**** BEGIN LOGGING AT Thu Apr 01 02:59:56 2021 Apr 01 08:35:07 * ldir ughs Apr 01 08:57:18 ldir: I almost hate to ask...? Apr 01 08:57:41 'mornin! Apr 01 08:58:31 'evenin! (too late) Apr 01 08:59:19 guidosarducci: I use my time zone as the reference. :) Apr 01 09:00:02 guidosarducci: oh it's 'just' morning and I'm still feeling 'a bit odd' from the covid vaccine I had last Saturday. Apr 01 09:00:22 ldir: maybe its a tumah? Apr 01 09:03:49 ldir: (ignore obscure ref) it does seem a long time to feel off though Apr 01 09:06:43 guidosarducci: oh the other half is feeling similar and many others I know have reported feeling a bit off for a surprisingly long time Apr 01 09:07:13 guidosarducci: and thanks for the email Apr 01 09:09:01 ldir: np, it was a while ago and a bit of a struggle to recall. Maybe it'll help though... Apr 01 09:09:17 ldir: which jab BTW? Apr 01 09:09:33 guidosarducci: like you, I no longer think the 'memory accounting' matches simon's hoped for 'COW' behaviour. Apr 01 09:10:40 ldir: at least you already have your vaccine :) Apr 01 09:11:27 guidosarducci: oxford astrazeneca - should have gone for the cambridge version lol Apr 01 09:12:33 ldir: still waiting for mine, so hopefully soon, but not really picky! Apr 01 09:12:50 Borromini: oh I couldn't agree more - I am very fortunate to have it. Apr 01 09:13:37 guidosarducci: cambridge version doesn't exist - it was a terrible joke on the rivalries between the great university towns of oxford, cambridge (& hull) Apr 01 09:27:12 my experience with GitHub just got even worse Apr 01 09:27:14 https://github.com/openwrt/openwrt/pull/3964#issuecomment-811731078 Apr 01 09:29:48 LMAO Apr 01 09:30:02 WTF…?! Apr 01 09:32:19 rmilecki: Some people take code criticism personally… Never understood why. Apr 01 09:32:37 i won't care too much ;) Apr 01 09:32:55 rmilecki: some people are just crazy. Apr 01 09:34:48 I like crazy people, especially the obsessive-compulsive kind, when it comes to code. :P Apr 01 09:34:49 could partly be a cultural thing too (and / or language barrier) - i mean there are many chinese ppl using and doing openwrt forks Apr 01 09:35:58 afaik sometimes "normal" "western" behaviour is too abrasive Apr 01 09:36:10 plntyk: check my comments in that code Apr 01 09:36:22 plntyk: if that wasn't being nice, I can't imagine what could Apr 01 09:36:29 *in that PR Apr 01 09:36:34 plntyk: i think in this case he has more a 'take it or leave it' mentality Apr 01 09:36:38 seen a few PR's like that Apr 01 09:36:40 Borromini: i think so Apr 01 09:36:47 'oh you won't merge my code as is? F it.' Apr 01 09:37:02 http://reviewthecode.blogspot.com/2016/01/wtf-per-minute-actual-measurement-for.html Apr 01 09:42:07 the comments were really normal - I dont know if / how much "friendlier" / "boilerplate" stuff really helps ("thank you for submitting, this seems to be your first contribution, please note ...) Apr 01 09:42:47 also NLS seems to be a little bit broken in the toolchain dunno Apr 01 09:43:34 Pro-tip: before contributing, try to understand the project culture first. :P Apr 01 09:54:34 rsalvaterra: I found the dts overlay for the er10x from Ubuntu.. It outlines the rt8367rb ethernet defines.. hmmmm Wonder if it'll work :) I had to move it to the new overlay format. ralink/overlays/ubnt_er_10x.dts: interrupt-names = "gsw", "rtl8367_int"; Apr 01 10:06:54 Grommish: Oooh, nice! Give it a spin! ;) Apr 01 10:07:18 My life is now refreshing target/subtarget kconfigs. :P Apr 01 10:07:50 Ooo Apr 01 10:08:31 Grommish: sounds like you're making progress :) Apr 01 10:08:38 Know any good ways to figure out what GPIO are for which LEDs? Apr 01 10:09:03 Borromini: I'm just avoiding trying to deal with the ebpf issue with suricata, or the python issue with suricata... Apr 01 10:09:34 and this is easier at the moment Apr 01 10:09:39 :P Apr 01 10:09:45 Grommish: i used this last time https://openwrt.org/docs/techref/hardware/port.gpio#script_gpio_in1 Apr 01 10:10:19 while trying to debug the R6800 which lost its LEDs - turned out that was in the 4.19 to 5.4 migration... the GPIO expander driver just got dropped >_> Apr 01 10:10:40 but i can confirm the script works. you need a bit of patience of course (and take notes) Apr 01 10:11:01 I'll try that.. i think I tried it and others.. but I never saw a blinky.. I'll run it again Apr 01 10:11:10 I know there are 100+ that were returned Apr 01 10:11:19 sorry - I used this one: https://openwrt.org/docs/techref/hardware/port.gpio#script_blink_all Apr 01 10:11:38 that just tries all GPIOs and needs you to jot down what blinks when Apr 01 10:11:55 it might be you need a driver as well, if the board has a gpio expander Apr 01 10:12:08 (i barely know what i'm talking about though) Apr 01 10:13:46 I'll check.. I have no idea what's on or not.. I see the lights all flash when the init happens Apr 01 10:15:22 Borromini: https://gist.github.com/Grommish/1c34c5cb2a8d994095297814ca560b47 :) time to hunt I guess Apr 01 10:26:52 Grommish: yeah, i see the same on my GS108T v3 e.g. at boot, but they don't work once in OpenWrt (at least not yet on vanilla OpenWrt) Apr 01 10:27:11 good luck! Apr 01 10:29:36 * ldir wonders if anyone else been brave enough to bump 5.10.27 Apr 01 10:31:16 big change? Apr 01 10:43:20 no not really - it's just that patches down in netfilter/iptables/bridging required manual intervention to refresh Apr 01 10:55:11 Yep. I have doubts in a condition which was added in br_switchdev, that's why I haven't bumped it. Apr 01 10:56:07 This one will require divine intervention. Apr 01 11:12:08 rsalvaterra: you mean nbd Apr 01 11:12:42 the local fdb check right? Apr 01 11:13:56 ldir: Yes and yes. ;) Apr 01 11:16:56 ldir: Not this kind of divine intervention, hopefully. ;) https://www.youtube.com/watch?v=qFhs5LCNTFY Apr 01 11:20:30 * ldir isn't quite sure what he just watched Apr 01 11:22:01 mind you, 2nd message from just re-subscribed dnsmasq mailing list was "f**k you" Apr 01 11:22:01 ldir, well its Peter Jackson (Lord of the Rings) trash "B" movies Apr 01 11:23:19 "I kick arse for the lord" is going to be stuck in the head for a little while now :-D Apr 01 11:23:56 Sam Raimi (Spider-Man) did Evil Dead - trash "B" movie too Apr 01 11:26:07 My work here is done. XD Apr 01 11:34:13 Why do some targets set HZ=250, while others set HZ=100? Any particular reason? Apr 01 11:54:40 probably old defaults from before tickless kernel etc. Apr 01 12:01:55 Timer frequency has nothing to do with tickless idle. Apr 01 12:03:08 In any case, since (I wager) most of the time is spent in the kernel, in case of OpenWrt, it probably should be a low value (I think everything supports HZ=100). Apr 01 12:14:00 nbd: can you please check https://github.com/openwrt/mt76/issues/316 Apr 01 14:26:12 I have a bunch of scripts that executes "uci set, uci commit" and reload firewalls. I'm having an issue with race conditions where two scripts that wants to add to the firewall is adding things simultanously and then f*cking up the config in such a way that I loose access to the router after a firewall reload (probably a bad config somehow) Apr 01 14:26:35 Before I begin adding some kind of lock mechanism in my own scripts, is there something already in openwrt I should use? Apr 01 14:44:13 ldir: I feel sorry for the poor bas^W person reviewing this… https://github.com/rsalvaterra/openwrt/commits/ubifs :P Apr 01 15:40:58 barhom: start with getting broken config file, it may help understanding what happens Apr 01 17:33:55 barhom: flock Apr 01 18:32:16 jow, thanks Ill check flock out Apr 01 19:19:53 Hi family Apr 01 19:21:50 I think I have a bug with netifd Apr 01 19:31:00 apparently the bridge vlans are not initialized at all on boot Apr 01 19:53:38 Hauke: now someone is reporting that all of the PHY auto-polling is not working for external PHYs and instead manual configuration is needed (similar to fixed links). is there any errata about this? Apr 01 19:54:16 Hauke: (not saying that the report is correct - just want to confirm that there's no known problem) Apr 01 20:00:46 mgiganto: the initial step in getting help is to describe the problem with details: maybe describe your device, pastebin your config and show the resulting state Apr 01 20:01:27 I have this: Apr 01 20:04:21 I have this in WRT32X: OpenWrt SNAPSHOT, r16341+5-bbb9c1c2be Apr 01 20:04:21 config interface 'lan' Apr 01 20:04:22         option proto 'static' Apr 01 20:04:22         option ipaddr '192.168.0.1' Apr 01 20:04:23         option netmask '255.255.255.0' Apr 01 20:04:23         option type 'bridge' Apr 01 20:04:24         list ifname 'lan1' Apr 01 20:04:24         list ifname 'lan2' Apr 01 20:04:25         list ifname 'lan3' Apr 01 20:04:25         list ifname 'lan4' Apr 01 20:04:26 config bridge-vlan Apr 01 20:04:26         option device 'br-lan' Apr 01 20:04:27         option vlan '2' Apr 01 20:04:27         list ports 'lan1:t' Apr 01 20:04:28 When I reboot, I expect to see vlan 2 with bridge vlan, but I don't: Apr 01 20:04:28 #root@OpenWrt:~# bridge vlan Apr 01 20:04:29 port              vlan-id Apr 01 20:04:29 lan4              1 PVID Egress Untagged Apr 01 20:12:58 mgiganto: you need to define the bridge as a 'config device' instead of implicitly creating it through the lan interface Apr 01 20:14:22 that is interesting, when I enable STP it is redefined as config device, but then the flag STP dissapear and i no longer have access to select bridged ports Apr 01 20:16:26 and if I create first a bridge with the ports and then an interface of the bridge, was kind of working, but i lost the ability to see the connected ports and statistics of each port in luci, i just see bridge Apr 01 20:18:33 I also use the bridge name itself, instead of bridge.1. I don't know if it is a big difference Apr 01 20:20:55 mgiganto: in future, use a pastebin site to share more than a few lines, via a url Apr 01 20:21:15 sorry about that Apr 01 20:33:55 xdarklight: I haven't found the errata sheet Apr 01 20:34:40 but this is strange, I thought that the PHY polling using the MDIO bus and there is one MDIO bus with the external and the internal phys Apr 01 20:34:50 is the pin configuration correct? Apr 01 20:35:01 in the pin ctrl for the MDIO bus Apr 01 20:35:26 is this person able to manually ready data from the PHY Apr 01 20:36:03 coudl also be that it is broken and only works for the internal phys becasue we use the default avlues Apr 01 20:36:05 *values Apr 01 20:44:20 Hauke: ok, thanks for checking. for the user who reported that setting the link parameters manually: I don't know what he's seeing from the PHY. the other user who reported issues is Martin Schiller and he reports (that one one of his boards Ethernet is not working at all) "Another point that points to a missing initialization of the phys" Apr 01 20:45:11 I'll need to investigate this in the next days in more detail Apr 01 22:11:31 What functions from the init.d script get called when a "stop" happens? Apr 01 23:37:08 Hmm anyone have any insight on how to check for installed opencl-headers? Apr 01 23:37:53 it seems mesa just syncs them to the latest version Apr 02 01:02:18 russell--: Hey Russell, could you take a look at an NLS fix in iproute2 please? https://github.com/openwrt/openwrt/pull/4038. Thanks! Apr 02 01:17:52 ldir: great, two pointless responses on the dnsmasq list. One, a straight up "FU", and the other mostly ignoring my email to make irrelevant pedantic points demonstrating how smart they are. That sort of noise just kills discussion. :-/ Apr 02 01:49:46 guidosarducci: yeah, i've been busy (my build tree has been tied up doing deployment compiles), i'll try to look at it tonight. can you remind me of what testing you've done? Apr 02 02:09:13 russell--: I've built and run tested the 3 NLS-sensitive packages with interdependencies: binutils, bpftools and iproute2, using both musl and glibc as libc. Apr 02 02:10:01 russell--: this was on QEMU/malta/be32 with some run-time regression testing as well. **** ENDING LOGGING AT Fri Apr 02 02:59:57 2021