**** BEGIN LOGGING AT Mon May 04 02:59:57 2020 May 04 05:27:12 aparcar[m]: https://git.openwrt.org/?p=openwrt/staging/lynxis.git;a=commitdiff;h=a889fc5995c5d0b8c9a41b5974f5722d52cdb2c5;hp=342de5cfa1f3bd75f8ae3238959de501f302d572 May 04 05:45:20 lynxis: nice. I think you can use be version 0.9.1 which contains the fixes May 04 05:45:35 Also the tar2sqfs comment is obsolete May 04 06:57:47 lynxis: hi, BTW there are two votes ongoing (google docs & SFC), can you please vote? May 04 06:58:04 ynezz: i voted on both, didn't I? May 04 07:05:15 lynxis: I don't register your vote(s), otherwise I wouldnt bother :) May 04 07:05:35 check yourself http://lists.infradead.org/pipermail/openwrt-adm/ May 04 07:07:34 ynezz: strange. I definitely sent them for docs: Mon, 4 May 2020 01:10:47 +0200, for sfc: Mon, 27 Apr 2020 01:29:02 +0200 May 04 07:07:49 ynezz: maybe they pile up, if I'm CC them. let me check my mail logs. May 04 07:09:16 hey guys May 04 07:09:24 hi Borromini May 04 07:10:52 ynezz: maybe they land somewhere moderated? my mailserver have delivered the mail to infradead.org. my vote is both "yes" btw. May 04 07:11:24 ynezz: Hauke sad in #openwrt-adm he have the same problem May 04 07:46:31 nbd: ping. Could you add https://github.com/LorenzoBianconi/mt76/commit/e1cb717eb6c5f0f5f616edfd77670e00dcad1391 to mt76? May 04 08:03:10 Borromini: will do May 04 08:03:42 xback: pling May 04 08:07:42 lynxis: please check my latest push on the GitHub PR. It uses 0.9.1 and also adds the blocksizes which decreases the file size by like 2mb May 04 08:07:43 nbd: thanks a lot :) May 04 08:09:14 aparcar[m]: there aren't new commits. May 04 08:09:44 aparcar[m]: include/image.mk is missing the selection of the compression. May 04 08:10:46 lynxis: https://github.com/openwrt/openwrt/pull/2916/files I see it there May 04 08:11:27 aparcar[m]: yes, but where do you select the compression there? gensquashfs "-c $()" May 04 08:12:40 aparcar[m]: it doesn't match it correctly, but take a look into include/image.mk here https://github.com/openwrt/openwrt/blob/9d93d8426f54d021459feda2898a1a8d30c71894/include/image.mk#L79 May 04 08:13:52 I'm currently in Mobile so can't fix it, should I fix that up or will you do in your staging tree? May 04 08:15:09 aparcar[m]: we cannot directly take $(SQUASHFSCOMP) into it, because it doesn't support all -X options. but we have to take them. May 04 08:15:17 aparcar[m]: we have to merge some of my staging tree into your PR May 04 08:15:34 aparcar[m]: it's 10am here, have to go to bed :) May 04 08:18:37 it's 10pm here, have to go to bed ;) May 04 08:58:00 Hauke I am out for today, please sign the docs one approved via votes May 04 09:01:10 aparcar[m]: which does? May 04 09:02:09 Hauke: the document which is currently voted on May 04 09:02:39 ok May 04 09:11:58 kernel and hostapd bumps in my staging tree May 04 09:30:24 ynezz: which hostapd ? May 04 09:30:35 latest stable or git HEAD May 04 09:30:42 i have 11ax script support in my tree May 04 09:34:26 blogic: https://git.openwrt.org/97bb5987429f0d48d0d3bdb85049cdd39e65c30 hostapd: update to hostap_2_9-1113-gc54a5e96b505 May 04 09:34:57 blogic: https://git.openwrt.org/f75e45c6668735957c9d24aef78b24945e172511 hostapd: disable support for Wired Equivalent Privacy by default (in case you need WEP) May 04 09:36:08 and I didn't bothered with those mesh patches, it's PITA May 04 09:36:55 so I expect people using those to rebase them May 04 09:38:18 so far it's only compile tested (run testing on TODO), I plan to send those two patches to the list for review and push tem eventually later this week May 04 09:46:21 ynezz: I like that you removed the mesh patches May 04 09:48:43 ynezz: which new features do we get with this update? May 04 10:09:13 Hauke: tl;dr :) 1113 commits, 537 files changed, 37874 insertions(+), 14159 deletions(-) May 04 10:10:26 my main motivation is preparation for next release, start testing one of core components ASAP, since it wasnt updated for quite some time May 04 10:21:12 ok May 04 10:21:38 blogic: what is current state/plans with 11ax in OpenWrt? Do you've first device in the works? :) May 04 10:21:46 ynezz: yes May 04 10:21:51 and the 2nd May 04 10:22:08 and i am told fmac ax drivers will be out soon aswell May 04 10:22:14 something affordable for use mere mortals? May 04 10:22:21 no idea May 04 10:22:22 s/use/us/ May 04 10:22:33 the HW i have here costs liver + kidney May 04 10:22:38 :D May 04 10:25:01 Hauke: BTW any plans for next 19.07 service release? May 04 10:25:26 bots are all green May 04 10:25:32 * ldir thinks that isn't a good long term development strategy. Liver+kidney != infinite May 04 10:25:55 he didnt said it was his own May 04 10:26:28 without a liver, you're a dier May 04 10:26:54 ynezz: this is true. Hmmmm May 04 10:28:05 https://www.youtube.com/watch?v=Sp-pU8TFsg0 May 04 10:28:19 blogic: what do you mean with fmac ax driver? May 04 10:29:10 ynezz: I ordered the xiaomi ax device with qca chips, I hope I can circumvent the security May 04 10:29:17 it is sold for ~120$ May 04 10:29:50 Hauke: yeah, looks like a nice device May 04 10:29:53 good luck :) May 04 10:37:11 Ubiquiti also seems to have their first ax APs on the way: https://fccid.io/SWX-UAP6L May 04 10:44:53 I know I shouldn't be surprised by this, however I recently got facetime traffic classification working. Upstairs Mrs ldir is having a loooooong & in depth discussion over facetime with neice. I started a network backup (full on egress), then I started a full on download. Not a flicker on the 'Video' classified facetime call. May 04 10:45:24 The sqm stats luci graphing, especially on cake, showing it all 'just working'. May 04 10:49:34 Well there be any guide for usecases like that? Or will default SQM "just work" once you polish it? May 04 11:17:18 ynezz: I think we should do a 19.07.3 soon May 04 11:18:09 jow: ynezz ldir what about planning 19.07.3 for Thursday this week? May 04 11:18:24 I would like to check if we should backport some more stuff May 04 11:18:30 shouldmbe done by today May 04 11:18:42 ldir: what is the status on the security updates? May 04 11:19:39 Hauke: fine with me, just updating kernel in 19.07 to 4.14.178 May 04 11:20:04 ynezz: thanks May 04 11:20:19 I would like to update the wireless drivers to 4.19.120 in the evening May 04 11:20:27 one of the changes looks like a security update May 04 11:20:46 there's bunch of PRs with 19.07 label May 04 11:21:15 I've merged some last week, unable to decide about the others, so please review and close/reject those if you can May 04 11:22:31 ynezz: #2943 has elements that are candidates for a backport. I was waiting for master merge to prepare the backport bits, as I understand this is the expected code trajectory May 04 11:23:06 indeed May 04 11:24:25 i think xback is hands on there, I expect movement soon but if you want to release .3 in a couple days that might be a bit close call May 04 11:24:55 if someone wants something backported from master to the 19.07 branch please send a mail to the mailling list May 04 11:31:58 Hauke: umdns & relayd are done, fixed with approval from reporter. May 04 11:32:12 ldir: thanks May 04 11:33:41 Hauke: I've heard nothing further about umbim, ie reporter hasn't offered any more clues/fixes. It's an 'unusual' attack vector in that physical access is required to plug a malicious modem in. May 04 11:35:06 Hauke: it needs fixing/improving nonetheless. jow had some ideas. I would need handholding through what is required and don't have a means of testing anyway. May 04 11:36:20 ldir: do the umdns & relayd fixes need to be backported to 18.06 as well? May 04 11:37:42 zorun: If we're doing 18.06 still, then technically yes. May 04 11:38:18 there are already several fixes in openwrt-18.06 since the last release, so yes :) May 04 11:38:45 ah, ok! I'd forgotten about 18.06. May 04 11:38:53 * f00b4r0 hasn't :) May 04 11:39:36 I'll cherry-pick the relevant bits this afternoon. May 04 11:40:22 thanks! May 04 11:40:48 https://openwrt.org/docs/guide-developer/security#support_status May 04 11:42:43 ldir: do you feel like writing security advisories, using https://openwrt.org/docs/guide-developer/security#security_advisories as a model? May 04 11:44:05 I'm starting the 19.07.3 release notes May 04 12:09:37 hello! I 'd like to use SYNPROXY with OpenWRT - Source code seems to be in place (kernel & iptables), but I need someone to "school me a bit" how to add it to /include/netfilter.mk & other required files :) Is there anyone who could give me a hand? May 04 12:21:34 does anyone understand the relation between https://forum.openwrt.org/t/gcc-was-broken-on-mvebu-armada-370-device-after-commit-on-2019-03-25/43272/25 and https://git.openwrt.org/dee8fcfe9f84b584073ca28349c3c04634650744 ? May 04 12:22:00 the commit touches the tegra target while the forum post is about mvebu May 04 12:23:03 ah right, there is also a commit that fixes it on mvebu May 04 12:23:06 it makes sense now May 04 12:43:40 should be mostly good, I haven't looked at LuCI yet: https://openwrt.org/releases/19.07/notes-19.07.3 May 04 12:46:41 is there a plan to backport luci-theme-openwrt-2020 to 19.07? May 04 12:59:27 [12:22] and i am told fmac ax drivers will be out soon aswell May 04 12:59:28 blogic: can you share what hardware for? May 04 13:40:50 hello! I 'd like to use SYNPROXY with OpenWRT - Source code seems to be in place (kernel & iptables), but I need someone to "school me a bit" how to add it to /include/netfilter.mk & other required files :) Is there anyone who could give me a hand? May 04 14:55:34 Hauke: need some time for luci and opkg fixes May 04 15:29:28 jow: how long? May 04 15:29:55 or just tell me when you are finished May 04 15:54:41 oops! - lol - can someone delete the mine branch from openwrt please May 04 15:56:19 too many windows open and I believed the prompt telling me which branch I was in when I hit push May 04 16:04:14 that argp-standalone lib is an accident waiting to happen. that fix I pushed yesterday was included in glibc in 2006 May 04 16:05:43 yikes May 04 16:15:54 Hauke: I think I'll need until tomorrow night May 04 16:36:02 stintel: is it possible to update the code to match current glibc? May 04 16:36:12 jow: ok no problem May 04 16:36:15 take your time May 04 16:37:18 Hauke: I guess, but who's going to do it. I don't have the will, nor enough C experience/knowledge to maintain something like that May 04 17:21:50 stintel: I had a quick look at argp-standalone code and the current glibc code and there are big differences May 04 17:21:55 but I do not plan to udpaet this May 04 17:21:58 *update May 04 17:34:44 aparcar[m]: I accepted the "Alternative Administrator Registration" regstriation, is there more needed? May 04 17:36:01 aparcar[m]: so I assume this is now done May 04 17:36:06 lets hope we get accepted May 04 17:41:28 Hauke: I do remember a different implementation of argp May 04 17:43:45 https://github.com/xhebox/libuargp May 04 18:00:06 hello! I 'd like to use SYNPROXY with OpenWRT - Source code seems to be in place (kernel & iptables), but I need someone to "school me a bit" how to add it to /include/netfilter.mk & other required files :) Is there anyone who could give me a hand? May 04 18:01:57 RobertP: I think you might try starting and when you get a specific question about the build system or anything, just ask. May 04 18:11:21 inyeresting...this isnmaintained: https://github.com/sabotage-linux/gettext-tiny May 04 20:51:06 blogic: should this be backported to 19.07: https://git.openwrt.org/6786dc26a205da55ec2d9771693cdfb99e756e59 ? May 04 21:02:44 stintel: should your argp-standalone fix be backported to 19.07? May 04 21:07:04 Hauke: I hit this bug while packaging readsb which is not in any feeds yet afaik May 04 21:07:18 but it has the potential to segfault programs using argp-help May 04 21:08:05 there don't seem to be many packages using it May 04 21:08:06 your call May 04 21:16:26 huh, I just stubled an "interesting" problem with openssh-client in openwrt when ssh_config has "VerifyHostKeyDNS yes", DNSSEC is enabled on dnsmasq and I set manually an hostname in openwrt settings (thus dnsmasq again I think), ssh (client) segfaults May 04 21:17:01 ..when I try to ssh with openwrt device into dns that is set manually May 04 21:17:36 "debug3: verify_host_key_dns" --> "Segmentation fault" May 04 21:19:20 connecting to address that dns resolves from upstream works normally May 04 21:19:46 and also on "intranet" names when disabling said option May 04 21:20:41 idea is that few dns-names would resolve to local IP inside the NATted network, hence hostnames set within openwrt :) May 04 21:20:59 "Split DNS" I'd guess the naming of approach is May 04 21:23:18 ..and some names existing only on lan side anyways, same deal there (so problem is exactly combination of local DNS name and openssh-client) May 04 21:26:08 using ssh from another computer on LAN, having same options for SSH, ssh works fine May 04 21:26:38 "debug3: verify_host_key_dns" --> "DNS lookup error: data does not exist" and carries on asking trust etc May 04 21:33:18 build #201 of mediatek/mt7629 is complete: Failure [failed pkgbuild] Build details are at http://buildbot.openwrt.org/master/images/builders/mediatek%2Fmt7629/builds/201 blamelist: Petr ?tetiar , Hans Dedecker , Josef Schlehofer , John Crispin May 04 21:34:04 Obviously LAN computer using router as DNS-resolver and using same hostname given in openwrt for the host to be reached :) May 04 21:40:29 x86-64 master from not 2 weeks ago, to ballpark the version May 04 21:40:31 self-compiled May 04 22:21:43 aparcar[m]: what about making base64 in busybox be a standard requirement? May 04 22:21:59 frankly I’m surprised it wasn’t already. May 04 22:22:14 seems to be ubiqitous enough. May 04 23:09:11 philipp64: currently nothing needs it. I do, but I use openssl to provide that functionality (which I need for other purposes anyways) May 04 23:43:50 Tue May 5 02:43:32 2020 daemon.err seccomp-trace: /usr/sbin/umdns[10783] tried to call non-whitelisted syscall: epoll_pwait (see /etc/seccomp/umdns.json) May 04 23:44:18 epoll_wait is in the whitelist in /etc/seccomp/umdns.json May 04 23:46:04 stintel: seccomp is the breakage that keeps on breaking May 04 23:46:29 apparently. this is on a first install of umdns package May 04 23:46:47 I just commented out the seccomp param in the init script May 04 23:51:25 Well it is powerfull feature in that sense it strips powers away ;) May 05 00:11:07 The OpenBSD guys are right. seccomp is a PITA May 05 00:16:24 I'd love to use it, but last time I tried, local DNS resolution went down the drain (not quite sure if that was just about IPv6 or also IPv4, it's been a while) - sadly too much breakage May 05 00:18:29 ugh. ubuntu is awfully slow with llvmpipe May 05 01:24:56 philipp64: I think it is about 4kb extra space, not sure if we should add that **** ENDING LOGGING AT Tue May 05 02:59:57 2020