**** BEGIN LOGGING AT Sat Jun 14 02:59:58 2014 Jun 14 04:03:35 <5EXAA8SU8> build #194 of mxs is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/mxs/builds/194 Jun 14 04:18:52 I'm trying to troubleshoot an issue with LEDs on an WRT54GS, 'echo x > brightness' has no effect, anyone know how leds-gpio gets loaded on brcm47xx? Jun 14 04:23:20 <5EXAA8SU8> build #91 of bcm53xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/bcm53xx/builds/91 Jun 14 04:28:25 ah, leds-gpio-register.c Jun 14 04:47:30 ugh Jun 14 04:47:42 seems wlan0 overrides leds-gpio control of power and WLAN LEDs Jun 14 04:48:07 cool LED circuit bro Jun 14 05:13:34 <5EXAA8SU8> build #88 of omap is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/88 Jun 14 05:47:42 <5EXAA8SU8> build #89 of adm8668 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/89 Jun 14 07:34:34 <5EXAA8SU8> build #670 of brcm63xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/670 Jun 14 07:37:35 pretty sure trunk is broken on WRT54GS v1.0 Jun 14 07:38:10 both a prebuilt image from snapshots and one I built myself fail to bring up the ethernet and/or switch Jun 14 07:38:18 don't have a serial console on this device yet Jun 14 07:44:03 eigma: i think it should work - maybe you use the wrong profile ? Recently there was a legacy profile introduced https://dev.openwrt.org/changeset/41041/ ; and hauke one of the devs has one wrt54gs with the ADM switch and it was reintroduced with https://dev.openwrt.org/changeset/38699 Jun 14 07:46:07 <5EXAA8SU8> build #89 of brcm2708 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm2708/builds/89 Jun 14 07:48:08 <5EXAA8SU8> build #472 of mpc52xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/472 Jun 14 08:02:44 <5EXAA8SU8> build #493 of adm5120 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/adm5120/builds/493 Jun 14 09:09:07 <5EXAA8SU8> build #90 of mpc83xx is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/90 Jun 14 12:44:04 <5EXAA8SU8> build #91 of realview is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/realview/builds/91 Jun 14 13:25:52 any perl expert here? Jun 14 14:24:59 <5EXAA8SU8> build #89 of cns3xxx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/cns3xxx/builds/89 Jun 14 17:40:31 <5EXAA8SU8> build #595 of rb532 is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/595 Jun 14 17:40:32 <5EXAA8SU8> build #564 of sibyte is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/564 Jun 14 17:41:08 <5EXAA8SU8> build #595 of ppc44x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/595 Jun 14 17:41:11 cyrus r41187 packages/utils/tmux * tmux: moved to github Jun 14 17:41:13 cyrus r41188 packages/utils/lsof * lsof: moved to github Jun 14 17:41:14 cyrus r41189 packages/utils/haserl * haserl: moved to github Jun 14 17:41:15 cyrus r41190 packages/utils/smartmontools * smartmontools: moved to github Jun 14 17:41:16 cyrus r41191 packages/utils/stm32flash * stm32flash: moved to github Jun 14 17:44:34 <5EXAA8SU8> build #502 of iop32x is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/502 Jun 14 18:02:17 plntyk: hmm, I tried building with the legacy profile, no luck; I also tried rolling back to a version that had worked, and that was broken now too, so maybe I have a dirty build tree. but the pre-built images didn't work either, and I think that's really strange. maybe something wrong with my hardware? Jun 14 18:18:32 dirty build tree is possible - "make distclean" makes the tree clean like on the first day Jun 14 18:20:23 in case this will not help - hauke or zajec do/did the work and commits on these targets and they probably can say what additional data they need to detect and fix the problem Jun 14 18:22:18 I've been in touch with zajec a little: https://lists.openwrt.org/pipermail/openwrt-devel/2014-June/025865.html Jun 14 18:23:16 it would be great if one of them could confirm that trunk still works on a WRT54GS 1.0 Jun 14 18:25:00 r41004 is what worked on my device (wrong GPIO layout but that's not a big deal, ethernet+switch were working), now it's not working Jun 14 18:48:26 hm Jun 14 18:48:48 I disabled kmod-b44 along the way at some point.. turned it back on, ethernet seems working Jun 14 18:48:54 will have to try to sync to trunk now Jun 14 18:53:13 any idea why brcm47xx_generic and brcm47xx_legacy have different build dirs? Jun 14 18:58:22 ahhhh ok Jun 14 18:58:31 r41181 is working too now Jun 14 19:01:07 <5EXAA8SU8> build #530 of au1000 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/530 Jun 14 19:01:12 'swconfig dev switch0 show' still makes the device freeze though :( Jun 14 19:01:35 must have something to do with the GPIO layout Jun 14 19:01:43 eigma: brcm47xx_generic runs on all devices of the bcm47xx line and the brcm47xx_legacy subtarget only on old devices Jun 14 19:02:22 yes, but why separate build dirs in build_dir/target-x/linux-brcm47xx_{generic,legacy} Jun 14 19:02:29 makes incremental builds less effective Jun 14 19:04:11 it builds different kernels Jun 14 19:04:36 some options are deactivated in the legacy subtarget Jun 14 19:04:38 <5EXAA8SU8> build #635 of cobalt is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/635 Jun 14 19:04:48 for example BCMA bus, right? Jun 14 19:04:57 <5EXAA8SU8> build #566 of xburst is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/566 Jun 14 19:05:03 yes Jun 14 19:05:35 https://dev.openwrt.org/browser/trunk/target/linux/brcm47xx/legacy/config-default Jun 14 19:06:45 maybe I don't understand the build system too well Jun 14 19:07:10 this kernel config gets added to the defailt brcm47xx kernel config Jun 14 19:07:14 but if you had these different profiles, but always build into build_dir/target-x/linux-brcm47xx, and switch between profiles, incremental build couldn't adjust the output? Jun 14 19:07:37 yes, I understand that you want different configs for the subtargets. I'm just asking why separate *build dirs* Jun 14 19:07:52 <5EXAA8SU8> build #641 of orion is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/641 Jun 14 19:08:19 OpenWrt has to recompile all the packages creating kernel modules Jun 14 19:08:46 and build_dir/target-x/linux-brcm47xx_{generic,legacy} should just contain packages generating kernel modules Jun 14 19:09:04 <5EXAA8SU8> build #630 of lantiq is complete: Failure [failed shell compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/630 Jun 14 19:09:32 <5EXAA8SU8> build #201 of imx6 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/imx6/builds/201 Jun 14 19:10:49 sorry, I don't really understand Jun 14 19:11:37 build_dir/target-x/linux-brcm47xx_{generic,legacy} does not contain only kernel modules, it's the full kernel tree that builds vmlinuz, and then later, modules Jun 14 19:11:54 well, build_dir/target-x/linux-brcm47xx_{generic,legacy}/linux-3.10.36 Jun 14 19:12:22 I think it is possible to change kernel .config from openwrt .config Jun 14 19:12:24 when a kernel config is changed this could and often will also change the kernel abi, so all the kernel modules have to get recompiled Jun 14 19:13:04 you can add modules through the openwrt build system Jun 14 19:13:31 but why do you want to build both subtargets in the same dir? Jun 14 19:13:41 to improve incremental builds Jun 14 19:13:48 <5EXAA8SU8> build #606 of uml is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/606 Jun 14 19:14:32 what do you mean by " incremental builds"? Jun 14 19:14:56 incremental build means you type 'make' a second time, after you built the first time Jun 14 19:15:08 some files have not changed Jun 14 19:15:14 like... b44*.c Jun 14 19:15:26 because subtargets are in different directories, now I have to rebuild that Jun 14 19:15:33 but I already built is Jun 14 19:15:35 *built it Jun 14 19:16:43 and even if config.h changes, meaning change in ABI as you say, the kernel makefiles will see the dependency from b44.c -> config.h and rebuild as needed Jun 14 19:16:48 <5EXAA8SU8> build #547 of kirkwood is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/547 Jun 14 19:16:50 I don't understand why openwrt manually does two separate builds Jun 14 19:19:49 say I change openwrt .config CONFIG_KERNEL_SWAP Jun 14 19:19:53 that is a pretty big change Jun 14 19:20:08 a subtarget is like a target Jun 14 19:20:16 and a target gets its own directory Jun 14 19:20:41 what's the reason for that though Jun 14 19:20:45 yes one could put some subtargets into the same directory but not all Jun 14 19:21:21 how often do you change the subtarget? Jun 14 19:21:40 well, I changed it yesterday because of your changes in trunk Jun 14 19:22:18 it's not the end of the world, it's an extra few minutes of the build when I change it Jun 14 19:22:21 it's fine.. Jun 14 19:22:30 but I want to understand the reason Jun 14 19:22:47 because if incremental builds aren't important at all, then we would always do "make distclean && make" each time Jun 14 19:23:06 but we don't, because incremental builds reduce size of build dir, improve build time Jun 14 19:23:11 distclean remove a lot more stuff Jun 14 19:23:16 and this separation of subtargets in dirs hurts incremental builds Jun 14 19:23:31 we share the compiler and the userspace stuff between all targets when it is possible Jun 14 19:23:37 yes Jun 14 19:23:54 so why not kernels between subtargets? what makes it unshareable Jun 14 19:25:06 for example if I change openwrt CONFIG_KERNEL_SWAP and run "make", all the right things will be rebuilt that are necessary. without cleaning the build dir or changing build dir path Jun 14 19:25:16 same thing should apply to subtargets? Jun 14 19:25:25 it's just a few different kernel .config values Jun 14 19:27:15 maybe it is to speed up builds after changing subtargets Jun 14 19:27:22 for people who build to many subtargets Jun 14 19:27:35 then each of the build trees are already there Jun 14 19:27:37 space-time tradeoff Jun 14 19:28:19 ie. optimization, not correctness issue Jun 14 19:31:15 ~. Jun 14 19:33:45 technically a subtarget is the same as a normal target Jun 14 19:34:13 the different targtes are getting their own dir because they get different patches in most cases Jun 14 19:34:22 /usr/lib/lua/luci/dispatcher.lua:211: /etc/config/luci seems to be corrupt, unable to find section 'main' Jun 14 19:34:31 luci is broken Jun 14 19:35:04 somehow it installs /root/etc/config/luci, after copied it to /etc/config still not working Jun 14 19:38:20 ahh different patches, interesting Jun 14 19:41:01 feed update -a and rebuilding luci Jun 14 20:32:05 <5EXAA8SU8> build #490 of mcs814x is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/490 Jun 14 20:36:59 Hauke: does swconfig work on your WRT54GS? Jun 14 20:41:37 <5EXAA8SU8> build #575 of avr32 is complete: Failure [failed shell compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/575 Jun 14 20:51:07 'swconfig dev switch0 load network' seems to work Jun 14 20:51:17 'swconfig show' is broken though Jun 14 21:06:57 eigma: the last time I check the switch in the WRT54GS v 1.0 worked Jun 14 21:13:05 I think it's related to the GPIO layout Jun 14 21:13:44 when you have a chance to plug in your WRT54GS v1.0 with trunk, I would really be interested to see your /sys/kernel/debug/gpio Jun 14 21:14:04 and again, the switch does work, 'swconfig dev switch0 load network' works, it's "show" that is broken (hangs the device) Jun 14 21:14:26 also 'swconfig dev switch0 port 0 get link' does not hang the device but prints incorrect information Jun 14 21:14:32 'dev switch0 port 0 get mib' hangs Jun 14 21:22:23 I will try OpenWrt on my device tomorrow Jun 14 21:32:01 hm, I'd like to toggle the adm6996 GPIOs to confirm that they're mapped correctly Jun 14 21:32:06 but adm6996.c has already claimed those pins Jun 14 21:32:18 so I can't echo N > /sys/class/gpio/export Jun 14 21:32:23 is there another way I can control them? Jun 14 21:54:04 <5EXAA8SU8> build #90 of pxa is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/90 Jun 14 21:57:21 procd: Instance sysntpd::instance1 s in a crash loop 6 crashes Jun 14 21:57:34 Failed to connect to ubus Jun 14 22:14:25 why did blogic break asiarf awm002evb? Jun 14 22:15:16 also, trying to add back support in asiarf.mk's profiles doesn't show up in the menuconfig Jun 14 22:18:30 r41177 Jun 14 22:23:04 ah, left over cruft in tmp Jun 14 22:28:21 hmm, adm6996 eecs, eesk, edi seem to be mapped correctly. not sure about edo, can't seem to toggle it from /sys/class/gpio Jun 14 22:30:16 ..and that pin seems unused in adm6996.c Jun 14 22:36:25 we have a patch to restore awm002evb Jun 14 22:41:20 wrt54 GS + G, we have many of them.. i will also make a test with trunk next days... most of them are running with white russian here, and our years-old testarea in a village is still running with backfire/b43 :-( Jun 14 22:42:13 ah, thank you Jun 14 22:43:07 beware, sysupgrade saving of config is broken, so you will most likely have to connect to 192.168.1.1 Jun 14 22:49:49 oh, right! my gpio 6 is actually reset button, not adm6996 EDO (called adm_eerc in adm6996.c) Jun 14 22:51:18 thats one of the easiest problems we have to deal with :-) i know, configs are changing all the times, so i always use sysupgrad with -n option.. firmwares are nowadays coming from imagebuilder stuff (freifunk meshkit) with build-in config... Jun 14 22:51:41 tharvey: ping Jun 14 22:53:12 found it.. EDO is on gpio 4. now I think maybe eerc means RESETL Jun 14 23:12:45 hmm, 'port [1-5] get pvid' does work and return correct info Jun 14 23:13:02 as does 'vlan N get ports' Jun 14 23:13:10 so probably just some small parts of the driver are wrong Jun 15 00:22:02 anyone around with commit access to github.com/openwrt/packages? would it be possible to include a LICENSE file in the root stating the license of any contributions? Jun 15 00:45:25 <5EXAA8SU8> build #184 of mvebu is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mvebu/builds/184 Jun 15 01:01:34 <5EXAA8SU8> build #107 of sunxi is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/sunxi/builds/107 Jun 15 01:33:53 Hello, everyone! I just built from trunk and noticed that there's no longer an option to select HT40+/-, just HT40. I haven't been following the project for the last few weeks so I'm a little confused... **** ENDING LOGGING AT Sun Jun 15 02:59:58 2014