**** BEGIN LOGGING AT Thu Oct 24 02:59:58 2013 Oct 24 10:02:42 jow r38525 packages/libs/libpam/Makefile * libpam: remove autoreconf call from configure macro Oct 24 10:39:12 hi Oct 24 10:39:34 where can i find the build logs from the build bots? Oct 24 10:41:06 tripolar: http://buildbot.openwrt.org:8010/ Oct 24 10:41:13 KanjiMonster: thanks Oct 24 10:41:44 tripolar: logs of failing packages are collected in http://buildbot.openwrt.org:8010/broken_packages/ Oct 24 10:43:03 KanjiMonster: someone write on the ml that libpam's build on the build bots is broken, but it compile fine here - i just wanted to the see the breakage on the build server Oct 24 10:43:48 ahh found it Oct 24 10:43:49 then you should find it in broken packages Oct 24 10:44:36 but it seems to link against libdb Oct 24 10:46:38 probably sees it during configure and then enables whatever uses it Oct 24 10:48:29 i knoe Oct 24 10:48:31 know Oct 24 10:48:49 was the same with licap when it was introduced Oct 24 10:48:56 licap/libcap Oct 24 10:49:27 tripolar: in the future you need to do more tests when introducing a new library Oct 24 10:49:56 jow_laptop: i tested it - but i'd inlcude all packages to see if it tries to link against any one of them Oct 24 10:50:36 tripolar: with more tests I mean you actually need to build everything Oct 24 10:50:42 okay Oct 24 10:51:04 jow_laptop: whats the way the buildbot is building it? Oct 24 10:51:12 i mean how does it select all packages Oct 24 10:51:50 rm .config Oct 24 10:51:57 echo CONFIG_TARGET_xyz=y > .config Oct 24 10:52:03 echo CONFIG_ALL=y >> .config Oct 24 10:52:06 make defconfig Oct 24 10:57:55 jow_laptop: what about net/proftpd and libpam dependency? add it or use --disable-auth-pam ? Oct 24 10:58:43 plntyk: personally I'd disable pam support everywhere for now Oct 24 10:58:52 then selectively revise in the future Oct 24 10:59:07 jow_laptop: that was not the correct patch Oct 24 10:59:08 I do not like dozens of packages suddenly gaining new dependencies just because a libpam package was created Oct 24 10:59:14 the autoreconf didn't break anything Oct 24 10:59:23 tripolar: its crap Oct 24 10:59:35 do not hack manual autoconfig invocations into Build/Configure Oct 24 10:59:43 thats what PKG_FIXUP:=autoreconf is for Oct 24 10:59:46 it does it properly Oct 24 10:59:55 i know i missed to remove it Oct 24 11:00:03 it also still used in pythong Oct 24 11:00:07 python Oct 24 11:01:15 jow_laptop: no i ment the manual call didn't break the package Oct 24 11:01:36 tripolar: maybe not for you Oct 24 11:01:50 it does here, in particular since such manual invocations will pull in host autoconf tools Oct 24 11:02:02 if they happen to be in a mismatching version, stuff will break down Oct 24 11:02:12 also shell mismatches Oct 24 11:04:12 tripolar r38526 packages/libs/libpam/Makefile Oct 24 11:04:12 [packages] libpam: add "--enable-db=no" to the configure macro to avoid linking Oct 24 11:04:12 against libdb Oct 24 11:04:12 Signed-off-by: Peter Wagner Oct 24 11:08:09 so mine patch for libpam was correct or not? :-) it was breaking all of mine builds even if libpam was not selected (for all targets I use) and I missed the point to call autoreconf twice. Oct 24 11:08:26 slachta: yes your patch was correct Oct 24 11:08:59 jow_laptop: do you have a list of packages that try to link against libpam? Oct 24 11:09:16 tripolar: so far I am aware of proftpd and at Oct 24 11:09:26 tripolar: but since libpam itself was broken I can't tell yet Oct 24 11:09:40 I suspect once it builds okay on the buildbot, you'll see a lot more packages breaking down Oct 24 11:10:18 should i mark it as broken und build and fix everything here first? Oct 24 11:10:35 @jow_laptop Oct 24 11:10:48 no Oct 24 11:10:51 I would wait for a day or two to get buildbot results Oct 24 11:10:53 you can also download http://buildbot.openwrt.org:64112/broken_packages/ar71xx/logs.tar.gz Oct 24 11:10:59 this contaisn all configure logs Oct 24 11:11:16 rgrep for libpam checks there and you can probably get a list of packages Oct 24 11:16:08 okay thanks Oct 24 11:24:22 jow_laptop is there a way to select a package from the shell with its dependencies, as it would be selected when i enable it in menuconfig? Oct 24 11:25:42 tripolar: I usually do echo CONFIG_PACKAGE_foo=m >> .config; make defconfig Oct 24 11:25:49 ahh okay Oct 24 11:25:50 thanks Oct 24 11:26:14 this works even if there already is a "# CONFIG_PACKAGE_foo is not set" Oct 24 11:29:07 because the old DEVELOPERS=1 make "package that is not selected in menuconfig" doesnt work anymore Oct 24 11:36:55 it's DEVELOPER=1, not DEVELOPERS Oct 24 11:39:55 tripolar r38527 packages/net/proftpd/Makefile * [packages] proftpd avoid linking against libpam Oct 24 11:41:02 typo - KanjiMonster still doesn't work most of the time when a package is not select selected which has also unselected dependencies Oct 24 11:41:57 tripolar: the last time I tried it did work, but it doesn't play nice with multiple packages or config option Oct 24 11:41:59 s Oct 24 13:18:44 build #388 of rb532 is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/388 Oct 24 13:20:51 build #388 of ppc44x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/388 Oct 24 13:51:40 kaloz r38528 trunk/target/linux/ (107 files in 2 dirs) * [generic:] preliminary 3.12 support Oct 24 13:52:38 Kaloz: \o/ Oct 24 13:53:58 wow Oct 24 13:54:20 kernel devel is going so fast.. Oct 24 13:54:47 faster than owrt, which I find extremely fast Oct 24 14:28:30 blogic r38529 trunk/package/ system/procd/Makefile system/procd/files/log.init * procd: update to latest git head Oct 24 15:08:40 build #421 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/421 Oct 24 16:44:29 nbd: getting UCI build problems due to -Werror on fedora 19: http://pastebay.net/1333337 Oct 24 16:46:44 sorry, I lie, on ubuntu 1204, Oct 24 16:51:12 build #364 of sibyte is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/364 Oct 24 18:03:41 hi. i wrote a collectd plugin that collects splash leases with uci. I made a patch and tried (several times with diff options/Makefiles) to compile it with make/package/collectd/{clean,preapare,compile}. The problem: It seems that my plugin is ignored by make although it is selected in make menuconfig. Here is my patch http://pastebin.com/5isQZLUa , the make log http://pastebin.com/GS6NZU6G and make Oct 24 18:03:42 errors http://pastebin.com/7cKsAQ5u Oct 24 18:04:49 could anyone give me a tip whats going wrong here? Oct 24 20:08:45 nbd: seen this? https://lists.openwrt.org/pipermail/openwrt-devel/2013-October/022031.html Oct 24 20:09:18 "AR9344 wifi dying on WD N600 routers" Oct 24 20:10:22 if you can also fix this: https://lists.openwrt.org/pipermail/openwrt-devel/2013-October/022002.html ("WD MyNet N750: ar9344 + ar8327 switch, unclear on phy/mdio/mac wiring") I could send you a device to test, if you need one Oct 24 20:30:39 fyi: michiwends problem solved, so don't try to find the error (except you are bored) ;) Oct 24 21:19:08 build #353 of kirkwood is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/353 Oct 24 21:26:39 heffer: yes, i've seen it. not sure yet, where the bug is Oct 24 21:26:50 karlp: please git pull and try again Oct 24 21:28:52 that switch on the N750 however sure is a bitch :D Oct 24 21:29:26 have you tried simply configuring it like similar devices that have ar9344 + ar8327? Oct 24 21:29:46 you don't need to send me a device though, at the moment i don't have the time to deal with new toys ;) Oct 24 21:30:21 haha okay. yeah i tried several different devices (TL-WDR4300 DB 120 reference board) Oct 24 21:30:42 however the switch is deactivated by the bootloader, possibly to prevent LAN/WAN leakage Oct 24 21:30:54 as WAN and LAN are on the same switch, separated by VLANs Oct 24 21:31:42 i get [ 1.440000] ag71xx ag71xx.0 eth0: connected to PHY at ag71xx-mdio.1:00 [uid=004dd042, driver=Generic PHY] Oct 24 21:32:06 also: [ 2.010000] eth1: Found an AR934X built-in switch but that switch has nothing connected to it Oct 24 21:33:18 ag71xx-mdio.1 has 4 PHY all with the same phy_id (004dd042) Oct 24 21:34:31 with what image was that? Oct 24 21:35:00 i made one for the device specifically, copied from the device's smaller sibling, the My Net N600 Oct 24 21:35:30 the My Net N600 however utilizes the AR9344's internal switch directly while the N750 has an additional AR8327N Oct 24 21:36:04 right Oct 24 21:36:15 you need to copy the switch/mdio stuff from the wdr4300 Oct 24 21:36:23 and ethernet Oct 24 21:36:26 also the N600 uses GMAC0 for WAN and GMAC1 for the switch Oct 24 21:36:40 i'll try that, although I believe i did that before Oct 24 21:37:29 ok. when you try again, make a pastebin of the mach-file ;) Oct 24 21:37:58 nbd is it correct that brcm947xx uses the same mac for wan and lan? I have a wrt310n and it reports the same mac for both Oct 24 21:38:46 nbd: will do :) Oct 24 21:40:28 Devastator: maybe you should ask somebody that has actually used or worked on brcm947xx recently ;) Oct 24 21:41:34 nbd sorry, i did, it if I understand correctly, it's how is designed Oct 24 21:41:41 I just find it odd.. Oct 24 21:42:21 i won't comment, i haven't looked at that stuff in years Oct 24 21:42:47 nbd got it, what's your main platform? atheros? Oct 24 21:44:11 http://pastebin.com/MuqLd2UZ mach file Oct 24 21:44:20 Devastator: ar71xx Oct 24 21:44:21 http://pastebin.com/5NqDrEXs boot log Oct 24 21:44:49 note line 31 where the boot loader does "GIGA: athrs17_link_down all port..." Oct 24 21:44:58 after that the switch is essentially dead Oct 24 21:45:34 when i interrupt autoboot by pressing a key it is brought back up again Oct 24 21:47:00 heffer: maybe change the code to make it scan and dump all PHY IDs on the bus Oct 24 21:49:30 sounds like an idea. but first i need to find out how to do that. do you have a hint? Oct 24 21:50:00 something in the direction of ag71xx_phy_connect_multi? Oct 24 21:50:28 but basically i can read the phy ids from sysfs, right? Oct 24 21:53:15 is automount 'remove' in hotplug not working again? Oct 24 21:53:29 https://dev.openwrt.org/ticket/12043 Oct 24 21:53:51 I unplug my USB drive and it keeps everything mounted.. the remove command in hotplug script is 'unmount -l /dev/$device' Oct 24 21:54:07 where device=`basename $DEVPATH` Oct 24 21:55:49 build #401 of uml is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/401 Oct 24 21:57:47 phy ids: http://pastebin.com/tNdXiJET Oct 24 21:58:04 build #315 of iop32x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/315 Oct 24 21:58:40 heffer: nothing in there strikes me as a valid phy id Oct 24 21:58:46 for mdio0 Oct 24 21:58:50 mdio1 is the built-in switch Oct 24 21:59:14 okay so we have a problem there :) Oct 24 21:59:21 any gpl source of the boot loader? Oct 24 21:59:58 build #65 of imx23 is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/imx23/builds/65 Oct 24 21:59:59 there's a gpl source tarball Oct 24 22:00:02 i'll check it Oct 24 22:00:30 the bootloader isn't even part of the tarball i believe Oct 24 22:00:37 maybe it's not necessary Oct 24 22:00:39 at least i didn't see anything Oct 24 22:00:42 maybe the kernel is enough Oct 24 22:02:15 thanks for taking a look, nbd. it's much appreciated Oct 24 22:14:16 funny how the GPL snapshot includes OpenWRT patches :D Oct 24 22:15:03 most of them do these days, and, surprisingly, some of them don't even seem to realize it Oct 24 22:16:44 i guess it's mostly chinese OEMs that make the firmware as well anyway Oct 24 22:16:56 even if WD wanted to help I'm pretty sure they couldn't Oct 24 22:31:19 3.12 support looks busted -> net/netfilter/nf_conntrack_standalone.c:289:4: error: too few arguments to function 'nf_ct_iterate_cleanup' Oct 24 22:39:43 nbd: still looking into the GPL tarball? Oct 24 22:52:41 i'm off for today. thanks a lot for your help so far. if you find anything in the meantime write it and i will read it in the morning. Oct 24 22:52:43 good night :) Oct 24 23:35:10 build #374 of avr32 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/374 Oct 24 23:57:57 more 3.12 breakage -> a lingering parse_err call in 440-block2mtd_init.patch, https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=a1c06609b37c304c58f6d1d3b8e987edaf3391f3 Oct 25 01:49:03 florian r38530 trunk/package/kernel/linux/modules/netdevices.mk * kernel: add mii library kernel module package Oct 25 01:49:13 florian r38531 trunk/package/kernel/linux/modules/netdevices.mk * kernel: make kmod-pcnet32 depend on kmod-mii Oct 25 01:49:22 florian r38532 trunk/target/linux/generic/config-3.12 * kernel: add some missing 3.12 config symbols Oct 25 01:49:31 florian r38533 trunk/include/kernel-defaults.mk * include: handle LZ4 compression scheme Oct 25 01:49:39 florian r38534 trunk/Config.in * buildroot: add lz4 initramfs compression option Oct 25 01:49:48 florian r38535 trunk/target/linux/generic/config-3.12 * kernel: handle Cavium Octeon Kconfig rename Oct 25 01:49:57 florian r38536 trunk/target/linux/malta/config-3.12 * malta: add kernel configuration for 3.12 **** ENDING LOGGING AT Fri Oct 25 02:59:58 2013