**** BEGIN LOGGING AT Sat Jul 07 03:00:00 2018 Jul 07 04:14:18 hi Jul 07 04:14:29 need hel Jul 07 04:14:32 p Jul 07 05:57:03 Are the diagnostics supposed to work out of the box after initial install? I keep getting a ping: bad address 'lede-project.org' Jul 07 05:58:23 i think additional packages are needed, mentioned on the diagnostics page Jul 07 05:59:08 in particular, any of the packages that provide a "host" command Jul 07 06:00:14 DonkeyHotei, nothing mentioned on my diagnostics page. I also tried to update the package lists on the software page and got a bunch of could not resolve issues Jul 07 06:00:41 perhaps your wan settings are not functioning? Jul 07 06:00:45 I was looking at the wiki and noticed that my device has a interface for the CPU... is my CPU not connected to a switch port by default? Jul 07 06:01:02 depends on the hardware Jul 07 06:01:34 https://openwrt.org/toh/hwdata/tp-link/tp-link_tl-wdr4300 Jul 07 06:02:15 I also don't have anything plugged into the WAN port because after I congfigured the static IP it wasn't being picked up when plugged in via WAN Jul 07 06:02:27 I tried to disable the VLAN's... but not sure that's working Jul 07 06:03:17 are you trying to create a "dumb ap" device? if so, the diagnostics page will not function Jul 07 06:05:24 DonkeyHotei, Ah... yes I did configure it as only a switch and not a router Jul 07 06:05:28 https://openwrt.org/docs/guide-user/network/wifi/dumbap Jul 07 07:21:43 Ola. What exactly needs to be enabled in .config for lxc support? https://pastebin.com/XXf3KNSF Jul 07 07:23:18 https://pastebin.com/L6hFBXBJ from kernel config Jul 07 07:28:51 https://pastebin.com/v91LKW3q Jul 07 07:29:08 lede-17.01 was fine, openwrt-18.06 is not by default Jul 07 07:30:18 does lxc-checkconfig checking for wrong modules? Jul 07 07:39:59 ah http://lists.infradead.org/pipermail/lede-commits/2017-February/002779.html Jul 07 07:41:55 DonkeyHotei: im not sure.. Jul 07 07:42:43 muhaha: btw, lxc does not compile in 18.06 Jul 07 07:42:48 i thought it/i was good to go, when i got that notification about "someone" logging in to the cloudflare account.. but i dont understand what im missing in the ddns setting Jul 07 07:43:15 DonkeyHotei: where is problem? I have lxc in 18.06 Jul 07 07:43:30 build from yesterday Jul 07 07:44:03 muhaha: apparently something broke with the migration to gcc7: https://pastebin.com/Vwygdjfa Jul 07 07:45:43 meffe: typically, ddns update services use a special "access" url, and curl/wget is a way to trigger it manually Jul 07 07:46:24 yeah thats right.. Jul 07 07:47:46 i did look at it just a minute ago - but "all" i found was an bash script that did the update by first looking for an IP change, and then changed it - but the curl did retrieve at least 2-3 api.cloudflare.com sites.. Jul 07 07:48:12 ill try to figure out what one liner openwrt's ddns-cloudflare script is using, and see if i can get it to work with only curl Jul 07 07:48:32 DonkeyHotei: ah. I am fine with old gcc in Ubuntu 17.10. But still, problem is that at least kmod-bridge was removed from kernel and lxc-checkconfig is still checking for this module .. There are othere problems like vlan module, etc. Jul 07 07:48:55 muhaha: i'm on 14.04 Jul 07 07:49:32 the gcc in your distro is not that relevant Jul 07 07:49:48 openwrt now uses gcc7, at least for x86 Jul 07 07:50:15 I succesfully compiled Powered by LuCI openwrt-18.06 branch (git-18.163.60998-c379f00) / OpenWrt 18.06-SNAPSHOT r7074-b72bced2d7 Jul 07 07:50:34 which target? Jul 07 07:50:42 x86/64 Jul 07 07:50:47 odd Jul 07 07:52:11 what is CONFIG_GCC_VERSION in your .config? Jul 07 07:54:27 zcat /proc/config.gz |grep -i CONFIG_GCC_VERSION Jul 07 07:54:28 root@OpenWrt:/# Jul 07 07:54:39 Iam using g++ instead of gcc, is it possible? Jul 07 07:54:40 not the kernel .config Jul 07 07:54:47 the openwrt .config Jul 07 07:56:31 if global config is not somewhere in builed artifact -> I can not tell it. I am building it in docker containers, so its lost right now:] Jul 07 07:57:33 daemon.err hostapd: Failed to set beacon parameters Jul 07 07:57:47 daemon.notice hostapd: handle_probe_req: send failed Jul 07 07:59:22 muhaha: g++ is just the C++ compiler of the GCC collection. doesn't matter. Jul 07 07:59:44 btw you can use zgrep as well instead of zcat $foo|grep $bar Jul 07 08:00:06 zgrep on busybox? Jul 07 08:00:37 oh :( Jul 07 08:00:39 seeing those hostapd messages on a meraki mr24 with recent image Jul 07 08:00:39 nevermindn me Jul 07 08:01:24 Can anyone point me to the lxc-checkconfig source code, please? Jul 07 08:03:33 DonkeyHotei: i've been trying to configure and set it up in luci, so i thought i should give it a try in the cli - this is what happens when running '/etc/init.d/ddns start' https://paste.fedoraproject.org/paste/jVygs7uO348KaVSDLj0pdA Jul 07 08:03:44 morning Jul 07 08:06:11 o/ ultito Jul 07 08:06:11 misconfiguration in luci, looks like Jul 07 08:07:37 meffe: you don't have it disabled still by any chance? pastebin your config? Jul 07 08:07:56 ill remove everything, and skip the luci part.. and try to reconfigure it Jul 07 08:08:04 Borromini: no it's enabled Jul 07 08:08:41 maybe some typo since sh is picking up operands that shouldn't be there apparently Jul 07 08:14:10 something is definitely up with yesterdays snapshot Jul 07 08:15:36 I uninstalled all ddns-scripts and luci-app-ddns and so on, rebooted the router.. and ssh'd in, and ran 'opkg update; opkg install ddns-scripts' and this is what i get https://paste.fedoraproject.org/paste/pcQnDzhbAfFwl7OOZ5xFQw Jul 07 08:18:12 anyone? Jul 07 08:18:19 I created an issue for lxc-checkconfig if its ok... https://github.com/openwrt/packages/issues/6409 Jul 07 08:18:56 meffe: i have a 'plug-in' for No-IP e.g. since my provider is No-IP Jul 07 08:19:15 you need to install a similar dependency probably Jul 07 08:19:50 my provider is he.net which does not need a dependency Jul 07 08:19:57 ok Jul 07 08:20:29 uhm, what? but that "WARN : Service section disabled! - TERMINATE" when installing the ddns-scripts ? Jul 07 08:20:51 probably refers to your config Jul 07 08:21:02 what config? Jul 07 08:21:06 ddns? Jul 07 08:21:19 yes Jul 07 08:21:39 I removed that before installing Jul 07 08:21:52 the default is disabled Jul 07 08:22:22 uh, what? Jul 07 08:27:08 rebuilding hostapd fixed it Jul 07 08:29:27 https://openwrt.org/docs/guide-user/network/traffic-shaping/sqm doesn't work in 18.06, the package luci-all-sqm doesn't exist in https://downloads.openwrt.org/releases/18.06.0-rc1/packages/mipsel_24kc/luci/ Jul 07 08:29:36 luci-app-sqm Jul 07 08:32:29 muhaha: i forgot, you can look in /proc/version Jul 07 08:34:14 a__pi: which target is it you're using? Jul 07 08:35:04 i don't know if one can check the buildbot logs for package failures Jul 07 08:35:05 DonkeyHotei, Borromini: here's the second attemp.. everything (ddns-scripts. luci-apps and configs removed first), then reinstalled ddns-scripts and ddns-scripts_cloudflare.com-v4 (and got that 'Service section disabled!' when installing ddns-scripts).. here's after i configured the ddns config and start it, config included. https://paste.fedoraproject.org/paste/sw0V5xms-KjT1J6heUrZow (the 'sh: 1: Jul 07 08:35:07 unknown operand' keep echoing out in the terminal every 60sec or so) Jul 07 08:35:10 TP-Link TL-WR841N/ND v9 Borromini Jul 07 08:38:53 meffe: try running /usr/lib/ddns/dynamic_dns_updater.sh -v 1 Jul 07 08:44:33 https://paste.fedoraproject.org/paste/KoCUr6fdD2GRLl5IG~10IA Jul 07 08:53:16 well that sheds some more light i think? line 33? Jul 07 09:01:20 maybe, but then it's the script itself that's wrong i guess Jul 07 09:02:24 ill have to get back to this issue when i get back home again.. i need to go out and my me some IPA's for the game later today (swe vs eng).. ill check in later, thanks. Jul 07 09:11:50 good luck Jul 07 09:43:02 i got it working now, but it's updating the wrong domain. i'll be back later, to receive some help from you guys :) Jul 07 09:43:03 later Jul 07 09:46:41 still rc1... :-( Jul 07 10:03:06 Is there any solution for https://bugs.openwrt.org/index.php?do=details&task_id=1516 ? Jul 07 10:28:26 * ldir wibbles Jul 07 10:42:08 * russell-- saw something weird ... /etc/config/dropbear was corrupted Jul 07 10:42:21 which caused dropbear to not start Jul 07 10:42:46 it was filled with 130-some bytes of null bytes Jul 07 11:37:26 ashkan: how have you done the initial installation? Jul 07 11:52:36 Borromini: maybe you can help me out now.. i got the ddns-script working-ish.. problem now is that if i have the subdomain that i want to have updated, in the 'domain' field in the ddns config - it complain "Could not detect 'zone id' for domain.tld: 'svartgrind.meffe.se'". so i figure that i need to set the domain field to 'meffe.se' which is the correct zone. and when i do this, the script runs Jul 07 11:52:39 successfully - and changes the ip on the _wrong_ domain.. it changes on 'meffe.se' instead of 'svartgrind.meffe.se' -- any thoughts about this? Jul 07 11:53:31 ashkan: forget the question, I can see it was done via initramfs kernel and your error message is "[ 3.176639] UBI error: no valid UBI magic found inside mtd6" Jul 07 11:54:01 meffe: sorry, no idea about subdomains. you're managing meffe.se yourself I presume? Jul 07 11:54:02 ashkan: which means the fist few bytes of your mtd6 partition are not "UBI#" for some reason Jul 07 11:55:56 Borromini: hm. well, this script must be able to do that - i mean, everyone is using subdomains for something.. so im pretty sure it should be able to do that. -- and yes, i'm managing meffe.se myself. Jul 07 11:57:19 Borromini: what confuses me.. is that the 'lookup_host' option/field - is the domain it is comparing my current WAN ip against.. and it's a subdomain (svartgrind.meffe.se) Jul 07 11:57:23 meffe: yeah, in fact i'm using a no-ip subdomain myself Jul 07 11:58:14 https://paste.debian.net/1032739/ < how it looks for me Jul 07 12:04:37 yeah, exactly like that is what i was doing first - when it says it cant detect the zone id for domain.tld Jul 07 12:04:59 i believe i need to use the zone id somehow, instead of my username Jul 07 12:05:11 ima google a bit Jul 07 12:05:29 maybe try the forum as well? i think the ddns-scripts maintainer is more active there than on IRC Jul 07 12:05:59 that might be a good idea Jul 07 12:06:17 are we talking the forum.openwrt.org site now? Jul 07 12:07:21 mangix: that pci renumbering doesn't do good things.... https://zerobin.net/?a502a6ae1eab9ff7#Cw4fUG0t29g53JFB0QsVK0yzMF++jtQ0YhL/N/sdInI= Jul 07 12:07:22 no lede forum Jul 07 12:07:29 openwrt forum is read-only Jul 07 12:07:45 well the original one is AWOL actually Jul 07 12:08:09 ok Jul 07 12:10:35 mangix: and even though iw list still shows the radios now, now they won't turn on. Jul 07 12:16:27 Borromini: i need to modify this script a bit i think.. if you look at this one https://gist.github.com/lifehome/eb3f7d798f9bc6720cdc7d2be1238d4f you can see that it uses the "zone_identifier" which is a "api"-like ID for the zone "meffe.se", and the ddns-script for cloudflare doesn't seem to have that.. ill see what i can do Jul 07 12:19:19 ok good luck Jul 07 12:20:28 meffe: when i look at the ddns scripts there's cloudflare and cloudflare.com v4 Jul 07 12:20:31 did you try both? Jul 07 12:20:35 https://downloads.openwrt.org/releases/18.06-SNAPSHOT/packages/mipsel_24kc/packages/ Jul 07 12:25:49 no i didn't try both, cause i believe cloudflare changed something so the v1 script doesnt work Jul 07 12:25:51 but i can try Jul 07 12:27:25 i'm not familiar with cloudflare so i don't know what does what, but it can't hurt to try Jul 07 12:37:04 ok, well, went back in time to 512c57e7f3 and now wifi stays working. weird. Jul 07 12:37:15 no time to bisect this now, wife needs internet :| Jul 07 12:37:41 :) Jul 07 13:01:20 Borromini: ok i got it working now, i had to modify one line in update_cloudflare_com-v4.sh. i changed "__RUNPROG="$__PRGBASE --request GET '$__URLBASE/zones?name=$__DOMAIN'"" to "__RUNPROG="$__PRGBASE --request GET '$__URLBASE/zones?name=meffe.se'"" Jul 07 13:01:25 and now it works flawless Jul 07 13:02:12 good to hear Jul 07 13:07:05 thank you for your time and help, i really appreciate it man! Jul 07 13:13:02 well i didn't do much more than pointing and suggesting ;) Jul 07 13:22:25 thats all it takes some times :) Jul 07 13:43:38 Hello ! Jul 07 13:44:27 I'm trying to implement the A+P protocol in a internet-facing openwrt router (ie sharing IP address by fixing ports range per clients) Jul 07 13:45:07 I'm looking into the code of firewall3 , but I don't see any possible options for masq = "1" Jul 07 13:45:58 however, iptables's target MASQUERADE do accept port ranges : root@modemtim:/usr# iptables -j MASQUERADE --help Jul 07 13:46:07 MASQUERADE target options: Jul 07 13:46:07 --to-ports [-] Jul 07 13:46:07 Port (range) to map to. Jul 07 13:46:07 --random Jul 07 13:46:07 Randomize source port. Jul 07 13:46:08 --mode Jul 07 13:46:10 NAT mode. Jul 07 13:46:45 my goal is to have each client having, for example, 2000 ports each Jul 07 13:46:56 (outgoing ports, that's it) Jul 07 13:47:15 somebody joined the #lede channel and asked about release status. i told them to try the rc, but they ran into a release blocker: Jul 07 13:47:16 [Sat 2018-07-07 02:45:52 AM PDT] openwrt on 18.06 does not build Jul 07 13:47:16 [Sat 2018-07-07 02:46:26 AM PDT] https://pastebin.com/r8zeeZ5G Jul 07 13:47:16 [Sat 2018-07-07 02:58:13 AM PDT] https://github.com/openwrt/packages/pull/6062 thats the bug i guess, unfortunately it seems the arch i have still uses that patch Jul 07 13:47:44 The idea here is to avoid any conntrack logging and avoid the administrative load Jul 07 13:56:44 obinou: lw4o6 does this, see https://github.com/openwrt/openwrt/blob/master/package/network/ipv6/map/files/map.sh#L143 for the code setting up firewall rules for that Jul 07 13:58:17 thanks, i'm checking this out Jul 07 14:30:35 no it's not what I was looking for, since I do not control anything on the IP provider side Jul 07 14:30:47 I need a purely client-side solution Jul 07 14:49:52 I've installed LEDE on a BT HomeHub 5 following the instructions here: https://openwrt.org/toh/bt/homehub_v5a Jul 07 14:50:50 I'm using the "lede-17.01.4-lantiq-xrx200-BTHOMEHUBV5A-squashfs-sysupgrade.bin" image at the moment. The device is booting and I can get to LuCI, but I don't see any wireless. Jul 07 14:52:09 It appears that probing the ath10k_pci driver fails, anyone else experienced this and know of a fix? Jul 07 16:03:55 karlp: weird. I have no idea then. Jul 07 16:06:07 maybe the pci numbering and the staging driver need to go together. who knows. i'm going to stop working on this soon. Jul 07 19:07:56 someone knows date...? Jul 07 19:10:02 probably about 2 weeks Jul 07 19:12:48 when its ready Jul 07 19:22:13 I hope it will be soon, my 6 routers will be very happy Jul 07 19:26:54 @ultito did you try the rc release Jul 07 19:28:12 right, the more people use rc, the sooner it's ready Jul 07 19:28:50 I am running OpenWrt 18.06.0-rc1 r7090-d2aa3a1b62 / LuCI openwrt-18.06 branch (git-18.163.60998-c379f00) on 1 of my routers right now Jul 07 19:29:35 somebody joined the #lede channel and asked about release status. i told them to try the rc, but they ran into a release blocker for them: Jul 07 19:29:35 [Sat 2018-07-07 02:45:52 AM PDT] openwrt on 18.06 does not build Jul 07 19:29:36 [Sat 2018-07-07 02:46:26 AM PDT] https://pastebin.com/r8zeeZ5G Jul 07 19:29:36 [Sat 2018-07-07 02:58:13 AM PDT] https://github.com/openwrt/packages/pull/6062 thats the bug i guess, unfortunately it seems the arch i have still uses that patch Jul 07 19:31:29 Westislander: no, I prefer regular final release Jul 07 19:31:41 right, the more people use rc, the sooner it's ready Jul 07 19:34:14 @ultito That is what I would normally do, however I wanted to determine what was different, alas I found a bug , still unfixed which makes loading startup, scheduled tasks , and firewall rules difficult to upload. Jul 07 19:34:52 So this might result in a fix prior to "final" release benefiting everyone. Jul 07 19:35:59 I do not think, that rc is good idea for 6 routers. Jul 07 19:36:18 even each is quite far away from me Jul 07 19:36:26 I agree, but you might want to pick one, say a spare and test that. Jul 07 19:36:38 Biggest UI difference I found was in wireless security tab: Jul 07 19:36:49 each is different hw :-) Jul 07 19:51:02 ultito: check the roadmap https://openwrt.org/releases/18.06/start Jul 07 19:53:24 mkresin: thanks Jul 07 20:04:17 Excuse the delay, my session blocked and I rebooted. Jul 07 20:04:37 So I was saying the biggest UI difference is in wireless security, new option Enable key reinstallation (KRACK) countermeasures Jul 07 20:05:45 Biggest UI difference I could find that is. 18 UI looks almost identical to previous lede 17 with the exception of the version numbers on software etc. displays Jul 08 01:11:17 https://pastebin.com/37Qkkq2Y who can tell me the result of this 10-line C snippet without running it Jul 08 01:11:46 i rarely used this function and could not make sense out of it based on manpage and online sample code Jul 08 01:12:09 strspn that is Jul 08 01:15:43 based on the manpage, should be zero Jul 08 01:20:41 you mean the 'initial' does not mean 'first'? Jul 08 01:21:38 then if you try it with strcspn, it behaves very differently, strcspn("7abc","7") will return 1, and strcspn("abc7","7") will return 4, so 7 can be initial or not initial Jul 08 01:21:59 thus "initial" means different across strspn to strcspn? am I missing something? Jul 08 01:22:26 huh :o i would've thought '0' too Jul 08 01:22:40 weirdo string functions Jul 08 01:23:29 strspn("7abc","7") will return 1, so should strspn("abc7","7") was my guess, but strspn("abc7","7") will return 0 instead Jul 08 01:23:44 so the key must be at the start of the string, which is odd Jul 08 01:24:04 just emailed the manpage maintainer anyways, i might be missing something stupid, whatever Jul 08 01:24:06 not odd Jul 08 01:24:42 the way i read the manpage is that strspn returns the position of the first byte of s1 that is not in s2 Jul 08 01:24:55 ^ Jul 08 01:24:56 if not odd, then why strcspn's initial means anywhere ... Jul 08 01:25:19 and strcspn returns the position of the first that is in s2 Jul 08 01:25:24 ok let me re-read... Jul 08 01:26:26 ausjke: what does the code you pastebin:ed return? I'd say from the man page it should return 4 ? Jul 08 01:28:04 i'd say from the manpage it should be zero Jul 08 01:28:13 ausjke: re-reading it again, strspn("banana", "ban") should return 3. Does it? Jul 08 01:29:05 and strspn("banana", "banfoo") should also return 3. Jul 08 01:29:58 and the strcspn would be the utf8 equivalent, so it'll return number of characters, UTF-8 characters Jul 08 01:30:38 hm, never mind that last statement. Jul 08 01:30:46 I just think the man page is inconsistent. Jul 08 01:36:29 SwedeMike: it's 0 Jul 08 01:37:32 ausjke: wow, weird. Jul 08 01:37:44 ausjke: oh, your pastebinn one returns 0? That makes sense. Jul 08 01:38:04 strspn("77abc","7")=2; strspn("abc77","7")=0; strcspn("77abc","7")=0;strcspn("abc77","7")=3 Jul 08 01:39:09 anyway it kind of making sense, just a bit strange Jul 08 01:39:45 yeah, that man page should have more text in it, explaining what's going on. Jul 08 01:40:27 i don't understand what there is not to understand in that manpage Jul 08 02:27:56 Is there a built in mechanism for ignoring "image too big" errors during compile so it doesn't stop? Jul 08 02:29:11 make image less big Jul 08 02:29:11 :D Jul 08 02:30:52 :/ Jul 08 02:32:30 i've already patched one image process to not crash, but now mkdapimg2 is giving me grief. Jul 08 02:33:02 if its too big, how are you going to flash it? Jul 08 02:34:04 I don't need to, i'm generating a lot of images with a particular set of packages and configurations. The ones that fit and work will be distributed. The ones that don't go in the bin. Jul 08 02:36:28 umm Jul 08 02:36:32 right Jul 08 02:36:32 lol Jul 08 02:39:50 isn't that the default behaviour (which you can tweak) anyways? Jul 08 02:40:19 not with mkdapimg2 Jul 08 02:41:50 image size check fails and them rm -f the file. Then mkdapimg2 fails because input file does not exist. Jul 08 02:43:28 it can't fit and work anyway if it's too big Jul 08 02:43:37 that really isn't the point. Jul 08 02:43:58 if you really want to waste your time, just comment out the check Jul 08 02:43:58 why should it stop all of the other images from compiling which may fit? Jul 08 02:47:35 it's not going to stop images that fit Jul 08 02:49:05 i mean i'm telling you it does Jul 08 02:49:21 I tell you! Jul 08 02:50:54 I want to go to bed Jul 08 02:50:58 but this race is just ridiculous Jul 08 02:51:01 :( Jul 08 02:51:10 it's nearly 4am **** ENDING LOGGING AT Sun Jul 08 03:00:00 2018