**** BEGIN LOGGING AT Mon Oct 01 03:00:02 2018 Oct 01 04:42:41 Hello all, looking to understand how exactly variables from /etc/config/wireless get to mac80211_hostapd_setup_base(). If i add a dummy variable "option foo 'bar'" why can i not use the value of it in that function? thx Oct 01 04:48:25 is it okay to have the same entry under gpio-export and gpio-led's in a .dtsi file? Oct 01 04:50:08 I can't control the LED if it's only under gpio-leds, but I can't control it from LuCI it's only under gpio-export. Looking for a good solution to let me program it, but also let the user configure it via LuCI Oct 01 05:37:14 answer to my own question: no Oct 01 06:48:40 philipp64|laptop: following what Oct 01 07:52:02 what should I be looking at when ath9k oopses on ath9k_hw_setup_statusring? https://gist.github.com/stintel/88fd5ed7e8aa148c6bfaa9a38c5cab60 Oct 01 09:03:43 stintel: databus error Oct 01 09:03:50 stintel: is this a mmio radio ? Oct 01 09:03:58 any mikrotik gurus around ? Oct 01 09:04:16 i have a helite, pass it bootfile via dhcp put it does not tftp boot anything Oct 01 09:04:58 blogic: it the radio of a QCA9531 SoC Oct 01 09:05:17 stintel: then the core is not powered or you are using a bad addr Oct 01 09:05:33 data bus error means that the SoC threw an error trying to write that memory addr Oct 01 09:05:43 normally this means a AHB/APB bus error Oct 01 09:06:02 which boils down to writing to a non powered/reset/clocked or simply invalid memory range Oct 01 09:06:17 hmm, maybe this thing doesn't even have a wifi radio Oct 01 09:06:25 lulz Oct 01 09:07:08 ok, it has Oct 01 09:09:23 the thing doesn't seem to have an art partition, so I did ath79_register_wmac(NULL, wmac); Oct 01 09:14:06 are there antennas on the unit ? Oct 01 09:14:19 that is an indicator for a radio being present Oct 01 09:15:35 blogic: yeah, there are Oct 01 09:16:11 it has a wifi radio, that was just me not being awake / not had breakfast / forcefully joined in an unplanned meeting I know jack shit about :D Oct 01 09:16:34 monday morning meetings are the best Oct 01 09:16:44 maybe caldata is in the nvram partition Oct 01 09:16:53 they indicate that someone had too much time to think on the weekend Oct 01 09:31:15 stintel: such meetings are the best Oct 01 09:31:32 just throw some random techbuzz into the round and watch the blank stares Oct 01 09:31:50 i think we should run the code on the internet fabric Oct 01 09:32:08 this will reduce the cloud latency specially when our backend server hits peak saturation Oct 01 09:32:09 yes, we should run serverless code on the fabric of the internet itself Oct 01 09:32:15 jow: :-D Oct 01 09:32:37 it'll allow us to reach IPO faster Oct 01 09:48:01 which is all that matters. certainly isn't anything like "delivering actual value to customers" or anything :) Oct 01 10:04:25 mention blockchains Oct 01 10:50:58 blogic: the caldata was in the "Factory" partition Oct 01 10:51:10 it's working now, should have figured that earlier Oct 01 10:51:20 I do way too little of this stuff :( Oct 01 10:55:35 sorry to repeat a question: Oct 01 10:55:41 looking to understand how exactly variables from /etc/config/wireless get to mac80211_hostapd_setup_base(). If i add a dummy variable "option foo 'bar'" why can i not use the value of it in that function? thx Oct 01 13:57:47 i hate our image building code Oct 01 13:57:53 its simply not usable Oct 01 13:58:05 trying to add a new board, did all the rocket science magic Oct 01 13:58:21 and am now failing to get ze c0de to generate a valid dts name Oct 01 13:58:28 it's a struggle for me every time :) Oct 01 13:58:47 its just crap Oct 01 13:58:56 oops, we are not allowed to say crap anymore Oct 01 13:59:04 lets call ti a brown nugget Oct 01 13:59:25 eh Oct 01 13:59:32 this isn't #openwrt-devel-pc is it ? :P Oct 01 14:09:06 stintel: fix it Oct 01 14:09:11 made a change, for testing Oct 01 14:09:26 that did not work so reverted the change and am at what i had 15 minutes ago and it works Oct 01 14:09:29 ftw ! Oct 01 14:09:36 pfff Oct 01 14:09:38 rm -rf tmp ? Oct 01 14:09:48 probably added a space at the right place or gawd knows Oct 01 14:22:21 12or - vs _ Oct 01 14:22:29 sometimes one's allowed, not the other, sometiems vice versa Oct 01 14:41:06 ive not been keeping up, has anything changed wrt sfp support on er-x and such? Oct 01 15:18:13 karlp: = vs := is also a good one Oct 01 15:18:30 and that combined in various versions berfore/after the include calls Oct 01 15:24:33 oh yeah, that one can be horrible Oct 01 15:28:40 blogic: I just hate anything involving Makefiles. :) Oct 01 15:29:38 * karlp hasn't seen anything yet that doesn't become equally gross in different ways. Oct 01 15:29:50 There's that. Oct 01 15:30:13 * Monkeh throws Maven at karlp Oct 01 15:30:17 at least there's no autotools int he openwrt buildchains Oct 01 15:30:45 Monkeh: yeah, that's one of the other ones that's "great at first glance, and falls apart nastily when it gets big and custom enough" Oct 01 15:31:12 I can't say it looks great at first glance Oct 01 15:31:30 It looks a little ugly at first glance and then rapidly the rotting skin sheds and you see the evil underneath Oct 01 15:52:02 oh yeah, "my way or the highway" maven Oct 01 15:52:30 anything not supported by it is simply defined as outside of the problem scope by the java elders Oct 01 15:53:23 and I love how after a few years the java application repo of a mediocre developer team consists of 90% "prepare for next development release" automated pom.xml updates Oct 01 15:54:01 jow, mkresin ever around. been trying to pester him about the archer c7 v4 LED issue for two weeks now Oct 01 15:54:04 ? Oct 01 15:55:14 he used to. Maybe he's busy with real life stuff Oct 01 16:39:37 nbd: I believe this was fixed upstream: https://github.com/openwrt/openwrt/blob/master/target/linux/generic/pending-4.14/100-MIPS-fix-cache-flushing-for-highmem-pages.patch Oct 01 16:40:01 I can't get it to crash on my mt7622 512MB device Oct 01 16:40:07 *21 Oct 01 16:49:51 I can't build master right now Oct 01 16:49:52 Package libiptc is missing dependencies for the following libraries: Oct 01 16:49:52 libip6tc.so.0 Oct 01 16:50:04 it errors there, not sure why Oct 01 17:03:15 mangix: do you have a pointer/clue to the upstream commit? Oct 01 17:26:50 blogic: can you please look at PR #1419? Oct 01 17:27:05 jow: ping Oct 01 18:15:04 ldir: let me fine it Oct 01 18:16:26 ldir: i think https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v4.19-rc6&id=55a2aa08b3af519a9693f99cdf7fa6d8b62d9f65 Oct 01 18:17:58 That commit fixes a bug introduced in kernel 4.8. The OpenWrt patch was introduced in 4.9. Oct 01 18:56:25 exit Oct 01 18:59:11 mangix: so you dropped the 100-MIPS-fix-cache patch and nothing has blown up? I can see the upstream fix for *something* mips & cache related has made it in, but I wouldn't like to say whether that fixes the highmem stuff or not Oct 01 20:26:00 ldir: hi. just wanted to say that I managed to figure things out with cake and 6rd: https://forum.openwrt.org/t/low-speeds-on-ipv6-with-sqm-cake/22403/5 Oct 01 20:46:53 huaracheguarache: great! moeller0 is better at the debugging than me. It's curious that you should be running a version of cake with the workaround...which clearly doesn't work for you. kernel 4.14.73 has the official kernel fix in it, I'd expect openwrt to have that in the next few days (I've a kernel bump in my tree, but Koen does such a good job with the bumps that I'd wait for his) Oct 01 20:47:58 ldir: I'm going to give your bump and cake patch a spin tomorrow to see if it fixes the issue Oct 01 20:50:42 cool! I'll be really interested to know the result. Oct 01 20:53:31 I'll let you know =) Oct 01 21:20:47 blogic: what is the purpose of the FUNC("-", ... lines in arch/mips/ralink/mt7620.c? git blame says i should ask you. Oct 02 00:09:21 ldir: no nothing has blown up. I can submit a patch to the mailing list. **** ENDING LOGGING AT Tue Oct 02 03:00:01 2018