**** BEGIN LOGGING AT Mon Apr 11 02:59:57 2011
Apr 11 03:26:50 <CIA-83> nbd * r26585 /trunk/target/linux/atheros/patches-2.6.37/100-board.patch: atheros: remove bogus wlan device mask checks
Apr 11 03:26:55 <CIA-83> nbd * r26586 /trunk/package/mac80211/patches/474-ath5k_fix_aes_crypto_support.patch: ath5k: fix eeprom capability checks that disabled AES crypto support where it should have been enabled
Apr 11 03:27:00 <CIA-83> nbd * r26587 /trunk/package/mac80211/patches/475-ath5k_fix_ar5312_5ghz_cap.patch: ath5k: disable the unusable 2 GHz mode on the first radio of a dual-band AR5312 device
Apr 11 03:27:48 <CIA-83> nbd * r26588 /trunk/target/linux/atheros/Makefile: atheros: use ath5k instead of madwifi
Apr 11 04:11:06 <OutBackDingo> nbd: ping atheros: use ath5k instead of madwifi <- Really ? Ill let you know if its gunna work on Ubiquiti devices
Apr 11 04:11:55 <nbd> I tested it on AR2315, AR2317 and AR5312
Apr 11 04:11:59 <nbd> it worked well on all of those
Apr 11 04:12:13 <nbd> (after my fixes)
Apr 11 04:13:01 <OutBackDingo> cool.. ill test some nanos and such
Apr 11 04:13:12 <nbd> sounds good
Apr 11 04:13:23 <OutBackDingo> be great to get off that madwifi
Apr 11 04:13:29 <nbd> yeah
Apr 11 04:13:40 <nbd> i'm really looking forward to getting rid of that one
Apr 11 04:13:44 <nbd> bbl
Apr 11 04:13:50 <OutBackDingo> though i only have 60 atheros based units left deployed
Apr 11 04:13:58 <OutBackDingo> nbd: caio
Apr 11 04:36:42 <russell--> i need adhoc + ap though, madwifi can't die until after that
Apr 11 04:49:35 <OutBackDingo> russell--: heh well its dead in trunk now :P
Apr 11 04:49:57 <OutBackDingo> i would think ath5k will do adhoc / ap
Apr 11 04:57:02 <russell--> didn't yesterday
Apr 11 05:13:23 <russell--> or i should say, ath9k didn't yesterday
Apr 11 05:14:42 <philipp64|laptop> doh...  too bad I missed nbd.  had a question about running udhcpd on the wan interface.
Apr 11 05:24:21 * russell-- is surprised he was up as late as he was
Apr 11 05:33:50 <russell--> OutBackDingo: "ath5k phy0: Only one single ad-hoc interface is allowed."
Apr 11 07:10:50 <builder_> build #4 of cobalt is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/4
Apr 11 07:57:49 <swalker> jow_laptop: cmake.mk makes a path assumption about ccache as well, http://paste.openwrt.org/iLPQrBOn/, https://dev.openwrt.org/browser/trunk/tools/ccache/Makefile#L20
Apr 11 08:54:08 <xMff> swalker: I don't know it and don't use it so I  will not touch it
Apr 11 09:40:48 <builder_> build #3 of mpc52xx is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/mpc52xx/builds/3
Apr 11 12:45:17 <builder_> build #6 of at91 is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/6
Apr 11 13:03:34 <builder_> build #6 of ubicom32 is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/6
Apr 11 14:19:52 <CIA-83> jow * r26589 /trunk/include/autotools.mk: [include] autotools.mk: disable autoreconf recursion if more than one dir is given in PKG_AUTOMAKE_PATHS
Apr 11 14:20:30 <CIA-83> jow * r26590 /branches/backfire/include/autotools.mk: [backfire] merge r26589
Apr 11 14:27:51 <CIA-83> jow * r26591 /packages/libs/cyrus-sasl/ (3 files in 2 dirs): [packages] cyrus-sasl: fix and use autoreconf to get rid of crappy libtool, extend compile fix to cover automake templates, solve some other automake braindeadness as well
Apr 11 14:39:05 <CIA-83> jow * r26592 /packages/libs/gd/Makefile: [packages] gd: remove manual autoreconf invocation, it pulls in host autofail crap
Apr 11 16:07:48 <CIA-83> jow * r26593 /trunk/rules.mk: [buildroot] rules.mk: properly populate $(LIBGCC_S) for external toolchains
Apr 11 16:23:30 <builder_> build #3 of kirkwood is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/3
Apr 11 17:15:17 <tripolar> hi
Apr 11 17:16:22 <tripolar> i have a ext4 partition and when i mounted it is mounted as ext2 partion - is that a bug or is this normal when the CONFIG_EXT4_USE_FOR_EXT23=y is used?
Apr 11 17:16:39 <tripolar> i have a ext4 partition and when i mount it, it is mounted as ext2 partion - is that a bug or is this normal when the CONFIG_EXT4_USE_FOR_EXT23=y is used?
Apr 11 17:19:00 <tripolar> and another question: how can i test if a package is installed during the makeprocess or with opkg while the system is allready running in postinst?
Apr 11 17:30:41 <xMff> your last question makes no sense, can you rephrase?
Apr 11 17:33:02 <tripolar> i want i know if i install the package withing openwrt with opkg  - or if the package is installed while creating a new image - at postinst time
Apr 11 17:33:10 <tripolar> @ xMff
Apr 11 17:33:33 <tripolar> do you understand what i mean?
Apr 11 17:33:41 <xMff> you mean on-target vs. within the buildroot?
Apr 11 17:33:46 <tripolar> yes
Apr 11 17:34:16 <tripolar> xMff: i sent a patch a few minutes ago that sames ~300kb when installing usbutils and pciutils
Apr 11 17:34:20 <xMff> if its installed within the buildroot, the $IPKG_INSTROOT environment variable is set
Apr 11 17:34:56 <tripolar> sames=saves
Apr 11 17:36:43 <tripolar> thanks
Apr 11 17:36:46 <tripolar> ping swalker
Apr 11 17:37:59 <CIA-83> juhosg * r26594 /trunk/target/linux/ar71xx/files/arch/mips/include/asm/mach-ar71xx/ar71xx.h: ar71xx: fix AR934X_WMAC_SIZE
Apr 11 17:38:04 <CIA-83> juhosg * r26595 /trunk/target/linux/ar71xx/ (23 files in 2 dirs): ar71xx: merge dev-ar9{13,34}x-wmac into dev-ar9xxx-wmac
Apr 11 17:38:08 <CIA-83> juhosg * r26596 /trunk/target/linux/generic/ (2 files in 2 dirs): generic: add support for the MX25L1606E SPI flash
Apr 11 17:40:24 <philipp64|laptop> xMff: I've sent the solos patches to "stable"... but it might be a while before they appear in 2.6.37.x
Apr 11 17:40:34 <xMff> ok
Apr 11 17:48:21 <philipp64|laptop> then again, it might be a while before they appear in openwrt :-)
Apr 11 18:21:34 <builder_> build #3 of au1000 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/3
Apr 11 21:07:50 <PrimasMike> help
Apr 11 21:08:07 <PrimasMike> Is anyone ourt here?
Apr 11 21:08:22 <xMff> maybe
Apr 11 21:08:27 <PrimasMike> lol
Apr 11 21:08:34 <PrimasMike> thanks for answering
Apr 11 21:09:16 <PrimasMike> I am trying to locate the official site for openwrt developers
Apr 11 21:09:45 <trine_> its secret
Apr 11 21:09:49 <xMff> https://dev.openwrt.org/wiki/people ?
Apr 11 21:10:19 <PrimasMike> thanks
Apr 11 21:11:03 <PrimasMike> are there rules of engagement listed there?
Apr 11 21:11:51 <xMff> no, this depends on what you want to engange in
Apr 11 21:11:54 <PrimasMike> In other words, if we want to add some code to the base and need some help in the start up phase, does it give us some guidence?
Apr 11 21:12:27 <PrimasMike> Simple explanation: ...
Apr 11 21:12:38 <xMff> so you seek commit access?
Apr 11 21:12:49 <xMff> or a way to get your changes included in openwrt?
Apr 11 21:13:08 <PrimasMike> Let me explain
Apr 11 21:13:41 <PrimasMike> we are in need of a product that will allow us to do "Remote Sniffing" on a clients network
Apr 11 21:14:23 <PrimasMike> our initial thought was to install two routers with OpenWRT software and VPN them together to extend the traffic
Apr 11 21:14:48 <PrimasMike> in other words ... a "repeater"
Apr 11 21:15:46 <PrimasMike> the software will basically function like a wireless repeater, except it will be between the WAN
Apr 11 21:16:02 <PrimasMike> well
Apr 11 21:16:33 <PrimasMike> it seems like we can't get the code to do this in it's present form - so we wil have to modify it ourselves
Apr 11 21:17:18 <PrimasMike> assuming we successfully get this completed, we would be willing to submit it for others to use
Apr 11 21:17:22 <xMff> so you want do something like port mirroring but in software with the mirrored data stream sent via vpn
Apr 11 21:17:38 <PrimasMike> WOW, you get it
Apr 11 21:17:40 <PrimasMike> exactly
Apr 11 21:18:18 <xMff> hm
Apr 11 21:18:50 <xMff> and it is not possible with libpcap and some random vpn solution?
Apr 11 21:18:57 <PrimasMike> So now we can ship our client a router with Openwrt software and have it homed back to our Sniffer (with our special decoders their protocols)
Apr 11 21:19:31 <xMff> hm
Apr 11 21:19:38 <PrimasMike> It might be, by my developer has been having trouble getting the packets to route
Apr 11 21:20:19 <PrimasMike> it appears that there may be some filters or routing algorithms that he doesn't understand
Apr 11 21:20:20 <xMff> well there's also a nice recursion problem if you send away the captured data via the same channel you capture the data on :)
Apr 11 21:20:47 <PrimasMike> shouldn't be
Apr 11 21:20:55 <xMff> true you can filter it out
Apr 11 21:21:05 <PrimasMike> the LAN port is promiscuous and listens to the data
Apr 11 21:21:20 <PrimasMike> It then should route it on the VPN across the WAN
Apr 11 21:21:36 <xMff> okay, I though you want to observe wan
Apr 11 21:21:48 <PrimasMike> no
Apr 11 21:22:02 <PrimasMike> we want to use the VPN to Secure the data across the internet
Apr 11 21:22:43 <PrimasMike> For instance...
Apr 11 21:23:14 <PrimasMike> A client has a manufacturing facility that is experiencing some network troubles
Apr 11 21:23:32 <PrimasMike> the manufacturing protocol is proprietery
Apr 11 21:23:41 <PrimasMike> but it is OIP based
Apr 11 21:23:46 <PrimasMike> IP based
Apr 11 21:23:49 <xMff> I don't know, there are a couple of techniques I'd read into:  iptables TEE (layer 3 mirroring), libpcap (layer 2 capturing) and/or some custom kernel space solution
Apr 11 21:24:36 <xMff> it is for example very easy to set up tcpdump to listen on lan and then send the captured stream across the network
Apr 11 21:25:22 <PrimasMike> that is exactly what is needed
Apr 11 21:25:24 <xMff> a simple proof of concept could be done with tcpdump and netcat, a more sophisticated solution with libpcap and some custom C deamon to do the network io
Apr 11 21:26:02 <xMff> aside from the vpn implementation which can be anything, pptp, ipsec, openvpn, ...
Apr 11 21:26:31 <PrimasMike> that is what I was thinking as well...
Apr 11 21:27:37 <xMff> the switch driver might be a problem as it enables learning and all that stuff that isolates ports from each other
Apr 11 21:27:39 <PrimasMike> the only issue we are having is we are not versed in the source code so we were thinking of something that was open supported to get help
Apr 11 21:28:10 <PrimasMike> in exchange for the help, we could submit our development
Apr 11 21:28:51 <xMff> hmm, did you consider to hire local unix/linux programmer? Your problem does not sound openwrt specific at all
Apr 11 21:29:05 <PrimasMike> Yes,
Apr 11 21:29:14 <PrimasMike> we posted and only got one response
Apr 11 21:29:43 <PrimasMike> the guys was very knowledgeable but wanted to make a long project of it
Apr 11 21:30:12 <PrimasMike> FYI this was only 1 week kago
Apr 11 21:30:18 <xMff> I see
Apr 11 21:30:56 <PrimasMike> I wanted to check some more forums on my own ( to help my developer)
Apr 11 21:31:38 <PrimasMike> I wanted to find someone with good knowledge of the Openwrt software to make sure there wasn't something in there that could be easily adapted
Apr 11 21:32:16 <PrimasMike> this guy didn't have any knowledge of openwrt
Apr 11 21:32:22 <PrimasMike> he was only unix
Apr 11 21:33:13 <PrimasMike> my thought was to adapt the wireless repeater section
Apr 11 21:33:29 <xMff> thats not applicable at all
Apr 11 21:34:54 <PrimasMike> from an architecture point of view, my htought was to intercept the packet before it goes out the wireless client and redirect it to a VPN client
Apr 11 21:35:23 <PrimasMike> since I am totally in the blind on the internal software, I can only visualize this
Apr 11 21:35:33 <xMff> hjowever there is one very simple but extremely unoptimized way to do what you want
Apr 11 21:35:56 <PrimasMike> where do I send the check????
Apr 11 21:35:56 <xMff> you can build a tap (layer2) openvpn tunnel
Apr 11 21:35:56 <PrimasMike> please tell me
Apr 11 21:36:19 <xMff> then build a linux bridge over lan and the tap interface
Apr 11 21:36:22 <xMff> and finally disable learning on the linux bridge
Apr 11 21:36:36 <xMff> then it will flood the tap interface with all traffic occuring on lan
Apr 11 21:36:50 <PrimasMike> ok
Apr 11 21:37:36 <PrimasMike> is promiscuous set up by default?
Apr 11 21:38:24 <PrimasMike> are you available to consult on this?
Apr 11 21:38:25 <xMff> yeah, afair the bridge will force all its member ports into promiscuous mode
Apr 11 21:39:04 <xMff> however I still think that a libpcap solution would be best
Apr 11 21:39:31 <xMff> then you have at least the option to filter traffic, in case that is needed in the future
Apr 11 21:39:50 <PrimasMike> I know my developer was working with that, but he ran into "lack of knowledge" issues
Apr 11 21:40:15 <PrimasMike> My thought was that there were some filters intercepting the traffic
Apr 11 21:40:25 <xMff> in what area? on how to use libpcap or on how to port his solution to openwrt?
Apr 11 21:41:04 <PrimasMike> libpcap
Apr 11 21:41:09 <PrimasMike> I think he loaded openwrt and it may have been creating conflicts
Apr 11 21:41:46 <PrimasMike> are you suggesting that we run tcpdump or libpcap without openwrt?
Apr 11 21:41:48 <xMff> libpcap has python bindings, maybe this helps to implement a prototype
Apr 11 21:42:28 <xMff> without knowing the actual problem details I am not suggesting for or against openwrt
Apr 11 21:42:52 <xMff> but it might be helpful to first implement it on a pc if your developer is less experienced in embedded development and/or cross compiling
Apr 11 21:43:27 <PrimasMike> he is very comfortable in custom development
Apr 11 21:43:43 <PrimasMike> and has created a PC environment
Apr 11 21:44:21 <PrimasMike> he is also comfortable / but not extremely experienced / in Linux
Apr 11 21:45:05 <PrimasMike> we use it throughtout the year, but only as an OS to boot the application that we spend most of our time with
Apr 11 21:45:58 <PrimasMike> unfortunately he is working from Thailand now, otherwise I would get him on this line right now.  I am tempted to wake him up ...
Apr 11 21:46:37 <PrimasMike> If I wake him up, will you be around on this for a while?
Apr 11 21:46:49 <xMff> no, its almost midnight here
Apr 11 21:46:58 <PrimasMike> ah... where are you?
Apr 11 21:47:03 <xMff> germany
Apr 11 21:47:18 <PrimasMike> wow... your english is perfect
Apr 11 21:47:46 <PrimasMike> I am used to communicating with my team in broken english
Apr 11 21:50:29 <PrimasMike> anyway, is there a way we can reach you again?
Apr 11 21:50:30 <xMff> sure, I am around all the time
Apr 11 21:50:30 <xMff> here and in #openwrt
Apr 11 21:50:32 <PrimasMike> on this site or would you prefer we chatted on email?
Apr 11 21:50:32 <xMff> this irc channel is better as other people can contribute their opions as well
Apr 11 21:50:32 <xMff> opinions
Apr 11 21:50:32 <PrimasMike> I like opions better...
Apr 11 21:50:32 <PrimasMike> lol
Apr 11 21:50:33 <PrimasMike> ok, then I can arrange for him to join this tomorrow.
Apr 11 21:50:40 <xMff> yes.
Apr 11 21:55:43 <CIA-83> heil * r26597 /packages/net/haproxy/Makefile:
Apr 11 21:55:43 <CIA-83> packages: haproxy
Apr 11 21:55:43 <CIA-83> - bump to version 1.4.15
Apr 11 21:55:43 <CIA-83> - Diese und die folgenden Zeilen werden ignoriert --
Apr 11 21:55:43 <CIA-83> M haproxy/Makefile
Apr 11 21:56:04 <xMff> ignore fail
Apr 11 21:59:08 <KanjiMonster> lol
Apr 11 22:00:17 <cshore> Now I wish my mom had taught me german (she was born there)...
Apr 11 22:02:34 <KanjiMonster> cshore: "This and the following lines ones will be ignored"
Apr 11 22:02:59 <xMff> *"This line and the following ones will be ignored"
Apr 11 22:04:23 <KanjiMonster> eh, right. somewhere my english got buggy
Apr 11 22:05:20 <cshore> thanks....though I mean in a more general sense, just being reminded of it because of the german commit
Apr 11 22:09:26 <KanjiMonster> hah! the perfect solution for the current pending patches problem is obviously throwing more patches at -devel =) (at least now the backlog of patches I wanted to clean up and post is almost empty)
Apr 11 22:09:37 <KanjiMonster> thepeople: ping
Apr 11 22:09:42 <CIA-83> jow * r26598 /trunk/package/opkg/ (5 files in 2 dirs): [package] opkg: update to r618
Apr 11 22:09:58 * russell-- has a patch!
Apr 11 22:10:42 <xMff> heh, now that the fs overlay stuff was merged to upstream opkg nobody needs it anymore...
Apr 11 22:11:57 <russell--> there was a ruby-ssl build problem i found an existing patch for (from debian or ubuntu)
Apr 11 22:12:11 <xMff> russell--: v1.0.0 api breakage?
Apr 11 22:12:30 * russell-- looks for it
Apr 11 22:13:37 <russell--> +From openssl 1.0, SSLv2 becomes disable by default.
Apr 11 22:13:37 <russell--> +If you want to use SSLv2 in ruby, you need config with --enable-opensslv2.
Apr 11 22:13:40 <russell--> +The SSLv2 support is disable by default.
Apr 11 22:13:52 <xMff> ah ok
Apr 11 22:14:54 <russell--> http://www.personaltelco.net/~russell/+From openssl 1.0, SSLv2 becomes disable by default.
Apr 11 22:14:58 <russell--> gah
Apr 11 22:15:00 <russell--> +If you want to use SSLv2 in ruby, you need config with --enable-opensslv2.
Apr 11 22:15:03 <russell--> +The SSLv2 support is disable by default.
Apr 11 22:15:08 <russell--> http://www.personaltelco.net/~russell/ruby-ssl.patch
Apr 11 22:32:41 <xMff> russell--: this patch solves a compile issue?
Apr 11 22:33:37 <russell--> xMff: yeah
Apr 11 22:34:00 <xMff> k. just noticed that the ruby zlib extensions seems to be broken as well
Apr 11 22:34:06 <xMff> but it does not crash the build
Apr 11 22:34:07 <russell--> on one of my build machines (debian-unstable).  another one (gentoo) works fine without it
Apr 11 22:34:18 <xMff>  /bin/tar: usr/lib/ruby/1.9/*/zlib.so: Cannot stat: No such file or directory
Apr 11 22:34:59 * russell-- just googled for my error message, found someone had already fixed it, tried it and it worked
Apr 11 22:35:06 <builder_> build #4 of ps3 is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/4
Apr 11 22:36:24 <CIA-83> jow * r26599 /packages/lang/ruby/patches/ (. 0001-Add-option-which-enable-SSLv2-support.patch): [packages] ruby: fix compilation against OpenSSL v1.0.0, thanks russell
Apr 11 23:25:10 <builder_> build #4 of pxcab is complete: Failure [failed compile_3]  Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/4
Apr 12 00:35:04 <builder_> build #6 of s3c24xx is complete: Failure [failed compile_10]  Build details are at http://tksite.gotdns.org:8010/builders/s3c24xx/builds/6
Apr 12 01:03:08 <builder_> build #3 of adm5120 is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/3
**** ENDING LOGGING AT Tue Apr 12 02:59:57 2011