**** BEGIN LOGGING AT Wed Nov 14 03:00:00 2018 Nov 14 10:53:52 jow: purely fyi. it seems one of the docker buildslaves (fsl-02) ran out of space during build: :Cannot write: No space left on device" Nov 14 10:58:56 does a git fetch --all and discovers the xback-kernel-bump-bot is still broken :-) What no 4.14.81??? :-) Nov 14 10:59:12 xback: I increased the tmpfs sizes Nov 14 10:59:17 jow: thank you Nov 14 10:59:23 ldir: nearly finished :) Nov 14 10:59:43 lol!!! It's been 15 hours! just what is going on ;-) Nov 14 11:00:00 my kid kept me up all night ;) started late today Nov 14 11:00:08 rofl Nov 14 11:00:36 you know I'm joking right - I have no idea how you do it so quickly...well actually I do... the hardware helps Nov 14 11:00:55 ldir: no worries :) I know Nov 14 11:01:24 this bump takes some effort .. lots of patches didnt apply Nov 14 11:01:43 and it's pretty tedious to alter some patches regarding voltage control Nov 14 11:04:27 ldir: if my buildserver ever gets obsolete here, I'll request my boss to send it over to you :) Nov 14 11:05:14 so make sure to reserve a day somewhere in 2029 :) Nov 14 11:37:59 anyone here experienced with the qmi/wwan proto handler? Nov 14 11:41:01 you mean qmi.sh ? Nov 14 11:41:07 yes Nov 14 11:41:25 I've done commits on it Nov 14 11:41:32 it seems to spawn virtual interfaces like wwan_4 or wwan_6 Nov 14 11:41:53 problem is that those do not seem to inherit the firewall zone membership of the parent wwan Nov 14 11:43:25 yes. it's been reported a few times iirc Nov 14 11:43:51 which could be solved by passing `{ "zone": "$(fw3 -q network wwan)" }` in the interface data Nov 14 11:43:58 but I lack a suitable testbed Nov 14 11:44:24 I'm able to provide the testing for you Nov 14 11:44:31 got a device up and running with QMI in my desk here Nov 14 11:44:40 on my desk :p Nov 14 11:51:11 xback: http://luci.subsignal.org/~jow/0001-uqmi-inherit-firewall-zone-membership-to-virtual-sub.patch Nov 14 11:51:29 jow: noted. I'll provide feedback later today Nov 14 11:51:54 thanks! Nov 14 12:22:41 hmm, interesting Nov 14 12:22:59 upon attempting to ssh to a vanilla 18.06.1 box: Error relocating /usr/bin/ssh: explicit_bzero: symbol not found Nov 14 12:23:15 xback: didn't you post something similar a few days ago? Nov 14 12:23:47 hm Nov 14 12:23:54 jow: i use wwan Nov 14 12:23:59 nvm, brainfart Nov 14 12:24:16 that issue I remembered was about ld.so Nov 14 12:24:24 i thought it worked ok.. or is it working by accident? Nov 14 12:25:06 jwh: good question; the way it works atm I'd assume that it simply uses default firewall policies Nov 14 12:25:38 it could also be that wwan_4 and wwan_6 share the same actual l3 netdev with wwan in which case its likely fine Nov 14 12:26:05 my patch is solving the problem that the firewall is not aware of wwan_4 and wwan_6, preventing it from doing things like nat loopback or subnet matching Nov 14 12:26:34 netdev based rules will likely work as expected as long as "wwan" has a proper l3_device in "ifstatus wwan" Nov 14 12:29:10 masq isnt on by default though right? Nov 14 12:29:21 ahhh Nov 14 12:29:59 jwh: can you paste me some (redacted) outputs of "ifstatus wwan"; "ifstatus wwan_4" and "ifstatus wwan_6" ? Nov 14 12:31:10 uh yeah give me a few, just trying to get the sim recharged (tmo seem to suck everywhere) Nov 14 12:32:17 I am grateful for your efforts :) Nov 14 12:33:14 i dont think anyone is manning the english bit of the call center :( Nov 14 12:47:31 Hi, are there any plans for a release in the next time? Nov 14 12:51:29 Justanick: 19.01 is due in january / february Nov 14 12:51:49 Justanick: an 18.06.2 probably in the same time frame Nov 14 12:52:02 maybe earlier Nov 14 12:54:03 jow: That would be nice. Thanks for the information. Nov 14 12:54:57 ah cool, their card payments are broken i need to go to a shop Nov 14 12:58:14 BTW, what's the current status of ar71xx/ath79? If I would like to add support for the new board to the master, should I add the support to the both targets? Nov 14 12:59:34 I would say ideally both, until we deprecate ar71xx Nov 14 12:59:59 I would have said ath79 only :) Nov 14 13:00:26 good, so it's 1:1 in 12' Nov 14 13:00:47 12'? Nov 14 13:01:26 I wanted to express 12 minute as in a (place your favorite sport game) match Nov 14 13:01:46 but it was only 2' ;) Nov 14 13:01:52 that's what threw me :) Nov 14 13:02:24 was 19.01 going to have ar71xx still and ath79 unmasked? if so, then yeah, both I guess. Nov 14 13:02:28 as a first step we could start building ath79 as binaries Nov 14 13:02:34 in snapshots Nov 14 13:02:47 blogic: is there anything speakign against it? Nov 14 13:02:48 we're probably at that point now aren't we? Nov 14 13:06:19 jow: cant paste easiky but _4 has the correct netde so your guess is correct :D Nov 14 13:06:24 netdev* Nov 14 13:07:57 why the _(4|6) jooi? Nov 14 13:10:20 jow: correct. I posted that too Nov 14 13:10:21 jwh: I guess to be able to spwan dedicated udhcpc / odhcp6c instances Nov 14 13:10:45 it's currently occuring on a device in the field too, which is inaccessible due to that Nov 14 13:10:53 hoping for a spontaneous reboot .. Nov 14 13:11:13 jow: ahhh makes sense Nov 14 13:11:37 jwh: ok so the firewall will mostly work by accident but fail in some edge cases Nov 14 13:11:50 yeah Nov 14 13:11:56 basically when trying to figure out the wan ip Nov 14 13:13:21 hmmm Nov 14 13:14:12 ldir: kernel bumps in my staging for master and 18.06, but compile-tests just started Nov 14 13:15:10 jow: taken live a few seconds ago: Nov 14 13:15:13 [ Casino Blankenberge | node-200 ] ssh root@10.254.253.31 Nov 14 13:15:14 root@10.254.253.31's password: Nov 14 13:15:16 Error relocating /bin/ash: qsort: symbol not found Nov 14 13:15:17 Error relocating /bin/ash: readdir: symbol not found Nov 14 13:15:19 Error relocating /bin/ash: regerror: symbol not found Nov 14 13:15:20 Error relocating /bin/ash: getmntent_r: symbol not found Nov 14 13:15:21 Error relocating /bin/ash: setenv: symbol not found Nov 14 13:15:23 Error relocating /bin/ash: kill: symbol not found Nov 14 13:15:25 Error relocating /bin/ash: getgrouplist: symbol not found Nov 14 13:15:26 Error relocating /bin/ash: ntohs: symbol not found Nov 14 13:16:00 ow Nov 14 13:19:20 is uh, is there any bluetooth pan support or do i need to figure it out myself? Nov 14 13:42:26 Hei folks, I've a Raspberry Pi Zero WH to play with :). Raspian loads brcmfmac and I'm trying to figure out, that this chip can do. Ideally, it like to run STA + 802.11s at the same time. Unforturnatly, I havn't found much docs, yet. https://openwrt.org/docs/techref/driver.wlan/brcmfmac is almost blank. https://wikidevi.com/wiki/Brcmfmac says 802.11s is unsupported, but that's from 2014... Nov 14 13:43:08 Do you know some helpful, docs? Is the statements on 802.11s still up-to-date? Nov 14 14:16:15 Mornin all Nov 14 14:16:58 Looking for install guidance that might lean towards the dev side of the house Nov 14 14:20:12 yanosz: "iw list" and check "Supported interface modes" Nov 14 14:23:04 Back Nov 14 15:18:50 xback: it's all working for me :-) Nov 14 15:19:32 sorry, been really busy of late so irc response time erratic Nov 14 15:28:46 ldir: thanks. just pushed a few seconds ago :) Nov 14 15:29:08 ldir: I see you managed to tame hostapd.sh Nov 14 15:30:04 jow: lol - yes, thanks to you - the missing bit in the end turned out to be 'netifd needed a kick up the backside' Nov 14 15:31:22 jow: currently testing your uqmi patch Nov 14 15:31:25 jow: I'm thinking a wiki page with your words of netifd wisdom would be good. Nov 14 15:31:39 hope I still make it as I need to leave in about 20 minutes :) Nov 14 15:32:00 stintel: Thx Nov 14 15:35:52 ldir: I need someone who is able to structure the information for me, I'll gladly provide the content Nov 14 15:40:46 xback: ah great; currently thinking about a suitable testcase... I guess comparing /var/run/fw3.state before/after might be enough already. If this is inconclusive, then add a port forward from wwan to lan and see if reflection rules containing the wan ip get generated Nov 14 15:44:15 jow: ok. The more the info is 'out there' the more people can use it/tweak it/enhance it. I'll volunteer.....but just not right at this second Nov 14 15:47:29 all this so I can put 𓂸𓂹𓂺 in an SSID - one year I'll grow up, but evidently 2018 is not that year :-) Nov 14 15:49:17 jow: I won't make it .. :( got to pickup my kids. Next window to check it is Friday morning. srry Nov 14 15:50:32 the kernel bumps required a lot of time .. Nov 14 15:51:18 thats fine, don't worry :) Nov 14 15:51:56 cu all xx Nov 14 15:53:25 * ldir waves Nov 14 15:53:48 jow: is the utf8_ssid commit ok? Nov 14 15:54:19 ldir: looks sane to me Nov 14 15:54:42 * ldir fools jow again - sane?!!!??!?!?!?! :-) Nov 14 15:55:38 hilarity ensues - 𓂸𓂹𓂺 is censored on my macbook - but not my iphone Nov 14 15:56:41 ¯\_(ツ)_/¯ Nov 14 15:57:07 right - time to 'refresh' qnap support. locates brick - hits head against repeatedly Nov 14 15:58:45 unreferenced object 0xffff880008634aa8 (size 1024): Nov 14 15:58:46 comm "charon", pid 7396, jiffies 4295863817 (age 1997188.780s) Nov 14 15:59:36 size 1024, is that bytes or what (kmemleak) Nov 14 17:17:43 I am trying to build a software from outside the repo, but when I run make it complains about #include which is missing. Where would this file be? I've followed https://openwrt.org/docs/guide-developer/crosscompile Nov 14 17:17:46 I successully built openwrt for my box, including openssl, but I see that openssl/ssl.h is inside openwrt/staging_dir/target-mips_24kc_musl and not inside toolchain-mips_24kc_gcc-7.3.0_musl which is the directory I set as STAGING_DIR as the guide says to Nov 14 18:49:30 allizom: use -I and -L Nov 14 18:52:37 thanks jow, I've been searching meanwhile and I that's what I've just done. Going to try again Nov 14 18:53:35 note that the crosscompile article is mainly a simple guide for using the toolchain for one-off compiles outside the buildroot Nov 14 18:54:08 our cross gcc is patched to expect a STAGING_DIR env variable, hence its set to the toolchain dir itself to make it not complain anymore Nov 14 18:54:33 if you want to link with stuff produced by buildroot, it is better to set STAGING_DIR to staging_dir/target-* in the buildroot Nov 14 18:55:14 this should probably already fix all your issues Nov 14 18:56:18 jow: I believe I tried, but then it wouldn't find the compiler. Anyway, with -I and -L it seems to be working Nov 14 20:09:06 heh, I think I managed to work around the UDOO USB bug Nov 14 20:18:33 env: 'time': No such file or directory. wtf? Nov 14 20:20:27 see #openwrt Nov 14 20:25:54 sweet Nov 14 20:33:29 I might even submit this upstream Nov 14 21:28:40 * russell-- just sent a patch to add support for the Skylab SKW92A in their eval board to the openwrt-devel mailing list Nov 14 21:31:49 ooo utf8 ssid Nov 14 21:31:59 wonder how many devices this will break Nov 14 21:37:23 karlp: is this you? http://false.ekta.is/2015/11/zhone-6748-a2-teardown-bootlog-and-openwrt-thoughts/ Nov 14 21:50:18 karlp: i'm particularly interested in this fragment: "The Zyxel tricks for ATEN and ATSE to get into full mode don’t seem to work" Nov 14 21:50:36 * russell-- has a zyxel ;-) Nov 14 21:54:25 https://pastebin.com/be8DSv2f any idea what i'm missing here? Nov 14 21:54:33 mangix: dunno - at least it can be switched off Nov 14 22:10:56 util-01 is on outlook.com block list \o/ Nov 14 22:11:30 https://sendersupport.olc.protection.outlook.com/snds Nov 14 22:16:27 jow: can you forward me the mail you got from outlook? Nov 14 22:16:39 I tried configuring a mail forward but I don't understand jack shit of exim Nov 14 22:17:17 needlessly complex thing, almost reminds me of sendmail Nov 14 22:36:52 ldir How do I reset my config after messing around in make menuconfig? Nov 14 22:38:25 rm .config && make defconfig Nov 14 22:38:54 use both, or both do the same thing? Nov 14 22:38:58 weird, one of my UDOO Quads was spontaneously rebooting, u-boot claiming reason "por" Nov 14 22:39:01 (defconfig only really makes sense if you have preseeded your .config with at least the arch and target though) Nov 14 22:39:05 Or input that as one line? Nov 14 22:39:18 my other one boots fine on the same PSU, and the one that was rebooting now seems to work fine on the other PSU Nov 14 22:39:32 the deletion of your existing config is the crucial step, defconfig just an optimization/ refinement Nov 14 22:40:04 thanks pkgadd Nov 14 22:42:13 will git pull get me the last packages and luci or do I have to do scrypts update -a Nov 14 22:42:21 latest* Nov 14 22:43:34 worked it out thanks Nov 14 22:59:14 anyone still here? Nov 14 23:16:02 russell--: it is. Nov 14 23:16:29 nice Nov 14 23:17:45 I'd forgotten all abotu that one actually, Nov 14 23:17:59 I had to give it back, but put it out there in case it ever proved useful. Nov 14 23:18:11 i'm curious what the zyxel trick was Nov 14 23:18:28 presumably to get a fuller bootloader console Nov 14 23:18:39 oh, it was something I found on some other forums about these two commands and a serial number generator to enable "full" bootloader yeah Nov 14 23:19:08 I don't know if it's even real, it was just ideas from forum threads on getting more access. Nov 14 23:19:23 you should be able to search further with that? Nov 14 23:19:31 hmm, i think i saw that. but the links were dead :( Nov 14 23:19:52 but ... like you said, too many projects already. i put it back in a box Nov 14 23:20:12 yeah, they're probably all ewaste here. Nov 14 23:21:24 the cpu in my board is a dual core ARM, was yours arm or mips? Nov 14 23:22:15 hmmm ok, now it started rebooting on the other PSU as well :( Nov 14 23:44:40 fml. board fell off the desk with usb plugged into micro usb port and the port broke off Nov 14 23:47:27 same it's whatever you can tell from that page, that's all the data I'v egot sorry Nov 15 00:51:19 so, I'm interested in poking at the LEDs on the nanopi-neo2. Where would I even start? Nov 15 01:04:22 greearb_: /sys/class/leds/ Nov 15 01:09:24 Monkeh, er, ok...what can I do with that? Nov 15 01:09:54 I have this: /sys/class/leds/nanopi\:green\:pwr for instance Nov 15 01:12:12 greearb_: Check the files inside - you can select a trigger or simply set the brightness (which generally won't be adjustable, so just stick to 0 and 255). Just write to them. Nov 15 01:12:34 cat trigger Nov 15 01:12:44 will show the optionsi Nov 15 01:12:54 then echo the string you want into trigger Nov 15 01:15:17 find /sys -name trigger Nov 15 01:17:07 echo timer > trigger will blink the led and create magical sysfs files to set time on/off Nov 15 01:17:20 different triggers have different magical sysfs files Nov 15 01:23:58 thanks...will explore Nov 15 01:41:40 (and you only get this if the dts has all the leds listed properly) Nov 15 01:44:03 Which it does Nov 15 01:51:16 in this case :) Nov 15 02:02:15 yeah, seems to work...the 'blue' on is actually green, and the 'green' one is actually red, but works OK Nov 15 02:03:07 too bad there is no easy way to drive the NanoHat OLED (as far as I can figure) Nov 15 02:03:16 seems it needs 'bakebit' which has lots of dependencies Nov 15 02:22:19 ps Nov 15 02:51:43 so to enable/disable leds, you use the timer and then set delay-on/off to 500/0 accordingly? Or is there a better way? **** ENDING LOGGING AT Thu Nov 15 03:00:00 2018