**** BEGIN LOGGING AT Mon Oct 11 02:59:57 2010
Oct 11 04:54:11 <EqUaTe> neutronscott: submit patches? there's an article on the wiki about how to do it.
Oct 11 04:54:20 <EqUaTe> at least, i think it's on the wiki
Oct 11 08:18:59 <loswillios> xMff: I have written an hotplug.d/iface script and it gets executed twice. I only have one wifi-interface though (wan). Is that expected or a bug in current trunk?
Oct 11 08:31:35 <philipp64|laptop> did you try uncommenting the top of hotplugs-common.rules and rebooting?
Oct 11 08:32:03 <philipp64|laptop> loswillios: try that and then check your logs
Oct 11 08:36:51 <acoul> hello
Oct 11 08:48:35 <loswillios> hi acoul
Oct 11 08:48:49 <loswillios> philipp64|laptop: that doesn't give me IFUP events
Oct 11 09:00:33 <loswillios> ah, pebkac
Oct 11 09:16:59 <swalker> updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/uscan.shtml
Oct 11 09:26:59 <loswillios> philipp64|laptop, xMff: I had to check for the wan-interface. Apparently there's an event for lan even if there's nothing connected.
Oct 11 10:02:13 <CIA-39> acoul * r23391 /trunk/target/linux/au1000/au1500/ (config-2.6.34 config-2.6.35): linux/au1000: fix jffs2/squashfs au1500 boot issue (deals with ticket #8058)
Oct 11 10:26:32 <builder_> build #7 of ifxmips is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/ifxmips/builds/7
Oct 11 10:40:01 <diizzy> Hi Acoul, jow directed me to you for building issues on freebsd. I'm running into some odd issue where gmake cannot find install for some reason. Anything you can have a look at? I have a copy of the error here http://paste.openwrt.org/m27e77e0 . I'll be away for a few hours but I'll be back tonight if you want me to test something
Oct 11 10:40:50 <diizzy> This is trunk from about a day ago
Oct 11 10:42:27 <diizzy> ..on freebsd -current from yesterday but I can give it a go on 8.1 if you want to
Oct 11 10:42:32 <acoul> @diizzy: for what platform are you building
Oct 11 10:43:21 <diizzy> Ahh, bcrm47xx
Oct 11 10:43:55 <acoul> have you instelled the gnu tools/ports?
Oct 11 10:44:14 <acoul> also it's important to fix your path
Oct 11 10:44:21 <acoul> so /usr/local comes first
Oct 11 10:44:46 <diizzy> path is fixed
Oct 11 10:45:09 <diizzy> gnu coreutils are installed
Oct 11 10:45:25 <acoul> PATH=/sbin:/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin
Oct 11 10:46:38 <diizzy> Want me to try with that path?
Oct 11 10:46:38 <acoul> what: "which install"  gives you
Oct 11 10:46:59 <diizzy> /usr/bin/install
Oct 11 10:47:51 <diizzy> although gnu install is called ginstall and is /usr/local/bin
Oct 11 10:48:07 <diizzy> is in
Oct 11 10:48:36 <jow_laptop> acoul: we should check for ginstall, then install in include/host.mk and then use $(INSTALL) where applicable, like I did it for patch
Oct 11 10:48:48 <acoul> I see, I had to use the above "PATH" in order to get openwrt going on freebsd 8.1
Oct 11 10:51:05 <acoul> @jow: http://paste.openwrt.org/m27e77e0  <-- line 2 calls install not ginstall though
Oct 11 10:52:47 <jow_laptop> yes and ideally it should call $(INSTALL) which has been determined previously
Oct 11 11:02:04 <builder_> build #7 of brcm47xx is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/brcm47xx/builds/7
Oct 11 11:54:04 <builder_> build #7 of sibyte is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/sibyte/builds/7
Oct 11 12:24:45 <_lore_> ping nbd
Oct 11 12:25:15 <nbd> pong
Oct 11 12:25:17 <_lore_> hi
Oct 11 12:25:21 <nbd> hi
Oct 11 12:25:38 <_lore_> after some test I discovered the TPC works fine
Oct 11 12:26:02 <nbd> interesting
Oct 11 12:26:04 <nbd> what did you change?
Oct 11 12:26:42 <_lore_> I was wrong because I forgot to set the register AR_PHY_POWER_TX_RATE_MAX after a reset
Oct 11 12:26:50 <_lore_> :)
Oct 11 12:26:51 <nbd> ah
Oct 11 12:27:27 <_lore_> I set the register just at bootstrap
Oct 11 12:27:46 <_lore_> so
Oct 11 12:28:17 <_lore_> after setting the channel the configuration was lost
Oct 11 12:28:26 <_lore_> now works fine
Oct 11 12:28:29 <_lore_> :)
Oct 11 12:28:57 <_lore_> but I think the TPC should be enabled by default
Oct 11 12:30:38 <nbd> not much point as long as it's still unused
Oct 11 12:31:04 <_lore_> ok
Oct 11 12:32:12 <_lore_> but if someone wants to write his custom algorithm ..
Oct 11 12:32:27 <nbd> imho we need mac80211 support for tpc first
Oct 11 12:33:09 <builder_> build #7 of iop32x is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/iop32x/builds/7
Oct 11 12:33:26 <_lore_> I definitely agree
Oct 11 12:34:17 <_lore_> but till that moment imho we can just enable the support in the driver
Oct 11 12:34:43 <nbd> yeah
Oct 11 12:35:53 <builder_> build #7 of adm5120 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/7
Oct 11 12:39:26 <CIA-39> acoul * r23392 /trunk/target/linux/brcm47xx/patches-2.6.36/921-bootargs_fix.patch: linux/brcm47xx: fix an initialization issue on 2.6.36 kernel
Oct 11 12:40:39 <_lore_> I wrote this simple patch
Oct 11 12:40:53 <_lore_> it is just for ar9002 card
Oct 11 12:41:14 <_lore_> http://pastebin.ca/1959257
Oct 11 12:43:36 <_lore_> just for my tests :)
Oct 11 12:44:48 <builder_> build #7 of at91 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/at91/builds/7
Oct 11 12:45:04 <nbd> your placement of the tpc field in the tx info looks problematic to me
Oct 11 12:45:28 <nbd> increases the size of the control part by 4 bytes
Oct 11 12:45:57 <nbd> may cause buffer overruns
Oct 11 12:46:12 <_lore_> hm...
Oct 11 12:47:21 <_lore_> per packet it is not implemented yet in ieee80211_tx_info
Oct 11 12:47:27 <_lore_> *power
Oct 11 12:48:33 <_lore_> where do you suggest to insert this field?
Oct 11 12:48:42 <nbd> no idea right now
Oct 11 12:48:51 <_lore_> :)
Oct 11 12:49:28 <nbd> we have two padding bytes
Oct 11 12:49:36 <_lore_> yes
Oct 11 12:49:38 <nbd> i will need one of them for the aggregation stuff
Oct 11 12:49:43 <_lore_> ok
Oct 11 12:49:44 <nbd> and the other one is used by ath9k
Oct 11 12:50:02 <_lore_> ok
Oct 11 12:50:04 <nbd> maybe we can free up the ath9k one at some point
Oct 11 12:50:34 <_lore_> sounds good
Oct 11 13:04:16 <CIA-39> mb * r23393 /packages/lang/erlang/Makefile: erlang: Need extra -lz
Oct 11 13:25:57 <CIA-39> mb * r23394 /packages/lang/erlang/Makefile: erlang: Remove stray backslashes
Oct 11 13:31:05 <loswillios> mb__: yes, config.log says it fails on the current-dir-in-LD_LIBRARY_PATH-check. both eglibc and glibc are affected
Oct 11 13:32:17 <mb__> Well, I was interested in the actual value of LD_LIBRARY_PATH. That should be stored in config.log, too, loswillios
Oct 11 13:33:40 <loswillios> ah, ok gonna start another build
Oct 11 13:43:08 <CIA-39> mb * r23395 /packages/lang/erlang/Makefile: erlang: Bump release
Oct 11 13:51:12 <CIA-39> nico * r23396 /packages/net/aircrack-ng/Makefile: packages/aircrack: enable unstable binaries (closes: #8018), use PKG_INSTALL, bump release number
Oct 11 13:58:25 <CIA-39> mb * r23397 /packages/net/aircrack-ng/Makefile: aircrack-ng: Enable parallel build
Oct 11 14:06:47 <builder_> build #7 of cobalt is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/7
Oct 11 14:08:49 <loswillios> mb__: it doesn't print LD_LIBRARY_PATH: http://willies.info/tmp/config.log
Oct 11 14:22:29 <mb__> loswillios: hm, you could probably hack the configure script and add an echo.
Oct 11 14:56:33 <loswillios> mb__: /proj/ora/app/oracle/product/11.1.0/client_2/lib:/lib:/usr/lib:/usr/shlib::/usr/lib/X11:/usr/dt/lib:
Oct 11 14:57:00 <loswillios> hm, configure checks for "# - the path contains an element with a dot in it"
Oct 11 14:57:10 <loswillios> weird
Oct 11 14:57:10 <mb__> yeah I guessed that
Oct 11 14:57:31 <mb__> It does not like the empty elements
Oct 11 14:57:52 <mb__> see after shlib and at the end
Oct 11 14:58:02 <mb__> It's a pasting issue
Oct 11 14:58:36 <loswillios> you mean :: or 11.1.0
Oct 11 14:58:57 <loswillios> guess both are matching
Oct 11 14:59:42 <loswillios> I could of course overwrite the path when building
Oct 11 14:59:44 <mb__> mean the :: in the middle and the trailing :
Oct 11 15:00:12 <mb__> Well, why are there host elements in it anyway?
Oct 11 15:00:28 <mb__> I think you _must_ actually override it to get a clean cross environment
Oct 11 15:01:48 <loswillios> good question. it actually is the host-path. echo $LD_LIBRARY_PATH prints the same
Oct 11 15:02:14 <loswillios> so the question is why are glibc and eglibc taking the host LD_LIBRARY_PATH?
Oct 11 15:03:08 <mb__> I guess because nobody overwrote it
Oct 11 15:04:29 <mb__>  export LD_LIBRARY_PATH:=$(if $(LD_LIBRARY_PATH),$(LD_LIBRARY_PATH):)$(STAGING_DIR_HOST)/lib
Oct 11 15:04:40 <mb__> I don't really understand the $(if part here.
Oct 11 15:04:47 <mb__> Why do we want to preserve it?
Oct 11 15:05:27 <mb__> Probably for host tools...
Oct 11 15:05:38 <mb__> Well, anyway, you can still override the variable locally
Oct 11 15:21:41 <loswillios> I wonder why uclibc doesn't run into that
Oct 11 15:25:34 <dottedmag> mb__: $(if) is to avoid adding ':' to the front if LD_LIBRARY_PATH is unset
Oct 11 15:26:54 <loswillios> and uclibc doesn't need that?
Oct 11 15:27:03 <mb__> Yeah I know what it does semantically. I'm just unsure why we want to preserver it at all.
Oct 11 15:27:06 <{Nico}> the main problem is that the local LD_LIBRARY_PATH override our own
Oct 11 15:29:51 <{Nico}> i'm wondering in what cases this would be needed
Oct 11 15:30:30 <{Nico}> probably only when building with an external toolchain
Oct 11 15:31:30 <loswillios> can we ifdef it then or something?
Oct 11 15:32:17 <builder_> build #7 of ixp4xx is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/7
Oct 11 16:33:45 <CIA-39> claudio * r23398 /trunk/target/linux/at91/ (2 files in 2 dirs): [at91] Add support for FOXGM carrier board
Oct 11 16:33:49 <CIA-39> claudio * r23399 /trunk/target/linux/at91/image/Makefile: [at91] Copy rootfs(s) inside BIN_DIR
Oct 11 16:33:51 <CIA-39> claudio * r23400 /trunk/target/linux/at91/ (Makefile config-default patches/800-netus-support.patch): [at91] Update to 2.6.35.7
Oct 11 16:33:54 <CIA-39> claudio * r23401 /trunk/target/linux/ (at91/Makefile etrax/Makefile): Add me as maintainer
Oct 11 16:33:56 <CIA-39> claudio * r23402 /trunk/target/linux/at91/image/Makefile: [at91] Tag the uncompressed image
Oct 11 18:48:56 <CIA-39> tripolar * r23403 /packages/net/ctorrent/patches/ (. 100-CVE-2009-1759.patch):
Oct 11 18:48:56 <CIA-39> add CVE-2009-1759 patch to ctorrent
Oct 11 18:48:56 <CIA-39> for more info see
Oct 11 18:48:56 <CIA-39> http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2009-1759
Oct 11 18:52:03 <CIA-39> tripolar * r23404 /packages/net/ctorrent/patches/100-CVE-2009-1759.patch: fix a typo in the patch
Oct 11 20:02:56 <CIA-39> nico * r23405 /packages/net/chaosvpn/Makefile: packages/chaosvpn: fix build, remove redundant vars
Oct 11 20:05:25 <booo> i would like to add switch support for se515 aka RTA770W but i don't know where to start? getting roboswitch: Probing device eth0: No such device
Oct 11 21:01:34 <diizzy> acoul: your path seems to fixed the quirks
Oct 11 21:01:55 <diizzy> any idea how well package feed(s) works?
Oct 11 21:02:26 <{Nico}> diizzy: what do you mean by "how well" ?
Oct 11 21:03:09 <diizzy> Well hrm... do a lot more packages fail compared to a native linux host
Oct 11 21:04:36 <diizzy> Since everything is supposed to be build within the build environment I would expect same compatibility but I somewhat doubt thats the case
Oct 11 21:05:35 <{Nico}> sorry, i don't get your point
Oct 11 21:06:22 <diizzy> My question is (without actually trying myself yet) if more packages fails on freebsd as host compared to linux
Oct 11 21:06:31 <diizzy> for myself
Oct 11 21:07:27 <diizzy> since base needs/needed a bit of hand holding I would expect more packages to fail too in freebsd but that's maybe not the case
Oct 11 21:07:28 <{Nico}> ah, no idea, never tried building on freebsd
Oct 11 21:09:20 <diizzy> Great work getting base to build on freebsd
Oct 11 21:10:15 <diizzy> and I would also thank for the help so far
Oct 11 21:16:51 <{Nico}> diizzy: you're very welcome to try building the feeds and report anything weird
Oct 11 21:18:14 <diizzy> I'll give it a spin and report
Oct 11 21:18:39 <{Nico}> good, thanks!
Oct 11 21:34:09 <builder_> build #7 of kirkwood is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/7
Oct 11 21:55:52 <heilterminalcons> in include/netfilter.mk, line 138 is a 2.6.35 exception that prevents tproxy from working properly. so libxt_mark.so is built and this
Oct 11 21:55:52 <heilterminalcons> rule wont work anymore iptables -t mangle -A DIVERT -j MARK --set-mark 1. What was the reason for this?
Oct 11 21:57:52 <heilterminalcons> i tried this with a seagate dockstar
Oct 11 21:59:14 <heilterminalcons> also rules like iptables -A POSTROUTING -t mangle -j CONNMARK --save-mark wont work anymore. or is there a new syntax?
Oct 11 21:59:16 <xMff> the Linux guys love to move stuff around
Oct 11 21:59:27 <xMff> so iptables userspace goes out of sync with the kernel
Oct 11 21:59:41 <xMff> this exception probably tries to solve that
Oct 11 21:59:55 <heilterminalcons> iptables 1.4.9.1 is not recent enough?
Oct 11 22:00:21 <xMff> dunno, 1.4.9.1 probably made this exception obsolete
Oct 11 22:01:02 <xMff> so whats the exact issue? kmod not packaged or userpspace part missing?
Oct 11 22:01:57 <heilterminalcons> it searches for libxt_MARK.so but there is only libxt_mark.so
Oct 11 22:06:02 <diizzy> that went fast... :/
Oct 11 22:06:04 <diizzy> http://openwrt.pastebin.com/KsnFakjG
Oct 11 22:06:37 <xMff> diizzy: not enough context
Oct 11 22:07:01 <xMff> heilterminalcons: you can probably remove the exception but then it'll break for older kernels
Oct 11 22:08:15 <diizzy> http://openwrt.pastebin.com/QrmzG1Zm
Oct 11 22:08:29 <diizzy> That's the whole build for cyassl
Oct 11 22:09:00 <heilterminalcons> xMff: it wont breaks older kernels because the exection is ge,2.6.35
Oct 11 22:09:34 <xMff> iptables has xt_mark (lowercase)
Oct 11 22:09:35 <heilterminalcons> ive changed the execption to match kernels higher 2.6.36 and rebuild iptables. that works for me at the moment
Oct 11 22:09:41 <xMff> 2.6.35 has xt_mark.ko
Oct 11 22:09:53 <xMff> all other kernels have xt_MARK.ko
Oct 11 22:11:37 <heilterminalcons> thats right, but i need /usr/lib/iptables/libxt_MARK.so the kernel modul is xt_mark.ko
Oct 11 22:13:25 <diizzy> All files seems to be where they should be as far as I can tell
Oct 11 22:13:58 <xMff> heilterminalcons: so xt_MARK.so and xt_mark.ko ?
Oct 11 22:14:02 <xMff> for 2.6.35
Oct 11 22:14:30 <heilterminalcons> yes. i clean my build root a give them a new fresh try
Oct 11 22:16:09 <xMff> hm no wait
Oct 11 22:16:28 <xMff> they merged xt_MATCH.ko into xt_match.ko
Oct 11 22:16:33 <heilterminalcons> right
Oct 11 22:16:45 <xMff> same for iptables?
Oct 11 22:16:52 <xMff> or is it still two libs there?
Oct 11 22:17:33 <xMff> yes
Oct 11 22:17:35 <xMff> ok
Oct 11 22:17:45 <heilterminalcons> no only one
Oct 11 22:17:59 <heilterminalcons> libxt_match.so >= 2.6.35
Oct 11 22:18:12 <heilterminalcons> xt_match.ko is fine i think
Oct 11 22:18:13 <xMff> .so != kernel
Oct 11 22:18:32 <xMff> $ ls build_dir/linux-ar71xx_generic/iptables-1.4.9.1/extensions/libxt_* | grep -i libxt_mark.so
Oct 11 22:18:32 <xMff> -rwxr-xr-x 1 jow jow 8,3K 2010-10-07 00:37 build_dir/linux-ar71xx_generic/iptables-1.4.9.1/extensions/libxt_mark.so
Oct 11 22:18:32 <xMff> -rwxr-xr-x 1 jow jow  11K 2010-10-07 00:37 build_dir/linux-ar71xx_generic/iptables-1.4.9.1/extensions/libxt_MARK.so
Oct 11 22:19:25 <heilterminalcons> for kirkwood with 2.6.35 its ... only libxt_mark.so
Oct 11 22:19:53 <xMff> yeah because the netfilter.mk is broken in that regard
Oct 11 22:20:02 <heilterminalcons> right
Oct 11 22:20:09 <xMff> it should mask out MARK only for kmod builds, not iptables
Oct 11 22:20:19 <heilterminalcons> sounds great
Oct 11 22:24:11 <diizzy> I'm for me to catch some sleep, Google says that my issue is probably gmake vs make from most posts
Oct 11 22:24:19 <diizzy> time for me
Oct 11 22:25:56 <xMff> heilterminalcons: http://paste.openwrt.org/mb6a0f4e
Oct 11 22:36:00 <heilterminalcons> xMff: thanks a lot
Oct 11 22:54:14 <philipp64|laptop> xMff: so you weren't able to apply the patch "as is", I take it?  I'm going to have to downgrade to TB 3.0 until they get the problem with plain-text corruption fixed.  Either that, or send patches as attachments.
Oct 11 23:01:01 <xMff> yeah it didn't apply
Oct 11 23:01:18 <xMff> TB 3.0 is disappointing
Oct 11 23:01:34 <xMff> I get dozens of dupe mail since I upgraded to it
Oct 11 23:02:05 <philipp64|laptop> sorry, I'll use attachments.
Oct 12 00:08:47 <builder_> build #5 of ps3 is complete: Failure [failed compile_6]  Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/5
Oct 12 00:25:10 <builder_> build #7 of orion is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/orion/builds/7
Oct 12 00:28:30 <EqUaTe> xMff: i can honestly say i've had no such issues with tb3.. i'm fairly happy with it.. it's got apple mail style searching, but the client itself actually works, unlike mail. :P
Oct 12 00:34:45 <builder_> build #7 of rdc is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/rdc/builds/7
Oct 12 00:48:07 <builder_> build #5 of ubicom32 is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/5
Oct 12 01:18:10 <CIA-39> jow * r23406 /trunk/scripts/remote-gdb:
Oct 12 01:18:10 <CIA-39> [scripts] remote-gdb:
Oct 12 01:18:10 <CIA-39> - fix parsing of arch-libc tuples
Oct 12 01:18:10 <CIA-39> - automatically add sysroot to executable if needed
Oct 12 02:28:02 <builder_> build #7 of au1000 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/7
**** ENDING LOGGING AT Tue Oct 12 02:59:57 2010