**** BEGIN LOGGING AT Wed Nov 08 03:00:02 2017 Nov 08 04:13:22 build #596 of ramips.rt3883 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.rt3883/builds/596 Nov 08 04:16:44 build #599 of ipq806x is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ipq806x/builds/599 Nov 08 04:30:58 build #581 of realview is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/realview/builds/581 Nov 08 05:49:32 build #448 of ath25 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ath25/builds/448 Nov 08 07:40:58 build #440 of adm8668 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/440 Nov 08 07:45:18 build #443 of mvebu is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mvebu/builds/443 Nov 08 08:02:52 build #434 of mpc85xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mpc85xx/builds/434 Nov 08 09:42:26 build #489 of x86.xen_domu is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/x86.xen_domu/builds/489 Nov 08 11:11:17 build #552 of ramips.mt7620 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.mt7620/builds/552 Nov 08 11:28:43 build #586 of imx6 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/imx6/builds/586 Nov 08 12:00:14 build #579 of x86.64 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/x86.64/builds/579 Nov 08 13:12:14 build #584 of oxnas is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/oxnas/builds/584 Nov 08 14:27:28 build #527 of gemini is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/gemini/builds/527 Nov 08 14:27:55 build #612 of brcm47xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/612 Nov 08 14:28:48 build #467 of cns3xxx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/cns3xxx/builds/467 Nov 08 14:39:30 build #582 of ramips is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ramips/builds/582 Nov 08 14:50:55 ping mkresin Nov 08 14:57:20 build #535 of brcm63xx.smp is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm63xx.smp/builds/535 Nov 08 15:12:40 DonkeyHotei: pong Nov 08 15:13:28 hi. you seem to be the last person to have tried to get br2684ctl to set a proper macaddr/esi in owrt/lede Nov 08 15:14:55 DonkeyHotei: should work in LEDE head Nov 08 15:15:02 DonkeyHotei: for atm ethernet bridges Nov 08 15:15:15 DonkeyHotei: was confirmed by a test on a DSLAM Nov 08 15:15:16 nope, all zeroes with a 1 in the middle Nov 08 15:15:39 at least on the dgn3500 Nov 08 15:16:17 lan is not set either, but that's a separate issue that i will address myself Nov 08 15:17:38 DonkeyHotei: no idea than. was tested on xrx200 Nov 08 15:18:13 where exactly in br2684ctl is it set on xrx200? Nov 08 15:18:15 DonkeyHotei: which revision do you use Nov 08 15:18:31 revision of what? Nov 08 15:18:52 DonkeyHotei: it isn't set via br2684ctl. the mac address is applied to the ethernet bridge by netifd Nov 08 15:18:57 DonkeyHotei: revision of LEDE Nov 08 15:19:02 head Nov 08 15:19:38 \ DE \ / Reboot (SNAPSHOT, r1009+4267-9780b2d) Nov 08 15:19:38 Nov 08 15:19:57 DonkeyHotei: well, that is r1009 + 4267 custom commits Nov 08 15:20:04 hmm Nov 08 15:20:28 DonkeyHotei: do you have a dsl0 interface instead of nas0? Nov 08 15:20:37 i rebased somewhere, but it's synced to head Nov 08 15:20:42 yes, dsl0 Nov 08 15:22:02 DonkeyHotei: if you disconnect your adsl and load the ltq_atm_ar9 module by hand. what does "ip link" says about the link Nov 08 15:23:06 DonkeyHotei: expection is that atm module is loaded, netifd applies the mac address, link changes to up Nov 08 15:23:19 DonkeyHotei: and only if link is up, the dhcp client is started Nov 08 15:24:04 DonkeyHotei: please upload your /e/c/network to pastebin.com Nov 08 15:24:36 disclosure: as of 10 or 11 months ago, i no longer have a 2684 circuit at home, but i'm just getting back to this now, and it was not working then, and it is not fixed, judging by the dsl0 interface Nov 08 15:25:12 DonkeyHotei: ... Nov 08 15:25:30 DonkeyHotei: you might want to ask me again if you are actually really testing head Nov 08 15:25:46 DonkeyHotei: instead of wasting my time Nov 08 15:25:59 i now have access to an adsl line to test on the average of once a month for 4 days at a time, but it's pppoe. that said, i still have sync at home Nov 08 15:26:21 and it syncs and creates dsl0 Nov 08 15:26:47 the dhcp client is running Nov 08 15:28:00 DonkeyHotei: than show me your /e/c/network and the output of 'ip link' Nov 08 15:28:30 is that ip link with it connected to the line, or after manually loading the module? Nov 08 15:28:49 both please Nov 08 15:28:53 ok Nov 08 15:28:56 hang on Nov 08 15:31:25 build #562 of malta is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/malta/builds/562 Nov 08 15:37:10 https://transfer.sh/pzaDC/etc-config-network Nov 08 15:38:44 DonkeyHotei: according to your /de/c/network no custom mac address should be set Nov 08 15:39:09 yes, i thought br2684ctl itself was to set it Nov 08 15:39:21 previously, it was set via esi Nov 08 15:39:36 which was never the right way Nov 08 15:40:11 yes, but it should not rely on uci, either Nov 08 15:40:22 add something like https://gist.github.com/mkresin/157d84185db4a132924c93912997038e to your /e/c/network Nov 08 15:40:48 DonkeyHotei: we had different opinions about it in the past and I'm quite sure it will never change Nov 08 15:41:15 DonkeyHotei: It don't like the custom kernel/driver patches and prefer to have a clean driver code and fix stuff in userspace Nov 08 15:42:02 macaddr is always set in userspace for atm anyway Nov 08 15:42:12 for ptm, it is hardcoded Nov 08 15:42:56 for ptm it should be rather random. but need to verify it and doesn't really matter Nov 08 15:43:30 i think a compromise could be that lack of a specified macaddr in uci would be a fatal error Nov 08 15:43:42 agreed Nov 08 15:44:09 as far as I remember you can get your base mac address from u-boot Nov 08 15:45:07 the esi method was programmatically deriving it from lan, which in turn, came from uboot Nov 08 15:45:20 have a look at target/linux/lantiq/base-files/etc/board.d/02_network to set the mac from u-boot for $interfaces Nov 08 15:47:05 that iirc only initially sets the macaddr, but not when the interface went down and then back up Nov 08 15:47:53 and it did not handle the case of multiple atm links Nov 08 15:47:57 DonkeyHotei: the interface should keep the mac address Nov 08 15:48:09 nope multiple atm links are not supported Nov 08 15:50:48 i expect changing uci to error on no macaddr in atm/ptm would be a tough sell Nov 08 15:53:56 the other problem with doing it via /etc/board.d/02_network is that if the atm link is deleted and recreated in luci, or replaced with ptm, the default macaddr does not carry over Nov 08 15:54:34 due to the dsl0, there is no need to apply any changes because of atm/ptm switch Nov 08 15:54:45 it will just work Nov 08 15:54:51 ok Nov 08 16:24:58 mkresin: btw, sysupgrade used to work, but now the bootloader complains that the checksum is wrong Nov 08 16:26:03 (and fails) Nov 08 16:29:24 DonkeyHotei: strange. the only issue i'm aware of should be fixed with https://git.lede-project.org/06741411e020bb28160ddd42149d6a489800dbc7 Nov 08 16:34:23 the "factory" image works from the bootloader's web-ui, but the "sysupgrade" image from sysupgrade has a bad sum Nov 08 16:43:19 build #158 of archs38 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/archs38/builds/158 Nov 08 16:46:48 mkresin: ok, i added the relevant lines to 02_network and booted initramfs, but /e/c/network still has no macaddr Nov 08 16:53:21 DonkeyHotei: push your tree + changes somewhere Nov 08 16:54:17 there are no changes atm other than 02_network Nov 08 16:54:28 show me your changes Nov 08 16:55:17 and i have no idea why the sysupgrade image causes a "checksum is wrong" error Nov 08 16:56:08 https://transfer.sh/Hv50c/dgn3500 Nov 08 16:56:36 DonkeyHotei: please use pastebin.com in future Nov 08 16:58:27 the partition is named uboot-env and not uboot_env Nov 08 16:58:43 DonkeyHotei: ^^ Nov 08 17:07:56 build #430 of sunxi is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/sunxi/builds/430 Nov 08 17:14:25 ah. Nov 08 17:50:26 mkresin: strange thing is that even if i specify xfer_mode 'ptm', it still loads ltq_atm_ar9 when the line is connected unless i manually load ltq_ptm_ar9 first Nov 08 17:51:34 i know there is such a thing as ptm over adsl (as opposed to vdsl) Nov 08 17:56:48 teh atm/modules should be autoloaded based on the xfer mode negotiated by dsl_control Nov 08 17:56:55 the atm/ptm... Nov 08 17:57:54 could be that I missed to add the piece of code, which overrides the dsl_control negotiation Nov 08 17:59:38 DonkeyHotei: nope, the xfer_mode was never supported on !xrx200 Nov 08 17:59:57 build #534 of netlogic is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/netlogic/builds/534 Nov 08 17:59:57 ah, ok Nov 08 18:00:24 DonkeyHotei: only package/network/config/ltq-vdsl-app/files/dsl_control reads the uci var in stable Nov 08 18:02:13 DonkeyHotei: but should be easy to add. something similar to https://git.lede-project.org/2da6c85c8005c9b1ce8a3e626f644ee468305800 is required Nov 08 18:02:43 DonkeyHotei: but ti be honest, I've no idea if it is supported by the danube/ar9 dsl_daemon Nov 08 18:03:14 idk either, but the ptm driver for ar9 is there Nov 08 18:03:35 however, it appears not to be stable Nov 08 18:03:41 I would be surprised if it was ever used by someone Nov 08 18:04:02 adsl over ptm must be a really rare config Nov 08 18:09:36 iirc it's used by vdsl isp's when loop length/distance is too great for vdsl Nov 08 18:20:54 build #493 of mcs814x is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mcs814x/builds/493 Nov 08 19:26:47 build #518 of x86.kvm_guest is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/x86.kvm_guest/builds/518 Nov 08 19:41:06 mkresin: https://github.com/lede-project/source/pull/1502 Nov 08 20:01:35 DonkeyHotei: still around? Nov 08 20:01:41 yep Nov 08 20:02:15 DonkeyHotei: swconfig isn't required in device packages. it is already in target/linux/lantiq/xway/target.mk Nov 08 20:02:29 also, the commit msg is wrapped at 79 instead of 75 chars, just noticed Nov 08 20:03:09 strange that it doesn't get picked up from xway, though Nov 08 20:03:10 DonkeyHotei: commit message isn't really good Nov 08 20:03:24 DonkeyHotei: commit message should explain why something was done Nov 08 20:03:54 DonkeyHotei: maybe an old config. check ./script/diffconfig.sh Nov 08 20:04:27 i tried deleting .config and doing menuconfig, both before and after this change Nov 08 20:04:53 DonkeyHotei: we usually include switch driver in the kernel config Nov 08 20:05:18 DonkeyHotei: but no strong opinion about it Nov 08 20:05:21 there is no profile-specific kernel config here Nov 08 20:05:56 DonkeyHotei: you mean target/linux/lantiq/xway/target.mk is missing in your tree? Nov 08 20:06:53 it's there, but menuconfig did not have swconfig selected by default Nov 08 20:10:24 DonkeyHotei: works for me: https://gist.github.com/mkresin/cb00ccbb9c769b4a8bfd978635cabc9d Nov 08 20:11:46 ok, i'll remove that part and update the pr, but the switch driver really was missing Nov 08 20:12:24 build #493 of xburst is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/493 Nov 08 20:21:40 mkresin: better now? Nov 08 20:21:47 DonkeyHotei: I'm not sure about the switch driver Nov 08 20:22:17 DonkeyHotei: https://git.lede-project.org/?p=source.git;a=blob;f=target/linux/lantiq/config-4.9;h=a98509e5d660a76910ed524c7699864c4911889f;hb=HEAD#l170 Nov 08 20:22:22 DonkeyHotei: line 170 Nov 08 20:22:48 hmm Nov 08 20:23:18 DonkeyHotei: I suspect something is wrong with your tree Nov 08 20:23:44 hmm, i was only looking at modules Nov 08 20:23:46 DonkeyHotei: would you mind to test a precompiled image from https://downloads.lede-project.org/snapshots/targets/lantiq/xway/ Nov 08 20:24:00 ok, i'll remove that part too Nov 08 20:24:15 DonkeyHotei: I'm still concerned about the sysupgrade checksum failure thingy Nov 08 20:24:31 tried sysupgrade again and this time it worked Nov 08 20:24:38 could've been a stale image Nov 08 20:25:19 and my kernel tree does have =y for the switch Nov 08 20:25:49 DonkeyHotei: Am I right that we can finally get rid of the ethaddr related patches? Nov 08 20:26:18 i think part of that is already upstreamed, no? Nov 08 20:26:25 nope Nov 08 20:26:58 and I doubt that it will be accepted uptream Nov 08 20:27:46 if you just want to remove the setup= thing, i guess that makes some sense Nov 08 20:28:20 DonkeyHotei: setup thing and ltq_get_eth_mac Nov 08 20:29:02 wait, is no device using a macaddr from the dtb anymore? Nov 08 20:29:57 DonkeyHotei: that would be of_mac_address() Nov 08 20:30:29 DonkeyHotei: all the code in question is for parsing a mac address passed via kernel bootargs Nov 08 20:30:57 ah. then i suppose it can go. however, then there is the stale code upstream that ltq_get_eth_mac removes, one line iirc Nov 08 20:31:52 DonkeyHotei: mac address from platform data? https://git.lede-project.org/?p=source.git;a=blob;f=target/linux/lantiq/patches-4.9/0035-owrt-lantiq-wifi-and-ethernet-eeprom-handling.patch;h=a5ecd94c4a9c17b7c4374365b9fc64619800bd6f;hb=HEAD#l211 Nov 08 20:34:16 platform data is no longer used, correct? Nov 08 20:34:25 there is quite some code in that patch that can be easly replaced Nov 08 20:34:31 yes Nov 08 20:34:48 nope, it isn't. It is more or less device tree XOR pdata Nov 08 20:37:09 mhh. not sure if I it is a good idea to touch the ath5k stuff. exploded aweful the last time I did Nov 08 20:39:36 ath5k stuff uses bootargs too Nov 08 20:41:32 i might send a bootargs-append patch upstream regardless, as it could have some use there outside of owrt/lede Nov 08 20:51:02 mkresin: anyway, this is now down to where it doesn't even need a pr, but the pr is updated nonetheless Nov 08 20:52:24 build #544 of lantiq is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/544 Nov 08 21:32:41 build #513 of ar71xx.nand is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.nand/builds/513 Nov 08 22:54:04 build #491 of mxs is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/mxs/builds/491 Nov 08 23:07:54 build #423 of omap is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/omap/builds/423 Nov 08 23:33:48 build #461 of arm64 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/arm64/builds/461 Nov 09 00:18:03 build #434 of adm5120 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/adm5120/builds/434 Nov 09 02:21:28 build #559 of at91 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/at91/builds/559 Nov 09 02:38:23 build #614 of brcm47xx.legacy is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.legacy/builds/614 **** ENDING LOGGING AT Thu Nov 09 03:00:01 2017