**** BEGIN LOGGING AT Wed May 19 02:59:57 2021 May 19 08:30:38 Build [#110](https://buildbot.openwrt.org/master/images/#builders/65/builds/110) of `archs38/generic` failed. May 19 08:42:34 are those br-lan -> lan naming changes expected on latest snaphot builds? https://gitlab.com/ynezz/openwrt-device-runtime-testing/-/jobs/1275024161 (glinet-b1300) and https://gitlab.com/ynezz/openwrt-device-runtime-testing/-/jobs/1275024180 (tplink-c7v5) May 19 08:50:34 ynezz: yes for newly generated configs May 19 08:50:42 ynezz: for sysupgrade it should not change May 19 08:52:53 so why there is still br-lan on turris omnia? May 19 08:53:27 ynezz: i don't know, can you show me /etc/board.json ? May 19 08:53:31 the busybox nslookup app change is interesting. It breaks ddns for a start. May 19 08:57:07 rmilecki: ah, it seems like mvebu testing was performed on different build/image e884389976, and tplink/glinet on b36b8b6929 May 19 08:57:22 it is possible to specify the server port BUT the syntax has changed from #port to :port - which also appears to cause fun for ipv6 server addresses. May 19 09:04:32 ddns-scripts throws " 090358 CRIT : sanitize on domain found characters outside allowed subset - TERMINATE" May 19 09:23:08 rsalvaterra: nslookup with port works here (strace nslookup example.org 1.2.3.4:123 shows connect(3, {sa_family=AF_INET, sin_port=htons(123), sin_addr=inet_addr("1.2.3.4")}, 16)) May 19 09:24:21 jow: I assumed so, I saw aparcar[m]'s commit, but it's nice to be sure. Another patch dropped! :) May 19 09:26:06 jow: does it work for ipv6 ? May 19 09:26:36 ldir: not sure, I was testing nslookup_lede, not the upstream mangled version May 19 09:27:12 denys replaced a number of things with generic libbb calls which behave slightly differently (like the # vs. : thing you already mentioned) May 19 09:27:21 ahhhhh, I understand now :-) May 19 09:27:51 for IPv6 I would've expected [fdca:1234::1]:53 to work May 19 09:28:46 I suspect if I use the correct syntax maybe it works. Forgot about the [] May 19 09:29:25 ha! yes, my mistake May 19 09:36:13 rmilecki: https://git.openwrt.org/?p=openwrt/openwrt.git;a=commitdiff;h=f716c30241d5fd9d821560f58d0af0c3ffe78600;hp=43fc720657c6e3b30c6ed89d7227ee6e646c158b - this appears to lack a uci commit ? May 19 09:41:08 jow: uci-defaults don't need commit May 19 09:41:24 ah, right forgot about it May 19 09:41:24 see /etc/init.d/boot May 19 09:42:18 # grep "uci commit" /rom/etc/uci-defaults/* | wc -l May 19 09:42:20 4 May 19 09:42:31 some uci-defaults still do commit though ;) May 19 10:14:31 aha! It's not fallout from the nslookup change - it's fallout from a ddns scripts change. Looks like a regexp fubar to me. May 19 10:22:54 [@a-zA-Z0-9._-] got changed to [@a-zA-Z0-9._-*] May 19 10:24:24 - and * strike me as 'traps' for the unwary May 19 10:25:34 jow: could we switch github LuCI repo to fasf forward merges? so we don't have all those Merge pull request commits? May 19 10:25:58 jow: most pull requests have only 1 commit i believe and those merge commits only make git log messy May 19 11:30:33 ldir: 5.10.38 out, working on it. May 19 12:21:54 rmilecki: look at the issues, this was multiple times requested and multiple denied May 19 12:22:12 rsalvaterra: ldir so no fallout from my patching? May 19 12:22:27 aparcar[m]: ah, ok, I'll check the discussion May 19 12:23:08 aparcar[m]: Don't know yet. I pulled, but I'm also bumping Tor to the lastest release, as well as the 5.10 kernel. :) Going to build now, actually. May 19 12:23:56 rmilecki: i think dwmw2_gone had the strongest points on fast forward merges May 19 12:24:36 aparcar[m]: I was asking myself why we're choosing the internal DNS resolver instead of libc's, when I found this… https://www.linkedin.com/pulse/musl-libc-alpines-greatest-weakness-rogan-lynch/ May 19 12:24:39 Oh, well… May 19 12:29:19 why is dnsmasq still doing dhcp if odhcpd is there as well? May 19 12:29:39 i think dnsmasq can do some things odhcpd can't? May 19 12:30:25 * Borromini likes odhcpd May 19 12:30:39 odhcpd doesn't handle dns :) May 19 12:31:04 and having both would needlessly increase storage requirements May 19 12:31:53 * karlp mumbles about dnsmasq having far more testing and users than odhcpd... May 19 12:32:29 karlp: That sounds like an odhcpd problem… :P May 19 12:33:03 correct.... May 19 12:39:06 stintel: true. completely forgot that. I use Unbound for that May 19 12:39:40 https://bugs.openwrt.org/index.php?do=details&task_id=2328 is still relevant May 19 12:41:10 aparcar[m]: nothing has obviously/directly exploded from the time string & use of builtin nslookup - for me at least May 19 12:41:43 aparcar[m]: the coincidental explosion I did have is an unrelated change to ddns-scripts May 19 12:57:19 Yey May 19 13:14:34 i wrote a mini tutorial on DSA config: https://forum.openwrt.org/t/mini-tutorial-for-dsa-network-config/96998 May 19 13:14:36 comments welcome May 19 13:31:21 rmilecki: Amazing work, thanks! One question, how does one specify a port PVID? May 19 13:48:43 rsalvaterra: asterisk in bridge-vlan May 19 13:49:08 rsalvaterra: config bridge-vlan; option vlan 123; list ports eth1:*; list ports eth2:* May 19 13:49:11 jow: Nice. I think it's the only thing missing from the tutorial. :) May 19 13:49:17 vid 123 is PVID for eth1 and eth2 May 19 13:56:01 Time to massage my /etc/config/network, I guess… May 19 13:58:39 lipnitsk: I just noticed the erroneous dmesg BUG we had on mt7621 is gone, at least on Linux 5.10.38. May 19 14:02:22 rmilecki: nice, some typos: "Linux*'s*", "upstream*s*", "*A* switch can be setup" May 19 14:02:52 and maybe note that the "list ports" thingy is not in -rc1? May 19 14:03:37 ime to massage my /etc/config/network, I guess… May 19 14:03:37 14:58 < rsalvaterra> lipnitsk: I just noticed the erroneous dmesg BUG we had on mt7621 is gone, at least on Linux May 19 14:03:40 5.10.38. May 19 14:03:42 14:59 -!- daregap [~belovent@cr-01.anx09.lju.si.anexia-it.com] has quit [Quit: daregap] May 19 14:03:45 oops, copy-paste fail May 19 14:03:46 Uhh… It doesn't accept a bridge with no ports…? How do I bridge two wireless interfaces? o_O May 19 14:03:46 sorry May 19 14:08:00 Oh, I might have screwed up big time. May 19 14:10:06 Of course, forgot to specify the ifname on the interface configuration, because I'm an idiot. :P May 19 14:10:24 * stintel hands rsalvaterra the noob hat May 19 14:10:41 XD May 19 14:13:38 I must say I reeeeealy like this. Separation between L2 and L3 configuration, everything explicit. May 19 14:15:21 ldir: https://github.com/openwrt/openwrt/pull/4186 ;) May 19 14:44:38 rsalvaterra: what is the best test for these kernel bumps? If it's just compiling every kmod once I think we need a CI for this May 19 14:46:33 jow: do you have an idea about the LuCI acl handling in 19.07? May 19 14:47:08 aparcar[m]: Well, compiling is nice, but nothing beats run-testing. That's what I always do, at least on my (current, I switch all the time) production device. May 19 14:48:22 rsalvaterra: ack May 19 14:48:23 aparcar[m]: Right now it's the Redmi AC2100, since there's a lot of development going on in mt76. My other devices are more boring. :) May 19 14:49:27 yea everyone seems to love mt76... May 19 14:50:49 i don't actually think txbf or mu-mimo was implemented for mt7615 though May 19 14:51:22 runs well enough for me anyway though May 19 14:53:55 i'm very happy with MT7615, would like to see DFS support for MT7613 though. May 19 14:56:27 Namidairo: Really? I wasn't aware of that… nbd, do you confirm mt76 doesn't support MU-MIMO and TxBF on MT7615 (yet?) May 19 14:58:25 at least I don't think I saw it advertised when I last looked at iw May 19 14:59:45 Namidairo: It's true iw doesn't show any beamformer/beamformee info. May 19 15:03:30 i don't think either is all that important May 19 15:03:55 maybe the mu-mimo if you have a bunch of clients hanging off it May 19 15:12:46 any perl-people around? https://github.com/openwrt/openwrt/pull/3861 May 19 15:22:16 Any reaction to the Freenode compromise? May 19 15:23:20 s/compromise/takeover May 19 15:25:21 Anyone known how to test a buzzer using kmod-gpio-beeper? May 19 15:25:33 it creates an input0 device May 19 15:26:07 hurricos: what? May 19 15:26:48 aparcar[m]: https://lwn.net/Articles/856543/ May 19 15:30:00 Uhh May 19 15:31:15 blogic: I've got pcie working on rb912 May 19 15:31:25 but it seems it only works when using kernel 5.10 for now May 19 15:31:35 xback: judging by https://manpages.debian.org/testing/beep/beep.1.en.html there should be some "trivial" printf '...' > /dev/input/input0 command May 19 15:32:31 any idea what could be missing in 5.4? dmesg shows that PCIE link stays down (the bit never changes in the register) May 19 15:32:47 I would guess that deassert fails somewhere but could not pinpoint it atm May 19 15:32:54 rsalvaterra: yeah I fixed the BUG a few weeks back upstream and it was backported to stable releases May 19 15:35:28 lipnitsk: Oh, nice, thanks! May 19 15:37:16 PaulFertser: thanks for the suggestion. but the beep command doesn't accept that input. even when specified May 19 15:55:54 xback: I think you can try running https://gist.github.com/zed/9af4ebc7136348721677a35d7fd8cc84 on your host but tell it to use stdout and redirect to that input device via ssh. May 19 16:10:51 PaulFertser: thanks May 19 16:11:14 aparcar[m] https://blog.bofh.it/debian/id_461 May 19 16:11:16 reading that it could be that pcspkr mod is also needed next to kmod-gpio-beeper May 19 16:33:28 xback: no idea May 19 16:33:44 xback: can you push the v5.4/10 patches somewhere ? i'll play around with it May 19 17:24:21 blogic: see my staging tree May 19 17:27:08 rmilecki: nbd: I think something is wrong with the netifd changes etc May 19 17:27:19 I just updated an rebuild my local build with all these updates May 19 17:27:35 and when booting from ethernet, network interfaces are not configured at all May 19 17:27:50 I'll try to check the details later this week May 19 17:28:14 before all was working fine May 19 17:36:26 https://blog.bofh.it/debian/id_461 May 19 17:36:47 looks like freenode is dead and the admin team now manages libera.chat May 19 17:37:20 should we keep using it? May 19 17:38:52 we are currently working on getting the channels registered May 19 17:39:43 https://pastebin.com/mKh8eaMM May 19 17:40:00 the mail queue is loaded, so the register call might need to be tried several times May 19 17:40:06 took me 5 tries to get it to work May 19 17:44:45 Looks like there's gong to be a big split between Libera and OFTC in this shuffle.. May 19 17:51:17 Only wished Libra would have allowed a transition of registered nicks in NickServ May 19 17:51:55 I would think that brings additional legal issues they'd rather avoid. May 19 17:52:20 To spin up quickly there's two options: No database, copy database. The latter could be problematic. May 19 17:55:43 No registration for Gmail accounts ;/ May 19 17:57:27 xback: no, I think if the speaker input device is created then no extra modules should be needed. Thatk pcspkr is implementation for x86 legacy hardware. May 19 18:30:13 xback: crap :( please try to get some more details May 19 18:38:21 Grommish: I've seen other say so too, and yet others say that it works. May 19 18:39:30 Q_: THey've probably disabled it for the huge influx.. I ended up using another one May 19 18:39:49 Q_: But it specifically said they aren't accepting from gmail.com.. Ah well May 19 18:42:09 Grommish: Someone on HN said that using gmail worked. May 19 18:42:23 likely to avoid getting classified as spammers by gmail May 19 18:42:32 are we talking about Libera? May 19 18:42:37 that's the downside if 90% of the internet ends up using the same mail provider... May 19 18:42:41 Yes. May 19 18:42:48 i regged using Protonmail May 19 18:42:57 confirmation mail ended up in spam though May 19 18:45:37 The actual downside of 90% of the internet using the same mailproviders is that those 90% decide that mail not coming from within them must be spam. May 19 18:46:14 true May 19 18:46:19 While most spam seem to come from them. May 19 18:46:23 My wife is actually losing business over that and it can only be efficiently mitigated by paying any of those 90% providers to use them with a custom domain. May 19 18:47:25 And then you have to get over yourself to handle all your business comms through Google or Microsoft. May 19 18:56:31 i told my brother to get an MS or Gmail account for his business... i set up a well configured mail server (DKIM, SPF, ...) but it would still get flagged way too often because they don't know you May 19 19:06:33 @Borromini: I've got a GSuite account I had to use. and yeah, the registration confirmation ended up in spam May 19 19:07:34 But, it came very quickly at least.. So hopefully they are settling things out May 19 19:13:07 yes, same here May 19 19:29:48 dhewg: did i miss that 's after "Linux" in some single place? May 19 19:30:27 dhewg: or should that be in every occurence: "Linux kernel subsystem", "separated Linux interface", "using Linux bridge"? May 19 19:32:25 dhewg: "It's upstream replacement" - I thought of "upstream" as adjective - I think I shouldn't need 's then, or should it? May 19 19:32:27 *should I May 19 19:54:50 rmilecki: i'm your bughunter again May 19 19:55:08 blocktrr1: what did i break? ;) May 19 19:55:13 Your netifd bridge syntax changes breaks on several mt7621 devices May 19 19:55:19 namely the U6-Lite May 19 19:55:32 blocktrr1: sorry :( please provide more details May 19 19:55:40 because it now creates a bridge named "lan" while it's sole interface is called this way May 19 19:55:54 So you have DSA port "lan" and try to create a bridge named "lan" May 19 19:56:15 universe explodes, and you don't even get network ;) May 19 19:58:56 blocktrr1: where can I find U6-Lite DTS file? May 19 19:59:51 https://github.com/openwrt/openwrt/blob/master/target/linux/ramips/dts/mt7621_ubnt_unifi-6-lite.dts May 19 20:01:01 This is the list of devices affected in ramips-mt7621: https://github.com/openwrt/openwrt/blob/master/target/linux/ramips/mt7621/base-files/etc/board.d/02_network#L34 May 19 20:01:10 blocktrr1: i see, thanks May 19 20:01:29 y expectation would've been the bridge being called br-lan May 19 20:01:30 Yikes, the nanoHD! May 19 20:01:31 Hauke: have you seen https://github.com/openwrt/openwrt/pull/3085#issuecomment-843610783 ? it would get my Acked-by because "it looks OK and I don't know any better" May 19 20:01:33 jow: nbd: https://github.com/openwrt/openwrt/blob/master/target/linux/ramips/dts/mt7621_ubnt_unifi.dtsi#L67 - some devices have "lan" interface created by DSA driver May 19 20:01:41 jow: nbd: so creating "lan" bridge fails May 19 20:01:50 jow: nbd: should we switch back to the "br-lan" ? May 19 20:02:25 rsalvaterra: any personal dislike against that? ;) May 19 20:02:46 blocktrr1: Not at all, I just need to warn a friend of mine not to update to current master. :P May 19 20:03:05 It's most likely broken anyways ATM, at least they've switched to fit images and i do not have my unit anymore May 19 20:03:19 Ah okay. Then he was early to the party ;) May 19 20:04:02 blocktrr1: Actually, he's a sysadmin, so he's a bit… conservative with upgrades. Doesn't hurt to warn him, though. :) May 19 20:09:08 blocktrr1: Personally, I'm looking forward to get my hands on the UniFi 6 LR, hopefully you guys will have sorted out the 2.5 GbE PHY configuration by then… O:) May 19 20:10:17 rsalvaterra: Is there a datasheet for the mt7621? May 19 20:11:12 I've not looked, but you're usually handy with that sort of stuff, if not I'll do my own proper research :) May 19 20:22:50 Grommish: If there's a publicly available datasheet, I've never found it… :/ May 19 20:23:13 rsalvaterra: I just ran across a not public one, so I"m happy May 19 20:23:53 I want to know if I can use the external ports on the RTL8367 to connect to the mt7621, or if it HAS to go thru port@5.. gmac1 is just sitting there ;( May 19 20:29:44 so, i'm bringing up openwrt on a Linksys EA8100 (virtually identical to the EA7300 and EA7500v2) May 19 20:29:52 (ramips) May 19 20:30:50 on 097dc943f1f9b02835c6b249b2a035679da5619f (a few commits away from current master) - kernel is fine but ubus seems to fail to start, not sure why May 19 20:31:10 [ 156.548901] procd: Started instance ubus::instance1[834] / [ 156.554687] procd: Instance ubus::instance1 exit with error code 32512 after 0 seconds / [ 156.698696] procd: Connection to ubus failed May 19 21:40:30 anyone already over on the libera thing? May 19 21:40:37 me May 19 21:43:00 me! May 19 21:48:21 given that openwrt is a SPI project, I would suggest that _if_ a move is required, oftc might be a better choice, as a SPI sister org... May 19 21:48:31 also, completely sidesteps this entire debacle May 19 21:50:42 karlp: I've seen a few projects already go that route. Makes sense, why get trapped in the politics. Also, established and fully-functional, Libera is still a bit creaky. May 19 21:51:15 Hm… are we migrating to Libera, then? May 19 21:51:16 I've been on oftc for other things for years, has been way less drama and spam than any of this, even without current shenanigans May 19 21:52:10 rsalvaterra: there are mirror channels already but i have no idea how official that is (or if that will be). I suppose that will warrant some internal discussion May 19 21:52:51 karlp: Less spam? But.. that'd be so boring May 19 21:53:17 openwrt is no longer a SPI project May 19 21:55:07 didn't we just change _to_ SPI? May 19 21:55:16 sfc afaik May 19 21:55:26 oh, we moved to SFC. May 19 21:55:29 bah humbug May 19 21:55:42 here's me thinking that was a good thing. May 19 21:55:59 thinking what? May 19 21:57:38 that SPI being over oftc was an extra +1 for it, above just the "not part of this barrel of monkeys at freenode/libera" May 19 22:03:10 is Karel Kočí here? May 19 22:18:17 aparcar[m]: No, I don't think so. May 19 22:38:53 Pepe: thanks May 19 23:05:13 I added support for a new device (https://forum.openwrt.org/t/adding-openwrt-support-for-tl-wa1201-v2/96972). Can I please get access to the wiki to add information about it? May 19 23:08:07 bluew: ping PaulFertser he can get you wiki access May 19 23:08:46 Borromini: where do I contact him? forum? May 19 23:10:07 no. i just highlighted him here May 19 23:10:55 ah I didn't realize he was in here May 19 23:16:36 is the wiki not self signup? May 19 23:17:03 karlp: it's been locked because of excessive spamming apparently May 19 23:18:06 ah, the good old days I guess :) May 19 23:20:08 of the virgin internet? :P May 20 01:44:49 Freenode drama - https://gist.github.com/joepie91/df80d8d36cd9d1bde46ba018af497409 / https://blog.bofh.it/debian/id_461 May 20 01:58:38 yeah, freenode is on the way out, Strykar. expect everyone to migrate to a different server real soon now. May 20 01:59:24 * dorf looks at the state of the wallpaper and wonders where the millions spent on freenode went. **** ENDING LOGGING AT Thu May 20 03:00:43 2021