**** BEGIN LOGGING AT Mon Oct 22 03:00:00 2018
Oct 22 03:25:55 <biangbiangmian> Is there a specific channel for LuCI?
Oct 22 03:27:26 <DonkeyHotei> no
Oct 22 03:27:41 <DonkeyHotei> here
Oct 22 03:28:53 <biangbiangmian> I've made a new LuCI protocol for ModemManager https://bitbucket.org/mips171/luci-proto-modemmanager/wiki/Home
Oct 22 04:03:54 <uberushaximus> biangbiangmian: awesome
Oct 22 06:06:52 <blogic> hi
Oct 22 06:07:02 <blogic> we'll be in lisabon next week this time
Oct 22 07:42:28 <Tapper> Hi is there any one around to have a look at a patch I just sent to the email list pleas
Oct 22 07:42:44 <Tapper> I just want to know if i did it write?
Oct 22 08:20:09 <xback> responding on some chatter above regarding ar71xx
Oct 22 08:23:20 <xback> I'm working on fixing the issues regarding the kernel size increase
Oct 22 08:23:52 <xback> there are some solutions to make the partitions dynamic, or moving some to tiny.  but in the end, they all need to work and to be properly tested
Oct 22 08:23:58 <xback> so it takes some time ..
Oct 22 08:25:37 <ldir> rmilecki: ping
Oct 22 08:25:48 <rmilecki> ldir: ?
Oct 22 08:32:51 <ldir> rmilecki: sent you a pm :-)
Oct 22 08:38:14 <jow> biangbiangmian: yes, the luci channel is #luci-hackers on freenode
Oct 22 09:02:13 <Tapper> xback thanks for looking in to that.
Oct 22 09:18:30 <Tapper> Sent v2 of my patch
Oct 22 09:33:11 <stintel> hmmm, kmemleak initially points to memleak in conntrack
Oct 22 09:42:21 <stintel> if anyone wants to look at my memleak: https://gist.github.com/bb3aa6ee5e3ed6b7bfa723c55f8f94ac
Oct 22 09:43:56 <KanjiMonster> stintel: 404
Oct 22 09:44:23 <stintel> wut
Oct 22 09:44:33 <stintel> but I think I see it already
Oct 22 09:45:18 <stintel> ok github is still acting up after their booboo from last night I guess
Oct 22 09:45:32 <stintel> the problem is very likely in the rtcache patch
Oct 22 09:46:49 <stintel> I rmmod the module and poof panic
Oct 22 09:46:58 <stintel> something is definitely fishy there
Oct 22 09:57:04 <stintel> yeah, github gist is broken
Oct 22 09:58:23 <stintel> https://status.github.com/messages
Oct 22 09:58:58 <stintel> shit really hit the fan there
Oct 22 09:59:06 <stintel> microsoft took over operations? :P
Oct 22 10:00:01 <jow> maybe they tried to migrate the datastore to ntfs
Oct 22 10:00:07 <stintel> 😂
Oct 22 10:00:15 <jow> chkdisk still running
Oct 22 10:00:25 <stintel> hahaha
Oct 22 10:01:30 <stintel> thinking if we should just drop the rtcache patch, iirc it was never accepted upstream
Oct 22 10:01:49 <rmilecki> it wasn't, that's right
Oct 22 10:02:01 <rmilecki> nbd: ^^
Oct 22 10:02:11 <rmilecki> let's let Felix comment
Oct 22 10:02:23 <stintel> and wondering if it maybe clashes with flow_offload
Oct 22 10:05:59 <stintel> hmm. removed the module from /etc/modules.d/nf-conntrack, rebooted, but still seeing the memleak
Oct 22 10:10:22 <stintel> https://paste.pound-python.org/show/TMduNYGtbi6pPEdY5hp0/
Oct 22 10:35:28 <mangix> stintel: flow offload?
Oct 22 10:36:27 <rmilecki> mangix: you could'nt possible miss OpenWrt getting that feature ;)
Oct 22 10:38:51 <mangix> no i'm saying that could be it. has been a problem in the past.
Oct 22 10:38:53 <ldir> hmmm https://pastebin.com/5du6bnM0 is this our heisenbug back again
Oct 22 10:40:17 <abenz> jow: reverting the last wolfssl update fixes the compile error
Oct 22 10:40:22 <abenz> https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=ed0d5a1e609e0b39eff9f06e3522396581d0b06e
Oct 22 11:08:28 <jow> abenz: ?
Oct 22 11:30:53 <abenz> jow: https://pastebin.com/HrvGZnp5
Oct 22 11:32:33 <abenz> wolfssl fails again after the update (do you remember we did this not long ago, also for wolfssl)
Oct 22 11:59:58 <biangbiangmian> @jow thanks
Oct 22 12:12:22 <ldir> Tapper: ping - did you send in a v2 of your patch 'cos I can't find it.
Oct 22 12:25:51 <stintel> or maybe those conntrack leaks are false positives
Oct 22 12:26:00 <stintel> ugh, this stuff is way over my head
Oct 22 12:26:55 <stintel> mangix: flow offload was disabled on the machine where I am clearly seeing the problem (VM with only 256MB RAM, easier to notice than on an APU2 with 4GB)
Oct 22 12:40:04 <dedeckeh> jow:ping
Oct 22 12:41:46 <jow> dedeckeh: pong
Oct 22 12:44:19 <dedeckeh> jow:In PR https://github.com/openwrt/openwrt/pull/1454 a nftables-json variant is made
Oct 22 12:44:38 <dedeckeh> which relies on the package jansson
Oct 22 12:44:54 <dedeckeh> which is in feed/packages
Oct 22 12:46:04 <dedeckeh> since we want to avoid to depend on packages in feeds/packages should we move jansson to OpenWrt git repo ?
Oct 22 12:46:36 <dedeckeh> or move nftables to feeds/packages ?
Oct 22 12:46:51 <karlp> or make it stop using jansson and use json-c instead?
Oct 22 12:46:52 <jow> nftables is going to be a core components, I'd prefer it to stay in base
Oct 22 12:46:57 <stintel> +1
Oct 22 12:47:01 <stintel> (to stay in core0
Oct 22 12:47:07 <jow> and yes, I'd rather rework it to use json-c
Oct 22 12:47:17 <jow> until then simply do not offer json export
Oct 22 12:47:17 <dedeckeh> I'm certainly not in favor to move it to feeds/packages
Oct 22 12:48:31 <dedeckeh> Ok I will give this as comment to use json-c
Oct 22 12:49:00 <jow> haven't yet looked at the nft code base so not sure how simple/hard it would be
Oct 22 12:49:15 <dedeckeh> me neither
Oct 22 12:49:46 <jow> I also do not like the naming of the buildvariant
Oct 22 12:49:59 <jow> usually its $pkg-$feature, not $pkg-with{$feature}
Oct 22 12:53:51 <jow> from a cursorly look at the Jannson api it looks like it should be possible to build a Jannson compat layer for json-c
Oct 22 12:53:55 <jow> *cursory
Oct 22 12:54:42 <jow> so maybe it is feasible to write a libjannson-json-c which mimicks the api and uses json-c internally
Oct 22 12:55:13 <jow> the only "unusual" thin in jannson is the json_pack() family of functions
Oct 22 12:58:07 <dedeckeh> the libjansson-json-c api should than be added to the libjson-c package or do you rather see it as a stand-alone package ?
Oct 22 12:59:02 <jow> standalone
Oct 22 13:01:43 <karlp> how heavily wedded to jansson is the upstream code?
Oct 22 13:02:22 <jow> the first json call I found was json_pack()
Oct 22 13:02:29 <jow> so ... heavy I guess
Oct 22 13:50:28 <karlp> doh. :)
Oct 22 13:50:44 <karlp> that _pack api does look convenient for some usecases.
Oct 22 13:51:20 <jow> yeah, its a bit like printf for json
Oct 22 14:16:46 <xback> also like scanf :)  with the option to only convert it if it does exist
Oct 22 14:16:54 <xback> (i'm using jansson heavily in my app)
Oct 22 14:17:18 <jwh> I love jansson
Oct 22 14:17:47 <xback> only downside is the lib size and speed ..
Oct 22 14:17:54 <jwh> its not so bad
Oct 22 14:18:02 <jwh> convenience has its cost
Oct 22 14:18:03 <karlp> well, unless you want to conver tall of openwrt code to jansson, it doesn't matter much if you like it :)
Oct 22 14:18:12 <jwh> heh
Oct 22 14:36:48 <huaracheguarache> greearb: the firmware you provided for debugging seems to be pretty stable. I've tried to make it crash by running an iperf session between a computer connected to the wifi and one connected via ethernet. running for several hours without a crash
Oct 22 15:50:46 <greearb_> huaracheguarache, I suspect power-save is what triggers the corruption
Oct 22 16:12:26 <huaracheguarache> greearb_: should I turn on power management for my wifi on my laptop? I have it turned off since I had some issues with irc connection over ipv6 dropping out when it's on
Oct 22 16:17:59 <greearb_> huaracheguarache, sure, but either way, I appreciate the testing
Oct 22 16:28:02 <huaracheguarache> greearb_: no worries, I'll give it a try =)
Oct 22 18:36:26 <KevinDB> now that dnsmasq v2.80 has been released, what are the thoughts on bumping 18.06 which is currently on v2.80test3 ?
Oct 22 18:40:24 <stintel> ack from me
Oct 22 19:13:51 <Tapper> Hi stintel just out of intrest what does ack stand for?
Oct 22 19:14:50 <Tapper> I have seen it on github befor and I know that ack is a yes and nack is a no.
Oct 22 19:17:48 <DLange> acknowledge
Oct 22 19:18:04 <DLange> and _n_ot acknowledged = nack
Oct 22 19:27:22 <Tapper> thanks DLange
Oct 22 19:36:19 <Monkeh> Aff and neg are much cooler.
Oct 22 19:52:00 <dedeckeh> KevinDB:ack from me as well
Oct 22 19:57:53 <Hauke> stintel: we should try to upstream or drop more of the kernel patches
Oct 22 20:02:53 <Tapper> Hi any one know if there is a readme for luci-app-https_dns_proxy?
Oct 22 20:03:25 <mangix> mvebu build error stiol
Oct 22 20:03:35 <mangix> still hasn't been fixed
Oct 22 20:08:55 <Tapper> mangix the snapshots are still building. Or you on about something different
Oct 22 20:45:33 <mangix> Tapper: i'm talking specifically about the cortex a9 ones
Oct 22 20:46:29 <Tapper> OK I just looked and the snapshot for my wrt3200acm was upbated today
Oct 22 20:47:04 <Tapper> which is cortexa9/
Oct 22 20:48:14 <Tapper> Mon Oct 22 01:56:29 2018
Oct 22 21:13:16 <vk496> To enter into failsafe mode, it possible when you press any button?
Oct 22 21:13:24 <vk496> or only reset button?
Oct 22 21:19:45 <Redfoxmoon> vk496, depends on your device, and which button is wired up to be the failsafe button; if any
Oct 22 21:19:54 <Redfoxmoon> in *most* cases it will be the reset button though.
Oct 22 21:36:29 <russell--> mail to openwrt-devel is being held because "Message has a suspicious header" wut?
Oct 22 22:01:09 <greearb__> so if I want to disable whatever starts hostapd and configures the network and such, what do I disable?
Oct 22 22:14:37 <mangix> greearb__: netifd?
Oct 22 22:15:47 <greearb__> worth a try, if I can get my network testing SW to run on openwrt, I'll need to disable the normal network config
Oct 22 22:16:17 <stintel> wifi is disabled by default, I don't think you need to do anything
Oct 22 22:16:35 <greearb__> hmm, ok
Oct 22 22:17:05 <greearb__> first I need to build an image that boots, and then get my aplication cross-compiling with openwrt toolchain
Oct 22 22:18:00 <greearb__> think I'll target netgear r7800 as first attempt
Oct 22 22:23:50 <luaraneda> Is there a way to tell me to not manage an interface? Something like what NetworkManager does when you put the interface's MAC address into a file. I've seen people before asking something similar
Oct 22 22:24:12 <luaraneda> To tell netifd*
Oct 22 22:49:33 <owrt-1806-builds> build #142 of mvebu/cortexa72 is complete: Success [build successful]  Build details are at http://release-builds.openwrt.org/18.06/images/builders/mvebu%2Fcortexa72/builds/142
Oct 23 00:40:27 <jwh> luaraneda: proto=manual, maybe
Oct 23 00:40:38 <jwh> or something along those lines
**** ENDING LOGGING AT Tue Oct 23 02:59:59 2018