**** BEGIN LOGGING AT Wed Jun 06 02:59:58 2012
Jun 06 03:51:42 <builder_> build #30 of rb532 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/30
Jun 06 03:55:17 <builder_> build #27 of uml is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/27
Jun 06 04:26:34 <swalker> are there targets that benefit from busybox hwclock being enabled that don't already depend on util-linux hwclock?
Jun 06 05:05:58 <builder_> build #24 of au1000 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/24
Jun 06 06:01:50 <builder_> build #27 of x86 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/27
Jun 06 06:13:25 <builder_> build #25 of iop32x is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/25
Jun 06 07:23:20 <builder_> build #22 of ep93xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/22
Jun 06 07:42:23 <builder_> build #31 of s3c24xx is complete: Failure [failed compile_4 shell_15]  Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/31
Jun 06 08:15:33 <builder_> build #22 of etrax is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/22
Jun 06 08:20:43 <builder_> build #22 of gemini is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/22
Jun 06 08:21:24 * russell-- is seeing squashfs problems on both wgt634u and alix2
Jun 06 08:21:43 <loswillios> russell_: which revision?
Jun 06 08:22:10 <russell--> r32074
Jun 06 08:23:06 <russell--> [   29.317387] SQUASHFS error: xz_dec_run error, data probably corrupt
Jun 06 08:23:06 <russell--> [   29.336281] SQUASHFS error: squashfs_read_data failed to read block 0x32d10a
Jun 06 08:23:10 <russell--> [   29.357505] SQUASHFS error: Unable to read fragment cache entry [32d10a]
Jun 06 08:23:44 <russell--> i've seen something pretty close to identical on wgt634u from the same rev
Jun 06 08:28:21 <russell--> yeah, same message, only the block number differs
Jun 06 08:30:55 <loswillios> hm, can't reach my wgt at the moment. No ipaddress, but the wifi->lan bridge seems to be working. Probably something netifd related
Jun 06 08:31:47 <builder_> build #34 of brcm63xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/34
Jun 06 08:35:06 <russell--> also, the alix2 build is giving me a config question three times per build:   LED Support for Soekris net5501 series Error LED (LEDS_NET5501) [N/m/?] (NEW)
Jun 06 08:39:40 <russell--> interesting, reflashed, problem went away.  it seemed intermittent on wgt634u also.
Jun 06 08:52:40 <builder_> build #31 of at91 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/31
Jun 06 09:06:46 <builder_> build #33 of brcm47xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/33
Jun 06 09:41:08 <builder_> build #31 of atheros is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/31
Jun 06 09:48:13 <builder_> build #34 of cobalt is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/34
Jun 06 10:43:25 <builder_> build #27 of pxcab is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/27
Jun 06 10:47:41 <builder_> build #26 of ar7 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/26
Jun 06 10:55:07 <builder_> build #29 of ppc44x is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/29
Jun 06 11:12:04 <builder_> build #27 of sibyte is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/27
Jun 06 11:13:02 <builder_> build #29 of orion is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/29
Jun 06 11:28:18 <stintel> did anyone try to register openwrt here ? http://www.worldipv6launch.org/form/?q=3
Jun 06 11:30:12 <jow_laptop> our ipv6 support is not exactly mature
Jun 06 11:30:56 <jow_laptop> and I believe we do not meet the "IPv6 Ready CE Router (CPE) Interoperability Test Scenario"
Jun 06 11:31:14 <stintel> ok
Jun 06 11:32:03 <jow_laptop> so better keep it low profile until we can actually deliver :)
Jun 06 11:34:25 <stintel> bugger
Jun 06 11:35:23 <builder_> build #25 of xburst is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/25
Jun 06 11:36:27 <dape-work> jow_laptop:  your ipv6 support is excelent since november 2011, latest trunks are awesome
Jun 06 11:36:40 <jow_laptop> 6rd is unsupported
Jun 06 11:36:43 <jow_laptop> ds-lite is unsupported
Jun 06 11:36:52 <jow_laptop> dhcpv6-pd is unsupported
Jun 06 11:36:57 <dape-work> ok, let me rephrase: your pppoe dual stack works cool
Jun 06 11:37:03 <jow_laptop> that is like 90% of real world deployments
Jun 06 11:37:14 <stintel> especially dhcpv6-pd yes
Jun 06 11:37:15 <dape-work> dhcpv6 pd works for me
Jun 06 11:37:33 <stintel> that's how my ISP is planning on doing it for clients with own router
Jun 06 11:37:52 <stintel> jow_laptop: 6rd ... I had some script, but from before netifd time
Jun 06 11:37:58 <dape-work> i wish my isp will use openwrt routers..
Jun 06 11:38:13 <stintel> dape-work: CPE, maybe. core/edge/whatever else, please no
Jun 06 11:38:13 <jow_laptop> stintel: we need to fork 6to4 and make 6rd out of id
Jun 06 11:38:20 <stintel> jow_laptop: let me see
Jun 06 11:38:27 <stintel> I have something already
Jun 06 11:38:28 <jow_laptop> but it annoys me soo much that my isp is stuck in the dark ages
Jun 06 11:38:37 <jow_laptop> no ipv6 in sight
Jun 06 11:38:42 <stintel> my ISP has 6rd
Jun 06 11:38:50 <stintel> but it's employee only, non-production
Jun 06 11:39:10 <stintel> fortunately I work there for the moment ;)
Jun 06 11:39:16 <dape-work> stintel:  ofc cpe
Jun 06 11:39:19 <stintel> so I am actually able to test it :)
Jun 06 11:39:52 <stintel> unless maybe openvswitch starts having hardware  support and maybe we can someday run openwrt on switches too ;)
Jun 06 11:40:36 <stintel> ah meh, I'll have to build a recent trunk for my router @ home first
Jun 06 11:40:38 <dape-work> jow_laptop:  maybe change isp if possible
Jun 06 11:40:59 <jow_laptop> no it is not
Jun 06 11:41:06 <stintel> sucks :(
Jun 06 11:41:54 <dape-work> thats a shame, openwrt devs deserve that ipv6 native access
Jun 06 12:18:43 <stintel> is there a way to specify extra processes that shouldn't be killed on sysupgrade ?
Jun 06 12:18:57 <stintel> (other than editing the script)
Jun 06 12:19:08 <stintel> it just killed my bird4 and I have lost connectivity now
Jun 06 12:19:33 <jow_laptop> no its not possible
Jun 06 12:20:26 <stintel> could be handy to have a sysupgrade-do-not-kill.d directory where packages can drop a file if they shouldnt be killed? or too much hassle
Jun 06 12:20:34 <jow_laptop> too much hassle imo
Jun 06 12:21:01 <jow_laptop> there'll allways be weird setup which will still break even if we implement whitelisting
Jun 06 12:21:18 <stintel> so adding bird(4|6) (bgp|ospf|..)d in the script an option ?
Jun 06 12:21:24 <jow_laptop> but why does terminating bird breaks sysupgrade?
Jun 06 12:21:27 <stintel> I wouldn't call running routing protocols on a router weird setups
Jun 06 12:21:32 <jow_laptop> did you attempt to upgrade from remote?
Jun 06 12:21:37 <stintel> I did
Jun 06 12:21:42 <stintel> I am not within the same subnet
Jun 06 12:21:48 <jow_laptop> then I'd rather remove that feature
Jun 06 12:22:08 <jow_laptop> because it does not work asnyway
Jun 06 12:22:13 <stintel> what do you mean ?
Jun 06 12:22:15 <jow_laptop> at least not for combined images and the like
Jun 06 12:23:30 <jow_laptop> I mean that streaming remote images to the local write process always has been fragile and its fundamentally broken for verious platfrom/image combos
Jun 06 12:24:39 <stintel> jow_laptop: I ran sysupgrade /tmp/openwrt-something
Jun 06 12:24:41 <stintel> via ssh
Jun 06 12:24:58 <stintel> it actually flashed the new image, fortunately
Jun 06 12:25:14 <jow_laptop> yes, then sysupgrade should be changed to fork into background and lock out the user
Jun 06 12:25:29 <jow_laptop> imo its pointless trying to keep the controlling connection open, too fragile, too dangerous
Jun 06 12:25:44 <stintel> ah
Jun 06 12:26:06 <stintel> but you can't see if anything goes wrong then
Jun 06 12:26:14 <stintel> sensitive subject it seems :)
Jun 06 12:26:25 <jow_laptop> the thing is
Jun 06 12:26:47 <jow_laptop> service killing has been implemented to get rid of services causing rootfs i/o which leads to kernel oopses during write
Jun 06 12:27:25 <jow_laptop> if we now whitelist everything again, like bind, bird, openvpn, tinc, httpd, whatever we can as well ditch it
Jun 06 12:27:27 <stintel> didn't know that
Jun 06 12:27:49 <stintel> I'll just connect to a machine in the same subnet next time
Jun 06 12:27:54 <stintel> s/to/from/
Jun 06 12:28:00 <builder_> build #26 of ixp4xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/26
Jun 06 12:28:02 <stintel> or w/e, depends how you look at it
Jun 06 12:28:03 <jow_laptop> thats probably the most sensitive approach
Jun 06 12:35:12 <builder_> build #23 of rdc is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/rdc/builds/23
Jun 06 12:54:47 <builder_> build #25 of kirkwood is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/25
Jun 06 13:09:42 <builder_> build #22 of adm5120 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/adm5120/builds/22
Jun 06 13:27:13 <CIA-84> blogic * r32075 /trunk/ (5 files in 3 dirs): [tools] add ccache wrapper
Jun 06 13:30:05 <stintel> ah damn, looks like trunk r32074 no longer boots on my rspro
Jun 06 13:30:11 <stintel> I'll have to debug tonight
Jun 06 13:30:18 <jow_laptop> did you defconfig ?
Jun 06 13:31:17 <stintel> I did not, but after doing it and then diffing it to previous .config, I get only timestamp different
Jun 06 13:31:25 <stintel> (netifd was already enabled in my previous buuild)
Jun 06 13:31:26 <CIA-84> florian * r32076 /trunk/target/linux/brcm63xx/patches-3.3/ (4 files):
Jun 06 13:31:26 <CIA-84> [brcm63xx] increase led number in patch 301-led_count.patch
Jun 06 13:31:26 <CIA-84> So that board patches are only touching board-related code
Jun 06 13:31:32 <CIA-84> florian * r32077 /trunk/target/linux/generic/patches-3.3/047-spi_message_queue.patch: [generic/3.3] backport spi message queue infrastructure
Jun 06 13:31:35 <CIA-84> florian * r32078 /trunk/target/linux/brcm63xx/patches-3.3/ (3 files):
Jun 06 13:31:35 <CIA-84> [brcm63xx] backport upstream spi-bcm63xx fixes
Jun 06 13:31:35 <CIA-84> * message pump conversion
Jun 06 13:31:35 <CIA-84> * not using stopping state
Jun 06 13:31:35 <CIA-84> * setting spi driver mode bits
Jun 06 13:31:46 <stintel> should have connected that poe injector to a plugwise module
Jun 06 13:32:19 <stintel> when my poe switch finally arrives I can probably just powercycle it via the switch
Jun 06 13:45:55 <CIA-84> nbd * r32079 /trunk/include/download.mk: download.mk: refuse to download mirror tarballs from packages using version control if no mirror md5sum was provided
Jun 06 13:45:58 <CIA-84> nbd * r32080 /trunk/package/libubox/Makefile: libubox: add mirror md5sum
Jun 06 13:46:02 <CIA-84> nbd * r32081 /trunk/include/toplevel.mk: build: prevent the environment variable CROSS_COMPILE from leaking into builds (fixes #11603)
Jun 06 14:02:45 <dape-work> guys please help, make breaks at this: http://dpaste.com/755975/
Jun 06 14:03:00 <dape-work> Cloning into yaffs2_android...
Jun 06 14:03:00 <dape-work> android.git.kernel.org[0: 149.20.4.77]: errno=Connection refused
Jun 06 14:04:16 <stintel> I don't think anyone here can help with that
Jun 06 14:04:28 <nbd> actually we can
Jun 06 14:04:46 <nbd> that repo went away a while ago, unnoticed until my commit r32079 uncovered it
Jun 06 14:04:49 <stintel> ah
Jun 06 14:04:58 <dape-work> thx nbd
Jun 06 14:08:49 <CIA-84> nbd * r32082 /trunk/include/host-build.mk: build: add PKG_MIRROR_MD5SUM support for host builds
Jun 06 14:09:01 <CIA-84> nbd * r32083 /trunk/tools/yaffs2/Makefile: tools/yaffs2: add mirror md5sum - upstream repo went away
Jun 06 14:09:09 <nbd> dape-work: there, should work now
Jun 06 14:09:27 <dape-work> thx, compiling now !
Jun 06 14:47:52 <builder_> build #32 of s3c24xx is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/s3c24xx/builds/32
Jun 06 15:25:55 <CIA-84> nbd * r32084 /trunk/package/hostapd/files/wpa_supplicant.sh:
Jun 06 15:25:55 <CIA-84> hostapd: fix driver setting for wpa_s in IBSS/WPA-NONE
Jun 06 15:25:55 <CIA-84> This fix is needed to correctly set the driver to use with wpa_s in case of
Jun 06 15:25:55 <CIA-84> IBSS/WPA-NONE. The current setting is not made correctly.
Jun 06 15:25:55 <CIA-84> Signed-off-by: Antonio Quartulli <ordex@autistici.org>
Jun 06 15:25:58 <CIA-84> nbd * r32085 /trunk/ (Config.in target/linux/generic/config-3.3): kernel: make tracing support configurable
Jun 06 15:44:53 <CIA-84> florian * r32086 /trunk/target/linux/brcm63xx/patches-3.3/ (26 files): [brcm63xx] fix leds naming for all bcm63xx boards
Jun 06 15:44:58 <CIA-84> florian * r32087 /trunk/target/linux/brcm63xx/patches-3.3/513-board_livebox.patch: (log message trimmed)
Jun 06 15:44:58 <CIA-84> [brcm63xx] update livebox board support (#8677)
Jun 06 15:44:58 <CIA-84> Merge most fixes from #8677:
Jun 06 15:44:58 <CIA-84> - add basic hardware detection of Livebox Blue 5g revisions
Jun 06 15:44:59 <CIA-84> - register leds and gpio buttons
Jun 06 15:44:59 <CIA-84> - fix boot address location
Jun 06 15:45:00 <CIA-84> - properly parse mac addresses
Jun 06 15:45:06 <CIA-84> florian * r32088 /trunk/target/linux/brcm63xx/ (3 files in 2 dirs): [brcm63xx] add 96348w3 board (#10764)
Jun 06 16:02:17 <builder_> build #32 of at91 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/32
Jun 06 16:07:47 <CIA-84> florian * r32089 /trunk/target/linux/brcm63xx/patches-3.3/516-board_96348A-122.patch: [brcm63xx] fix alarm led color for board 96348A-122
Jun 06 16:08:57 <dape-work> im getting ready to order a TP-LINK TL-MR3040 from DX, if anyone had it working with openwrt trunk i appreciate if any hints if it needs any special care..
Jun 06 16:15:23 <builder_> build #30 of orion is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/30
Jun 06 16:17:06 <russell--> Set the kernel architecture to x86
Jun 06 16:17:06 <russell--> configure: error: The lib_ifxos include directory is not valid!
Jun 06 16:17:06 <russell--> make[3]: *** [/aux/src/openwrt-fresh/build_dir/linux-x86_alix2/drv_tapi-3.13.0/.configured_] Error 1
Jun 06 16:17:52 <nbd> russell--: blogic is working on it
Jun 06 16:18:08 <russell--> sweet
Jun 06 16:20:13 <builder_> build #35 of cobalt is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/cobalt/builds/35
Jun 06 16:20:25 <builder_> build #34 of brcm47xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/34
Jun 06 16:21:10 <builder_> build #32 of atheros is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/atheros/builds/32
Jun 06 16:55:12 <builder_> build #35 of brcm63xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx/builds/35
Jun 06 17:00:39 <builder_> build #28 of x86 is complete: Failure [failed compile_1]  Build details are at http://buildbot.openwrt.org:8010/builders/x86/builds/28
Jun 06 17:05:50 <builder_> build #28 of ramips is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/28
Jun 06 17:24:03 <CIA-84> nbd * r32090 /trunk/target/linux/ar71xx/image/Makefile: ar71xx: only build all loader images if the image builder is enabled - saves some time on make target/install
Jun 06 17:24:07 <CIA-84> nbd * r32091 /trunk/target/Makefile: target: do not make target/*/install depend on target/*/compile - removes one redundant kernel build dir call on target/install
Jun 06 17:24:11 <CIA-84> nbd * r32092 /trunk/target/linux/ar71xx/files/arch/mips/ath79/ (dev-eth.c dev-eth.h mach-tl-wr741nd-v4.c): ar71xx: add a helper function for setting up PHY4 swapping on ar933x
Jun 06 17:24:15 <CIA-84> nbd * r32093 /trunk/target/linux/ar71xx/files/arch/mips/ath79/mach-tl-mr3020.c: ar71xx: fix PHY setup on TL-MR3020 (fixes #11572)
Jun 06 17:30:14 <blogic> nbd: thx
Jun 06 17:30:29 <stintel> jow_laptop: heh, I rebooted the rspro and now comes up fine. weird
Jun 06 17:30:39 <stintel> never saw what happened the previous boot
Jun 06 17:56:35 <jow_laptop> swalker: what where the missing icmp types again?
Jun 06 17:57:06 <builder_> build #28 of uml is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/28
Jun 06 18:04:03 <CIA-84> nbd * r32094 /trunk/target/linux/generic/config-3.3: kernel: add a few more tracing related symbols
Jun 06 18:05:34 <stintel> hurray for backups, found my 6rd config
Jun 06 18:05:50 <stintel> now lets see if I can make 6rd.sh out of 6in4.sh
Jun 06 18:13:49 <stintel> but dinner first
Jun 06 18:24:54 <builder_> build #28 of pxcab is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/pxcab/builds/28
Jun 06 18:26:07 <builder_> build #28 of ppc40x is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc40x/builds/28
Jun 06 18:30:00 <builder_> build #31 of rb532 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/31
Jun 06 18:37:41 <builder_> build #30 of ppc44x is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/30
Jun 06 19:13:35 <dape> geez, seagate goflex openwrt support? that would be awesome !
Jun 06 19:18:58 <stintel> right, netifd was completely different, it doesn't use ip tunnel command in shell scripts
Jun 06 19:19:08 <stintel> so far my attempt at 6rd
Jun 06 19:24:20 <builder_> build #28 of sibyte is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/28
Jun 06 19:32:29 <stintel> ok I see now how the ip tunnel add works via netifd
Jun 06 19:33:21 <stintel> but how would I go about running "ip tunnel 6rd dev $interface 6rd-prefix $prefix
Jun 06 19:37:50 <CIA-84> nbd * r32095 /trunk/target/linux/ar71xx/image/Makefile: ar71xx: fix LZMA loader build regression from r32090 (thx, juhosg)
Jun 06 19:44:29 <stintel> nbd: can I bug you for a moment, about netifd ?
Jun 06 19:49:31 <builder_> build #27 of ar7 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/27
Jun 06 20:10:09 <stintel> ok, wonder why I didnt think of this earlier, I can just run that from the shell script I guess
Jun 06 20:30:29 <builder_> build #26 of xburst is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/26
Jun 06 20:30:36 <loswillios> dape: yes
Jun 06 20:31:20 <dape> loswillios: if you're replying about seage goflex.. i have one and eager to flash it.. somehow
Jun 06 20:33:23 <loswillios> I'm running openwrt on a dockstar for quite some time now, which is essentially the same as the goflex IIRC
Jun 06 20:34:10 <dape> uhm, i got the home 1 tb one, its not the dual slot..
Jun 06 20:34:42 <dape> anyways, im stalking here and the list for trunk support :P
Jun 06 20:36:01 <builder_> build #26 of kirkwood is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/kirkwood/builds/26
Jun 06 20:39:37 <builder_> build #27 of ixp4xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/27
Jun 06 20:39:59 <builder_> build #25 of au1000 is complete: Failure [failed compile_4]  Build details are at http://buildbot.openwrt.org:8010/builders/au1000/builds/25
Jun 06 20:44:15 <swalker> jow_laptop: at least {neighbour,router}-advertisement
Jun 06 20:44:41 <stintel> nice, 6rd working partially
Jun 06 20:53:38 <stintel> if anyone cares to have a look: http://stewie.be.tintel.eu/6rd.sh
Jun 06 20:56:21 <builder_> build #25 of ar71xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/25
Jun 06 20:58:49 <CIA-84> florian * r32096 /trunk/target/linux/brcm63xx/patches-3.3/553-board_96348w3.patch: [brcm63xx] fix 553-board_96348w3.patch patch wrt gpio-keys-polled
Jun 06 21:14:17 <nbd> stintel: sure
Jun 06 21:15:22 <stintel> nbd: ah well, I look a bit further and I found what I needed. wrote the 2nd part to configure the 6rd tunnel in the 6rd.sh
Jun 06 21:15:37 <stintel> hope this all doesn't look too bad: 2601:d:a980:3b:1904:b9f9:2568:6ab
Jun 06 21:15:39 <stintel> ehr
Jun 06 21:15:40 <stintel> lol
Jun 06 21:15:45 <stintel> http://stewie.be.tintel.eu/6rd.sh
Jun 06 21:16:20 <stintel> based it on 6in4.sh
Jun 06 21:16:50 <stintel> don't know who's IP that is but he/she also checked that script :)
Jun 06 21:18:08 <nbd> you should be able to just remove that ip line
Jun 06 21:18:19 <nbd> netifd creates the tunnel
Jun 06 21:18:27 <stintel> it does, but wait
Jun 06 21:18:32 <stintel> in my previous script, it had two line
Jun 06 21:18:47 <stintel>                 ip tunnel add $link mode sit local $local4 remote $remote4 ttl ${ttl:-64}
Jun 06 21:18:51 <stintel>                 ip tunnel 6rd dev $link 6rd-prefix $sixrdprefix::/$sixrdprefixlen
Jun 06 21:19:03 <stintel> so the tunnel needs to be create. then set the 6rd prefix
Jun 06 21:19:07 <nbd> ah
Jun 06 21:19:08 <nbd> ok
Jun 06 21:19:33 <nbd> i could put that into netifd
Jun 06 21:19:42 <stintel> this works for my isp's 6rd setup, would be nice if  someone else could test it
Jun 06 21:20:31 <nbd> i'll see if i can put the 6rd part into netifd tomorrow and then we could just commit the script afterwards and let people start testing it
Jun 06 21:20:53 <stintel> sounds great
Jun 06 21:22:11 <stintel> my isp is actually testing ipv6 deployment with openwrt routers
Jun 06 21:22:42 <nbd> cool
Jun 06 21:23:15 <stintel> I'm working there as a contractor currently, which is how I got hold of the 6rd server IP
Jun 06 21:23:27 <stintel> too bad I can't share it with friends
Jun 06 21:34:34 <stintel> nbd: in my case it even supports routed subnet, not sure if that is always the case though
Jun 06 21:35:26 <jow_laptop> routed subnet can be expressed as generic route6
Jun 06 21:35:28 <stintel> I just configured $sixrdprefix:ffff:ffff:ffff:ffff/64 on br-lan and announced $rdprefix with bird, and pcs behind router have ipv6 connectivity
Jun 06 21:35:35 <jow_laptop> that does not need to be handled in 6rd specifically
Jun 06 21:35:45 <stintel> ok
Jun 06 21:36:43 <stintel> but, in this case, the prefix has the v4 IP in hex in it. so if you have a dynamic IP, the prefix will change
Jun 06 21:36:47 <stintel> can that be handled somehow ?
Jun 06 21:36:56 <jow_laptop> hm
Jun 06 21:37:26 <jow_laptop> yeah, in this case it might make sense to handle it more like an interface route
Jun 06 21:37:36 <jow_laptop> means its implicitely created and destroyed
Jun 06 21:37:45 <jow_laptop> along with the 6rd state
Jun 06 21:37:54 <jow_laptop> s/state/interface/
Jun 06 21:38:59 <stintel> previously I did it like this in 6rd.sh:                 local lanip6="$sixrdprefix:$local4hex:ffff:ffff:ffff:ffff/64"
Jun 06 21:39:12 <stintel> but that's ugly ofc
Jun 06 21:40:22 <jow_laptop> why the fff ?
Jun 06 21:40:41 <jow_laptop> ah, just an arbritary choice
Jun 06 21:40:44 <stintel> with v4, I gave my routers the last usable ip in the subnet, took over that habit to v6
Jun 06 21:40:51 <stintel> slightly more typing though
Jun 06 21:40:55 <stintel> ;)
Jun 06 21:40:59 <jow_laptop> well
Jun 06 21:42:00 <jow_laptop> ok, if we substract $sixrdprefixlen from 64 and get a result > 0
Jun 06 21:42:19 <jow_laptop> then we know the number of avilable subnets
Jun 06 21:42:47 <jow_laptop> we wil lthen somehow need a mechanism to simpy assign one /64 slice to each internal iface
Jun 06 21:43:51 <stintel> btw, maybe someone in here has an ISP that supports 6rd ?
Jun 06 21:44:49 <jow_laptop> find_6rd_wanif() is a candidate for /lib/functions/network.sh
Jun 06 21:45:21 <jow_laptop> as  network_get_wandev()  and a complementary  network_get_wandev6()
Jun 06 21:45:37 <jow_laptop> it is also used by 6in4 and 6to4
Jun 06 21:45:41 <stintel> yeah
Jun 06 21:46:13 <jow_laptop> however for 6rd we should be able to get along without
Jun 06 21:46:31 <jow_laptop> we can register a route dependency on the ipv4 6rd relay server address
Jun 06 21:47:39 <jow_laptop> ( proto_add_host_dependency "$cfg" "$remote4" )
Jun 06 21:48:25 <jow_laptop> and then remove all the wan iface stuff
Jun 06 21:48:57 <jow_laptop> see pptp.sh for reference
Jun 06 21:49:28 <jow_laptop> ah, wait - we still need the wan iface handling to find the local ip
Jun 06 22:00:36 <stintel> jow_laptop: something like this for the first part ? https://gist.github.com/2885103
Jun 06 22:01:27 <danmackay> hmm, well I got my 6rd working
Jun 06 22:01:31 <stintel> have not tested this yet, just checking if that looks ok :)
Jun 06 22:01:37 <stintel> danmackay: with that script ?
Jun 06 22:01:55 <danmackay> something posted on the forum in whatever topic it was in
Jun 06 22:01:59 <stintel> ah
Jun 06 22:02:26 <stintel> are you running recent trunk with netifd on that openwrt ?
Jun 06 22:02:37 <danmackay> hmm, doesnt show under ipv6 wan status though -_-
Jun 06 22:02:56 <danmackay> truck build from the 31st of may
Jun 06 22:03:09 <jow_laptop> stintel: not quite, I am currently implementing the needs functions
Jun 06 22:03:25 <jow_laptop> stintel: especially I want to get rid of the subshells
Jun 06 22:04:32 <stintel> good that I checked :)
Jun 06 22:06:26 <stintel> are you saying you are already doing what you suggested above ? I didn't get what you meant with the needs functions
Jun 06 22:07:07 <stintel> brb, grabbing a beer
Jun 06 22:32:15 <jow_laptop> stintel: well I wanted to use jshn and ubus calls to find the wandev but that will fail if the wan iface is not a netifd managed interface
Jun 06 22:32:31 <stintel> ah
Jun 06 22:33:08 <jow_laptop> I feel uncomfortable scraping "ip"
Jun 06 22:35:19 <stintel> jow_laptop: do you feel more comfortable scraping netstat -rn or route -n ?
Jun 06 22:35:25 <jow_laptop> also I am not sure whetherr we should consider default routes from all tables
Jun 06 22:35:35 <stintel> in case of multiwan
Jun 06 22:35:42 <jow_laptop> for example, yes
Jun 06 22:39:49 <stintel> we could ald a simple option to let the user choose which wan device he wants the tunnel on
Jun 06 22:40:00 <jow_laptop> its already there
Jun 06 22:40:17 <jow_laptop> ops nope
Jun 06 22:40:27 <stintel> if you have static ip maybe
Jun 06 22:40:31 <stintel> no, not even
Jun 06 22:41:25 <jow_laptop> json_get_var wanif wandev || wanif=$(find_6rd_wanif)
Jun 06 22:43:49 <stintel> is there a specific reason to have wanif vs wandev ?
Jun 06 22:44:10 <jow_laptop> yeah, I try to clear up the temrinology a bit
Jun 06 22:44:15 <nbd> i think we should extend the netifd api so that these scripts do not have to look up stuff themselves
Jun 06 22:44:22 <jow_laptop> devices are the linux things
Jun 06 22:44:31 <jow_laptop> interfaces are the openwrt descriptions
Jun 06 22:45:02 <nbd> finding the wan device should be done in a way similar to the pptp host dependency thing
Jun 06 22:45:07 <jow_laptop> yes
Jun 06 22:50:42 <builder_> build #26 of iop32x is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/26
Jun 06 22:51:50 <stintel> dynamic routed subnet is pretty lame though
Jun 06 22:52:16 <stintel> dns horror, meh.
Jun 06 22:52:37 <stintel> maybe I can fix it with a mysql procedure on my pdns db, but wtf
Jun 06 22:54:40 <builder_> build #23 of gemini is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/23
Jun 06 22:54:59 <builder_> build #23 of ep93xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/23
Jun 06 22:57:55 <jow_laptop> stintel: I think for 6rd it is supposed to get pushed via dhcp(v4)
Jun 06 22:59:21 <stintel> makes sense, if user needs to configure it, the name rapid deployment is kinda fail
Jun 06 22:59:55 <jow_laptop> we haven't that figured out either yet
Jun 06 23:00:11 <jow_laptop> I mena how to propagate 6rd info from dhcpv4 to the 6rd iface
Jun 06 23:00:16 <stintel> meh, I hope that wont mean loads of isps will just rollout 6rd via pd
Jun 06 23:00:31 <stintel> there is the possibility to add custom dhcp options, afaik
Jun 06 23:00:55 <stintel> it's still tunneled then, I don't really care for that
Jun 06 23:01:01 <jow_laptop> however netifd can probably just store the dhcp 6rd option in the status json
Jun 06 23:01:04 <stintel> but for now is all I have
Jun 06 23:01:08 <jow_laptop> then when a 6rd iface is decalred without any options set
Jun 06 23:01:27 <jow_laptop> it will simply dump the status of all active ifaces and looks whether one has an 6rd tag set
Jun 06 23:01:37 <jow_laptop> and take the info from there
Jun 06 23:02:22 <stintel> and then if you have multiple internal layer 2 networks, you'll need the ability to specify how many prefixes you want
Jun 06 23:02:27 <stintel> meh
Jun 06 23:02:35 <jow_laptop> yes
Jun 06 23:03:06 <jow_laptop> for 6to4 I implemented the adv_interface
Jun 06 23:03:20 <stintel> sigh, 6rd is just another reason for isps not to start supporting native yet
Jun 06 23:03:22 <jow_laptop> that is basically a list of interfaces you want to announce v6 on
Jun 06 23:03:44 <jow_laptop> plus another adv_subnet option to pick the first prefix offset
Jun 06 23:04:11 <jow_laptop> then it iterates over the interfaces and incrementally assigns prefixes beginning with the start offset
Jun 06 23:04:15 <stintel> ic, well never used 6to4
Jun 06 23:04:19 <jow_laptop> that is felixble enough imo
Jun 06 23:04:24 <stintel> 6to4 is for v6 only clients to connect to v4 networks right ?
Jun 06 23:04:40 <jow_laptop> no, its just the passive version of 6rd
Jun 06 23:04:53 <builder_> build #24 of rdc is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/rdc/builds/24
Jun 06 23:04:55 <stintel> hmmm
Jun 06 23:04:56 <builder_> build #23 of etrax is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/etrax/builds/23
Jun 06 23:05:04 <jow_laptop> with anycast relay servers and v4-derived v6 subnet instead of explicitely configured values
Jun 06 23:06:09 <stintel> ah, what windows does automatically, teredo thing
Jun 06 23:06:32 <stintel> reminds me of a guy claming he had native v6 from my isp
Jun 06 23:07:20 <stintel> because the whois info returned some info about the v4 address mapped in his v6 address
Jun 06 23:09:49 <jow_laptop> 6rd is basically 6to4 with defined relay server(s) and defined ipv6 prefix
Jun 06 23:11:37 <jow_laptop> eventually we'll also need a "remote controllable" radvd
Jun 06 23:11:52 <jow_laptop> and/or a "remoate controllable" dhcpv6 server
Jun 06 23:12:19 <jow_laptop> means you can instruct the running process to dynamically add or remove prefixes to advertise
Jun 06 23:13:18 <Kevin`> jow_laptop: I got the package
Jun 06 23:13:27 <jow_laptop> Kevin`: good
Jun 06 23:14:31 <Kevin`> do you remember where that document the guy put together is?
Jun 06 23:14:50 <jow_laptop> http://bcm-v4.sipsolutions.net/mac-gbit
Jun 06 23:15:09 <builder_> build #23 of adm5120 is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/adm5120/builds/23
Jun 06 23:19:04 <Kevin`> ugh, looks nasty
Jun 06 23:19:37 <Kevin`> at least there's a pretty clean table of registers
Jun 06 23:27:57 <Kevin`> (aside from documenting what they all do :))
Jun 06 23:44:20 <stintel> jow_laptop: and if you want to support managed radvd, maybe also support managed bird6, quagga?
Jun 06 23:44:29 <stintel> who knows what else is there
Jun 06 23:44:45 <jow_laptop> yeah, but thats probably a lot of work
Jun 06 23:45:02 <stintel> if you already run quagga or bird for something, it makes no sence to run radvd next to it just for v6 advertising
Jun 06 23:45:03 <jow_laptop> maybe not for quagga
Jun 06 23:45:10 <jow_laptop> don't know bird6
Jun 06 23:45:29 <jow_laptop> radvd needs some control socket thing
Jun 06 23:45:41 <jow_laptop> and dnsmasq probably as well if it provides dhcpv6
Jun 06 23:48:14 <stintel> sounds like a lot of work
Jun 06 23:48:52 <jow_laptop> rewriting /var/etc/whatever and sending sighup is probably a workable intermediate solution
Jun 06 23:49:53 <jow_laptop> I mean if the wan ip changes and the network state is recalculated, every session is interrupted anyway so it doesn't matter if dnsmasq is gone for 2s as well
Jun 06 23:50:11 <stintel> with bird you can edit the config file, then run birdc(4|6) and run the configure command
Jun 06 23:50:14 <stintel> it reloads the config
Jun 06 23:51:12 <jow_laptop> for quagga and bird6 it probably makes sense to cat together the final config from various bits
Jun 06 23:51:23 <stintel> actually I like quagga more, but quagga can be a pita and so now I am trying bird for a change
Jun 06 23:51:44 <jow_laptop> e.g. one or more user fragments which are untouched and some dynamically generated sections, e.g. to describe dynamically received prefixesd
Jun 06 23:51:47 <jow_laptop> -d
Jun 06 23:52:27 <stintel> http://pastebin.com/SWk1TJZ1
Jun 06 23:52:33 <stintel> that's the bird6 config
Jun 06 23:52:41 <stintel> well, the part for rtadv
Jun 06 23:52:59 <jow_laptop> yes, looks workable
Jun 06 23:53:07 <jow_laptop> does it have some kind of rpc?
Jun 06 23:53:23 <jow_laptop> if I google for bird6 rpc here I only get weird stuff on google
Jun 06 23:53:38 <stintel> unix  2      [ ACC ]     STREAM     LISTENING       3630 /var/run/bird6.ctl
Jun 06 23:53:45 <stintel> thats for birdc6 to connect
Jun 06 23:54:05 <stintel> but I don't think you can actually change config values with that
Jun 06 23:55:01 <jow_laptop> yeah but you can query the state and instruct it to smoothly reload the config
Jun 06 23:55:07 <stintel> http://pastebin.com/33RsTxPQ
Jun 06 23:55:08 <jow_laptop> that seems good enough
Jun 06 23:56:03 <jow_laptop> hm. 161K
Jun 06 23:59:20 <stintel> yeah kinda big compared to radvd, but it does ospf and bgp as well
Jun 06 23:59:29 <jow_laptop> yep
Jun 07 00:00:22 <stintel> also supports multiple routing tables
Jun 07 00:00:33 <jow_laptop> well maybe we can build a -mini without ospf and bgp
Jun 07 00:00:59 <stintel> one might as well use radvd then
Jun 07 00:01:26 <stintel> but rather split it in bird6-(radv|ospf|bgp|whatelse) then ?
Jun 07 00:01:30 <stintel> if configure supports it
Jun 07 00:01:43 <jow_laptop> that would work too
Jun 07 00:01:53 <stintel> like quagga
Jun 07 00:02:25 <stintel>   --with-protocols=LIST   include specified routing protocols (default: all)
Jun 07 00:02:28 <stintel> hmmm
Jun 07 00:03:11 <stintel> it's a single binary
Jun 07 00:04:15 <stintel> wow 2am already, gonna call it a day
Jun 07 00:04:57 <jow_laptop> good night
Jun 07 00:05:24 <stintel> wonder when the ttl2usb convertors will arrive so I can mess with my rb2011l-in a bit
Jun 07 00:06:02 <stintel> thanks, you too, whenever that will be ;)
Jun 07 00:41:10 <builder_> build #23 of mpc52xx is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/mpc52xx/builds/23
Jun 07 02:00:08 <builder_> build #26 of octeon is complete: Failure [failed compile_6]  Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/26
**** ENDING LOGGING AT Thu Jun 07 02:59:58 2012