**** BEGIN LOGGING AT Sun Oct 13 02:59:57 2019 Oct 13 04:43:48 ldir: docker sdk image is just a debian base with unpacked sdk tarball for that particular target Oct 13 04:46:12 ldir: I'm not sure if it's possible (or even recommended) to use sdk for your use case, to me it seems like you just want to use base debian (or whatever you prefer) docker image and use that Oct 13 04:51:19 ldir: but it makes sense to use that sdk docker image as it should have all build deps already in place Oct 13 04:56:19 ldir: docker run --rm --interactive --tty --volume "$(pwd)":/home/build/my-openwrt openwrtorg/sdk /bin/bash Oct 13 04:57:15 so you would've in docker container in ~/my-openwrt your $pwd and in ~/openwrt the latest sdk (in order you want to check something under sdk as well) Oct 13 04:58:34 so lets see what is broken in docker on catalina :) Oct 13 05:00:47 don't forget, that --rm mean, that the container is going to be removed, so anything you do outside of ~/my-openwrt is going to be removed when you exit that container Oct 13 06:08:23 mangix: what wikidevi thing? Oct 13 06:09:52 "WikiDevi will be going offline 2019-10-31. For historical dumps of the database, [...]" Oct 13 06:45:23 jow: could you please create firmware-utils repo as well? thanks Oct 13 06:45:55 background: https://github.com/openwrt/openwrt/pull/2490#issuecomment-541391542 **** BEGIN LOGGING AT Sun Oct 13 06:59:50 2019 Oct 13 07:20:06 ynezz: thanks for the tips - I agree I effectively want a debian 'base' system but I didn't know how to have that with all the buildroot build dependencies already there (eg. basics such as c compiler & various libs etc). Currently cloning my build directory for testing :-) Oct 13 07:34:58 Hi ldir Oct 13 07:37:22 Tapper: depending on your needs/ wishes, ipq40xx and ipq806x are still interesting - sadly just no longer as much of a bargain as the ea6350v3 would have been Oct 13 07:37:24 Hi does any one know if I will have to redo my .config file for my wrt3200acm after this? https://github.com/openwrt/openwrt/commit/ace1bfb5541387ee4f4eca5d0372da8897023449 Oct 13 07:38:06 pkgadd ? Oct 13 07:39:13 and this https://github.com/openwrt/openwrt/commit/d2a1075973728b55231688809b68a0f0b93c9da4 Oct 13 07:39:16 Tapper: you'll have to "make oldconfig" it, but to get a proper start - it would be better to ./scripts/diffconfig.sh >/tmp/some_temporary_file - to clean that up a little further (removing implicit dependencies) and using that as new config --> to be extended via make defconfig oldconfig Oct 13 07:39:34 Tapper: you asked about a suggestiong for a ath10k device a couple of days ago Oct 13 07:39:59 pkgadd O yeah I did sorry lol Oct 13 07:41:25 I was thinking about how long my wrt3200acm will be good for now work has stoped on the wifi drivers. Oct 13 07:41:42 the output of ./scripts/diffconfig.sh is almost like config.seed, just the config symbols that differ from the defaults for your target <-- that's much easier to maintain long term/ across significant changes/ updates Oct 13 07:41:53 When master is bumped to k5.3 or somthing like that it mite stop working Oct 13 07:41:57 Tapper: you'll be the first to know ;) Oct 13 07:42:13 yeah! lol Oct 13 07:42:41 * ldir waves at tapper Oct 13 07:42:59 How gos it ldir. Oct 13 07:43:12 as sad as the situation around mwlwifi is, I don't expect the driver to be dropped that soon (as forward porting a wireless driver is possible, the deriver source is available, it just gets increasingly more difficult over time) Oct 13 07:44:03 a bigger issue will be firmware bugs and interoperability issues (e.g. fun like esp8266, with the whole lot of 802.11ax drivers we're going to see within the next 1-2 years) Oct 13 07:44:16 their is a lot of love for the wrt.xx routers, but I don't think menny people with the scills to do driver bumps have them. Oct 13 07:44:25 skills* Oct 13 07:45:40 yep, but I don't expect that to fail that soon Oct 13 07:45:57 long term, yes - go mainline or go bust Oct 13 07:46:12 short- to medium term there are likely to be solutions Oct 13 07:46:16 It's always to keep one eye on what's coming up. Oct 13 07:46:23 all ways best* Oct 13 07:46:57 Yeah go mainline or go bust is like go big or go home! lol Oct 13 07:53:10 it's qute remarkable how many issues tend to get found in new drivers, after they've been merged mainline - regardless of how 'old' or mature the driver has been out of tree before. as a rule of thumb I'd say it takes at least 3 upstream kernel versions (so e.g. v5.4 --> v5.7) until the bigger things have been fleshed out Oct 13 07:53:23 Tapper: been busy at work of late which takes openwrt out of the picture - Apple released catalina and I foolishly upgraded and broke my ability to build openwrt for 3+ days - have stumbled across a native workaround, but now trying out a docker based solution as a backup. Oct 13 07:59:54 pkgadd yeah Oct 13 08:00:15 ldir that update has messed things up for a lot of people Oct 13 08:00:44 it's nice to see that both intel (ax200 --> iwlwifi/ kernel v5.1) and QCA (ath11k, submitted, but not merged yet) are working on their drivers Oct 13 08:01:06 for their 802.11ax devices Oct 13 08:01:44 pkgadd lets hope that they have good drivers and not just big fuck off firmware blobs Oct 13 08:02:07 Tapper: to be fair the only things that broke were gpgmail (known issue, fix coming very soon) and 'building openwrt'...which is hardly a mainstream thing Oct 13 08:03:23 all my VMs, parallels, docker etc just work. Oct 13 08:05:18 By the way people are tallking on twitter Oct 13 08:05:44 you would think that apple declared war on the US lol Oct 13 08:06:15 they sort of have with their support of China ;-) Oct 13 08:06:27 hahaha! Oct 13 08:07:05 well my new build is dun see you on the flip side. lol Oct 13 08:07:16 good luck Oct 13 08:10:30 doesn't look too bad ;) Oct 13 08:11:53 OpenWrt SNAPSHOT r11212-10f5eb0398 / LuCI Master (git-19.284.27115-3688774) is flashed and running Oct 13 08:12:40 Tapper1: magic :-) Oct 13 08:13:11 I am the harry potter of routers hahahahah Oct 13 08:13:55 I'm kind of looking towards mvebu again, as I'll switch from VDSL2+vectoring (100/40 MBit/s) to ftth (400/200 MBit/s) soon (it's expected to arrive within the next 6-12 months); but I'll first have a look how my nbg6817/ ipq8065 behaves at those speeds Oct 13 08:14:36 pkgadd: do you have an openwrt running modem? Oct 13 08:16:21 If I don't like any one on the Internets I say exsplearmos and there antennas fall off there router. Pooff no wifi hahahaha. Oct 13 08:16:35 ldir: at the moment, with VDSL2 yes/no. I do have four lantiq VRX2xx devices running OpenWrt, but I'm using a Draytek Vigor 130 (lantiq VRX2xx, but the device unsupported as of now) at the moment. with ftth/ GPON, I'll get a plain modem from my ISP (the law requires that I could replace that with my own, but let's see) Oct 13 08:18:13 I would like more people to be on mvebu! Oct 13 08:19:29 pkgadd: ah the draytek - I have one of those spare in a box if my hg612 goes pop! APU2 as router, Archer C7 v2 as 'dumb AP' Oct 13 08:19:43 BT Business Hub 5 Type A, 2* O2 Box 6431 (Arcadyan VGV7510KW22), Vodafone Easybox 904 xDSL (Arcadyan VGV952CJW33-E-IR) Oct 13 08:21:05 the only reason why I'm sticking to the draytek is because it's bolted to the wall in the basement - and I already had it before I bought the others. the BThub5 is a great device, but a bit at its performance limit - and using it as a mere modem would be sad Oct 13 08:21:59 ok - I'm kinda curious as to VDSL reliability on openwrt firmware Oct 13 08:23:11 around a year ago, I tried it for a couple of weeks - flow-offloading caused some spurious reboots back then (not just on lantiq, also on ipq806x and ar71xx). other than that it was fine Oct 13 08:25:14 very good performance and latency (not using SQM, don't need it) Oct 13 08:26:00 What does this meen in my log? Sun Oct 13 09:10:25 2019 cron.err crond[2366]: time disparity of 618 minutes detected Oct 13 08:26:38 Tapper: I guess ntp updating the time Oct 13 08:26:39 It means that the clock changed... almost certainly as it got correct time from t'internet Oct 13 08:27:00 OK I will keep a glass eye on my logs. Oct 13 08:27:58 it's normal - especially for devices that don't have a real time clock that runs when device switched off/rebooting etc Oct 13 08:29:10 pkgadd: I use sqm (cake) not just as latency control but also as a fairness per host control. Oct 13 08:29:58 yep, well - it's been good enough (both) for me so far, so I didn't have a need to look into it Oct 13 08:31:04 may be different with the new ISP and ftth, we'll see Oct 13 08:31:40 https://www.thinkbroadband.com/broadband/monitoring/quality/share/3319b22c5d9770b168c9f076f5a872c92877835d-13-10-2019 Oct 13 08:31:40 and that's with a full bandwidth upstream backup going on most of the time and with various updates/browsing in the downstream. Oct 13 08:31:51 http://www.dslreports.com/speedtest/55302463 Oct 13 08:33:44 wish I had the 30 upstream :-) Oct 13 08:34:38 yep, I'm fine with my speeds - but I had to sign up for ftth now, or face paying 750 EUR setup fees if I'd decide that later Oct 13 08:35:20 they're doing the groundworks now, for free - but any house connected later will have to pay Oct 13 08:35:39 makes sense to do it now. Oct 13 08:36:23 DS-Lite (cgNAT for IPv4, plus a semi-static /56 IPv6 prefix) is the big problem though Oct 13 08:37:22 CGNATed v4 is the future. you're going to be stuck with it at some point or another Oct 13 08:38:05 hmmmm, yeah cgNAT not so hot. and depends how semi static the ipv6 is Oct 13 08:38:22 DS-Lite isn't a bad thing though -- it's a lot better than getting CGNATed v4 with no v6 Oct 13 08:38:42 what kind of convinced me in the end, was confirming that wireguard can connect via IPv6 (the strongswan android app can't) Oct 13 08:39:17 and my mobile ISP does provide a /64 prefix over LTE Oct 13 08:39:34 I'm not yet cgnat ipv4 but very dynamic! The ipv6 is relatively static, seems to be valid for a day before it changes if you don't refresh the lease. Oct 13 08:40:15 yep, I'll alsway get a new IPv4 address and IPv6 prefix after every reconnect/ router reboot Oct 13 08:40:34 while the ftth /56 prefix is supposed to be semi-static Oct 13 08:44:17 building on docker is slooooowwww Oct 13 09:12:20 but it is building (so far) so I have a fallback position :-) Oct 13 09:57:54 *yawn* Oct 13 10:05:40 is anyone else seeing appalling latency on ath10k-ct - it may also on ath10k..not tested **** BEGIN LOGGING AT Sun Oct 13 11:07:24 2019 Oct 13 11:57:06 blogic: no news about 802.11r? Oct 13 11:58:39 ldir: I'm seeing some weirdness when uploading from my laptop (intel 8265), speed dropping to ~30Mbit while wifi bitrates > 400Mbit on both ends Oct 13 11:58:45 did not look into latency though Oct 13 11:59:06 going to start with building a new image from master for my APs Oct 13 11:59:34 currently running "ancient" images Oct 13 11:59:37 well simple pings are varying wildly from 1ms to 1000+ms... that's when they're not being dropped completely Oct 13 11:59:54 built 31/08 and 01/09 Oct 13 12:00:40 it's come to something when you've switched to 2.4GHz 'cos it's better! Oct 13 12:05:50 that's no good :( used to be like that until I started using -ct Oct 13 12:08:40 ldir: 499 packets transmitted, 497 received, 0.400802% packet loss, time 1295ms Oct 13 12:08:40 rtt min/avg/max/mdev = 0.960/6.860/188.060/18.955 ms Oct 13 12:09:00 but that's with the old images still Oct 13 12:09:57 ok, you're not seeing it. I need to do some more investigating Oct 13 12:10:40 I also just wiped build_dir and staging_dir due to the gcc default version bump, so it might be a while before I get to test a newer image Oct 13 12:14:13 * ldir rebuilds the toolchain for breakfast :-) Oct 13 12:25:54 ldir: i find breakfast with bits in it too crunchy Oct 13 12:46:23 wow, it already built Oct 13 12:46:56 I can only imagine how that's going to be with a Ryzen 9 3950x :P Oct 13 12:48:18 ldir you should try building with gcc 9 I have bin using it for ages now. Oct 13 12:48:59 I am building with gcc9 - both macos native and also within a docker container Oct 13 12:49:08 stintel I upgraded from the core i920 to a amd risen 2600 and it is like nite and day Oct 13 12:49:34 * ldir remembers when the i920 was the mutts nutts Oct 13 12:49:37 :) Oct 13 12:49:37 ryzen* Oct 13 12:50:11 ldir I loved my 920 it was a gift from a friend of mine Oct 13 12:50:57 He had a computer shop and when I was moovin house he just knocked on and was like here mate I hird you need a new cpu lol Oct 13 12:51:03 I went from a Core 2 Duo E8400 to an i7 3930k, amazing upgrade, still rocking the latter machine, little over 7 years now Oct 13 12:51:24 but I'm now considering upgrading to Ryzen 9 3950x, or maybe even TR3 Oct 13 12:52:03 Yeah I went from the q6600 to the i920 to the ryzen 2600 Oct 13 12:52:28 I imagine you like it :) Oct 13 12:52:54 arf, my Unifi AP AC Pro is not coming back after flash Oct 13 12:53:02 maybe the image has become too big again :( Oct 13 12:53:15 grmbl Oct 13 12:55:15 adding a delegated consistency mode to host bind mounts on docker under macos seems to help performance - docker run --rm --interactive --tty --volume "$(pwd)/docwrt":/home/build/my-openwrt:delegated openwrtorg/sdk Oct 13 12:59:01 Sun Oct 13 10:08:11 2019 kern.err kernel: [ 10.965435] jffs2: Too few erase blocks (2) Oct 13 12:59:03 grrrr Oct 13 13:01:22 * ldir hands stintel an -ENOSPACE Oct 13 13:02:23 not even that much exotic packages added Oct 13 13:02:46 libnetsnmp seems to be to largest package Oct 13 13:05:00 aha, objdump was still enabled (dependency of perf), now I can deselect that and libbfd and some other lib Oct 13 13:05:17 should fix my problem. I can run perf on my other AP with 16MB flash I guess Oct 13 13:50:16 hi, I use OpenWrt v18.06.4 and try to get an an event when a cable is plugged in/out. I created a script in /etc/hostplug.d/iface/30-test with the line "env >> /tmp/test.log". Oct 13 13:51:07 when I unplug the cable, I get ACTION=ifup-failed Oct 13 13:51:20 when I plug the cable in, I get no event at all Oct 13 13:52:01 but the interface state (as seen with "ip a a") goes DOWN and UP, so that seems to be worjing Oct 13 13:52:05 *working Oct 13 13:52:26 Does someone has an idea on how to get an event when the interface goes up? Oct 13 13:52:53 is there anything I can try? Oct 13 14:13:40 mwarning: hotplug.d/iface Oct 13 15:07:15 blogic: that was only typo (otherwise I would have gotten not events at all) Oct 13 15:25:49 ynezz: https://git.openwrt.org/?p=project/firmware-utils.git;a=summary Oct 13 16:52:11 [ANNOUNCE] WireGuard Snapshot `0.0.20191012` Available https://lists.zx2c4.com/pipermail/wireguard/2019-October/004594.html **** BEGIN LOGGING AT Sun Oct 13 20:05:30 2019 Oct 13 21:08:25 jow: is that to move fwtool oot? Oct 14 02:19:08 updated openwrt/upstream, https://sdwalker.github.io/uscan/index.html **** ENDING LOGGING AT Mon Oct 14 02:59:57 2019