**** BEGIN LOGGING AT Wed Jun 19 02:59:58 2013 Jun 19 08:13:24 build #224 of gemini is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/224 Jun 19 09:54:11 cyrusff: so, jow sent me here with my question. I use 6relayd on openwrt to delegate a prefix to my laptop. my laptop assigns that to a virtual interface, so I have VM <--> laptop <--> openwrt Jun 19 09:54:37 cyrusff: while the VM knows how to route to openwrt, openwrt can't reach the VM Jun 19 09:55:12 cyrusff: so, the other day, I was able to use a neighbour discovery proxy on my laptop, and then openwrt could reach the VM, too Jun 19 09:55:29 cyrusff: so I thought, why not use 6relayd on my laptop Jun 19 09:56:13 cyrusff: the problem is, I can't get this to work with 6relayd: both server as a RD server on the laptop's downstream interface, and to relay the NDP requests from upstream to downstream Jun 19 09:56:24 cyrusff: question is, is it even intended to do that, and if so, how do I do it? Jun 19 10:24:04 Brain0: i wonder why your first attempt with the PD didnt work. Normally openwrt knows how to reach the vm in that case if you have configured the laptop for routing corrrctly Jun 19 10:24:41 Ah nk i know its missing a routing entry. I will fix that later. Jun 19 10:25:23 If you use proxying then you shoild proxy all 3. Ndp dhcpv6 and rd Jun 19 10:29:16 cyrusff: so, 6relayd should add a route and I shouldn't need to mess with NDP on the laptop? Jun 19 10:29:57 Well it should if it does pd but it doesnt atm. Will fix that later today / tomorrow Jun 19 10:31:08 good to know. another problem with 6relayd: whenever dhclient (on my laptop) has a lease recorded, then it requests that specific lease for pd. but then 6relayd always answers that there's no prefix available for delegation. if I delete dhclient's lease file, it works once Jun 19 10:32:25 Strange Jun 19 10:32:30 Will have a look Jun 19 10:33:15 cyrusff: thanks. I am afk for a while, but I'll stay online in this channel for the rest of the day. if you have anything to test for me or more questions on these matters, just ping me and I'll reply eventually Jun 19 10:34:06 Thx Jun 19 12:02:37 hi Jun 19 13:10:05 cyrus r36969 trunk/package/network/ipv6/6relayd/Makefile * 6relayd: correctly set downstream routes for DHCPv6-PD Jun 19 13:10:11 brain0: ^ Jun 19 13:11:55 couldn't reproduce your NOPREFIXAVAIL problem yet though Jun 19 13:12:56 cyrusff: I'll give you detailed output from dhclient later, but gtg again. thanks Jun 19 13:13:03 sure take your time Jun 19 13:13:33 brain0: if you have any pcaps / logs or so, you can send them to cyrus@openwrt.org if you like Jun 19 13:14:53 okay, I'll capture some later, see you Jun 19 13:30:36 cyrus r36970 branches/attitude_adjustment/package/odhcp6c/ files/dhcpv6.sh Makefile * AA: odhcp6c: backport r36959 Jun 19 13:30:40 cyrus r36971 branches/attitude_adjustment/package/6relayd/Makefile * AA: 6relayd: backport r36969 Jun 19 14:49:02 build #267 of rb532 is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/267 Jun 19 14:55:36 build #267 of ppc44x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/ppc44x/builds/267 Jun 19 16:54:39 KanjiMonster: found this patch http://patchwork.linux-mips.org/patch/5490/ Sounds that some bcm6328 could have the second core enabled in hardware, i'm right? Jun 19 17:33:10 <[florian]> Pteridium: some of them yes Jun 19 17:33:21 <[florian]> Pteridium: I am not sure if there are any devices out there where the second core is enabled though Jun 19 17:35:07 well, i have 6 routers with bcm6328 to test. I'll tell you the results. Jun 19 17:35:11 thanks Jun 19 18:07:11 Pteridium: AFAIK it's actually some bcm6329s that have it (the dsl-less variant of bcm6328) Jun 19 18:07:39 Ah Jun 19 18:07:39 Pteridium: and that patch is mine, btw ;p Jun 19 18:07:52 yes Jun 19 18:08:47 *melt* Jun 19 18:09:19 so the bcm6329 is detected as a 6328 Jun 19 18:10:04 :,-( Jun 19 18:20:14 Pteridium: no, the bcm6329 is likely detected as 6329 and deemed "unsupported" by the code ;) Jun 19 18:40:53 KanjiMonster: excuse me, i was helping my father to repair his car Jun 19 18:41:21 no problem, I'll be here the whole evening Jun 19 18:41:51 then, there isn't bcm6328s with the second core activated Jun 19 18:47:27 tomorrow i'll tell you the the tests results, thanks for the info and the patches Jun 19 18:52:39 Does somebody know where can be found the Fusiv Vx160 drivers? Jun 19 19:06:08 Pteridium: they are as closed as broadcom Jun 19 19:10:12 Pteridium: they are as closed as broadcom Jun 19 19:13:35 KanjiMonster: Broadcom share the important drivers, Ikanos doesn't. The Livebox 2 is a dead end. Jun 19 19:14:01 Asked because you know a lot more of this stuff Jun 19 19:14:45 i've searched the entire internet without success Jun 19 19:15:33 some avm fritzboxen (well, actually one) is fusiv based, so maaaybe you can find something in their gpl tarballs Jun 19 19:19:31 yes, but they has fusiv 180 and there are no drivers for the vx160. At least in the tarballs there are code of the puma5 if somebody is crazy enough to port it. Jun 19 19:26:01 vx160 doesn't even seem to be mips Jun 19 19:27:08 ah, lexra; mips without the patented bits Jun 19 19:28:20 yes, like some realtek SoCs Jun 19 19:53:26 build #255 of sibyte is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/sibyte/builds/255 Jun 19 19:58:38 hi Jun 19 19:58:58 which target system do i select for a Ubiquiti UniFi AP? Jun 19 19:59:01 Atheros AR7xxx/AR9xxx Jun 19 19:59:20 http://wiki.openwrt.org/toh/ubiquiti/unifi Jun 19 20:00:49 build #271 of ar7 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ar7/builds/271 Jun 19 20:02:23 i neeed glasses ;) Jun 19 20:19:23 build #296 of orion is complete: Failure [failed compile_8] Build details are at http://buildbot.openwrt.org:8010/builders/orion/builds/296 Jun 19 20:20:18 tripolar, it's an ar7xxx Jun 19 20:20:29 I believe there is a profile there already Jun 19 20:38:07 groz yeah i allready found it, therefore the " i neeed glasses ;)" ;) Jun 19 20:38:16 thanks Jun 19 20:41:06 ok Jun 19 20:41:19 so I'm working on a gadget with nand flash today Jun 19 20:41:38 fixing up a few things with regards to sysupgrade so it can write the nand correctly Jun 19 20:41:50 is there a clean way to add some busybox items to a profile for a gadget Jun 19 20:42:01 nandwrite and flash_eraseall in particular ? Jun 19 20:42:30 I'd like to get this all bundled up as a clean addition to trunk, but, I don't see a way to add busybox items inside a device profile Jun 19 20:53:46 groz just select them manually? Jun 19 20:54:00 yah, I can do that, but, what I was hoping to achieve Jun 19 20:54:03 once it's all merged Jun 19 20:54:13 select the correct profile,a nd that will bring them in Jun 19 20:54:29 so one doesn't have to manually configure busybox before building for this one Jun 19 20:54:48 that's the main stumbling block I have right now to a fully clean new profile Jun 19 20:54:49 you could craeate a busybox submenu entry that selects what you want as dependency Jun 19 20:55:18 hmm, I need to look into that Jun 19 20:55:20 or even a new package that has the busybox settings as dependency Jun 19 20:55:41 I'd like to end up with a way, that once I'm done here, and it gets merged in Jun 19 20:55:44 it's 'just magic' Jun 19 20:56:01 i for example us an aoverlay package thats configures all programs to the values i want them to use Jun 19 20:56:18 no net to manually tweak anything Jun 19 20:56:24 net=need Jun 19 20:57:39 so you can for example run the router without setting a password on first boot and you can only log in with an ssh key Jun 19 20:59:37 I have stuff like that in my own trees here, but, what I'm trying to figure out Jun 19 20:59:51 is if I can do this cleanly, so I can submit patches into trunk Jun 19 21:00:01 that make this 'just work' when you select the device Jun 19 21:00:10 rather than have to manually add a few things in busybox first Jun 19 21:00:19 what wht is the use case? i mean your perfect config fits you but maybe not veryone Jun 19 21:00:21 because if you build it without adding those things Jun 19 21:00:33 you will brick it on a sysupgrade Jun 19 21:00:40 create a package ans add you busybox options as dependency Jun 19 21:00:44 use case is 'finish' support for rb-751g in trunk Jun 19 21:00:55 so that it works correctly, builds correctly, etc Jun 19 21:01:06 but when you select the microtik nand target Jun 19 21:01:14 it adds the necessary nand utilities Jun 19 21:01:33 but why not add your mods to an overlay package and flash the package normal without sysupgrade? Jun 19 21:02:05 what I'm trying to accmplish, is end up with a clean build such that Jun 19 21:02:15 somebody can do an svn checkout from trunk Jun 19 21:02:19 build it, flash it onto the unit Jun 19 21:02:23 and sysupgrade works Jun 19 21:02:29 without bricking these things Jun 19 21:02:52 but you can't blindly use mtd write with the nand flash Jun 19 21:02:57 or mtd erase Jun 19 21:02:59 but why should a busybox "extention" break anything? Jun 19 21:03:26 well, if you check out from scratch with it the way it is Jun 19 21:03:26 build #230 of iop32x is complete: Failure [failed shell_12] Build details are at http://buildbot.openwrt.org:8010/builders/iop32x/builds/230 Jun 19 21:03:36 and don't manually add these utilities to busybox Jun 19 21:03:47 then you can't manipulate the flash correctly Jun 19 21:04:09 so I'm just trying to figure out a way to make it clean and magic Jun 19 21:05:06 what do you actually want to do? what would break when you add busybox modules? Jun 19 21:05:08 no serial console available on thise, so, if you break them, convoluted netboot setup to try recover Jun 19 21:05:20 it's not about _me_ adding them, that's easy Jun 19 21:05:30 it's about ending up so that if somebody else checks out trunk Jun 19 21:05:46 adding the necessary bits to busybox is magic, if they have selected the nand target Jun 19 21:13:06 groz: nand usually means plenty of space, why not just use the mtdutils versions? Jun 19 21:13:27 that is another options yes Jun 19 21:13:47 at least busybox' nandread (or *dump, whatever the name was) is quite broken Jun 19 21:22:13 never really paid attention to mtd utils packages before Jun 19 21:22:26 it's neatly broken down already, that's really the best / easiest way Jun 19 21:22:32 just make some dependancies on those Jun 19 21:31:04 hauke r36972 trunk/package/mac80211/patches/030-rt2x00_options.patch * mac80211: do not build RT2X00_LIB_XXX when XXX is not available in the kernel. Jun 19 21:35:00 hauke r36973 trunk/package/mac80211/patches/060-no_local_ssb_bcma.patch * mac80211: do not use CPTCFG_SSB and CPTCFG_BCMA Jun 19 21:37:51 hauke r36974 trunk/package/mac80211/Makefile * mac80211: copy the backport-include files to staging_dir Jun 19 21:38:21 hauke r36975 trunk/package/mac80211/Makefile * mac80211: libertas-sd depends on mmc Jun 19 21:39:25 hauke r36976 trunk/package/acx-mac80211/Makefile * acx-mac80211: make it compile with the new backports mac80211 Jun 19 21:44:01 KanjiMonster, [florian]: the bcm6328 can be easily underclocked editing clk.c now i go to test overclocking Jun 19 21:45:09 sorry, cpu.c Jun 19 21:51:22 Pteridium: might not necessarily run stable, as there is a bit of dram and bus clock syncing involved at boot/init time Jun 19 21:53:20 i thought that unstability could appear, so tests are strongly needed Jun 19 21:56:08 bcm6358 and never support dynamic clocking down to 1/2, 1/4 and 1/8 (but requires a bit of added code because you can't use the cpu internal timer anymore) Jun 19 21:58:56 cyrusff: just tested the 6relayd pd changes, it now seems to add routes properly Jun 19 21:59:15 cyrusff: do they expire at some point (when the lease expires, or something)? Jun 19 21:59:39 yes, in a source code found a file related power saving, but is really needed? Jun 19 22:00:48 time to test ;) **** ENDING LOGGING AT Thu Jun 20 02:59:59 2013