**** BEGIN LOGGING AT Tue Oct 30 02:59:59 2018 Oct 30 04:47:34 uberushaximus: after next release Oct 30 05:19:40 build #1097 of lantiq/xway is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/lantiq%2Fxway/builds/1097 **** BEGIN LOGGING AT Tue Oct 30 06:23:09 2018 **** BEGIN LOGGING AT Tue Oct 30 06:50:27 2018 **** BEGIN LOGGING AT Tue Oct 30 08:34:28 2018 Oct 30 09:52:13 * ldir wanders in Oct 30 09:57:40 ldir: Hi Oct 30 09:57:50 ldir: any idea what is wrong here? https://forum.openwrt.org/t/wireguard-fails-on-latest-turris-omnia-18-06rc2-openwrt/17444/5 Oct 30 10:07:37 cmon watchdog, kick in plz :) "sh: page allocation stalls for 388310ms" Oct 30 10:08:25 ldir ping Oct 30 10:09:08 I have tested this change and is working grate. Oct 30 10:09:09 https://github.com/openwrt/openwrt/pull/1507 Oct 30 10:09:42 anomeome did a better job of it than me! :-) Oct 30 10:13:33 build #456 of octeontx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/octeontx%2Fgeneric/builds/456 Oct 30 10:23:28 Tapper, thanks for the heads-up Oct 30 10:23:36 building as we speak Oct 30 10:24:01 what do you exactly mean by "working great"? Oct 30 10:24:33 better speed? better coverage? Oct 30 13:38:04 nitroshift Hi It seems to be holding on to devices at a higher rate for longer Oct 30 13:42:46 I must of misted a package from my build tho I am getting a lot of this in my logs Oct 30 13:42:49 Tue Oct 30 13:39:44 2018 daemon.warn odhcpd[2018]: DHCPV6 REQUEST IA_NA from 00010001236b03e2a8967598d84d on br-lan: not on-link Oct 30 14:25:08 build #1088 of brcm2708/bcm2710 is complete: Failure [failed sourceupload] Build details are at http://phase1.builds.lede-project.org/builders/brcm2708%2Fbcm2710/builds/1088 blamelist: Kevin Darbyshire-Bryant Oct 30 14:38:21 xback: Hi Koen, sorry for delay - at the conference so a little distracted :-) Oct 30 14:41:35 xback: I have no instant answer to the forum question and no matching hardware to test. I'm confused by "Turris Omnia OpenWRT" is this a fork? Oct 30 14:44:40 ugh Oct 30 14:44:49 ldir: yes Oct 30 14:46:18 AFAIK the protocol isn't stable stable which means mismatched wireguard builds might not be interoperable. Oct 30 14:52:04 ldir: the protocol is stable now Oct 30 14:52:54 the last time we had a change was Thu Apr 27 11:10:50 2017 +0200 Oct 30 14:52:58 with dc34c6f2e6f038f2943fff1057a8dd307d9193cd Oct 30 14:53:12 zx2c4: Yeah, I don't know when the Turris joined the wireguard train..... it's a thought...my only obvious thought at this time :-) Oct 30 14:53:16 which means there's only incompatibility pre-0.0.20170517 and post-0.0.20170517 Oct 30 14:53:19 oh yea Oct 30 14:53:22 i think turris' first build Oct 30 14:53:27 was based on the first ever snapshot of wireguard Oct 30 14:54:20 I shall try to remember the split at 20170517 - two days after my birthday :-) Oct 30 15:20:37 build #611 of layerscape/armv8_64b is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/layerscape%2Farmv8_64b/builds/611 Oct 30 15:23:36 Hello Oct 30 15:23:58 Is accepted special names for factory images? Oct 30 15:25:06 The firmware upgrade OEM web interface only accept files with special name, but openwrt buildsystem generates the factory image with prefix Oct 30 15:47:04 hi, luci keeps crashing (no browser output). Is there a way to see any error message on the router itself? Oct 30 15:47:50 vk496: just rename the file before sending it? Oct 30 15:57:29 vk496: they can't all be pre-renamed, because there'd be conflicts. Oct 30 16:10:43 Is better to indicate it in the wiki (rename before upload) instead of creating directly the ready-to-flash img? Oct 30 16:11:14 the wiki for the table of hardware for the device _should_ mention the name required if it needs one, yes. Oct 30 16:13:47 great. thnks karlp Oct 30 16:43:45 mwarning: logread Oct 30 16:56:37 ldir: I actually have some fix for the dnsmasq duplicate lease thing in the pipeline Oct 30 16:57:04 ldir: fur luci that is, not sure if its worthwhile to handle this in the init script too Oct 30 16:57:07 *for Oct 30 17:04:05 build #1068 of cns3xxx/generic is complete: Failure [failed sourceupload] Build details are at http://phase1.builds.lede-project.org/builders/cns3xxx%2Fgeneric/builds/1068 blamelist: Kevin Darbyshire-Bryant Oct 30 17:12:27 jow: I figured out that there was no error in logread because the generated html code was invalid. The browser was telling me that there was a server error. Oct 30 17:54:46 Is there some clean way to extract the extension of file in Makefile define? Oct 30 17:55:03 I want to modify the extension of $@ Oct 30 17:55:58 Doing this not work as expected: mv "$@" "$(dir $@)/$(echo $(notdir $@)| cut -d. -f1).bin" Oct 30 17:56:23 Not sure where to escape to execute inside pure shell Oct 30 18:16:21 Is there a way to keep the factory install from over-riding the /etc/config/wireless file that I am adding to the files dir? Oct 30 18:49:25 jow: excellent. Personally I think the dnsmasq.init script is hideous enough without doing checks there as well. and if you're editing uci config files directly then you can deal with your own mess when dnsmasq.init doesn't produce a working config. Luci shouldn't produce invalid uci config. IMHO Oct 30 20:12:11 build #1044 of lantiq/xrx200 is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/lantiq%2Fxrx200/builds/1044 blamelist: Kevin Darbyshire-Bryant Oct 30 22:06:52 so, how to make /etc/resolv.conf actually be configured? Oct 30 22:06:55 root@LF-R7800-6ab8:/# cat /etc/resolv.conf Oct 30 22:06:56 # This file is generated by the DHCPD service Oct 30 22:06:56 nameserver 127.0.0.1 Oct 30 22:06:56 root@LF-R7800-6ab8:/# uci get network.wan.dns Oct 30 22:06:56 8.8.8.8,9.9.9.9 Oct 30 22:07:27 I have disabled dhcp Oct 30 22:08:44 greearb_: you can set dns servers in /etc/init.d/dnsmasq and /etc/config/network Oct 30 22:09:21 meh, /etc/init.d/dhcp, not /etc/init.d/dnsmasq Oct 30 22:09:41 greearb_: you should find your configured nameservers lurking in /tmp/resolv.conf.auto which is used by dnsmasq as its resolver address source Oct 30 22:14:51 I don't think I want dnsmasq running at all, actually Oct 30 22:18:22 I disabled dnsmasq, rebooted. /tmp/resolv.conf.auto does show my 8.8.8.8 as expected, but /etc/resolv.conf is still set at 127.0.0.1 Oct 30 22:24:32 greearb_: usually /etc/resolv.conf is just a symlink pointing to /tmp/resolv.conf Oct 30 22:24:47 greearb_: it probably got replaced by an actual file Oct 30 22:25:00 root@LF-R7800-6ab8:/etc/init.d# ls -l /etc|grep reso Oct 30 22:25:00 lrwxrwxrwx 1 root root 16 Oct 26 16:59 resolv.conf -> /tmp/resolv.conf Oct 30 22:25:01 root@LF-R7800-6ab8:/etc/init.d# cat /tmp/resolv.conf Oct 30 22:25:01 # This file is generated by the DHCPD service Oct 30 22:25:01 nameserver 127.0.0.1 Oct 30 22:25:09 also the comment ("# This file is generated by the DHCPD service") indicates that Oct 30 22:25:33 ok, so are you using dhcpd ? Oct 30 22:26:01 I did previously, and it is installed on the system Oct 30 22:26:11 currently things are configured for static IP Oct 30 22:26:55 it looks as if dhcpd or its init script are overwriting /tmp/resolv.conf Oct 30 22:32:00 build #1028 of lantiq/falcon is complete: Failure [failed sourceupload] Build details are at http://phase1.builds.lede-project.org/builders/lantiq%2Ffalcon/builds/1028 blamelist: Matthias Schiffer , Kevin Darbyshire-Bryant Oct 30 22:37:06 jow, yes, that was it. I installed isc-dhcpd to get dhclient and to manually start dhcpd from my own application. I did not realize it was auto-starting and mucking with things. I disabled both dnsmasq and dhcpd and now /etc/resolv.conf looks as hoped Oct 30 23:24:46 so, I ran 'poweroff', but the netgear came back up! Oct 30 23:25:01 what is proper way to tell it to power down and stay down gracefully? Oct 30 23:39:14 halt? Oct 30 23:40:37 poweroff worked for me Oct 30 23:45:38 russell--, what platform? I found something on the web that indicates lots of systems may just not poweroff and stay down due to watchdogs Oct 30 23:45:43 no idea how true that is Oct 30 23:46:12 probably just means most of them don't hve support for doign the right thing with turning off watchdogs? Oct 30 23:46:23 could easily be that Oct 30 23:46:35 greearb_: mt76x8 Oct 30 23:53:35 poweroff was working on ipq8065 in the past, but it doesn't seem to work anymore (rebooting instead) - I haven't really debugged why so far Oct 30 23:55:10 not critical for me now, but in case someone has a fix worth trying, I'm happy to test. Oct 30 23:57:04 In general, the r7800 has been enjoyable...nice it can have a serial port w/out soldering, ran a 10 virtual-station test over the weekend w/out trouble, and today I got it to run pretty much my entire kernel and stack, with 64 virtual stations on one of the radios. Oct 30 23:57:23 So, maybe nice new affordable platform if someone wants 20+ devices for network testing **** ENDING LOGGING AT Wed Oct 31 02:59:58 2018