**** BEGIN LOGGING AT Mon Nov 12 02:59:59 2018 Nov 12 03:11:14 hello, is it possible to create a new taboption in a LuCI interface other than general, advanced, physical or firewall? Nov 12 03:11:43 do i need to add some boilerplate somewhere first? Nov 12 03:28:48 biangbiangmian: maybe check the nlbwmon package, which adds a new tab for "Bandwidth Monitor" for a reference implementation Nov 12 05:04:27 pkgadd: perfect, thanks! Nov 12 07:53:21 aloha Nov 12 07:54:10 howdy Nov 12 08:37:04 nbd: station-mode on an mt7628n-based board: https://pastebin.com/DjPh009E Nov 12 08:37:59 on r8448-81d7f82441 Nov 12 08:40:31 interestingly, i seem to lose the console afterwards. i can often type a command (like "ip a") but it never returns any information and ^c doesn't interrupt. Nov 12 09:07:28 morning Nov 12 09:31:23 * russell-- discovers wndr3800 seems to be adequately-fully-supported in ath79 nowadays Nov 12 10:43:36 ooh I have a booting kernel on my udoo quad (imx6), just need to figure out why it gets stuck and throws binary output on the serial console Nov 12 11:06:17 stintel: oscilloscope or cheapo logic analyzer? Nov 12 11:09:05 russell--: I don't have that nor do I know what to do with it :( Nov 12 11:09:19 https://sigrok.org/wiki/Noname_Saleae_Logic_clone Nov 12 11:10:52 sigrok will do protocol decoding for you, so for example, it would help you figure out if you had a baud rate problem Nov 12 11:26:11 yeah, sounds like somethign just opened the port and started writing at a different baud rate. Nov 12 11:57:32 that udoo thing has a micro usb port with a CP210x chip attached to it, so you can do serial console just with a microusb cable Nov 12 11:57:49 but I suspect the USB driver might try to take over this port and then freezes the whole thing Nov 12 11:59:58 unplug/replug? Nov 12 12:00:54 Hi people If I build a bin for a dumb AP without dnsmasq will still beable to log on to it at 192.168.1.1 Nov 12 12:05:48 Tapper: sure, dropbear handles SSH Nov 12 12:06:02 OK thanks Nov 12 12:14:05 Tapper: not necessarily Nov 12 12:15:01 in two circumstances: a) your client device doesn't have a compatible ip addr; b) if something else also claims to be 192.168.1.1 Nov 12 12:16:09 for example, if it's just you and the dumb ap, then you'll need to set a static ip on your client device Nov 12 12:16:41 I am OK with that. :-) Nov 12 12:19:09 generally, when i'm installing a dumb-ap on a network, i usually configure its image with a static ipaddr on the network i'll be installing it on (and document that with a sticky label on the device) Nov 12 12:23:34 russell-- how do you build in a static IP to Nov 12 12:23:55 a miage Nov 12 12:24:04 a miage Nov 12 12:24:36 a image* sorry Nov 12 12:27:21 going to build with chipidea usb controller disabled and see if I can get any further Nov 12 12:27:58 I suspect the upstream DTS file is bad and nobody cares because nobody uses this old thing anymore Nov 12 12:28:30 thinks already it is one day after the release of 4.14.80 and xback hasn't bumped the kernel - this would never have happened under the n***s - a little in joke for blogic ;-) Nov 12 12:29:36 Tapper: personally, i tend to use a uci-defaults script in the files overlay, but i think there is a CONFIG option too. /me looks Nov 12 12:30:21 * ldir remembers seeing one - can't remember where either russell-- Nov 12 12:30:24 I ldir Nov 12 12:31:25 I am back trying to brake shit again :-) dont get scaired tho mate it's just lokel stuff I am messing with for now! lol Nov 12 12:32:20 Tapper: as long as you fix it after :-) It's how you learn - okay it's how I learn... but occasionally have to ask an adult :-) Nov 12 12:32:25 CONFIG_TARGET_PREINIT_IP Nov 12 12:32:33 nice thanks Nov 12 12:33:01 brb Nov 12 12:33:03 russell--: ah, had a thought.... is that the *recovery mode" address ? Nov 12 12:33:19 maybe, i've never tried to use it Nov 12 12:34:19 * ldir seeks charger - macbook about to die Nov 12 12:34:33 luckily, Tapper is here it figure it out for us ;-) Nov 12 12:36:58 yep, with USB driver disabled it's fine, so the USB driver tries to take over the micro USB port that is used as serial console port Nov 12 12:40:05 OK ldir and russell-- I will be the ginnypig Nov 12 12:45:55 git grep 192.168.1.1 turns up some more hacky possibilities Nov 12 12:46:23 for example: package/base-files/files/bin/config_generate Nov 12 12:49:00 stintel: if you're talking about a serial port like that, there's no usb involved from the udoo side remember. Nov 12 12:50:02 didn't you say it's a usb port facing your host? and just a ttl uart on the imx internally? the cp210x's usb side is to you, not the imx? Nov 12 12:50:21 yes Nov 12 12:50:28 but this board is weird Nov 12 12:50:42 it shares GPIO pins between the imx6 and the embedded arduino too Nov 12 12:50:54 are some of those the uart pins? Nov 12 12:51:01 which udoo is this? Nov 12 12:51:12 udoo quad Nov 12 12:51:31 yes, some are uart pins but the documentation .. I can't make much out of it Nov 12 12:51:57 and the only serial port in the dts is obviously the one connected to the cn6 micro usb port Nov 12 12:53:01 udoo stuff is so weird Nov 12 12:53:32 If any one is goot any free time the mwlwifi driver has had a update Nov 12 12:53:58 has got* Nov 12 12:54:10 karlp: yeah :) Nov 12 13:00:38 looking at the schematics now Nov 12 13:01:16 yah, just dfound it. Nov 12 13:01:32 the cp2102 is tied to both imx uart _and_ the sam3x uart Nov 12 13:01:55 bototm right of page 15 Nov 12 13:02:32 karlp: what document are you looking at ? Nov 12 13:02:40 maybe http://download.udoo.org/files/schematics/UDOO_REV_D_schematics.pdf is old / wrong Nov 12 13:02:51 that's the one I'm looking at. Nov 12 13:03:21 ok Nov 12 13:03:50 but that doesn't explain why the board (or console on cn6) freezes when loading the chipidea module Nov 12 13:04:19 I'd just say it's unrelated suckage Nov 12 13:04:38 there's a jumper that looks like it can choose whether cp210x usb uart goes to the imx or to the sam3x Nov 12 13:04:42 yeah Nov 12 13:06:01 j18 jumper, it is shorted Nov 12 13:06:20 if it's open you can program the arduino, if it's closed -> cp210x Nov 12 13:06:29 ah, Nov 12 13:06:39 the sam3x is connected to the usb of the imx6 as well Nov 12 13:07:05 so you might be getting things caused by the sam3x part when you bringup usb on the imx? Nov 12 13:07:09 yeah you can program it from the udoo directly Nov 12 13:07:27 J18 jumper present ==> (GPIO settings ignored) ==> CN6 routes to i.MX6 unconditionally Nov 12 13:07:40 https://www.udoo.org/forum/threads/switching-cn6-between-imx6-and-sam3x-using-software.836/#post-9880 Nov 12 13:09:09 the HDMI doesn't work either Nov 12 13:09:24 and no recent changes in linux.git Nov 12 13:37:44 hah https://www.linux-ipv6.be/UDOO/index.html Nov 12 13:38:00 I used to be running Gentoo with mainline kernel Nov 12 13:38:04 so it did work at some point Nov 12 13:38:42 well, not exactly mainline Nov 12 13:41:11 ah Nov 12 13:41:11 rmilecki: seems like the ubifs patch it taking a long time to hit the stable kernel tree, maybe we should merge a local version in the meantime? Nov 12 13:41:41 karlp: apparently https://patchwork.kernel.org/patch/3180581/ was never accepted upstream and maybe that's the problem Nov 12 13:42:02 I vaguely recall something about a board with vague usb issues Nov 12 13:42:28 karlp: https://bugzilla.kernel.org/show_bug.cgi?id=186701 Nov 12 13:42:29 right Nov 12 13:42:49 2 years later Nov 12 13:42:56 nobody even bothered to look at it Nov 12 13:45:09 doesn't make me want to look at udoo hardware... Nov 12 13:45:33 still doesn't explain garbage on console, unless that's just a sideaffect of the lockup Nov 12 13:45:53 ah that only happened a few times, other times it just locked Nov 12 13:56:42 fell int o the too hard basket of "device tree is noble and pure" and "excep I can't describe my hardware!" Nov 12 13:57:38 karlp: :D Nov 12 13:58:08 thats why you don't use noble and pure devicetree on dirty filthy hardware Nov 12 13:59:04 there's no viability on doing what we've got with brcm63xx which uses part device tree, part board file is there? Nov 12 13:59:11 that's considered totally non-viable? Nov 12 13:59:39 such bastard may never exist! its neither noble, nore pure Nov 12 13:59:40 ok and that patch that I linked in the bugreport doesn't help Nov 12 13:59:41 bah Nov 12 14:01:05 ah but I should patch the dts Nov 12 14:09:28 jow: ever given any thought to installing lua as lua/5.1 like other ssystems? Nov 12 14:09:39 would make installing lua 5.3 side by side more sane, Nov 12 14:09:52 and packages like haproxy are already pulling in lua 5.3 on their own Nov 12 14:10:21 let me put it this way...stuff is currently breaking faster than I can keep up with, so I have no desire to touch Lua atm Nov 12 14:10:35 perfectly reasonable :) Nov 12 14:10:56 however, staging a liblua53.so will likely do no harm Nov 12 14:11:36 so someone can package Lua 5.3 as lang/lua53 and let it provide at least the shared library for now Nov 12 14:12:35 was just thinking that that exact thing would be neater if it was just /usr/lib/lua/5.3, like it is on other distros, but that means moving our existing /usr/lib/lua to /usr/lib/lua/5.1 Nov 12 14:12:37 switching LuCI to 5.3 would require porting the old setfenv() stuff to the hideous _ENV thing Nov 12 14:12:48 yeah, wasn't suggesting doing any of that, Nov 12 14:13:01 I know, but it would be the sane way forward Nov 12 14:14:07 ti's probably justmy dreams that the distros are consistent anyway, Nov 12 14:14:20 they can't agree on lua-5.1 or lua51 or lua5.1 and things, so... Nov 12 14:14:48 blame upstream for doing such incompatible changes Nov 12 14:15:01 hey, at least I'm not talking baout lua 5.2 :) Nov 12 14:15:09 its like Ruby all over again Nov 12 14:15:19 each somewhat serious Ruby user bundles its complete own prefix Nov 12 14:17:36 yay, I got it to boot with USB enabled Nov 12 14:17:42 now let's see if the ports actually work Nov 12 14:18:12 haha, nope Nov 12 14:18:53 but ok, if I disable USB by patching the DTS, it can be supported by OpenWrt just without USB ports Nov 12 14:21:04 stintel: what wifi module does that udoo quad use? Nov 12 14:21:13 some usb thingie Nov 12 14:21:18 which I cannot check because no usb :D Nov 12 14:21:20 ugh Nov 12 14:21:38 I was considering moving my domoticz instance to it Nov 12 14:21:56 russell--: i don't know... Nov 12 14:22:00 let me ping Richard again Nov 12 14:22:01 but wouldn't work with my zstick Nov 12 14:22:48 soundsd like -> bin would be almost better choice... Nov 12 14:25:03 karlp: or figure out how to address the comments on https://lore.kernel.org/patchwork/patch/572699/ Nov 12 14:26:01 yeah, those were the comments that lead to my "dt can't describe my hardware!" comment Nov 12 14:26:25 email maciej, see what they did next? Nov 12 14:27:49 I guess I can do that Nov 12 14:37:30 or maybe just wait for that La Frite I ordered, evaluate it and if I like it, order another one Nov 12 14:38:03 the nice thing is that I have 2 UDOO Quad boards and so I can use one in production and one to test new openwrt images before trying to upgrade the production one Nov 12 14:38:25 what was the goal with the quads, they seem like a really niche product, aand not real cheap. Nov 12 14:38:37 is there really a board called La Frite? Nov 12 14:38:53 https://libre.computer/products/boards/aml-s805x-ac/ Nov 12 14:38:57 * karlp turns openssl deprecated apis back on again and pretends he never walked into that garbage fire Nov 12 14:39:18 karlp: haha Nov 12 14:39:24 karlp: well I had never done anything with arduino before and thought it would be nice to have a soc with embedded arduino to mess around Nov 12 14:39:38 just sounds expensive and complicated honestly :) Nov 12 14:39:52 one of them used to control my first asic miner when I still lived in Belgium Nov 12 14:40:00 I wonder how many yet undiscovered bugs have been created by openssl's decision to not provide an officially endorsed compat wrapper Nov 12 14:40:05 never really did anything with the arduino :P Nov 12 14:40:19 every project more or less does the same wrappers nowdays Nov 12 14:40:23 err_remove_state() deprecated in 1.0.0, pls use err_remove_thread_state(). Nov 12 14:40:31 got to experiment with arduino and an esp8266 later Nov 12 14:40:34 err_remove_thread_state() deprecated in 1.1, please use nothing at all. Nov 12 14:40:35 the fun part comes when you try to support libressl at the same time Nov 12 14:40:48 so you need to check versions _and_ compat api defines Nov 12 14:41:20 karlp: essentailly you need an entirely new api layer which translates to the 0.9.x, 1.0.x, 1.1.x and libressl falvors Nov 12 14:42:06 and _some of us_can choose to drop 0.9.x Nov 12 14:42:26 trust arm to not screw up mbedtls? Nov 12 14:42:32 has it been any more api stable really? Nov 12 14:42:49 at least I don't recally ever having to change it drastically Nov 12 14:50:44 feckert: ping Nov 12 14:59:23 jow: pong Nov 12 14:59:48 feckert: I just came back to the openvpn proto handler topic... do you remember where we did end up there? Nov 12 15:00:28 I think you merged this in your stanging tree put it was never merged in openwrt master Nov 12 15:01:39 I still have this on patch set but did not do any further researches Nov 12 15:01:51 ok Nov 12 15:02:35 need to fig it out then Nov 12 15:02:40 *dig Nov 12 15:05:38 If you need this i could add this to a branch of my github account Nov 12 15:05:46 found it Nov 12 15:07:56 will hopefully get around to rework it the next days Nov 12 15:08:04 and merge it Nov 12 15:09:05 jow: ok thanks Nov 12 15:44:42 when building openwrt my computer wates so long between building parts of the toolchane that it gos to sleep. Why is it doing that? Nov 12 15:45:08 heh, imx6 only supports sysupgrade on boards using NAND Nov 12 15:47:40 stintel: I'm not sure why you'd willingly use anything else with such a CPU.. :) Nov 12 15:48:30 well, memory not shitting itself on read access has upsides too... Nov 12 15:49:07 Ahh, but that's boring Nov 12 15:49:46 Tapper: dunno - try adding V=s to your make command and see where the pauses are? Nov 12 15:50:07 Tapper: its likely that it is downloadling things Nov 12 15:50:12 Or just tell your computer not to go to sleep by itself, because it clearly isn't smart enough to make its own decisions. Nov 12 15:50:26 jow: I'd never have thought there would be a genuine market for write only memory but I was wrong! Nov 12 15:51:29 It was just paused if my lan is downloading the computer will not sleep. Nov 12 15:52:12 or if the disks are working or the cpu I need to save a log some how next time Nov 12 15:52:27 Or just turn off your sleep timer. Nov 12 15:52:50 it's just when building the toolchane a standard build is mutch faster Nov 12 16:01:09 Reset cause: POR Nov 12 16:01:10 wut Nov 12 16:01:53 hehe, the udoo is in a reboot loop Nov 12 16:02:33 udont ;-) Nov 12 16:02:41 it's getting closer to the trash Nov 12 16:16:40 Sounds like it belongs there Nov 12 17:58:30 Hi, I've tried to use a TV tuner card in my Openwrt box, but then I found out that the drivers for TV cards are not built. Therefore I'd like to get them into Openwrt builds. So far I've looked at https://git.openwrt.org/?p=openwrt/openwrt.git;a=tree;f=package/kernel/linux/modules;hb=refs/heads/master and tried to read them as examples. Here's what I cobbled up thus far: https://paste.debian.net/plain/1051428 Nov 12 17:58:36 I know it's incomplete and really horrible, but I'd like some guidance to understand how this works and what would be needed to do. Nov 12 18:00:17 looks good so far, but some of the devices will require loadable firmware, too Nov 12 18:00:32 yes, I know DonkeyHotei Nov 12 18:01:14 can it be included in openwrt? It is not included in some distributions such as Fedora etc. Nov 12 18:03:05 (I mean the firmware, the kernel driver is built) Nov 12 18:13:06 fyi, 4.14.80 bump in my staging tree, currently compiling for testing Nov 12 18:13:24 ldir: xback: before you bump kernels, please review and/or give a try to the patch that I sent to avoid merge conflicts Nov 12 18:13:25 http://lists.infradead.org/pipermail/openwrt-devel/2018-November/014566.html Nov 12 18:14:18 It just moves some lines and add comments to a makefile Nov 12 18:20:43 stintel: neat Nov 12 18:21:01 talk about a minimal bump lol. Nov 12 18:25:34 hello Nov 12 18:26:05 question on cron - is it running as a default? Nov 12 18:27:01 don't think so. Nov 12 18:27:29 although... i don't remember enabling it, and it runs here, so it might be. Nov 12 18:28:13 tranq: it is included by default but only started on boot if a crontab file is present Nov 12 18:45:04 make 3 c target linux install and my dumb computer gos to sleep again! Nov 12 18:52:10 must be tired Nov 12 18:57:46 Borromini the dam thing will be in the bin if it carries on! lol Nov 12 18:59:00 * Tapper says sorry mrs tapper I need a new computer! Nov 12 19:02:41 you're sure there's no power settings interfering? Nov 12 19:03:19 if i log on to my desktop (GUI) and then decide to handle things over SSH it might go to sleep after a few hours because it thinks nothing's going on... Nov 12 19:05:21 Should I get this when flashing from ar71xx to ath79? Nov 12 19:05:23 The uploaded image file does not contain a supported format. Make sure that you choose the generic image format for your platform. Nov 12 19:07:14 it might be, what with bigger kernel partitions and all Nov 12 19:11:01 it's quite likely that you need --force when going from ar71xx to ath79 Nov 12 19:11:31 with the switch to device tree, the device names are consolidated in many cases - which have traditionally been a mess on ar71xx Nov 12 19:13:05 just be sure that you really got the correct image and the right router when using -F | --force, otherwise the results could be... interesting... Nov 12 19:32:29 yeah. Nov 12 19:33:29 Should I file a bug about my issue? I'd like to actually write the patch, but my packaging skills are quite limited as of now, I'd be happy to get help Nov 12 19:34:05 scroll back if you've missed it please Nov 12 19:42:38 scroll back how far?! Nov 12 19:42:43 file a ticket Nov 12 19:42:50 it's way more trackable than something in the irc bakclog Nov 12 19:46:42 pkgad Borromini thanks we will see lol Nov 12 19:48:12 Tapper: like pkgadd says device names can quite differ between ar71xx and ath79 so it's certainly not abnormal behaviour Nov 12 19:48:46 allizom: not everyone was here before you posted your problem, or has irc bouncers etc. Nov 12 19:50:17 Borromini: ok, I'll quote myself Nov 12 19:50:59 Hi, I've tried to use a TV tuner card in my Openwrt box, but then I found out that the drivers for TV cards are not built. Therefore I'd like to get them into Openwrt builds. So far I've looked at https://git.openwrt.org/?p=openwrt/openwrt.git;a=tree;f=package/kernel/linux/modules;hb=refs/heads/master and tried to read them as examples. Here's what I cobbled up thus far: https://paste.debian.net/plain Nov 12 19:51:07 I know it's incomplete and really horrible, but I'd like some guidance to understand how this works and what would be needed to do. Nov 12 19:54:13 allizom: the link in your message was truncated Nov 12 19:54:48 carldani: https://paste.debian.net/plain/1051428 Nov 12 20:13:33 allizom: I think filing a ticket with the work you've done so far would be the best option Nov 12 20:14:18 carldani: yes, I'm registering to file the issue Nov 12 20:20:25 allizom: if you have a github account you might open a PR in openwrt/packages Nov 12 20:20:41 https://github.com/openwrt/packages/ Nov 12 20:21:16 Borromini: ah, so not in bugs.openwrt.org, I see Nov 12 20:23:32 Borromini: are you sure it belongs there? It is about building a driver in the linux kernel, that page says "Community maintained packages for OpenWrt" Nov 12 20:24:39 it's not needed for core openwrt stuff so it will end up in packages afaik. Nov 12 20:25:01 oh, ok :) Nov 12 20:26:53 you're right to ask though, it looks like there's very few kernel packages in there, so I'm not sure. maybe it's best to enquire on the mailing list first as to where to open your PR Nov 12 20:29:17 it's not a PR yet, more of a HR (help request) Nov 12 20:29:28 But thanks, let me see where I come up with Nov 12 20:31:50 allizom: https://github.com/openwrt/openwrt/pull/1451 Nov 12 20:32:00 take a look at this Nov 12 20:33:26 oh, thanks tmn505 Nov 12 20:40:49 luaraneda: I like the idea, but would like maybe an ack from someone else before merging it Nov 12 21:59:58 stintel: yeah, sure. I was asking because if there is a kernel bump, the patch will create a merge conflict after, the very thing that is meant to avoid Nov 12 22:08:46 I'm really confused - am trying to add a new wifi interface option, something like utf8_ssid into /etc/config/wireless Nov 12 22:09:10 this is being picked up into uci - wireless.@wifi-iface[2].utf8_ssid='1' Nov 12 22:11:37 I'm trying to work out how that gets into, I presume, hostapd.sh Nov 12 22:35:46 luaraneda: yeah, unfortunately I hadn't seen your patch yet when I was bumping to 4.14.80 Nov 12 22:43:06 luaraneda: I'm going to push 4.14.80 to master now, I would suggest you hold off on respinning your patch until you got some feedback on it, to avoid that you have to respin it once again Nov 12 23:52:02 stintel: Ok. It should be the last merge conflict related to kernel versions :) **** ENDING LOGGING AT Tue Nov 13 02:59:59 2018