**** BEGIN LOGGING AT Mon Jul 11 02:59:58 2016 Jul 11 05:10:40 build #252 of adm8668 is complete: Failure [failed shell_18] Build details are at http://buildbot.openwrt.org:8010/builders/adm8668/builds/252 Jul 11 05:10:52 build #245 of mpc85xx is complete: Failure [failed shell_18] Build details are at http://buildbot.openwrt.org:8010/builders/mpc85xx/builds/245 Jul 11 05:13:09 nbd: so if a company hires someone - they contribute code, it's still the devs? If someone works for the project and gifts the copyright assignment to an entity "Openwrt" they can't retrospectively change the copyright assignement to Openwrt UNLESS the same entity that is OPENWRT (I.e the copyright holder of the name / project openwrt which is not the german jow,blogic catlover gang) changes the name of the project the old copyright assign Jul 11 05:13:09 need to stay. Jul 11 05:14:25 you are welcome to appened Lede copyright tag but openwrt ones in the code need to stay not be renamed. That's GPL 2.0 in action right boys? Jul 11 05:14:48 build #368 of brcm47xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/368 Jul 11 05:15:29 the thing is, openwrt is not an entity Jul 11 05:16:18 then why are people the last 10 years assigning copyright to a non-enttity. Then if it isn't an organization it's a trademark of a the person who started openwrt. Jul 11 05:17:03 and that seems to be the basic reason lede split off Jul 11 05:18:04 Devastator: i need a larger scrollback buffer. great i'll add another remote to my Openwrt "Reboot" mirror and try and learn some git tricks. Jul 11 05:19:45 DonkeyHotei: if that is legal (and I doubt it) because all those years people gave copyright assignments to Openwrt (a trademark belonging to the person who started openwrt - what's his name again i read it before) - what they thought was a project - it shouldn't be ripped off and renamed, that's what the GPL 2.0 was to stop, rip it off but don't strip copyright assignments from the code. Jul 11 05:21:11 there were two founders, [mbm] and groz (who has gone to lede) Jul 11 05:21:14 DonkeyHotei: so you are saying it was just opportunisim, someone hasn't got around to registing a "foundation" - oh they did - the prpl foundation to help openwrt get enterprise ready - ... so we will make a new fork as a commercail opensource entity so we can cash in donations, hosting, rebranding. Jul 11 05:32:25 hojuruku: regardless of what copyright lines we put in the files over time, copyright itself can only be assigned via contract, and no such contract was ever made for assigning anything to openwrt Jul 11 05:56:38 build #330 of malta is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/malta/builds/330 Jul 11 06:47:03 build #367 of brcm47xx.legacy is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.legacy/builds/367 Jul 11 06:55:00 nbd: so how many other opensource projects do you think are vulnerable to the german gay catlover gangs rip-off attacks? ffmpeg, openwrt - what's next? I think this incident needs to be talked about more among the entire opensource community that GPL 2.0 doesn't always work and an opensource project isn't a project unless someone create sa legal entity for tax purposes? is that what you are saying? Jul 11 07:01:57 sorry about that it's my other world im involved in talking through me. on facebook it's the rainbow filter, on github it's hypenated-names and cats. on twitter other minority groups have taken to using brackets like shell (((((jewish))))) not that i have a problem with that, i have a problem with the heavy commercialization of lede. Jul 11 07:02:18 that cite their split on http://wiki.prplfoundation.org/wiki/Main_Page Jul 11 07:02:38 cat avatars that is. Jul 11 07:04:35 https://garryburnsantidiscriminationactivist.com/2016/04/27/luke-mckee-now-hacking-accounts/ i got called a hacker, because i did the crime of opening my mailbox, and recieving mail including contract negotiations with underage melbourne craigslists gay prostitutes and @garryburnsblog I'm not a hacker the mail was forwarded to me by his jilted partner. Just letting you know that I'm not a hacker as some are claiming. Jul 11 07:04:40 i have never noticed any pattern to who has a cat avatar and who doesn't Jul 11 07:05:23 has anyone else here been accused of being a "hacker" or "cracker" by the computer illiterate? Jul 11 07:05:31 yes Jul 11 07:05:41 i shrugged it off Jul 11 07:05:51 hojuruku: what do you mean by 'commercialization of lede'? Jul 11 07:06:32 DonkeyHotei: Garry is now trying to censor our lawyer for using material obtained from the "hack" or mass email forwarding using 3 email providers (bigpond.net.au gmail, yahoo) yet the mail message headers show Apple Mail MTA was used. Can apple mail MTA be remote controlled without a graphical GUI? Jul 11 07:07:46 nbd: venture capital funding by digital ocean, and mirrors on Chinese universities famous for reverse engineering American products. Jul 11 07:08:43 venture capital funding? Jul 11 07:08:44 wtf? Jul 11 07:09:10 where did you get that idea Jul 11 07:09:39 also, are you saying all chinese universities are famous for reverse engineering american products, or just some of them, or just the specific one that offered to host a mirror? Jul 11 07:11:01 DonkeyHotei: when you were accused of being a hacker did they run to the special minority GLLO gay police that contacted you, and then filed charges even though the police refused to collect forensic evidence from his system (including the underage gay child prositute negotiations my dad is including in evidence in his next submission to court) here's the email from the gay police http://pastebin.ca/3586599 See Garry's emails to the head o Jul 11 07:11:01 Police counter-terrorism unit Catherine Burns - who now is the state of going down. http://www.news.com.au/national/nsw-act/news/former-nsw-police-assistant-commissioner-clive-small-criticises-the-handling-of-the-lindt-cafe-siege/news-story/2efd3ee92b3e4c986d27a832413331c5 Here's gay media covering the special relationship between head of counter-terrorism and the gay nutjob calling me a hacker because I received all his forwarded emails s Jul 11 07:11:01 his jilted gay partner from sri-lanka he sponsored for a visa angry that he's hiring rent boys. http://www.altmedia.net.au/police-seek-help-on-potts-point-murder/64895 Jul 11 07:12:30 nbd: Digital Ocean is only VC funded. Digital ocean offered LEDE free hosting. Lede Mailing list ppl say - NO we don't want free hosting - go back to your boss and please give us a 2TB compute cluster so we can take over the continuous integration . Jul 11 07:12:37 i covered this before I don't want to rehash. Jul 11 07:13:35 'take over'? your theories are sounding a bit weird to me Jul 11 07:16:27 DonkeyHotei: the phone number of the special GLLO gay sex police officer in Rose bay is in that pastebin, as well as the head of NSW Police counter terrrorism Cath Burns writing back to the pedo suspect suing my dad for (no reshash to hardcore see here: http://linkis.com/www.quora.com/Who-do/uuscB ) Do you want to tell him before you charge people with cyber-terrorism for the crime of reviving emails from gay marriage activists including t Jul 11 07:16:27 cockpix and pictures of 17 year old asian (illegal) craigslist gay prostitutes rear ends - they might have to actually get evidence that i wasn't a victim of spamming, But no - the word of a gay activist is all they need to prosecute a cyber-crime, no cyber-evidence. Jul 11 07:17:00 i just asked for some details because you seem to be on a mission or something and i wanted to check if i could clear up some of the misunderstandings along the way Jul 11 07:17:40 in this country, "hate crime" laws cannot cover speech, because we have freedom of speech Jul 11 07:17:49 nbd: not all chinese universities, only Tsinghua - they offer majors in it and sell books about it. https://www.amazon.com/Reverse-engineering-tutorial-examples-Paperback/dp/7302262837?ie=UTF8&*Version*=1&*entries*=0 Jul 11 07:18:17 how is publishing a book about reverse engineering bad? Jul 11 07:18:25 i think that kind of knowledge is quite useful Jul 11 07:19:08 nbd: wouldn't it be great to have university students studying cisco reverse engineering making commits to openwrt, *gulp* I'm just saying.... Look at Huwei being banned from all government purchasing in America. What happens if they ruled all openwrt / lede derived commercial products could suffer the same fate. Don't tell me it's impossible. Jul 11 07:19:27 nbd: reverse engineering is useful knowledge, no problem with the science. Jul 11 07:19:33 i often enough have to use reverse engineering as well, because of american companies that don't give a crap about the license of code that they leech for free Jul 11 07:19:45 ;) Jul 11 07:19:46 nbd: Russia making nukes in the cold war was good because it ensured MAD - and no nuclear wars to date. Jul 11 07:20:42 nbd: but any science can be used for good or evil, and Tsinghua university who offered to give lede resources not openwrt has a history of saying hey chinese comany who wants to make a knock-off - give our reverse engineering department $$$ and we'll give you the know how you need. Jul 11 07:21:03 [citation needed] Jul 11 07:21:44 nbd: it's well documented that state owned companies in China even have a framework for this, Chinia's expionarge - i.e. real computer hacking as opposed to the crime of receiving child porn spammed in 100s of emails from a gay activist like me, is real and the information stolen is distributed to chinese industry and educational institutions. It's well documented. Jul 11 07:21:57 as opposed to the US? Jul 11 07:22:01 Well start with ChinaUncensored on youtube. Jul 11 07:22:28 what big nation isn't involved in pulling that crap? Jul 11 07:23:10 nbd don't get me wrong. Most of the cost of a CPU is jewish patent lawyers pay-offs, the OLPC project you were involved in - did they attempt to get cheaper hardware by getting the elites to say - we'll let you license all these patents to make low cost geodes as long as african kids get it? after all most cpu's are made of SAND, which is now being dredged up from the sea seeing we are running out of it on land now. Jul 11 07:23:59 A wafered grands of sand doesn't cost $400 dollars, it's the licensing and cross licensing. Jul 11 07:24:25 the main problem is that intellectual property has become a nasty cult Jul 11 07:24:28 especially in the US Jul 11 07:24:34 China's first consumer CPU the dragon was MIPS, but missing some patented instructions until they struck a deal. Jul 11 07:25:18 either way, i still fail to see how accepting a mirror from a chinese university somehow makes the project bad Jul 11 07:25:27 or commercialized Jul 11 07:25:29 nbd: you are right. McAfee suing microsoft over windows update - a program that downloads stuff to update itself is a patent. Look at cr.yp.to taking the piss of RSA and cyrpto patents. Wasn't BSD ports around before Windows update? Jul 11 07:26:03 nbd: so Openwrt gets banned from US government purchasing requests. It's happened to Huwaei it can happen to openwrt. Jul 11 07:26:48 pretty wild speculation in my opinion Jul 11 07:27:18 i fail to see what hosting a mirror has to do with anything Jul 11 07:27:28 and i don't particularly care about US government purchasing requests either Jul 11 07:28:02 look just one gentoo mirror got hacked, and it made a big deal. Tsingua university gets paid to do commercial espinarge and "hardware hacking" - and you want to give them a mirror. Go right ahead. Trust is a commodity that once lost you can't get back. Jul 11 07:28:27 just make sure it's a readonly mirror :P Jul 11 07:28:46 no write access is ever given out for hosting a mirror Jul 11 07:28:49 oh that doesn't make any difference if they have write access to the disks in China - I forgot :P Jul 11 07:29:32 nbd: THEY ARE FRIGGEN PUTTING MICROPHONES IN THE WINDOW PANES SOLD TO GOVERNMENT IN VIETNAM AND WERE CAUGHT DOING IT. GIVE THEM THE CHANCE.... Jul 11 07:29:47 at least they're not backdooring the entire internet ;) Jul 11 07:30:11 you'll have to be vigilant and think carefully about how opkg get's its signing data. Jul 11 07:30:30 that's why i wrote the signing infrastructure myself Jul 11 07:30:40 and made it compatible to the openbsd stuff Jul 11 07:30:45 the openbsd guys are quite paranoid Jul 11 07:31:00 too much caps Jul 11 07:31:24 i don't blame them, did you see what happened at blackhat and how the FBI shut down the speech from the guy who could demonstrate a cisco to cisco router hopping virus that could take down the net :) Jul 11 07:32:37 nbd: but the guy offering the mirror on lede right before I got banned obviously sucked up the communist chinese propaganda in school, especially about their war with Vietnam. I was doing that 30 mins before I got banned from #lede-dev to make a point - you need to be careful about how you partner with. Jul 11 07:33:09 i was there when you got banned Jul 11 07:33:27 and it was because you were rambling wildly about off-topic stuff and refused to stop when asked to Jul 11 07:33:38 it had nothing to do with your opinions Jul 11 07:33:40 only with your behavior Jul 11 07:34:39 https://bugzilla.mozilla.org/show_bug.cgi?id=432406 wow viruses injected into mozilla - only for vietnamese users Jul 11 07:39:17 https://www.stratfor.com/sample/situation-report/vietnam-hanoi-reportedly-eyeing-us-spy-planes I'm just saying the once upon a time Australian Unions went on strike because they didn't want to sell junk metal to the Japanese because they said it would come back as bombs. They were right. Who's side is LEDE on? Jul 11 07:41:16 those australian unions would have been contrary to the Open Source Definition, then Jul 11 07:41:51 i still fail to see what random links about vietnamese stuff have to do with LEDE or OpenWrt Jul 11 07:41:53 To think the Germans would be on the side of the Ruski's but according to the book the Pink Swastiska hitler was macho gay, and the commies were effeminate gay hence the rival in hitler's rize to power : http://www.thepinkswastika.com/ I looks like the Fabian left queer has taken over the EU from the conservative queer nazis. Watch the Conserative nazi's complain 6 years ago that europe would have a migrant invasion and that the country w Jul 11 07:41:53 implode, also talking about the history of the gay jew fabian socalists: https://www.youtube.com/watch?v=dWLBYjEJMXI Jul 11 07:42:58 hojuruku: stop this here, last warning Jul 11 07:42:58 China is the enemy faction that we may fight ourselves at war with soon. China isn't very happy about trump and visa versa. Jul 11 07:43:02 ok. Jul 11 07:44:35 it is very much against china's interests to go to war Jul 11 07:45:22 I was kidnapped in China by the PSB. If it wasn't for Australian expats to surround me i'd be dead. I did a truthnews.com.au interview about it in 2011 called "the pale shadow of democracy" Jul 11 07:46:00 China has gone to war against the country i'm now in and people fear they will do it again. Jul 11 07:46:11 but i better be quiet, so you guys have fun. Jul 11 08:15:14 git remote add nbd https://git.lede-project.org/lede/nbd/staging.git - or do I need to use origin? Jul 11 08:18:09 you should not use origin, because that one's already defined Jul 11 08:18:13 that command is fine Jul 11 08:20:09 mktplinkfw2 -c -B TD-W8970v1 -s -k /home/fld/dev/openwrt/build_dir/target-mips_34kc+dsp_musl-1.1.14/linux-lantiq_xrx200/tmp/openwrt-lantiq-xrx200-TDW8970-initramfs-kernel.bin -o /home/fld/dev/openwrt/build_dir/target-mips_34kc+dsp_musl-1.1.14/linux-lantiq_xrx200/tmp/openwrt-lantiq-xrx200-TDW8970-initramfs-kernel.bin.new Jul 11 08:20:12 [mktplinkfw2] *** error: kernel image is too big Jul 11 08:21:39 the -k file is ~7.68 Mi in size, how exactly is that too big? Jul 11 08:38:20 nbd: i'm trying out your two latest patches to ath9k Jul 11 08:40:39 nbd: cooking up a new brew now. let the missues and her iphone facebook addiction beta test to see if it works. please tell me what ath9k debug flags you want on, because this is a debug build with unstripped kernel. i'll write a script that monitors wifi signal strength from proc when it dropps off gzip 128k openwrt log buffer. and ill tell the missus to go hard and not disconnect until it dies. Jul 11 08:42:26 JyZyXEL: some of the tp-links have a u-boot uimage limit to how big the kernel can be it loads into memory. I had a lot of my kernel modules compiled in to the kernel, i had to take them out and use modules for everything the openwrt way because i got the same error as you, and that was with lzma compression. Jul 11 08:42:59 hojuruku: how do i see what the limit is? Jul 11 08:43:57 JyZyXEL: factory images of tp-links load the kernel from a seperate partition to the rootfs, which of course has a size limit. It's set by mktplinkfs2 in it's headers i think. I have to get around to adding my custom tp-link with 8mb flash instead of 4 (the factory did it not me, ISP special order) https://forum.openwrt.org/viewtopic.php?id=65786 Jul 11 08:45:08 JyZyXEL: you don't want to change those limits, you'll your brick your router. you want to make your kernel image smaller, and use modules for everything. Jul 11 08:46:13 JyZyXEL: i'm no expert but from my tinkering i see the tp-link routers have a modified uboot - and the kernel image is put in a a u-boot uimage format, and the size limit would come from that - set in the factory bootloader itself. Jul 11 08:48:45 what is the limit set to? Jul 11 08:49:41 is it the same as the size of the kernel partition? Jul 11 08:57:16 JyZyXEL: http://blog.gaku.net/building-a-custom-kernel-for-tp-link-tl-wr841n/ Jul 11 08:59:10 different hardware Jul 11 08:59:14 https://github.com/jtreml/firmware-mod-kit/blob/master/src/firmware-tools/mktplinkfw.c answering your question - it's in that file under firmware layout Jul 11 08:59:15 static struct flash_layout layouts[] Jul 11 08:59:55 yeah but mktplinkfw command line probably the model of your router with the mktplinkfw2 version is going to set the file size limits. Jul 11 09:00:31 that is some older utility that doesn't even support this hardware? Jul 11 09:01:41 yeah you are using a newer one, what arch do you have again? Jul 11 09:02:04 it works on the same principals though. Jul 11 09:02:37 it's not ar71xx is it? Jul 11 09:02:45 no its TDW8970 Jul 11 09:02:51 xrx200 Jul 11 09:03:09 lantiq Jul 11 09:03:43 IMAGE_SIZE := 7680k Jul 11 09:04:08 that's you buddy - you got 8m flash and you are not getting over that image size limit, or you will overwrite your art partition and kill your wifi forever :P Jul 11 09:04:13 ahhh, so i was right at the limit :D Jul 11 09:04:54 JyZyXEL: on top of that you need 3x JFFS2 blocks after your squashfs or it's not going to work, so add 3x 64k to the maximum possible size of the image it generates to be safe. Jul 11 09:05:22 i guess i can not have Luci then Jul 11 09:05:23 target/linux/lantiq/image/tp-link.mk is where i got your limit from. Jul 11 09:05:52 JyZyXEL: yeah guess not, but i'm experimenting with mounting an extra readonly overlay with NBD (network block device) Jul 11 09:06:04 loaded up from busybox, when it's all working nice i'll make a blog post Jul 11 09:09:27 JyZyXEL: i see benefits for openwrt firewall, such as procd init scripts modififying a running file as things change, very cool. I'm pondering weather to go with fwbuilder gui support of clusters (openwrt running vrrp again it's in oldpackages)... Jul 11 09:11:10 but... i have recently changed my mind, once the wifi is debugged i can strip the kernel - i'll go back to openwrt firewall. the annoying problem I have is /etc/config/network not letting you define a physical interface to be both pppoe and have an ip address on it :P i have to do a hack via /etc/ppp/if-up.d to get around it. Jul 11 09:12:51 having ash scripts being able to use json to manipulate the openwrt firewall is pretty cool, so i'll be playing with that. i'm working on a naughty version of openwrt's multi-interface script mixed up with pppoe that did the same thing the fon wireless pinapple did only with fibre ethernet if you catch my drift :P Jul 11 09:13:21 oh interesting, i just did a fresh build with default settings with only "luci" added and got an image of 4.3 Mi Jul 11 09:13:50 JyZyXEL: well you are in business with 8m flash. 4mb flash blows. Jul 11 09:14:29 in my previous configuration i spent around 1 hour carefully selecting all kinds of stuff that i think i might need Jul 11 09:15:30 i just assumed it would have been the luci that bloated my image, but it must have been something else Jul 11 09:23:06 i wonder if i could just solder in a bigger flash chip :P Jul 11 09:24:31 probably would need to change values in all kinds places after that hehe Jul 11 09:28:33 though it looks like completely doable to stay within the 7.6 Mi limit by very carefully selecting only what i really need Jul 11 09:32:47 JyZyXEL: then you have to move the art partition to the right place at the end of your new flash chip, and reconfigure the image building system accordingly Jul 11 09:33:06 first of all you need a backup of all the partitions on your router's flash Jul 11 09:33:55 luckily for time being it looks like i can get by with the 7.6 Mi :) Jul 11 09:40:47 guesses: 0 time: 11:12:47:59 0.00% c/s: 142758 trying: C2hyec! - C2hye7k some great cyber-hacker i am. I don't know how long john incremental will take to crack tp-links new default password in their factory roms Jul 11 09:41:28 a million years or less? what say you guys? nbd you seem to be the crypto guru here :P root:$1$GTN.gpri$DlSyKvZKMR9A9Uj9e9wR3/:0:0:root:/root:/bin/sh john -i=All passwd.1 Loaded 1 password hash (FreeBSD MD5 [128/128 AVX intrinsics 12x]) Jul 11 09:44:34 JyZyXEL: back up your full rom using u-boot via serial. Instructions here: https://forum.openwrt.org/viewtopic.php?pid=330161#p330161 then knock yourself out, just don't reflash the u-boot and you wont be screwed. Jul 11 09:44:44 JyZyXEL: requires a 3.3v working serial connection though Jul 11 09:46:15 also do a printenv in uboot as well, that will be very useful knowing where to put your art partition. according to the datasheet for the chipset you have it will probably be at the end of the rom - 20000 hex or something. you have to rtfm on that. Jul 11 09:47:11 rather than trying to crack the root password i'm just going to flash back the factory squashfs, only with a new root password :P cracking their password is taking too long. Jul 11 09:48:05 you guys need to make a KEXEC and or network boot option in buildroot - helps people add new routers . Jul 11 09:49:06 it's ok i'll just patch base-files to make preinit open a shell and do no more. Jul 11 09:49:17 after setting up devs etc. Jul 11 10:08:11 build #267 of ramips.rt305x is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.rt305x/builds/267 Jul 11 11:25:02 build #279 of ar71xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx/builds/279 Jul 11 12:16:47 build #271 of brcm47xx.mips74k is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.mips74k/builds/271 Jul 11 12:31:01 blogic 49399 /branches/chaos_calmer/target/linux/ramips/patches-3.18/0061-SPI-ralink-add-mt7621-SoC-spi-driver.patch ramips: fix timing issues when using MT7621 spi Jul 11 12:31:02 blogic 49400 /branches/chaos_calmer/target/linux/ramips/patches-3.18/0037-USB-phy-add-ralink-SoC-driver.patch ramips: fix usb phy initialisation Jul 11 12:48:28 build #276 of ixp4xx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ixp4xx/builds/276 Jul 11 13:06:11 nbd: i'm running your latest, i need to know what ath9k kernel debug logs you want and i'll set up a script to dump the 128k logbuffer when the wifi dies. Jul 11 13:10:17 i don't have any good ideas of what debug logs to capture, i'd mainly like to know if it still dies Jul 11 13:14:51 fair enough, i'll keep you posted Jul 11 13:16:00 actually, you may want to build again today Jul 11 13:16:07 i found some more bugs Jul 11 13:16:11 and i'm still fixing some stuff Jul 11 13:18:37 i did build today.. you can see what i'm running at github.com/hojuruku - did i miss anything when i merged what you are running? Jul 11 13:18:48 i did the build before i rode to work about 4 hours ago. Jul 11 13:19:03 you may want to rebuild some time later today Jul 11 13:19:06 i'm working on stuff right now Jul 11 13:19:12 so not all of it is in yet Jul 11 13:24:16 sounds reasonable, it's 8pm now, no rush but if you do it in the next 4 hours let me know. i'll give you root if you don't have something to play with, and throw in busybox screen Jul 11 13:25:15 nbd: quick question - so the rpcd/fstools/other stuff are mirrored from the lede git to openwrt git, right ? in that case, i'll run through those patches as well, if you're ok with it Jul 11 13:27:41 wigyori: it's mirrored, yes. Jul 11 13:28:03 great, thx Jul 11 13:42:01 ah just reboot testing my pull request. finally it's done. https://github.com/openwrt/packages/pull/2941 you can merge it now i think i made everyone happy. Jul 11 13:43:27 nbd: i learn't something very interesting about procd, it can do interface monitoring at two places, once per service and once per instance Jul 11 14:50:32 build #260 of uml is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/uml/builds/260 Jul 11 15:39:49 build #267 of cns3xxx is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/cns3xxx/builds/267 Jul 11 15:56:02 build #258 of ar71xx.mikrotik is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.mikrotik/builds/258 Jul 11 15:56:07 hey nbd Jul 11 15:56:15 ey Jul 11 15:56:22 it just crashed, and guess what iw wlan0 station dump said it was still connected Jul 11 15:56:37 so i ran over to my wifes iphone and verified it - it was offline Jul 11 15:56:58 ok Jul 11 15:57:06 please try the updated version with the fixes in the staging tree Jul 11 15:58:38 surely no ath9k debug logs will help you, like interrupts or bss? Jul 11 15:59:19 i'll cook up the new version now, at least you know through incremental testing you didn't hit the mark with everything up to to what you did 6 hours ago. Jul 11 16:03:33 nbd: apart from the benefit of being able to use out of tree kernel modules, you got no problem if I strip the kernel? Jul 11 16:12:19 nbd: HEAD is now at 30416b4... ath9k: implement temperature compensation support for AR93xx and newer is that better? Jul 11 16:12:43 yes Jul 11 16:14:52 IEEE 802.11: handle_action - unknown action category 4 or invalid frame ---- that's a red herring is it not? Jul 11 16:15:05 from hostapd Jul 11 16:19:37 you can ignore that Jul 11 16:24:27 is that problem you guys are dealing with in ath9k ? Jul 11 16:24:32 or another radio ? Jul 11 16:34:00 build #355 of realview is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/realview/builds/355 Jul 11 16:40:20 -initramfs-kernel.bin is not uImage compatible? Jul 11 16:40:37 i tried tftpbooting it and got "Uncompressing ... LZMA: uncompress or overwrite error 1 - must RESET board to recover" Jul 11 16:42:25 ah, perhaps my ramdisk image Compression has to be set to (lzma) Jul 11 16:48:52 yup, that was it Jul 11 16:57:38 groz: yep Jul 11 16:58:20 groz: i'm building it again now to see if it can fix the problem. my wife's facebook addiction is the tester. me grumbles about being baited to post a lede link, oh well here goes: https://bugs.lede-project.org/index.php?do=details&task_id=13 Jul 11 16:58:23 am I reading correctly, the symptom is, runs well for a while, then, altho ap shows client still connected, no data is flowing anymore Jul 11 16:58:43 haha, no, i'm not baiting anybody Jul 11 16:58:49 groz: yep, until it times out. the ssid goes offline, and the interface can still scan for other ssids, but it has stopped broadcasting bss. Jul 11 16:59:00 no it's my joke on the others who know the backstory, dont' go there. Jul 11 16:59:04 ok, that's different that the one I'm saying Jul 11 16:59:09 or rather, thinking of Jul 11 16:59:16 the issue I ran into, is with ath10K Jul 11 16:59:23 and a wds link Jul 11 16:59:37 similar symptom tho, both ends, the client and the ap show connected via iw Jul 11 16:59:39 it takes about 1-3 mins before the actual iw wlan0 station dump catches up with reality - that's what i just found out tonight. Jul 11 16:59:41 but no data flowing anymore Jul 11 17:00:00 Mine are in a bridge, so, I can catch it quite quickly Jul 11 17:00:00 but after 3 mins - or 5 mins or something it eventually goes offline? Jul 11 17:00:12 because stp should be flowing over the link Jul 11 17:00:19 I tested a kludge on the weekend Jul 11 17:00:23 stp over wifi? Jul 11 17:00:31 you mean with WDS? Jul 11 17:00:34 yes, it's a bridged setup Jul 11 17:00:36 yes Jul 11 17:00:42 ok that would work. Jul 11 17:00:46 so what I tested over the weekend, and it worked Jul 11 17:01:03 watch the client side of the bridge, the one in 4addr mode Jul 11 17:01:15 and when it goes 20+ seconds with no increases in the data flow count Jul 11 17:01:20 you are using WDS 4byte macs whatever not relayd or something? WDS seems to ad more bugs, then there is the MUlti SSID bugs Jul 11 17:01:22 down the adapter, and up the adapter right away Jul 11 17:01:27 here's my favorite ath9k bugs... Jul 11 17:01:36 https://dev.openwrt.org/ticket/12372 Jul 11 17:01:48 this monster, though i don't get it anymore thank god. https://dev.openwrt.org/ticket/11862 Jul 11 17:01:53 yes, I'm using a strait wds connection to bridge two hardwire networks Jul 11 17:02:13 what i was talking about a few mins ago - the openwrt link - BEFORE i found the lede boys complaining with the same Jul 11 17:02:14 https://dev.openwrt.org/ticket/22709#comment:4 Jul 11 17:02:31 the primary link between them is an optical fibre link,, but, on occaision it has issues with power going out at one of the fibre bridge boxes Jul 11 17:02:39 so I need it to fail over onto a wifi link for backup Jul 11 17:03:11 ahhh, that ticket, wrt160NL Jul 11 17:03:18 those have NEVER been fully stable Jul 11 17:03:29 I've got a few of them here Jul 11 17:03:33 dont use them anymore Jul 11 17:05:16 I still have one in an out of the way spot, connected in as a client Jul 11 17:05:28 it's got a usb temperature / humidity gadget plugged in Jul 11 17:05:42 script on it writes readings to a database on a server in the closet every 5 minutes Jul 11 17:23:18 whois github chaptar, is he here? Jul 11 17:25:05 afaik no - he regularly posts to mailing list Jul 11 17:25:52 what are your thoughts on making the mini_snmpd service start by default, now that new script binds to the lan (internal) interface by default. If there is no lan interface it wont start. Jul 11 17:25:54 https://github.com/openwrt/packages/pull/2941#issuecomment-231795208 Jul 11 18:26:01 https://github.com/openwrt/packages/pull/2941 blogic i think wrote ubox, he should have a look at that, /sbin/validate_data casts default variables to the wrong type in shell - casts them as a string not as a number. Jul 11 18:39:40 nbd: that's me running your latest version Jul 11 18:40:04 Reboot (HEAD, r931+32) Jul 11 18:57:04 build #253 of mvebu is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/mvebu/builds/253 Jul 11 19:12:49 ubus_strerror - does that include different exit $? responses for shell scripts? Jul 11 19:29:08 build #331 of malta is complete: Failure [failed shell_18 compile_9] Build details are at http://buildbot.openwrt.org:8010/builders/malta/builds/331 Jul 11 19:29:28 build #323 of rb532 is complete: Failure [failed shell_18 compile_9] Build details are at http://buildbot.openwrt.org:8010/builders/rb532/builds/323 Jul 11 19:29:44 build #273 of octeon is complete: Failure [failed shell_18 compile_9] Build details are at http://buildbot.openwrt.org:8010/builders/octeon/builds/273 Jul 11 19:33:43 build #303 of ar71xx.nand is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ar71xx.nand/builds/303 Jul 11 20:03:13 it would be nice if someone documented the exit levels from ubus command. Jul 11 20:03:14 https://github.com/openwrt/packages/pull/2941#issuecomment-231842265 Jul 11 20:38:45 yay its' merged. Jul 11 20:39:49 build #304 of x86.xen_domu is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/x86.xen_domu/builds/304 Jul 11 20:52:54 https://github.com/troglobit/mini-snmpd/issues/5 nbd ath9k doesn't update /proc/net/wireless - what is the proper way to get wifi stats via snmp, don't say scraping iw :P they hate that. Jul 11 20:59:00 hojuruku: the fun with ubus, particularly if you have ubus->rpcd->shjell is that you can get different things depending on wher eit failed. Jul 11 21:00:45 i see in the source there libubus has something to do with the error but i haven't worked out all the possible error levels, as a package maintainer now my users will tell me if it's a problem :P Jul 11 21:12:03 build #311 of ramips.mt7628 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ramips.mt7628/builds/311 Jul 11 22:11:08 build #273 of arm64 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/arm64/builds/273 Jul 11 22:23:57 build #260 of ath25 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/ath25/builds/260 Jul 11 22:34:47 build #368 of brcm47xx.legacy is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx.legacy/builds/368 Jul 11 22:56:53 build #369 of brcm47xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/brcm47xx/builds/369 Jul 11 23:02:47 build #363 of bcm53xx is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/bcm53xx/builds/363 Jul 11 23:48:30 build #246 of adm5120 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org:8010/builders/adm5120/builds/246 Jul 12 00:12:43 build #334 of lantiq is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq/builds/334 Jul 12 00:15:21 build #264 of ep93xx is complete: Failure [failed shell_13] Build details are at http://buildbot.openwrt.org:8010/builders/ep93xx/builds/264 Jul 12 00:34:35 build #288 of xburst is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/xburst/builds/288 Jul 12 00:58:59 build #11 of socfpga is complete: Failure [failed compile_4] Build details are at http://buildbot.openwrt.org:8010/builders/socfpga/builds/11 Jul 12 02:12:09 build #275 of lantiq.xrx200 is complete: Failure [failed compile_5] Build details are at http://buildbot.openwrt.org:8010/builders/lantiq.xrx200/builds/275 **** ENDING LOGGING AT Tue Jul 12 02:59:58 2016