**** BEGIN LOGGING AT Sat Jul 28 02:59:59 2018 Jul 28 04:39:19 hrm, in 02_network i added ubnt,rspro and it's being detected properly (i tested board_detect in functions.sh and added a "touch /tmp/is_rspro" to the case) Jul 28 04:39:26 yet it's not creating eth1.1 :/ Jul 28 04:39:51 identical to ar71xx aside from the board name Jul 28 04:41:35 i want to say it worked at some point in ath79, idk what i messed up heh Jul 28 04:51:36 eh nevermind. it's because i switched from redboot mac address to local-mac-address = [...]; i thnk. works after reverting that Jul 28 04:53:37 tmn505: btw, i get that irq domain error as well on both my ath9k cards, but i just tested 'iw dev wlan1 scan' and it worked so idk what it affects Jul 28 05:41:31 tmn505: hey i got redboot partition detection working... Jul 28 05:42:07 i only defined "rbc: partition@fff000 {" for the mac addresses Jul 28 05:42:23 then put this above it: linux,part-probe = "RedBoot"; Jul 28 05:42:50 no other code changes necessary... linux,part-probe is used by like, 2 dts files in the upstream tree heh Jul 28 05:42:56 but luckily it exists ;) Jul 28 05:44:46 tmn505: https://paste.ee/p/WWEG6#nAJaTEH2JXceTVI6OEhUc8uTBsb492y8 is my current dts Jul 28 05:47:56 KanjiMonster: ^ Jul 28 06:23:21 Guy's please give me link for download LEDE/OpenWRT on Zyxel Giga 3 Jul 28 06:25:54 Guy's please give me link for download LEDE/OpenWRT on Zyxel Giga 3 Jul 28 06:27:55 Zyxel: doesnt seem to be supported in main openwrt Jul 28 06:28:16 what kind of hardware is it? Jul 28 06:28:52 https://wikidevi.com/wiki/ZyXEL_Keenetic_Giga_III Jul 28 06:28:58 mt7621 Jul 28 06:31:27 several other keenetic models are supported Jul 28 06:33:28 It's means I can't intall LEDE or OpenWRT? Jul 28 06:37:30 Everbody thank's. Goodbay) Jul 28 06:45:28 hrm, actually pci wifi does *not* work :( no interrupts mapped per /proc/interrupts Jul 28 06:45:52 i guess it works fine to send commands to/poll but everything else is dead Jul 28 06:46:11 e.g. scan works, iw works, nothing via mon0 and no ssid is broadcast in AP mode Jul 28 06:47:25 * m4t recalls wifi not working on wndr3700v2 w/ath9k as well Jul 28 06:47:30 er, ath79 :) Jul 28 07:58:55 m4t: please don't use "linux,partprobe", it's a deprecated, non-standard way Jul 28 07:59:30 so write something that needs to be maintained in the openwrt tree? Jul 28 08:00:19 there's no other way that i could see to use the existing mtd parsers Jul 28 08:00:47 there's always the option of sending the added dts-support upstream ;p Jul 28 08:01:40 but in general having linux,part-probe work on spi-flashes is already due to a patch in the openwrt tree, upstream it only works for physically mapped flashes Jul 28 08:02:03 also where are you seeing it is deprecated? Jul 28 08:02:39 Documentation/devicetree/bindings/mtd/nand.txt doesn't mention that Jul 28 08:04:24 the fact that linux,part-probe isn't documented at all Jul 28 08:04:42 i don't understand Jul 28 08:05:04 why reinvent it? Jul 28 08:05:30 also it's a linux proprietary option, device tree files should be OS-agnostic Jul 28 08:05:58 i'll use it until something better comes along Jul 28 08:06:03 not trying to make extra work for myself Jul 28 08:06:14 remember this is only for redboot, which doesn't even support devicetree... Jul 28 08:06:29 then don't expect your work being accepted until you do it the right way Jul 28 08:06:43 what's the right way? Jul 28 08:07:22 also i don't see any other upstream partition parsers using of_* Jul 28 08:08:03 because it's basically bleeding edge, and has just been added in 4.17 or 4.18 Jul 28 08:08:41 the part-probe was rejected as a proper solution from the mtd linux maintainers Jul 28 08:11:46 i'm more worried about getting the thing booting from flash at the moment Jul 28 08:11:50 getting pci working would be nice too Jul 28 08:14:39 sure, but if you plan to submit it, you will need to use the proper parser binding Jul 28 08:14:53 KanjiMonster: also i don't see what new stuff you're talking about Jul 28 08:14:57 wouldn't it be here? https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/mtd Jul 28 08:15:02 having an example would be nice Jul 28 08:15:43 m4t: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/mtd/bcm47xxpart.c Jul 28 08:18:22 i might give a stab at it if i can get pci working Jul 28 08:22:37 [Sat 2018-07-28 01:05:29 AM PDT] also it's a linux proprietary option, device tree files should be OS-agnostic <------ commit c9c4b2116c0984eb8baeff03e806030ee9c6175f implies the contrary Jul 28 08:23:30 DonkeyHotei: which git? Jul 28 08:23:39 master Jul 28 08:24:07 I mean which git tree are you talking about Jul 28 08:24:30 openwrt Jul 28 08:24:31 linux (torvalds, mtd-next), openwrt, ... ? Jul 28 08:26:20 DonkeyHotei: yeah, I'm pretty sure that aliases thing won't fly upstream. (a better place would be the chosen node, which *is* intended for configuration) Jul 28 08:27:05 ususally these kind of things get sorted out once you start upstreaming Jul 28 08:32:17 i'm staging support for a board i don't have, and that commit made a rebase fail, so i had to convert it to the aliases thing Jul 28 08:32:52 KanjiMonster: btw, idk about upstreaming the rspro dts. there's a boatload of other stuff that'd need to get merged for the devicetree to even boot Jul 28 08:32:59 unless you mean just into the openwrt tree Jul 28 08:33:11 m4t: I mean just the openwrt tree Jul 28 08:37:51 i still don't see how it makes more sense to patch drivers/mtd/redboot.c and use that patch to probe partitions, and then having to refresh that patch between kernel bumps and fix conflicts (tho i doubt that redboot parser will be touched much...), rather than just using a single line in the .dts which works fine as-is Jul 28 08:44:40 m4t: because that patch can be sent upstream, and then it doesn't need to be refreshed anymore, and the rs (pro) dts is in a more upstream compatible format Jul 28 08:45:13 m4t: note that on vanilla linux linux,part-probe doesn't work on spi-flashes Jul 28 08:45:26 yeah i see the patch and the brcm boards using it Jul 28 08:46:37 that's because the brcm47xx target is maintained by rmilecki, who has also done most of the work upstream to find a upstream accepted solution Jul 28 08:46:50 so they are the first to use it Jul 28 08:53:43 .compatible = "brcm,bcm947xx-cfe-partitions" Jul 28 08:54:13 who would be the vendor ? ...,redboot-fis-partitions Jul 28 08:54:36 ecoscentric? Jul 28 08:56:15 Copyright (C) 2000, 2001, 2002, 2003, 2004 Red Hat, Inc. Jul 28 08:56:17 hrm Jul 28 08:56:34 (from 'version' in redboot) Jul 28 09:00:19 m4t: I'd probably use "ecos,redboot-partitions" Jul 28 09:01:44 Hi any one able to bump mwlwifi to driver version to 10.3.8.0-20180716. Jul 28 09:01:51 https://github.com/kaloz/mwlwifi/commit/bf22f642181b013ae1481f743afa9141892fac6b Jul 28 09:03:05 m4t: that is probably something cleared up when submitting the binding upstream Jul 28 09:25:54 Isn't there any interest in a honeypot system designed for OpenWrt? I put so much effort in it for my BA thesis. :/ Jul 28 09:28:43 m4t: but your proposal is also fine (ecoscentric,redboot-fis-partitions), just decide what makes most sense to you Jul 28 09:29:29 lnslbrty: do you have a link for it? Jul 28 09:29:45 KanjiMonster: https://github.com/lnslbrty/potd Jul 28 09:29:54 KanjiMonster: pull reuqest alrdy out Jul 28 09:30:02 KanjiMonster: for the owrt pkg Jul 28 09:31:01 well, pull requests can take a while to be looked at, especially with stuff that requires more specialized knowledge Jul 28 09:31:35 do you have a link to the PR? Jul 28 09:35:26 lnslbrty: ^ Jul 28 09:35:54 KanjiMonster: y, mom :D Jul 28 09:36:11 https://github.com/openwrt/packages/pull/6511 Jul 28 09:56:21 KanjiMonster: thanks for instructing m4t Jul 28 09:56:28 m4t: KanjiMonster is right Jul 28 09:56:40 we're going to drop linux,probe support in next kernels Jul 28 09:57:05 rmilecki: currently doing a build to test if dropping my patches is enough Jul 28 09:57:15 KanjiMonster: THANKS! Jul 28 09:57:33 ok thx Jul 28 10:00:51 less patches is always a plus Jul 28 10:00:57 lnslbrty: there, now happy? ;p Jul 28 10:01:35 KanjiMonster: appreciated :) Jul 28 10:19:45 KanjiMonster: fixed Jul 28 10:25:43 Are there some other C/C++/go/whatever packages that need help or some nice2have pkgs that haven't been ported yet? Jul 28 10:30:52 lnslbrty: will take a closer look later Jul 28 10:31:41 lnslbrty: also porting a package is easy (most of the time), keeping a package maintained is a lot harder; the PKG_MAINTAINER isn't just for the looks Jul 28 10:33:19 woohoo, i fixed pci-ar71xx.c on ath79 Jul 28 10:34:28 KanjiMonster: Why should it be harder to maintain an already ported pkg compared to a fresh one? Jul 28 10:34:40 KanjiMonster: don't get it Jul 28 10:36:46 ar71xx_pci_irq_init(apc); was getting called before apc->np = pdev->dev.of_node; Jul 28 10:37:40 traced it to __irq_domain_add where it was oddly returning 'unknown-1' as the domain name: [ 0.494311] irq: Added domain unknown-1 Jul 28 10:38:02 it's supposed to be the name of the dts node with /'s replaced with : Jul 28 10:38:27 blogic: ^ i'll submit a pull req Jul 28 10:44:18 lnslbrty: because you need to invest time and energy in the long term to update the package when new versions come out, react to bug reports, etc. Jul 28 10:46:21 lnslbrty: will you still care about the package in 6 months? how about a year or two? Jul 28 11:44:25 KanjiMonster: Thats what I'm doing with atop and netatop. I care about pkg's especially these ones I use on my devices. Jul 28 12:43:04 process started under procd - where does stderr go? Jul 28 12:48:33 ldir: by defualt to /dev/null unless you do procd_set_param stderr Jul 28 12:48:47 then it'll end up in logread via syslog Jul 28 12:49:23 ha - ok. I'm trying to get my head around https://github.com/openwrt/openwrt/pull/1209 Jul 28 12:50:08 and have finally figured what he's actually doing is effectively implementeing a '--quiet' option Jul 28 12:50:54 by setting a verbos flag to 3. which IMHO is just yuck. Jul 28 12:51:10 what is -d ? Jul 28 12:51:17 debug? Jul 28 12:51:36 thne verbose 3 -> -d(ebug) instead of -v(erbose) -v(erbose) kinda makes sense Jul 28 12:52:09 hm, no does not make sense at all Jul 28 12:52:12 -d is log to stderr Jul 28 12:52:16 yeah Jul 28 12:52:26 -v increases logging verbosity Jul 28 12:52:43 maybe the idea is to print to the interactive shell while restarting Jul 28 12:52:53 so he's using -d to redirect log to /dev/null to shut it up. Jul 28 12:53:04 -d also implies -n (run in foreground) Jul 28 12:53:34 so again -d is aimed at debugging Jul 28 12:54:00 that change makes no sense Jul 28 12:54:15 if you want to debug igmpproxy just start it manually Jul 28 12:54:52 he wants to turn all output off. Jul 28 12:55:16 yeah but verbose 1 -> log, verbose 2 -> more log, verbose 3 -> no log is kinda counter intuitive Jul 28 12:55:26 implement verbose 0 Jul 28 12:55:39 or option nolog Jul 28 12:55:41 so by default igmpproxy will log some ("too much") stuff to syslog. So to stop it logging to syslog you use -d Jul 28 12:55:52 or quiet 1 Jul 28 12:57:09 anyone a preference nolog v quiet. - I tend for latter as in 'set something to make the option stick' Jul 28 12:57:49 a quiet = 1 makes slightly more sense to me that nolog = 1. ie a negative option as it were. Jul 28 12:58:18 I would perform a quick "grep -r quiet package/" vs. "grep -r silent package/" vs. "grep -r nolog package/" etc. survey and use whats most ocmmon Jul 28 12:58:32 'k Jul 28 13:10:43 jow: how happy would you be with 0=no log, 1= some logging level (default), 2= verbose, 3= v verbose - the default config file doesn't spec verbose level so code currently defaults to 'some logging'. Where this potentially causes problems is if someone has specified verbose '0' rather than letter the init code go to the default. Jul 28 13:11:06 In that case they'd get no logging at all. Jul 28 13:11:34 the package config file has "# option verbose [0-2]" Jul 28 13:12:25 see I'm not sure that having 2 variables e.g. verbose and quiet/nolog/silent is *that* intuitive either :-) Jul 28 13:31:27 ldir: verbose 0..n would work well for me Jul 28 13:32:14 great. Jul 28 13:37:19 rmilecki: with the revert and my patch removed (and the linux,part-probe removed, as it seems to take precedence with your patch): https://pastebin.com/m4wnB4ik Jul 28 13:39:04 rmilecki: actually, I just saw, it seems the of stuff seems now to have lowest precedence, just saw it testing for redboot Jul 28 13:40:42 rmilecki: the full one https://pastebin.com/hdLWxVyn Jul 28 13:43:16 m4t: cool Jul 28 13:55:02 merits of changing what would be 3 ifs to a 3 way case https://pastebin.com/NiBUU2CT ? Jul 28 14:58:33 I have TP-Link VR200V, has anyone managed to get this to work with annex a adsl or is it impossible? vdsl_cpe_control --console prints "DSL_CPE: Wrong combination of DSL PHY Firmware and hybrid type used! Please change one of it." continuously Jul 28 15:12:00 ldir: looks sane but make a noop 1-case Jul 28 15:12:12 otherwise people will think that its a mistake Jul 28 15:14:11 yeah it does look a little odd. Anyway the PR guy did a new version, just using [[ $blah ]] 3 times, which I've committed, and unless there's a burning desire to change to case will leave alone :-) Jul 28 15:14:45 ok Jul 28 16:40:28 from #openwrt: Jul 28 16:40:29 [Sat 2018-07-28 09:36:46 AM PDT] Hello guys! I didn't expect so many users here. Anyway, I have created the pull request for updating package owfs (https://github.com/openwrt/packages/pull/6471) and maintainer is jow, but I see he is active, but it seems like he doesn't care about my pull request. Is there anything what can I do? Jul 28 20:56:28 KanjiMonster: so... it works? Jul 28 20:56:30 seems OK to me Jul 28 20:56:51 DT has low priority on purpose, it's a new feature and we can't break existing setups Jul 28 20:57:01 i already cuased regression like that once Jul 28 20:57:07 and my patch had to be dropped Jul 28 21:02:42 Noltari: do you have Asus AC-87U? Jul 28 21:02:54 Noltari: could you check the latest master if it boots correctly? Jul 28 21:03:24 Noltari: https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=a07730472c49c1f7bb56afa3eb8be23e6e87b4f1 Jul 28 23:46:52 bah, server crashes when on holiday suck Jul 29 02:18:29 can anybody tell me which port on the internal switch on the mt7620a is connected to the flowoffload engine? nbd? blogic? **** ENDING LOGGING AT Sun Jul 29 03:00:00 2018