**** BEGIN LOGGING AT Mon Mar 23 02:59:59 2015 Mar 23 04:01:21 build #239 of ar71xx.mikrotik is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.mikrotik/builds/239 Mar 23 04:03:28 build #252 of ar71xx.nand is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.nand/builds/252 Mar 23 04:07:56 build #914 of atheros is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/914 Mar 23 05:12:44 build #881 of uml is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/881 Mar 23 05:28:35 build #369 of malta is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/malta/builds/369 Mar 23 06:42:48 build #362 of cns21xx is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/cns21xx/builds/362 Mar 23 06:59:53 build #908 of orion is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/908 Mar 23 07:07:48 build #342 of brcm2708 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm2708/builds/342 Mar 23 07:07:49 build #361 of pxa is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/pxa/builds/361 Mar 23 07:39:39 build #804 of ar71xx is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/804 Mar 23 10:14:54 nbd r44946 trunk/target/linux/ bcm53xx/config-3.14 bcm53xx/patches-3.14 * bcm53xx: remove linux 3.14 support Mar 23 10:15:00 nbd r44947 trunk/target/linux/bcm53xx/config-3.18 * bcm53xx: add missing cache controller errata options Mar 23 12:01:26 build #346 of sunxi is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/sunxi/builds/346 Mar 23 12:32:01 nbd r44948 trunk/package/network/utils/iwinfo/Makefile * iwinfo: update to the latest version, fixes 802.11ac capability reporting Mar 23 12:45:04 kaloz r44949 trunk/target/linux/mvebu/files/arch/arm/boot/dts/armada-xp-mamba.dts * mvebu: update mamba dts Mar 23 12:45:52 kaloz r44950 trunk/package/kernel/ mwlwifi/patches/001-pwrtable_path.patch mwlwifi/Makefile * mwlwifi: update to the new revision Mar 23 13:29:25 kaloz r44951 trunk/target/linux/generic/patches-3.18/ (5 files) * generic/3.18: refresh patches Mar 23 14:03:16 uml build breaking for anyone else? Mar 23 14:03:37 it has been broken week or two for me already I think, due to smp stuff being referenced on non-smp build in the compat-wireless module of kernel Mar 23 14:08:49 jow r44952 branches/barrier_breaker/package/ (5 files in 2 dirs) * BB: openssl: update to v1.0.2a (14 CVEs) Mar 23 14:10:19 idli: yes - your error is probably the same like buildbots: http://buildbot.openwrt.org:8010/builders/uml Mar 23 14:11:19 ah, seems same Mar 23 14:11:25 cyrusff managed to compile it somehow semirecently Mar 23 14:14:49 well if you dont built wireless modules it should compile probably - the buildbots compiles ALL available packages - even with some which might not make sense / dont know if thats the case with uml Mar 23 14:17:01 some new dependency forces that for me, I guess I could try recreating .config from scratch without it Mar 23 14:17:10 funny enough it worked for me even with that enabled Mar 23 14:17:16 i'll try with new clean openwrt tree Mar 23 14:17:54 I did nuke+build, but with same .config couple of weeks ago and it failed same way Mar 23 14:18:28 so who knows what's broken, my guess is that something is making it think it is SMP enabled (the package) while the env is not, but who knows :p Mar 23 14:18:48 did you intentionally enable SMP? Mar 23 14:19:11 or disable Mar 23 14:19:20 don't knwo wahts default Mar 23 14:19:43 I didn't do anything to it Mar 23 14:19:50 okay Mar 23 14:19:52 will see what happens Mar 23 14:19:53 I would like it to be disabled, what option is it? Mar 23 14:20:06 the error sounds like reference to SMP symbol that is defined only on SMP build Mar 23 14:20:23 hmm will have to look Mar 23 14:21:45 compiling right now Mar 23 14:37:32 cyrusff, I see that odhcp6c_update_entry() in odhcp6c.c always returns true. Could this be the cause for https://forum.openwrt.org/viewtopic.php?pid=269643 ? Mar 23 14:38:26 this might cause ra_process() to always return true as well for every received RA. Mar 23 14:39:05 which will result in /lib/netifd/dhcpv6.script being called with ra-updated Mar 23 14:41:38 mwarning: nobody should send RS as response to RAs Mar 23 14:41:47 especially the kernel shouldn't send out any RS Mar 23 14:41:51 if it does, its broken Mar 23 14:42:05 (at least with openwrt's defaults) Mar 23 14:42:33 yes, but I have forwarding disabled. Mar 23 14:42:46 so the system behaves like a host Mar 23 14:43:19 and sends out RS packets - 3 by default I think Mar 23 14:43:42 yep but if you disable forwarding (and still use our architecture) Mar 23 14:43:47 you should disable the kernel sending RAs Mar 23 14:44:01 RSs you mean? Mar 23 14:44:11 right Mar 23 14:44:41 yes, but isn't this a valid setting? (to disable forwarding) Mar 23 14:45:25 mwarning: on openwrt we generally disable all kernel facilities and do everything in userspace Mar 23 14:45:37 (or do it not in case you want to disable it) Mar 23 14:50:34 * mwarning has a flaky connection Mar 23 14:51:27 jow_laptop, I want the router get an address automatically based on RAs on the network Mar 23 14:52:18 setting the kernel option seemed to be the way to go - or I missed the appropriate setting in /etc/confgi/network Mar 23 14:52:30 cyrus r44953 trunk/package/kernel/mac80211/Makefile * mac80211: disable for uml Mar 23 14:52:31 you need a wan6 interface with proto dhcpv6 Mar 23 14:52:39 and leave the kernel sysctls alone Mar 23 14:52:54 dhcpv6 will do either slaac or dhcpv6, whatever is available Mar 23 14:53:04 you cna restrict it to slaac only with further options Mar 23 14:53:20 jow_laptop, ah, how can I do that? Mar 23 14:53:26 so assuming you want to receive ipv6 on lan, you add a new interface lan6 Mar 23 14:53:30 with ifname set to @lan Mar 23 14:53:41 this will stack the ipv6 on top of the lan Mar 23 14:53:49 ack Mar 23 14:54:11 recommended way is to do it in parallel though ;) Mar 23 14:54:17 http://wiki.openwrt.org/doc/uci/network#protocol_dhcpv6 Mar 23 14:54:20 cyrusff, wut? xD Mar 23 14:54:57 well ah its a bridge Mar 23 14:55:02 cyrusff: what was the option for slaac only again? Mar 23 14:55:29 option reqaddress none Mar 23 14:55:31 build #842 of avr32 is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/avr32/builds/842 Mar 23 14:55:46 jow_laptop, ah, I see it now: reqprefix 'no' Mar 23 14:55:59 this will disable the prefix acquisition Mar 23 14:56:16 if you also add reqaddress none it will also disable the acquisition of a dhcpv6 /128 Mar 23 14:56:32 alright, I think I understand now Mar 23 14:56:37 thanks a lot Mar 23 14:57:46 cyrusff, anyway, I think it would be safe for odhcp6c to check before claiming it to be an ra-update :) Mar 23 14:58:19 there was filter at some point under certain conditions Mar 23 14:58:28 but 3 second RA interval is ... excessive Mar 23 14:58:38 build #349 of mpc83xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc83xx/builds/349 Mar 23 14:59:25 well, this way nobody will miss it :) Mar 23 15:00:35 well, it is a funny finding. Really a bit like in a nuclear reactor, but with RSs/RAs :) Mar 23 15:02:43 does anyone know why ipv6 udp socket behavior might have changed between 3.14 and 3.18? Mar 23 15:02:52 russell--: define "changed" Mar 23 15:02:53 build #755 of ep93xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/755 Mar 23 15:04:45 cyrusff: this is olsrd, it opens two sockets, one apparently for listening and the other for transmitting. on 3.14 only the listening socket receives the packet (select returns for that fd), on 3.18 both sockets get alerted and both read the same message. Mar 23 15:05:06 multicast? Mar 23 15:05:10 apparently Mar 23 15:05:18 also there are no "listening only" sockets Mar 23 15:05:40 * russell-- isn't an expert on the internals of olsrd, just has stared at strace output Mar 23 15:05:50 so any software depending on such "features" is inherently broken Mar 23 15:08:01 cyrusff: I got other reports about weird dhcp behaviour Mar 23 15:08:19 someone sent me a pcap dump that appeared to contain duplicate dhcp responses Mar 23 15:08:31 v6? Mar 23 15:08:36 I think there might be a general issue with udp, nothing specifically tied to ipv6 Mar 23 15:08:42 ah Mar 23 15:08:53 well maybe other daemons depend on weird behavior as well Mar 23 15:14:04 udhcpc does weird socket fiddling too, not sure if it would be affected by the same problem Mar 23 15:14:37 cyrusff: shall i email you the strace log excerpts? Mar 23 15:14:43 russell--: so I guess olsrd need some event filtering on the write-only socket Mar 23 15:15:13 it's just kind of weird that it doesn't happen on 3.14 Mar 23 15:15:37 well, nothing in the kernel surprises me anymore Mar 23 15:15:41 lol Mar 23 15:15:47 new round, new regressions Mar 23 15:16:02 they made it very clear that QA work is up to the downstream users now Mar 23 15:16:04 i thought userspace was sacrosanct Mar 23 15:16:30 "don't break userspace" etc Mar 23 15:16:47 mabye its not even an upstream problem but a nasty side-effect of some openwrt specific patches Mar 23 15:16:56 that's possible Mar 23 15:17:00 would be interesting to know if its replicable with a vanilla kernel Mar 23 15:17:09 yes Mar 23 15:17:15 i've been meaning to try that Mar 23 15:17:19 I see Mar 23 15:17:21 but haven't yet Mar 23 15:17:26 damn, need to run again. bbl Mar 23 15:44:45 build #776 of au1000 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/776 Mar 23 15:53:51 build #363 of x86_64 is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/x86_64/builds/363 Mar 23 16:14:03 jow_laptop: still around? Mar 23 16:25:32 jow_laptop: I wanted to ask you about luci2 Mar 23 17:00:29 build #95 of ramips.mt7628 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.mt7628/builds/95 Mar 23 17:00:29 build #275 of brcm47xx.legacy is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.legacy/builds/275 Mar 23 17:11:48 build #230 of x86.xen_domu is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/x86.xen_domu/builds/230 Mar 23 17:28:02 build #922 of brcm47xx is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/922 Mar 23 18:29:42 build #897 of ppc40x is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/897 Mar 23 18:39:50 rmilecki r44954 trunk/target/linux/generic/patches-3.18/076-net-phy-export-fixed_phy_register.patch * kernel: backport fixed_phy_register EXPORT fix from 3.19 Mar 23 18:42:03 build #718 of mpc52xx is complete: Failure [failed shell_10] Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/718 Mar 23 20:16:01 build #729 of gemini is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/729 Mar 23 22:08:19 build #325 of omap is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/325 Mar 23 22:22:37 build #840 of xburst is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/840 Mar 23 22:56:17 cyrusff, I tested the solution we talked about: leaving forwarding on default "1" and using /etc/config/network with dhcpv6 Mar 23 22:56:37 the problem persists and I get a lot of RS packets.. Mar 23 22:57:02 is there something I could try to pinpoint the problem? Mar 23 22:58:02 as I have said, /lib/netifd/dhcpv6.script is called with "ra-updated" as argument every few seconds non stop Mar 23 23:52:39 build #236 of lantiq.xrx200 is complete: Failure [failed compile_2] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq.xrx200/builds/236 Mar 24 02:57:47 build #240 of ar71xx.mikrotik is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.mikrotik/builds/240 **** ENDING LOGGING AT Tue Mar 24 02:59:58 2015