**** BEGIN LOGGING AT Tue Aug 28 03:00:03 2018
Aug 28 05:52:08 <blogic> hi
Aug 28 05:52:45 <nitroshift> hey blogic
Aug 28 07:00:36 <wigyori> morning
Aug 28 07:15:31 <blogic> wigyori: hi there
Aug 28 07:19:05 <nitroshift> hey wigyori
Aug 28 07:31:30 <blogic> https://patchwork.ozlabs.org/patch/961068/  anyone got feelings regarding this ?
Aug 28 07:59:39 <KanjiMonster> blogic: the get_big_endian() is unneeded and can be replaced by ntohl()
Aug 28 07:59:54 <KanjiMonster> also seperate => separate
Aug 28 08:07:27 <blogic> KanjiMonster: not htonl() ?
Aug 28 08:09:01 <KanjiMonster> blogic: right. In my head that was a "get the value of a big endian int", not "get the big endian version of an int"
Aug 28 08:09:16 <blogic> ok, just wanted to be sure, thought i was dense
Aug 28 08:10:55 <KanjiMonster> also of course it would be nice if the ras-thing would support long options so the switch to it would only require changing the called executable, but eh
Aug 28 08:11:19 <blogic> tradeoff between getting stuff merged and not
Aug 28 08:11:29 <blogic> its better than before so fine with me :-)
**** BEGIN LOGGING AT Tue Aug 28 09:18:13 2018
Aug 28 10:22:47 <johan> i get the error "Error: Unknown error (INVALID_PROTO)" when trying to start wireguard on my wrt3200acm. ive installed wireguard, wireguard-tools, kmod-wireguard and luci-app-wireguard. anyone know why it complains on a unknown protocol?
Aug 28 10:25:57 <KanjiMonster> johan: have you tried doing a /etc/init.d/network restart or a reboot? IIRC netifd caches known protocols on start, so if you install new protocols while netifd is up it won't recognize them
Aug 28 10:26:36 <johan> ah ok.. no i didnt try that yet. ill be back - thank you.
Aug 28 10:27:41 <johan> KanjiMonster: that solved it.
Aug 28 10:28:20 <johan> im sorry for the stupid question then. totally ----
Aug 28 10:28:23 <johan> thank you again
Aug 28 10:28:37 <KanjiMonster> not a stupid question at all, as it's quite non-obvious behaviour
Aug 28 10:29:25 <johan> but i should've tried that before begging for help :)
Aug 28 11:04:54 <stintel> would be nice if the package installing a new proto handler would reload netifd automatically
Aug 28 11:05:00 <stintel> can we do that easily?
Aug 28 11:09:34 <johan> i got it work now anyhow.. but the ipv6 via WG isn't - and i dont know why. the WG iface and client(s), have got their own /64 to work with.. but still it doesn't work. Im running just about the same setup on a VPS.. and i've got both ipv4 & ipv6 working via WG on the clients. i wonder if i need to add some other routes or something?
Aug 28 11:16:32 <Rene__> johan: can both sides ping eachother ipv6 address?
Aug 28 12:34:42 <huaracheguarache> greearb: should I document the behavior and create an issue thread in your repo on git?
Aug 28 12:54:23 <blogic> stintel: yes, postinst trigger
Aug 28 12:54:40 <blogic> stintel: ideally a small helper function
Aug 28 12:58:00 <huaracheguarache> stintel: did you try out the new ath10k-ct driver and firmware?
Aug 28 12:58:14 <stintel> huaracheguarache: yes, I pushed it to master
Aug 28 12:58:17 <stintel> it works like a charm
Aug 28 12:58:21 <stintel> never going  back to stock
Aug 28 12:58:34 <blogic> stintel: ok so we will drop stock in that case
Aug 28 12:58:42 <blogic> already spoke with felix about it
Aug 28 12:58:45 <stintel> blogic: this is only for qca998x though
Aug 28 12:59:00 <huaracheguarache> blogic: one sec, qca9984 still has some bugs that need to be fixed
Aug 28 12:59:24 <stintel> blogic: but it's probably the way to go forward
Aug 28 12:59:27 <huaracheguarache> 802.11w clients have very low speeds even on the latest firmware
Aug 28 12:59:40 <stintel> because nobody at QCA gives a flying fuck about bugs in Linux
Aug 28 13:00:36 <stintel> huaracheguarache: did you tell Ben / filed ticket at his github?
Aug 28 13:00:39 <huaracheguarache> but yes, I agree that it's the right way to move forward
Aug 28 13:01:00 <huaracheguarache> stintel: I only briefly spoke with him here on irc, but I'm probably going to file a ticket
Aug 28 13:01:04 <stintel> ok
Aug 28 13:01:06 <stintel> please do so
Aug 28 13:01:11 <huaracheguarache> ok
Aug 28 13:04:30 <stintel> I have closed FS#333
Aug 28 13:06:22 <huaracheguarache> stintel: how much uptime do you have on the latest driver and firmware?
Aug 28 13:06:40 <stintel> huaracheguarache: the latest? not that much really
Aug 28 13:06:52 <stintel> but it's in master so hopefully more people will test it
Aug 28 13:07:24 * mamarley is testing it.  It works well so far with no problems to report.
Aug 28 13:07:27 <huaracheguarache> I thought that greearb said that he didn't manage to find the root cause of that issue and solve it
Aug 28 13:07:56 <mamarley> He did solve it.  The issue was that block-ack wasn't being used with .11w.
Aug 28 13:07:57 <stintel> huaracheguarache: I am not seeing it with ath10k-ct so far
Aug 28 13:08:06 <stintel> mamarley: FS#333 is about something else
Aug 28 13:08:15 <mamarley> Oops, sorry, wrong issue.
Aug 28 13:08:28 <stintel> but with -ct I have yet to see it
Aug 28 13:08:48 <stintel> and since I have my Gemini PDA (mediatek + android), ath10k usually crapped itself within 24h
Aug 28 13:08:49 <huaracheguarache> stintel: my understanding is that it craps out far less often than with stock
Aug 28 13:08:55 <mamarley> I haven't seen any FW crashes either.
Aug 28 13:08:55 * mamarley knocks on a nearby wooden surface.
Aug 28 13:09:57 <stintel> either way, I wrote in the note for closing FS#333 to use -ct driver + FW
Aug 28 13:10:12 <huaracheguarache> but then again I've never had this issue
Aug 28 13:10:34 <mamarley> I started off with -ct, so I have no data at all with the stock FW>
Aug 28 13:11:29 <stintel> I have seen uptimes of over a month before finally running into #333 again
Aug 28 13:11:49 <stintel> but as I said, I now have the perfect combination of clients that can crash it usually within 24h-48h
Aug 28 13:12:08 <mamarley> Mine never runs that long, I update the FW too much. :P
Aug 28 14:27:52 <mirko> is Hirokazu MORIKAWA hanging around here (maintainer of libicu)?
Aug 28 14:28:17 <jwh> mmmmm, so RPS is enabled by default now everywhere?
Aug 28 16:18:43 <philipp64> what the hell?  did everyone switch to defaf and fall asleep at their desks?
Aug 28 16:18:50 <philipp64> * decaf
Aug 28 16:24:42 <stintel> I usally don't drink coffee :)
Aug 28 16:34:30 <philipp64> mmmm… geneveer
Aug 28 16:35:48 <ntd> I have a few sites where the LTE USB modem borks every 14 days or so. ifdown/ifup doesn't do the trick and a power cycle of the router isn't optimal. is there any way to power cycle the usb ports? (ar71xx)?
Aug 28 16:40:13 <Zero_Chaos> ntd: most devices don't support such things, and this is the main project that I am aware of https://github.com/mvp/uhubctl
Aug 28 16:40:46 <Zero_Chaos> ntd: to get around a similar issue, I was issuing the hard reset command to the cell modem over serial
Aug 28 16:41:10 <DonkeyHotei> ntd: some devices export a gpio that does exactly that
Aug 28 16:45:00 <ntd> AT^reset?
Aug 28 16:45:13 <ntd> did not think of that.  thanks
Aug 28 16:49:59 <ntd> anyone know of a watchdog script i can lift?
Aug 28 16:51:01 <ntd> perhaps something like nslookup &host &vpnresolver, if fail then echo "at^reset" > /den/ttyUSB0?
Aug 28 16:51:09 <ntd>  /dev even
Aug 28 16:53:51 <Zero_Chaos> ntd: you have to send the reset BEFORE it goes out to lunch though. so that's the issue.
Aug 28 16:54:02 <Zero_Chaos> ntd: so maybe once a day on a timer
Aug 28 16:54:42 <ntd> iirc the serial interface works still works after it borks
Aug 28 16:54:53 <Zero_Chaos> ntd: it didn't for me
Aug 28 16:55:01 <Zero_Chaos> ntd: also this is my command "at+cfun=1,1"
Aug 28 16:55:10 <ntd> something to do with modem<->sim or the radio
Aug 28 16:55:16 <Zero_Chaos> ntd: possibly different per radio, but it worked well for me
Aug 28 18:07:49 <andy2244> is anyone using the snapshots sdks to compile packages? I get "/bin/sh: mkhash: not found" since a few days constantly?
Aug 28 18:08:22 <profmac> I am looking at https://bugs.openwrt.org/index.php?do=details&task_id=1743 and I am thinking "how hard could it be really for me to diff from 15.05.1 and pull the problematic 5G radio support out of the build."
Aug 28 18:45:13 <ultito> gninrom
Aug 28 18:51:24 <huaracheguarache> gnineve
Aug 28 19:37:05 <jwh> SwedeMike: hm, what did you say needed to be installed for lw4o6?
Aug 28 19:38:32 <jwh> o, hm
Aug 28 19:43:28 <KanjiMonster> jwh: map provides lw4o6
Aug 28 19:43:48 <jwh> yeah I found it, think it just got lost amongst all the other matches in / (search) for 'map'
Aug 28 19:43:51 <jwh> heh
Aug 28 19:44:30 <carldani> meh. another reboot, and no crashlog :-(
Aug 28 19:44:36 <jwh> I can't find any concrete for map-* on the wiki, if odhcpc6 is presented with the attributes will it just work?
Aug 28 19:44:50 <jwh> or do I need to specify map in the config?
Aug 28 21:51:29 <tripolar> hi
Aug 28 21:53:32 <tripolar> nbd jow Floppe is it known that packages set with PKG_BUILD_DEPENDS inside a Makefile are not selected in as module in make menuconfig when it is selected for build?
Aug 28 21:53:40 <tripolar> [florian]:
Aug 28 21:54:54 <tripolar> so packge A has PKG_BUILD_DEPENDS=B and package A is selected in menuconfig package B isn't selected but needed to get package A build
Aug 28 21:55:14 <tripolar> is this a bug or am i missing something
Aug 28 21:55:48 <tripolar> the packages repo has
Aug 28 21:55:49 <tripolar> grep -R PKG_BUILD_DEPENDS|wc -l
Aug 28 21:55:49 <tripolar> 117
Aug 28 21:57:25 <tripolar> Makefiles that use it
Aug 28 22:05:47 <mkresin> carldani: xrx200 by any chance?
Aug 28 22:07:15 <mkresin> carldani: if you don't use xdsl, you have to disable dsl_control. known issue that it causes a high load and triggers the watchdog. reason unknown so far.
Aug 28 22:12:52 <jwh> hmmmm
Aug 28 22:13:34 <jwh> network reload causes interface reset :(
Aug 28 22:13:41 <jwh> caused, rather
Aug 28 22:14:09 <jwh> maybe related to setting mtu I guess, as its changed back to 1500, then what I specify after
Aug 28 22:27:34 <jwh> soooo, is there no netifd support for dummy interfaces or am I missing something?
Aug 28 22:31:20 <jwh> bird6 is really helpful and ignores addresses on loopback interfaces, sigh
Aug 28 22:38:17 <jwh> "oops"
Aug 28 23:16:54 <nyt> https://github.com/openwrt/packages/pull/6820
Aug 28 23:17:02 <nyt> would be nice to do fresh builds without having to patch every time :(
Aug 28 23:51:49 <jwh> nyt: taking a break from peering? :D
Aug 29 01:46:06 <jwh> hm, having to set kernel config options is tripping me up
**** ENDING LOGGING AT Wed Aug 29 03:00:03 2018