**** BEGIN LOGGING AT Wed Nov 03 02:59:57 2010 Nov 03 03:05:16 build #20 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/20 Nov 03 08:29:22 acoul * r23815 /packages/libs/libtasn1/Makefile: libs/libtasn1: fix installation issue. (closes #8165) Nov 03 09:34:24 acinonyx * r23816 /packages/net/srelay/ (Makefile patches/): [packages] srelay: Update to v0.4.8b3 (#8150) Nov 03 10:44:29 acinonyx * r23817 /trunk/package/hostapd/ (Makefile files/wpa_supplicant.sh): [package] hostapd: Add 'identity' config option when eap_type is 'tls' on wpa_supplicant (#8164) Nov 03 11:55:48 [florian], ping Nov 03 11:59:52 <[florian]> Wipster: pong Nov 03 12:02:11 [florian], afternoon I got the dg834g v3 on the bench last night and saw the IRQ issue Ollie is seeing. Not sure why Narge wasn't getting it though as we have the same model, hmm Nov 03 12:02:27 <[florian]> Wipster: okay Nov 03 12:05:17 <[florian]> Wipster: hopefully we should be able to finalize the switch patch right? Nov 03 12:08:13 [florian], hopefully yeh. Going to try and see where that irq fail creeps in this evening Nov 03 12:16:59 [florian], I see in the log that it says eth%d: Marvell 88E6060 PHY driver attached., is it trying to come up before the cpmac interface is up? Could that have something to do with it? Nov 03 12:18:24 <[florian]> it is because we have not called netdev_register() yet, so the device does not have its number yet Nov 03 12:18:27 <[florian]> but that's harmless Nov 03 12:18:47 <[florian]> what I do not understand is how he could possibly get irq = -1 without modifying the current code Nov 03 12:24:30 [florian], the irq is specified in the platform isn't it, I dont know what would cause it to 'fail to obtain' Nov 03 12:25:00 <[florian]> Wipster: yes it is specified as a resourece, which is bound to the cpmac platform_device Nov 03 12:25:17 <[florian]> unless there is a bigger issue like stack overflow or something like that, this should not fail Nov 03 12:26:40 <[florian]> Wipster: anyway, there is an issue in the cleaning path when cpmac is failing to obtain the irq, this must be fixed as well Nov 03 12:46:30 mb * r23818 /trunk/scripts/deptest.sh: deptest: Support specifying certain packages to test on the commandline. Nov 03 12:54:38 mb * r23819 /trunk/scripts/deptest.sh: deptest: Add sanity check to avoid trouble with illegal package names Nov 03 13:34:42 libhistory.so.5.2 may need rpath, but --root not specified Nov 03 13:35:06 I receive this message from the toolchain. How can I setup rpath ? Nov 03 13:35:13 alias --root ? Nov 03 13:35:29 blogic: ping ;) Nov 03 13:40:24 mb * r23820 /feeds/desktop/apps/gpxview/Makefile: gpxview: Add missing dependencies Nov 03 13:43:49 pong Nov 03 13:44:07 what package are you building ? Nov 03 13:44:15 that is a libfool error Nov 03 13:44:31 blogic: openwode. Nov 03 13:46:44 blogic: and how can I move around? Nov 03 13:47:59 blogic: yesterday I spoked with Izma about openwode. I am ready to star with this target, how can we share sources and more? The problems are the same or very similar to these. Nov 03 13:48:33 blogic: Now I pastebin the compilation output Nov 03 13:51:24 can you remember me the pastebin address ? Nov 03 13:56:08 pastebin.com Nov 03 13:56:18 ok Nov 03 14:49:31 mb * r23821 /trunk/scripts/deptest.sh: deptest: Add option for lean test Nov 03 15:00:54 nbd * r23822 /trunk/ (13 files in 3 dirs): ar71xx: improve the wndr3700 quirks - move most of the code out of ath9k and instead allow the platform device to specify gpio overrides - fixes 5ghz signal strength issues Nov 03 15:01:00 nbd * r23823 /trunk/package/mac80211/patches/572-ath9k_xmit_queue_cleanup.patch: ath9k: fix queue pending frame counter tracking by cleaning up tx queue selection Nov 03 15:01:06 nbd * r23824 /trunk/package/mac80211/patches/590-ath9k_fix_cycle_counter_ps.patch: ath9k: fix cycle counter tracking Nov 03 15:03:32 mb * r23825 /trunk/ (include/host-build.mk scripts/deptest.sh): deptest: Install of host tools has to be forced for every test, because the host staging dir is cleaned prior a test. Nov 03 15:06:03 mb * r23826 /trunk/scripts/deptest.sh: deptest: Add --force to force a test Nov 03 15:08:47 nbd: I think it's pretty much usable now :) Nov 03 15:09:12 We should probably run the script for every new package. Nov 03 15:10:14 What I currently don't like in the script is that it clobbers the staging/host directory. I'll probably change that to move the old directory elsewhere and move it back after tests finished. Nov 03 15:12:15 mb__: how about just overriding STAGING_DIR_HOST on the command line Nov 03 15:13:48 hm, if that works, that would be cool Nov 03 15:19:19 i didn't try it, but i think it should work Nov 03 15:19:43 anybody seen the mklibs fix in #7590 ? Nov 03 15:19:51 * nbd looks Nov 03 15:19:59 it is supposed to resolve the missing awk symbol when mklibs is used Nov 03 15:20:45 seems like a good idea regardless of the mklibs issues Nov 03 15:20:57 so that busybox doesn't pull in the libgcc code which we already have in a shared lib Nov 03 15:23:54 mb * r23827 /trunk/scripts/deptest.sh: deptest: --force also overrides blacklist stamps Nov 03 15:24:35 yeah Nov 03 16:18:35 acoul * r23828 /trunk/target/linux/generic/ (63 files in 2 dirs): target/linux: add preliminary 2.6.37 kernel support Nov 03 16:23:40 acoul * r23829 /trunk/package/madwifi/473-mutex_fix.patch: package/madwifi: Use sema_init() instead of init_MUTEX() (based on: http://madwifi-project.org/changeset/4118) Nov 03 17:00:52 acoul * r23830 /trunk/package/madwifi/ (473-mutex_fix.patch patches/473-mutex_fix.patch): package/madwifi: fix r23829 commit Nov 03 18:11:44 ping xMff Nov 03 18:14:26 pong cshore Nov 03 18:15:34 Is it possible to build a luci app out-of-tree (i.e. so I can have a profile for a specific service provider kept as proprietary, but have the generic configuration in-tree)? Nov 03 18:16:40 yes Nov 03 18:17:05 how would I do that? Nov 03 18:17:35 need to find it Nov 03 18:17:38 moment Nov 03 18:19:39 http://luci.subsignal.org/~jow/luci-app-standalone.tar.gz Nov 03 18:20:28 ty! - I'm just about read to start doing the LuCI part of the freeswitch UCI stuff Nov 03 18:20:34 *ready Nov 03 18:25:18 also is it possible to do an OR for dependencies? Nov 03 18:25:36 arteq * r23831 /packages/net/nzbget/Makefile: [package] nzbget: add curses output mode Nov 03 18:26:52 cshore: nope Nov 03 18:28:42 in theory it should be doable through a metapackage that is provided by all OR'd dependies Nov 03 18:29:39 xMff: ok...I'm trying to figure out how to use one configuration (and set of modules) for proprietary and another for the stuff that will be in OpenWRT...the code's the same so maintaining two entirely separate packages would be a pain Nov 03 18:30:52 KanjiMonster: how do you provide a metapackage in OpenWRT? Nov 03 18:31:29 mb * r23832 /trunk/tools/mkimage/Makefile: ensure an empty BUILD_DIR variable is passed down u-boot's makefile Nov 03 18:32:17 mb * r23833 /packages/net/aircrack-ng/Makefile: aircrack-ng is not 100% parallel safe. Disable parallel build. Nov 03 18:32:53 cshore: iirc there is something like PROVIDES that was commited ages ago and never used anywhere, so I doubt it will really work Nov 03 18:43:39 [florian]: ping Nov 03 18:46:16 obsy * r23834 /packages/net/proftpd/Makefile: [packages] proftpd: update to 1.3.3c Nov 03 18:53:04 mb * r23835 /trunk/scripts/deptest.sh: deptest: Do not clobber the base build and staging dirs Nov 03 19:01:43 i have a question: Nov 03 19:02:11 when after hours and hours of testing, I want to settle on a specific openwrt release version, I know what to do... Nov 03 19:02:34 svnversion -> write it down and then whenever I want to check out from the repo this specific version , I can get it Nov 03 19:02:39 and add my .config file Nov 03 19:02:47 and everything will be reproducible and fine Nov 03 19:02:53 but, how does this work for the packages? Nov 03 19:03:01 sorry, for the feeds? Nov 03 19:03:19 you can add "@revnr" to the feed url in feeds.conf Nov 03 19:03:25 so far I only fetched them via scripts/feeds update -a ; scripts/feeds install -a Nov 03 19:03:31 ahhh Nov 03 19:03:33 ok :) Nov 03 19:03:59 gosh, i should get started with reading more of the source (and then document stuff) Nov 03 19:04:05 instead of asking questions here Nov 03 19:04:14 thx! Nov 03 19:04:32 its more an obscure svn feature that works by accident Nov 03 19:04:44 grr fucking tb3 is stealing focus all the time Nov 03 19:07:32 blogic * r23836 /trunk/target/linux/ifxmips/ (125 files in 33 dirs): Nov 03 19:07:32 [ifxmips]: Nov 03 19:07:32 * bump kernel to 2.6.35.8 Nov 03 19:07:32 * merge arcadyn mach files Nov 03 19:07:32 * fixes ar9 Nov 03 19:07:33 * adds hack for tapi drivers Nov 03 19:11:41 blogic * r23837 /trunk/package/ifxmips-dsl-api/ (5 files in 3 dirs): Nov 03 19:11:41 [ifxmips] Nov 03 19:11:41 * make dsl driver work with new kernel Nov 03 19:12:44 blogic * r23838 /trunk/package/ifxmips-dsl-control/Makefile: Nov 03 19:12:44 [ifxmips] Nov 03 19:12:44 * remove broken flag from dsl-control app Nov 03 19:16:00 blogic * r23839 /trunk/target/linux/generic/patches-2.6.35/251-atm.patch: Nov 03 19:16:00 [ifxmips] Nov 03 19:16:00 * adds patch to 2.6.35 series to make dsl work Nov 03 19:17:37 blogic * r23840 /trunk/package/ (93 files in 14 dirs): Nov 03 19:17:37 [ifxmips] Nov 03 19:17:37 * adds a rewrite of the tapi drivers + sip app. this is the result of lars' gsoc 2010 project, Thanks ! Nov 03 19:26:08 blogic: ping Nov 03 19:49:32 hlu: hi Nov 03 19:54:45 hi Nov 03 19:55:43 can you tell me what i have to do, to get amazon pci work? Nov 03 19:56:33 it is not trivial Nov 03 19:56:41 it is rather complicated actually Nov 03 19:56:41 :) Nov 03 19:56:50 the whole io space needs to be swapped Nov 03 19:57:11 ermmm Nov 03 19:59:10 oh, that sounds realy complicated Nov 03 20:00:41 why? Nov 03 20:32:34 is there some kind of information to do that? Nov 03 20:43:58 jow * r23841 /branches/backfire/scripts/patch-kernel.sh: [backfire] partly merge r23314 (#7745) Nov 03 21:06:04 lars * r23842 /trunk/target/linux/generic/config-2.6.36: [kernel] Add even more missing symbols to the 2.6.36 config Nov 03 21:12:45 arteq * r23843 /packages/net/nzbgetweb/ (5 files in 2 dirs): [packages] nzbgetweb: update to 1.4, add external xmlrpc library, add uhttpd configuration Nov 03 22:20:47 mb * r23844 /packages/ipv6/send/Makefile: send: uClibc-0.9.31 fix Nov 03 22:48:57 updated my ar71xx trunk image to binutils 2.20.1 uclibc 0.9.31 and compiled with gcc 4.4.1-cs+graphite, seems to boot fine Nov 03 22:50:54 mb * r23845 /packages/net/dsniff/patches/ (5 files): dsniff: uClibc-0.9.31 fix Nov 03 23:23:24 mb * r23846 /packages/sound/mpd/patches/ (100-compile-fixes.patch diff.patch): mpd: uClibc-0.9.31 fix Nov 03 23:25:27 mb * r23847 /packages/sound/mpd/Makefile: mpd: Enable parallel build Nov 03 23:54:41 build #19 of ps3 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/ps3/builds/19 Nov 04 00:13:19 http://xkcd.com/814/ hahaha Nov 04 00:17:10 I'm loading up a module and it's crapping out with: leds_net5501: Unknown symbol platform_device_register (err 0) Nov 04 00:17:33 not sure what I'm missing... insmod gives a lot less useful info than modprobe and modinfo. Nov 04 00:24:28 that's strange, platform_device_register should be always available Nov 04 00:27:06 that's what I thought. Nov 04 00:51:10 philipp64|laptop: do you have by chance strip unneccesary kernel exports enabled? Nov 04 00:55:26 which flag is that? Nov 04 00:55:51 mb * r23848 /packages/net/strongswan4/patches/ (202-clone.patch 300-compile-fixes.patch): strongswan4: uClibc-0.9.31 fix Nov 04 00:56:59 mb * r23849 /packages/net/strongswan4/Makefile: strongswan4: Enable parallel build Nov 04 00:57:32 philipp64|laptop: CONFIG_STRIP_KERNEL_EXPORTS - you probably knew if you did, but just to be sure Nov 04 01:02:43 it's off... but I have DEBUG_KSYMS_ALL... hope that's not interfering. **** ENDING LOGGING AT Thu Nov 04 02:59:57 2010