**** BEGIN LOGGING AT Tue Jun 19 03:00:04 2018
Jun 19 03:07:20 <Fishman> Rhcev6: give them back and ask if this person hates you
Jun 19 03:07:47 <Fishman> from the info i see on the wiki, those are bcm wifi.  ancient at that, too.
Jun 19 04:59:39 <landwork> hi everyone, I am having trouble with using netifd/proto/wwan.sh… in the script it checks for the usb devices using the following approach -> ttys=$(ls -d /sys/bus/usb/devices/$devicename/${devicename}*/tty* | sed "s/.*\///g" | tr "\n" " ")
Jun 19 04:59:54 <landwork> this only works for usb->serial devices that are option devices, not cdc devices.
Jun 19 05:00:25 <landwork> for example CDC devices end up with another level under tty (at least under 4.4.92).
Jun 19 05:00:30 <landwork> For example -> /sys/bus/usb/devices/1-1.1.2/1-1.1.2:1.2/tty/ttyACM11
Jun 19 05:02:28 <landwork> as opposed to (serial option) -> /sys/bus/usb/devices/1-1.1.2/1-1.1.2:1.0/ttyUSB0
Jun 19 05:03:10 <landwork> note the extra tty folder for ACM (CDC) devices.
Jun 19 05:03:21 <landwork> they actually don't work with wwan without modifying the script.
Jun 19 05:23:31 <owrt-snap-builde> build #32 of mediatek/mt7623 is complete: Failure [failed images]  Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7623/builds/32  blamelist: Vincent Wiemann <webmaster@codefetch.de>, Magnus Kroken <mkroken@gmail.com>, John Crispin <john@phrozen.org>, Koen Vandeputte <koen.vandeputte@ncentric.com>, Magnus Fr?hling
Jun 19 05:23:31 <owrt-snap-builde> <skorpy@frankfurt.ccc.de>, Lucian Cristian <lucian.cristian@gmail.com>, Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>, Hannu Nyman <hannu.nyman@iki.fi>, Daniel Golle <daniel@makrotopia.org>, Paul Spooren <mail@aparcar.org>, Jeremiah McConnell <miah@miah.com>, Daniel Engberg <daniel.engberg.lists@pyret.net>, Hauke Mehrtens <hauke@hauke-m.de>, Weijie
Jun 19 05:23:31 <owrt-snap-builde> Gao <hackpascal@gmail.com>, Andy Boyett <agb@agb.io>, Chen Minqiang <ptpt52@gmail.com>, Thomas Nixon <tom@tomn.co.uk>, Krystian Kozak <krystian.kozak20@gmail.com>, David Bauer <mail@david-bauer.net>, Marty E. Plummer <hanetzer@startmail.com>, Rosy Song <rosysong@rosinson.com>, Mirko Parthey <mirko.parthey@web.de>
Jun 19 05:40:26 <owrt-1806-builds> build #22 of mediatek/32 is complete: Failure [failed kmods]  Build details are at http://release-builds.openwrt.org/18.06/images/builders/mediatek%2F32/builds/22  blamelist: Vincent Wiemann <webmaster@codefetch.de>, Sven Eckelmann <sven@narfation.org>, Rosen Penev <rosenp@gmail.com>, John Crispin <john@phrozen.org>, Thomas Nixon <tom@tomn.co.uk>, Hannu
Jun 19 05:40:26 <owrt-1806-builds> Nyman <hannu.nyman@iki.fi>, Daniel Golle <daniel@makrotopia.org>, Paul Spooren <mail@aparcar.org>, Jeremiah McConnell <miah@miah.com>, Andy Boyett <agb@agb.io>, Chen Minqiang <ptpt52@gmail.com>, Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>, Mathew McBride <matt@traverse.com.au>, Martin Schiller <ms@dev.tdt.de>, Mirko Parthey
Jun 19 05:40:26 <owrt-1806-builds> <mirko.parthey@web.de>, Hauke Mehrtens <hauke@hauke-m.de>
Jun 19 06:13:25 <blogic> Hauke: https://patchwork.ozlabs.org/patch/927426/  fails to apply
Jun 19 06:41:40 <luaraneda> blogic, Hauke: That patch is trying to edit the file ¨target/linux/zynq/config-4.4¨, which was removed with 32ab2e1 (zynq: remove support for kernel 4.4)
Jun 19 06:42:49 <luaraneda> I think if you remove the hunk that edits the file ¨target/linux/zynq/config-4.4¨, the patch should apply
Jun 19 07:01:12 <werty> hi
Jun 19 07:02:00 <werty> make package/foo/clean cleans the package but how can i clean the "host part" of a package?
Jun 19 07:02:36 <blogic> package/host/foo/clean afair
Jun 19 07:02:42 <werty> okay thanks
Jun 19 07:03:45 <werty> blogic, do you also know where the include files are defined when a package is build? i mean which include files get copied to the target include folder
Jun 19 07:04:33 <blogic> include/*.mk
Jun 19 07:04:43 <werty> thanks
Jun 19 08:04:14 <wigyori> Hauke: the ethernet on the linksprite pcduino3 (a20) doesn't work. i'm yet to test the a10/a13 boards.
Jun 19 08:04:46 <wigyori> Hauke: the bananapro (other a20) is fine
Jun 19 08:53:32 <wilson8345> seems like something from Felix Fietkaus' tree stops ntfs-3g from being packaged
Jun 19 09:17:10 <rmilecki> blogic: i'd suggest not to accept code without licensing info
Jun 19 09:17:12 <rmilecki> ath79: add initial nand support for gl-ar300m
Jun 19 09:17:23 <rmilecki> see qca9533_glinet_ar300m_nand.dts
Jun 19 09:18:28 <rmilecki> same for the "ath79: initial gl-ar300m support" and the qca9533_glinet_ar300m.dtsi
Jun 19 09:19:01 <blogic> rmilecki: well i see it differently
Jun 19 09:19:13 <rmilecki> ok, how?
Jun 19 09:19:21 <blogic> i'd rather force people to participate in the review process and fix stuff post moprtem
Jun 19 09:19:33 <blogic> see, when the patch was in patchwork no one cared
Jun 19 09:19:38 <blogic> now in tree people look at it
Jun 19 09:19:46 <blogic> same goes for the other 50 patches i merged
Jun 19 09:19:54 <rmilecki> were these ath79 in the patchwork?
Jun 19 09:20:02 <blogic> i was not active for 10 days, no review/merging happened and patches piled up
Jun 19 09:20:21 <blogic> i would need to look it up, the nand one came from github i think
Jun 19 09:20:34 <blogic> i'll simply mail the guy and ask if i can add the license
Jun 19 09:20:46 <blogic> and i am 100% certain he'll be fine with it
Jun 19 09:20:53 <rmilecki> i don't see them there https://patchwork.ozlabs.org/project/lede/list/?series=&submitter=&state=*&q=ath79&archive=&delegate=
Jun 19 09:20:59 <rmilecki> that situation with github sucks
Jun 19 09:21:11 <blogic> well, you aree the only person refusing to use it
Jun 19 09:21:17 <blogic> but yes i hate it also
Jun 19 09:21:19 <rmilecki> oh, great ;/
Jun 19 09:21:34 <blogic> i fully share your feeling about github though
Jun 19 09:21:38 <rmilecki> i don't have time to play with visiting another web page/service
Jun 19 09:21:53 <rmilecki> i reply whenever someone pings me directly
Jun 19 09:24:33 <blogic> rmilecki: sent him an email, turns out ist hanetzer, a regular all will be fine, thanks for pointing this out
Jun 19 09:24:56 <rmilecki> thanks, I was just going to write him
Jun 19 09:24:57 <rmilecki> thank you
Jun 19 09:25:42 <blogic> it was the only way to fight backlog, i was at the kbd for 17 hours yesterday and stayed up until all builders finished to clean up the fallout
Jun 19 09:26:11 <rmilecki> blogic: i'm not blaming you for merging these patches
Jun 19 09:26:20 <blogic> i know, just explaining the context
Jun 19 09:26:29 <blogic> i actually spoke to jow aout it today
Jun 19 09:26:38 <blogic> we need to agree on some patternn for merging stuff
Jun 19 09:26:40 <rmilecki> blogic: just meant to note on that, you may reject such patches if you happen to notice missing license
Jun 19 09:27:02 <blogic> i actually manually fixed up about 30% of patches yesterday on the way
Jun 19 09:27:08 <rmilecki> i don't expect you to notice everything with this amount of work
Jun 19 09:28:26 <blogic> we are basically in the trap again where we have the project on the shoulders of a few people doing the day2day chores
Jun 19 09:28:52 <blogic> luckily we have many more regulars
Jun 19 09:29:02 <rmilecki> i'm aware of this
Jun 19 09:29:07 <blogic> we are way more inclusive now than before but could do with 2-3 more janitors
Jun 19 09:29:08 <rmilecki> and I have no idea for a solution
Jun 19 09:29:14 <blogic> we need janitors
Jun 19 09:29:43 <rmilecki> we need full time employee to just handle incoming patches :/
Jun 19 09:29:49 <blogic> not really
Jun 19 09:30:04 <blogic> its a matter of 10 hours / week
Jun 19 09:30:20 <blogic> its all ok as long as jow and myself are online
Jun 19 09:30:32 <blogic> but just being away for 1 week makes patches pile up
Jun 19 09:30:35 <rmilecki> 10 hours / week is still quite a lot for 1 person
Jun 19 09:30:47 <blogic> i am doing about 25 at the moment
Jun 19 09:30:54 <blogic> and have been doing sor for several years
Jun 19 09:31:39 <rmilecki> i have no idea how you manage to do that
Jun 19 09:31:55 <rmilecki> i can't afford spending an extra hour daily
Jun 19 09:32:38 <blogic> well the show must go on
Jun 19 09:32:39 <biboc> Hi, I'm using LINKIT7688 board, I would like to add a new mtd partition "data" so I can flash again my board and partition stay unchanged. I added a partition in LINKIT7688.dts:
Jun 19 09:32:47 <biboc> 				partition@50000 { 					label = "firmware"; 					reg = <0x50000 0x1f90000>; 				};  				partition@1fe0000 { 					label = "data"; 					reg = <0x1fe0000 0x20000>; 				};
Jun 19 09:32:58 <blogic> biboc: pastebin.com
Jun 19 09:33:02 <blogic> dont paste int he channel
Jun 19 09:33:20 <biboc> https://pastebin.com/VgKYa5Td
Jun 19 09:33:52 <blogic> ok looks good
Jun 19 09:33:59 <blogic> and the problem is .... ?
Jun 19 09:35:20 <biboc> Problem: https://pastebin.com/kVLTPYRb
Jun 19 09:36:07 <biboc> I use TFTP from uboot to flash it
Jun 19 09:36:08 <blogic> ok, the firmware partition is not being split
Jun 19 09:36:41 <biboc> And to recompile image, I use: make target/linux/{clean,compile,install} -j8 && make -j8
Jun 19 09:37:09 <rmilecki> biboc: did it work before adding "data"?
Jun 19 09:37:21 <biboc> If I use default configuration and recompile as I tell, it works
Jun 19 09:37:28 <blogic> biboc: what file did you flash ?
Jun 19 09:37:31 <biboc> So yes rmilecki, it works
Jun 19 09:37:47 <biboc> openwrt-ramips-mt7688-Linkit7688-squashfs-sysupgrade.bin
Jun 19 09:38:08 <rmilecki> biboc: did you sysupgrade?
Jun 19 09:38:31 <biboc> Because kernel crashed, I use tftp from uboot
Jun 19 09:38:35 <rmilecki> hm
Jun 19 09:38:37 <rmilecki> it should work
Jun 19 09:38:46 <biboc> It is what I think also :-)
Jun 19 09:39:00 <rmilecki> i thought that "firmware" got shrunk not-nicely and maybe it can't be mounted anyore
Jun 19 09:39:05 <rmilecki> s/mounted/split/
Jun 19 09:39:26 <rmilecki> biboc: can you pastebin a log from working boot?
Jun 19 09:39:32 <rmilecki> without "data" partition as I guess
Jun 19 09:40:05 <blogic> biboc: and a full bootlog please
Jun 19 09:40:44 <biboc> yes sure
Jun 19 09:43:39 <biboc> Here: https://pastebin.com/YMjf6qJE
Jun 19 09:43:50 <biboc> Early boot, OK/KO kernel
Jun 19 09:44:36 <rmilecki> is that OpenWrt build?!
Jun 19 09:45:31 <rmilecki> "Linux version 3.18.29 #2 Tue Jun 19 11:12:21 CEST 2018" this doesn't seem like OpenWrt string for Linux version
Jun 19 09:45:53 <rmilecki> weird, I expectec gcc info there
Jun 19 09:46:34 <rmilecki> so basically the problem is missing: "2 uimage-fw partitions found on MTD device firmware"
Jun 19 09:46:34 <karlp> Linux eg-037BCC 4.9.106 #0 Mon Jun 11 13:59:12 2018 mips GNU/Linux
Jun 19 09:46:38 <karlp> not on mine either?
Jun 19 09:46:43 <rmilecki> ok, nevermind
Jun 19 09:46:49 <rmilecki> i see that on bcm53xx
Jun 19 09:46:50 <biboc> (gcc version 4.8.3 (OpenWrt/Linaro GCC 4.8-2014.04 unknown)
Jun 19 09:46:58 <biboc> Yes I removed computer name + gcc
Jun 19 09:47:30 <biboc> So here is gcc: (gcc version 4.8.3 (OpenWrt/Linaro GCC 4.8-2014.04 unknown)
Jun 19 09:47:44 <rmilecki> biboc: my suggestion is to add debugging into uimage spliter code
Jun 19 09:48:01 <rmilecki> see why it doesn't split "firmware" anymore after your change
Jun 19 09:48:06 <biboc> Where is located uimage splitter code?
Jun 19 09:48:18 <rmilecki> target/linux/generic/files/drivers/mtd/mtdsplit/mtdsplit_uimage.c
Jun 19 09:48:29 <rmilecki> or directly in build_dir depends how you prefer to work
Jun 19 09:48:41 <biboc> thanks
Jun 19 09:48:47 <biboc> I'll let you know
Jun 19 09:50:36 <biboc> @blogic, do you another idea before I try to dig into mtdsplit_uimage?
Jun 19 10:01:56 <blogic> no idea
Jun 19 10:03:39 <karlp> blogic: I'm pretty much ath79/ar71xx only.  I've got a mt7621 for personal use at home, but no marvel or anything else exotic.
Jun 19 10:03:52 <blogic> karlp: oki
Jun 19 10:03:56 <blogic> i was not sure
Jun 19 10:04:00 <karlp> (the bastard child broadcom/ralink thing I'm poking a bit doesn't count until it works)
Jun 19 10:04:03 <karlp> no problem.
Jun 19 10:04:12 <blogic> hannu replied so all good
Jun 19 10:07:40 <ldir> blogic: I'm looking at this https://github.com/openwrt/openwrt/pull/1073/commits/2e79a5861188113e3db7ebbcb90ffdb4ed4f659e#diff-31d5103a53dd49aa9b8e1c039dde81f3R90 and wondering about this whole export gpio for usb power v get the usb driver to use the usb voltage regulator mechanism.
Jun 19 10:07:53 <biboc> How to enable pr_debug in mtdsplit_uimage.c?
Jun 19 10:08:45 <karlp> echo "f:mtsplit_uimage.c" > prdebug thing or something
Jun 19 10:08:47 <ldir> other devices you've ported use the fixed regulator mechanism.... which I think is the technically correct solution, and I recently modified the MIR3G to do the same thing.
Jun 19 10:10:33 <biboc> karlp, where should I do this? in cmd line at openwrt root repository?
Jun 19 10:11:43 <karlp> have you enabled pr_debug stuff in your make menuconfig?
Jun 19 10:11:54 <ldir> I personally think that at the very least a per-arch preferred mechanism should be indicated before we get further down the 'different boards do different USB gpio things' road.
Jun 19 10:12:15 <karlp> ldir: I think we should all move to the regulator stuff, I think it was just to get it in so it gets attention.
Jun 19 10:13:11 <karlp> biboc: global build settings -> kernel build options -> dynamic printk is what you need I think.
Jun 19 10:15:09 <biboc> I do not have a dynamic printk line but a "[*] Compile packages with debugging info" and "[*] Enable support for printk"
Jun 19 10:16:22 <karlp> pres / and search prink or pr_debug or something
Jun 19 10:16:30 <karlp> what tree are you working on?
Jun 19 10:17:08 <karlp> "Compile the kernel with dynamic printk" ? it was there on CC iirc, it _may_ have moved ifyou're back on BB era or something, but...
Jun 19 10:20:59 <ldir> karlp: I've been bitten before by the 'let's get it in and fix it later approach' - I have scars on the MIR3g to prove it :-)
Jun 19 10:24:34 <karlp> ldir: oh yeah, me too :)
Jun 19 10:24:40 <karlp> damned if you do, damned if you don't
Jun 19 10:25:26 <karlp> when you make it a regulator, you can still turn it on and off from userspace right?
Jun 19 10:25:46 <karlp> just you do it using special files in sys or something instead of echo > gpio ?
Jun 19 10:27:34 <biboc> I'm on CC
Jun 19 10:27:42 <ldir> I'm unsure of that - I know it gets controlled by driver load/unload
Jun 19 10:29:22 <biboc> Found it sorry
Jun 19 10:30:05 <karlp> ldir: becasue that's basically the only opposition to having it be controlled by regulators properly like that.
Jun 19 10:30:47 <karlp> as long as you cna show people how they can still turn off and on their usb ports again, to nuke badly behaving external modems, there'll be no objections to making them proper regulators.
Jun 19 10:30:54 <ldir> karlp: yes, I can clearly see that concern.
Jun 19 10:31:06 <karlp> the gpio thing is just carry over from when gpio controlled regulators weren't well setup in dt i think
Jun 19 10:32:15 <mkresin> what would be better than unloading/reloading the usb driver to restart a badly behaving external modem
Jun 19 10:32:36 <tmn505> blogic: please cherry-pick these two commits to 18.06: 4ec8c8c23ed50dbe9cabbfc544ae2780f0287fd0, 772258044b48036699302840abf96cd34c4e5078
Jun 19 10:33:59 <ldir> I need to poke into that.... later today probably tomorrow.  I can see the attraction of being able to remove power... but I'm not so sure that doing it from 'underneath the USB drivers vision' is a nice either.
Jun 19 10:34:43 <karlp> ldir: I don't hve a board with it, but on your mir3g, can you can see things like /sys/class/regulator and turn it on/off?
Jun 19 10:35:12 <karlp> https://stackoverflow.com/a/30041862 seems to imply that you had to register yet another layer to be able to turn it on /off
Jun 19 10:35:18 <ldir> it's upstairs without power... and I have to deal with real life right now - later :-)
Jun 19 10:35:19 <karlp> ldir: well, there's different "usb power" too,
Jun 19 10:35:42 <karlp> there's gpios that just turn on/off teh regulator for external vbus, and there's ones that turn on/off the actual usb host
Jun 19 10:44:00 <karlp> ldir: might want to try enabling REGULATOR_VIRTUAL_CONSUMER and /or REGULATOR_USERSPACE_CONSUMER and seeing whatyou get in sysfs/debugfs?
Jun 19 10:46:45 <karlp> userspace_consumer looks pretty gross. needs to add a bunch of extra DT nodes that are supplied by your regulator :|
Jun 19 10:54:32 <blogic> tmn505: done
Jun 19 10:55:15 <tmn505> Thanks
Jun 19 12:04:46 <owrt-snap-builde> build #868 of ramips/rt305x is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/ramips%2Frt305x/builds/868
Jun 19 12:07:49 <Spamcoll_> Hi there,
Jun 19 12:08:39 <Spamcoll_> Need help in achiving same solution in openwrt - https://community.netgear.com/t5/Mobile-Routers-Hotspots-Modems/Nighthawk-M1-Linux-RNDIS-USB-Connectivity-Issues-After-Latest/m-p/1515552
Jun 19 12:16:47 <jow> Spamcoll_: replace bin/bash with bin/sh and it should work
Jun 19 12:31:56 <f00b4r0> jow: ping?
Jun 19 12:34:07 <jow> f00b4r0: pong
Jun 19 12:41:33 <muhaha> Ola. https://bugs.openwrt.org/index.php?do=details&task_id=1593
Jun 19 12:41:58 <Borromini> Hauke: ping
Jun 19 12:42:43 <owrt-snap-builde> build #48 of gemini/generic is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/gemini%2Fgeneric/builds/48
Jun 19 12:46:01 <muhaha> Do I need to create patch with quilt, or is enough to do git diff file.c edited-file.c ?
Jun 19 12:48:19 <karlp> muhaha: as long as it applies cleanly and doesn't have any misleading commentary :)
Jun 19 12:53:20 <muhaha> karlp: unfortunatelly not :(
Jun 19 12:53:43 <muhaha> I dont get a difference between quilt and classic patching
Jun 19 12:55:11 <Borromini> muhaha: is this a git checkout?
Jun 19 12:55:18 <Borromini> if it is you can just use git to create your patch.
Jun 19 12:56:57 <karlp> what is "classic patchgin" ?
Jun 19 12:57:16 <karlp> quilt just helps you maintaina  sequence of htem and helps you apply/unapply in order.
Jun 19 12:57:30 <karlp> and gives you helpers for refreshing them and so on.
Jun 19 12:57:57 <muhaha> yes, git checkout openwrt-18.06 ; git diff gpio-nct5104d.src gpio-nct5104d-updated.src >> package/kernel/gpio-nct5104d/patches/301-mypatch.patch
Jun 19 12:58:26 <muhaha> Does not work.... Hunk #1 FAILED....
Jun 19 12:58:40 <Borromini> means you need to fix it up before trying to apply it again
Jun 19 12:59:33 <muhaha> Some guy tell it to me too, but I dont know what it means ... Unfortunatelly I dont know wrt build system
Jun 19 12:59:54 <muhaha> say hello to ldir :)
Jun 19 12:59:55 <Borromini> muhaha: git status will show you any tracked files that have been modified; you can commit them locally and use 'git format-patch -1' to generate your patch
Jun 19 13:01:01 * ldir is *REALLY REALLY* busy dealing with $REAL_LIFE
Jun 19 13:02:12 * karlp suspects you are trying to patch teh wrong files
Jun 19 13:02:18 <muhaha> Borromini: And will it be compatible withou you patching system? It should be merged to master and openwrt-18.06
Jun 19 13:02:35 <karlp> build will checkout a clean upstream kernel and then apply patches.
Jun 19 13:02:41 <karlp> yours need to be the right place in that system.
Jun 19 13:03:17 <karlp> basically, https://openwrt.org/docs/guide-developer/build-system/use-patches-with-buildsystem
Jun 19 13:03:49 <karlp> you can play with doing things "your way" when you can make the documented way work.
Jun 19 13:04:55 <Borromini> muhaha: if the patch can be dropped in the right place, gets picked up by the build system, and applies cleanly, that's all you need
Jun 19 13:04:58 <muhaha> oh gosh :( I am just a simple user, not developer
Jun 19 13:06:31 <karlp> congratulations, you're a developer now.
Jun 19 13:06:43 <karlp> if you want to start patching things, you are by definition a "developer"
Jun 19 13:06:50 <blogic> karlp: its really misovervalued
Jun 19 13:07:31 <karlp> for variouslevels of developer sure :)
Jun 19 13:13:03 <Borromini> blame balmer for devaluing it :P
Jun 19 13:13:25 <Borromini> muhaha: i'm a simple user as well... that's how you pick things up
Jun 19 13:13:39 <Borromini> the moment you're applying patches yourself you're not a simple user anymore :)
Jun 19 13:14:08 <muhaha> I just want to repair my router to get my internet back to another important things :)
Jun 19 13:14:29 <muhaha> *do
Jun 19 13:16:08 <muhaha> I used to do git diff > /patches/file.patch  .. And it always worked, now I want to do the same thing with different patch and bam... hunk failed ... :D
Jun 19 13:16:37 <karlp> so you're probably not patching the right files.
Jun 19 13:17:04 <karlp> you're not being very concrete with your examples.
Jun 19 13:17:13 <muhaha> https://github.com/lukasmrtvy/docker-lede-apu2c4/tree/snapshots-18.06
Jun 19 13:17:16 <muhaha> here it is..
Jun 19 13:17:24 <muhaha> Dockerfile describes what I am doing
Jun 19 13:18:10 <karlp> lede is no longer, long live openwrt, for starters....
Jun 19 13:19:00 <karlp> ok, none of that has "git diff > patches && fail"
Jun 19 13:19:08 <karlp> so... what _else_ is going wrong?
Jun 19 13:21:06 <karlp> strange way to customize a build too.
Jun 19 13:21:08 <muhaha> i was manually created these files and uploaded to git
Jun 19 13:21:43 <muhaha> Everytime I upgrade firmware(coreboot) on APU, half of thing are broken in wrt
Jun 19 13:22:34 <muhaha> its not strange, I am builing it with codefresh, its i a cloud... I have resource free PC to work on
Jun 19 13:22:43 <muhaha> *in
Jun 19 13:22:59 <karlp> it is a very strange way to customize  a build (I don't care about you doing in docker, that's yoru party)
Jun 19 13:23:40 <ausjke> massive mediatek updates on 1806, why mt7622/mt7623 under target/linux/mediatek while mt7620/mt7621/mt76x8 under ramips?
Jun 19 13:23:53 <muhaha> Of course If you want to do it automagically... You have to do strange things...
Jun 19 13:24:49 <muhaha> And with strange things you mean configuring kernel .config?
Jun 19 13:26:51 <Borromini> ausjke: because ramips is MIPS and mediatek/ is ARM afaik.
Jun 19 13:28:11 <ausjke> Borromini: makes sense, thanks
Jun 19 13:34:22 <muhaha> wel..
Jun 19 13:34:37 <muhaha> karlp: so? Any idea what is wrong?
Jun 19 13:35:35 <karlp> no, you still didn't say where you trying touse "git diff"
Jun 19 13:35:50 <karlp> and still haven't provided anyt indication of having tried the documented steps for making patch files.
Jun 19 13:39:44 <tmn505> muhaha: why not directly applying the patches instead of relying on OpenWrt buildroot?
Jun 19 13:44:02 <aparcar> jow can golang still go on 18.06? https://github.com/openwrt/packages/pull/6314
Jun 19 13:46:24 <karlp> you have _ages_ for this, why did you bring this up so late?
Jun 19 13:46:59 <aparcar> karlp: me?
Jun 19 13:47:24 <karlp> yes you
Jun 19 13:48:41 <aparcar> You should know be by now, it's the third time you comment on my backport PRs :)
Jun 19 13:48:58 <karlp> yes, and you keep doing the same things :)
Jun 19 13:49:17 <karlp> has this even landed in master yet?
Jun 19 13:49:48 <jow> yes, I merged it
Jun 19 13:49:49 <aparcar> Is there a RC1 yet?
Jun 19 13:49:51 <aparcar> Yes!
Jun 19 13:52:28 <aparcar> Sorry regarding the delays, I promise to speed up in future
Jun 19 13:53:00 <karlp> you're goign to be happy with this version of go for 18.06 now right? you're not going to want to change go versions and things are you?
Jun 19 13:53:57 <ausjke> does go have a stable mips version, or non-x86 version, in openwrt?
Jun 19 13:54:17 <ausjke> still feel go-binary is big in size, so is that rust lang, C rules...
Jun 19 13:58:01 <aparcar> karlp: I think I'm happy with it!
Jun 19 13:58:55 <aparcar> ausjke: I'm using it on some arm devices and it feels stable
Jun 19 14:01:21 <karlp> how many apps can you have though? it's all static right? every go binary is huge with everything in it?
Jun 19 14:02:13 <ausjke> that's go's selling point, but it does not scale on 8M/16M flash... I do think there are places it will shine, but not on low to mid-range routers
Jun 19 14:03:13 <jow> yeah, go is a total no-go (no pun intended) for space cosntrained devices
Jun 19 14:04:30 <ausjke> same to rust-lang, the 'system' language, that looking at its size, it should be "bloated system language" instead, there is no way for it to fly in embedded space
Jun 19 14:04:42 <shalzz> Hi! could this patch be backported to linux-4.9 for ar71xx? https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=36148c2bbfbe
Jun 19 14:04:56 <jow> nobody cares about sub-1G devices anymore
Jun 19 14:05:00 <shalzz> also maybe for the 18.06 release?
Jun 19 14:05:20 <jwh> yup, 1G+ or bust
Jun 19 14:05:46 <jow> todays notion of "embedded" is either android phones or iot thingies running node.js
Jun 19 14:06:09 <jwh> blockchain enabled fridges
Jun 19 14:06:09 <karlp> shalzz: the mechanisms it refers to, do they existin 4.9?
Jun 19 14:06:10 <ausjke> what about that lovely LoraWAN thing that is sub 1G?
Jun 19 14:06:23 <karlp> (i think jow menat <1G ram, not freq)
Jun 19 14:06:46 <jow> I mean <1G flash/storage / <1G ram
Jun 19 14:06:50 <jow> not cpu
Jun 19 14:07:19 <ausjke> oops, sub-1G also make me think about 4xx/8xx/9xx Mhz
Jun 19 14:08:22 <ausjke> the running-blockchain on the router thingy is called a fad, it will back to the ground some day
Jun 19 14:08:50 <jwh> theres actually an ad by a world cup sponsor that is about blockchain enabled fridges
Jun 19 14:09:10 <shalzz> karlp: I'm not sure
Jun 19 14:09:17 <ausjke> oh shit, blockchain will make that fridges really hot, should it be run within an oven instead?
Jun 19 14:09:37 <jwh> https://www.techradar.com/news/what-is-hdac-the-world-cup-blockchain-advert-decoded
Jun 19 14:09:38 <shalzz> karlp: if it isn't at least the value could be backported
Jun 19 14:10:50 <jwh> you must wait until all participating devices on your network have confirmed it
Jun 19 14:10:53 <muhaha> tmn505: you mean directly editing source files?
Jun 19 14:10:55 <jwh> before you can open the fridge door
Jun 19 14:10:58 <jwh> :D
Jun 19 14:11:02 <karlp> shalzz: sounds like something you should have alook at :)
Jun 19 14:11:25 <karlp> ausjke: they're already selling wall heaters that are bitcoin miners
Jun 19 14:11:48 <shalzz> karlp: this lwn article describes the history of the patch https://lwn.net/Articles/757643/
Jun 19 14:11:52 <ausjke> friends at amazon told me that they're using android for anything, including the echo-dot, that's like forcing a 500 pound guy to ride a kid's bike. thanks god Amazon bought freertos, otherwise it will put android on to MCUs, I'm sure about that.
Jun 19 14:12:21 <ausjke> s/anything/everything/
Jun 19 14:12:23 <karlp> shalzz: I'm not really interested personally, you may wanmt to try it out yourself and send it as a patch, or file a ticket to ask for it, with all the links.
Jun 19 14:14:02 <shalzz> karlp: okay thanks I will
Jun 19 14:28:56 <aparcar> karlp: My intention is to run some Go stuff on APU boards which I rather flash with OpenWrt than Debian. Having the possibility to easily have Go (bloat) on OpenWrt may increases the latter's use cases, making it available as "cloud" OS  to run services rather than having (again) a full grown Debian running underneath. Does that make sense?
Jun 19 14:37:13 <karlp> aparcar: I don't care what you do with it, I'm just against your general desire to continually churn versions of things in teh stable branches :)
Jun 19 14:38:16 <ausjke> aparcar: honestly i will just do debian to make life easier
Jun 19 14:39:12 <ausjke> aparcar: minimal size is probably 300MB on debian without the gui bloat
Jun 19 14:39:35 <aparcar> ausjke: sorry what do you mean?
Jun 19 14:40:02 <jwh> hnng
Jun 19 14:40:16 <ausjke> i mean cloud-os might be easier when just staying with debian, if you mean container that debian is too big
Jun 19 14:40:24 <ausjke> s/that/then/
Jun 19 14:40:28 <jwh> gonna have to upgrade my edge routers now :(
Jun 19 14:40:42 <jwh> Tue Jun 19 14:41:09 2018 daemon.err bird4: lonap_google: Socket error: Kernel does not support TCP MD5 signatures
Jun 19 14:40:45 <jwh> sadface
Jun 19 14:40:49 <ausjke> believe edge-routers run debian
Jun 19 14:40:57 <jwh> no
Jun 19 14:41:00 <jwh> edge routers
Jun 19 14:41:02 <jwh> not edgerouters
Jun 19 14:41:02 <jwh> :D
Jun 19 14:41:41 <jwh> despite their name, ubnt edgerouters are not sufficient edge routers
Jun 19 14:41:42 <jwh> :D
Jun 19 14:41:59 <ausjke> ubnt stole the general term for its product, indeed
Jun 19 14:42:32 <stintel> jwh: not enough memory for full BGP table?
Jun 19 14:42:49 <jwh> on the smaller models yeah
Jun 19 14:42:58 <jwh> and nowhere near enough cpu power
Jun 19 14:43:08 <jwh> the bigger models are slightly better but lack the ports
Jun 19 14:43:36 <jwh> the "infinity" is somewhat better in that regard
Jun 19 14:45:33 <stintel> heh, nice price too
Jun 19 14:45:43 <jwh> yeah
Jun 19 14:45:51 <jwh> the "beta" was only 500 bucks
Jun 19 14:46:24 <jwh> I still use x86 for most things
Jun 19 14:46:32 <jwh> reasonable density vs cost
Jun 19 14:46:36 <jwh> fast enough with DPDK etc
Jun 19 14:46:47 <jwh> just need to get it into my openwrt builds
Jun 19 14:47:22 <jwh> but first, I need to enable tcp md5
Jun 19 14:47:26 <jwh> isn't on by default it would appear
Jun 19 14:47:47 <stintel> it isn't, doesn't really make sense for the typical openwrt use case ;)
Jun 19 14:48:16 <jwh> I guess
Jun 19 14:49:21 <jwh> just need to uh, find it in the menu now :D
Jun 19 14:49:52 <stintel> CONFIG_TCP_MD5SIG
Jun 19 14:49:56 <jwh> <3
Jun 19 14:50:36 <jwh> currently running build is from a different incarnation of my tree too, probably have to force it and hope it doesn't break :D
Jun 19 14:51:29 <jwh> can uh, can I list /etc/config/* paths in sysupgrade.conf?
Jun 19 14:51:36 <blogic> Hauke: ping
Jun 19 14:51:39 <blogic> Hauke: c807a7737f1b034d4776312bd1dee12ce7860595
Jun 19 14:51:44 <blogic> how good is your memory ?
Jun 19 14:52:00 <jwh> coz last time I checked the bird packages unconditionally overwrote the config there
Jun 19 14:53:07 <stintel> jwh: sysupgrade -l -> check if they aren't there already
Jun 19 14:53:18 <stintel> looks like bird doesn't have /etc/config/bird* in conffiles
Jun 19 14:53:23 <jwh> theres err
Jun 19 14:53:26 <jwh> bird?-openwrt
Jun 19 14:53:27 <stintel> which would explain why it's being overwritten
Jun 19 14:53:47 <jwh> as that contains the init.d scripts, rather than the main package
Jun 19 14:53:52 <jwh> no idea if its been fixed since though
Jun 19 14:53:56 <stintel> ah
Jun 19 14:54:00 <stintel> I just use /etc/bird*.conf
Jun 19 14:54:04 <jwh> yeah me too
Jun 19 14:54:32 <jwh> at least, at the time, unless I installed -openwrt packages there was no /etc/init.d/bird4 etc
Jun 19 14:55:00 <jwh> probably nearly a year ago now though
Jun 19 14:55:13 <jwh> it looks like bird2 package contains init script though, so thats fine
Jun 19 14:55:55 <jwh> may as well move to that at the same time
Jun 19 14:56:30 <stintel> oh there is bird2 now?
Jun 19 14:56:35 <stintel> haven't been following recently
Jun 19 14:57:03 <jwh> yeah, been using it for ages on generic vms, but it looks like someone added it to the routing packages
Jun 19 14:57:17 <jwh> no seperate v4/v6 binaries now
Jun 19 14:57:33 <jwh> and some fancy stuff like vrf support
Jun 19 14:58:54 <stintel> interesting
Jun 19 15:01:13 <jwh> I don't know C or netifd/procd well enough to add VRF support though :(
Jun 19 15:01:16 <jwh> or uci, really
Jun 19 15:03:02 <owrt-snap-builde> build #33 of mediatek/mt7623 is complete: Success [build successful]  Build details are at http://phase1.builds.lede-project.org/builders/mediatek%2Fmt7623/builds/33
Jun 19 15:12:42 <ausjke> where to find wireless-config for 11ac (mediatek, qca,...), the openwrt doc is mostly just 11n
Jun 19 15:13:34 <jow> ausjke: its the same as 11n
Jun 19 15:13:48 <jow> ausjke: just with htmode VHTxx instead of htmode HTxx
Jun 19 15:14:14 <jow> band is still a (5ghz) or g (2.4ghz)
Jun 19 15:14:24 <jow> mode, encryption etc. is the same
Jun 19 15:14:57 <Borromini> ausjke: https://openwrt.org/docs/guide-user/network/wifi/basic
Jun 19 15:14:58 <ausjke> so hwmode will be 11a or 11g always(only 11a 11b 11g available), and htmode will be HT20, HT40-, HT40+, HT40, NONE or VHT20, VHT40, VHT80, VHT160?
Jun 19 15:15:07 <jow> ausjke: correct
Jun 19 15:15:14 <ausjke> i see, will use VHT then
Jun 19 15:15:50 <ausjke> i'm trying to benchmark zbt1326 here, 1m direct of sight, iperf to 512kb size, HT40 etc
Jun 19 15:16:08 <Borromini> jow: it seems this could be handy in 17.01.5 as well but hasn't been backported so far https://github.com/openwrt/openwrt/commit/2ea98fc39b21392c1556d6768604f54a578278d4
Jun 19 15:16:17 <Borromini> related forum topic: https://forum.lede-project.org/t/17-01-1-2-3-qca988x-ath10k-5ghz-firmware-crashed-zyxel-nbg6716/1911/50
Jun 19 15:21:36 <ausjke> is 802.1x ready for non-enterprise usage? seems like all OS and cellphone provide supplicant client these days and it is not hard to set up
Jun 19 15:21:59 <ausjke> i'm thinking to enable it for the house to experiment
Jun 19 15:35:17 <jwh> mmm
Jun 19 15:35:27 <jwh> does ipt-offload need conntrack?
Jun 19 15:35:45 <jwh> oh, hm
Jun 19 15:36:30 <Borromini> ausjke: seems ath10k has issues with it
Jun 19 15:36:42 <jwh> doesn't appear to, wonder what is selecting nf-conntrack
Jun 19 15:37:04 <jwh> oooh, offload selects nf-flow which selects conntrack
Jun 19 15:37:05 <jwh> nm
Jun 19 15:38:13 <jwh> what should I select, ipt or nft offload?
Jun 19 15:58:27 <tmn505> muhaha: yes
Jun 19 16:00:33 <ausjke> Borromini: good to know, thanks
Jun 19 16:07:53 <stintel> so if I want to port my ar71xx devices to ath79 ... where do I start
Jun 19 16:08:17 <stintel> Unifi AP AC Pro should be very similar to AC Lite
Jun 19 16:09:01 <blogic> stintel: figure out if the SoC dtsi is present
Jun 19 16:09:05 <Monkeh> You could start with the new DTS for the AC Pro and compare the existing mach files to see what changes
Jun 19 16:09:16 <blogic> then make an initramfs with flash / leds / buttons
Jun 19 16:09:19 <blogic> make it buildable
Jun 19 16:09:30 <ausjke> using lighttpd with luci, works fine on firefox, only https works under chrome, http will also show 403 after login, i.e. keep redirecting back to login page
Jun 19 16:09:31 <blogic> then try to bring up ethernet, pci/e, usb, wmac
Jun 19 16:09:37 <stintel> Monkeh: yeah, there's really not a lot of differences
Jun 19 16:09:41 <blogic> and tell me whats broken
Jun 19 16:10:07 <stintel> Monkeh: https://gist.github.com/537cbdafc16310f8a9bf0ff31f35eeee
Jun 19 16:10:22 <stintel> so I copied the ac-lite dts to ac-pro
Jun 19 16:11:23 <stintel> and changed the phy-mask for mdio0 from <4> to <0>
Jun 19 16:12:01 <stintel> should figure out how to boot an initramfs first, but I suspect this will require serial console
Jun 19 16:12:04 <ausjke> lighttpd+luci(http) will not login on chrome, https is fine, firefox fine both http/https
Jun 19 16:12:07 <ausjke> what the hack
Jun 19 16:12:45 <stintel> guess I should start with my DAP-2695 first
Jun 19 16:12:56 <stintel> it has a console port <3
Jun 19 16:17:49 <jwh> stintel: the unifi stuff usually have headers at least for ttl uart
Jun 19 16:18:15 <jwh> ubnt are kinda good like that, they leave pins on
Jun 19 16:19:11 <czakey> yeah, ubnt makes nice gear
Jun 19 16:19:55 <czakey> although i prefer their firmware on newest aps... for now ;)
Jun 19 16:20:40 <jwh> what on earth, build completes
Jun 19 16:20:48 <jwh> forget to enable vmdk images
Jun 19 16:20:50 <jwh> has a shitfit
Jun 19 16:21:05 <jwh> spews oldconfig at me, opkg fails to install
Jun 19 16:21:07 <jwh> packages
Jun 19 16:21:37 <jwh> 0 * satisfy_dependencies_for: Cannot satisfy the following dependencies for kmod-scsi-core:
Jun 19 16:21:40 <jwh>  *      kernel (= 4.14.48-1-5fdc9976c281cbee425c46a57472c039) *
Jun 19 16:21:40 <jwh>  * opkg_install_cmd: Cannot install package kmod-scsi-core.
Jun 19 16:21:41 <jwh> umm
Jun 19 16:24:45 <jwh> sigh, god damn merge conflicts for the sake of it
Jun 19 16:24:53 <jwh> there is no conflict, git
Jun 19 16:24:55 <jwh> wke up
Jun 19 16:24:56 <jwh> wake*
Jun 19 16:26:18 <jwh> I give up
Jun 19 16:26:27 <EvilRick> I need someone to add board support to ath79 for me.. the board is the same as the atheros DB120 reference design with some minor GPIO changes and additional 128 NAND on top of the 8M NOR. I looked at doing it myself but feel that its best left to someone with more experience. I am currently able to run the board using firmware release 17 ar71xx. I had to mangle the default DB120 gpio
Jun 19 16:26:27 <EvilRick> board definition though to get the GSM/CDMA modem to power up. There is some vendor documentation around it as well which will help even though its in chinglish.
Jun 19 16:26:37 <salcedo> "Looks like you're dealing with a merge conflict!" -- M$ Clippy
Jun 19 16:26:42 <jwh> lol
Jun 19 16:26:50 <jwh> <<<<<<< HEAD
Jun 19 16:26:50 <jwh> vpi=0
Jun 19 16:26:50 <jwh> vci=38
Jun 19 16:26:50 <jwh> =======
Jun 19 16:26:50 <jwh> vpi=1
Jun 19 16:26:50 <jwh> vci=32
Jun 19 16:26:51 <jwh> >>>>>>>
Jun 19 16:27:05 <jwh> that is it
Jun 19 16:28:13 <EvilRick> those look to be ADSL Frame relay paramaters. usually the end user has to define those as they vary from provider to provider. Maybe just leave them the same :)
Jun 19 16:28:20 <jwh> unsurprisingly, the patch its trying applies cleanly
Jun 19 16:28:24 <jwh> therefore, git is garbage
Jun 19 16:29:04 <endoedgar> Hello.I am currently running openwrt on Ubnt Unifi v2 with success, and I would like to pull request.
Jun 19 16:29:25 <karlp> file it on github, or send to the mailing list
Jun 19 16:29:36 <endoedgar> ok
Jun 19 16:30:05 <jwh> think I'll definitely just hand pick commits from now on
Jun 19 16:30:15 <jwh> rebasing is just a chore
Jun 19 16:30:38 <jwh> breaks ff, dumbass merge commits because it can't replay stuff in the proper order
Jun 19 16:30:45 <jwh> err, conflicts
Jun 19 16:33:26 <endoedgar> well, there is just one problem lol I've used a custom version of uboot to sucessfully write to the flash because ubnt needs that signed firmware thingy
Jun 19 16:35:31 <endoedgar> and the only tool i've found to correcly write the firmware is a ddwrt one: here: https://github.com/mirror/dd-wrt/commit/076f9aae746ac033370a26394ccd41bb3a505219
Jun 19 16:35:43 <jwh> sigh, new repo *again* it is then, won't even rebase between my own branches now either, despite it having fixed conflicts
Jun 19 16:36:09 <karlp> jwh: don't know what you're doing, things just work fine here.
Jun 19 16:36:20 <karlp> only fiddles are when kernel version changes and files move to new directories
Jun 19 16:36:21 <jwh> karlp: it consistently fucks it up
Jun 19 16:36:28 <jwh> so like
Jun 19 16:36:42 <karlp> it consistently works for me, I was _just_ doing a "git pull --rebase upstream openwrt-1806"
Jun 19 16:37:12 <blogic> karlp: it has worked like that for years now
Jun 19 16:37:13 <jwh> well, one of the (many) conflicts was changed to the vpi/vci in my local branch
Jun 19 16:37:18 <jwh> except
Jun 19 16:37:28 <jwh> the 2 lines it is complaining about apply cleanly
Jun 19 16:37:34 <jwh> yet git thinks they don't
Jun 19 16:38:10 <jwh> master still has the same content so there is no reason why it can't replay my commit ontop
Jun 19 16:39:45 <jwh> the others are pretty similar, its just not replaying in the correct order
Jun 19 16:39:56 <karlp> blogic: I wasn't complaining :)
Jun 19 16:40:15 <blogic> i was agreeing with you
Jun 19 16:40:36 <blogic> jwh: well then you are doing something fundamentally wrong
Jun 19 16:40:43 <jwh> obviously
Jun 19 16:40:47 <blogic> jwh: it simply works
Jun 19 16:40:52 <jwh> by following the rebase workflow and prompts it gives me D
Jun 19 16:40:53 <jwh> :D
Jun 19 16:42:02 <jwh> so, if I fix conflicts in a branch and then rebase it into another, it *should* also have the fixed conflicts
Jun 19 16:42:05 <jwh> but it doesn't
Jun 19 16:42:36 <jwh> since fixing conflicts with edit, git add, rebase --continue just complains about me forgetting to do add, maybe not though
Jun 19 16:43:12 <jwh> it all works fine until something diverges enough that it gets upset
Jun 19 16:43:19 <blogic> </cares>
Jun 19 16:48:05 <jwh> oh well heh
Jun 19 17:30:01 <hitech95> Hi guys is there here someone that have a bit of experience on alsa development? I would start as developer (I have to fix a driver) but i cant find good reference online. is there a book like linux device drivers but for alsa?
Jun 19 17:30:09 <jwh> well, I got ipt-netflow working
Jun 19 17:30:56 <jwh> it builds anyway
Jun 19 17:31:12 <Hauke> blogic: pong
Jun 19 17:31:51 <Hauke> Borromini: pong
Jun 19 17:33:28 <Hauke> blogic: https://patchwork.ozlabs.org/patch/927426/ is already applied
Jun 19 17:33:40 <Hauke> + an upgrade to 4.4.137
Jun 19 17:33:56 <Hauke> blogic: I haven't loked into the gemini build issue
Jun 19 17:34:52 <Hauke> wigyori: is the ethernet on the linksprite pcduino3 only broken in u-boot or also linux
Jun 19 17:35:17 <Hauke> wigyori: thanks for testing
Jun 19 17:36:05 <Hauke> blogic: what about c807a7737f1b034d4776312bd1dee12ce7860595 ?
Jun 19 17:39:36 <Borromini> Hauke: hi
Jun 19 17:39:37 <Borromini> .
Jun 19 17:40:19 <Borromini> i already asked jow about this in the meantime, but i don't think he replied (i went offline as well), so here goes
Jun 19 17:40:44 <Borromini> this might be useful for 17.01.5: https://github.com/openwrt/openwrt/commit/2ea98fc39b21392c1556d6768604f54a578278d4
Jun 19 17:40:56 <Borromini> related forum topic: https://forum.lede-project.org/t/17-01-1-2-3-qca988x-ath10k-5ghz-firmware-crashed-zyxel-nbg6716/1911/44
Jun 19 17:42:19 <Hauke> ok I will have a look
Jun 19 17:42:47 <ntd> looking at the archer ct v4
Jun 19 17:43:01 <ntd> they actually made the new rev a better product, owrt-wise?
Jun 19 17:43:22 <ntd> more flash, ext antennae are now both 2.4/5ghz?
Jun 19 17:43:50 <ntd> and an extra 50mhz
Jun 19 17:44:17 <ntd> no spam intended, sorry. *and* radios are AC wave 2?
Jun 19 17:44:28 <stintel> hmm, new symbol? THERMAL_EMERGENCY_POWEROFF_DELAY_MS
Jun 19 17:44:41 <ntd> must be getting rid of old stock in prep for wpa3
Jun 19 17:44:49 <Borromini> Hauke: ty.
Jun 19 17:45:19 <stintel> seems to be set in several 4.14 configs for targets
Jun 19 17:45:22 <Borromini> is WPA3 going to require new hardware then?
Jun 19 17:45:26 <Borromini> ntd: ^^
Jun 19 17:45:29 <stintel> we should probably move it to generic/config-4.14
Jun 19 17:46:00 <ntd> according to jow they're prolly gonna come up with arbitrary explanation/bs as to new hw reqs
Jun 19 17:46:15 <ntd> been a few months since he said that, though
Jun 19 17:46:17 <Hauke> I think some parts of WPA3 can be done in SW
Jun 19 17:46:19 <jwh> hm, https://gist.github.com/joeholden/b11f5de547b47e36ed5220304c5770e2
Jun 19 17:46:27 <jwh> I can't figure out what its trying to access
Jun 19 17:46:36 <blogic> Hauke: got a sec ?
Jun 19 17:46:41 <Hauke> yes
Jun 19 17:46:46 <jwh> optional feature, but would be nice given the default openwrt config
Jun 19 17:47:32 <Hauke> did someone look more closely into WPA3?
Jun 19 17:48:57 <ntd> been bugging jow about it, haven't found much in terms of actual documentation
Jun 19 17:49:53 <Borromini> is it already a final standard?
Jun 19 17:50:29 <ntd> asus and d-link are prepping for a q3 2018 launch afaik
Jun 19 17:50:50 <ntd> and that is consumer gear
Jun 19 17:52:01 <ntd> what i've come over thus far: 11w required, non-passphrase protected networks are also encrypted (...), something about a 192-bit sec suite
Jun 19 17:52:43 <ntd> and somehow smart lightbulbs and toilet plungers with no screen/input device will somehow be able to automagically be set up
Jun 19 17:53:18 <Borromini> smells like more WPS eh
Jun 19 17:53:31 <ntd> no wps
Jun 19 17:53:36 <stintel> ah, it's probably not that simple
Jun 19 17:56:17 <Hauke> they use better hash algorithms no MD5 or SHA1 any more
Jun 19 17:56:22 <Hauke> https://www.heise.de/ct/ausgabe/2018-12-WPA3-loest-die-loechrige-WLAN-Verschluesselung-WPA2-ab-4056749.html
Jun 19 17:56:30 <jwh> ntd: the 192bit is the one the NSA proposed, can't remember what its called now
Jun 19 17:56:47 <Hauke> "simultaneous authentication of equals"
Jun 19 17:57:13 <jwh> https://tools.ietf.org/id/draft-jenkins-cnsa-cert-crl-profile-00.html
Jun 19 17:57:25 <jwh> umm, no
Jun 19 17:59:00 <jwh> https://www.iad.gov/iad/programs/iad-initiatives/cnsa-suite.cfm
Jun 19 17:59:40 <Hauke> it looks like it is not clear how open access points should be secured
Jun 19 17:59:46 <Monkeh> It is highly amusing to me that I can't actually load that page due to a certificate error.
Jun 19 17:59:48 <Hauke> in wpa3
Jun 19 18:00:04 <jwh> Monkeh: not really an error, you just don't trust the DOD
Jun 19 18:00:04 <jwh> D
Jun 19 18:00:08 <jwh> :D
Jun 19 18:00:17 <Monkeh> Who does?
Jun 19 18:00:20 <jwh> heh
Jun 19 18:00:51 <Hauke> this is NSA
Jun 19 18:01:36 <Borromini> i'm sure trump has been developing his own ecryption variety for WPA3+ :)
Jun 19 18:01:48 <stintel> anyone care for a quick review? https://git.openwrt.org/5a31a627
Jun 19 18:01:50 <Hauke> simultaneous authentication of equals (SAE) is already supported by hostapd since version 2.1
Jun 19 18:02:09 <mamarley> Borromini: Knowing him, it would probably be rot13.  The strongest encryption!
Jun 19 18:02:15 <Borromini> :D
Jun 19 18:03:11 <stintel> you forgot rot26
Jun 19 18:03:59 <Monkeh> Nobody expects the rot26 encryption!
Jun 19 18:06:24 <jwh> hmmm, I suspect my build problem is musl related, wonder why though
Jun 19 18:08:11 <jwh> skbuff.h is just a mess anyway
Jun 19 18:09:25 <stintel> hmmmz, unknown LG Electronics device on my network
Jun 19 18:13:00 <ultito> morning
Jun 19 18:13:28 <gzc__> hah, the gl-ar150 uses a full watt less power than the ar750. sure, it's only got a 2.4ghz radio, but dang! pretty decent for ~.6W
Jun 19 18:15:55 <stintel> oh, nexus 5x is lg, nevermind :P
Jun 19 18:17:22 <Borromini> lol
Jun 19 18:17:58 <stintel> Borromini: well someone gave it to me, if I had bought it myself I would have known it was LG :)
Jun 19 18:19:01 <Borromini> fair enough. nexus branding isn't always too prominent ;)
Jun 19 18:19:38 <jwh> aw, no feeds mirror on github
Jun 19 18:20:21 <jwh> oh just packages
Jun 19 18:34:54 <luaraneda> I read somewhere that WPA3 open networks will use Opportunistic Wireless Encryption ( https://tools.ietf.org/html/rfc8110 ) for encryption
Jun 19 18:39:21 <jwh> ok now I'm confused
Jun 19 18:39:35 <jwh> it doesn't build with glibc toolchain either, but builds fine natively
Jun 19 18:40:56 <ca4ti_> Hi i try compile lirc for use on raspberry py 3 (lede 17.01.4) , but i not understain i need to compile one app (in this case lirc)
Jun 19 18:41:04 <ca4ti_> thanks atention.
Jun 19 18:42:20 <stintel> ca4ti_: unfortunately it's not that simple, especially with lirc
Jun 19 18:42:33 <stintel> ca4ti_: you're going to have to add kernel modules, and revive the lirc package from the oldpackages feed
Jun 19 18:43:17 <ca4ti_> thanks your help , i try found how to , but i found for old versions ...
Jun 19 18:43:36 <stintel> yes, lirc is currently not available in openwrt
Jun 19 18:44:04 <ca4ti_> if is possible compile , i have time to try :) , but i need more details to build :( ...
Jun 19 18:44:56 <stintel> I started on it at some point but it's nowhere near finished
Jun 19 18:45:13 <ca4ti_> I mount this fisical project : http://www.raspberry-pi-geek.com/Archive/2015/10/Raspberry-Pi-IR-remote
Jun 19 18:45:45 <ca4ti_> but in software ... (in 17.01.4) more thinks is missing
Jun 19 18:47:29 <ca4ti_> stintel , if i open one "post" on forum lede , is possible i have , a minimun help ?
Jun 19 18:49:11 <Monkeh> You should perhaps use a more genreal purpose distro.
Jun 19 18:49:15 <Monkeh> general, bah, my typing.
Jun 19 18:49:58 <jwh> :D
Jun 19 18:50:04 <ca4ti_> i like minimum specs of openwrt :)
Jun 19 18:50:17 <Monkeh> Well, it's coming back to bite you right now
Jun 19 18:50:18 <Monkeh> ;)
Jun 19 18:50:20 <jwh> so minimum its not suitable for your use
Jun 19 18:50:20 * ldir thinks monkey is spelled with a 'y'
Jun 19 18:51:12 <Monkeh> ldir: See? I can't even type my name right.
Jun 19 18:51:30 <stintel> here's my lirc WIP: https://git.openwrt.org/?p=openwrt/staging/stintel.git;a=shortlog;h=refs/heads/lirc https://github.com/stintel/lede-wip/tree/lirc
Jun 19 18:51:46 <stintel> but as I said it's nowhere near finished
Jun 19 18:51:51 * ldir has a spontaneously powering up qnap nas box.  No discs, no network, no lan..... just decides to switch itself on. WTF!
Jun 19 18:52:06 <stintel> if you want a quick solution you should probably do as Monkeh suggested
Jun 19 18:52:34 <Monkeh> ldir: I do have a fix for that
Jun 19 18:52:53 <stintel> that said, I really want to get lirc_rpi working on openwrt and hope that I can use it to control my ACs in the apartment
Jun 19 18:53:10 <stintel> maybe I should work on that instead of messing with ath79 and dts
Jun 19 18:53:17 * Monkeh throws some ESPs at stintel
Jun 19 18:53:26 * ldir suspects it involves a hammer.
Jun 19 18:53:36 <Monkeh> ldir: I could use the mallet instead?
Jun 19 18:53:44 <Monkeh> Or if you like I can just shoot it.
Jun 19 18:53:51 <stintel> Monkeh: did a PoC with an ESP to control my soundbar at my parents' place
Jun 19 18:53:51 <ca4ti_> No i need more knowledge , one , i need to learn how to compile applications for openwrt and I take advantage of doing this infrared project :)
Jun 19 18:54:09 <stintel> ca4ti_: you can't just compile applications for openwrt
Jun 19 18:54:25 <stintel> https://openwrt.org/docs/guide-developer/packages
Jun 19 18:54:29 <stintel> is probably a good start
Jun 19 18:54:58 <stintel> if you start going at it, have a look at the 2 links I just posted
Jun 19 18:55:02 <stintel> earlier
Jun 19 18:55:13 <stintel> so you don't start from scratch but continue on my work
Jun 19 18:55:30 <philipp64> new PR (as a couple of virtual packages, instead of modifying base-files and openssh) to (a) install a default root password, and (b) turn off password authentication in openssh-server... anyone want to review it?
Jun 19 18:55:31 <ca4ti_> thanks , i try but is need crosscompile ? is not possívle compile inside raspberry over openwrt (or chroot debian on openwrt) ?
Jun 19 18:56:45 <philipp64> drmr: what's the question to which "Panama" is the answer?
Jun 19 18:57:18 <ca4ti_> stintel , big thanks , in this weekend , i try compile source of lirc you indicate , thanks again
Jun 19 18:57:27 <philipp64> "Who makes the best hot sauce in the world?" maybe?
Jun 19 18:58:44 <jwh> weeeeeeeee, ubnt finally released 4.9 firmware
Jun 19 18:59:25 <ultito> what does it mean?
Jun 19 18:59:59 <ausjke> worked at ubnt for a while, not that exciting for me, a fine company
Jun 19 19:00:00 <jwh> not much really, but it means I can use the switch ports on some of the edgerouter models
Jun 19 19:00:05 <drmr> wat?
Jun 19 19:00:08 <drmr> who?
Jun 19 19:00:11 <drmr> where?
Jun 19 19:00:21 <jwh> (with openwrt, that is)
Jun 19 19:00:21 <ausjke> drmr: ?
Jun 19 19:00:23 <ca4ti_> stintel lol you push to repository your pie code of lirc ! youre nice !
Jun 19 19:00:28 <Monkeh> why?
Jun 19 19:00:31 <Monkeh> when?
Jun 19 19:00:34 <jwh> previous 3.10 kernel was way too much hassle to patch to make work with f2fs etc
Jun 19 19:00:36 <ausjke> what?
Jun 19 19:00:51 <drmr> oh. panama. yeah. that joke went by a long time ago. it fathered three children and died of old age.
Jun 19 19:00:51 <stintel> ca4ti_: I am even testing it atm to see what still needs to be done
Jun 19 19:01:06 <stintel> drmr: in a day? nice
Jun 19 19:01:21 <drmr> stintel: jokes age quite rapidly.
Jun 19 19:02:22 <ca4ti_> is you need , i try test for you, i have components of article a send in up , and one raspberry pi 3 (and banana pi m2+ zero )
Jun 19 19:02:23 <philipp64> ausjke: I interviewed there and they lowballed me on a salary offer of 60% my current salary.
Jun 19 19:03:19 <stintel> ca4ti_: well the first part is getting the lirc software and kernel modules in order
Jun 19 19:03:22 <philipp64> this was after flying me to Austin and having me talk to the CEO...
Jun 19 19:03:41 <stintel> ca4ti_: but I do have 3 rpi0w here, and a few IR emitters
Jun 19 19:03:46 <ausjke> philipp64: don't think its CEO is at Austin though
Jun 19 19:04:32 <ca4ti_> stintel , i have one ... you have 3 and more .... : ( , rsrsrsrs
Jun 19 19:04:33 <ausjke> philipp64: for me it's more of a rack of devices that I once a while need get close to debug kernel, kills me
Jun 19 19:04:40 <philipp64> no, it was to meet their architect.
Jun 19 19:04:56 <ca4ti_> stintel : my simple project (and one another channel to call me : http://forum.banana-pi.org/t/frist-use-banana-pi-m2-zero-using-on-lapdock-atrix/5067/4)
Jun 19 19:05:50 <ausjke> i mean 37 devices, some of them are wireless, some of them are so noisy, and i had to stay close to that rack once a while, home office, I can not sustain that working env
Jun 19 19:06:09 <ausjke> philipp64: i know who you talked with them, starting with c
Jun 19 19:06:35 <ausjke> s/them/then/
Jun 19 19:07:23 <ausjke> can't type, the typo also render the meaning totally different, always typo at the worst letter
Jun 19 19:10:40 <jwh> https://gist.github.com/joeholden/1c4a0a07975123c15f792281cd83f0ad
Jun 19 19:10:47 <jwh> thoughts? comments? (build and runtime tested)
Jun 19 19:11:33 <jwh> I have no idea how to create a proper package, so thats the upstream version with added things to unbreak it
Jun 19 19:13:01 <philipp64> ausjke: yeah
Jun 19 19:13:23 <ausjke> jwh: compare to collectd/etc what does netflow buy us? more granuality in traffic logging?
Jun 19 19:13:52 <jwh> an actual standardised method of flow analysis :P
Jun 19 19:14:11 <jwh> isn't collectd just arbitary metric ingest though?
Jun 19 19:15:38 <ausjke> it collects everything via a plugin, including wired/wireless traffic and collections
Jun 19 19:15:41 <jwh> the one thing that doesn't work in that one I just posted is the one thing that may actually be kinda useful
Jun 19 19:15:54 <ausjke> i heard netflow a lot but never used it, any open source collector for netflow reports?
Jun 19 19:15:59 <jwh> loads
Jun 19 19:16:15 <jwh> nfdump/nfsen is one of the more traditional ones
Jun 19 19:16:27 <jwh> but theres collectors for all the hipster crap
Jun 19 19:17:05 <jwh> ipt-netflow uses a netfilter target though and its done in kernel
Jun 19 19:17:15 <jwh> the only options currently are pcap based in userland
Jun 19 19:17:16 <ausjke> for me collectd or its cousins are more than enough, netflow is more like for 'higher-end-net-gears', but i actually have no idea about it
Jun 19 19:17:20 <jwh> which is fine for a handful of pps
Jun 19 19:18:14 <jwh> but also, don't really want to have to run an agent on the device when it can be done as part of conntrack
Jun 19 19:19:13 <ausjke> you never know, conntrack could be memory hungry, while an agent might be very light-weight instead
Jun 19 19:19:34 <jwh> if you're running openwrt you're already using conntrack anyway
Jun 19 19:19:35 <jwh> :D
Jun 19 19:19:51 <ausjke> jwh: why so? I did not
Jun 19 19:20:03 <jwh> the default config has it
Jun 19 19:20:12 <jwh> and its loaded even without nat
Jun 19 19:20:27 <stintel> yes because stateful filter
Jun 19 19:20:34 <jwh> yup
Jun 19 19:21:10 <jwh> the only way to do useful flow export is either pcap, or something in-kernel that has access to netdevs etc
Jun 19 19:21:17 <ausjke> so you're saying it's by default on, if i have lots of connections it can eat up my memory without me knowing it...
Jun 19 19:21:18 <jwh> pcap is really heavy though
Jun 19 19:21:30 <jwh> ausjke: you said you worked for ubnt, surely you know all this? :D
Jun 19 19:21:54 <ausjke> jwh: not really, more of a bring-up job there
Jun 19 19:21:59 <jwh> ah
Jun 19 19:22:36 <ausjke> u-boot, ddr, some kernel and debian user land, not much network actually
Jun 19 19:22:44 <jwh> heh
Jun 19 19:23:02 <jwh> is it your fault the erlite was plagued for so long then? :D
Jun 19 19:23:19 <ausjke> no i was doing their enterprise line
Jun 19 19:23:23 <jwh> I think out of the original 8 I bought when they were first available
Jun 19 19:23:28 <jwh> 7 had to be replaced within 3 months
Jun 19 19:23:52 <jwh> ah
Jun 19 19:24:04 <ausjke> the UniFi!
Jun 19 19:24:13 <jwh> "enterprise"
Jun 19 19:24:15 <jwh> :P
Jun 19 19:24:19 <ausjke> :)
Jun 19 19:24:48 <jwh> anyway, netflow for everyone
Jun 19 19:25:18 <jwh> you should try running userland bpf type stuff when doing many gigabits of traffic :D
Jun 19 19:26:01 <jwh> I wonder how it will work with offload though, can you chain targets?
Jun 19 19:26:09 <ausjke> which, not something openwrt is designed for...
Jun 19 19:26:17 <jwh> why not?
Jun 19 19:26:28 <jwh> it runs on boxes capable of it
Jun 19 19:26:50 <ausjke> just listened to a seminar by a facebook kernel guy oon epbf and xdp
Jun 19 19:27:06 <jwh> they'd go even faster if not having to waste cpu time on packet captures
Jun 19 19:27:07 <jwh> heh
Jun 19 19:27:20 <ausjke> and was told you can only do 4000 instructions in pbf, and there is absolutely no loop in the instruction
Jun 19 19:27:35 <jwh> ebpf is kinda interesting though
Jun 19 19:27:59 <jwh> better late than never I guess
Jun 19 19:28:30 <ausjke> xdp+bpf is basically trying to replace dpdk
Jun 19 19:29:11 <jwh> meh, DPDK has far more utility
Jun 19 19:29:19 <ausjke> it's said the bpfilter(?) is going to replace iptables and nftables, before nftables take afoot
Jun 19 19:29:40 <jwh> I suspect we're talking about a different bpf heh
Jun 19 19:29:59 <ausjke> no
Jun 19 19:30:02 <jwh> I meant bpf in the sense of berkeley packet filters (a la tcpdump), no idea what they are on linux these days
Jun 19 19:30:09 <jwh> ebpf is something entirely differet
Jun 19 19:30:12 <jwh> different*
Jun 19 19:30:28 <tripolar> hi
Jun 19 19:30:44 <tripolar> whats the offical way to get a patch for librpc into the main openwrt repo?
Jun 19 19:30:54 <tripolar> creating a pull request on github won't make it i gues
Jun 19 19:31:04 <ausjke> ebpf that is
Jun 19 19:31:43 <ausjke> anyways
Jun 19 19:32:07 <jwh> xdp has limited uses tbh, you can do some simple stuff within the constraints of ebpf
Jun 19 19:32:25 <jwh> but for a lot of the supposed use cases (packet filtering, routing etc)
Jun 19 19:32:38 <jwh> you can do them on nic cpus
Jun 19 19:33:17 <jwh> or if the soc has packet i/o abilities and can do arbitary filtering, there instead
Jun 19 19:34:12 <jwh> theres no support for it in openwrt at least, yet of course
Jun 19 19:34:15 <jwh> but its doable
Jun 19 19:34:45 <ausjke> for conntrackd, there are kmod for that, so iptables can use it to set up rules, what's the user space conntrackd for then? i thought I need that userspace daemon to interact with
Jun 19 19:35:04 <jwh> theres ipt conntrack
Jun 19 19:35:08 <jwh> which is the netfilter stuff
Jun 19 19:35:14 <jwh> conntrackd is just for interacting with it
Jun 19 19:35:30 <jwh> and pretty printing of things already available
Jun 19 19:35:45 <jwh> most of it lives in /proc anyway
Jun 19 19:35:50 <ausjke> i did not install the conntracd-utils, but then just realized I can still do iptables --state NEW for stateful filtering
Jun 19 19:37:37 <ausjke> i c, can conntrackd set iptables? or just monitoring/report? checking...never used conntrackd but did use -m state with iptables
Jun 19 19:37:54 <jwh> its just reporting iirc
Jun 19 19:38:00 <jwh> may have the ability to flush tables and stuff
Jun 19 19:38:02 <ausjke> not sure how much memory they consume, say 1000 connections, 10000 connections, how bad will it be
Jun 19 19:38:05 <jwh> but you can do all that without it
Jun 19 19:38:18 <stintel> it can sync conntrack table between multiple devices
Jun 19 19:38:23 <jwh> oh, that too yeah
Jun 19 19:38:42 <jwh> you'd think there would be something built in by now
Jun 19 19:40:20 <ausjke> stintel: for HA failover?
Jun 19 19:40:29 <ausjke> interesting to know
Jun 19 19:40:32 <stintel> ausjke: yes
Jun 19 19:40:38 <jwh> on a sort of related note, wonder if I can enable flow director at module compile time
Jun 19 19:40:42 <jwh> rather than mangling cmdline
Jun 19 19:40:56 <jwh> for uh, ixgbe
Jun 19 19:41:46 <stintel> ausjke: I should still write my part 3 blog post for "Building a redundant router setup with Open Source Software"
Jun 19 19:41:51 <stintel> which is about conntrackd
Jun 19 19:42:30 <stintel> but since I'm on a fulltime contract again I don't really find a lot of motivation to do anything outside of day job
Jun 19 19:42:36 <jwh> heh
Jun 19 19:43:43 <jwh> so... nobody has any suggestions on how to make that makefile less terrible? :(
Jun 19 19:43:49 <jwh> coz it certainly seems pretty terrible
Jun 19 19:47:45 <ausjke> stintel: where is part 1,2 to read?
Jun 19 19:47:59 <ausjke> and...please find some time to write part3 :)
Jun 19 19:48:28 <ausjke> even a brief one with details helps, just googled and find this: https://www.linuxjournal.com/article/10964
Jun 19 19:48:44 <stintel> ausjke: https://stijn.tintel.eu/blog/2017/09/20/building-redundant-router-setup-open-source-software-part-1
Jun 19 19:48:44 <ausjke> s/with/without/  typo swapped the meaning again
Jun 19 19:49:57 <stintel> building lirc 0.10.0
Jun 19 19:51:15 <stintel> apparently lirc depends on python now
Jun 19 19:53:06 <stintel> and no configure flag to override it
Jun 19 19:54:13 <ausjke> maybe just a few lines in code , why lirc mandates python without a turnoff flag
Jun 19 19:55:42 <stintel>   python3-yaml is missing. lirc-setup will not work, table.html
Jun 19 19:55:43 <stintel>   and lirc.hwdb will not be created.
Jun 19 19:56:12 <stintel> just going to add the dependency for now
Jun 19 19:58:20 <stintel> https://sourceforge.net/p/lirc/mailman/message/36015796/
Jun 19 19:59:21 <jwh> mmm, anyone have a link to a netfilter behaviour doc or something?
Jun 19 19:59:59 <jwh> I can't figure out what happens if I add two rules with different targets... will it just math the first, or will the first one return the packet and will netfilter hand it off to the second?
Jun 19 20:00:03 <jwh> etc
Jun 19 20:00:20 <jwh> so like, as an example I might have -A FORWARD -j OFFLOAD, then -A FORWARD -j NETFLOW
Jun 19 20:00:24 <jwh> is that going to work?
Jun 19 20:02:50 <ausjke> maybe termating target takes precedence?
Jun 19 20:03:10 <jwh> well, neither of those should swallow the pacet
Jun 19 20:03:12 <jwh> packet
Jun 19 20:03:21 <jwh> but the question is, do they return it to be processed further
Jun 19 20:04:07 <ausjke> terminating target means a final decision will be made by the target, e.g. DrOP, reject, forward, while log etc are not terminating targets, just an add-on side action
Jun 19 20:04:12 <jwh> yeah
Jun 19 20:04:35 <jwh> but generally a target can take any action it wants
Jun 19 20:04:38 <jwh> swallow, return, whatever
Jun 19 20:04:48 <ausjke> so netflow is not a terminating, in that case i would assume it will work like 'log'?
Jun 19 20:04:59 <jwh> I'd expect so yeah
Jun 19 20:05:11 <jwh> thats what (appears) to be happening
Jun 19 20:05:17 <jwh> I'm not sure how offload behaves though
Jun 19 20:05:34 <ausjke> hope to have sometime play with netflow here
Jun 19 20:05:36 <jwh> whether it skips any further processing
Jun 19 20:05:52 <jwh> since the point of offload is to be fast
Jun 19 20:06:01 <jwh> maybe I should just pass through netflow first
Jun 19 20:09:14 <jwh> tinkering for another day anyway, got other stuff to do
Jun 19 20:09:29 <jwh> now it builds I can work out how to plug it into netifd
Jun 19 20:09:40 <jwh> uh
Jun 19 20:09:42 <jwh> fw3 even
Jun 19 20:39:59 <stintel> bin/packages/arm_arm1176jzf-s_vfp/stintel_wip/lirc_0.10.0-1_arm_arm1176jzf-s_vfp.ipk
Jun 19 20:40:01 <stintel> \o/
Jun 19 20:44:31 <Monkeh> RFC from anyone interested in tplink-safeloader devices before I send to the ML: https://gist.github.com/eamaclean/80137539868f80c63b0e129e7ad18740
Jun 19 21:20:52 <jwh> so, firewall3 wizards? :D
Jun 19 21:21:29 <stintel> jwh: probably ask again during European office hours
Jun 19 21:21:41 <stintel> or mailing-list
Jun 19 21:21:41 <jwh> heh
Jun 19 21:21:55 <jwh> its C related (want to add support for a new target)
Jun 19 21:23:52 <jwh> guess I could mostly copy/paste the offload stuff
Jun 19 21:24:31 <stintel> or you could look at a specific commit that added a new target
Jun 19 21:24:35 <stintel> git log / git show are your friend
Jun 19 21:24:39 <jwh> yeah
Jun 19 21:24:40 <stintel> ooooh
Jun 19 21:24:46 <jwh> offload commit seems like the most sensible
Jun 19 21:25:13 <stintel> built an rpi image with lirc included o_O
Jun 19 21:25:19 <jwh> o
Jun 19 21:25:20 <jwh> o
Jun 19 21:25:21 <jwh> ffs
Jun 19 21:25:23 <jwh> o/
Jun 19 21:25:27 <stintel> curious to see how hard it will break :D
Jun 19 21:26:09 <jwh> I need to see if I can get ubnts 4.9 kernel booting
Jun 19 21:26:45 <jwh> should be able to build stock openwrt against it since its reasonably new
Jun 19 21:27:00 <jwh> gave up trying to backport stuff to 3.10 to make it work
Jun 19 21:28:48 <jwh> sucks that they won't publish their code for the switch
Jun 19 21:28:59 <jwh> even though its already supported by open source projects
Jun 19 21:29:15 <jwh> however they won't publish how its connected to the soc, so no idea how to attach
Jun 19 21:29:22 <jwh> all hidden in their kernel module
Jun 19 21:29:31 <stintel> ugh ubnt ... nice hw but asshole GPL violators
Jun 19 21:29:36 <jwh> heh
Jun 19 21:29:42 <jwh> well they do publish what they can tbf
Jun 19 21:29:48 <stintel> no, they don't
Jun 19 21:29:52 <jwh> qualcomm probably got them on the switch
Jun 19 21:30:14 <jwh> well, they publish the bits that they have to for GPL (userland and kernel sources)
Jun 19 21:30:16 <stintel> they told me to request GPL tarbal for US-8 on the "community" site
Jun 19 21:30:32 <jwh> but thats why they keep the useful stuff in out of tree moduels
Jun 19 21:30:36 <jwh> modules
Jun 19 21:30:40 <jwh> mmm, isn't the US-8 broadcom?
Jun 19 21:30:42 <stintel> but to get access to that I have to agree with a truckload of bullshit
Jun 19 21:30:59 <jwh> no chance of any code there heh
Jun 19 21:31:03 <stintel> and I'm pretty sure that saying "i agree" there includes "I will not request source code"
Jun 19 21:31:20 <jwh> yeah its err
Jun 19 21:31:21 <jwh> ranger?
Jun 19 21:31:22 <jwh> or something
Jun 19 21:31:22 <stintel> so I found the email of guy I was supposed to contact in some github repo and mailed him
Jun 19 21:31:29 <jwh> the cheap trident knock off
Jun 19 21:31:34 <stintel> he told me "don't share this URL"
Jun 19 21:32:07 <stintel> "Please do not share this URL. We release these on an as-needed basis."
Jun 19 21:32:11 <jwh> heh
Jun 19 21:32:15 <jwh> interesting
Jun 19 21:32:21 <stintel> stuff like that really grinds my gears
Jun 19 21:32:31 <jwh> wonder if there is some clause for non-public distribution in their contract with the vendor
Jun 19 21:32:37 <Monkeh> They're allowed to do that, but it's just silly
Jun 19 21:32:41 <jwh> you could of course, just publish it
Jun 19 21:32:42 <jwh> :D
Jun 19 21:32:44 <Monkeh> And once you have it they have no basis to stop you sharing it
Jun 19 21:32:53 <stintel> and then the same guy wrote on his introductionary post on the ubnt community site "i'm an open source enthousiast"
Jun 19 21:33:05 <stintel> no, you're not
Jun 19 21:33:17 <Monkeh> stintel: He's very enthusiastic about using open source code!
Jun 19 21:33:25 <jwh> having seen the sorts of NDAs signed previously for code to support hardware we purchased outright
Jun 19 21:33:27 <Monkeh> Not so much about the obligations.
Jun 19 21:33:31 <jwh> I wouldn't be surprised about anything
Jun 19 21:33:34 <stintel> if you were you wouldn't agree with those kind of practices, you would insist on publishing your work based on open source stuff instead
Jun 19 21:33:50 <jwh> stintel: yeah but... job/sued into oblivion
Jun 19 21:34:10 <stintel> bah
Jun 19 21:34:16 <stintel> rpi not coming back
Jun 19 21:34:21 <jwh> :(
Jun 19 21:34:27 <Monkeh> stintel: Did you threaten it with a camera?
Jun 19 21:34:27 <stintel> From 192.168.54.10 icmp_seq=384 Destination Host Unreachable
Jun 19 21:34:28 <Monkeh> :)
Jun 19 21:34:31 <stintel> Monkeh: 😂
Jun 19 21:35:14 <jwh> so actually, this code I need to add
Jun 19 21:35:17 <jwh> is quite literally
Jun 19 21:35:22 <Monkeh> I had a.. very fun experience trying to save one remotely with a bad SD in it.
Jun 19 21:35:28 <jwh> git diff | sed 's/flow/netflow/g'
Jun 19 21:35:29 <jwh> :D
Jun 19 21:36:00 <stintel> jwh: for "git sed ..." add this in .gitconfig: [alias]
Jun 19 21:36:01 <stintel>         sed = ! git grep -z --full-name -l '.' | xargs -0 sed -i -e
Jun 19 21:36:11 <stintel> very handy
Jun 19 21:36:22 <stintel> Monkeh: SD cards suck
Jun 19 21:36:29 <Monkeh> They really do.
Jun 19 21:36:39 <jwh> stintel: heh nice
Jun 19 21:36:52 <stintel> an rpi0w with eMMC would get met excited
Jun 19 21:36:57 <stintel> me*
Jun 19 21:37:08 <Monkeh> Until you realise they picked the cheapest eMMC so it's just as bad as SD cards.
Jun 19 21:37:24 <stintel> :D
Jun 19 21:37:29 <Monkeh> These are the same people who won't put a little underfill on their boards to stop cameras resetting them.
Jun 19 21:37:46 <stintel> well I've never been a fan of rpi, so I decided to try odroid c2
Jun 19 21:37:51 <stintel> it smoked itself
Jun 19 21:37:55 <Monkeh> Hah
Jun 19 21:38:03 <stintel> at least my rpi0w don't appear to do that
Jun 19 21:38:09 <Monkeh> I've got a Banana Pi.. somethingoranother
Jun 19 21:38:10 <Monkeh> It's okay.
Jun 19 21:38:19 <stintel> icmp_seq=610
Jun 19 21:38:24 <stintel> safe to say it's not coming back I guess
Jun 19 21:38:36 <Monkeh> Only half functional because you either pick an old kernel with broken proprietary crap or a patched mainline one with partial support
Jun 19 21:39:00 <Monkeh> Sorry, Orange Pi. Lite.
Jun 19 21:39:19 <karlp> my odroidc2 is fine, but it doesn't do much heavy lifting.
Jun 19 21:39:19 <stintel> well my replacement odroid c2 is in Belgium with my parents
Jun 19 21:39:20 <Monkeh> But it's made it to a month and a half without dying.
Jun 19 21:39:29 <karlp> runs a nextcloud instance and.... that's about it.
Jun 19 21:39:34 <stintel> so soon I should be able to finish the new meson target for openwrt
Jun 19 21:40:00 <stintel> https://git.openwrt.org/?p=openwrt/staging/stintel.git;a=shortlog;h=refs/heads/meson
Jun 19 21:40:15 <jwh> wtb odroid c2 with >1 nic
Jun 19 21:40:20 <Monkeh> I particularly love the experience of booting it: https://gist.github.com/eamaclean/b75c5068a48f8e066ae671ec90b56830
Jun 19 21:40:34 <Monkeh> Just scroll down a bit
Jun 19 21:40:35 <stintel> fffffffffuuuuuuu
Jun 19 21:40:36 <Monkeh> .. and a bit more
Jun 19 21:40:37 <Monkeh> ... and some more
Jun 19 21:40:42 <jwh> Monkeh: you there, you test my things
Jun 19 21:40:43 <jwh> thx
Jun 19 21:40:45 <jwh> :D
Jun 19 21:40:50 <Monkeh> jwh: I no test nothing.
Jun 19 21:40:53 <Monkeh> All tested out.
Jun 19 21:40:58 <stintel> ah so there's lirc stuff as well
Jun 19 21:41:13 <stintel> so the lirc work I'm currently doing might benefit more users / boards
Jun 19 21:41:28 <stintel> ugh
Jun 19 21:41:38 * stintel goes to grab the dead rpi0w from the bedroom
Jun 19 21:41:53 <Monkeh> Anyone want to take a guess what or who is spewing that stuff on my OPi? :)
Jun 19 21:42:05 <jwh> Monkeh: bleh
Jun 19 21:42:09 <jwh> Monkeh: but is nice :(
Jun 19 21:42:16 <Monkeh> jwh: Vhat?
Jun 19 21:42:35 <jwh> netflow :D
Jun 19 21:42:54 <Monkeh> I don't need netflow. :P
Jun 19 21:43:03 <jwh> bah
Jun 19 21:43:04 <jwh> consumers!
Jun 19 21:43:19 <stintel> go flow yourself! :D
Jun 19 21:43:22 <karlp> Monkeh: usb looks pretty borked too
Jun 19 21:43:25 <jwh> lol
Jun 19 21:43:38 <stintel> [   19.122512] brcmfmac: brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50
Jun 19 21:43:39 <jwh> I can't be the only one actually using openwrt in a non-residential setting
Jun 19 21:43:50 <Monkeh> karlp: What's b0rked about the USB?
Jun 19 21:43:56 <stintel> rmilecki: do you happen to know anything about that ^
Jun 19 21:44:01 <karlp> jwh: you seem to be the only one failing to use git, so anythigns possible ;)
Jun 19 21:44:09 <jwh> karlp: pssssht
Jun 19 21:44:17 <jwh> karlp: I'm pretty sure gitlab is partly to blame also
Jun 19 21:44:28 <stintel> so the rpi boots fine, even tries to load lirc module
Jun 19 21:44:31 <jwh> as I don't see it on locally kept repos
Jun 19 21:44:31 <stintel> but the wifi is broken
Jun 19 21:44:46 <karlp> Monkeh: were you just pulling out the usb-serial?
Jun 19 21:44:52 <rmilecki> stintel: unfortunately I don't :(
Jun 19 21:44:56 <Monkeh> karlp: Powering the other end off, but yes.
Jun 19 21:45:05 <rmilecki> my only suggestion is to write e-mail to brcm80211 ML + linux-wireless
Jun 19 21:45:13 <Monkeh> karlp: It has a massive frickin' relay controlled by GPIO which kills the power to that whole device.
Jun 19 21:45:39 <karlp> to what device?
Jun 19 21:45:54 <Monkeh> The one that CH340 thing is in
Jun 19 21:46:20 <stintel> rmilecki: seems firmware related, and that was recently updated, no ?
Jun 19 21:46:26 <karlp> you have a usb serial connector that is self powered (from the far side) not bus powered?
Jun 19 21:46:28 <karlp> cute.
Jun 19 21:46:32 <Monkeh> Indeed
Jun 19 21:46:40 <karlp> (why?)
Jun 19 21:46:41 <Tapper> what devices are on the meson target?
Jun 19 21:46:46 <Monkeh> I wish it were bus powered, it'd work around some software stuff I don't want to fix.
Jun 19 21:46:51 <Monkeh> They.. built it that way
Jun 19 21:46:52 <karlp> heh, ok.
Jun 19 21:47:37 <karlp> were they trying to copy carambola dev board style and just didn't understand it?
Jun 19 21:47:58 <Monkeh> *shrug*
Jun 19 21:48:00 <karlp> (bus powered usb serial connected to the console pins of the soc, so it's not actually a usb device on the board at all)
Jun 19 21:48:04 <Monkeh> Chinese 3D printer board.
Jun 19 21:48:06 <Monkeh> Go figure.
Jun 19 21:48:33 * karlp decides to try and diaganose this rt2860 wifi again.
Jun 19 21:48:34 <karlp> one more try
Jun 19 21:48:43 <stintel> sigh, U2410 SD card reader appears to be dead
Jun 19 21:48:49 <stintel> great
Jun 19 21:48:53 <jwh> Tapper: amlogic
Jun 19 21:49:11 <Tapper> k
Jun 19 21:49:21 <jwh> odroid, few other vendors
Jun 19 21:49:39 <Monkeh> karlp: Sorry, may not have been clear - the USB device is a 3D printer this thing is bolted to the side of
Jun 19 21:49:39 <karlp> lot of set top boxes
Jun 19 21:49:43 <jwh> yeah
Jun 19 21:50:50 <stintel> grmbl
Jun 19 21:51:02 <stintel> damn sd cards
Jun 19 21:51:06 * Monkeh looks at tub full of SD cards and readers
Jun 19 21:51:16 <Monkeh> I think you're outside throwing range..
Jun 19 21:53:47 <stintel> and the bad thing is I don't have a hammer here
Jun 19 21:54:23 <stintel> so I can't fix it, and soon I will try the same SD again and experience the same frustrating breakage
Jun 19 21:55:00 <Monkeh> Flush it down the toilet?
Jun 19 21:55:21 <stintel> yeah was considering the garbage container in the street
Jun 19 21:55:26 <stintel> toilet would do
Jun 19 22:01:35 <stintel> oh, my first reply from ubnt about source for us-8 was "Unfortunately, we don't have a GPL package for UniFi. "
Jun 19 22:01:56 <jwh> heh
Jun 19 22:03:29 <stintel> I should order another rpi0w
Jun 19 22:04:00 <Monkeh> Can you order 10 of them yet?
Jun 19 22:04:04 <stintel> having to grab this one from the bedroom for the 2nd time in a short period makes me grumpy
Jun 19 22:04:09 <stintel> not with my supplies :(
Jun 19 22:04:12 <stintel> only 1 per order
Jun 19 22:04:20 <jwh> heh
Jun 19 22:04:24 <stintel> so the price effectively increases by ~50%
Jun 19 22:04:27 <stintel> for shipping
Jun 19 22:04:47 <stintel> but still I kinda prefer it over an ESP
Jun 19 22:05:09 <stintel> as you can run a real OS on it and do other stuff than arduino or so
Jun 19 22:05:40 <jwh> http://ix.io/1dYw
Jun 19 22:05:42 <Monkeh> Still no way to buy more than 1 at a time
Jun 19 22:05:44 <jwh> how much do I suck?
Jun 19 22:05:56 <jwh> (and it isn't going to work either, is it?)
Jun 19 22:06:38 <Monkeh> stintel: I do like the 0W - but you still can't really buy them
Jun 19 22:06:41 <Monkeh> So it's still a classic fail.
Jun 19 22:07:47 <stintel> gzip: warning: bin/targets/brcm2708/bcm2708/openwrt-brcm2708-bcm2708-rpi-squashfs-sdcard.img.gz: trailing junk was ignored
Jun 19 22:07:50 <stintel> metadata I guess?
Jun 19 22:08:19 <Monkeh> I believe so
Jun 19 22:14:21 <stintel> sigh, now my workstation keeps mounting the sd as ro
Jun 19 22:15:14 <allotrope94> quick question - are there native lex/yacc equivalents for OpenWRT?
Jun 19 22:15:37 <allotrope94> I want to compile a small set of ncurses programs either natively or by cross-compiling
Jun 19 22:16:01 <allotrope94> frotz, gforth, nethack, bsdgames, and lynx
Jun 19 22:16:19 <allotrope94> edbrowse is in perl, and has no other dependencies, so that should work fine :-)
Jun 19 22:16:55 <allotrope94> I tried cross-compiling nethack 3.4.3, but its makefile requires running the makedefs executable on the host
Jun 19 22:16:59 <allotrope94> So that is a problem
Jun 19 22:17:04 <allotrope94> Any help would be appreciated
Jun 19 22:17:31 <karlp> plenty of packages have host build portions,
Jun 19 22:17:56 <karlp> https://openwrt.org/docs/guide-developer/packages#host_tools_required
Jun 19 22:18:10 <karlp> you can use those targets to do things on the host.
Jun 19 22:19:15 <stintel> might have been the sdcard feature PR that was partially causing my sdcard problems
Jun 19 22:19:46 <allotrope94> karlp: thanks
Jun 19 22:20:22 <allotrope94> I still think I have to compile nethack natively
Jun 19 22:20:26 <allotrope94> if there is a native flex/bison, that would be great
Jun 19 22:20:36 <allotrope94> or if I can cross-compile flex/bison
Jun 19 22:21:22 <karlp> if you can't crosscompile nethack, take it up with nethack, I'm 100% sure that's done repeatedly.
Jun 19 22:21:45 <allotrope94> OK. I will check with them
Jun 19 22:23:19 <Tapper> If we ever get a R-Pi with a gigabit Ethernet port it will be grate for openwrt
Jun 19 22:23:34 <jwh> odroid exists
Jun 19 22:23:36 <jwh> :D
Jun 19 22:23:45 <jwh> it will actually do 600meg
Jun 19 22:23:49 <jwh> probably more
Jun 19 22:24:08 <karlp> those machiens are all also all capable of running "bigger" off the shelf distros
Jun 19 22:24:13 <jwh> yeah
Jun 19 22:24:21 <Tapper> Is odroid exists like a rpi?
Jun 19 22:24:39 <jwh> its actually a worthwhile board
Jun 19 22:25:03 <Tapper> can you just flash a SD card and away you go like a pi?
Jun 19 22:25:07 <jwh> yup
Jun 19 22:25:11 <jwh> or emmc if you desire
Jun 19 22:25:34 <jwh> they supply ubuntu and android images
Jun 19 22:25:36 <Tapper> How strong is the CPU?
Jun 19 22:25:47 <jwh> pretty fast
Jun 19 22:25:59 <Tapper> as fast as a wrt3200acm?
Jun 19 22:26:14 <jwh> whats in that?
Jun 19 22:26:34 <jwh> https://www.hardkernel.com/main/products/prdt_info.php?g_code=G145457216438
Jun 19 22:26:37 <Tapper> ARMv7 Processor rev 1 (v7l)
Jun 19 22:27:27 <jwh> the odroid-c versions are the same soc used in a large majority of android/stbs
Jun 19 22:27:32 <jwh> and do a limited degree, some phones
Jun 19 22:27:37 <jwh> to*
Jun 19 22:27:41 <karlp> I'd expec tthe odroid c2 to be fairly substantially more powerful than that marvel router soc
Jun 19 22:28:20 <xdarklight> jwh: Amlogic chips haven't been used in mobile devices for a long time - the last "mobile" devices were tablets (but that was with the old, 32-bit chips)
Jun 19 22:28:21 <Tapper> Ok cool I will take a look at were i can find one. So I run 2 APs would i have to use a switch?
Jun 19 22:28:34 <jwh> xdarklight: ah
Jun 19 22:28:35 <karlp> xdarklight: kwh didn't say mobile, tehy said STBs
Jun 19 22:28:50 <karlp> Tapper: no wifi on an odroid
Jun 19 22:29:05 <karlp> well, yeah, they did on the next line, my bad :)
Jun 19 22:29:11 <jwh> odroids are really more htpc/small desktop things
Jun 19 22:29:17 <jwh> karlp: yeah, I'm just out of date :D
Jun 19 22:29:24 <Tapper> Yeah so ethernet to a switch and then plugin the 2 APS?
Jun 19 22:29:32 <Tapper> plug in
Jun 19 22:29:39 <jwh> well they also only have one ethernet port
Jun 19 22:29:45 <karlp> ^^^
Jun 19 22:29:51 <jwh> so they're not really useful as a network device
Jun 19 22:30:03 <Tapper> O
Jun 19 22:30:08 <jwh> their single disk nas things are kinda cool though
Jun 19 22:30:08 <karlp> theyr'e a node for doing things, like mine runs nextcloud on sits on the fibre at home.
Jun 19 22:30:09 <jwh> stackable
Jun 19 22:30:34 <jwh> https://www.hardkernel.com/main/products/prdt_info.php?g_code=G151505170472
Jun 19 22:31:05 <jwh> however, the samsung ones all have usb ethernet which kinda sucks
Jun 19 22:31:48 <karlp> but via usb3,
Jun 19 22:31:58 <jwh> yeah
Jun 19 22:32:03 <karlp> wasn't rpi doing gigaethernet on usb2 in the past?
Jun 19 22:32:04 <jwh> certainly better
Jun 19 22:32:14 <jwh> I thought they were only 100meg
Jun 19 22:32:22 <jwh> gig would be a waste since they were only capable of about 30 meg anyway
Jun 19 22:32:53 <jwh> only ever used them in the office for noc tvs
Jun 19 22:32:59 <jwh> and they're not very good at that either
Jun 19 22:34:37 <jwh> https://www.hardkernel.com/main/products/prdt_info.php?g_code=G150152508314 :D
Jun 19 22:35:06 <karlp> that kinda feels like a gimmmick though
Jun 19 22:35:17 <stintel> the fan says it all
Jun 19 22:35:25 <jwh> yeah it is kinda
Jun 19 22:35:32 <jwh> but the form factor and case is kinda cute
Jun 19 22:36:06 <jwh> I want something like that but with 2 nics heh
Jun 19 22:37:06 <Tapper> yeah me to
Jun 19 22:37:11 <jwh> they claim >900mbps out of those usb3 nics which is pretty impressive
Jun 19 22:38:07 <schm0> hmm how to get hostapd qos_map_set working? i added to the hostapd.sh script. the generated config contains the string but the log shows: daemon.err hostapd: Line 12: unknown configuration item 'qos_map_set'
Jun 19 22:40:37 <schm0> and i have the feeling without that option wmm doesnt work. when i invoke tc qdisc show wlan1 only 2 queues have packets going through. I think those are directly mapped in mwlwifi
Jun 19 23:00:05 <jwh> going mad, can't find fw3 in packages/
Jun 19 23:00:54 <jwh> nm I suck
Jun 19 23:04:17 <EvilRick> so where can I find openwrt developers? I want a quote to add a device definition. aka board support.
Jun 19 23:04:28 <EvilRick> please pm me.
Jun 19 23:05:16 <jwh> they're in here, but its midnight or later for most
Jun 19 23:05:27 <jwh> send to mailing list or hang around
Jun 19 23:05:51 <jwh> what board?
Jun 19 23:10:25 <ldiamond_> `opkg install block-mount` -> Unknown package 'block-mount'
Jun 19 23:10:41 <EvilRick> its a Wlink R220, the vendor. the Manufacturer says to use the Atheros DB120 reference board setup and then gives instructions to tweak some gpio's nad info on how to access extra NAND
Jun 19 23:10:42 <ldiamond_> Is this package broken? I also did include it in my `make menuconfig`,
Jun 19 23:10:57 <EvilRick> problem is the docs are circa 2016 and a lot has changed
Jun 19 23:11:11 <EvilRick> I got an ar71xx image running on the device
Jun 19 23:11:17 <EvilRick> but without nand support
Jun 19 23:12:04 <ldiamond_> Or maybe my build is broken. opkg gets a 404 on this: http://downloads.lede-project.org/snapshots/targets/mvebu/generic/packages/Packages.gz
Jun 19 23:12:23 <ldiamond_> The rest of the repos are fetched under this url http://downloads.lede-project.org/snapshots/packages/arm_cortex-a9_vfpv3/luci/Packages.gz
Jun 19 23:13:00 <ldiamond_> mvebu generic doesn't exist anymore.
Jun 19 23:14:58 <ldiamond_> I modified the distfeeds.conf file to point to cortexa9 and it works. Idk why my build generated that though
Jun 19 23:21:56 <jwh> EvilRick: ar71xx sucks, check ath79
Jun 19 23:23:38 <karlp> that's hardly going to help with getting nand support working
Jun 19 23:24:18 <stintel> [   14.022750] lirc_rpi lirc_rpi: lirc_dev: driver lirc_rpi registered at minor = 0
Jun 19 23:24:32 <stintel> unfortunately the wifi on my rpi0w is still broken :(
Jun 19 23:25:41 <stintel> reverted a couple of brmcfmac changes but still no lukc
Jun 19 23:25:57 <jwh> karlp: no, but if its just a case of having to describe the extra nand, dts might be easier heh
Jun 19 23:26:11 <jwh> obviously if it needs a driver or modifications to one then no, it won't help at all
Jun 19 23:29:26 <schm0> do i need hostapd(full? + wpa supplicant) instead of wpad to make qos map work?
Jun 19 23:30:22 <karlp> wpad is the "both"
Jun 19 23:39:39 <EvilRick> jwh: yes I would like to use ath79. I checked out the existing DTS files and felt somewhat out of my depth as there was no db120 reference board to work off. the chip on the board is a 9334 and I was looking at those files and includes but as I said, felt a bit out of my depth when it comes to things like memory registers and ethernet PLL data.
Jun 19 23:41:52 <schm0> i know
Jun 19 23:41:57 <EvilRick> also ar9344.dtsi seems to specifically make reference to "TP-Link WDR4300" which does not feel very generic to me
Jun 19 23:42:31 <schm0> but im trying to figure out why hostapd doesn't take qos_map_set option
Jun 19 23:43:55 <stintel> rmilecki: starting to suspect 037ef13a161c6167b328aaf6ca0e259a5b7db037 broke wifi on my rpi0w
Jun 19 23:45:32 <stintel> ugh, 2:45AM :(
Jun 19 23:46:59 <stintel> dangole isn't lurking around on IRC usually?
Jun 19 23:47:05 <schm0> but hostapd-full also gives  unknown configuration item 'qos_map_set'
Jun 19 23:48:11 <stintel> qos_map_set is within #ifdef CONFIG_INTERWORKING #endif
Jun 19 23:48:18 <schm0> but this is a valid option according to https://w1.fi/cgit/hostap/plain/hostapd/hostapd.conf and in the source (config_file_c or something) there is also a function to parse qos_map_set
Jun 19 23:48:29 <stintel> I don't believe we have CONFIG_INTERWORKING
Jun 19 23:48:29 <schm0> hmm
Jun 19 23:49:01 <stintel> #CONFIG_INTERWORKING=y in hostapd-full.conf
Jun 19 23:49:11 <stintel> package/network/services/hostapd/files/hostapd-full.config
Jun 19 23:49:21 <stintel> uncomment it, rebuild and try again
Jun 19 23:49:50 <schm0> ty
Jun 19 23:50:23 <stintel> yw
Jun 19 23:55:03 <stintel> yep, that 037ef13a161c6167b328aaf6ca0e259a5b7db037 is the culprit
Jun 19 23:58:41 <EvilRick> ok wow.. just dicovered this device has a http based bootloader. "breed"
Jun 19 23:59:01 <EvilRick> no wonder I couldnt find firmware upload option sin the cli
Jun 19 23:59:34 <EvilRick> the web interface actually has references to TP-LINK in it
Jun 20 00:23:55 <EvilRick> why does does downloads.openwrt.org 17.01.04 not contain any ath79 targets?
Jun 20 00:23:55 <jwh> hm, nearly
Jun 20 00:25:17 <jwh> oh I suck, space in the path for fopen :(
Jun 20 00:26:12 <jwh> probably not built for 17
Jun 20 00:26:17 <jwh> check 18 or master
Jun 20 00:26:19 <stintel> EvilRick: ath79 only exists in master
Jun 20 00:29:07 <ausjke> mt76 is for mediatek 2.4g/5g softmac in general, what is exactly ath79?
Jun 20 00:29:35 <ausjke> you had ath5k/ath9k/ath10k, consistent namings
Jun 20 00:29:39 <jwh> https://i.imgur.com/4IMzeRE.png
Jun 20 00:29:41 <jwh> did I do it right?
Jun 20 00:30:50 <stintel> ausjke: ath79 is a target, not a wifi driver
Jun 20 00:31:09 <ausjke> CONFIG_ATH79: Atheros AR71XX/AR724X/AR913X based boards
Jun 20 00:31:12 <stintel> it is the upstream equivalent of ar71xx, but using DTS while ar71xx uses mach files
Jun 20 00:31:19 <ausjke> i see, thanks
Jun 20 00:31:31 <stintel> which is why it's going to take a while until ath79 can fully replace ar71xx
Jun 20 00:31:41 <stintel> as all mach files needs to be converted to dts
Jun 20 00:32:08 <stintel> also for that reason we're not going to bump ar71xx to 4.14 anymore
Jun 20 00:32:35 <ausjke> i was against dts in the old days while in motorola when dts was initially pushed into kernel.org, did not realize it is so popular these days, nearly all non-x86 use it
Jun 20 00:32:50 <ausjke> it started as a powerpc thing
Jun 20 00:50:41 <jwh> https://github.com/joeholden/fw3/commit/afc95a95a202006d2ee4ba392f52e0f2a5dbd92b
Jun 20 00:50:47 <jwh> it works, dunno if I did it right though :d
Jun 20 01:13:08 <jwh> now to teach netifd about macsec
Jun 20 01:13:31 <jwh> deffo gonna need some help with this one heh
Jun 20 01:19:26 <stintel> yay, got a running lircd \o/
Jun 20 01:20:35 <jwh> o/
Jun 20 01:21:24 <stintel> cleaning it up is going to be quite some work though
Jun 20 01:21:54 <stintel> bloody wrapper scripts to binaries library directories, wut
Jun 20 01:23:08 <jwh> heh
Jun 20 01:23:13 <schm0> back after some nightwalk. enjoying the silence
Jun 20 01:23:31 <schm0> just comeback to see the build has failed
Jun 20 01:24:43 <schm0> https://pastebin.com/edit/PyxNh1uv
Jun 20 01:27:11 <jwh> hm
Jun 20 01:27:37 <jwh> what examples of ethernet interfaces with parents are there that openwrt supports?
Jun 20 01:27:41 <jwh> trying to find a good example
Jun 20 01:27:57 <jwh> I mean I guess .1q is probably a reasonable example
Jun 20 01:27:58 <jwh> hm
Jun 20 01:29:26 <jwh> ah hm, actually macvlan is probably a better example
Jun 20 01:34:34 <schm0> nvm i forgot to enable internetworking also in wpa_supplicant conf?°
Jun 20 01:48:40 <schm0> mwlwifi with mesh disabled still seems broken
Jun 20 01:49:48 <schm0> https://github.com/kaloz/mwlwifi/issues/302 how yuhhaurlin fixed everthing else but this ignored he^^
Jun 20 01:50:07 <schm0> *but this he ignored
Jun 20 01:50:33 <schm0> makes more sense?
Jun 20 01:51:08 <schm0> what makes sense anyway ¯\_(ツ)_/¯
Jun 20 01:53:45 <schm0> enabling internetworking also in wpa supplicant also didnt work now im screwed
Jun 20 02:00:54 <ubermaster> Hi I have a problem compiling a package which uses "#include <array>" and when I cross compile it it says "fatal error: array: no such file or directory" and it points to the include line
Jun 20 02:01:47 <ubermaster> Does anyone know why it can't find a basic include? I'm sorry I do not know C much but I assume <array> is a basic include
Jun 20 02:02:16 <jwh> umm
Jun 20 02:02:27 <jwh> what is "array" ?
Jun 20 02:03:48 <ubermaster> I believe that is what the compiler is asking me, "what is this include <array>"
Jun 20 02:04:00 <ubermaster> I believe that is the actual error
Jun 20 02:04:30 <ubermaster> I grepped other packages in the SDK and there are others that have "#include <array>" in them but there is no array file
Jun 20 02:04:42 <jwh> examples?
Jun 20 02:04:45 <ubermaster> I assume it is standard library
Jun 20 02:05:01 <ubermaster> What example? are you trolling me?
Jun 20 02:05:25 <jwh> if you like
Jun 20 02:05:48 <ubermaster> You are trolling me.... If you dont know what an array is why are you in #lede-dev?
Jun 20 02:06:26 <jwh> same reason the other 300 or so others who don't do any development are i'd imagine
Jun 20 02:07:03 <jwh> but at least I know what #include does heh
Jun 20 02:07:11 <ubermaster> Ok well I would like to wait for another fellow developer to help out, or someone who programs the SDK
Jun 20 02:07:30 <ubermaster> Yes as do I, you moron, it includes something
Jun 20 02:07:34 <jwh> sure
Jun 20 02:07:38 <jwh> it includes files
Jun 20 02:07:40 <ubermaster> I know other languages.... just not C
Jun 20 02:07:43 <jwh> C is not C++
Jun 20 02:08:09 <ubermaster> I dont know C++ either so please stop being a complete dick
Jun 20 02:08:27 <jwh> #include <stdlibthing> is ++ behaviour
Jun 20 02:08:35 <ubermaster> there is a standard library not being linked in because it is missing somehow
Jun 20 02:08:40 <jwh> hence why I asked for examples
Jun 20 02:08:46 <ubermaster> or something like that
Jun 20 02:08:49 <jwh> so I can see what you're looking at when you said its in other packages
Jun 20 02:08:58 <ubermaster> example: #include <array>
Jun 20 02:09:18 <jwh> yes, those are C++ stdlib includes
Jun 20 02:09:19 <jwh> not C
Jun 20 02:09:35 <ubermaster> I suspect its something to do with stdc++ or libClibcpp or whatever
Jun 20 02:09:56 <ubermaster> Yes so why would it not be there to be included?
Jun 20 02:10:08 <jwh> because I'd imagine you're trying to compile it as C, not ++
Jun 20 02:10:23 <ubermaster> I am using a standard CMAKE_INSTALL:=1 with only one additional flag
Jun 20 02:10:35 <jwh> is it actually trying to use g++?
Jun 20 02:10:39 <ubermaster> should the SDK not select the right std libraries to include?
Jun 20 02:10:59 <ubermaster> No, I am using CMake and Make
Jun 20 02:11:07 <jwh> cmake should probably be the one to do that
Jun 20 02:11:11 <ubermaster> inside of the package Makefile
Jun 20 02:11:25 <ubermaster> yes that is why I am confused why CMake didnt add it
Jun 20 02:11:31 <jwh> don't really do cmake, but presumably you'd have to hint it to use a c++ compiler instead
Jun 20 02:11:36 <ubermaster> one sec I pastebin the makefile
Jun 20 02:12:11 <ubermaster> Does it use C by default? I thought it autodetected all that stuff....
Jun 20 02:13:05 <ubermaster> but it was not complaining about syntax, it complained about a standard library...weird.
Jun 20 02:14:33 <jwh> well #include is common between both, so it probably bailed on that before the rest
Jun 20 02:14:51 <jwh> are your files names .cc?
Jun 20 02:14:54 <jwh> named
Jun 20 02:15:03 <jwh> or whatever the proper name is these days
Jun 20 02:17:18 <jwh> bleh
Jun 20 02:17:24 <jwh> I'm sucking at netifd
Jun 20 02:17:45 <jwh> debug suggests its not actually reading the config
Jun 20 02:18:13 <ubermaster> no they are cpp
Jun 20 02:18:48 <ubermaster> this is after I resolved all other dependancies like jansson
Jun 20 02:19:07 <ubermaster> then it started complaining about the array include
Jun 20 02:19:24 <ubermaster> +libuuid +jansson +libuv
Jun 20 02:19:37 <ubermaster> I added those dependancies because they errored first ^
Jun 20 02:20:04 <jwh> should be able to see debug
Jun 20 02:20:07 <ubermaster> Wole line - DEPENDS:=+libpthread +libstdcpp +libc +libuuid +jansson +libuv +uclibcxx
Jun 20 02:20:17 <jwh> if its actually trying to build as c or c++
Jun 20 02:20:25 <jwh> if its the latter then maybe need to set version?
Jun 20 02:20:37 <ubermaster> would like pastebin of Makefile or Debug or both?
Jun 20 02:20:48 <ubermaster> It is the latter :P C++
Jun 20 02:21:02 <ubermaster> I look at cmake flags and see whats there
Jun 20 02:21:09 <jwh> don't think I'll be much help tbh, I've avoided C++ as much as possible
Jun 20 02:21:42 <jwh> also kinda distracted, trying to figure out how to netifd heh
Jun 20 02:22:34 <jwh> does not do what it is supposed to do
Jun 20 02:22:42 <ubermaster> It's ok thanks for help anyways you may have shoved me in right direction
Jun 20 02:23:10 <jwh> the only bit I know about cmake is 'cd build; cmake ..'
Jun 20 02:23:11 <jwh> :D
Jun 20 02:27:24 <jwh> argh
Jun 20 02:28:55 <jwh> it doesn't even try :(
Jun 20 02:30:17 <jwh> I don't really get what its trying to do either though and have no inclination to step through the entirety of the code to figure it out heh
Jun 20 02:38:42 <jwh> at least a crash would indicate something heh
Jun 20 02:39:41 <ubermaster> you don't have to, the compiler points right at the include line and complains it doesnt exist. just need to know why it is not finding something that sould be included by default....
Jun 20 02:39:58 <jwh> oh, I'm talking about my attempt to add nice things to netifs :P
Jun 20 02:40:00 <jwh> netifd
Jun 20 02:40:09 <jwh> it starts, but it does... nothing
Jun 20 02:40:12 <ubermaster> I suspect this is an issue with the SDK because I use a simple Makefile
Jun 20 02:40:19 <ubermaster> I go sleep on it
**** ENDING LOGGING AT Wed Jun 20 03:00:08 2018