:17] <synth> sure, for my code
[01:17] <synth> not for my tools :D
[01:17] <[mbm]> code yes, utils no..
[01:18] <[mbm]> I'd say something as simple as rsync would probably handle what you want
[01:18] <synth> this coffee is crapppp
[01:19] <synth> yeah
[01:19] <synth> i hadnt thought of rsync..
[01:19] <synth> HM.
[01:31] <nbd> [Oleg]: hi
[01:32] <[Oleg]> nbd: hi
[01:33] <[Oleg]> nbd: does buildroot uses new wbx patches already?
[01:35] <nbd> [Oleg]: yes
[01:35] <[Oleg]> nbd: fine, how et.o is managed now? does it still not in cvs?
[01:36] <nbd> [Oleg]: we won't include stuff with 'unpublished proprietary' headers
[01:36] <nbd> [Oleg]: the montavista sources contain a free version of et with support for both the adm and the robo switch, but it doesn't contain vlan support
[01:37] <[Oleg]> nbd: oh, do you've a link?
[01:37] <nbd> [Oleg]: so we need a switchcfg utility first, before we can use that or b44
[01:37] <[Oleg]> nbd: robocfg will work with montavista stuff for sure. :)
[01:37] <nbd> [Oleg]: you have to register and get a 'preview release' with the kerne sources in an rpm contained in an encrypted loopback image :)
[01:37] <nbd> [Oleg]: but i can send you the source that i have :)
[01:38] <[Oleg]> nbd: would be great
[01:39] <nbd> [Oleg]: http://openwrt.inf.fh-brs.de/~nbd/linux-2.4.20_mv131.tar.bz2
[01:39] <nbd> [Oleg]: i can make it public, it's gpl after all
[01:39] <[Oleg]> nbd: size?
[01:39] <nbd> 28m
[01:39] <[Oleg]> ok
[01:40] <frop> lo
[01:40] <nbd> i think when we get an utility that supports both the adm and the robo switch like robocfg does, then we can switch to a completely free driver
[01:41] <[Oleg]> nbd: adm could be managed from userspace, is not it?
[01:42] <synth> damnit, i tried to flash a jffs2 image and it didnt work, so now i'm flashing a squash one over it but it's not coming to life after the tftp..
[01:42] <nbd> [Oleg]: should be possible somehow.
[01:42] <nbd> [Oleg]: i don't know much about that kind of stuff
[01:43] <synth> should the DMZ light be solid? or off
[01:43] <nbd> [Oleg]: spent a while yesterday fixing the ethtool call, without even knowing what it is or what it does
[01:43] <[Oleg]> nbd: et is GPLed by broadcom... weird... same sources, different copyright notice...
[01:43] <nbd> [Oleg]: because the new wl utility uses it to probe for the wireless interface
[01:43] <nbd> [Oleg]: i managed to make it work without breaking robocfg
[01:44] <[Oleg]> nbd: this et.o contains admtek support, as well as robo
[01:44] <nbd> but no real vlan support, right?
[01:44] <[Oleg]> nbd: why do you think it does not support vlans?
[01:44] <[Oleg]> nbd: everything
[01:45] <nbd> i got the impression from looking at a diff between your driver and the free driver
[01:45] <nbd> maybe i was too tired, i'll have a look again
[01:45] <[Oleg]> nbd: oops, they do not support vlans on the robo switch...
[01:45] <[Oleg]> nbd: they just use older et sources
[01:46] <[Oleg]> nbd: adm vlan support is here
[01:46] <nbd> but with a different structure than the older et sources that we got
[01:46] <nbd> hmm
[01:46] <[Oleg]> nbd: robo vlan support is not needed, as robocfg provides this
[01:46] <nbd> but it would be good to have the same mechanisms for controlling both switches
[01:47] <nbd> having one use robocfg and letting the driver take care of the other is not that great
[01:47] <nbd> better have a switchcfg utility which can handle both
[01:48] <[Oleg]> nbd: the switches are really different, so single utility is not practical
[01:48] <nbd> can't you at least set the vlan stuff the same way?
[01:49] <[Oleg]> nbd: no, this is switch dependent
[01:49] <[Oleg]> nbd: it's like different CPUs
[01:49] <nbd> yeah, i know. i meant having at least the same command line syntax for setting vlan on both switches
[01:49] <nbd> and then having different options for the extra stuff
[01:49] <macsat> [Oleg], is there big differences in building drivers from source, and other packages? I mean...drives are usually made either in kernel, or modules....I am considering modules here of course :-) What I want to try, it to make the spca drivers work (http://mxhaard.free.fr/camera.html)
[01:49] <[Oleg]> nbd: yes
[01:50] <[Oleg]> macsat: spca (both drivers and hw) is a peice of crap... really.
[01:51] <macsat> hehe
[01:51] <macsat> well....I have a cam
[01:51] <macsat> so...
[01:51] <macsat> and I bet a lot of ppl have...since the Creative NX Pro is working with that driver
[01:52] <nbd> [Oleg]: btw. can you do a stability test of this image http://openwrt.inf.fh-brs.de/~nbd/test/ on your asus wl500g?
[01:52] <[Oleg]> macsat: you need to adjust makefile for mips kernel first
[01:52] <macsat> maby the NX Pro uses OV9620 or the HT1621 I dunno, but it does however work with that driver...
[01:52] <[Oleg]> macsat: or add it directly to kernel source tree (simplest)
[01:52] <macsat> yeah...I kindda figured that much :-(
[01:53] <synth> every few seconds, all the ethernet port lights light up and then off
[01:53] <nbd> [Oleg]: i ported over some workarounds in the mips cache code from the montavista sources, i want to know if it helps in any way
[01:53] <synth> is that it rebooting? i cant get in so i cant tell
[01:53] <[Oleg]> nbd: ok
[01:53] <nbd> [Oleg]: thx
[01:53] <cliffstah> morning :]
[01:54] <synth> and i can ping it but i cant telnet in. i think something in my old configs is screwing the new .bin
[01:55] <[mbm]> synth: switch is resetting .. happens on powerup and again when the ethernet driver takes control over the switch .. if you're seeing it repeatedly then yes it is rebooting
[01:55] <cliffstah> I still havent installed openwrt on my wrt54g :P
[01:56] <[Oleg]> nbd: does mvista sources contains toolchain?
[01:56] <cliffstah> lemme get this straight tho .. wrt* .. BOOT_WAIT is off by default, right?
[01:56] <cliffstah> I have a 2.2
[01:56] <nbd> [Oleg]: somewhere, yes.
[01:56] <nbd> [Oleg]: i'll try to find it
[01:56] <[Oleg]> nbd: I mean famous patches for 4710 rev 0
[01:57] <nbd> [Oleg]: those that get disabled for the kernel :)
[01:57] <[Oleg]> nbd: yes
[02:01] <nbd> [Oleg]: ftp://delphi.mvista.com/YLZ4Yzatxaw5tdhthT3hfDnhtDD9bq/previewkit-mvl3.1.0-source.img
[02:01] <nbd> [Oleg]: it's an iso image
[02:06] Action: synth tries the jffs2 again
[02:06] <synth> i think i might just slap that SD on today..
[02:06] <synth> how slow could it be?
[02:07] <nbd> somewhere between 150 and 300 kbyte/s
[02:07] <synth> not tooooo bad
[02:07] <nbd> with max. cpu load
[02:10] <nbd> [Oleg]: so. how is the image on the asus?
[02:14] <[Oleg]> nbd: will flash now, I've problems finding ASUS ip...
[02:15] <nbd> k
[02:15] <Dabian> [Oleg]: You have wrt also!?
[02:16] <Dabian> [Oleg]: 192.168.1.1
[02:16] <[Oleg]> Dabian: I've lot of devices, and the ASUS in question was 192.168.2.1
[02:16] <nbd> :)
[02:16] <Dabian> [Oleg]: Then its a different asus from mine.
[02:16] <[Oleg]> was not used for 3 weeks or so
[02:17] <Dabian> oh ok
[02:17] <Dabian> Maybe its the same after all. :-)
[02:17] <Dabian> [Oleg]: I have a wl-500gd.
[02:18] <[Oleg]> Dabian: I've deluxe box too
[02:18] <Dabian> ok
[02:19] <Dabian> [Oleg]: Gobopisch?
[02:19] <[Oleg]> Dabian: what?
[02:19] <Dabian> never mind
[02:19] <Dabian> Just wondered if you speak Russian with that hostmask.
[02:19] <pp_> [C
[02:19] <pp_> oops
[02:20] <Dabian> pp_: [C] ?
[02:20] <[Oleg]> Dabian: yes, i'm russian
[02:20] <Dabian> [Oleg]: Cool
[02:20] <[Oleg]> but i can't understand what are you talking about...
[02:21] <Dabian> [Oleg]: I am not very good at russian, I took a little russian about 15 years ago.
[02:21] Action: pp_ ponders when he'll next dare to try using experimental :-)
[02:22] <Dabian> pp_: Check the topic.
[02:22] <pp_> (removing wlcompat didnt stop the crashes)
[02:22] <pp_> dabian: it just doesn't like my 54g 1.1
[02:22] <cliffstah> bbiam
[02:22] <Dabian> pp_: OK
[02:26] <nbd> pp_: do you want to test something?
[02:26] <nbd> pp_: maybe it'll work
[02:26] <pp_> maybe later :-)
[02:27] <pp_> Like in the evening
[02:27] <nbd> pp_: ok. when you feel like testing something, here's an image for you to test: http://openwrt.inf.fh-brs.de/~nbd/test/
[02:27] Action: pp_ is happily using wireless on his bed with his new X31 laptop :-)
[02:28] Action: nbd hopes that someone will send him a wl-500g for testing someday
[02:28] Action: nbd doesn't have any v1 hardware
[02:28] <Dabian> nbd: You have only the d-lux?
[02:28] <nbd> yes
[02:29] <Dabian> I will eat
[02:29] <Dabian> then I am going to ask some pretty stupid newbie questions, if its ok?
[02:29] Action: pp_ notes new wrt hardware is cheaper than soap these days ;)
[02:29] <nbd> Dabian: try to ask intelligent questions :)
[02:30] <nbd> pp_: depends on the sope and the amount of it :)
[02:30] <[Oleg]> nbd: does not boot
[02:30] <nbd> damn
[02:31] <[Oleg]> nbd: will check on the v2.2
[02:31] <nbd> [Oleg]: it boots on the deluxe
[02:33] <Dabian> nbd: I will try, but I give no warrenties :)
[02:33] <nbd> Dabian: on what device?
[02:33] <Dabian> nbd: delux
[02:33] <Dabian> e
[02:33] <nbd> Dabian: no need to. it works there
[02:33] <nbd> Dabian: and the deluxe is stable
[02:34] <Dabian> It seems to be
[02:34] <Dabian> I've been running it for days.
[02:34] <Dabian> In fact, maybe I should make a backup and upgrade before asking questions?
[02:35] <nbd> you can ask...
[02:35] <Dabian> Hmm .. there is also the problem that my internet connection is running over the wrt now..
[02:36] <nbd> what problem?
[02:36] <Dabian> oh
[02:36] <nbd> :)
[02:36] <Dabian> well, if I upgrade, I guess I need to drop the connection. :)
[02:36] <nbd> just ask your questions ... :)
[02:37] <Dabian> Then again, i guess I could just go around the wrt
[02:37] <nbd> you don't have to upgrade right now
[02:37] <Dabian> Right .. need to do it sooner or later though.
[02:38] <[Oleg]> nbd: v2.2 booted fine
[02:39] <[Oleg]> nbd: ... with lots of iptables errors
[02:42] <nbd> [Oleg]: oh yeah... must be some change that wbx did to the kernel configuration. will fix that
[02:42] <nbd> [Oleg]: btw. the build system is ready for additional hardware support (2.6 kernel or wag54g) now
[02:43] <[Oleg]> nbd: i've mounted mvista img - there is no kernel sources
[02:43] <nbd> [Oleg]: right. you asked for the toolchain
[02:43] <nbd> [Oleg]: the kernel source is in that encrypted thing
[02:43] <[Oleg]> nbd: but it has toolchain as seems - gcc-3.3.1 binutils 2.14
[02:49] <Ge0rG> some of the iptables kernel modules needed for stateful inspection are missing, which freaks out the firewall too - e.g. ipt_state.o isn't there
[02:49] <[Oleg]> nbd: as for et from mvista - the same as wrt54gv2_2.00.8 uses
[02:50] <nbd> k
[02:50] <cliffstah> would it make more sense for me to compile the buildroot with fewer modules, and add modules with ipkg later?
[02:50] <[Oleg]> nbd: so it's just outdated a bit, probably someone could drop them a message to update driver
[02:50] <cliffstah> just found myself sitting there enabling module after module (package)
[02:50] <haa42> i successfully separated the ports into interfaces on experimental :)
[02:51] <cliffstah> Oh
[02:51] <cliffstah> like, 4 ports as 4 seperate interfaces?
[02:51] <haa42> 4 + 1 wan = 5 interfaces
[02:51] <cliffstah> nice
[02:51] <cliffstah> why would you do that? :]
[02:51] <cliffstah> but cool anyway ;)
[02:51] <haa42> then add them all to br0
[02:52] <haa42> so you get a very intelligent bridge ( arp filtering , ipx filtering ...)
[02:52] <cliffstah> ah ic
[02:52] <[Oleg]> nbd: btw, as soon broadcom released it as GPL we could just patch it to the latest version
[02:52] <haa42> the latest tc ipkg doesnt contain the tc binary ! anyone realize that ?
[02:53] <nbd> [Oleg]: compiling a new image for your asus. fixed the iptables stuff and took out all the patches except for one that adds an extra nop somewhere
[02:53] <nbd> haa42: fixed in cvs
[02:53] <nbd> haa42: you can get an updated package here: http://openwrt.org/downloads/nbd/gcc34/
[02:53] <haa42> is there a snapshot or do i have to pull from cvs ?
[02:54] <cliffstah> starting to think I shouldnt have built all those modules
[02:54] <haa42> nbd : thank you very much
[02:55] <[Oleg]> nbd: in my opinion we should use et (patched), not b44
[02:56] <nbd> [Oleg]: for 2.4 at least
[02:56] <[Oleg]> nbd: exactly
[02:56] <nbd> [Oleg]: it probably doesn't work on 2.6
[02:56] <[Oleg]> nbd: why? looking at the code, it appears, that broadcom is playing with 2.6 too
[02:56] <nbd> [Oleg]: when we get a switch control utility for the adm switch, we can just rip out all the vlan stuff in the free et driver
[02:57] <nbd> [Oleg]: the osl stuff needs to be ported, right?
[02:57] <pp_> getting et into upstream will never happen :-)
[02:57] <nbd> that's not important anyway
[02:57] <[Oleg]> pp_: sure, but we've bunch of hacks already, which would never be accepted
[02:58] <[Oleg]> and I do not care of accepting this to the mainstream
[02:58] <pp_> for 2.4 anyway
[02:58] <[Oleg]> nbd: most likely broadcom already has 2.6 working
[02:59] Action: cliffstah is in over his head here
[02:59] <[Oleg]> nbd: in the mean time asus should start selling wl700g, which is probably uses bcm4780 cpu
[02:59] <pp_> they do I think
[02:59] <pp_> I've heard rumors of docs for the wireless potentially getting out too
[02:59] <pp_> (finally)
[03:00] <pp_> still work in progress in getting them, but some movement at least
[03:00] <[Oleg]> pp_: you mean brcm wlan stuff?
[03:00] Action: nbd is starting to work on a wlconf replacement today
[03:00] <wbx> moin
[03:00] <nbd> wbx: moin
[03:01] <pp_> yup
[03:01] <[Oleg]> wbx: hi, I've changed et a bit (one guy sent me a patch adding mii ioctls)
[03:01] <pp_> there's a small corner of clue in bcom ;)
[03:01] <pp_> they recently released a linux driver for their new gige chip
[03:01] <pp_> notice the word "linux driver"
[03:02] <pp_> I didn't say "driver that compiles on linux"
[03:02] <wbx> [Oleg]: can you update the package it is in downloads/
[03:02] <pp_> :-)
[03:02] <[Oleg]> wbx: need some testing yet
[03:02] <[Oleg]> wbx: as for et from mvista - the same as wrt54gv2_2.00.8 uses
[03:04] <nbd> when you update the et sources, you can add this patch for openwrt: http://openwrt.inf.fh-brs.de/~nbd/netdev-random-et.patch
[03:05] <wbx> nbd: what is the status of my buggy commits? :}
[03:05] <nbd> wbx: what buggy commits do you mean?
[03:05] <nbd> wbx: fixed ethtool with the new sources
[03:05] <wbx> nbd: i mean what bugs are still open?
[03:06] <nbd> wbx: don't really know. i fixed the two new bugs that i knew of (buggy ethtool, buggy wlcompat)
[03:06] <synth> Unlikely Microsoft tells users to uninstall Netscape 8, saying it could cause IE problems. In other news, your shoe's untied
[03:06] <synth> <3 fark
[03:06] <wbx> nbd: ah, okay. thx
[03:06] <nbd> wbx: i'm currently trying out some patches i ported from the montavista source
[03:07] <nbd> wbx: all of them at once make the wl500g fail at boot time (at least on oleg's router)
[03:07] <nbd> wbx: could you try this test image, too? http://openwrt.inf.fh-brs.de/~nbd/test/
[03:07] <wbx> nbd: strange. my images worked on wl500g..
[03:07] <nbd> wbx: i thought they still had the wireless crash
[03:07] <nbd> wbx: you said it was unresolved
[03:07] <[Oleg]> nbd: keep in mind that mvista sources are outdated (same as et)
[03:08] <nbd> [Oleg]: i know
[03:08] <nbd> [Oleg]: i'm only grepping for 4710 and 4704 in arch/mips and looking if all the workarounds are present in our current sources
[03:09] <nbd> [Oleg]: the advantage that the mvista source has is, that it's not specifically for the bcom stuff alone, so i have a better chance of finding crash workarounds that way
[03:09] <wbx> nbd: yeah, the reboots are still there. i meant the image boots up..
[03:10] <nbd> wbx: can you try my test image?
[03:10] <nbd> wbx: btw. i enabled some conntrack stuff again (stuff that you changed to compile as module)
[03:10] <wbx> nbd: just flashing..
[03:11] <wbx> nbd: into the kernel?
[03:11] <nbd> wbx: it's much more efficient to keep this in the lzma compressed kernel, and at least 90% of the people flashing openwrt are going to use it somehow
[03:11] <wbx> nbd: please not. i like to have it as modules, so you can workaround a bug...
[03:11] <nbd> wbx: let's work around this bug differently
[03:12] <wbx> nbd: hmm...
[03:12] <nbd> wbx: you mean the asterisk-behind-wrt thing, right?
[03:12] <wbx> i forgot that it take so much space..
[03:12] <wbx> nbd: yes. at the moment i reboot every night.
[03:12] <nbd> wbx: and i removed 3 useless jffs bbc compressors
[03:13] <nbd> wbx: size mode in bbc is a lot faster now (on my wrt it's 1m10 for an asterisk install instead of 1m45)
[03:13] <nbd> wbx: and it uses exactly the same amount of flash space
[03:13] <wbx> nbd: great.
[03:18] <[Oleg]> wbx: http://openwrt.org/downloads/kernel-source-et-0.9.tar.gz
[03:19] <[Oleg]> nbd: could you please update cvs to use http://openwrt.org/downloads/kernel-source-et-0.9.tar.gz?
[03:19] <nbd> sure
[03:20] <[Oleg]> this new version allows programs like mii-tool to operate with individual ports on the v2.2, v3.0 and deluxe boxes
[03:20] <[Oleg]> i.e. forcing speeds, reading ports status, etc.
[03:20] <nbd> [Oleg]: nice
[03:20] <wbx> tar: Cannot identify format. Searching...
[03:20] <wbx> tar: Cpio file name length 15713 is out of range
[03:21] <[Oleg]> wbx: to me?
[03:21] <wbx> strange, forget it.
[03:21] <nbd> lol
[03:21] <[Oleg]> :)
[03:22] <[Oleg]> oops...
[03:22] <nbd> what?
[03:23] <[Oleg]> downloaded stuff and tar fails
[03:23] <nbd> what stuff?
[03:23] <[Oleg]> et-0.9
[03:23] <nbd> it works here
[03:24] <[Oleg]> it's 28 bytes larger than original
[03:24] <[Oleg]> used http://openwrt.org/downloads/kernel-source-et-0.9.tar.gz bla-bla-bla
[03:25] <[Oleg]> rsync -t kernel-source-et-0.9.tar.gz bla-bla-bla
[03:25] <[Oleg]> should be 35520 bytes
[03:26] <nbd> it is
[03:26] <nbd> in my download
[03:27] <[Oleg]> firefox sux
[03:27] <nbd> for files like that i always use wget :)
[03:28] <[Oleg]> yes, wget worked fine
[03:29] <nbd> wbx: i'd like to write the my wlconf replacement in c, working as a drop-in replacement for the old wlconf, but using the wireless extension for standard stuff like ssid, etc
[03:29] <nbd> wbx: that way it will work with broadcom and non-broadcom stuff
[03:29] <nbd> wbx: and i don't have to implement every single thing in the wireless extension first
[03:30] <nbd> wbx: though i need to add wep support before that :)
[03:30] <wbx> :}
[03:30] <nbd> wbx: any differnet suggestion, or do you think i should do it that way?
[03:30] <wbx> [Oleg]: what do you think about making a general switch configuration tool? for robo and adm?
[03:31] <wbx> [Oleg]: if it is not on your todo, i will ask a friend to try it.
[03:31] <nbd> wbx: btw. how is my image doing in your asus?
[03:31] <wbx> it boots.
[03:31] <wbx> :}
[03:31] <nbd> wbx: it boots? does it crash?
[03:32] <wbx> nbd: need to check. i lost my test machine yesterday
[03:32] <nbd> how did that happen?
[03:32] <wbx> nbd: i needed an asterisk server by my parents. (phone flatrate :})
[03:32] <nbd> ah :)
[03:33] <marcus> hi, does somebody might want to help me installing openwrt on my wrt54gs?
[03:33] <[Oleg]> nbd: would you like to save ~10k in the compressed kernel? :)
[03:33] <nbd> marcus: the documentation surely does. and if that's not enough, you can ask questions here :)
[03:33] <nbd> [Oleg]: sure
[03:34] <[Oleg]> nbd: add -lc 1 -lp 2 -pb 2 to the lzma invocation
[03:34] <nbd> i'll do that, thx
[03:34] <[Oleg]> nbd: I've played with compression options
[03:34] <wbx> [Oleg]: any comment on my question?
[03:35] <[Oleg]> wbx: just noticed it
[03:35] <[Oleg]> wbx: adm switch is really different, so nothing common except command line
[03:35] <nbd> still useful
[03:36] <nbd> then there's at least some shared code
[03:36] <nbd> and we don't have to duplicate our script stuff for configuring vlan
[03:36] <[Oleg]> the command line parser is tightly coupled with real code
[03:36] <nbd> hmm
[03:36] <[Oleg]> i.e. it executes once it read command
[03:37] <wbx> [Oleg]: hmm. this could be modular so that we can add other switch support in the future.
[03:37] <marcus> ok, i guess openwrt-wrt54gs-squashfs.bin flashed, installed and booted fine. but i can't telnet into 192.168.1.1 afterwards. (boot_wait was set, as i had some alchemy-version before) error message: telnet: Unable to connect to remote host: No route to host
[03:37] <[Oleg]> wbx: this requires too much code to be added
[03:38] <nbd> marcus: go into failsafe mode (see documentation) and run the firstboot script
[03:38] <[Oleg]> wbx: to store args passed from command line
[03:38] <marcus> ok, i'll do this
[03:38] <nbd> marcus: was your wrt54gs configured to a different ip before?
[03:39] <[Oleg]> alternatively we could use switch "class", just like in the c++
[03:39] <nbd> [Oleg]: lzma e -si -so -eos -lc 1 -lp 2 -pb 2 input_file output_file
[03:39] <nbd> [Oleg]: what did i do wrong?
[03:40] <[Oleg]> nbd: lzma e piggy piggy.$lc$lp$pb -lc$lc -lp$lp -pb$pb
[03:40] <[Oleg]> nbd: move switches to the end
[03:40] <marcus> yes, 10.1.1.100 or 10.1.1.101. but these don't work either.
[03:40] <[Oleg]> nbd: leave e -si -so -oes as is
[03:40] <nbd> marcus: maybe you should erase your nvram when failsafe works
[03:41] <nbd> [Oleg]: i left everything as it was, just added your options
[03:41] <[Oleg]> nbd: does it work?
[03:41] <[Oleg]> :)
[03:41] <nbd> marcus: alchemy (like all the other sveasoft shit) pollutes it heavily
[03:41] <nbd> [Oleg]: no
[03:41] <nbd> [Oleg]: Error: incorrect command
[03:41] <[Oleg]> nbd: move -lc 1 -lp 2 -pb 2 to the end
[03:42] <nbd> [Oleg]: wait.. it was actually cat input_file | lzma e -si -so -eos -lc 1 -lp 2 -pb 2 > output_file
[03:42] <nbd> [Oleg]: makes more sense with -si and -so :)
[03:43] <nbd> [Oleg]: still your switches are at the end and it doesn't work
[03:43] <nbd> [Oleg]: got it
[03:43] <[Oleg]> remove spaces
[03:43] <nbd> [Oleg]: i needed to remove the spaces
[03:43] <nbd> :)
[03:43] <marcus> nbd: i think i can erase nvram, but this will turn boot_wait off, right?
[03:44] <nbd> marcus: if openwrt works, you can turn it on again later
[03:44] <marcus> ok, i'll be back
[03:45] <[Oleg]> wbx: any thoughts?
[03:46] <[Oleg]> nbd: do you've test image for my asus?
[03:47] <nbd> [Oleg]: i don't know why it crashes on your asus and works on wbx'
[03:47] <nbd> [Oleg]: any idea?
[03:47] <[Oleg]> wbx: used different image :)
[03:47] <nbd> [Oleg]: well... i gave him the same url
[03:47] <[Oleg]> nbd: for you
[03:48] <[Oleg]> nbd: have no idea then
[03:48] <nbd> [Oleg]: i want to know if the same crash is still there in wbx' unit before i do more testing
[03:48] <nbd> [Oleg]: if it doesn't work there, it's no point testing it on other units
[03:50] <[Oleg]> my wl500g is really old, so it's probably different (cpu revision or whatever)
[03:52] <nbd> yeah, but it's probably the same bug on both units
[03:52] <nbd> and if my patch doesn't fix one, then it probably won't fix the other
[03:52] <nbd> it seems to be a general v1 hardware bug
[03:53] <[Oleg]> nbd: unlikely, as stock firmwares work fine
[03:53] <nbd> i mean a general bug in our v1 hardware support
[03:53] <wbx> nbd: need to resolve some routing issues before i can try to crahs the asus
[03:53] <nbd> k
[03:55] <nbd> [Oleg]: http://openwrt.inf.fh-brs.de/~nbd/test2/
[03:57] <wigyori> re
[03:59] <nbd> [Oleg]: but there's definitely a bug in v1 hardware
[03:59] <nbd> [Oleg]: and the reason the stock firmware works and ours doesn't is probably that we haven't integrated the right workaround for it yet
[04:00] <[Oleg]> nbd: same thing with test2
[04:00] <[Oleg]> nbd: it crashes with wl.o only
[04:00] <[Oleg]> nbd: and works stable without it
[04:00] <nbd> k
[04:01] <[Oleg]> nbd: do you've older image?
[04:01] <nbd> no
[04:01] <nbd> i will look at my patches again
[04:02] <wbx> [Oleg]: my machine crash without wl.o , too
[04:02] <wbx> [Oleg]: i only need to transfer 2gb from wan to lan.
[04:02] <wbx> nbd: still not checked your image
[04:05] <nbd> [Oleg]: compiling another test image...
[04:06] <marcus> nbd: i've started into failsafe, erased nvram, had to start into failsafe once more to run firstboot. after rebooting o can now telnet into 192.168.1.1
[04:07] <marcus> nbd: thank you
[04:07] <nbd> marcus: no problem :)
[04:10] <Kaloz> re
[04:10] <[Oleg]> nbd: your gcc34 image from May 25 boots
[04:11] <nbd> [Oleg]: but it's not stable with wl.o, right
[04:11] <Kaloz> nbd: what if we probe for the board, and use that free et for adm switch, and b44 + robocfg for roboswitch?
[04:11] <nbd> Kaloz: we can use the free et + roboswitch, too
[04:11] <Kaloz> nbd: well, imho we should go with b44
[04:12] <nbd> Kaloz: because it's slower?
[04:12] <[Oleg]> Kaloz: b44 is not so good
[04:12] <Kaloz> nbd: as the difference between 4400 and 47xx is minimal, and one driver for all would be better
[04:12] <[Oleg]> Personally I prefer et.o
[04:12] <nbd> afaik et support 4400, too
[04:12] <[Oleg]> b44 is not tested with 47xx carefully
[04:12] <Kaloz> well, who will port et to 2.6? :)
[04:13] <[Oleg]> 2.6 will use b44
[04:13] <Ge0rG> is it possible to deactivate reverse lookups in dropbear?
[04:13] <wbx> Ge0rG: there is a patch for it in freifunk 1.0
[04:13] <Ge0rG> wbx: ah, thanks :)
[04:14] <[Oleg]> Kaloz: broadcom will do the port (and they probably done that already)
[04:14] <nbd> maybe we should use it for our dropbear package
[04:14] <[Oleg]> nbd: this is a config.h option
[04:14] <nbd> ah, ok
[04:15] <nbd> [Oleg]: got a new image to test: http://openwrt.inf.fh-brs.de/~nbd/test3/
[04:15] Action: nbd wants an asus wl-500g for testing
[04:15] <Ge0rG> nbd: deluxe or normal?
[04:15] <nbd> normal
[04:15] <nbd> deluxe works just fine
[04:16] <nbd> and i have a deluxe already
[04:17] <[Oleg]> nbd: wl500g v2 is coming :)
[04:17] <nbd> nice
[04:18] <[Oleg]> bcm5350 based
[04:18] <[Oleg]> nbd: same with test3
[04:18] <nbd> might buy one, when it's out. does it still have a parport?
[04:18] <nbd> hmm
[04:19] <[Oleg]> nbd: looks like this
[04:20] <nbd> [Oleg]: compiling another test image
[04:21] <wbx> nbd: 25 % / 260 mb..
[04:21] <wbx> nbd: in the meantime, what you changed?
[04:23] <nbd> wbx: added some workaround patches from the montavista source
[04:24] <macsat> [Oleg], what is the difference between comming V2 and current versions? You write that the chipset is different...but is it faster, or...?
[04:24] <wbx> hmm. 419 mb stalled. but not chrashed.
[04:24] <wbx> we have the tx/rx et patch from oleg, because last time i used his tgz
[04:25] <nbd> wbx: let's see if you can make it crash
[04:25] <wbx> will try to ping flood
[04:26] <[Oleg]> wbx: it still stalls sometimes
[04:27] <[Oleg]> wbx: do you've et binary?
[04:27] <wbx> ifconfig eth1 down/up recovered it.
[04:27] <wbx> [Oleg]: on the router, sure. or in nbd image. why?
[04:27] <[Oleg]> wbx: it would be great if you try to collect some info for me
[04:28] <wbx> [Oleg]: okay.
[04:28] <[Oleg]> wbx: I mean user space et binary
[04:28] <wbx> ah. no. can you provide one?
[04:28] <[Oleg]> wbx: you need to use latest kernel module too
[04:29] <[Oleg]> wbx: http://wl500g.dyndns.org/bin/et
[04:29] <wbx> nbd: can you prepare a new image with newest et driver with mii code?
[04:30] <[Oleg]> macsat: it should be faster, just like deluxe version, but it still has 16MB of ram as seems
[04:30] <nbd> wbx: sure. just use the next test image that i send to oleg
[04:31] <nbd> [Oleg], wbx: http://openwrt.inf.fh-brs.de/~nbd/test4/
[04:31] <wbx> nbd: ah, okay. first crash. ping flood on wan site, downloading 1 gb of data and doing ps;ps;ps;ps
[04:31] <nbd> hmm
[04:31] <[Oleg]> wbx: you will need to type "et msglevel 1" to turn on messages
[04:31] <nbd> wbx: is it at least harder to crash?
[04:32] <wbx> nbd: it is harder to crash, definetly
[04:32] <nbd> good
[04:32] <nbd> then it's a step in the right direction...
[04:33] <nbd> maybe i find more stuff like that in the other sources
[04:33] <Kaloz> [Oleg]: you mena the 520g? or an 500g v2?
[04:38] <[Oleg]> 500g v2
[04:38] <[Oleg]> 520g has 8mb and no lpt/usb
[04:38] <[Oleg]> 2mb flash
[04:39] <wbx> [Oleg]: et1: rx stuck suspected
[04:39] <[Oleg]> wbx: stalled?
[04:39] <wbx> [Oleg]: not yet.
[04:39] <wbx> [Oleg]: 5%
[04:40] <nbd> [Oleg]: does the test4 image boot?
[04:40] <[Oleg]> wbx: stall it and check if it still outputs above message
[04:40] <[Oleg]> nbd: no :(
[04:40] <wbx> [Oleg]: some more crc error
[04:41] <[Oleg]> wbx: that's ok, wait for stall
[04:41] <[Oleg]> wbx: it just indicates, that you've problems on your link
[04:42] <nbd> [Oleg]: ok. the patch that breaks disables the create_dirty_excl operations in the mips cache code
[04:43] <nbd> [Oleg]: it's in the montavista sources in a different form. seems like the old stuff doesn't like it
[04:43] <[Oleg]> wbx: could you please try http://wl500g.dyndns.org/bin/mii-tool and launch it like this: mii-tool -vv eth1 ? and send an output to me then
[04:43] <[Oleg]> nbd: try creating image without this patch to be sure
[04:43] <nbd> [Oleg]: the test3 image was without that
[04:43] <wbx> [Oleg]: it stalls on 372 mb, no kernel message
[04:43] <[Oleg]> nbd: it does not work for me
[04:44] <nbd> [Oleg]: you said the test3 was like test2
[04:44] <wbx> [Oleg]: can your remove your existin image first, mtd -e linux ?
[04:44] <nbd> [Oleg]: and you said that one booted but didn't fix the crash
[04:44] <[Oleg]> wbx: et msglevel 3 and try pinging your http source
[04:44] <[Oleg]> wbx: flashed from recovery mode, tried mtd -e linux too
[04:45] <[Oleg]> nbd: nothing works at athe moment
[04:45] <nbd> [Oleg]: so all the images didn't boot?
[04:45] <[Oleg]> nbd: yes
[04:45] <nbd> [Oleg]: except for gcc34
[04:45] <nbd> damn
[04:45] <nbd> then it's the new system code
[04:47] <wbx> very strange. why all the other machines work, even my asus wl500g.
[04:47] <wbx> may be the pci stuff..
[04:47] <[Oleg]> your wl500g is rev 2.40
[04:47] <[Oleg]> wbx: my is 1.40
[04:48] <nbd> ok. i'll add my workaround patch to cvs as it seems to improve stability on wbx' asus
[04:48] <[Oleg]> nbd: we need to create at least one bootable image first
[04:50] <nbd> [Oleg]: too bad that the wl500g doesn't have an uart
[04:52] <[Oleg]> nbd: yes, unfortunatly
[04:52] <wbx> [Oleg]: but i thought someone added a uart to wl500g..
[04:53] <[Oleg]> wbx: yes, several guys on the chupa has it. but they've rev 2.40 too
[04:53] <[Oleg]> as 1.40 uses other header
[04:53] <wbx> why you use so old hardware ;P
[04:54] <nbd> for testing :)
[04:54] <[Oleg]> wbx: ok, send me new unit :)
[04:55] <wbx> [Oleg]: http://openwrt.openbsd-geek.de/debug-et.txt
[04:56] <[Oleg]> wbx: so, ping works?
[04:58] <nbd> found another missing nop :)
[04:59] <[Oleg]> nbd: do you've a link for your patches?
[04:59] <nbd> [Oleg]: i'll put them online when i have integrated the stuff i'm currently looking at
[05:05] <[Oleg]> wbx: does ping works?
[05:06] <nbd> [Oleg]: http://openwrt.inf.fh-brs.de/~nbd/006-bcm47xx_workarounds.patch
[05:06] <wbx> [Oleg]: no.
[05:07] <[Oleg]> according to your output eth1 works fine for both tx and rx
[05:07] <[Oleg]> do you've tcpdump on the unit?
[05:07] <[Oleg]> nbd: these are bcm4704 related...
[05:08] <nbd> [Oleg]: yes
[05:09] <[Oleg]> wbx: et_start: len 98 is the echo request, et_sendup: 102 bytes is the reply received
[05:09] <caldari> any idea what this is ? wl rssidump
[05:09] <Dabian> OpenWRT is based on the "original firmware" of Linksys wrt54g?
[05:10] <z3ro> hmm... anyone want to take a guess what JP2 is on the wag54g? http://z3ro.geek.nz/wag54g_left_side.jpeg
[05:10] <nbd> Dabian: only some of the system stuff and the wireless and ethernet driver is from the original broadcom stuff
[05:10] <nbd> Dabian: the rest is different
[05:10] <z3ro> I think JP1 is jtag.
[05:10] <[Oleg]> wbx: please start pinging and check with tcpdump...
[05:11] <[Oleg]> wbx: another thing is that it's probably running in loopback mode ???
[05:12] <[Oleg]> wbx: also, try pinging non existent host too (this should trigger reset logic)...
[05:12] <wbx> gimme some seconds..
[05:12] <wbx> just trying to gte some info about ibm x40 :}
[05:12] <wbx> i like to buy a new computer
[05:17] <nbd> wbx: got a new image for you to test: http://openwrt.inf.fh-brs.de/~nbd/test/
[05:17] <nbd> wbx: it has the patches from before plus one additional nop found in some linksys source
[05:29] <wbx> nbd: is wlc obsolete?
[05:29] <nbd> wbx: i guess so. i haven't ported everything over to wlcompat yet
[05:29] <nbd> wbx: but when i have, i'll remove it
[05:35] <[Oleg]> wbx: also, run http://wl500g.dyndns.org/ettool to get additional info for me. "ettool 1"
[05:35] <wbx> nbd: i can still crash it.
[05:35] <nbd> wbx: how much data were you able to transfer until it crashed?
[05:36] <wbx> nbd: it is not important how many data i transfer. i simply need to do transfer and do ps;ps;ps;ps;ps
[05:36] <nbd> ok
[05:37] <nbd> can you crash it with ps;ps;ps... alone?
[05:38] <phedny> while /bin/true; do ps; done
[05:40] <RItalMan> yeah I can hear you now !
[05:40] <nbd> we could always hear you :)
[05:41] <wbx> [Oleg]: http://openwrt.openbsd-geek.de/mii.txt
[05:41] <wbx> [Oleg]: http://openwrt.openbsd-geek.de/ettool.txt
[05:42] <RItalMan> has anyone got a working gcc4 kernel ?
[05:43] <nbd> no
[05:43] <wbx> nbd: it does not crash, but do some memory cooruption.
[05:43] <wbx> nbd: strange things..
[05:44] <nbd> wbx: wait a sec. i think i got something
[05:44] <nbd> wbx: there's some system stuff that you didn't port over
[05:45] <nbd> wbx: some bcm4710 mmu and cache code
[05:45] <wbx> nbd: may be. i never said it is perfect :}
[05:45] <nbd> wbx: or did you integrate stuff from these files into the standard sources?
[05:45] <nbd> wbx: e.g. arch/mips/mm/c-bcm4710.c
[05:45] <nbd> and loadmmu.c
[05:48] <[Oleg]> wbx: you missed 1 in the ettool cmd line
[05:48] <[Oleg]> wbx: please run again, if it's still dead
[05:48] <[Oleg]> nbd: have you tried doing mvista vs broadcom diff?
[05:49] <nbd> [Oleg]: not yet. first i'll try to have a look at the 4710a0 specific mmu code
[05:49] <[Oleg]> nbd: check the patch - 4710 related
[05:50] <[Oleg]> you should be happy with yet another nops
[05:50] <[Oleg]> got it?
[05:51] <wbx> [Oleg]: ah shit. sorry
[05:55] <frop> mmm
[05:57] <Dabian> wbx: Watch your language, please.
[05:58] <wbx> Dabian: i do my best, shithead :p
[05:58] <Dabian> wbx: Thanks.
[06:08] <wbx> [Oleg]: hmm. no stuck.
[06:08] <wbx> [Oleg]: if i do not do anything else then download. no crahs tries at the moment. no stuck.
[06:09] <[Oleg]> wbx: the stuck seems to be with broadcom kernel too
[06:10] <[Oleg]> wbx: but it looks strange on your wl500g - packets are get received ???
[06:11] <[Oleg]> i really need ettool output :)
[06:11] <wbx> [Oleg]: i changed from crossover cable to hub based network
[06:11] <[Oleg]> wbx: this only occures with some ethernet cards
[06:11] <wbx> [Oleg]: may be that is the reason. i will do some structured data gathering.
[06:11] <[Oleg]> wbx: so, connect it directly
[06:11] <nbd> wbx: this is really weird stuff
[06:12] <nbd> wbx: but i'm porting some small 4710 things now, let's see if it works ... :)
[06:12] <[Oleg]> wbx: also, hubs are curing the problems - as they're fixing data errors, which are seems to be causing stuck
[06:12] <nbd> wbx: some BCM4710_FILL_TLB(addr) calls -- i don't even know what a tlb is :)
[06:13] <wbx> nbd: you need to buy "Mips Run" ;=)
[06:13] <[Oleg]> nbd: TLB is a vital element of the MMU - maps addresses
[06:13] <wbx> nbd: i can tell you what it is, but not more ;)
[06:14] <nbd> i'm currently diffing c-mips32.c and c-bcm4710.c, both in the asus source
[06:14] <nbd> this seems to produce some useful information
[06:14] <wbx> nbd: then you get the changes brcm has done.
[06:14] <nbd> yeah, but there's a lot of stuff there
[06:14] <nbd> most of it seems not to be applicable
[06:14] <nbd> but we'll see, what i get out of it
[06:15] <wbx> [Oleg]: okay. the problem only occurs when i directly connect to a 3com card. without hub. normal cable.
[06:15] <[Oleg]> send this card to me. :)
[06:16] <wbx> lol
[06:16] <[Oleg]> wbx: ettool please and tcpdump packets coming back
[06:16] <wbx> [Oleg]: okay.
[06:16] <wbx> [Oleg]: tcpdump on which side?
[06:16] <[Oleg]> also, check RX counter in the ifconfig eth1 output
[06:16] <[Oleg]> tcpdump on the wl500g itself
[06:17] <[Oleg]> also, ahve your tried ping on the wl500g itself?
[06:17] <[Oleg]> RX counter should increase if it gets packets (i.e. no real stuck)
[06:18] <wbx> when i ping the 3com ip, it works.
[06:18] <frop> why? why eth driver doesn't show MAC address of interface on loading?
[06:18] <wbx> but the http download does not recover
[06:18] <frop> imho, it would be useful if eth driver loaded by firmware print REAL mac addr
[06:19] <[Oleg]> wbx: tried restarting download?
[06:19] <[Oleg]> wbx: this means eth port works, something else failed?
[06:21] <wbx> [Oleg]: it seems to recover as soon as i ping heise.de
[06:21] <wbx> http://openwrt.openbsd-geek.de/ettool-working.txt
[06:21] <nbd> building a test image...
[06:21] <wbx> http://openwrt.openbsd-geek.de/ettool-bug.txt
[06:21] <wbx> http://openwrt.openbsd-geek.de/ifconfig-rx.txt
[06:22] <[Oleg]> then you should have rx stuck in the log
[06:23] <wbx> [Oleg]: it does not.
[06:24] <[Oleg]> wbx: then it worked all the time :)
[06:24] <wbx> [Oleg]: no. now the port is dead-
[06:24] <[Oleg]> have you typed et msglevel 1 before testing?
[06:24] <wbx> [Oleg]: crc error. but not stuck.
[06:24] <wbx> [Oleg]: and no data anymore
[06:24] <[Oleg]> wbx: magic
[06:25] <[Oleg]> the reset are performed for errors only...
[06:25] <wbx> [Oleg]: very strange. all the stuff.
[06:25] <[Oleg]> wbx: does ettool-working is grabbed with hub?
[06:25] <wbx> [Oleg]: after i tried to log with tcpdump, eth1 works again
[06:26] <[Oleg]> use tcpdump -n to avoid dns lookups
[06:26] <wbx> [Oleg]: eh. yes. and ettool-bug as eth1 stops, but with direct connection
[06:26] <[Oleg]> ok. i need working with direct :)
[06:26] Action: ralf is away ( http://www.autistici.org.nyud.net:8090/shockraver/ph0t0.htm)
[06:26] <wbx> [Oleg]: tell me which register is important and what you need to know
[06:27] <[Oleg]> your hub seems to be half-duplex
[06:27] <wbx> sure, 10mbit *g*
[06:27] <[Oleg]> wbx: all registers are ok, that's the problem
[06:28] <wbx> http://openwrt.openbsd-geek.de/et-direct.txt
[06:28] <wbx> after fresh boot
[06:28] <[Oleg]> wbx: try pinging non-existent host when it fails
[06:29] <[Oleg]> wbx: this should trigger rx stuck
[06:29] <wbx> now i start downloading again and set msglevel to 1
[06:29] <[Oleg]> wbx: the idea of my fix is that it transmitted a packet and does not receive anythign in 3 seconds
[06:30] <[Oleg]> then it prints "rx stuck suspected" and resets chip
[06:30] <nbd> ok. first good sign: kernel image compiled without error :)
[06:30] <[Oleg]> wbx: so, to trigger reset you could ping non-existent host (it will send arp queries, but will not receive anything back)
[06:31] <wbx> maybe it work, but the machine crashed again. so i never can say, if it is a et problem, or the generic bug
[06:31] <wbx> so before we fixed the generic bug, it is useless to debug et driver.
[06:31] <[Oleg]> wbx: when you're wgeting something it should listen for tcp retrasmissions and also trigger resets
[06:31] <wbx> jiauka: z3ro is doing porting efforts.
[06:32] <nbd> jiauka: yeah. z3ro is the guy who made a working patch agains linux-mips 2.4.17
[06:32] <wbx> [Oleg]: i use plain openbsd ftp client.
[06:32] <[Oleg]> wbx: it still use tcp, right? :)
[06:32] <jiauka> wickus and i have done some efforts
[06:32] <z3ro> someone needs to kill that bot...
[06:32] <[Oleg]> wbx: the idea is to listen for simmetrical communication
[06:32] <[Oleg]> symmetrical
[06:33] <z3ro> jiauka: what have you guys got?
[06:33] <z3ro> at the moment, we have a working 2.4.17 kernel
[06:33] <jiauka> we have our own compiled firmware, and we have been to use the standard busybox init iso the rc/pcmd binary stuff from linksys
[06:34] <wbx> my last free day. i will try wag.
[06:34] <z3ro> hi, wbx
[06:34] <jiauka> i hvae managed to make it work except the wifi part, i need more time to fix thw eifi part
[06:34] <z3ro> what are you trying on the wag54g? the latest patch on my site?
[06:35] <z3ro> that was for wbx, btw.
[06:35] <wbx> z3ro: yes.
[06:35] <jiauka> nope, our own stuff -basically the linksys gpld with some hacks/parts from other places
[06:35] <z3ro> wbx: okay. we need to start porting to 2.4.30...
[06:36] <jiauka> why do you need 2.4.30?
[06:36] <z3ro> however, I think some of the structures changed... so we'll have to find out the "new" ones.
[06:36] <z3ro> jiauka: higher kernel version is good, and openwrt uses 2.4.30.
[06:37] <wbx> z3ro: first i like to bootup your 2.4.17 and get it compiled.
[06:37] <z3ro> wbx: sure.
[06:37] <nbd> wbx: can you try another image on the asus?
[06:37] <nbd> wbx: it boots fine on the deluxe
[06:38] <nbd> wbx: http://openwrt.inf.fh-brs.de/~nbd/test2/
[06:38] <wbx> nbd: new cache patches? can you add some printf first, so that we see in dmesg if your code is actually used..
[06:38] <nbd> i'm pretty sure it gets called
[06:39] <wbx> i just like to see it. to be really sure. i have the correct image, the patch do sth... just lost to much time for this fucking router :}
[06:40] <nbd> ok
[06:40] <jiauka> [z3ro] i also setup a serial console on it
[06:41] <z3ro> jiauka: okay
[06:41] <z3ro> I also found with the help of others here (what we are pretty sure is) a jtag connector.
[06:42] <jiauka> [z3r0] jp4 on the right hand side is a JTAG
[06:43] <z3ro> that's odd, I found that JP1 is a perfect match to the arm 14-pin jtag connector.
[06:44] <jiauka> uhmm, i will test -i haven't tested the jp4 either-
[06:45] <jiauka> U need a MAX232 between JP3 and the PC. The PINout is as follows
[06:45] <jiauka> PIN 1: GND (opossite to the JP3 marking)
[06:45] <jiauka> PIN 2: NC
[06:45] <jiauka> PIN 3: RX
[06:45] <jiauka> PIN 4 : TX
[06:45] <jiauka> PIN 5: VCC (3v3)
[06:45] <nbd> wbx: compiling an image with printk inserted
[06:45] <wbx> nbd: thx
[06:45] <z3ro> jiauka: I know that, I already have a serial console.
[06:46] <z3ro> wbx: oops, I found one file is missing from my patch: /home/z3ro/wag54g/linux_2_4_17/include/asm/avalanche/sangam/sangam.h:16: haltypes.h: No such file or directory
[06:46] <z3ro> wbx: I did not touch hal_types.h though, so just get it from the actiontec or linksys wag54g sources.
[06:48] <nbd> wbx: yeah, it gets called
[06:48] <nbd> wbx: it gets called enough
[06:49] <nbd> wbx: http://openwrt.inf.fh-brs.de/~nbd/test2/
[06:49] <jiauka> back in a while
[06:51] <z3ro> alright, this seems strange. I just flashed my kernel and a (empty) jffs2 image onto the wag54g.
[06:51] <z3ro> request_module[block-major-31]: Root fs not mounted
[06:51] <z3ro> VFS: Cannot open root device "" or 1f:00
[06:51] <z3ro> Please append a correct "root=" boot option
[06:51] <z3ro> Kernel panic: VFS: Unable to mount root fs on 1f:00
[06:51] <z3ro> the kernel has the jffs2 driver compiled in.
[06:51] <nbd> it needs to know what its root fs is
[06:51] <nbd> is it really 1f:00?
[06:52] <z3ro> /dev/mtdblock/0, yeah.
[06:52] <nbd> hmm
[06:52] <nbd> it's obviously not a filesystem issue
[06:52] <nbd> it can't open the device
[06:52] <z3ro> when you say "it needs to know what it's rootfs is", what did you mean?
[06:52] <nbd> the device
[06:53] <z3ro> hmm. it has the mtd map driver... maybe something is wrong with it's config.
[06:53] <z3ro> I'll check the linksys .config and make sure the mtd stuff is correct.
[06:54] <nbd> maybe you need to port drivers for the flash chip
[06:55] <z3ro> the flash chip is a standard intel flash chip...
[06:55] <nbd> wbx: i really hope that this image works...
[06:55] <z3ro> and it has the avalanche map file, which has the correct values in .config and autoconfig.h
[06:56] <nbd> hmm
[06:59] <[Oleg]> z3ro: it should print mtd partitions before any mounting, etc.
[06:59] <[Oleg]> z3ro: does it?
[06:59] <z3ro> [Ogle]: you mean this:
[06:59] <z3ro> Creating 4 MTD partitions on "Physically mapped flash":
[06:59] <z3ro> 0x000e0000-0x007e0000 : "mtd0"
[06:59] <z3ro> 0x00020000-0x007e0000 : "mtd1"
[06:59] <z3ro> 0x00000000-0x00020000 : "mtd2"
[06:59] <z3ro> 0x007e0000-0x00800000 : "mtd3"
[07:00] <[Oleg]> z3ro: exactly
[07:00] <[Oleg]> z3ro: do you you mtdblock in the kernel config?
[07:01] <z3ro> no, but the avalanche code added this:
[07:01] <z3ro> #ifdef CONFIG_MTD
[07:01] <z3ro> ROOT_DEV = name_to_kdev_t("/dev/mtdblock/0");
[07:01] <z3ro> #endif
[07:01] <z3ro> which is also in my kernel.
[07:01] <[Oleg]> z3ro: also jffs should be flashed, otherwise jffs2 will not initialize (missing superblock or whatever)
[07:01] <z3ro> the jffs image is flashed to the router, as is the current kernel.
[07:02] <[Oleg]> ok, debug rootfs mounting then
[07:03] <[Oleg]> init/do_mounts.c
[07:04] <z3ro> [Ogle]: $ ls init/
[07:04] <z3ro> main.c main.o version.c version.o
[07:04] <wbx> nbd: hmm
[07:04] <z3ro> no do_mounts.c
[07:05] <wbx> nbd: while true;ps ;done segfaulst. but i see no dmesg changes
[07:05] <nbd> wbx: the printk is not in your image
[07:05] <nbd> wbx: but the kernel change is
[07:05] <wbx> nbd: hmm. i have a oops.
[07:06] <nbd> :(
[07:06] <wbx> nbd: give it up. i will try to get an asus for you.
[07:06] <nbd> does this image 'feel' more buggy or the same?
[07:06] <wbx> nbd: what is the city you were living?
[07:06] <nbd> hamburg
[07:07] <wbx> nbd: shit, yesterday i have seen a firend who lives there.
[07:07] Action: nbd is doing some more extensive diffing now
[07:08] <leOn> anybody knows of any good .mk file for djbdns ?
[07:08] <nbd> leOn: you can make one yourself :)
[07:08] <nbd> leOn: it's easy
[07:08] <z3ro> brb
[07:10] <[Oleg]> nbd: I've another strategy for porting 2.4.30 :)
[07:10] <leOn> nbd: was hoping someone already had it .. no point on re-inventing the wheel
[07:10] <nbd> [Oleg]: what strategy?
[07:11] <[Oleg]> nbd: probably wbx could do that - doing a diff between 1.9.4.1 gpl and 2.4.30 and the applying a patch to 1.9.4.1
[07:11] Action: leOn wished djb would change his licensing policy
[07:11] <[Oleg]> nbd: after cleaning up it...
[07:11] <nbd> [Oleg]: that sounds like a lot of work
[07:11] <[Oleg]> nbd: porting, done by wbx is also not so simple
[07:11] <nbd> i know
[07:11] <[Oleg]> but this way we do not miss anything
[07:12] <nbd> but this means going through several mb text
[07:12] <[Oleg]> yes, exactly
[07:12] <leOn> nbd: still about djb software and cross compiling .. i don't believe it's that easy ... the build process is rather different from other software
[07:13] <nbd> leOn: our build system is easy to use and we have a lot of example packages
[07:13] <nbd> leOn: it should work somehow
[07:13] <jiauk1> z3ro, how i can help with the wag54g?
[07:13] <wbx> i have no time to do this again..
[07:14] <z3ro> jiauk1: disassemble tiatm.o ;)
[07:14] <jiauk1> jejeje
[07:14] <jiauk1> any good tool?
[07:14] <z3ro> or help find why the current kernel does not mount it's rootfs.
[07:15] <jiauk1> ooops, that's related to the lzma/bzip stuff on the squashfs
[07:15] <jiauk1> probably
[07:15] <nbd> hmm... there's much more tlb stuff in the include file
[07:16] <jiauk1> [z3ro] where is the kernel image to test?
[07:16] <[Oleg]> z3ro: do_mounts contains this message and error code as well - just recompile it with debug output, that's it
[07:17] <leOn> nbd: have you ever build djbware ?
[07:17] <leOn> djbdns, qmail, etc ..
[07:17] <jiauk1> [off topic] any one knows how to get my nick jiauka back? I have registered it with nickserv but now i show up as jiauk1
[07:17] <nbd> leOn: long time ago
[07:18] <leOn> if you remember, all the building is done by binaries making system calls
[07:18] <nbd> leOn: should still be possible somehow
[07:18] <leOn> meaning that in order to invoke different building tools, one has to patch the sources
[07:18] <nbd> leOn: if you instruct it to use a different compiler and different cflags
[07:19] <z3ro> [Ogle]: thanks
[07:19] <nbd> leOn: should even work through environment variables
[07:19] <nbd> leOn: just try it
[07:19] <z3ro> jiakua1: a patch is on my site, z3ro.geek.nz
[07:19] <nbd> leOn: grab the sdk for openwrt experimental, copy over an example directory and start hacking
[07:19] <z3ro> it's missing hal_types.h though... I forgot to include it, just get it from the linksys tarball.
[07:19] <nbd> leOn: use this: http://openwrt.inf.fh-brs.de/~nbd/gcc34/
[07:21] <olli_04> hi, howto compile rrdtool for openwrt experimental? i always get an error when configure comes to math and ieee checking
[07:21] <leOn> btw ... are there still issues with early wl500g and the drivers ?
[07:22] <wbx> leOn: it has nothing to do with the drivers.
[07:22] <leOn> wbx: that's what i was told before
[07:22] <leOn> what's the problem then.. ?
[07:22] <jiauka> [z3r0] i'm new to openwrt, to which tree should i apply the patch?
[07:23] <nbd> leOn: do you have a wl500g?
[07:23] <wbx> leOn: we have no idea. find it out for us :}
[07:23] <leOn> yep
[07:23] <wbx> leOn: or seent nbd a machine.
[07:23] <leOn> but not running owrt atm
[07:23] <nbd> leOn: when i build the next image for testing, can you test it for me?
[07:23] <nbd> leOn: or is it in 'production' use?
[07:23] <leOn> nbd: sure
[07:23] <nbd> leOn: thx. i'm currently doing some kernel work and i hope this helps with the stability problems
[07:24] <leOn> nbd: it's in production use, but it's here at home .. there's no problem if i re-flash it a couple of times
[07:24] <wbx> nbd: maybe you should compile one image with the original compiler provided by asus.
[07:25] <wbx> nbd: so we can be sure it does not do this stuff, because we did not add the noops stuff, binutil will add.
[07:25] <nbd> wbx: this is no nop stuff
[07:25] <leOn> btw ... wbx
[07:25] <nbd> wbx: this is different
[07:25] <Piega`WrK> z3ro, ag241 is a wag54g without wireless and the linksys source is the same ( apparenly)
[07:25] <wbx> nbd: the binutils patch?
[07:25] <nbd> wbx: it gets disabled for the kernel -> useless
[07:25] <leOn> did you do that stdin thing to the flashing utility ?
[07:26] <leOn> brb
[07:26] <jiauka> [piega] i have succefullyt compiled and used the kernel from ag241 into a wag54g v2
[07:26] <nbd> wbx: the toolchain patch has no effect whatsoever on the kernel
[07:26] <wbx> nbd: are you sure broadcom does not use it?
[07:26] <Piega`WrK> yeah
[07:26] <Piega`WrK> :)
[07:26] <nbd> wbx: they added an extra flag to the kernel compile to disable it
[07:26] <jiauka> [z3r0] to which kernel image should i apply the patch?
[07:26] <z3ro> Piega`WrK: same chipset.
[07:27] <z3ro> jiauka: linux-mips.org 2.4.17.
[07:28] <wbx> z3ro: which toolchain you use?
[07:28] <z3ro> wbx: I'm using the one from the linksys sources at the moment, but I need to compile a newer one.
[07:29] Action: z3ro just hasn't got around to doing it yet.
[07:29] <[Oleg]> wbx: when you created initial diff, which mips kernel you've used?
[07:29] <wbx> [Oleg]: 2.4.29
[07:30] <wbx> [Oleg]: the broadcom diff was from mbm
[07:30] <wbx> [Oleg]: he cleaned it up for me.
[07:30] <wbx> that is 3-4 month ago.
[07:31] <wbx> i removed the bridge and mtd backports broadcom has added.
[07:31] <[Oleg]> well, i thinked, that the diff was broadcom vs linux-mips 2.4.20
[07:31] <wbx> and tried to get most of the other stuff included.
[07:31] <wbx> [Oleg]: yeah.
[07:32] <[Oleg]> wbx: ok
[07:32] <wbx> [Oleg]: may be Asus vs. linux-mips will have more stuff?
[07:33] <[Oleg]> no, checked asus vs linksys several times
[07:33] <[Oleg]> is there a way to download tarball from linux-mips?
[07:33] <leOn> back
[07:33] <wbx> use cvs -rlinux_2_4_20
[07:33] <jiauka> [z3r0] ftp://ftp.linux-mips.org/pub/linux/mips/kernel/ it's only up to 2.3 kernel, i'm missing something to download 2.4.17?
[07:34] <z3ro> cvs
[07:34] <leOn> wbx: was saying about mtd
[07:34] <[Oleg]> wbx: really slow for me
[07:34] <wbx> leOn: nothing changed.
[07:35] <leOn> wbx: input from stdin would be really nice i think
[07:35] <wbx> leOn: i accept patches :}
[07:35] <leOn> =P
[07:35] Action: nbd is building another image
[07:36] <wbx> leOn: send your asus wl500g to nbd and i make the feature *g*
[07:36] <leOn> i would, but like i said .. it's in production
[07:37] <leOn> using it as printserver
[07:37] <leOn> =P
[07:37] <leOn> but i actually mentioned the mtd stuff because of nbd
[07:38] <leOn> if there was stdin input, i'd give him remote access to it and a machine in the same network
[07:38] <nbd> i don't need remote access
[07:39] <leOn> and he could easily reflash it
[07:39] <nbd> i'd still need help from you if the thing crashes (which it will do, if my patches don't work)
[07:39] <wbx> remote machine is no option for developing.
[07:39] <wbx> or debugging.
[07:40] <wbx> we need a sponsor for a wl500g..
[07:40] <nbd> yeah
[07:40] <leOn> hmmm
[07:40] <leOn> for how much time do you think you'd need it?
[07:41] <nbd> don't know
[07:41] <wbx> 70 euro is not much.
[07:41] <leOn> wife (the actual user of the wl500g) and i are going on vacation in august and first weeks of september
[07:42] <leOn> i think i could spare the thing in that time
[07:42] <wbx> you pay 5-6 porto in one direction.
[07:42] Action: Ge0rG might get a wl500g sponsor before linuxtag
[07:42] <wbx> so use paypal to give nbd 10 euro.
[07:43] <nbd> Ge0rG: cool
[07:44] <Kaloz> :P
[07:44] <Ge0rG> nbd: are you going to be there btw? :)
[07:44] <leOn> i could send some EUR to nbd
[07:45] <leOn> not the 70 though .. i'd like to, but i really can't
[07:46] <leOn> say nbd .. you from germany right?
[07:47] <wbx> z3ro: do you have a working kernel config for me?
[07:48] <nbd> Ge0rG: i don't know yet. i think so
[07:48] <z3ro> wbx: after you apply the patch: "cp arch/mips/defconfig-wag54g .config"
[07:48] <nbd> leOn: yes
[07:48] <z3ro> then do "make oldconfig"
[07:49] <wbx> z3ro: and then? make all?
[07:49] <z3ro> make dep; make ram_zimage
[07:49] <wbx> ram_zimage, okay.
[07:49] <z3ro> it will make a file, ram_zimage.bin which can be flashed to the wag54g.
[07:49] <z3ro> or loaded into ram
[07:49] <wbx> z3ro: and this could be done via adam2? or directly from client via ftp?
[07:50] Action: wbx always forget this stuff..
[07:50] <wbx> if i never used it, i always forget it..
[07:50] <z3ro> wbx: I am pretty sure when the router boots up, adam2 will wait for 5seconds for an incoming ftp connection.
[07:50] <wbx> z3ro: you have already told me several times, i remember :}
[07:50] <wbx> z3ro: Adam2_AR7WRD >
[07:50] <wbx> z3ro: can i use here any commands to load it.
[07:51] <z3ro> okay, so your at the adam2 prompt. open another term and do "ftp 192.168.1.1"
[07:51] <z3ro> login with adam2/adam2
[07:51] <wbx> hmm. ping works, but ftp is closed
[07:52] <z3ro> hmm, someone else said they could not ftp either.
[07:52] <jiauka> it is tftp, not ftp
[07:52] <leOn> nbd: have you ever heared of a company called invair ?
[07:52] <z3ro> they had a different version of the wag54g.
[07:52] <nbd> leOn: no
[07:52] <wbx> bootloaderVersion 0.22.12
[07:52] <z3ro> jiauka: on my adam2, it uses *ftp* not tftp.
[07:52] <leOn> nbd: care to take a look at a link ?
[07:52] <jiauka> mine uses tftp for sure
[07:52] <jiauka> tftp 192.168.1.1
[07:53] <jiauka> bin
[07:53] <z3ro> ADAM2 Revision 0.22.06
[07:53] <jiauka> put upgrade_code.bin (or whatever firm)
[07:53] <z3ro> wbx: you have a newer version of adam2 then...
[07:54] <jiauka> mine is ADAM2 Revision 0.22.12
[07:54] <z3ro> well, version 0.22.06 uses ftp, .12 must use tftp.
[07:55] <jiauka> [z3r0] a quick guide to apply the patch?
[07:56] <z3ro> jiauka: just like any other patch
[07:56] <z3ro> patch <~/patch
[07:56] <z3ro> probably need -p0 or -p1 in there too
[07:57] <wbx> hmm. 2.4.17 and gcc3.4.3 seems not to be a good idea.
[07:57] <wbx> z3ro: can you give me a kernel binary?
[07:58] <z3ro> wbx: yes.
[07:58] <wbx> z3ro: just for testing this procedure.
[07:58] <z3ro> sure
[07:58] <z3ro> give me a second, I'll scp ram_zimage.bin to my server,
[07:59] <z3ro> wbx: z3ro.geek.nz/ram_zimage.bin
[07:59] <chaimj> any experiences using the wsb24 (signal booster) with wrt54g ?
[08:00] <jiauka> [z3r0] which toolchain are you using?
[08:00] <z3ro> jiauka: the one from the linksys wag54g tarball, for now.
[08:01] <wbx> nbd: btw: 104-netfilter_maxconn.patch is obsolete if conntrack is in kernel
[08:01] <jiauka> uhmm, linux tree looks for mipsel-linux-gcc but the one from linksys is mipsel-uclibc-gcc
[08:02] <nbd> wbx: it is?
[08:02] <wbx> nbd: yes, old version had a mistake, missing ;
[08:02] <z3ro> jiauka: symlink it.
[08:02] <wbx> nbd: after i used modules, i get a compile error
[08:02] <wbx> nbd: then i fixed the patch.
[08:03] <nbd> good
[08:03] <wbx> nbd: there are different c files for standalone and module
[08:04] <wbx> put upgrade_code.bin
[08:04] <wbx> Error code 4: Wrong Code Pattern
[08:04] <wbx> Code Pattern should be:WA21 , but Upgrade Code Pattern is:Búíþ!!!
[08:04] <wbx> so i can only add firmware, but do not load kernel to ram?
[08:04] <wbx> with this version of adam2?
[08:05] <z3ro> wbx: I do not know, sorry.
[08:06] <z3ro> I don't think the adam2 0.22.12 sources are released?
[08:06] <jiauka> yes wbx you can only upload full images
[08:06] <wbx> jiauka: no chance to fetch a kernel? go ...
[08:06] <jiauka> i haven't been able to do it
[08:07] <jiauka> what i have done is setup a fs tree from the original image and a small script that builds the image all together
[08:07] <z3ro> wbx: you might have to look into the linksys wag54g sources, and find how the firmware image is constructed.
[08:08] <wbx> jiauka: how i make the image?
[08:09] <jiauka> you need a filesystem extracted from the original binary file
[08:11] <wbx> jiauka: i mean only the command to add the header..
[08:11] <jiauka> ../tools/bin/merge ram_zimage.bin target.squshfs 1
[08:11] <jiauka> chmod 644 code.bin
[08:11] <jiauka> ../tools/bin/addpattern -i code.bin -o upgrade_code.bin -p WP21
[08:11] <jiauka> ../tools/bin/tichksum upgrade_code.bin
[08:12] <jiauka> merge add_pattern and tichksum from the linksys tarball
[08:13] <jiauka> and target.squashfs extracted from WAG54GV2-EU_AnnexA_ETSI_v1.00.19_001_code.bin
[08:13] <wbx> jiauka: can you send me target.squashfs?
[08:14] <jiauka> yup, hold on
[08:21] <jiauka> [wbx] www.escudellers.com/target.squashfs.orig.tar.gz
[08:24] <jiauka> [z3ro] the kernel is not able to mount the fs because the squashfs is missing on the kernel tree
[08:24] <wbx> ah, good :}
[08:25] <wbx> only 18 rejects on linux 2.4.30
[08:25] <z3ro> jiauka: I was trying jffs2, not squashfs.
[08:25] <z3ro> my router has a jffs2 root on it.
[08:25] <z3ro> wbx: yeah, but the avalanche cpu init is wrong.
[08:26] <z3ro> because the structures have changed.
[08:26] <wbx> z3ro: just beginning with the fun ;==
[08:26] <z3ro> hehe
[08:27] <jiauka> [z3ro], do you want to use jffs2?
[08:27] <z3ro> ye
[08:27] <z3ro> *yes
[08:28] <jiauka> uhmm, then i need to add on my kernel tree
[08:33] <wbx> jiauka: do i need a special addpattern binary?
[08:33] <wbx> jiauka: i still get Code Pattern should be:WA21 , but Upgrade Code Pattern is:WAG2!!!
[08:33] <wbx> ---------- add code pattern --------
[08:33] <wbx> input file is [code.bin]
[08:33] <wbx> output file is [upgrade_code.bin]
[08:33] <wbx> code pattern is [WA21]
[08:34] <jiauka> the pattern is hard coded on the addpattern binary
[08:34] <wbx> -p is useless?
[08:34] <jiauka> yes
[08:34] <wbx> ah
[08:34] <nbd> weird stuff
[08:34] <jiauka> you can manually change it before the tichksum command
[08:35] <wbx> jiauka: how?
[08:36] <jiauka> use you favourite binary editor
[08:36] <nbd> sed
[08:36] <nbd> like in openwrt :)
[08:36] <nbd> with the W54G and W54S stuff
[08:37] <jiauka> anyone has a quick guide to make a jffs2 fs with kernel 2.6?
[08:38] <jiauka> http://www.escudellers.com/tools.bin.tar.gz
[08:38] <jiauka> [wbx] http://www.escudellers.com/tools.bin.tar.gz there is a proper addpattern
[08:38] <wbx> 404 Not found
[08:39] <jiauka> http://www.escudellers.com/tools.tar.gz sorry
[08:40] <ralf> hola
[08:41] <synth> whats this 'wl' command?
[08:43] <wbx> Code Pattern Correct! juhu
[08:44] <wbx> jiauka: thx
[08:44] <leOn> but
[08:45] <jiauka> wbx, if you use the image from z3ro won't work for sure
[08:45] <cliffstah> Eagle_Fire: you there? :]
[08:45] <synth> anyone know of the binary named wl? kismet wont go w/out it
[08:46] <z3ro> I never said my kernel was completly working.
[08:46] <jiauka> the target.squashfs is a squashfs filesystem, but the kernel does not have support for it
[08:46] <jiauka> yes z3ro, i'm not complaining at all
[08:46] <z3ro> do you guys want me to compile support for squashfs into it?
[08:46] <wbx> jiauka: no problem.
[08:46] <wbx> jiauka: i just want to have a routine to flash my test kernels.
[08:46] <wbx> Linux version 2.4.17 (z3ro@lianli)
[08:47] <jiauka> z3ro could be interesting for testing, but i prefer to use jffs2
[08:47] <wbx> at least this is working, so everything fine for me
[08:47] <z3ro> wbx: good. :)
[08:47] <wbx> ind . -name \*rej|wc -l
[08:47] <wbx> 13
[08:47] <z3ro> wbx: you have my patch too, so you can compile your own kernel and test things. :)
[08:47] <wbx> just trying to merge it to 2.4.30 with all openwrt patches.
[08:48] <nbd> wbx: when you tried my image and it crashed, then networking was dead too, after that. right?
[08:48] <wbx> nbd: it reboots most of the time.
[08:48] <nbd> wbx: because when leOn flashed my image, the device was still pingable
[08:48] <nbd> wbx: he tried it two times
[08:48] <nbd> wbx: could be an improvement ...
[08:49] <z3ro> wbx: yeah, I'll start working on that tonight, after I sleep for the day. It's 4:51am here.
[08:49] <wbx> i just do not want to spent my last free day for hacking on buggy hardware. ;=)
[08:49] <nbd> wbx: my new image contains more workarounds in the mmu stuff
[08:49] <nbd> wbx: of course not
[08:49] <nbd> wbx: just wanted to tell you
[08:49] <z3ro> I need to go out in the morning for jtag parts (basically connectors and some ribbon cable)
[08:49] <wbx> z3ro: or wait, i send you a mail after i finish..
[08:49] <z3ro> wbx: sure
[08:49] <wbx> nbd: okay sure.
[08:54] <[Oleg]> nbd: give me your last image
[08:54] <nbd> [Oleg]: it will almost certainly still crash on your asus
[08:55] <leOn> hey [Oleg] ... you the one from wl500g.dyndns.org ?
[08:55] <nbd> [Oleg]: but if you want, you can still try
[08:55] <[Oleg]> anyway i would like to try it
[08:55] <nbd> [Oleg]: http://openwrt.inf.fh-brs.de/~nbd/test.trx
[08:55] <[Oleg]> leOn: yes
[08:55] <nbd> [Oleg]: the patches are on that page, too
[08:56] <leOn> using your stuff till owrt is not working
[08:56] <leOn> =P
[08:56] <leOn> ops
[08:56] <leOn> s/not//
[08:58] <[Oleg]> nbd: looking at your mmu patch - does wbx missed this?
[08:58] <nbd> [Oleg]: yes
[08:58] <nbd> [Oleg]: he used the generic mmu stuff
[08:58] <[Oleg]> oh...
[08:58] <nbd> [Oleg]: i ported parts of the broadcom mmu stuff into the generic file with ifdefs
[08:59] Action: wbx hides behind a stone
[08:59] <nbd> [Oleg]: it's almost certainly not all of it, but i haven't made sense of ther rest, yet
[08:59] <jiauka> [z3r0] the jffs2 filesystem gets too big for my 4MB flash wag54g
[08:59] <nbd> [Oleg]: it's completely different from the stuff in 2.4.30
[08:59] <leOn> wbx: shame on you ... and for that you shall give stdin capability to mtd!
[08:59] <wbx> leOn: :p
[08:59] Action: leOn grins
[09:00] Action: leOn goes for a smoke
[09:00] <leOn> brb
[09:01] <[Oleg]> nbd: it does not boot as usual...
[09:01] <nbd> [Oleg]: might be the pci stuff
[09:02] <nbd> [Oleg]: i'll have a look at it...
[09:05] <nbd> kernel hacking is a bit hard if you don't have a clue about either the kernel itself or the hardware it's running on
[09:08] <nbd> ah, found it (i think)
[09:09] <nbd> wbx: did you comment this out? -- if ((chip == BCM4310_DEVICE_ID) && (chiprev == 0)) pci_disabled = TRUE;
[09:10] <nbd> lool
[09:10] <nbd> it's just a typo
[09:10] <nbd> a /* that should really be a */
[09:11] <nbd> i guess someone didn't use syntax highlighting :)
[09:11] <wbx> :p
[09:12] <nbd> [Oleg]: i'll build an image now that actually has a chance of working on your asus
[09:12] Action: wbx should stop hacking on this stuff. to not break more ;)
[09:12] <nbd> wbx: no, don't stop. it's fun :)
[09:15] <nbd> wbx: btw. you also disabled the check for (boardflags & BFL_NOPCI) -- is this necessary or was it just a test?
[09:15] <wbx> it was neccessary to work on gs and deluxe.
[09:15] <nbd> ah, ok
[09:15] <nbd> thought so
[09:15] <wbx> nbd: deluxe will not work, because boardflags is empty
[09:16] <wbx> nbd: and so pci is disabled.
[09:16] <nbd> k
[09:16] <wbx> nbd: deluxe without pci, does not boot
[09:16] <nbd> wait... why does it disable pci, when boardflags is empty? is boardflags set to -1?
[09:16] <wbx> nbd: may be you find a better solution..
[09:18] <nbd> wbx: if it works, i'll leave it that way :)
[09:18] <leOn> nbd: just say when
[09:18] <leOn> =)
[09:19] <nbd> leOn: i'll do that. first i want oleg to check something that has to do with a different problem ...
[09:19] <leOn> k
[09:20] <nbd> [Oleg]: http://openwrt.inf.fh-brs.de/~nbd/test.trx
[09:20] <nbd> [Oleg]: that one should at least boot
[09:38] <Piega`WrK> can I compile only a package??
[09:38] <SeRi> inh thanks for the help! Got it working
[09:40] <nbd> [Oleg]: so. how is the image doing?
[09:49] <ixs> anyone got an idea what kind of problem this could be? the power led is flashing, dmz is lit, wlan is off. The wrt54gs is that way since powerup
[09:50] <nbd> ixs: what openwrt version did you flash?
[09:50] <nbd> ixs: what filesystem?
[09:52] <ixs> nbd: right now, I flashed the WRT54GS_3.37.2_ETSI_code.bin file.
[09:53] <nbd> ixs: don't know anything about the original firmware
[09:53] <caldari> from the web or with a tftp ?
[09:53] <ixs> nbd: but it doesn't really matter which firmware one flashes. always the same. right now, the led status changed though. power is flashing, dmz is changing between on and off every 3 seconds.
[09:53] <ixs> caldari: tftp
[09:53] <nbd> ixs: that's a different story... :)
[09:53] <caldari> tftp dont accept bioses bigger than 2mb
[09:54] <nbd> ixs: if dmz is flashing slowly, then the firmware on the device is corrupt
[09:54] <caldari> download from ftp.linksys.com the 2.00.0 version
[09:54] <nbd> ixs: you need to flash a smaller image (like openwrt)
[09:54] <caldari> and from that version try a web upload the new one
[09:54] <ixs> okay. let's try openwrt again.
[09:54] <nbd> ixs: don't get the 2.00 version
[09:54] <nbd> ixs: it might not work
[09:54] <caldari> hmmm yes nbd is right
[09:54] <caldari> 3.37 = gs not wrt54g.... might not work
[09:55] <caldari> nbd can you plz check something at the original jffs2 image for me ?
[09:55] <ixs> nbd: okay. right now, I just flashed the openwrt-wrt54gs-jff2.bin from my own build of the experimental cvs checkout on friday.
[09:55] <nbd> caldari: what?
[09:56] <jiauka> [z3ro] 2.4.17 with your patch working on my wag54g
[09:56] <ixs> nbd: for the record. This firmware runs fine on 13 other wrt-54gs and one wrt-54g
[09:56] <jiauka> using jffs2 ;)
[09:56] <caldari> ntpclient is running before telnetd/dropbear ?
[09:56] <z3ro> huh? hmm, maybe it was because I was testing with an empty jffs2 image.
[09:56] <z3ro> does yours have anything in it?
[09:56] <jiauka> yes
[09:56] <z3ro> filex extracted from the linksys squashfs?
[09:57] <z3ro> *files
[09:57] <jiauka> i guess you need at least the lib/modules
[09:57] <jiauka> yes, files extracted for orig 1.0.19 linksys binary image
[09:57] <nbd> caldari: it is
[09:57] <z3ro> okay, I'll do that and test it.
[09:57] <jiauka> also i have used a different .config file
[09:57] <caldari> nbd what will happen if no internet connection ? ntpclient will stay there waiting forever ?
[09:58] <nbd> caldari: seems like a bug
[09:58] <nbd> seems like a bug that i will fix now
[09:58] <caldari> i had that problem and had to safeboot and add a & at the end of the ntpclient
[09:58] <caldari> so the boot process will continue normaly
[09:59] <jiauka> [z3r0] you need to add the MTD stuff and the JFFS2 on your .config file
[09:59] <nbd> caldari: changed it from S41 to S60 and added the missing &
[09:59] <nbd> caldari: next snapshot will contain a fixed package
[10:00] <caldari> nbd :)
[10:00] <jiauka> [z3ro] the problem is that jffs2 does not compress as good as the squashfs, and the image gets too large
[10:01] Action: nbd has a slight headache from hacking on the kernel stuff
[10:01] <caldari> also the official bios set the nvram param wlan_gateway. you think would be nice to set something like: route add default gw $(nvram get wan_gateway)
[10:01] <caldari> ?
[10:01] <leOn> nbd: i was getting something like that, but i'm better now ... beer does wonders .. ;)
[10:02] Action: nbd doesn't have any beer right now
[10:02] <nbd> why does it have to be so hot outside?
[10:02] <nbd> that sucks
[10:03] <z3ro> nbd: heh, middle of the nigh here... nice and cold. ;)
[10:03] <z3ro> well, early morning now... 6am.
[10:03] <jiauka> [z3ro] do you have a patch to add the squashfs into the 2.4.17 tree?
[10:03] <caldari> <-- air condition
[10:03] <z3ro> jiauka: no
[10:03] <jiauka> ok
[10:05] <ixs> nbd: hmmm. this device here is still the same. power is flashing, dmz is lit. since about 10 min.
[10:06] <wbx> nbd: now you know how i am feeling the last week ;=)
[10:07] <leOn> nbd: european clima is getting weird, don't you feel that?
[10:07] <jiauka> [z3ro] the problem was the MTD stuff, do you want my .config?
[10:08] <wbx> jiauka: show us the vars you needed to activate
[10:08] <z3ro> jiauka: in my .config the mtd stuff is fine, I probably just forgot to cp .config arch/mips/defconfig-wag54g after I changed it.
[10:08] <z3ro> so you have the old .config.
[10:09] <jiauka> nope, it's a mix
[10:09] <jiauka> from ag241 & wag54g v.1.0.19
[10:11] <nbd> did you try the wag54g-de sources?
[10:11] <nbd> maybe they contain some newer stuff
[10:12] <jiauka> could be
[10:12] <jiauka> at least you need
[10:12] <jiauka> CONFIG_MTD_CFI=y
[10:12] <jiauka> CONFIG_MTD_GEN_PROBE=y
[10:12] <nbd> it's 1.0.23
[10:12] Action: z3ro is thinking about that "extra" ram chip solder pads on the wag54g...
[10:13] <z3ro> if I was to (be able to obtain and) solder another ram chip, with the same specifications onto this pad... it would still work, but maybe not "see" the extra memory without some bootloader hacking?
[10:15] <wbx> nbd: do we have it in /archiv?
[10:15] <nbd> yes
[10:19] <ixs> hmmm. I think this unit is toast.
[10:19] <nbd> openwrt does not boot?
[10:20] <ixs> nope. same. power flashing, dmz lit. for about 25 mins now
[10:20] <nbd> can you return it to the store?
[10:21] <ixs> nbd: probably. was mailorder though. but this unit is not even a week old
[10:22] <z3ro> jiauka: hmm, with jffs2 the rootfs gets too big.
[10:22] <z3ro> what's needed to make it boot
[10:23] <z3ro> just init, I guess?
[10:24] <jiauka> yes
[10:24] <jiauka> well, init and the lib folder!!
[10:29] <jiauka> [z3ro] the minimum image is: /dev /proc /lib /lib/modules and the sbin/init (i'm using busybox init)
[10:29] <z3ro> okay
[10:31] <z3ro> okay... I just tried it with a cramfs image containing the contents of the linksys squashfs, which did not work.
[10:32] <z3ro> I'm thinking something is wrong in the kernel... it's not mounting the rootfs.
[10:34] <jiauka> cramfs or jffs2?
[10:35] <jiauka> [z3ro] your .config did have the jffs2 on it
[10:35] <z3ro> I tried a cramfs image, and obviously changed the kernel options.
[10:39] <jiauka> [z3ro] i have added squashfs
[10:39] <z3ro> that's not going to help if my kernel won't mount a rootfs.
[10:39] <z3ro> it should have mounted the cramfs image.
[10:40] <jiauka> what's the message?
[10:40] <z3ro> request_module[block-major-31]: Root fs not mounted
[10:40] <z3ro> VFS: Cannot open root device "" or 1f:00
[10:40] <z3ro> Please append a correct "root=" boot option
[10:40] <z3ro> Kernel panic: VFS: Unable to mount root fs on 1f:00
[10:41] <jiauka> do you have /dev on the filesystem?
[10:42] <z3ro> the directory is there, yes.
[10:54] <jiauka> z3ro www.escudellers.com/www/linux/linksys/wag54gv2/fs.tar.gz this is my fs
[10:54] <z3ro> 404
[10:55] <jiauka> http://www.escudellers.com/linux/linksys/wag54gv2/
[10:56] <z3ro> hmm, I guess it's setup to disallow direct linking.
[11:03] <jiauka> did you got it?
[11:03] <z3ro> yes
[11:03] <z3ro> I'll try it tonight, after I sleep for the day.
[11:03] <z3ro> need to go out and get the parts for a jtag cable, then sleep, then more hacking. :)
[11:04] <jiauka> i have uploaded the squashfs patch
[11:04] <z3ro> and of course: make the jtag cable and test it.
[11:04] <jiauka> i have some experience with jtag (but not on mips processors)
[11:05] <z3ro> I'm going to use the openwince jtag tools.
[11:05] <wbx> just one link error and i can do the first 2.4.30 runtime debug :}
[11:05] <z3ro> wbx: you've been busy. :)
[11:06] <jiauka> ;)
[11:06] <nbd> wbx: wow. you're fast
[11:06] <z3ro> did you fix the errors in the sangam code?
[11:06] <z3ro> because the structures changed.
[11:06] <wbx> but may be it will fail and i need another 30 hours to get further ;=)
[11:06] <z3ro> such is the joy of hacking. :)
[11:06] <jiauka> anyone knows how to have days that last 40 hours? :-D
[11:07] <wbx> i need a dayextent fix ;=)
[11:07] <z3ro> hehe
[11:07] <nbd> when i had a lot of free time, i had an almost perfect 36 hour cycle
[11:07] <nbd> i didn't care about day or night that time
[11:09] Action: z3ro just stays awake at night.
[11:11] <z3ro> though I have to stay awake until about 10am today...
[11:11] <z3ro> need parts for jtag. :(
[11:11] <wbx> z3ro: are you a student?
[11:11] <z3ro> damn buisness hours!
[11:12] <jiauka> FYI, the squashfs patch works fine http://www.escudellers.com/linux/linksys/wag54gv2/squashfs2.4.17.diff
[11:12] <z3ro> wbx: yes, but I work from home.
[11:13] <wbx> btw: what network driver is needed for wag?
[11:13] <z3ro> wbx: *shrug* there are sources for it, though.
[11:13] <z3ro> avalanche_cpmac and avalanche_vmac
[11:15] <jiauka> [z3ro] what file you wanted to dissam?
[11:15] <z3ro> tiatm.o
[11:15] <z3ro> the adsl driver.
[11:16] <jiauka> i can dissam it, but without a proper xref
[11:16] <z3ro> jiauka: ???
[11:16] <z3ro> "proper xref"?
[11:17] <jiauka> without references to memory places, etc...
[11:17] <jiauka> then it is very hard to follow :(
[11:18] <z3ro> well, do what you can... any help is good help, especially when trying to dissemble something.
[11:18] <jiauka> ok will do l8r or tomorrow, i have to go now
[11:18] <z3ro> okay. cya
[11:18] <jiauka> cu
[11:19] <jiauka> BTW, what's the intended roadmap?
[11:20] <z3ro> ?
[11:20] <z3ro> for the wag54g?
[11:20] <jiauka> yes
[11:20] <z3ro> basically, port to 2.4.30, and then get that working with openwrt.
[11:21] <nbd> and then have fun :)
[11:21] <z3ro> disassembling the tiatm.o as soon as possible, if we can
[11:21] <jiauka> i can dissasembly it
[11:22] <z3ro> and get reasonably good c source?
[11:22] <nbd> let's try to make it work with the new kernel first
[11:22] <nbd> reverse engineering a driver takes a while
[11:22] <nbd> and if we can make it work like it is first...
[11:22] <z3ro> nbd: no harm in doing parallel development on the kernel and tiatm.o
[11:22] <jiauka> nop, i will get a horrible asm source
[11:22] <z3ro> well... objdump can do that.
[11:23] <jiauka> well, far better than objdump
[11:23] <z3ro> okay. that would be good.
[11:24] <jiauka> i will do it with a windows soft, don't blame me ;)
[11:24] <jiauka> cu
[11:29] <wbx> z3ro: can i use an uncompressed kernel without inflate stuff?
[11:29] <mac|gyver> great found the cause of it
[11:35] <z3ro> wbx: nope
[11:35] <wbx> z3ro: is inflate not only the loader which decompress the gzipped kernel?
[11:36] <z3ro> wbx: it's job is to decompress the gzipped kernel, yes.
[11:36] <z3ro> and provide a adam2 compatible binary...
[11:37] <wbx> so if i use an uncompressed kernel, i need no inflate..
[11:37] <z3ro> adam2 would not accept it.
[11:37] <z3ro> wbx: read: http://www.seattlewireless.net/index.cgi/DlinkDslG604t#head-598c178f857d16afe9f2369086aa0fd257315bdb
[11:37] <wbx> should'nt adam2 accept SREC kernels?
[11:38] <z3ro> no, it does not.
[11:38] <z3ro> adam2 is damn annoying.
[11:38] <z3ro> which is why I want to (eventually) replace it with something better, like yamon.
[11:38] <Dabian> z3ro: Maybe you want to punch his nose a little?
[11:39] <z3ro> Dabian: and how would one punch a bootloader?
[11:39] <z3ro> ;)
[11:39] <Dabian> z3ro: Uhm ...
[11:39] <wbx> z3ro: can you give me your uncompressed vmlinux?
[11:39] <Dabian> z3ro: That might be diffecould :-P
[11:40] <z3ro> wbx: yes
[11:40] <z3ro> wbx: give me a second
[11:40] <wbx> just for a test..
[11:41] <z3ro> wbx: z3ro.geek.nz has my vmlinux and System.map now.
[11:45] <Demonic> is there a way to see all connections that are made through the router... rather than just the connections TO the router?
[11:46] <wbx> Demonic13: iptables -J LOG and maybe iptables -vL
[11:46] <Demonic13> wbx: what about the netstat command?
[11:48] <wbx> z3ro: mipsel-linux-objcopy -S -O srec vmlinux-2.4.17 srec
[11:48] <wbx> z3ro: ./scripts/srec2bin srec ram_zimage.bin
[11:49] <wbx> z3ro: is enough.
[11:49] <wbx> z3ro: bad for me. my 2.4.30 does not boot :}
[11:49] <nbd> ... which was to be expected
[11:49] <nbd> what does it say?
[11:49] <wbx> sure :}
[11:49] <wbx> |
[11:50] <wbx> not even a single byte.. maybe serial code merged badly..
[11:50] <z3ro> wbx: really, it boots the kernel with just a srec image put through srec2bin?
[11:50] <wbx> z3ro: yes.
[11:51] <wbx> z3ro: verify it.
[11:51] <z3ro> wbx: sure
[11:51] <wbx> z3ro: so we can be sure.
[11:52] <z3ro> wbx: yes, that worked for me.
[11:52] <z3ro> but... is it worth keeping the inflate stuff? for gzip?
[11:53] <z3ro> the kernel is much smaller gzipped.
[11:53] <wbx> z3ro: we can use lzma loader later.
[11:53] <z3ro> okay
[11:53] <wbx> z3ro: just had some issues compiling inflate..
[11:54] <wbx> z3ro: so it was kind of workaround i tried *g*
[11:54] <z3ro> :)
[11:54] <wbx> can you explain me, what you needed to change so that serial worked with your patch?
[11:56] <z3ro> all the changes are in drivers/serial.c and include/asm-mips/serial.h
[11:56] <wbx> z3ro: okay. important fact for me
[11:57] <z3ro> I suppose you could just replace them with the ones from the wag54g 2.4.17 kernel... for a quick hack.
[11:57] <z3ro> that might work.
[11:57] <z3ro> unless the serial structs have changed.
[12:02] <z3ro> I need sleep now... I asked someone to pickup the jtag components for me, and wrote there component numbers down so (hopefully) I should get the right components.
[12:02] <wbx> z3ro: good nigt and thanks
[12:03] <z3ro> wbx: no problem, and good night. :)
[12:06] <wigyori> re
[12:41] <beattie> I seem to have bricked my version v2, shorting pins 15&16 of the flash causes the Diag led to flash, but it does not respond to pings on 192.168.1.1 I'll probably try the JTAG aproach but can't get the part till tomorrow, I'm looking for ideas/information
[12:41] <beattie> wrt54g-v2
[12:44] <hrzags11> Diag LED ?
[12:44] <beattie> sorry th epower
[12:44] <beattie> led
[12:44] <wbx> beattie: new machine? or what you have flashed?
[12:45] <hrzags11> Aha ... What happend if you reboot, without shorting a pins ?
[12:45] <beattie> I had a modified version of the linksys software about 6 mons old, tried to flash ewtr 0.3
[12:46] <beattie> seemd to work, but could not connevt after reset
[12:46] <wbx> ähem.
[12:46] <wbx> what is ewtr?
[12:46] <nbd> beattie: unplug, press and hold reset, plug in again, while holding reset for another 10-20 sec
[12:46] <nbd> beattie: maybe this helps
[12:46] <beattie> normal reboot the leds look normal, (power WLAN, LAN) but can not get any response
[12:46] <nbd> wbx: he means ewrt
[12:46] <beattie> tried that
[12:47] <wbx> nbd: oh. wrong channel i think.
[12:47] <beattie> right ewrt
[12:47] <nbd> beattie: so tftp doesn't work either?
[12:47] <beattie> nope
[12:47] <beattie> and no arp entry
[12:48] <nbd> can you find the device over wlan?
[12:48] <beattie> kismet does not show it
[12:49] <beattie> just my functional one
[12:49] <nbd> well.. don't flash ewrt next time :)
[12:49] <beattie> good advice :)
[12:49] <nbd> is there even anyone still working on this thing?
[12:50] <hrzags11> nbd: FYI i did web interface for ppoe :) ... and work nice :) ... wasnt to hard to do it
[12:50] <nbd> k
[12:53] <nbd> beattie: are you sure that you shorted the right pins for your device?
[12:53] <beattie> yep did it a number of times
[12:54] <beattie> I also saw mention of 18&19 20&19 and tried that too
[12:54] <nbd> if you get it right, the dmz light should blink very slowly
[12:54] <Dabian> Is there a package for a webinterface? (J/W)
[12:55] <hrzags11> Debian : It should be very soon :)
[12:55] <beattie> well I'm not getting that, my power light keeps blinking, and no WLAN juts LAN
[12:55] <Dabian> OK .. just wondering about it, since there is a default /www/index.html talking about the lack of a package. :-)
[12:55] <nbd> beattie: maybe you don't need jtag, maybe a serial console will be enough
[12:56] <beattie> jtag would be easier, no IC needed
[12:56] <hrzags11> Debian: I just create partialy web pages using standard httpd, to help user using some standard function like pppoe
[12:56] <Dabian> I forgot where the config for the wondershaper package is.
[12:56] Action: fishhead decides that if his government decides to use a very weak allogation that terrorists are using piracy to make $$$$, and as a result they pass copyright laws that ban reverse engineering for interoperability, ban knowing the hardware details of hardware for propose of software developement, and make repairing and upgrading pc's illegal, then he is going to WAR with his governement
[12:57] <nbd> fishhead: good idea
[12:57] <synth> SUV drivers support terrorism
[12:57] <beattie> yeah but they support the Oil companies too, or do I repeat myself
[12:57] <synth> the white house spends MILLIONS on forcing drug war propaganda down our throats, even when mountains of research show its for naught
[12:58] <synth> the terrorism <-> piracy thing is just the latest
[12:58] <synth> they'll link terrorism to anything they want to f over today
[12:58] <Dabian> I havn't heard of that?
[12:58] <synth> so i say: SUV drivers support terrorism! rar
[12:58] <beattie> just another way to funnel money to corporate interests
[12:59] <synth> anyways while im here - where does udhcpd get started from?
[12:59] <synth> i've grepped through /etc and found nothing
[12:59] <synth> what puts resolv.conf in /etc ? udhcpd writes it to /tmp
[12:59] <nbd> we don't use udhcpd at the moment
[12:59] <nbd> we use dnsmasq
[13:00] <synth> uhh, when i ifup i see it and this is a untar and make of the experimental tar
[13:00] <synth> i have done no mods :D
[13:00] <crashev> nbd: are you going to switch to udhcpd?
[13:00] <synth> ifup on wan
[13:00] <synth> its using it for wan, but yes dnsmasq is there for lan
[13:00] <nbd> crashev: should we? there are no plans for doing so :)
[13:00] <nbd> synth: we have udhcpc, but not udhcpd
[13:01] <synth> err, sorry
[13:01] <synth> i do that with that program too much.
[13:01] <crashev> nbd: i suppose u shouldnt thats why I wonder u wrote "we dont use.. at the moment" :)
[13:01] <synth> ahhh
[13:01] <synth> so you guys use bk huh.. i need to apply some patches :D
[13:01] <nbd> crashev: well... we'll stay with dnsmasq :)
[13:02] <nbd> synth: send them to me :)
[13:02] <hrzags11> crashev: Do you know how exactly dnsmasq work ? :) ... I have one problem
[13:02] <beattie> nbd how long should I need to keep the pins shorted?
[13:02] <crashev> hrzags11: not exactly - havnt looked into sources :)
[13:02] <crashev> but I configured it couple times
[13:02] <crashev> hrzags11: just ask
[13:02] <nbd> beattie: only a few seconds max. guess
[13:03] <synth> nbd: i see the docs for the init.d standards, is there any docs for /etc/ itself ? so things like this can be configured, ie "i want to reference dnsmasq as my primary dns and then have udhcpd write its resolv elsewhere"
[13:03] <synth> maybe have /sbin/net source a config or something?
[13:03] <synth> that includes udhcpc setup
[13:03] <hrzags11> crashev: pppoe connection, i got ip address of ISP DNS , router work nice, but clients on WiFi can resolv domain name (used 192.168.1.1)
[13:04] <synth> (notice im just mashing the letters, not trying to spell it)
[13:04] <synth> heh
[13:04] <crashev> hrzags11: its nasty solution,but try to run it as root
[13:05] <crashev> user=root or similar in config file,dont remember exactly
[13:05] <crashev> and restart dnsmasq
[13:05] <hrzags11> crashev: ups may I know way ? ... or beter question way is nobody by default ?
[13:06] <crashev> nobody is always there where its possible to drop priviliges - for security reasons
[13:06] <crashev> but I suppose in this case there are some problems
[13:06] <hrzags11> creshev: I understand way root , it is obviusly, resolve conf is in root ownership
[13:07] <crashev> hrzags11: not really resolv.conf should be world-readable
[13:07] <crashev> :)
[13:07] <hrzags11> creshev: Thanks, I will try, one more question must I restart dnsmasq every time when resolve.conf change ?
[13:07] <crashev> hrzags11: uhh Im not sure, but probably yes,I think it loads entries from resolv.conf on startup
[13:08] <crashev> and on the other hand how ofter do u change your resolv.conf?! ;]
[13:08] <crashev> I usually set it up once a system install
[13:08] <hrzags11> crashev: I dont change it at all , my pppoe chang it, and ISP almost with every new connection change DNS entry :)
[13:09] <hrzags11> crashev: I know that isnt good ISP, but it is only on market :)
[13:09] <Dabian> Where can I find out how to configure wondershaper?
[13:10] <crashev> thats very strange that they change DNSes u use so often - there is crontab u can always make some script to restart dnsmasq once a day or similar :)
[13:12] <hrzags11> crashev: it is pppoe script for starting/stopin pppoe connectio, and I will put dnsmasq restart in it
[13:12] <synth> ahh i see
[13:12] <synth> smart shellscript ;)
[13:13] <hrzags11> synth : Just upgreat for standard S50pppoe :)
[13:13] <beattie> Thanks for the info
[13:14] <jopi> Dabian: in wondershaper
[13:14] <jopi> its a shellscrips
[13:14] <Dabian> yeah
[13:14] <Dabian> in /etc/init.d
[13:14] <Dabian> S65
[13:15] <Dabian> btw
[13:15] <Dabian> does a /real/ less really take up that much space on the wrt?
[13:17] <jopi> didn´t understand, sorry
[13:17] <jopi> less program, you mean ?
[13:17] <jopi> lol
[13:17] <Dabian> yeah
[13:17] <Dabian> of course
[13:17] <jopi> compile it yourself in busybox, if you really need it
[13:17] <Dabian> jopi: Are you english?
[13:18] <jopi> no, I´m catalan
[13:18] <coder> he is spanish
[13:18] <Dabian> ok :)
[13:18] <jopi> coder: more or less
[13:18] <Dabian> coder: Catalan not really Spanish .. they have their own language.
[13:18] <coder> no se por qué sabia que ibas a tocar los cojones con eso
[13:18] <jopi> lol
[13:18] <synth> cojones!
[13:18] <coder> Dabian are you going to explain me what catalan is? I live in spain and I speak catalan
[13:19] <Dabian> coder: In that case it would be in vain :)
[13:19] <Dabian> coder: I didn't know you were spanish/catalan. :)
[13:19] <jopi> never mind, that´s an offtopic in this channel
[13:19] <coder> catalunya is spain wether you like it or not
[13:19] <Dabian> True
[13:19] <jopi> coder: as of now
[13:19] <jopi> ^^
[13:19] <coder> as of now, right
[13:20] <coder> pero con actuaciones como las del borracho y de carod, por mucho tiempo xD
[13:20] Action: fishhead is going to turn into that guy from the movie "mystery men" and design a whole ton of non-lethal weapons
[13:20] <Dabian> catalunya wants to seperate? (Or some wants?)
[13:20] <fishhead> CANNED TORNADO!
[13:20] <coder> catalunya wants to separate
[13:20] <Tornado> hey, don't can me!
[13:21] <jopi> yes, a growing percent (but minoritary) of population
[13:21] <fishhead> heh
[13:21] <Dabian> Do you know how to set up wondershaper for SIP VoIP phone?
[13:21] <Dabian> I have a Sippura
[13:21] <jopi> Dabian: personnaly I don´t recommend wondershaper
[13:22] <jopi> I use my own qos scripts
[13:22] <Dabian> What do you recommend instead?
[13:22] <jopi> using htb is more accurate than cbq
[13:22] <Dabian> Well, wondershaper is just a script, right?
[13:22] <Dabian> htb?
[13:22] <Dabian> token buffer
[13:22] <Dabian> bucket
[13:22] <coder> jopi tienes mal la terminal, no tienes utf ni tp iso
[13:22] <coder> es raro
[13:22] <Dabian> hirakial?
[13:22] <jopi> coder: iso8859-15
[13:23] <jopi> hierarchical tocken bucket
[13:23] <coder> seguro?
[13:23] <jopi> coder: espera, que lo miro
[13:23] <Dabian> hierarchical token bucket :)
[13:23] <jopi> 8859-1
[13:23] <Dabian> jopi: I know token bucket a little .. very nice algorithm.
[13:23] <coder> 23:24 < jopi> Dabian: personnaly I don´t recommend wondershaper
[13:23] <Dabian> jopi: I can imagine what htb is ..
[13:23] <coder> lo digo por eso
[13:24] <jopi> por ?
[13:24] <coder> el apostrofo
[13:24] <jopi> ´`' ?
[13:24] <jopi> le di a uno cualquiera
[13:24] <coder> ``'''
[13:24] <jopi> o no lo ves?
[13:24] <Dabian> I have 8859-1 also.
[13:24] <coder> it does appear a Z with an accent
[13:24] <jopi> Dabian: do you see an apostroph ?
[13:24] <jopi> ´
[13:25] <Dabian> jopi: How do I learn to use it?
[13:25] <jopi> ?
[13:25] Action: fishhead agrovates his hital hernia and passes out on the floor
[13:25] <coder> wondershaper?
[13:25] <Dabian> jopi: yes .. a
[13:25] <Dabian> I cannot make one myself ..
[13:25] <jopi> Dabian: the best tutotiral I have seen is in lartc.org
[13:25] <jopi> in the recipes secton
[13:25] <Dabian> ahh yes
[13:25] <Dabian> recipes ok
[13:25] <jopi> is from a good friend, pedro larroy
[13:25] <Dabian> darn thing that my printer is broken
[13:25] <jopi> the recpie called "a complete nat installation"
[13:26] <Dabian> I need to fix it so I can print this stuff.
[13:26] <jopi> o something simillar
[13:26] <Dabian> Whats cbq ?
[13:26] <jopi> class based queing
[13:26] <Dabian> hmm
[13:26] <Dabian> don't ring any bells.
[13:26] <jopi> another qos algorithm
[13:27] <Dabian> Yeah
[13:27] <Dabian> but I don't know it.
[13:27] <Dabian> jopi: Which device do you have?
[13:27] <jopi> well, it´s also explained in lart
[13:28] <jopi> Dabian: sorry ?
[13:28] <jopi> repeat your last sentence, please
[13:28] <Dabian> jopi: Which device do you have?
[13:29] <Dabian> A wrt54g, an ASUS WL-500g deluxe, or something else?
[13:29] <jopi> wrt54g, wrt54gs, buffalo wbr-g54 and wbr-g54-2
[13:29] <Dabian> ok :)
[13:32] <jopi> Dabian: i have "debridged" my linksys
[13:32] <Dabian> I have an ASUS wl-500gd
[13:32] <Dabian> I bridge
[13:32] <Dabian> wan with lan
[13:32] <Dabian> I have turned off the radio.
[13:33] <Dabian> But I wonder what the ports are called.
[13:33] <Dabian> vlan0 and vlan1
[13:33] <jopi> I can send you my script
[13:33] <Dabian> cool
[13:33] <Dabian> I need a different client
[13:34] Action: leOn wants a wl500gd too
[13:35] <DabianII> Hey
[13:36] <jopi> DabianII: hi again
[13:36] <Dabian> jopi: That client can take dcc.
[13:36] <jopi> btw, where did you purchase your asus ?
[13:36] <lqb> one questionnnn
[13:36] <DabianII> mail order
[13:36] <DabianII> edbpriser.dk
[13:36] <DabianII> pc-netto.dk
[13:37] <lqb> Are there airplay precompiled for openwrt?
[13:37] <jopi> Dabian: yes, my client can, but my computer with the scripts is off now, I can´t connect it till tomorrow
[13:37] <DabianII> ahh ok
[13:38] <jopi> give your email in a private
[13:44] <LiquidNerd> has anybody seen this problem before? I have iptables forwarding port 80 from the router to a server. When connecting from the internet to the server, it works perfectly. But if you're behind the router, and try to connect to the server with the internet IP, iptables wont forward it, but it will return it back to the IP making the request
[13:45] <iwtf> LiquidNerd, theres something you need to do about the masqurading, can remeber what
[13:46] <LiquidNerd> hmm
[13:49] <frop> LiquidNerd: are you with a static or a dynamic ip?
[13:50] <LiquidNerd> pretty much static
[13:50] <LiquidNerd> it's dynamic, but changes about.. once or twice a year
[13:50] <LiquidNerd> last year it didn't change at all
[13:51] <frop> iptables -t nat -A POSTROUTING -o ppp0 -s your.lan.net.internal/dontknow -j MASQUERADE
[13:52] <frop> it should be something like this then
[13:52] <LiquidNerd> -s should be my external?
[13:53] <frop> ohhh
[13:53] <frop> wait
[13:53] <frop> wrong channel
[13:53] <frop> are you using openwrt...
[13:53] <LiquidNerd> yeah
[13:54] <frop> ...well, it should work with default
[13:54] <frop> then
[13:54] <LiquidNerd> that's what I thought too
[13:54] <LiquidNerd> lol
[13:55] <frop> there're already ### MASQ rules in /etc/init.d/S45firewall
[14:05] <lqb> anyone Know if exist aireplay for openwrt?
[14:06] <LiquidNerd> MASQUERADE all -- any vlan1 anywhere anywhere
[14:07] <LiquidNerd> hmmm... that should change my outgoing IP to my external IP address rather than 192,168.1.100 should it no
[14:08] <Eagle_Fire> i think i mighta found the second serial port on my WR850G
[14:11] <frop> Eagle_Fire: i've neiter soldered the first...
[14:11] <frop> there's no space to place it in a "beauty" way
[14:12] <lqb> I'm searching aireplay for my openwrt :( any one know something?
[14:13] <jopi> I think you cannot use it
[14:13] <jopi> "I think"
[14:13] <Eagle_Fire> correct
[14:13] <Eagle_Fire> openwrt is not currently capable of packet injection
[14:14] <jopi> as it would need access to some management frame capabilites wrt driver doesn´t have
[14:14] <lqb> no????????
[14:15] <lqb> :((((((((
[14:15] <lqb> thanks....
[14:16] <Eagle_Fire> sweet
[14:16] <pp_> you could technically plug in a minipci prism2.5 into a v1.0 and use that ;)
[14:16] <pp_> but that's really theoretical ;)
[14:16] <Eagle_Fire> there *is* a second serial port on the WR850GV2
[14:16] <Eagle_Fire> woot woot
[14:17] <lqb> packet injection imposible in wrt54g or imposible in any kind of prism chip hardware?
[14:17] <Eagle_Fire> prism chip hardware *does* support injection
[14:18] <Eagle_Fire> wrt has broadcom hardware
[14:18] <Eagle_Fire> there's a project going on to get packet injection working in wrt
[14:18] <lqb> ajam......
[14:18] <lqb> Eagle_Fire what's project?
[14:18] <lqb> what's page
[14:18] <lqb> ?
[14:18] <lqb> please... xD
[14:19] <Eagle_Fire> no page
[14:19] <Eagle_Fire> it's a private project
[14:19] <lqb> Eagle_Fire and can you say me any mail or something for contact with this persons?
[14:19] <Eagle_Fire> i kind of started it
[14:19] <Eagle_Fire> there's no results as of yet
[14:20] <Eagle_Fire> but we're working on it
[14:20] <lqb> ajam......
[14:20] <lqb> Eagle_Fire do you think in the future you'll get it?
[14:26] <lqb> Eagle_Fire If I can help you in something you only has say it........
[14:27] <Eagle_Fire> we've got it pretty much figured out
[14:27] <Eagle_Fire> just give it a week or two
[14:28] <nbd> heh... this is weird:
[14:28] <nbd> performance in current experimental with the new code is not as good as it should be
[14:28] <nbd> but somehow b44 manages to be faster in bridging mode
[14:28] <nbd> while some people told me that it was slower before
[14:29] <Eagle_Fire> sounds pretty subjective anyways
[14:29] <Eagle_Fire> have you run a double-blind trial
[14:29] <lqb> Eagle_Fire So in two weeks you can run aireplay in wrt? jeje
[14:29] <nbd> i tested it using a scp from wlan to lan, i compared the average speed
[14:29] <nbd> speed with et: 2,1 mb/s
[14:30] <nbd> speed with b44: 2,3-2,4 mb/s
[14:30] <Eagle_Fire> lqb: no promises
[14:30] <Eagle_Fire> until then you'll have to crack your neighbor's wep the old-fashioned way
[14:30] <lqb> Eagle_Fire jejeje ok ok.....
[14:30] <lqb> yeah.... but there aren't any IF in my air space dfsdfsdf!!
[14:30] <lqb> *IV
[14:31] <lqb> 20IV maybe.....
[14:31] <lqb> jeje
[14:32] Action: ralf is away ( http://www.autistici.org.nyud.net:8090/shockraver/ph0t0.htm)
[14:35] <jiauka> wbx, i have the dissam for tiatm.o
[14:37] <lqb> I wish you'll get it.... for your father!! xDDD
[14:39] <Eagle_Fire> that's not the kind of thing that's acceptable here, lqb
[14:39] <lqb> here where? in this channel?
[14:41] <Death_INC> wtf is he attempting to talk about?
[14:41] <wbx> jiauka: hmm.
[14:41] Action: Death_INC cant figure out why it would be acceptable or even why not
[14:41] <wbx> jiauka: LINUX started... but then nothing more. any idea why printk does not work? :}
[14:41] <lqb> Death_INC use aireplay in wrt54g
[14:42] <jiauka> quick & dirty dissam, but could somewhat useful
[14:42] <Death_INC> oh hey has anybody actually gotten any of the MIPS emulators to work with kernel and fs etc?
[14:42] <jiauka> you should compile the serial console on the kernel
[14:43] <nbd> wbx: did you read my stuff above?
[14:43] <wbx> jiauka: did you have any .config working for 2.4.17 i can look at?
[14:43] <wbx> nbd: performance on b44.
[14:43] <jiauka> and have the /dev device on the filesystem
[14:43] <wbx> nbd: very strange.
[14:43] <jiauka> yes
[14:43] <nbd> wbx: and i wonder why the performance of experimental sucks that way
[14:43] <nbd> wbx: it's not the new system code. last snapshot that i have on my wrt is just as slow
[14:44] <nbd> something is very wrong
[14:44] <wbx> nbd: are you sure it is faster anytime before?
[14:44] <asmodemon> nbd try a ping from the wrt to its own ip
[14:45] <jiauka> wbx: i have to reboot into the linux part, back in 2 min.
[14:45] <nbd> wbx: i don't know if it was in experimental
[14:45] <nbd> wbx: but i know when i first got the device, it could