**** BEGIN LOGGING AT Sun Apr 12 03:00:14 2020 Apr 12 03:04:14 rmilecki: I see in brcm47xx the PS-1208MFG Apr 12 03:04:43 should this be in the tree? A quick google search shows it as having 2MB flash Apr 12 03:17:25 oh interesting. there are actually quite a few devices on the legacy subtarget Apr 12 04:11:11 upgraded my tplink archer a5/c7 to 19.07.2, I noticed the UI is slower for almost all the pages, it will show "Loading view..." for each visit(i.e. not cached?) for most of the pages Apr 12 04:11:32 I did not notice this slowness with old lua-based luci though Apr 12 04:11:56 and I'm supposed to have a very powerful PC/browser running javascript comparing to tplink runs lua Apr 12 04:41:13 nginx+ssl is slower than uhttpd+ssl, but still even uhttpd+mbedtls shows the lagging Apr 12 05:28:29 build #316 of layerscape/armv8_64b is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/layerscape%2Farmv8_64b/builds/316 Apr 12 08:12:53 happy easter thing everyone Apr 12 08:14:46 rr1231: its a known bug in ustream-ssl Apr 12 10:43:21 ldir: you are the magician, what does this line mean? buildmaster-phase1_1 | /master/source.git/include/target.mk:211: CPU_TYPE "cortex-a7" doesn't correspond to a known type Apr 12 12:31:50 rr1231: have you tried uhttpd-mod-ubus ? Apr 12 13:25:52 What is the meaning of the LUCI wireless signal dbm / quality? I.e.: is it the link between the router and the phone I am using t look at luci? Apr 12 13:26:04 I have more than a client connected here Apr 12 13:28:22 I know the anser is - no; but still can't make complete sense of the values Apr 12 14:23:22 champtar: what difference will that make? I'm enabling it Apr 12 14:27:34 Hi, I'm adding a new target (rockchip) and want to create a dataentry / device for the supported devices. The target isn't listed in the target dropdown yet. Can I add targets myself or is that only editable by a wiki admin? Apr 12 14:28:14 I've created a PR for adding the HW support: https://github.com/openwrt/openwrt/pull/2917 so I hope it's OK to add the wiki pages about the devices now? Apr 12 14:29:09 (for reference, I'm talking about this page: https://openwrt.org/meta/create_new_dataentry_page ) Apr 12 14:32:59 manawyrm: i think a wiki admin. Apr 12 14:33:10 tm ping Apr 12 14:33:17 can you help out ^^ Apr 12 14:57:29 weird day Apr 12 14:58:06 its totally silent outside but the birds Apr 12 14:58:27 i live next to a natural conservation area, right infront of my office window Apr 12 14:59:05 due to the dorp in pollution i think the bird are blossoming like made Apr 12 14:59:20 and making a ridiculous noise level Apr 12 14:59:24 its pretty cool Apr 12 14:59:39 * blogic goes back to tidying up his office Apr 12 15:04:26 on average https://luci takes a few seconds to load a page, trying luci+uhttpd+openssl, luci+uhttpd, ustream-ssl is mostly a wrapper why is it causing this slow-down Apr 12 15:15:48 jow: ping Apr 12 15:16:17 jow: luci-app-ksmbd is broken because of missing uci acl Apr 12 15:19:26 ok I only installed luci(nothing luci-ssl-*), but chrome still showed a https://192.168.1.1?? and it worked, why is that Apr 12 15:22:03 is this a bug? I don't have luci+ssl enabled, just plain luci, then /etc/config/uhttpd has 'redirect_https 1' and my site visit is always https? where is the https from Apr 12 15:25:58 I disabled https and redirect in /etc/config/uhttpd and now I can use http finally, still the pageload is lagging, but it is faster than when ssl is used Apr 12 15:27:38 so two questions: 1. no luci-ssl installed but I'm still been forced and able to access https:// until I disabled ssl from /etc/config/uhttpd, that does not make sense. 2. without https/ustream-ssl the page loading still is slow Apr 12 16:01:22 rr1231: I had performance problem without mod-ubus https://bugs.openwrt.org/index.php?do=details&task_id=2613 Apr 12 16:23:09 jow: using snapshot IB fails because "Public usign key for unattended snapshot builds" is present in openwrt-keyring :( as it also gets included in base-files by the buildhost, the result is a file collision Apr 12 16:23:32 dangole: yet another clusterduck of the using/ucert fiasco Apr 12 16:23:41 imho we should nuke the lot Apr 12 16:24:10 blogic: that's a place where ucert is NOT used and that's the problem here imho Apr 12 16:24:18 dangole: did aparcar[m] catch up with you ? otherwise we should try to sync tonight or in the morning Apr 12 16:24:28 because build key should be delegated by a key in the keyring Apr 12 16:25:40 but that for now only works for sysupgrade but not for opkg Apr 12 16:26:03 imho we should nuke opkg (on target) in favour of apacar's chef :) Apr 12 16:28:28 jow: anyway, the keys for unattended build should either no longer be included in base-files or dropped from keyring Apr 12 16:49:22 champtar: interesting, in my case that did not help though Apr 12 17:59:09 You'll like this one - to improve security a manufacturer has set the default password to the mac address of the device. Apr 12 17:59:30 the device supports ipv6 Apr 12 17:59:45 well, this is harder to guess remotely comparing to admin/admin or root/root or 123456 Apr 12 17:59:49 it supports ipv6 slaac Apr 12 18:00:07 without privacy extensions. Apr 12 18:00:29 So the ipv6 address advertises the default password. Brilliant. Apr 12 18:00:46 it will take a magnifier and 30 minutes for the elder to type though Apr 12 18:53:14 ldir: this manufacturer has a name ? Apr 12 18:53:24 qnap Apr 12 18:54:46 thanks Apr 12 18:57:12 Is it just on networking gear, or also all the NAS equipements ? Apr 12 19:17:55 dangole: this must be something recent Apr 12 19:18:02 dangole: this used to work just fine in the past Apr 12 19:18:13 so whatever is broken there regressed in buildroot and must be fixed there Apr 12 19:19:08 jow: we need better CI Apr 12 19:19:56 blogic: people will invent new problems not covered by CI Apr 12 19:20:10 jow: i also only sporadically get to see that problem, but it has been there for a while, @aparcar mentioned it to me months ago Apr 12 19:21:01 jow: seems that maybe only some of the build-slaves are affected Apr 12 19:21:09 yup, yet covering the eniveitably will help. Apr 12 19:21:24 jow: ie. i downloaded snapshot IB for oxnas/ox820 a week ago and it was fine, now it is not,. Apr 12 19:22:56 dangole: not sure what I am supposed to do now Apr 12 19:23:02 dangole appears ... Apr 12 19:23:30 iirc if we drop the key from keyring it is also problematic, just differently Apr 12 19:23:56 maybe someone can tweak the base-file packaging to not copy the build key if CONFIG_BUILDBOT=y Apr 12 19:24:02 had 3 vodkias ping be on phone tomorrow but not not early Apr 12 19:36:08 jow: done. 0aa2ecf5b2 (base-files: don't ship local build key when on buildbot) Apr 12 19:37:53 thanks Apr 12 19:51:27 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html Apr 12 19:59:47 I am trying to use Comfast CF-912AC on OpenWrt 19.07 with kmod-rtl8812au-ct, but it seems that I can not get it working with AP mode. In log I have "Driver does not support authentication/association or connect commands" or am I wrong? Apr 12 20:00:27 jow: still around? Apr 12 20:02:09 greearb: Any hints? Seems you are using it. https://github.com/greearb/rtl8812AU_8821AU_linux Did you manage it to setup it as AP? Apr 12 20:12:54 aparcar[m]: yes Apr 12 20:14:37 dangole: please see this patch for signature/verification sepparation https://patchwork.ozlabs.org/patch/1207614/ Apr 12 20:15:33 jow: I looked briefly into staging tree builds and was wondering if we should introduce some kind temporary meta file which determines which targets are build Apr 12 20:15:42 say I'm working on x86 in my staging tree I don't want to build every target right? Apr 12 20:16:18 also I mentioned some fixes to your 2.4.1 patch, should I create a real patch and send it directly to you via mail? Apr 12 20:16:32 * now I get a coffee Apr 12 20:16:43 aparcar[m]: ah, so that will have to be rebased now after my !BUILDBOT patch... Apr 12 20:23:09 dangole: Yea well, I send it so long ago I'm not sure if it required anymore Apr 12 20:23:25 dangole: regarding replacing opkg, you may wonna follow here https://gitlab.alpinelinux.org/alpine/apk-tools/issues/10684 Apr 12 20:42:08 Did OpenWrt modified seccomp support stuff? Apr 12 20:42:27 Currently, proccesses that use seccomp have issues Apr 12 20:42:33 e.g. umdns Apr 12 20:42:56 Sun Apr 12 20:35:03 2020 daemon.warn procd: Seccomp support for umdns::instance1 not available Apr 12 20:44:39 nick[m]2: maybe a warning because you don't have ujail Apr 12 20:46:53 champtar: thanks, I will build new image with support and try again Apr 12 20:47:27 https://git.openwrt.org/?p=project/procd.git;a=blob;f=service/instance.c;hb=HEAD#l379 Apr 12 20:48:52 you need SECCOMP_SUPPORT, you might need to enable some kernel options Apr 12 20:52:57 but, there should be no problem on openwrt withput seccomp support running processes that "require" seccomp? Apr 12 20:57:26 have a good night guys!! See you soon! Apr 12 21:03:18 nick[m]2: this is a warning, so it should run just fine Apr 12 21:05:15 we started a conversation on whether and how to declare the differences between "supports seccomp" and "requires seccomp" and "desires seccomp" sort of thing Apr 12 21:06:33 same for all ujail features I imagine ? Apr 12 21:07:35 champtar: I check seccompo support and it was compiled in kernel. installed procd-ujail and seccmp via opkg. and restarted umdns and now it is even worse ;) Apr 12 21:08:02 Failed to parse message data Command failed: Request timed out Apr 12 21:08:31 it's securely preventing exploits... ;) Apr 12 21:09:18 aparcar[m]: uhm, mail sounds good Apr 12 21:11:59 xD Apr 12 21:13:00 jow: are you in favor of version 2.4.1 or can it also be 2.7.0? Apr 12 21:13:22 depends on what the udpate to 2.7.0 entails Apr 12 21:13:39 I'm not fond of the idea to rewrite everything again Apr 12 21:13:57 or come up with yet enother emulation of a force all scheduler Apr 12 21:14:05 I'm running it locally and it works currently Apr 12 21:14:30 ok Apr 12 21:14:39 in fact, even tho you specified 2.4.1 in the Dockerfile via pip, when runnin the buildmaster it turns out to be 2.7.0 anyway Apr 12 21:15:05 ah okay... well I'm too stupid for pip Apr 12 21:15:05 oh and we should stick with buildworker everywhere, currently it is a mix of slave and worker Apr 12 21:15:56 :) Apr 12 21:16:09 I only replaced enough "slave" with "worker" to get things going again Apr 12 21:16:20 most senseless change ever Apr 12 21:16:51 jow: don't get blogic started on this again Apr 12 21:17:48 maybe "worker" will become inacceptible too in a few years, because it symbolizes structural oppression of humans or something Apr 12 21:18:28 yea we should now come up with something futureproof Apr 12 21:18:35 !1 Apr 12 21:21:55 jow: any objections against the CDN mirror patch? Apr 12 21:27:45 jow: regarding pip, you install buildbot via apt-get from testing which happens to be 2.7.0 Apr 12 21:29:57 buildbot is only of secondary concern Apr 12 21:30:04 buildmaster is more important Apr 12 21:31:52 the pip and apt package is called buildbot, it includes the master Apr 12 21:33:57 I don't follow Apr 12 21:35:31 https://git.openwrt.org/?p=buildbot.git;a=blob;f=docker/buildmaster/Dockerfile;hb=d05f752e54bc614e4627a7ce623991e44cf80e16 Apr 12 21:35:34 -> pip Apr 12 21:36:15 https://git.openwrt.org/?p=buildbot.git;a=blob;f=docker/buildslave/Dockerfile;h=96919bdbb3dada21eb0a0e8ad564a0f7fb135f50;hb=d05f752e54bc614e4627a7ce623991e44cf80e16 Apr 12 21:36:20 -> apt, whatever debian ships Apr 12 21:36:40 buildmaster = important, buildslave = don't care Apr 12 21:38:05 26 apt-get install -t testing -y \ Apr 12 21:38:05 27 buildbot && \ Apr 12 21:38:19 this is in your buildmaster dockerfile Apr 12 21:38:27 which installs whatever debian ships Apr 12 21:39:41 ok Apr 12 21:40:21 then your question answers itself Apr 12 21:40:37 and you need to bump the pip versions to match the debian package Apr 12 21:42:58 You got a patch mail Apr 12 21:44:23 thanks Apr 12 21:44:34 there is more to come, be warned Apr 12 21:52:03 jow: really good updated setup btw, the docker-compose works nearly instantly. Maybe a default value should be activated so only a single arch is build for local testing Apr 12 21:55:01 yeah, a space separated env variable of arches maybe Apr 12 22:44:10 build #262 of layerscape/armv7 is complete: Success [build successful] Build details are at http://buildbot.openwrt.org/master/images/builders/layerscape%2Farmv7/builds/262 **** ENDING LOGGING AT Mon Apr 13 02:59:57 2020