**** BEGIN LOGGING AT Wed May 26 03:00:34 2021 May 26 08:36:30 netifd changes pushed to the 21.02 May 26 08:38:33 rmilecki: I've seen some reports about the luci migration breaking connectivity... May 26 08:38:50 jow: can you link me to that? May 26 08:38:57 did you push the luci change before netifd was updated or could that be users upgrading luci-mod-network (or the feed) without updating netifd? May 26 08:39:13 jow: first i pushed netifd, then LuCI May 26 08:39:29 I was thinking about checking whether netifd is new enough before triggering the migration logic May 26 08:39:42 because it is quite ocmmon that users upgrade luci components without upgrading base May 26 08:39:50 i see May 26 08:40:05 jow: is there some existing code checking for package version already somewhere? May 26 08:40:13 so i could copy & paste May 26 08:40:57 nope May 26 08:41:14 but to keep it lean I thought about fetching /usr/lib/opkg/info/netifd.control May 26 08:41:24 and checking the Version: line in there May 26 08:41:31 calling anything opkg is far too heavy May 26 08:41:54 the above may fail for users stripping opkg metadata, but thats fine, such users are unlikely to need any migration May 26 08:55:02 rmilecki: http://sprunge.us/90oe4E May 26 08:55:05 rmilecki: untested May 26 09:28:15 Hauke, jow, aparcar[m], etc.… What in blazes happened to mangix's musl 1.2.2 bump? He just gave up and deleted the whole branch. :( May 26 09:29:35 https://github.com/openwrt/openwrt/pull/3004#issuecomment-841758942 May 26 09:30:28 rsalvaterra: uhm, no idea. I didn't follow OpenWrt development during the last three weks or so May 26 09:34:05 jow: Would you like to me take over and shepherd it until it's merged? I *really* want to get this in (and mangix told me to resubmit it)… May 26 09:38:05 But, to be honest, I don't know what else is needed. I've been running musl 1.2.2 on all my systems since early March, without any issues. May 26 09:39:23 rsalvaterra: sure why not. Maybe noone wanted to merge the musl bump pre-branch May 26 09:39:47 also the amount of people willing to merge toolchain changes in relatively small in general May 26 09:40:01 lots of regression potential. comparatively high effort to track and fix regressions May 26 09:40:31 jow: Hmm… so now it would be the perfect time to merge, since 21.02 has just been branched. May 26 09:40:51 Lots of time to fix any possible fallout. May 26 09:40:53 the perfect time to merge is when 21.02 is fianlized May 26 09:41:19 otherwise backports become unnecessarily complicated May 26 09:41:26 jow: Fair enough. May 26 09:42:12 I'll wait until 21.02 final to resend. May 26 09:43:22 thank you! May 26 09:44:49 Don't mention it. :) May 26 09:44:49 jow: did you see my request for help on LuCI 21.02 branch? May 26 09:44:50 could one day be a bit harder about backports, instead of treating the branch point as a suggestion too :) May 26 09:55:15 rmilecki: nope May 26 09:56:38 jow: my request is: can you backport important stuff to the LuCI 21.02 branch from its master branch? May 26 09:57:38 i'm sure we need faad7464a8ed88a251c621dc92e8cbfb146bcd7f ("luci-mod-network: add support for network.device sections") and all later fixes May 26 09:57:59 i'm not sure what else & also what patches are exactl fixes for above May 26 09:58:13 will check later May 26 09:58:17 i could try cherry picking all luci-mod-network changes, but I'll get a bigger chance of breaking sth than you May 26 09:58:21 jow: thank you May 26 10:01:55 can I at least request that this time we don't fundamentally change it's behaviour ont he branch, after the first release? :) May 26 10:23:58 jow: there is one big patch, the last step if my network cleanup, that i'd appreciate you looking at: [PATCH luci] treewide: use "device" option in UCI "interface" sections https://patchwork.ozlabs.org/project/openwrt/patch/20210524161741.17920-1-zajec5@gmail.com/ May 26 11:06:17 * enyc meows May 26 11:09:28 jow: your diff http://sprunge.us/90oe4E works great, can you commit that? May 26 11:09:40 jow: i could but it's missing Signed-off-by May 26 11:24:09 please find a new irc home, the recent reports about freenode really want me to rm this connection May 26 12:00:02 rmilecki: I wasn't sure about the source date stamp to use May 26 12:00:19 rmilecki: feel free to push with a suggested-by of mine May 26 12:00:29 jow: thanks May 26 12:13:00 Hi any one seen CVE-2021-27821. May 26 12:13:06 https://www.sesin.at/2021/05/25/openwrt-luci-up-to-19-07-web-interface-cross-site-scripting/ May 26 12:14:37 man that sesin.at is link farming. May 26 12:15:42 I love the "links" section on https://nvd.nist.gov/vuln/detail/CVE-2021-27821 too. May 26 12:15:46 100% professional May 26 12:16:55 karlp I know nuthing about it mate. Just got posted to me on the OpenWrt twitter. May 26 12:24:30 can't open that page here May 26 12:44:43 sesin.at is just SEO. May 26 12:48:56 rsalvaterra: I looked into the gcc 10 update first and there is still a problem in umbim May 26 12:56:38 Hauke: Unrelated to this? https://github.com/rsalvaterra/openwrt/commit/fe2f19e680197b8262df226a9412d06988be72c9 May 26 13:23:41 Crap, isl.gforge.inria.fr is down, can't download tools/isl. May 26 14:35:53 nbd: Wed May 26 14:34:59 2021 daemon.notice hostapd: handle_probe_req: send failed May 26 14:36:36 nbd: Just built a new image and my log is flooded with these. May 26 14:36:58 nbd: Only my unencrypted Wi-Fi network came up. May 26 14:47:59 So, uh… yeah, that latest hostapd changes killed wireless for me. May 26 14:49:34 rsalvaterra: there should be mirrors, right? May 26 14:50:43 zorun: For tools/isl? Heh… I just disabled GRAPHITE. I wasn't building with polyhedral transforms anyway. May 26 14:51:21 The hostapd breakage is… a bit more urgent, I wager. May 26 14:58:35 zorun: So surprisingly, removing mostly just device target items changes the vermagic. Here is a diff of two diffconfigs: https://paste.c-net.org/FreshmenBagpipes May 26 15:00:05 I'm going to continue to bisect these options and see if I get something more interesting. Maybe CONFIG_IB is touching some of the options too. May 26 15:00:35 interesting... May 26 15:00:50 but in any case, you should indeed distribute kmods together with your builds May 26 15:01:40 zorun: Yes, it's true, it's true. And thank you very much for the info on ALL_KMODS, that's much faster / less space-consuming to build. May 26 15:07:17 nbd: I narrowed it down to sae-mixed. If I downgrade to psk2, the same configuration works. WPA3 is broken. May 26 15:19:02 hello! if I submit an update for packages/net/dnsdist to version 1.6.0, when will users get it? on the next major openwrt release? earlier? May 26 15:22:13 Habbie: for snapshots and the next release by default May 26 15:22:18 ok May 26 15:22:31 Habbie: if that is something important (crash fix / security fix) we may backport it for existing release May 26 15:22:39 thanks, it's not May 26 16:08:58 https://openwrt.org/contact#irc_channels needs to be updated May 26 17:19:59 rsalvaterra: yes this change May 26 17:20:13 I would perfer ifit wokrs without ignoring the warning May 26 17:21:33 jow: when will you have time to backport the LuCi changes to 21.02 May 26 17:22:32 If it is only in a week or so, I would like to do an 21.02.0-rc2 today to get the other fixes tested and then we can do an 21.02.0-rc3 later May 26 17:31:58 rsalvaterra: is this with mt7615? May 26 17:32:17 nbd: MT7603 and MT7615 both. May 26 17:32:38 Hauke: I'll see what I can come up with. :) May 26 17:36:54 Hauke: However, I don't have MBIM hardware for testing (only NCM). May 26 17:37:24 rsalvaterra: it looks to me like the reason is this line: May 26 17:37:25 wpa_key_mgmt=WPA-PSK WPA-PSK-SHA256 SAE May 26 17:37:46 log says using more than 2 AKM suites is refused by nl80211 May 26 17:38:27 nbd: So… No WPA/WPA2/WPA3, right? Which means sae-mixed should only support WPA2/WPA3. May 26 17:40:34 hm, if i get rid of that error, something else fails May 26 17:40:47 I'd love to just drop WPA2 and move to WPA3-only, buy my cell phone isn't having it… :) May 26 17:41:01 don't worry, i'll get it fixed soon May 26 17:41:42 ndb: No worries, I just wanted to give you the heads-up. Thanks! May 26 17:41:53 thanks for reporting this May 26 17:50:30 Hauke: do the rc today May 26 17:50:36 not sure when I get around to it May 26 17:51:12 jow: Hauke: let me see those network commits May 26 17:54:53 git log --oneline --no-merges origin/openwrt-21.02..origin/master modules/luci-mod-network/ May 26 17:55:01 30% of commits (the recent ones) are mine May 26 17:55:31 jow: can you have a quick look at above git log, see if I could just try cherry picking all those changes? May 26 17:56:26 jow: here is above git log output if you prefer: https://pastebin.com/xWM4rxcp May 26 17:57:19 they should be all backported May 26 17:57:28 but they will introduce unsolved regressions May 26 17:59:54 specifically https://github.com/openwrt/luci/issues/5009 May 26 18:10:05 jow: ok May 26 18:10:18 I have no problem do do an rc3 one week after the rc2 May 26 19:08:49 rsalvaterra: please try the latest version, i just pushed a fix May 26 20:45:31 Does anyone know if there's an existing firmware image that natively resolves Handshake domains (https://handshake.org) -- e.g., by bundling hnsd and using it to resolve names? May 26 21:12:09 gateway: https://github.com/handshake-org/hsd May 26 21:12:17 Javascript? May 26 21:13:03 I see, you want this: https://github.com/handshake-org/hnsd May 26 21:13:38 check out https://openwrt.org/docs/guide-developer/packages May 26 21:16:41 And then create an image with that package preinstalled? May 26 21:17:27 I'm not just looking to install it locally for myself... I'm looking to make it incredibly easy for any OpenWrt user to use it May 26 21:17:28 first create the package; then, it is very easy to create such an image, yes May 26 21:17:37 creating the package will make it easy May 26 21:17:42 creating images might be more than some users want May 26 21:18:37 cool... I'll read up on the packages... thanks for the help May 26 21:19:28 np, good luck May 26 21:28:39 is it true channels are getting delisted here for mentioning libera.chat? May 26 21:31:24 dorf: for mentioning in the topic, yes, the bot renames them to ##, gains control and changes the topic on its own. May 26 21:31:44 PaulFertser: cute May 26 21:32:17 won't be long before this network's Lee and his buddies alone. May 26 21:35:03 have we settled on a new network yet? oftc seems to be the choice? May 26 21:35:43 dorf: yes we are moving to oftc May 26 21:35:47 see the mailing list May 26 21:37:20 Hauke: yeah, seen that, just hadn't reviewed it in the last 24 hours. May 26 21:50:15 Hauke: give me one day (tomorrow0 to get LuCI into shape for 21.02 May 26 21:50:20 i'll start early morning May 26 21:50:45 it's better to hava pptp regression and test the rest May 26 21:50:57 rather than not having basic network setup tested at all May 26 21:51:38 rmilecki: ok fine with me May 26 21:51:48 thanks May 26 21:51:55 tomorrow evening I do not have time to do the release May 26 21:52:01 then it will be friday May 26 21:52:14 sounds ok May 26 21:55:50 does the direction of LUCI ever get brainstormed? **** BEGIN LOGGING AT Wed May 26 22:03:25 2021 May 26 23:53:13 nbd: Thanks! It's a bit late for that now (and I was afk for the evening), but will do so tomorrow. **** ENDING LOGGING AT Thu May 27 02:59:56 2021