**** BEGIN LOGGING AT Fri Aug 15 02:59:56 2008
Aug 15 04:13:34 <CIA-8> thepeople * r12316 /trunk/package/opkg/files/postinst: fix symlinks when installing opkg into a image
Aug 15 05:16:13 <frogonwheels>  thepeople - I just selected opkg into the image - and then saw that -
Aug 15 05:20:17 <sn9> i saw that too, but thought it was a fluke
Aug 15 05:21:01 <sn9> there should be an "ipkg" symlink in the pkg, too, in case busybox is built without it
Aug 15 05:21:31 <frogonwheels> yes - most probably
Aug 15 05:21:40 <sn9> vim-full is slightly broken in a similar way
Aug 15 05:21:46 <frogonwheels> oh?
Aug 15 05:21:54 <sn9> as is webif
Aug 15 05:22:26 <frogonwheels> sn9: you mean there should be a symlink to vi ?
Aug 15 05:22:59 <frogonwheels> apache is broken worse - it actually removes the symlink - which then breaks webif!
Aug 15 05:25:06 <sn9> no, to ipkg
Aug 15 05:25:21 <frogonwheels> sn9: sorry - I mean on vim-full
Aug 15 05:25:38 <frogonwheels> (I added the vim-full package)
Aug 15 05:26:21 <frogonwheels> so it should create the symlink if there isn't one there
Aug 15 09:44:58 <sn9> package/broadcom-wl/files/lib/wifi/broadcom.sh is broken! can someone please fix it?
Aug 15 09:45:13 <sn9> look at lines 149-150
Aug 15 09:45:21 <sn9> there is a line missing
Aug 15 09:46:01 <sn9> with vif_do_up
Aug 15 09:47:08 <sn9> should have: append vif_do_up "vif $_c" "$N"
Aug 15 09:47:55 <sn9> otherwise, only the last virtual interface is operated on
Aug 15 09:48:44 <sn9> it took me quite some time to track that down
Aug 15 09:49:26 <blogic> send a patch on -devel please
Aug 15 09:54:15 <sn9> great, i gotta deal with e-mail at this hour?
Aug 15 09:54:24 <sn9> it's just one line
Aug 15 09:54:41 <blogic> add it to trac then
Aug 15 09:54:47 <blogic> thats only 2 clicks
Aug 15 09:55:12 <sn9> then i gotta search for existing tickets...
Aug 15 09:55:19 <blogic> nope
Aug 15 09:55:24 <blogic> just throw it in there
Aug 15 09:56:20 <sn9> what's so bad about just adding the line to svn?
Aug 15 09:56:34 <blogic> i would have to test it first
Aug 15 09:56:43 <blogic> and i have no brcm hw at hand
Aug 15 09:56:58 * sn9 grumbles some more
Aug 15 09:57:41 <blogic> i noticed
Aug 15 09:58:26 <blogic> sn9: you might want to repeat your question now
Aug 15 09:58:40 <sn9> ok
Aug 15 09:58:47 <sn9> package/broadcom-wl/files/lib/wifi/broadcom.sh is broken! can someone please fix it?
Aug 15 09:58:55 <sn9> look at lines 149-150
Aug 15 09:59:07 <sn9> there is a line missing
Aug 15 09:59:24 <sn9> should have: append vif_do_up "vif $_c" "$N"
Aug 15 09:59:48 <sn9> otherwise, only the last virtual interface is operated on
Aug 15 10:00:28 <blogic> mb___: ?
Aug 15 11:01:26 <Kevin`> blogic: are you here? have a question
Aug 15 11:01:41 <blogic> sure
Aug 15 11:01:56 <Kevin`> blogic: how do you flash an olpc image
Aug 15 11:02:18 <blogic> ah, good question
Aug 15 11:02:22 <blogic> you have an olpc ?
Aug 15 11:02:24 <Kevin`> yes
Aug 15 11:02:32 <blogic> B or C rev ?
Aug 15 11:02:37 <Kevin`> C
Aug 15 11:02:42 <blogic> developer key ?
Aug 15 11:02:45 <Kevin`> yes
Aug 15 11:03:29 <blogic> ok
Aug 15 11:03:34 <blogic> in that case ...
Aug 15 11:04:15 <blogic> currently we test with usb stick. there is a new guy from germany, that has been really active with the olpc over the last few weeks and managed to get sugar running
Aug 15 11:04:30 <blogic> he is working on the avahi, networkmanager, python f00 atm
Aug 15 11:04:32 <Kevin`> boot directly from usb?
Aug 15 11:04:36 <blogic> yep
Aug 15 11:04:42 <blogic> so ...
Aug 15 11:04:46 <Kevin`> ah, that would explain some things
Aug 15 11:05:20 <Kevin`> do you by chance know how to fix the flash bad-block table, semi-related? :)
Aug 15 11:05:32 <Kevin`> somehow it got messed up and everything shows up as bad
Aug 15 11:05:35 <Kevin`> (it's not)
Aug 15 11:05:41 <blogic> weird
Aug 15 11:07:56 <Kevin`> http://kwzs.be/~kevin/pic2/p1110001.jpg :/
Aug 15 11:08:27 <blogic> 1 sec
Aug 15 14:19:00 <arosch1> hi
Aug 15 14:19:06 <wrt> arosch1: huh?
Aug 15 14:19:27 <arosch1> hi russell_
Aug 15 14:19:33 <wrt> hi
Aug 15 14:20:02 <blogic> hi arosch1
Aug 15 14:20:09 <wrt> hey
Aug 15 14:23:59 <arosch1> anybody tested the mac8011 in master mode on atheros ? does it work well enough?
Aug 15 14:28:33 <blogic> arosch1: [florian] has been playing with it, i know he tested ralink and b43
Aug 15 14:28:36 <blogic> dunno about ath
Aug 15 14:48:08 <arosch1> thx
Aug 15 14:57:19 <[florian]> arosch1: you mean ath5k or ath9k ?
Aug 15 16:01:13 <sn9> [florian]: you tested b43?
Aug 15 16:02:36 <[florian]> sn9: yes I did, it works fine without encryption
Aug 15 16:02:50 <[florian]> and most of the time requires manual starting of hostapd
Aug 15 16:02:59 <[florian]> other than that it is fine, with a bit of packet loss
Aug 15 16:03:14 <sn9> [florian]: would you be able to put a 2.4 image back on whatever hardware you tested b43 on?
Aug 15 16:03:18 <[florian]> so I would not recommend using it for production setups, but for testing, that is good
Aug 15 16:03:38 <[florian]> sn9: using mtd yes, so should you :)
Aug 15 16:03:53 <sn9> [florian]: i mean, to test something
Aug 15 16:04:05 <[florian]> sn9: not right now, but if you wait for 2-3H yes
Aug 15 16:04:14 <sn9> [florian]: read the log of this channel from six hours ago
Aug 15 16:05:10 <[florian]> so between line 149 and 150 "append vif_do_up "vif $_c" "$N"" should be inserted ?
Aug 15 16:05:34 <sn9> right
Aug 15 16:05:51 <[florian]> this might be the cause of the bugs I met
Aug 15 16:06:12 <sn9> i tested it with a 350n, but it needs to be verified on a 54gl or something
Aug 15 16:06:12 <[florian]> allright, will test that when I get back home
Aug 15 16:06:17 <[florian]> sure
Aug 15 16:06:27 <[florian]> 54g, wl500deluxe should be fine for testing
Aug 15 16:07:11 <sn9> yes, anything non-mimo
Aug 15 16:07:48 <[florian]> right
Aug 15 16:08:09 <[florian]> did you read my question about you having ssb-gige enabled in your kernel config ?
Aug 15 16:08:14 <[florian]> (for 2.6)
Aug 15 16:09:04 <sn9> [florian]: i did, but i didn't quite understand what you asked. could you please rephrase?
Aug 15 16:09:55 <sn9> tg3 refuses even to compile if ssb-gige is turned off, btw
Aug 15 16:10:22 <[florian]> sn9: sure, 2.6.25 for brcm47xx has CONFIG_SSB_DRIVER_GIGE which is a bridge between the tg3 driver and ssb, which should make your ethernet work on wrt350n
Aug 15 16:11:13 <sn9> [florian]: reread my last line
Aug 15 16:11:24 <[florian]> sorry, did not see it
Aug 15 16:18:30 <sn9> [florian]: anyway, it turned out that b43 refuses to do anything at all with 11n hardware atm, so whatever is causing the tg3 failure is far from the most pressing issue with brcm47xx
Aug 15 16:19:39 <sn9> so i built a 2.4 image, and ran into the problem i complained about six and a half hours ago
Aug 15 17:25:12 <aroscha_> hi
Aug 15 17:25:18 <wrt> hello
Aug 15 17:25:29 <aroscha_> cyrusFF: (trunk) + luci feeds give me: Safari can’t open the page.
Aug 15 17:25:29 <aroscha_> Safari can’t open the page “http://172.16.1.1/cgi-bin/luci” because the server unexpectedly dropped the connection, which sometimes occurs when the server is busy. You might be able to open the page later.
Aug 15 17:26:17 <aroscha_> cyrusFF: is that a well know bug in trunk right now?
Aug 15 17:26:48 <xMff> only in Safari?
Aug 15 17:27:19 <aroscha_> nope
Aug 15 17:27:21 <aroscha_> same in firefox
Aug 15 17:27:31 <CyrusFF> which platform?
Aug 15 17:27:36 <aroscha_> or any other piece of software that connects to port 80 :)
Aug 15 17:27:40 <aroscha_> i386
Aug 15 17:27:44 <aroscha_> alex
Aug 15 17:27:46 <aroscha_> alix
Aug 15 17:28:52 <aroscha_> I selected the full luci-admin-core  server (not the mini) in make menuconfig...
Aug 15 17:29:05 <CyrusFF> aroscha_: could you please do the following: open a telnet/ssh console and type /www/cgi-bin/luci and have a look at the output
Aug 15 17:29:11 <xMff> try to run /www/cgi-bin/luci on the shell and look if it prints a backtrace or sth. like that
Aug 15 17:29:17 <xMff> oops, same idea
Aug 15 17:30:08 <aroscha_> ok. pastebin is slow...
Aug 15 17:30:52 <xMff> but you haven't selected the experimental single-threaded server?
Aug 15 17:31:26 <arosch1> no
Aug 15 17:31:28 <arosch1> http://pastebin.com/m263934d5
Aug 15 17:31:47 <arosch1> so seems like lua is missing
Aug 15 17:31:52 <xMff> aroscha, ok known problem, fixed some time ago
Aug 15 17:31:52 <arosch1> or it did not pull it in
Aug 15 17:32:00 <arosch1> ok
Aug 15 17:32:16 <arosch1> did I not have the correct feed update?
Aug 15 17:32:20 <CyrusFF> hmm please do the following; ./scripts/feeds update luci; ./scripts/feeds install bitlib; make menuconfig; make
Aug 15 17:32:26 <arosch1> ahh
Aug 15 17:32:27 <arosch1> ok
Aug 15 17:33:09 <CyrusFF> on fresh checkout this is done automatically but dependency is sometimes lazy with updated dependencies
Aug 15 17:33:21 <CyrusFF> *dependency checking
Aug 15 17:33:36 <aroscha_> i see
Aug 15 17:33:53 <aroscha_> because I actually did make an scripts/feeds update -a before
Aug 15 17:34:51 <xMff> yep, but this doesn't detect new packages
Aug 15 17:35:13 <xMff> sometimes at least
Aug 15 17:38:32 <xMff> if you don't want to reflash your board, it should be enough to locate the bitlib ipk and install it manually
Aug 15 17:42:02 <loswillios> hi CyrusFF
Aug 15 17:42:08 <wrt> loswillios: what?
Aug 15 17:42:14 <CyrusFF> loswillios: hey
Aug 15 17:42:56 <loswillios> I noticed some escaping error in luci with my wireless psk-key
Aug 15 17:43:29 <loswillios> that's the key
Aug 15 17:43:40 <CyrusFF> thanks I will have a look at it
Aug 15 17:43:57 <loswillios> thanks
Aug 15 17:44:54 <loswillios> I setted it via commandline and when I looked at it with luci, it was only the first x characters
Aug 15 17:52:04 <arosch1> cyrusFF: works ! thanks. LuCi is really cool
Aug 15 17:52:23 <CyrusFF> nice to hear ;-)
Aug 15 17:52:26 <CyrusFF> thanks
Aug 15 17:53:25 <arosch1> yeah. the work is very much appreciated!
Aug 15 18:14:40 <arosch1> cyrusFF: I get another error (dont know if it is important): when I set the wifi -> networks tab to some other value (something, like the WPA key) , then I get lots of red error lines: "Error: This field is mandatory" next to actually filled out fields
Aug 15 18:14:50 <arosch1> but it seems to work anyway
Aug 15 18:15:33 <CyrusFF> arosch1: yes i have just noticed this one by myself, it is caused by missing ( ) in a condition, will also be fixed n a moment
Aug 15 18:18:02 <arosch1> ok, no prob. just wanted to report
Aug 15 18:19:08 <CyrusFF> arosch1, loswillios: both issues should be fixed now
Aug 15 18:19:25 <loswillios> thanks CyrusFF
Aug 15 18:56:06 <Yanira> http://openwrt.pastebin.com/d7b380d0 -- any idea what's wrong with uci firewall?
Aug 15 18:57:27 <blogic> Yanira: what error do you get ?
Aug 15 18:57:36 <blogic> ah sorry
Aug 15 18:57:42 <blogic> the dport module is missing
Aug 15 18:57:56 <blogic> i will fix this later today or tomorrow
Aug 15 18:58:04 <Yanira> okie
Aug 15 19:09:10 <nbd> blogic: seems like either -p tcp or -p udp is missing
Aug 15 19:09:16 <nbd> blogic: --dport doesn't need special matching modules
Aug 15 19:10:38 <blogic> i know
Aug 15 19:10:43 <blogic> i am looking at it now
Aug 15 19:10:53 <Yanira> :)
Aug 15 19:16:37 <blogic> nbd: CONFIG_NETFILTER_XTABLES=y
Aug 15 19:16:41 <blogic> don't get it
Aug 15 19:16:47 <nbd> ?
Aug 15 19:16:59 <blogic> leaving out --dport and the line works
Aug 15 19:17:01 <blogic> wtf
Aug 15 19:36:17 <arosch1> re: firewall
Aug 15 19:36:47 <arosch1> I would *love* to have a global setting: firewall_default_policy={ACCEPT, DROP} or something like that
Aug 15 19:44:56 <thepeople> arosch1: there is
Aug 15 19:46:51 <arosch1> thepeople: ah! where? I did not see it in uci_firewall.sh or in package/firewall/files
Aug 15 19:47:55 <thepeople> arosch1: config defaults, the first rule
Aug 15 19:49:04 <arosch1> sorry, /me is removing the tomatoes from my eyes
Aug 15 19:50:05 <thepeople> :-P
Aug 15 20:18:24 <CIA-8> blogic * r12317 /trunk/package/firewall/files/uci_firewall.sh: trigger error if dport is used when no proto is defined
Aug 15 20:37:41 <Yanira> blogic: i changes /lib/firewall/uci_firewall.sh but i get the same error
Aug 15 20:40:21 <arosch1> cyrusFF: bug in status -> wifi scan:
Aug 15 20:40:23 <arosch1> Status: 500 Internal Server Error     	
Aug 15 20:40:38 <CyrusFF> ok
Aug 15 20:40:40 <arosch1>   /usr/lib/lua/luci/util.lua:202: attempt to index local 'value' (a nil value)
Aug 15 20:42:20 <CyrusFF> fixed.
Aug 15 20:43:14 <Yanira> CyrusFF: does port forwarding with uci firewall and luci works for you?
Aug 15 20:43:16 <_trine> can someone fix firstboot cos that's not working
Aug 15 20:44:05 <arosch1> cyrusFF: ok, thx, will test
Aug 15 20:46:04 <CyrusFF> Yanira: i haven't tested it with the latest builds but it worked a few days ago, I will retry with the latest UCI firewall revisions
Aug 15 20:48:27 <Yanira> i have latest fix from blogic, but still Unknown arg `--dport'
Aug 15 20:51:56 <CyrusFF> blogic: I think you messed up something with protocol and proto, especially when i look at fw_redirect
Aug 15 20:53:21 <CyrusFF> and i did the same thing in LuCI -.-
Aug 15 20:58:54 <arosch1> cyrusFF: feedback - wifi scan "sort of" works now . It at least does not throw an error . But there are no scan results.  Do I have to put the device into monitor mode first?
Aug 15 20:59:34 <xMff> hmm the webif just parses the ouput of iwlist sc
Aug 15 21:01:55 <arosch1> hm lets test that then
Aug 15 21:02:56 <CyrusFF> Yanira: i have fixed a typo in LuCI firewall page but I think there is still something wrong with the UCI firewall itself
Aug 15 21:03:18 <Yanira> CyrusFF: ok
Aug 15 21:06:35 <CyrusFF> arosch1: was any wifi mode (ap/client/ad-hoc/...) enabled when you scanned?
Aug 15 21:06:46 <arosch1> jup
Aug 15 21:06:48 <arosch1> ap
Aug 15 21:07:07 <xMff> okay, we had the same problem in the old FF firmware
Aug 15 21:07:08 <Yanira> CyrusFF: with your and blogic's fix /etc/init.d/firewall restart runs without an error now
Aug 15 21:07:26 <xMff> fix was to temporily switch to client mode while scanning
Aug 15 21:07:40 <arosch1> hm, let's try that now then
Aug 15 21:07:41 <xMff> but I'm not sure whether this behaviour is desired...
Aug 15 21:07:51 <CyrusFF> Yanira: ok fine then
Aug 15 21:08:40 <Yanira> hmm, but the port is still stealth according to grc.com > shields up!
Aug 15 21:09:04 <arosch1> xMff: yeah, when I change to sta mode, it scans
Aug 15 21:09:40 <arosch1> but it seems it needs a ifconfig ath0 down; wlanconfig ath0 destroy; wlanconfig ath0 create wlandev wifi0 wlanmode sta; iwlist scan
Aug 15 21:10:02 <xMff> argh
Aug 15 21:10:40 <_trine> wlanconfig ath0 create wlandev wifi0 wlanmode sta works but opton
Aug 15 21:10:55 <_trine> opps
Aug 15 21:11:02 <arosch1> but opton?
Aug 15 21:11:17 <xMff> hard to code around this Problem, at least without introducing quirks for different chip types
Aug 15 21:11:20 <_trine> option sta does not
Aug 15 21:11:41 <arosch1> jup
Aug 15 21:11:53 <arosch1> scanning also works when it is in adhoc mode
Aug 15 21:12:03 <arosch1> btw
Aug 15 21:12:12 <xMff> yes, afair ap mode is the only one affected
Aug 15 21:12:27 <arosch1> well, xMff: does that apply to *all* atheros chips?
Aug 15 21:12:54 <arosch1> if so -> if $DEVICE_NAME ~= ath* then switch mode
Aug 15 21:12:57 <arosch1> ugly but works
Aug 15 21:12:57 <_trine> using option mode sta crashes my fonera on the snapshot
Aug 15 21:13:07 <arosch1> hmmm
Aug 15 21:13:08 <xMff> since it applies to broadcom too i believe its a general thing
Aug 15 21:13:51 <_trine> and if very annoying when that happens that fisrtboot is not working
Aug 15 21:14:11 <_trine> because then I have to reflash the router
Aug 15 21:15:13 <_trine> if=it's
Aug 15 21:15:30 <arosch1> xMff, cyrusFF: suggestion for at least the easy user interface: if you can show some "wlan strength" symbols next to the scan results it will be easier for folks to understand
Aug 15 21:16:26 <xMff> general wifi interface improvements are definitely on the roadmap
Aug 15 21:16:33 <arosch1> k
Aug 15 21:16:49 <arosch1> i might have a cool script for streaming out the SNR to the userinterface
Aug 15 21:18:01 <CyrusFF> ok, then what is it?
Aug 15 21:21:56 <arosch1> ahm.. it will need some glue
Aug 15 21:22:29 <arosch1> but basically the SNR is continually checked and thus you can draw graphs in the client. sort of AJAX thingy
Aug 15 21:22:33 <arosch1> still very beta
Aug 15 21:23:14 <xMff> hmm, you could just use collectd, enable its network plugin and stream the snr data to a local collectd installation
Aug 15 21:23:38 <CyrusFF> grml btw interestingly on my wgt634u it does scan in master mode oO
Aug 15 21:24:07 <thepeople> broadcom can scan in ap mode
Aug 15 21:24:12 <thepeople> atheros can not
Aug 15 21:24:18 <thepeople> currently anyhow
Aug 15 21:24:42 <CyrusFF> but the wgt634u is atheros
Aug 15 21:24:48 <CyrusFF> and it scans ^^
Aug 15 21:24:51 <thepeople> o...O
Aug 15 21:25:15 <arosch1> hmmm
Aug 15 21:25:19 <thepeople> hmmm
Aug 15 21:25:21 <CyrusFF> but don't ask me how and why
Aug 15 21:25:31 <arosch1> could the bgscan iwpriv have anything to do with it?
Aug 15 21:25:54 <arosch1> xMff: ok, sounds like that would fit in better
Aug 15 21:25:58 <thepeople> arosch1: that is only valid in client mode I believe
Aug 15 21:28:06 <arosch1> jup. now the CM9 also scans in ap mode
Aug 15 21:28:50 <CyrusFF> hmm to be sure it did not cache things i rebooted the wgt634u (its actually a broadcom board with an atheros wifi) and it is up in ap mode and still scans
Aug 15 21:28:58 <arosch1> jup
Aug 15 21:29:14 <arosch1> I /sbin/wifi started' again and it stopped scanning in ap mode
Aug 15 21:29:49 <arosch1> *mad*wifi
Aug 15 21:30:00 <xMff> arosch1: but we also plan a rpc like api which you can use to retrieve data, perhaps this is the thing you need as "glue"
Aug 15 21:30:06 <thepeople> I just confimed it here as well on my NS2
Aug 15 21:30:19 * thepeople wonders when that changed
Aug 15 21:30:44 <xMff> arosch1: sry for x-topic
Aug 15 21:31:40 <arosch1> xMff: yeah, sounds good
Aug 15 21:32:09 <CyrusFF> hmmn ok so I think we can leave this wlan-scan page as it is and maybe add a notice that it sometimes "just doesn't work"
Aug 15 21:32:21 <CyrusFF> ^^
Aug 15 21:32:26 * xMff say ok :)
Aug 15 21:32:59 <xMff> at least for now
Aug 15 21:33:05 <arosch1> cyrusFF: great idea ;-)
Aug 15 21:33:49 <thepeople> CyrusFF: or just detect if it doesn't work and create a sta interface if one doesn't exist and can be made
Aug 15 21:34:23 <arosch1> cyrusFF: actually that is pretty essential for all the community free networks for tuning antenna directions etc. Very often we need to scan while we are actually connected to that device via wlan
Aug 15 21:34:37 <arosch1> just my 2 cents/experience
Aug 15 21:35:26 <CyrusFF> arosch1: i know that, i am also part of such a community wifi network
Aug 15 21:35:33 <CyrusFF> ;-)
Aug 15 21:35:41 <CyrusFF> thepeople: yes seems like a good idea
Aug 15 21:36:05 <arosch1> cyrusFF: I know :)
Aug 15 21:37:29 <arosch1> thepeople, cyrusFF: that is what I just did . created a adhoc ath1 in addition and scanned there. works
Aug 15 21:38:01 <CyrusFF> grml why is wifi always such a pita
Aug 15 21:38:20 <arosch1> yeah
Aug 15 21:38:33 <arosch1> actually: it works when you type in iwlist ath1 scan
Aug 15 21:38:36 <Yanira> wireless freedom .. lol
Aug 15 21:38:42 <thepeople> CyrusFF: what fun would it be if it just worked?
Aug 15 21:38:50 <arosch1> but not when you go "iwlist sc"
Aug 15 21:39:17 <arosch1> cyrus: ok... maybe that is a compromise:
Aug 15 21:39:26 <arosch1> for i in $INTERFACES; do
Aug 15 21:39:33 <arosch1>   iwlist $i scan
Aug 15 21:39:38 <arosch1>   present the result
Aug 15 21:39:39 <arosch1> done
Aug 15 21:39:59 <arosch1> then the user can (for atheros) create another interface when he/she wants to scan
Aug 15 21:40:04 <arosch1> without changing the main interface
Aug 15 21:40:19 <CyrusFF> hmm ok
Aug 15 21:40:46 <CyrusFF> thepeople: grml hmm you unfortunately are right
Aug 15 21:40:58 <thepeople> :-P
**** ENDING LOGGING AT Sat Aug 16 02:59:56 2008