OpenWrt

[00:07] <RItalMan> hi all
[00:09] <[mbm]> .
[00:10] <RItalMan> I want to use my openwrt as a nas as I have already set up a radius
[00:10] <RItalMan> I'm wondering if I need to provide a wep key for the first time ?
[00:11] <[mbm]> donno, never messed with that aspect of it
[00:12] <RItalMan> I used to have a dlink dwl900+ and even if the radius provides a dynamic wep key, this access point needed a wep key
[00:13] <RItalMan> any news about a netbsd version ?
[00:15] <[mbm]> haha
[00:16] <RItalMan> really ;) ?
[00:17] <[mbm]> what do you think?
[00:17] <RItalMan> well few weeks ago I read someone posting here there was netbsd version planned by a people from sveasoft
[00:18] <[mbm]> they've been saying that for the last 6 months
[00:18] <RItalMan> wooa I should come here more often
[00:19] <RItalMan> and are they working about ?
[00:19] <[mbm]> there's more to it then just compiling netbsd for the mips .. not impossible given the right people but I doubt they'll get anywhere, atleast not any time in the next 6 months
[00:20] <[mbm]> aside from os wars, there really isn't much of a point
[00:21] <RItalMan> isn't there a problem using the broadcom drivers ? as they are provided just for linux ?
[00:21] <[mbm]> no sources for the wireless
[00:21] <RItalMan> reverse engineering ?
[00:22] <[mbm]> ...
[00:22] <[mbm]> again, what's the point of the exercise?
[00:22] <RItalMan> provide a driver for any linux or bsd versions
[00:23] <[mbm]> no, I mean .. what difference does it make if it runs linux or netbsd?
[00:23] <RItalMan> there is absolutely no difference, and I think it's much more a trip than a real need
[00:24] <[mbm]> I figured it was just a cry for attention
[00:24] <RItalMan> sure
[00:24] <[mbm]> sveasoft isn't doing too well these days
[00:24] <RItalMan> do you think it's partly dued to openwrt ?
[00:25] <[mbm]> possibly, but don't forget the whole gpl debate, subscription fees and closed forum
[00:25] <[mbm]> they're trying to milk as much as they can now
[00:26] <RItalMan> what I don't understand is why at the beginning their firmwares were accessible and now they are closed to the public
[00:26] <[mbm]> greed?
[00:27] <[mbm]> last I heard they were tagging all the releases and even looking into ways that the firmware wouldn't run without some activation process
[00:28] <[mbm]> pay per router
[00:28] <[zombie]> I don't see what all the fuss is about. their firmware really isn't anything special
[00:28] <[mbm]> thats been my thought
[00:29] <RItalMan> the only one thing that is really common for the end user with their firmware is the web interface
[00:30] <[zombie]> there's other projects that are similar to sveasoft
[00:30] <[zombie]> and aren't run by a loon
[00:31] <RItalMan> what do you mean with loon ?
[00:31] <[zombie]> but it's a different target audience to openwrt. if you want a flashy gui (and happy to be restricted to what they squeezed in the firmware) run one of those projects
[00:31] <[mbm]> I dont understand why everyone's so keen on a web interface .. limits the capability
[00:31] <KarlNapf_> hi, playing around with nas I thin I figured out you won't need a wep key prior to auth provided the wlx_ is in open system mode (wlx_auth=0)
[00:31] <[zombie]> if you want to do anything you can put your mind to, run openwrt
[00:33] <RItalMan> everyone is keen on web interfaces as they are similar to the consumption of IT's (windows inhabits), everything must be simple, reliable, well done, accessible, powerful ... is that possible ? I don't think
[00:34] <[zombie]> RItalMan: by loon I mean this: http://wrt54g.atw.hu/
[00:34] <[mbm]> easy, powerful, stable .. pick any two
[00:34] <[zombie]> as I said, different target audiences
[00:34] <[zombie]> it's the usual GUI vs CLI argument
[00:35] <[zombie]> most people prefer GUI. those that know what they are doing and want real power choose CLI
[00:36] <[mbm]> every time I try to use a gui it treats me like an idiot
[00:37] <RItalMan> lol do not make it as a general case
[00:37] <[mbm]> don't suppose anyone else has noticed, but I've seen a number of openwrt users where after they change anything they reboot .. :)
[00:37] <RItalMan> sometimes might be required ;)
[00:38] <[mbm]> umm ... no
[00:38] <[mbm]> unless you're changing the firmware there's no rebot required
[00:38] <RItalMan> it's a joke of course
[00:38] Action: [mbm] gets the feeling those people are windows users
[00:38] <[mbm]> dunno why
[00:41] <RItalMan> I know it's a wide debate, and it's not the right place to talk about it here. Even if I'm using linux and openbsd I understand my little sisters who don't want to use it when playing a game as it is not really convenient playing Harry potter or something else using wine
[00:43] Action: [mbm] got a playstation for that .. easier to just insert a game and play without dealing with windows and driver issues
[00:44] <RItalMan> true even if some "(àç'(é' developpers thought "Let's code a really specialized application for windows"
[00:45] <RItalMan> well back to openwrt
[00:45] <RItalMan> I want to port reaim to the openwrt
[00:46] <RItalMan> so that I created a reaim.mk
[00:46] <RItalMan> when compiling it, it's still a x86 binary any idea ?
[00:46] <RItalMan> shall I send by mail to someone my reaim.mk file ?
[00:47] <[zombie]> don't create a .mk file, better off creating an ipkg/rules file
[00:47] <[zombie]> that way it will build you a nice ipk file you can install on an openwrt
[00:47] <[mbm]> nod. also sounds like you created the .mk wrong
[00:47] <[zombie]> rather than having to reflash it with your specific firmware
[00:47] <[zombie]> there's some great documentation for creating ipk files on the wiki
[00:48] <[zombie]> I figured out how to create packages in a couple of hours, and I don't have much coding or makefile-hacking experience
[00:48] <[mbm]> donno if the ipkg-buildpackage stuff is documented yet or not
[00:48] <[zombie]> it is
[00:48] <RItalMan> ok
[00:48] <coder> -.
[00:48] <[zombie]> not in the wiki directly I think, but in one of the forum posts that is linked to
[00:48] <RItalMan> well I really wanted to just compile the binary and then ipk it
[00:49] <[zombie]> yup, so drop the .mk file and read the docs on creating ipkg files :)
[00:49] <[mbm]> coder: ..-. ..-
[00:49] <coder> :)
[00:50] <RItalMan> ok let's do it right now ;)
[00:50] <RItalMan> you guys are really nice and patient, rare to find some on the ircs
[00:51] <coder> on the internets
[00:52] <RItalMan> I apologize for my english
[00:53] <coder> your english is good
[00:53] <RItalMan> thanks
[00:56] <RItalMan> there is a huge work putting on the wiki all the posted howtos on the forum
[00:56] <RItalMan> has anyone a "role" developping for openwrt ?
[00:56] <coder> RItalMan I think that the best thing to play games is a videoconsole nor windows or wine
[00:57] <coder> RItalMan [mbm] is the man
[00:59] <RItalMan> alone doing this huge work ?
[00:59] Action: [mbm] smacks coder .. now I'm gonna get pestered with questions about the process
[00:59] <RItalMan> it's a full time job ?
[01:00] <[mbm]> not a job -- not paid
[01:00] Action: coder will try to answer the questions
[01:01] <Ge0rG> how can I install windows on the wtr????
[01:01] <coder> RItalMan at the beginning it was a big bang, then linksys made the source available and groz and [mbm] started a new firmware
[01:01] <RItalMan> Ge0rG: type /quit and this should work ;)
[01:02] <Ge0rG> RItalMan: nah, tried that already :(
[01:02] <coder> Ge0rG do you know about the w2k stolen code?
[01:02] <RItalMan> it will be really interesting porting windows to the mips :D
[01:03] <Ge0rG> coder: got it right here ;)
[01:03] <coder> ok then you can compile paintbrush and launch it through wine on xorg-x11 using the wrt as platform
[01:03] <Ge0rG> WinCE had support for mips many years ago
[01:03] <RItalMan> finally I found the howto : http://www.openwrt.org/forum/viewtopic.php?t=107&highlight=ipk+howto
[01:05] <RItalMan> I have seen a sharp zaurus last week running linux it's really amazing
[01:05] <coder> still too expensive
[01:06] <RItalMan> yes but really fun to wardrive
[01:07] <RItalMan> coder: you told me previously to use ipkg instead of a .mk file, but I still need to compile my bin using a mk file ?
[01:07] <[zombie]> I've had linux on my iPAQ for a couple of years now.. I can't believe I ever put up with CE on that thing
[01:08] Action: coder still wonders what makes wardriving atractive
[01:09] <coder> RItalMan you can write a mk, put it in make/ and add the package to the TARGET so it will get compiled automagically
[01:09] <RItalMan> ok, but if I get a x86 binary then my mk file is wrong
[01:09] <[zombie]> but it won't build an ipkg that way. I thought we'd been through that :)
[01:10] <RItalMan> yes, currently I just need the binary to test it
[01:10] <RItalMan> then I would make my ipk available
[01:10] <[zombie]> you don't need a .mk to create your binary
[01:10] <[zombie]> just an ipkg/rules file
[01:10] <RItalMan> ah ok
[01:11] <RItalMan> and how do I proceed from source ?
[01:11] <RItalMan> because there is no configure script for the app reaim
[01:11] <[zombie]> get that ipkg.skel.tar.gz file mentioned on that howto
[01:11] <[zombie]> and use that to create your ipkg/rules and ipkg/control files
[01:12] <RItalMan> ok
[01:12] <Blackvel> morning all
[01:12] <RItalMan> morning Blackvel
[01:14] <[zombie]> edit those, and edit your /usr/local/bin/ipkg-build as described so it creates tar files not ar files
[01:15] <[zombie]> then make sure your cross-compile tools are in your PATH, and run ipkg-buildpackage -c
[01:16] <RItalMan> it can be possible to run a complete debian distro using mips if I understand the howto ?
[01:17] <RItalMan> ok and that's all ?
[01:17] <[zombie]> that's all
[01:17] <RItalMan> eh eh thanks
[01:17] <[zombie]> if I want to create a new package all I do is create a new rules file and then run ipkg-buildpackage
[01:17] <[zombie]> of course the creation of the rules file is the tricky bit :)
[01:17] <RItalMan> it's pretty fun
[01:17] <[zombie]> and you learn much about Makefile hacking in the process
[01:18] <RItalMan> much more than creating a debian package ;)
[01:18] <[zombie]> if the program uses ./configure and has support for cross-compilation then it's not so much a problem
[01:18] <[mbm]> makefiles really aren't that hard
[01:18] <[zombie]> but compiling (for example) asterisk was a bitch
[01:19] <RItalMan> ok, my program has no configure
[01:19] <[zombie]> no problem, just make sure it uses the right GCC and LD etc
[01:20] <RItalMan> ok
[01:20] <RItalMan> I just have to indicate the full path to my gcc in the buildroot dir ?
[01:22] <[zombie]> the ipkg.skel/rules file is pretty well set up
[01:22] <RItalMan> yes it is
[01:22] <[zombie]> TARGET_CC=$(TARGET_CROSS)gcc
[01:23] <Blackvel> morning mbm :)
[01:23] <[zombie]> you might be able to get away with just setting the vars up the top correctly
[01:23] <Blackvel> crosscompiling asterisk ?
[01:23] <Blackvel> ;)
[01:23] <[zombie]> Blackvel: it was an interesting excercise
[01:23] <Blackvel> i thought there was a ipkg out somewhere already ?
[01:23] <[zombie]> not that I could find
[01:23] <[zombie]> and I searched long and hard
[01:23] <Blackvel> how is the performance on the WRT ? :)
[01:24] <[zombie]> both before I started, and also while I was strugglign with it
[01:24] <[zombie]> well umm.. it was an interesting excercise with a not very satisfactory result... :)
[01:24] <[zombie]> performance is pretty bad
[01:24] <[zombie]> I expected that, but I wanted to try it anyway
[01:25] <Blackvel> ah, its bad ? had you an ATA connected to the asterisk then and tried to phone out ?
[01:25] <Blackvel> you and the other person got choppy sound ?
[01:25] <RItalMan> asterisk is a SIP server is that true ?
[01:27] <Blackvel> yes
[01:27] <Blackvel> that's true
[01:27] <Blackvel> but it is more
[01:27] <Blackvel> sip client and sip gateway too
[01:28] <RItalMan> ok so that it manages sip proxying ?
[01:28] <[zombie]> I didn't really play much with it. with tweaking it might improve
[01:28] <[zombie]> but I hooked it up to my main asterisk server via IAX2
[01:28] <[zombie]> and then pointed a SIP phone at the WRT and connected to the demo (1000) on the main server
[01:28] <[zombie]> and the sound was very very very choppy
[01:29] <Blackvel> too bad
[01:29] <Blackvel> was that a WRT54GS with 200mhz yet ?
[01:29] <[zombie]> yes
[01:29] <[zombie]> a GS
[01:30] <Blackvel> well to be true
[01:30] <Blackvel> i have a dedicated epia V server for asterisk
[01:30] <Blackvel> because i have also connected asterisk to my pbx with some interface card
[01:37] <RItalMan> has ipac-ng been ported ?
[01:39] <Blackvel> can be
[01:40] <Blackvel> saw that yet i believe
[01:40] <Blackvel> but had it support for a db ?
[01:41] <RItalMan> absolutely don't know
[01:44] <Blackvel> i setup my own solution
[01:44] <Blackvel> wrt sends out the accounting stuff to some collector
[01:45] <Blackvel> which collects the packets and writes them into some db
[01:45] <Blackvel> little bug still, but i move on to a new version :)
[01:58] <RItalMan> do you think it is possible to run a distcc on openwrt ? and use it to cross compile via nfs for x86 :D
[01:58] <RItalMan> ?
[01:59] <PolarWolf> Why?
[01:59] <RItalMan> just for fun
[01:59] <[zombie]> of course it is possible, it just would be very pointless
[01:59] <[zombie]> even more pointless than my asterisk experiment :)
[02:00] <RItalMan> lol sure
[02:00] <PolarWolf> You'll spend a lot of time getting it to work, try to compile something once, and than give up
[02:00] <RItalMan> I'm a professionnal of useless things ;) and stupid questions
[02:02] <warfare> hi there
[02:04] <warfare> I have a small question regarding vlans on linksys wrt54g.. Is it possible to define multiple ssid's on the wireless interface and map them to vlans on the ethernet interface, so they can be send by an 802.1q trunk to a cisco switch?
[02:05] <RItalMan> I don't think the 802.11provide multiple ssid support
[02:06] <warfare> bummer. So I have to shell out the money for a cisco aironet.
[02:07] <RItalMan> may be
[02:07] <RItalMan> cisco supports multiple ssid on only one wireless interface ?
[02:08] <warfare> Yes. And it allows mapping to vlans on the ethernet Interface which acts as a 802.1q trunk link.
[02:08] <RItalMan> really great
[02:09] <warfare> ok, I'd get about 3 or 4 wrt's for the price of one aironet, so I'd implement this just with a stack of wrts ;)
[02:09] <RItalMan> lol sure
[02:15] <HeliumGas> I wonder if hostapd supports multiple ssids
[02:25] <warfare> hostap can't as the prism firmware won't allow it. Just googled about it.
[02:29] <warfare> I wonder how cisco does it..
[02:42] <turbostaat> hi @ll
[02:43] <RItalMan> hi
[02:43] <RItalMan> I have successfully compiled and ran reaim !
[02:43] <turbostaat> i need some info on openwrt's mtd-partitioning (arch/mips/brcm-boards/bcm-947xx/setup.c
[02:44] <coder> it adds just a new partition, that's all
[02:45] <turbostaat> i've been reading it for three days now, but i cant figure out why its done the way it is
[02:45] <coder> what you don't understand?
[02:46] <turbostaat> for example pmon size & offset get set several times
[02:46] <turbostaat> line 243/244 and 292
[02:47] <turbostaat> but i cant figure out the difference between both statements, they seem to set the same value
[02:52] <coder> mmm, my src is different from yours
[02:52] <coder> 0"%s: Minix filesystem found at block %d\n",
[02:52] <coder> that's line 244
[02:53] <turbostaat> hm ok, im looking @ the first few lines in init_mtd_partitions
[02:54] <turbostaat> my code comes directly from wrt54g.2.07.1
[02:55] <coder> i'm playing with 09 etsi
[02:56] <turbostaat> hm, im using a four day old buildroot-snapshot, did u change these in the meantime ??
[02:59] <turbostaat> ok, im fetching 09 etsi, ill get back to u as soon as i got it.
[03:00] <turbostaat> hm its missing @ linksys gpl-center
[03:00] <turbostaat> coder, u still here ?
[03:04] <JoeJoe> hi, what's the maximum size of the image file for Linksys wrt54g?
[03:05] <JoeJoe> df on the linksys shows :
[03:05] <JoeJoe> Filesystem 1k-blocks Used Available Use% Mounted on
[03:05] <JoeJoe> /dev/root 896 896 0 100% /rom
[03:05] <JoeJoe> /dev/mtdblock/4 2240 2108 132 94% /
[03:05] <turbostaat> rom is always 100%, i dont think there is any "maximum size"
[03:06] <JoeJoe> hm could the image 3MB ? so it is only 1MB for jffs2?
[03:06] <JoeJoe> is that possible?
[03:07] <turbostaat> sure, afaik there are no limitations regarding size
[03:08] <[zombie]> the G has 4MB flash total
[03:08] <JoeJoe> ok... next week i get a new wrt54g and than i will test it
[03:08] <[zombie]> once you take into account the size of the kernel, and the size of the nvram, what you are left with is the 3MB you see above
[03:21] <PolarWolf> JoeJoe: The maximum image size is ~1.5MB which you can flash, iirc
[03:21] <PolarWolf> JoeJoe: Something in that area anyway.
[03:42] <turbostaat> if i somehow destroy all mtd-partitions except pmon, is it possible to revive the wrt54g ?
[03:43] <turbostaat> (or, is it still possible ...)
[04:16] <Blackvel> all mtd ?
[04:16] <Blackvel> dunno
[04:16] <Blackvel> if it is the data partition, maybe firstboot ? but when you have erased everything including system files (rom) you are lost
[04:18] <turbostaat> so the tftp depends on rootfs and linux-kernel ?
[04:21] <Blackvel> tftp includes the rom image
[04:21] <Blackvel> which writes pmon + rom files (hey i am no expert on this topic!)
[04:21] <turbostaat> :)
[04:21] <Blackvel> re-tftp image and you can boot
[04:22] <Blackvel> it may happen that you corrupt existing flashfs fs
[04:22] <turbostaat> thats what i hoped to be true, like, as long as pmon is still there, i can tftp a new image up
[04:27] <burton27_> hi, can anyone suggest me what to use bridge or ip route between three wrt54g's? what to use for better performance?
[04:28] <[zombie]> routing is better than bridging
[04:47] <Dave`> does anyone know where I could get hold of the ipkg-utils package?
[04:47] <Dave`> All websites I go to seem to be down
[05:05] <phil> jon_kare: still there?
[05:14] <jopi> hi
[05:23] <jon_kare> phil: yes, back now
[05:45] <Blackvel> who is a networking / ip specialist and can tell me something about the MTU ?
[05:45] <turbostaat> im not a specialist but the mtu sets the maximum packet size
[05:46] <turbostaat> maximum transfer unit i think
[05:46] <Blackvel> Do you know QOS a bit ?
[05:46] <Blackvel> Do you know VOIP ?
[05:47] <turbostaat> omg been messing with that few years ago
[05:47] <turbostaat> what do u want to know ?
[05:47] <Blackvel> what happens if i have voip + qos active and a p2p application sends tcp data ?
[05:47] <Blackvel> what does mtu 1500 or 1492 for adsl mean ?
[05:48] <Blackvel> can it be that a tcp data packet uses my full adsl uplink just because of mtu ?
[05:48] <turbostaat> mtu has nothing to do with voip/qos. mtu 1492 is the maximum packet size for adsl connections (1492 + 8 byte pppoe-capsulation = 1500)
[05:50] <turbostaat> if you set mtu to higher values than 1492, packets will eventually become fragmented, and because of security issues in the past a lot of routers drop fragmented or oversized packets
[05:50] <Blackvel> so there is no calculation like this : 1500 MTU * X = 256kbit ?
[05:51] <turbostaat> every electronically transmitted data packet uses the full bandwith. u cant transfer two packets at the same time.
[05:51] <jon_kare> You can solve that equation, but it doesn't mean anything.
[05:51] <jon_kare> If ur MTU is too high, some web pages will load excriatingly slowly or not at all.
[05:52] <Blackvel> so it may happen that a p2p tcp packet makes use of the full uplink bandwidth so with even QOS the voip packet can not be sent out immediately ?
[05:52] <turbostaat> MTU has nothing to do with kbit at all. MTU is just a size, and 256kbit is a transfer rate
[05:53] <turbostaat> qos just handles the way packets are processed / prioritized. if there is a packet beeing transferred, every other packet has to wait
[05:53] <Blackvel> i see
[05:53] <Blackvel> lets think about this then
[05:53] <Blackvel> qos sends a p2p packet
[05:54] <Blackvel> which makes use of the total or half uplink bandwidth
[05:54] <turbostaat> no no no
[05:54] <turbostaat> no-go
[05:54] <Blackvel> there would be one voip packet which should be sent out (by qos)
[05:54] <turbostaat> stop blackvel
[05:54] <Blackvel> but qos simply can't prioritize it because a p2p tcp packet is already on the way out
[05:54] <Blackvel> that p2p packet is processed already
[05:54] <Blackvel> can there be a delay then ?
[05:55] <Blackvel> hmm i should stop ? why ? :)
[05:55] <turbostaat> 1st there is no such thing as a comparision of packet & bandwith
[05:55] <turbostaat> a packet always uses the whole adsl link
[05:56] <turbostaat> so, regarding your example (lets assume qos-rules set accordingly):
[05:56] <Blackvel> it does ? e.g i know voip is only max of 90kbit
[05:56] <Blackvel> right
[05:56] <turbostaat> thats for connections, not for packets
[05:56] <turbostaat> ok,
[05:56] <Blackvel> look slike i am a tcp networking noob
[05:56] <Blackvel> please go on :)
[05:57] <turbostaat> you have a p2p connection ongoing
[05:57] <turbostaat> so the qos-rule for this connection says minimum priority, maximum bandwith
[05:58] <turbostaat> now you recieve a voip-packet, processed by "minimize delay"
[05:59] <turbostaat> then qos should handle the voip-packet instantly, meaning it will get transferred/routed as soon as the line is clear.
[05:59] <Blackvel> right that should be that way
[05:59] <Blackvel> and it can be done ?
[05:59] <turbostaat> so, in worst case the packet would have to wait for 1499 bytes to be transferred, plus maybe 1 waitbyte
[05:59] <Blackvel> what happens if the p2p packet is on its way ? it cant be stopped anymore and be replaced by the voip packet ?
[05:59] <jon_kare> you want to slow down the incoming p2p. the only way to do that is to not send back acks
[06:00] <jon_kare> so fast.
[06:00] <Blackvel> nonope
[06:00] <Blackvel> i am not talking about the downlink but uplink!
[06:00] <jon_kare> I don't know what mechanism you have to do that, though.
[06:00] <jon_kare> incoming p2p = uplink, right?
[06:00] <Blackvel> turbostaat why are there 1500 ?
[06:01] <Blackvel> because of mtu ?
[06:01] <Blackvel> hehe hopefully noone laughs in this channel :)
[06:01] <turbostaat> yes, well, i forgot ethernet capsulation in that equation, but that would be 8 byte or so more
[06:01] <jon_kare> it's a magic number decided long long ago by our ancestors
[06:01] <turbostaat> lol
[06:01] <jon_kare> Who thought 1500 bytes was a big packet.
[06:02] <turbostaat> (uh i laughed :) )
[06:02] <jon_kare> That was when T1 was a backbone link.
[06:02] <Blackvel> ok
[06:02] <Blackvel> that 1500 has nothing to do with the MTU
[06:02] <turbostaat> actually long long ago there were and still are packets with sizes over 1000k
[06:02] <Blackvel> i understand
[06:02] <Blackvel> so its the standard to send out 1500 bytes ?
[06:03] <Blackvel> yes i understand
[06:03] <Blackvel> that the voip packet would have to wait a little bit until the p2p packet of size 1500+overhead was sent out
[06:03] <Blackvel> the question I ask myself is
[06:03] <turbostaat> yes
[06:03] <Blackvel> will i notice that delay in anyway ? i mean it is linux and somehow the kernel and tcp stuff magically does that behind the scene :)
[06:04] <Blackvel> do i have to care about this ?
[06:04] <Blackvel> or comming back to my original question
[06:05] <Blackvel> can it happen that with p2p+voip running both together, that this works worser on a 128k/196k/256k uplink as a 384/512 uplink would do ?
[06:05] <Blackvel> for all of these scenarios QOS would be turned on (and would also be configured apropriately)
[06:06] <turbostaat> no, you wouldnt notice any stuttering or something like that, but there may be a slightly increased delay if your upload is permanetly cranced up to maximum
[06:07] <turbostaat> depending on implementation that can range from , dunno, 10us to 1/2 s
[06:07] <AzzIzzA> see if your isp will provide qos as well
[06:07] <turbostaat> it shouldn't be noticeable over the overall delay voip has by itself
[06:07] <AzzIzzA> that would make things flow a bit smoother
[06:08] <turbostaat> i think he is talking about his very own router pushing packets to his isp (upload).
[06:08] <Blackvel> AzzIzzA: as i said, uplink!
[06:08] <Blackvel> not downlink
[06:08] <AzzIzzA> i know
[06:09] <AzzIzzA> but voip is typically a two way thing
[06:09] <AzzIzzA> and if your leeching it would be nice if the voice traffic took priority
[06:09] <turbostaat> ok, yes, i was assuming that he uses the voip-service provided by his isp
[06:10] <AzzIzzA> oh righto
[06:10] <Blackvel> dunno how well my isp implemented qos for his voip
[06:10] <Blackvel> i dont care
[06:10] <AzzIzzA> thats not so common over here
[06:10] <AzzIzzA> although it is taking off
[06:10] <Blackvel> but i care much about my own router ;)
[06:10] <AzzIzzA> Blackvel: yes but how your isp treats it will effect you
[06:10] <AzzIzzA> it doesnt matter how great your router is
[06:11] <AzzIzzA> if your isp isnt prioritising what they send you
[06:11] <turbostaat> blackvel, ur going to use arcor VOiP service ?
[06:11] <AzzIzzA> you wont get the best realtime performance
[06:12] <sleon> hi
[06:12] <Blackvel> no i dont
[06:12] <sleon> i have a question to senao cards
[06:12] <AzzIzzA> i had qos on my router but when i start leeching, you can feel it over ssh
[06:12] <Blackvel> i use other providers
[06:12] <sleon> can i set the power lower
[06:12] <Blackvel> hmm leeching means ACK packets back on uplink ?
[06:12] <sleon> is it possible?
[06:13] <sleon> are there 54mbts senao cards?
[06:13] <sleon> i ask it here cause ppl here suggested me to use senao cards cause they have best signal
[06:14] <turbostaat> well, leeching usually means cranking up full up & download speeds
[06:14] <AzzIzzA> yes if your leeching your also sending acks upstream, if your using a tcp connection and you max out your downstream link so acks dont reach your quickly enough, windowing will kick in and slowdown your uploads
[06:14] <turbostaat> AzzIzzA is right, that if your isp doesnt support prioritized traffic shaping, you will notice delay a lot
[06:15] <sleon> hallo???
[06:15] <sleon> :)
[06:15] <sleon> do you notice me?
[06:15] <sleon> leeching?
[06:15] <turbostaat> (sleon): ive got no idea
[06:15] <AzzIzzA> which is why im trying to drive home the point that uploads and downloads arent two differnt things
[06:15] <AzzIzzA> they are very closely related
[06:15] <AzzIzzA> sleon: dont know
[06:16] <sleon> how can i reduce power of a card?
[06:17] <AzzIzzA> iwpriv maybe, with some drivers
[06:17] <AzzIzzA> why do you want to reduce the power?
[06:17] <Blackvel> uhhh
[06:17] <Blackvel> i get lost
[06:17] <Blackvel> leeching means sending ACK back on uplink
[06:17] <Blackvel> i understand this well
[06:18] <turbostaat> (Blackvel): point is, if you cant upload, u cant download and otherwise.
[06:18] <Blackvel> that is the reason i would want to set uplink to 80%
[06:18] <Blackvel> and prioritze ACK packets in QOS
[06:18] <AzzIzzA> that will help to maximise your downloads
[06:18] <turbostaat> why prioritize ack ? voip uses udp
[06:18] <Blackvel> but getting back what you have said earlier turbostat, do i have also to resize ACK packets on the downlink which may be maxed out ?
[06:18] <AzzIzzA> the tcp stack has this thing called windowing
[06:19] <AzzIzzA> oh
[06:19] <AzzIzzA> k in that case windowing is less of an issue
[06:19] <AzzIzzA> udp seems logical i guess
[06:19] <Blackvel> i understand when my downlink is filled, that i cant resize any ACK packet in time
[06:19] <Blackvel> but what does that mean to me ?
[06:19] <Blackvel> right
[06:20] <AzzIzzA> nothing
[06:20] <AzzIzzA> well
[06:20] <turbostaat> nothing i guess
[06:20] <turbostaat> :)
[06:20] <AzzIzzA> not much as turbostaat says its udp
[06:20] <Blackvel> it means my uplink slows down (for ftp e.g ) ?
[06:20] <AzzIzzA> for tcp connections yes
[06:20] <AzzIzzA> you cant have more than
[06:20] <turbostaat> ok, resizing doesnt happen to ack packets, they usually asre way smaller then 1500 bytes
[06:20] <AzzIzzA> x packets in transit
[06:20] <AzzIzzA> if you dont recieve an ack or nack
[06:20] <AzzIzzA> it counts as in transit
[06:21] <Blackvel> is is useful to slow down downlink to 80% ?
[06:21] <AzzIzzA> when you hit x, it stops sending
[06:21] <Blackvel> will I really resize ACK packets then ?
[06:21] <Blackvel> bah
[06:21] <AzzIzzA> you cant slowdown your downstream link
[06:21] <Blackvel> receive i mean
[06:21] <AzzIzzA> and if you could you wouldnt want to
[06:21] <Blackvel> i know the wondershaper
[06:21] <AzzIzzA> the only reason you slow your upstream is to keep the buffer on the router rather than the modem
[06:22] <Blackvel> ingress polici
[06:22] <Blackvel> # filter *everything* to it (0.0.0.0/0), drop everything that's
[06:22] <Blackvel> # coming in too fast:
[06:22] <Blackvel> #tc filter add dev $DEV parent ffff: protocol ip prio 50 u32 match ip src \
[06:22] <Blackvel> # 0.0.0.0/0 police rate ${DOWNLINK}kbit burst 6k drop flowid :1
[06:22] <Blackvel> i have this turned off
[06:22] <Blackvel> i have not yet IMQ running
[06:23] <AzzIzzA> i dont know what the point of that is, maybe some sort of dos protection?
[06:23] <Blackvel> do you think that is useful (for my problems I talked about my uplink because it happens to ppl that the called party can't hear them not vice versa)
[06:23] <AzzIzzA> i dont think its relevant to qos stuff
[06:23] <Blackvel> k
[06:23] <Blackvel> hmm okay
[06:23] <turbostaat> yup its only DOS
[06:23] <AzzIzzA> what exactly is the problem you are experiencing?
[06:23] <Blackvel> so the answer to my question is
[06:24] <turbostaat> (AzzIzzA): purely theoretical discussion so far
[06:24] <Blackvel> if i have turned on, a 128kbit uplink should behave like a 256 or 384k uplink ?
[06:24] <AzzIzzA> righto
[06:24] <Blackvel> i mean with qos
[06:24] <AzzIzzA> um...not really
[06:24] <Blackvel> if qos is turned off, 384 will be probably better than 128 ;)
[06:24] <Blackvel> hm not ?
[06:24] <Blackvel> what is the issue with that then ?
[06:25] <AzzIzzA> you cant really compare the two
[06:25] <AzzIzzA> qos reduces latency
[06:25] <AzzIzzA> it doesnt increase bandwidth
[06:26] <AzzIzzA> you cant say x cps with qos will outperform y cps without
[06:28] <turbostaat> (Blackvel): if you prioritize voip traffic as in "minimize delay" it will definetly be processed earlier then without priotizing
[06:28] <turbostaat> this is independent from bandwith
[06:28] <Blackvel> so all links should behave the same with QOS
[06:29] <turbostaat> the bandwith doesnt have too much impact on delay, especially on dsl
[06:29] <Blackvel> because VOIP is 1:10 and default and p2p is 1:40 or 1:50 while even p2p gets rate limited to 1k
[06:29] <turbostaat> you shouldn't rate-limit voip
[06:30] <turbostaat> there are targets like "minimize delay" and "maximize throughput"
[06:30] <Blackvel> so that scenario with minimizing the voip packets delay should work on all adsl bandwidths
[06:30] <AzzIzzA> all dsl uplinks will have similar latencies regardless of bandwidth
[06:30] <turbostaat> but thats kernel-level
[06:30] <Blackvel> rate limit voip ? NO! rate limit p2p
[06:30] <AzzIzzA> bandwidth simply makes it harder to max out a link\
[06:30] <Blackvel> rate 90k ceil 307k
[06:30] <Blackvel> that is for voip
[06:30] <turbostaat> sounds reasonable
[06:31] <turbostaat> no !
[06:31] <turbostaat> you're trying to minimize delay by changing around bandwith options
[06:31] <turbostaat> that wont work
[06:31] <Blackvel> not only that
[06:32] <Blackvel> but voip queue 1:10 is prio 0
[06:32] <Blackvel> and default and p2p is prio 3-5
[06:32] <Blackvel> THAT will affect the delay really
[06:32] <AzzIzzA> dont put a bandwidth limit on voip
[06:32] <AzzIzzA> give it the highest priority
[06:32] <Blackvel> are you kidding me ?
[06:32] <turbostaat> i'd suggest that too
[06:32] <Blackvel> voip is prio 0
[06:32] <Blackvel> ;)
[06:32] <AzzIzzA> and if you like put rate limits on other stuff
[06:33] <AzzIzzA> but make sure only stuff like voip, acks, dns and other realtime stuff has a high priority
[06:33] <Blackvel> nothing has high priority besides voip
[06:34] <Blackvel> because if you mix voip + ACK on queue 1:10 or 1:20 (like wondershaper does) you run into a p2p ACK problem
[06:34] <Blackvel> because p2p ACK = data
[06:34] <Blackvel> btw
[06:34] <AzzIzzA> righto keep acks higher than most stuff though
[06:34] <Blackvel> LALALALALAL!!!!!!!!!!!!!! 3mbit flatrate RUNNING! :)
[06:35] <AzzIzzA> they are small packets and it will improve performance on other tcp based stuff
[06:36] <turbostaat> Blackvel u gone haywire ? :-)
[06:37] <Blackvel> i have that connection since summer, but not with flatrate. it got even 15EUR cheaper than before :)
[06:37] <turbostaat> id suggest maximum priority on voip, least priority on p2p. general traffic including ACK's somewhere between.
[06:38] <turbostaat> im currently considering having a part-time-job to pay my 3mbit flat .)
[06:39] <AzzIzzA> :( your both jerks, leased line is the only thing that can get 3mbit over here
[06:39] <AzzIzzA> and that costs a few hundred a month easy
[06:39] <AzzIzzA> :(
[06:39] <turbostaat> well, actually i am over 100EUR atm
[06:39] <AzzIzzA> i need to move to a country with proper internet
[06:40] <turbostaat> if you've got the cash, russia is supposed to be a good choice
[06:40] <Blackvel> costs of 3mbit flatrate where 75EURO until october, 45EUR for flexiflat 20GIG and now 30EUR for 3mbit fullflat
[06:40] <Blackvel> i am soooo happy
[06:41] <Blackvel> i had isdn 64k flatrate until summer ;)
[06:41] <AzzIzzA> we pay 120$au at work for a 512/256
[06:41] <AzzIzzA> i thing 1$au is reasonable close to 1eur
[06:42] <Blackvel> wow that is expensive
[06:42] <turbostaat> uh, ok ... but including flat i hope .)
[06:42] <Blackvel> france seems to be good too
[06:42] <AzzIzzA> i know
[06:42] <Blackvel> free.fr roxx
[06:42] <Blackvel> 6mbit down for almost nothing ;)
[06:42] <AzzIzzA> it should come down soon but only to about 90$
[06:42] <ursa> does anyone here know a lot about VRAM in the wrt54g?
[06:42] <AzzIzzA> they are redoing all their plans atm
[06:43] <AzzIzzA> ursa: vram?
[06:43] <AzzIzzA> you mean nvram?
[06:43] <ursa> NVRAM
[06:43] <ursa> yes
[06:43] <turbostaat> gg considering the prices, they _should_ redo their plans
[06:43] <AzzIzzA> its a partition on the mtd
[06:43] <AzzIzzA> i think
[06:43] <turbostaat> yup
[06:43] <ursa> it stores a lot of settings correct?
[06:43] <AzzIzzA> i dont know what format the data is stored in
[06:43] <turbostaat> (AzzIzzA): plain ascii
[06:43] <AzzIzzA> most of the important ones
[06:44] <AzzIzzA> turbostaat: that makes life easy :)
[06:44] <turbostaat> aye :) who needs compression anyway
[06:44] <ursa> the reason why i ask. i flashed with alchemy 5.3 and ever since i can't get anything to work right. I put linksys firmware back on and I can only recieve a dhcp on my laptop from the wan port
[06:45] <AzzIzzA> have you dont a factory reset?
[06:45] <ursa> with the reset button?
[06:45] <AzzIzzA> yeah or the web interface if you are running the linksys fw
[06:47] <ursa> when i was running linksys fw i couldn't get to the web interface or anything. The only thing i could do was load new fw images in the first 5 secs using tftp. None seemed to work! I did get openwrt to work, but i want to go back to Samadhi2
[06:47] <ursa> the only thing i want to use this router for is with kismet
[06:47] <AzzIzzA> yeah i have the same plan
[06:47] <AzzIzzA> why not stick with openwrt
[06:48] <AzzIzzA> ipkg install kismet
[06:48] <AzzIzzA> or something
[06:48] <AzzIzzA> and it "just works(tm)"
[06:48] <turbostaat> anyone knows if it is a necessety to have rootfs directly after kernel ?
[06:49] <AzzIzzA> im not going to tell you how to do it, because it could cause major problems
[06:49] <AzzIzzA> but if you erase the nvram
[06:49] <AzzIzzA> and then upload a linksys fw
[06:49] <AzzIzzA> it will reset to defaults
[06:49] <AzzIzzA> in my experience
[06:49] <AzzIzzA> it may also turn your router into a brick
[06:49] <AzzIzzA> so dont do it
[06:49] <ursa> well i have tried the shorting of the pins
[06:49] <ursa> to set boot_wait
[06:50] <ursa> how would i erase the nvram?
[06:50] <AzzIzzA> turbostaat: i dont think so, but the kernel would have to be told where to look for it
[06:50] <AzzIzzA> on an x86 box that would be done with rdev i think or in the config file at compile time
[06:50] <ursa> the router is just sitting on the table now in peices i have wrt54g V 1.0
[06:50] <AzzIzzA> ursa: sounds rather drastic
[06:51] <turbostaat> yes, root=/dev/mtdblock2 or something
[06:51] <AzzIzzA> what firmware is it running atm?
[06:51] <AzzIzzA> can you ping it
[06:51] <AzzIzzA> *shrug* i havent played with mtd much
[06:51] <ursa> i would use ipkg but i can't seem to get an address on wan port
[06:52] <ursa> I'm not sure. what is atm?
[06:52] <AzzIzzA> but it sounds reasonable, is there a KERNEL_PARAMETERS or something in the config file?
[06:52] <AzzIzzA> ursa: atm = at the moment
[06:52] <AzzIzzA> what firm ware is currently loaded onto the router
[06:53] <AzzIzzA> and can you ping the router
[06:53] <ursa> i am using openwrt
[06:54] <ursa> and i do have a dhcp address on the lan side to my laptop
[06:54] <AzzIzzA> ok, can you telnet to the router?
[06:54] <ursa> and my wireless card is up and running now
[06:54] <ursa> yes i am in now
[06:54] <AzzIzzA> ok
[06:55] <AzzIzzA> i seriously advise against this
[06:55] <AzzIzzA> for line in `nvram show | cut -d= -f 1` ; do nvram unset $line; done
[06:55] <AzzIzzA> nvram set boot_wait=on
[06:55] <AzzIzzA> nvram commit
[06:55] <AzzIzzA> dont do it yet
[06:56] <ursa> will nvram erase work also?
[06:56] <AzzIzzA> get someone else to look at it
[06:56] <AzzIzzA> oh
[06:56] <AzzIzzA> prolly
[06:56] <ursa> i'm not yet. you've got me a little nervous
[06:56] <ursa> lol
[06:56] <AzzIzzA> i never even noticed that was a command
[06:56] <AzzIzzA> good
[06:56] <ursa> yeah i found it on the web
[06:56] <AzzIzzA> cos i dont think it was ever intended to have everything wiped out
[06:57] <ursa> so, when you update the firmware it doesn't normally reset the nvram?
[06:57] <AzzIzzA> but if your already resorting to shorting out chips
[06:57] <AzzIzzA> i dont think this would be any worse
[06:57] <AzzIzzA> i just dont want you to blame me
[06:57] <AzzIzzA> if it kills your router
[06:57] <AzzIzzA> i think it does reset some stuff
[06:57] <AzzIzzA> but not everything
[06:57] <AzzIzzA> i dont know
[06:57] <AzzIzzA> i havent played with the linksys fw much
[06:57] <ursa> i think it was alchemy that did the most damage to my router
[06:58] <AzzIzzA> when you upload the linksys fw over tftp did you name it code.bin
[06:58] <ursa> from sveasoft
[06:58] <ursa> yup
[06:58] <AzzIzzA> what version did you use?
[06:58] <AzzIzzA> iv found the smaller the file size the more likely it is to work
[06:58] <AzzIzzA> since it only listens for a short period of time
[06:59] <ursa> i tried the newest from linkssys 3
[06:59] <ursa> but i have also tried 2
[06:59] <AzzIzzA> if your running the v1 hardware
[06:59] <AzzIzzA> try a 1.4 series fw
[07:00] <ursa> ok i will try
[07:02] <ursa> thanks for your help
[07:02] <AzzIzzA> np
[07:02] <AzzIzzA> let me know how it goes
[07:18] <Fabe> does someone kbow when i can buy an asus wl500gx in europe?
[07:27] <RItalMan[away]> re
[07:33] <Dave``> are the control scripts in ipkgs the same as in dpkgs ?
[07:34] <RItalMan> Fabe: for sure there are some sold on www.materiel.net (french site)
[07:34] <Dave``> ie. do preinst, postinst, prerm, postrm scripts in the control subpackage get executed on install/uninstall?
[07:35] Action: Dave`` wants to package a hacked central nvram map mod that backs up any files it replaces and puts them back on uninstall
[07:47] <AzzIzzA> i think the ipkg binary is a shell script
[07:47] <AzzIzzA> so you can see exactly how it works
[07:50] <Dave``> hmm
[07:54] <Dave``> got it; thanks
[08:00] <RItalMan> ipkg binary is quite simple
[08:06] <PolarWolf> It's not a binary
[08:06] <Dave``> it would seem it can be
[08:07] <Dave``> andhelds.org talk of a C version of ipkg
[08:07] <PolarWolf> Yeah, but [mbm] made a shell script of it
[08:07] <Dave``> though yes, the version on openwrt is a shell script
[08:08] <whazilla2> will this open modular stuff be inported into other routers to ?
[08:08] <whazilla2> like my router ? :p
[08:10] <Blackvel> hey polarwolf
[08:10] <Blackvel> you know what ?
[08:16] <RItalMan> what ?
[08:29] <Mobilici> How should I report a suspected bug in OpenWRT scripts?
[08:31] <PolarWolf> Using the forum
[08:31] <Mobilici> Ok.
[08:32] <Blackvel> polarwolf: i got 3mbit dslflat :)
[08:33] <RItalMan> wooaa ;) got a 15mbits adsl @ home
[08:35] <Blackvel> you lucky man :)
[08:36] <RItalMan> for only 30 ¤ per month I have also free call everywhere in france on wired lines and 60 television channels
[08:37] <Blackvel> 60 tv ?
[08:37] <Blackvel> that is cable then ?
[08:37] <RItalMan> no still adsl
[08:37] <Blackvel> hm
[08:38] <RItalMan> but its voip and terrestrial numeric television (is that expression valid)
[08:38] <Blackvel> how do you get 60 tv channels ?
[08:38] <Blackvel> is there free pay tv yet coming with those 60 tv channels ? :)
[08:38] <RItalMan> you plug your tv to a peritel connector on the freebox
[08:39] <Blackvel> ah
[08:39] <Blackvel> cool
[08:39] <Blackvel> what is your upstream with 15mbit down ?
[08:39] <RItalMan> 512 kbits up
[08:39] <RItalMan> a little bit more I don't exactly remember how many
[08:39] <RItalMan> it's quite interesting to host a personnal homepage
[08:41] <Dave``> RItalMan: through Free? :)
[08:41] <RItalMan> yes Dave``
[08:42] <RItalMan> of course ;)
[08:42] <RItalMan> but I think a WMAN should be preferable
[08:42] <Blackvel> hm
[08:42] <Dave``> nice; though outside big cities the dslams don't offer that kind of service yet :(
[08:42] <Blackvel> with 512kbit upstream it is intersting to do webhosting ? reallly ?
[08:43] <Dave``> we only get 2048/128 here
[08:44] <Dave``> grmbl
[08:44] <Dave``> I can't get ipkg to install my .ipk
[08:44] <RItalMan> what's the proble
[08:44] <RItalMan> pb ?
[08:44] <Dave``> I get a short read from tar when ipkg tries to get control.tar.gz from the archive
[08:44] <RItalMan> and ?
[08:45] <Dave``> and ipkg fails saying that there is no control.tar.gz, and I don't get it ;)
[08:45] <RItalMan> I got a similar problem
[08:45] <Dave``> the archive is valid on my desktop machine
[08:45] <RItalMan> when trying to extract files, I got that :
[08:46] <RItalMan> ok I said nothing
[08:47] <Dave``> weird, if I extract it manually with tar zxf, everything workds
[08:47] <RItalMan> try ipkg-extractfiles
[08:47] Action: Dave`` explores /bin/ipkg to find out what it is doing
[08:50] <Dave``> oh, got it
[08:50] <Dave``> stupid path problem in the archive's file listing
[08:51] <Dave``> control.tar.gz exists, but ./control.tar.gz doesn't
[08:53] <Dave``> yay, it works
[09:10] <RItalMan> I got a problem with my ipk
[09:11] <RItalMan> when trying to extract files, the binary is located in /sbin
[09:11] <RItalMan> but the ipkg looks for it at the top directory
[09:12] <Dave``> which binary? The one in your ipk, in data.tar.gz?
[09:12] <RItalMan> the ipkg-extractfile looks for it at the top directory
[09:13] <RItalMan> everything is extracted, but I wonder why it returns an error
[09:14] <Dave``> hmm, dunno
[11:51] <Myren> sshhiitt
[11:51] <Myren> my OpenWRT box just lost eth1
[11:51] <Myren> its not there any more
[11:51] <Myren> wlan led is out
[11:52] <Myren> iwconfig lists no wireless devices
[11:52] <Myren> i unplugged it for 30s, same deal when i plugged it in
[11:52] <Myren> i dont know what t odo
[11:53] <Dave``> insmod wl
[11:53] <Dave``> then check iwconfig again
[11:53] <Dave``> if not, your hardware probably just flaked out :/
[11:54] <Myren> insmod: init_modules: wl: No such device
[11:55] <Myren> :/
[11:55] <Myren> omg, wtf
[11:56] <Myren> this has to be a cruel joke somehow
[12:01] Action: Myren cries like a baby
[12:01] <Myren> whats the proper way to reset this hting?
[12:02] <Dave``> if the driver doesn't find the card, I doubt you can reset it or anything
[12:02] <Myren> whats the way to reset the unit as a whole?
[12:03] <Dave``> I don't know; try looking up the forum
[12:05] <alxhh> wich is the last known working version? the build today and yesterday doas not enable telnet
[12:07] <Dave``> I have one from 2004-11-17 running on my wrt, works fine
[12:10] <RItalMan> re
[12:10] <Myren> why would wireless just stop working? :( i was using default 33mW power.
[12:12] <Dave``> Myren: simple bad luck. Sometimes hardware just dies.
[12:14] <Myren> i dont undrestand how linksys can put out such crappy hardware
[12:14] <Myren> just failing is bullshit
[12:15] <Dave``> just failing happens
[12:15] <Dave``> just like shit
[12:17] <Myren> computers dont fail
[12:17] <Myren> you make them run
[12:17] <Myren> then they run forever
[12:17] <Myren> you really have to make a piece of crap for such spontaneous failure
[12:17] <Dave``> no you don
[12:17] <Dave``> don't
[12:17] <Myren> how do i figure out my firmware date?
[12:18] <Dave``> you simply have to build computers out of electronic components manufactured to the million and set them running
[12:18] <Dave``> all electronic components have a mean time before failure
[12:18] <Myren> and they all involve millions of hours
[12:18] <Myren> i got a couple thousand hours
[12:19] <Dave``> more like thousands
[12:19] <Dave``> and "mean time" means what it means
[12:19] <Myren> its a gaussian distribution
[12:19] <Dave``> you forget that probability figures are not absolute
[12:19] <Myren> your odds of getting 1% the mean time should be a lot less than 1%
[12:20] <Dave``> indeed
[12:20] <Dave``> except that it still happens
[12:20] <Dave``> bad luck, it probably happened to you
[12:20] <Myren> i could've juts won three lotteries with three tickets given the 2.5 months this @!#$@#$ has been running
[12:20] <Myren> @#$@#$!@#$!@#%$!@$%
[12:20] <Myren> @#$!@#$!@#$
[12:20] <Myren> $80 is WAY more money than i have to blow on a 2.5month router
[12:20] <Dave``> don't I know it
[12:20] <Dave``> I spent 1300$ on a laptop that died after a week
[12:21] <Myren> i hope i still have the sales slip
[12:21] <Myren> so i can send this back to linksys
[12:21] <Myren> like you undoubtedly could do with your laptop
[12:21] <Dave``> yep
[12:22] <Dave``> note however that flashing your router with an unofficial firmware voids your warranty :/
[12:22] <Myren> can they detect that
[12:22] <Myren> i'll flash it back first
[12:22] <Ge0rG> they see everything ;)
[12:22] <Dave``> (you could probably flash the original back and they wouldn't see a thing, but I'm just laying down the theory)
[12:22] <Myren> way to console a starving colleg estudent.
[12:29] <turbostaat> does anyone know exactly how a .trx-file is composed ?
[12:30] <turbostaat> like, which file(s) are used with trx ?
[12:31] <Myren> what tftp should i use?
[12:35] <turbostaat> is it safe to add lots of 0x00 to the end of vmlinuz ??
[12:50] <Myren> ok, so now what
[12:50] <Myren> boot_wait=on
[12:50] <Myren> but tftp isnt connecting to it
[12:50] <Myren> i've tried 192.168.1.1 and the default address it boots up to
[12:52] <Myren> is there an ftp client anywherE?
[12:53] <turbostaat> try 192.168.0.1
[12:53] <turbostaat> and dont forget to set your own ip accordingly
[12:53] <Myren> yeah
[12:56] <Myren> nothing. i'm going to try using mtd to rewrite the flash from within the shell
[12:58] <Myren> if i can figure out how to get vsftpd working
[12:59] <turbostaat> use wget
[12:59] <turbostaat> if you have a ftp-server handy
[13:47] <Adze> The new OpenWRT phone?: http://www.nttdocomo.com/presscenter/pressreleases/press/pressrelease.html?param[no]=499
[13:48] <Adze> It runs linux, according to different sites
[13:48] <Adze> A small pic on a dutch site: http://www.tweakers.net/nieuws/35136
[13:49] <Dave``> nice
[14:00] <f15h> anybody here knows what's the difference between "Annex A" and the "Annex B" used for DSL
[14:04] <seb_> hi
[14:04] <seb_> what is default login for wrt54g ?
[14:05] <seb_> it's not flash
[14:05] <seb_> i've just keep it out of box
[14:06] <snuva> no login name
[14:06] <snuva> pw is admin
[14:07] <seb_> thanks, i try
[14:07] <seb_> it doesn't work
[14:08] <snuva> then rtfm
[14:08] <Thus0> try the contrary
[14:08] <Thus0> admin as login and no password
[14:08] <snuva> The default on mine was no login and pw admin
[14:08] <snuva> hehe
[14:08] <seb_> doesn't work
[14:09] <Thus0> anyway, must be on the manual :)
[14:09] <snuva> So like I said, rtfm :)
[14:27] <f15h> is the restore routine via tftp part of the flash memory or is it stored inside the cpu
[14:28] <snuva> flash
[14:30] <[mbm]> the bootloader has a tftp server that you can access if you've set boot_wait .. bootloader is the first 128k of flash
[14:31] <[mbm]> firmware image is a trx file stored immediately after the bootloader .. last 128k of flash is nvram
[14:33] <f15h> thanks
[14:35] <Dave``> and the bit in the middle, in openwrt's case, is squashfs, then jffs2 ?
[14:37] <[mbm]> not quite
[14:37] <[mbm]> oops, bootloader is 256k
[14:37] <[mbm]> pmon on the v1.x and cfe on the v2.x
[14:37] <[mbm]> different code but identical as far as the user is concerned
[14:38] <[mbm]> after that you have the HDR0 signature of a trx image
[14:38] <[mbm]> trx image is the actual firmware .. the code.bin files are just trx files with an extra header (32 or so bytes added to the start)
[14:39] <[mbm]> trx contains the kernel and squashfs
[14:39] <Dave``> ah, right
[14:39] <[mbm]> at the next 64k boundary after the firmware is the jffs2 partition
[14:40] <[mbm]> all calculated dynamically at bootup
[14:40] <Dave``> nice
[14:41] <[mbm]> at some point we'll be getting rid of squashfs entirely and just having one jffs partition for the filesystem
[14:41] <Dave``> hmm
[14:41] <Dave``> how would failsafe mode work with that?
[14:42] <[mbm]> depends on what you mean by failsafe
[14:42] <[mbm]> failsafe actually does two things .. forces it to boot off the squashfs partition, and triggers a script to override nvram variables
[14:42] <Dave``> so I saw in the init script
[14:42] <[mbm]> obviously you can't force it to boot off of a jffs2 partition, but you can still override nvram variables
[14:43] <Dave``> but all that works because you're loading from a squashfs partition which the user can't bork
[14:43] <[mbm]> er can't force it to boot off squashfs
[14:43] <Dave``> true
[14:43] <Dave``> the problem's I've had so far is me being stupid and messing up the bootup scripts (implementing nvmaps)
[14:43] <[mbm]> so it'll basically mean that if the user does anything truely stupid that they'd better set boot_wait so they can tftp a new firmware and jffs2 image
[14:43] <coder> Dave`` btw your maps do not work with 1.x hardware
[14:44] <Dave``> coder: ouch; what happens?
[14:44] <coder> the ap gets inaccesible on a v1.1
[14:45] <Dave``> coder: hmm, does openwrt seem to boot very quickly?
[14:45] <coder> ?
[14:45] <coder> I think your problem is that you don't know that on v1.x the vlans look a bit different
[14:45] <Dave``> very likely
[14:46] <Dave``> I've had this router for something like three days :)
[14:46] <seb_> is there any way to flash a wrt with linksys firmware that i don't have password ?
[14:46] <Dave``> I was trying to get the nvmaps hack to work so that I could get my vlans up the way I wanted without hardcoding stuff in the initscripts
[14:46] <Dave``> coder: how do the vlans differ between v1.x and v2.0 ?
[14:47] <coder> wan is vlan on both but lan is vlan0 on 1x
[14:47] <coder> err, 2x
[14:47] <Dave``> hmm, strange
[14:47] <coder> lan 1x: vlan2, lan 2x: vlan0
[14:47] <Kaloz> coder: wrong
[14:48] <Dave``> coder: if the nvram value for lan_ifname points to the right vlan, things should be fine
[14:48] <Kaloz> coder: v1 - eth0 (lan), eth1 (wan)
[14:48] <coder> mmm
[14:48] <Kaloz> coder: v2 - vlan0 (lan), vlan1 (wan)
[14:48] <Dave``> my scripts don't have any hardcoded values after package installation
[14:48] <coder> 1x: vlan2->eth1->lan
[14:48] <coder> [mbm] : resolv
[14:49] Action: coder doesn't have v2 or gs :_/
[14:50] <Dave``> coder: by "booting quickly" I meant that before openwrt runs the initscripts it sets /proc/sys/diag = 0x05 (power flashing + DMZ led on), and sets it back to 0x00 (power on solid, DMZ off) in the last initscript
[14:50] <Dave``> coder: and when I booted on invalid scripts, as nothing worked properly the bootup phase was near instantaneous
[14:50] <Dave``> it was the only way that I knew something was wrong (except that no interfaces came up, so no telnet)
[14:51] <Dave``> coder: does admcfg work on v1.x hardware?
[14:51] <coder> it does
[14:51] <[mbm]> coder: resolv? I'm not going to get into this argument .. I really don't care about nvmaps
[14:52] <Dave``> right, so if lan and wan are vlans, everything should come up properly
[14:52] <coder> [mbm] I meant, 1x: vlan2->eth1->lan is that right?
[14:53] <[mbm]> vlan1 is always wan .. lan is vlan0 or vlan2 depending on hardware or if you've changed the vlans with admcfg
[14:53] <[mbm]> but the vlans aren't eth1 .. vlans are always on eth0
[14:54] <Dave``> coder: if admcfg works, then if you admcfg ports 1-4 into vlan0, vconfig add eth0 0 and ifconfig vlan0 xxx.xxx.... up, the interface should come up; that is exactly what my scripts do
[14:54] <coder> Dave`` yay
[14:55] <Dave``> coder: what is nvram get vlan0hwname with you?
[14:55] <Dave``> I've just realised I set that value explicitely in my interface definition, instead of pulling it from nvram
[14:56] <Dave``> and with a wrong setting, all interfaces would get ignored as "invalid"
[14:56] <[mbm]> guessing you really didn't do any testing before releasing?
[14:57] <Dave``> [mbm]: I'll readily admit that the package is a quick hack I put together and released as "worksforme". I probably didn't explicit that enough on the forum.
[14:59] <Dave``> I kind of assumed that talking about it on the dev forum and stating that running a dhcp-configured interface through the scripts would probably fsck them was warning enough :P
[14:59] <[mbm]> we're not yet to the point of banning newbies from reading the dev forums
[15:01] <f15h> anybody knows an irc channel where people talk about the netgear dg834
[15:01] <seb_> what is default wrt54g login with linksys firmware on 1.1 hardware_version ?
[15:02] <coder> default login? maybe you interrupted your compilation?
[15:02] <seb_> i want to put openwrt on it but the login "admin" "admin" doesn't work
[15:02] <[mbm]> any username you want, admin as password
[15:03] <coder> wtf?
[15:03] <[mbm]> it doesn't check the username .. manual suggests blank
[15:03] <coder> false sense of security?
[15:03] <coder> ahh, httpd
[15:03] <[mbm]> coder: read. linksys firmware.
[15:03] <coder> fuck, I didn't realize that he said 'linksys wifmrea'
[15:03] <coder> wifmrea=firmware
[15:04] Action: coder must be retired
[15:04] Action: coder is obsolete
[15:04] <[mbm]> retarted?
[15:04] <Kaloz> :DDD
[15:04] <coder> maybe I need to be restarted and you are retarded, who knows
[15:05] <seb_> in fact i've already flash some wrt, but this one is very special
[15:05] <seb_> i can't make tftp
[15:05] <[mbm]> seb_: so you changed the password and forgot?
[15:05] <seb_> neither linux or windows
[15:05] <Kaloz> anyone has ops here? :P
[15:05] <Kaloz> hehe
[15:05] <seb_> i fon't change anything
[15:05] <seb_> maybe i can put the wrt in default setting ?
[15:06] <Kaloz> seb_: v1 or v2?
[15:06] <seb_> 1.1
[15:06] <[mbm]> yeah .. with the linksys firmware if you hold down the reset button for something like 5 sectonds it resets
[15:06] <Kaloz> hold down reset
[15:06] <seb_> so i try, thank you
[15:06] <[mbm]> don't forget to set boot_wait
[15:06] <seb_> sure
[15:07] <coder> [mbm] too much love will kill you
[15:07] <seb_> great,
[15:07] <[mbm]> I'll die a happy man ...
[15:07] <seb_> it works
[15:07] <seb_> :)
[15:08] <coder> http://www.yonkis.com/imagenes01/grua_hero.htm
[15:08] <InfiL00p> Hey, what sort of screwdriver do people usually use to short pins
[15:08] <[mbm]> a metal one?
[15:08] <Dave``> and one with a small head, preferably
[15:08] <Dave``> best not short 5 pins together
[15:09] <[mbm]> it's not the size of the head, it's how you use it :P
[15:09] <Dave``> heh
[15:09] <coder> [mbm] : castlevania lament of the inocent is crap, don't buy it or you'll loose your money
[15:09] <[mbm]> I've used straight pins before too
[15:09] <InfiL00p> Hmm..I am just wondering why the screwdriver that I'm using to short pins out isn't shorting out anything
[15:10] <coder> >>> emerge (16 of 125) gnome-base/eel-2.8.0 to /
[15:11] Action: coder wonders how many beers will he have consumed before portage gets 100% updated
[15:12] <Dave``> ?!
[15:12] <coder> wrt : do not lie
[15:12] <[mbm]> hmm guess wrt doesn't know the difference between a new post and a deleted one
[15:13] <InfiL00p> Cool. There are some screwdrivers that are coated with some stuff. I just grabbed a flathead and I managed to debrick a GS no problem
[15:13] <InfiL00p> That's been bugging me for months why my debricking attempts have been failing
[15:14] <[mbm]> not to pick on you, but the instructions to debrick the wrts has been posted to death and still I get emails from people asking how to do it or why it doesn't work, asking for further instructions .. etc
[15:15] <InfiL00p> I know, I know, I blame coated screwdrivers for not allowing people to short out pints
[15:15] <InfiL00p> pins
[15:15] <coder> Hi I just bricked my wrt unit, is there a howto?
[15:15] <InfiL00p> Yes, yes there is
[15:16] <InfiL00p> http://openwrt.org/forum/viewtopic.php?t=580
[15:16] <InfiL00p> It's about 7 pages of forum replies consisting of "But when I shorted the pins, nothing happens"
[15:16] <coder> InfiL00p /ignore coder
[15:18] <Kaloz> bah, it's 0:30 here... and i'm compiling a kernel for a server instead to be drunken like coder
[15:18] <Kaloz> :p
[15:18] Action: coder chose the dark side today
[15:18] <snuva> So, you're doing something wrong :)
[15:18] <Kaloz> yeah
[15:18] <coder> the dark side is easy, has beer and hentai
[15:18] <Kaloz> :p
[15:18] <coder> you must resist
[15:18] <snuva> wtf hentai?
[15:18] <coder> resist you must
[15:18] <[mbm]> tentacle porn
[15:19] <Kaloz> [mbm] :DDDDDDD
[15:19] <y354C> Hi.
[15:19] <Kaloz> tentacle as the animal? :p
[15:19] <Kaloz> yo y354C
[15:19] <InfiL00p> Beer leads to hentai, heatai leads to tentacle pr0n, tentacles lead to bad sushi, bad sushi leads to suffering
[15:19] <[mbm]> never seen japanese anime?
[15:20] <Kaloz> i was just kidding :)
[15:20] <coder> y354C I read on the internets that you have hacked your wrt54g and you are using wince inside, right?
[15:20] <Kaloz> coder: there is a wrt running wince :p
[15:20] <[mbm]> http://www.pornblography.com/daily_grind/images/internet_porn.mov ... warning, this will offend people easily
[15:21] <coder> then we must download it
[15:21] Action: [mbm] thinks it explains kaloz's behavior
[15:21] <y354C> Is that the video of the kid having sex with his dog?
[15:22] <coder> that kid named ewing?
[15:22] <[mbm]> y354C: yeah
[15:22] Action: [mbm] wonders where the video came from
[15:23] <Kaloz> [mbm]: i do this with girls, not with dogs :p dunno what's your problem then :P
[15:23] <Kaloz> coder: lol
[15:23] <[mbm]> Kaloz: what, never heard of "man's best friend" ?
[15:23] <coder> Kaloz the pc is not a girl, told you like 200 times
[15:23] <coder> Kaloz and your mom does not count
[15:23] <coder> nor your sister does
[15:24] <seb_> how can i remove br0 ,
[15:24] <coder> lan_ifname
[15:24] <Kaloz> coder: you're lucky that you're so far away :P
[15:24] <coder> Kaloz because you wanna sex with me?
[15:25] <Kaloz> coder: i wanna kick you in the ass :P don't be a moron when you're drunk :p
[15:25] <Kaloz> coder: btw, you should be in bed with a girl, not near the computer with a keyboard :p
[15:25] <coder> that was yesterday, today I want to be with a pc
[15:26] <RItalMan> how much time does it take to upgrade from a previous openwrt firmware using mtd ?
[15:27] <[mbm]> should finish in less than 10 seconds .. sometimes it will hang and you have to unplug and use boot_wait
[15:27] <RItalMan> erf ok I retry
[15:27] <coder> Hi, does openwrt get compiled with -msse2 and -m3dnow ?
[15:28] <Kaloz> coder: yeah, but only with -m64
[15:29] <coder> does anyone know good parameters for hdparm on wrt v1 ?
[15:29] <jon_kare> what does an upgrade do to the customizations you've already made?
[15:29] <coder> -d1 -x69 seems quite good but I want to improve the performance
[15:29] <[mbm]> jon_kare: nothing
[15:31] <y354C> http://wireless.uoc.gr/forum/viewtopic.php?t=575
[15:32] <y354C> Look at that neat bootup script I found.
[15:32] <jon_kare> the text is a bit hard to follow in places
[15:33] <coder> admcfg on sveasoft?
[15:33] <coder> mmm suppose copied it from them
[15:33] <coder> suppose mbm copied it from them
[15:33] <y354C> No, the script is for OpenWRT.
[15:34] <y354C> ifconfig br0 down #disables default bridge br0
[15:34] <y354C> brctl delbr br0 #deletes default bridge br0
[15:34] <Ge0rG> coder: did you ever thought of helping out on the sveasoft forums?
[15:34] <Ge0rG> ugh. *think
[15:34] <coder> Ge0rG I got banned there and my posts deleted
[15:34] <y354C> route delete -net default #deletes default gateway route that might have been entered by nvram through S40network
[15:34] <coder> that was just after he stole my money
[15:34] <Ge0rG> coder: maybe you can try it with a new alias :)
[15:34] <coder> I do not want
[15:35] <y354C> He doesn't want to lose another $20.
[15:35] <coder> yeah
[15:54] <Jakdaw> Anyone have any experience running WRT54G's in Ad-Hoc mode?
[15:55] <Jakdaw> I have: Wired-LAN A < OpenWRT Bridge > Ad-hoc mode 802.11g < OpenWRT Bridge > Wired-LAN B
[15:56] <Jakdaw> but from a machine on Wired-LAN A I can't get more than 50k/s to Wired-LAN B over the ethernet bridge
[15:56] <Jakdaw> any ideas?
[15:58] Action: [mbm] uses wds
[15:59] <Jakdaw> can you point me in the direction of a decent guide to wds?
[15:59] <Jakdaw> are you using it with openwrt?
[16:00] <[mbm]> reset both to be access points, set wl0_wds on each one to be the mac address of the other
[16:00] <[mbm]> they need to be configured to the same channel, and if using encryption the encryption settings need to be identical
[16:02] <Jakdaw> that's it?
[16:02] <[mbm]> yep
[16:02] <Jakdaw> you end up with a pair that are both access point and route sensibly to one another?
[16:02] <[mbm]> yep
[16:02] <Jakdaw> cool
[16:02] <[mbm]> you can plug devices into the ports and it'll bridge too :)
[16:03] <Jakdaw> is there anyway to filter what goes over the WDS bridge?
[16:03] <Jakdaw> ie so that I can have an AP in one of two houses - linked over the air
[16:03] <Jakdaw> both houses have ADSL
[16:03] <Jakdaw> so a DHCP request in each house should return an IP local to that house (and a default route of the local ADSL)
[16:04] <Jakdaw> but the machines in one house should be able to route to the machines in the other house - over the air
[16:05] <Jakdaw> I presumably need to stop DHCP requests going from one AP to the other
[16:05] <Jakdaw> heh or perhaps rely upon the most local AP replying fastest
[16:06] <[mbm]> what happens is that a wds0.x device is created as a point to point link, openwrt then adds that device to br0 and it gets routed the same as the wireless + lan ports
[16:06] <Jakdaw> gotcha
[16:06] <[mbm]> you can filter the br0 with ebtables, or you can remove the link from the bridge and set your own routing
[16:07] <Jakdaw_> (screen'd client)
[16:07] <Jakdaw_> can that work between WAP54G's and WRT54G's?
[16:07] <[mbm]> sure
[16:08] <Jakdaw> cool I'll go have a play... thanks! :)
[16:08] <[mbm]> can even work with the old wap11 devices
[16:09] <[mbm]> doesn't always work with other access points .. not all have full wds support
[16:17] <Jakdaw_> hmmm can only get 21 mbit of throughput between a pair of WRT54G's
[16:17] <Jakdaw_> without bridging now - just LAN <WRT54G> Ad-Hoc <WRT54G> LAN - no bridging on the routers just static routes
[16:18] <Jakdaw_> I'll try it using WDS now
[16:18] <solar> [mbm]: you only had one 802.11a device at your place right?
[16:18] <[mbm]> I have a card that does a/b/g in my laptop, but that's the only 11a device here
[16:19] <solar> I can send you a wrt55ag if you would like to add that into the openwrt support devices.
[16:19] <[mbm]> sure
[16:19] Action: [mbm] never objects to free hardware
[16:20] <solar> I'd like to use them but don't have the freetime/space to poke around at making it work
[16:20] <Jakdaw_> WAP54G's support WDS but not Ad-Hoc - right?
[16:21] <[mbm]> you can do ap, client, adhoc or wds
[16:21] <[mbm]> they're almost identical hardware to the wrt54g
[16:22] <[mbm]> solar: don't suppose you've checked for boot_wait and tried openwrt ?
[16:22] <Jakdaw_> oh :) website suggested AP mode only... cool; I'll give it a go
[16:22] <solar> to be honest I've not even plugged the thing in :)
[16:22] <[mbm]> :)
[16:23] <solar> just picked up open. ripped it open and looked inside
[16:23] <solar> looked at gpl source code and said.. hrmm. gonna need mbm's help with this
[16:23] <solar> I'm sending you a mail now with a few details
[16:23] <[mbm]> spiffy
[16:24] <[mbm]> hmm .. hostap got merged in 2.6.10-rc2-mm2
[16:34] Action: [mbm] wonders why that isn't just one printf
[16:35] <[mbm]> bug is probably the return 1
[18:09] <y354C> Is psybnc the best IRC bouncer?
[18:17] <DynamicStability> does openwrt allow me to backup my configuration?
[18:17] <[mbm]> nvram show > /tmp/backup
[18:17] <[mbm]> have fun.
[18:20] <DynamicStability> thanks
[20:57] <y354C> Hmm.
[20:58] <y354C> The ntpclient ipkg doesn't work on the wrt54g.
[20:59] <y354C> Basically, the crystal in the WRT54G drifts beyond ntpclient's ability to correct it.
[20:59] <y354C> It may be possible to correct it though.
[21:02] <[mbm]> what crystal?
[21:02] <[mbm]> there's no clock
[21:02] <y354C> Well whatever the kernel's using.
[21:02] <[mbm]> cycle counter basically
[21:02] <[mbm]> depends on cpu frequency
[21:03] <y354C> Well I followed the directions here: http://doolittle.faludi.com/ntpclient/HOWTO
[21:03] <y354C> old frequency 0 ( 0 ppm)
[21:03] <y354C> new frequency 31998663 ( 488.261 ppm)
[21:03] <[mbm]> donno if the ntpclient was compiked with adjtime support
[21:03] <y354C> And got that.
[21:04] <[mbm]> the linksys solution was to run ntpclient every 5 minutes, then reset the new time to account for timezones
[21:04] <[mbm]> .. rather nasty
[21:04] <y354C> If the frequency offset (absolute value) is greater than about
[21:04] <y354C> 140 ppm (9175040), you have a problem: you may be able to fix it with
[21:04] <y354C> the -t option to adjtimex, or you need to hack phaselock.c, that has a
[21:04] <y354C> maximum adjustment extent of +/- 150 ppm built into phaselock.c (change
[21:04] <y354C> the #define MAX_CORRECT and rebuild nptclient). I'd like to suggest that
[21:04] <y354C> you replace the defective crystal instead, but I understand that is rarely
[21:04] <y354C> practical.
[21:04] <y354C> That's the relevant snippet form the HOWTO.
[21:15] <y354C> We should build adjtimex into busybox.
[21:17] <[mbm]> first step would be for you to see if that fixes anything
[21:17] Action: [mbm] doesn't bother setting time on the wrt since none of my apps are time critical or even aware of time
[21:18] Action: y354C neither
[21:22] <[zombie]> I have some issues with openvpn, but I fix that by setting the clock with rdate once
[21:22] <[zombie]> openvpn sends uses the timestamp in its packets, and the other end rejects packets it has "already seen"
[21:23] <[mbm]> also trying to avoid the issues of someone's horrible hack making it into cvs and causing issues for internet time servers
[00:00] --- Sun Nov 21 2004