**** BEGIN LOGGING AT Mon Oct 29 03:00:00 2018 Oct 29 03:04:24 the drivers are a mess and out of tree. there are a few developers working on cleaning up the code to be less buggy. he who shall not be named does not approve of their efforts because he must clean up after them. Oct 29 03:04:49 a few significant bugs got backported but still. Oct 29 03:05:53 still not fixed [Sun Oct 28 14:02:13 2018] NETDEV WATCHDOG: eth0 (mtk_soc_eth): transmit queue 0 timed out Oct 29 03:06:02 the I/O on all of these devices is slow because of the driver. kernel folks fixed that issue. I think with 4.19 their code will start being used instead of the openwrt specific stuff. we'll see. Oct 29 03:09:06 sounds like a blast, can't blame you for leaving Oct 29 03:09:50 well, more specifically, my router is ARM based Oct 29 03:10:19 I tried putting in an mt76 minipcie card in it but it didn't work Oct 29 03:10:27 it probably still doesn't Oct 29 03:10:38 so atheros it is Oct 29 03:11:17 do you have any other parting words for those of us who are still stuck on the platform? Oct 29 03:12:07 erm, it's a mess currently. it will get better in about 6 months Oct 29 03:12:42 i'm going through the DS for MT7621 and setting up my DTS file as best I can given that i can't find a lot of documentation about dts Oct 29 03:13:28 i've been able to set up gpio according to my required sharing scheme, but i guess as you say it's more to do with the mt76 drivers than any kind of registers we could possibly set, right? Oct 29 03:14:10 probably. the kernel folks did a lot of modifications to the GPIO driver Oct 29 03:14:28 you could try backporting to openwrt Oct 29 03:14:39 it's just a patch? Oct 29 03:14:56 yes and no Oct 29 03:15:42 lol, i'm about to get rekt, aren't i? Oct 29 03:16:19 oh....... Oct 29 03:16:29 the gpio driver is out of staging Oct 29 03:17:36 so the commit adding it to the kernel is this: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/drivers/gpio/gpio-mt7621.c?h=next-20181019&id=4ba9c3afda41213ec98c30053e32963892e6dc7c Oct 29 03:17:49 the openwrt patch is https://github.com/openwrt/openwrt/blob/master/target/linux/ramips/patches-4.14/0028-GPIO-ralink-add-mt7621-gpio-controller.patch Oct 29 03:18:17 i THINK you can just replace it. Oct 29 03:18:36 do note that the upstream driver additionally requires DTS changes Oct 29 03:19:21 https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/log/drivers/staging/mt7621-dts?h=next-20181019 Oct 29 03:19:27 note the GPIO changes Oct 29 03:32:00 sussing that all out atm Oct 29 03:49:37 sorry i don't think i understand -- the patch is applied to openwrt back in februrary, and to use the new driver we have to make changes in our device tree files? Oct 29 03:58:05 no Oct 29 03:58:17 the openwrt patch and the kernel patch are different Oct 29 03:58:58 the dts changes are needed in addition to replacing the openwrt driver Oct 29 04:19:02 ok Oct 29 04:19:23 side note: know of any good reading material on DTS? Oct 29 04:19:39 or supporting your own devices post-dts kernel? Oct 29 04:29:41 ehhh not really Oct 29 04:30:08 what does the openwrt patch do? Oct 29 04:30:17 if it's different than the kernel patch Oct 29 04:38:28 nvm Oct 29 05:15:25 Does this multicast(?) problem sound familiar? I have several routers connected by various Ethernet & WiFi. I set up quagga and rip so i wouldn't need static routes. Rip packets are being sent to a multicast address. Oct 29 05:15:56 However, they are only received if I have a tcpdump process going to look for them (I think it's because tcpdump puts the interface in promiscuous mode). Oct 29 05:16:22 If tcpdump is not running, the route expires and gets removed, which I think means that the local ripd did not see the announcement. Oct 29 05:34:51 spacewrench: that used to be relatively common, ages ago when very few things used multicast and it could easily sit broken with nobody noticing for a long time. Oct 29 05:35:23 These days, people tend to notice, and I haven't heard of that recently. Oct 29 05:36:28 you can try 'ip link set FOO allmulticast on' as a workaround. Oct 29 05:36:48 or leave it in promisc mode with 'ip link set FOO promisc on' Oct 29 05:37:53 kthx, will try that. Oct 29 05:38:47 it would be good to capture your hardware details and kernel version/driver details in a bug report somewhere. Oct 29 05:42:19 hmmm, looks like I have the busybox version of ip. Command was ip link set allmulticast. But the route just went away. Oct 29 05:44:47 ifconfig says the interface is in PROMIS (and MULTICAST) mode, but my routing table is still bare. Oct 29 05:44:58 sometimes allmulti isn't a big enough hammer, if multicast is really broken. Oct 29 05:45:05 hm Oct 29 05:45:12 I'm watching the outbound RIP packets on the sending machine, and it says it's sending. Oct 29 05:45:29 Let me start up tcpdump again, see if the route reappears. Oct 29 05:45:49 bang, almost instantly. Oct 29 05:47:28 Could I bungle something in the build that woudl cause this? I'm running a self-built image from v18.06.1 on an R7800 and NBG6817. Oct 29 05:50:39 Actually, this might explain some other weirdness with printers not found, Chromecast only working if you're on the same WiFi, etc. Oct 29 05:50:53 That's all multicast, I think. Oct 29 05:53:27 possibly. though you can't really turn off multicast in the build. Oct 29 05:54:20 That's what I was afraid I'd done somehow, but I didn't see very many "MULTICAST" things that could be switched. Nothing obvious, anyway. Oct 29 05:55:10 I don't want to ask people to think about a problem that I created myself by stupid configuration. Oct 29 05:55:42 I guess I can load the standard build off the website. Oct 29 07:01:12 anyone know if the summit is going to be streamed? Oct 29 07:07:36 hm, it was last time. Good point, I want to watch as well since I couldn't go. Oct 29 08:26:45 philipp64: mostly Oct 29 08:27:31 i feel like deaty h and no stores are open but i am progressing towards the venue Oct 29 08:36:32 russell--: thanks for the update Oct 29 09:20:25 ugh, traffic Oct 29 11:14:08 stintel: ping Oct 29 12:09:00 I love it when stuff breaks at inconvenient times Oct 29 12:11:17 jow: pong Oct 29 12:20:35 As an audio engineer I get annoyed with a) lectern mics with inadequate high pass filtering so the popping blows you to the back of the room. b) RF mics running right at the limit of RF range - neither fixed/tweaked to my knowledge during the break - grrrrrr ;-). Oct 29 12:20:58 Checking location of receiving antennas would be a good start. Oct 29 12:23:32 lol Oct 29 12:24:01 was hoping to catch the current talk, alas sat in the lobby thinger fixing stuff Oct 29 12:25:42 must be time for a smoke break Oct 29 12:26:36 it's ddid someone share a stream link yet? or is there oneeven? Oct 29 12:27:20 food is looking food :F Oct 29 12:27:21 :D Oct 29 12:27:32 ugh, good Oct 29 12:27:35 E_NOSLEEP Oct 29 12:28:58 karlp, no Oct 29 12:29:11 ldir: are you around in lisbon? Oct 29 12:29:41 yep Oct 29 12:38:13 jwh: I would be quite worried if the food wouldn't be looking like food ;P Oct 29 12:38:20 lol Oct 29 12:56:16 hey, I have serial access to a wifi router; where can I find the LED GPIO configuration for a stock firmware? Oct 29 13:02:42 The device appears to be running AA-rc1 as the stock firmware Oct 29 13:03:22 can confirm, grub is good Oct 29 13:03:25 https://forum.openwrt.org/t/support-for-tp-link-archer-c2-v3/15617 Oct 29 13:11:41 what a lovely day Oct 29 13:16:09 ldir: are you staying for beer after? Oct 29 13:57:34 build #968 of pistachio/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/pistachio%2Fgeneric/builds/968 Oct 29 13:59:41 blloody sysupgrade keeps failing. fails to kill netdata nd just aborts all the time Oct 29 13:59:53 :( Oct 29 14:01:06 karlp: I had something similar with hostapd hanging due to wrong combination of ath10k with ct-htt firmware Oct 29 14:01:08 bright side, it never corrupts anything, just looks good and then reboots into the old image unless you're paying attention Oct 29 14:01:28 it's really annoying that you can't see what goes wrong unless you're connected on serial console Oct 29 14:01:37 it does this "sending kill to remaining processes..." and seems to give up in less than a second Oct 29 14:06:10 build #1116 of brcm63xx/generic is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/brcm63xx%2Fgeneric/builds/1116 blamelist: Radek Dost?l , Koen Vandeputte , David Bauer Oct 29 14:23:24 huh? Oct 29 14:23:51 Reason: targetupload Uploading target files failed Oct 29 14:26:10 soundsd like a infra failure, and something defaulted to just "blaming" the committers Oct 29 14:26:18 the build succeeded, but copying artifacts didn't Oct 29 14:26:53 hi, random question maybe, but are the freifunk-* packages in OpenWrt still in use? Oct 29 14:27:27 s/in use/maintained/ Oct 29 14:31:18 and how to I show errors in luci? Oct 29 14:53:12 build #1072 of at91/legacy is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/at91%2Flegacy/builds/1072 Oct 29 15:14:30 xback: karlp: the tah builder is experience connectivity issues causing these errors, will kick it out of the pool for now Oct 29 15:16:09 nbd: https://forum.openwrt.org/t/netfilter-flow-offload-hw-nat/10237/186 Oct 29 15:24:25 jow: just pushed the final 2 fixes for all partition errors on ar71xx/generic Oct 29 15:24:51 jow: any chance of kicking a buildbot to try building ar71xx/generic again asap? Oct 29 15:27:19 Hi people is there any streams from lisbon? Oct 29 15:27:28 xback: query Oct 29 15:27:39 lynxis: ping, will need your laptop once again ;) Oct 29 15:28:10 stintel: un-ping. Wanted to ask you to provision another do droplet because I don't have my 2fe with me in lisbon but lynxis already helped Oct 29 15:30:09 jow: ok! Oct 29 15:35:56 Tapper: AFAIK sadly no. Oct 29 15:41:41 Tapper: bit late today :P Oct 29 15:42:02 Thanks ldir Oct 29 15:42:16 jwh kids off school Oct 29 15:42:23 ah Oct 29 15:42:46 put them in cages, there is internet at stake Oct 29 15:43:13 If there was any streems was going to put one on well cooking dinner Oct 29 15:43:19 jwh: where are you then? Oct 29 15:43:33 ldir: about to have a smoke Oct 29 15:43:42 Ah, outside! Oct 29 15:43:46 but uh, coffee place atm Oct 29 15:43:50 just leaving Oct 29 15:43:57 Yeah could get one of those big dog cages, but I would need 5 Oct 29 15:44:20 kernel is what 4.14.78 now? Oct 29 15:44:30 Slimey yeah Oct 29 15:45:13 ldir: i didnt realise there was a bit out the back, ive been sitting on the steps like a hobo Oct 29 15:46:04 jwh: I just don't go out :-) Oct 29 15:46:09 lol Oct 29 15:46:28 are you doing workshops? Oct 29 15:47:12 I'm hoping that I'm in the right place for 'birds of a feather'! Oct 29 15:47:25 which I thought was in the main auditorium. Oct 29 15:48:20 oh Oct 29 15:48:33 ah, that's later... I can't read! Oct 29 15:49:14 yeah i was about to say Oct 29 15:49:16 lol Oct 29 15:54:42 build #1096 of lantiq/xway is complete: Failure [failed targetupload] Build details are at http://phase1.builds.lede-project.org/builders/lantiq%2Fxway/builds/1096 blamelist: Radek Dost?l , Koen Vandeputte , David Bauer Oct 29 16:08:09 build #1011 of zynq/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/zynq%2Fgeneric/builds/1011 Oct 29 16:46:22 stintel: still around? Oct 29 16:47:42 jow: yes Oct 29 16:48:09 stintel: can you switch over the git.archive.openwrt.org cname to git-archive-01.infra.openwrt.org ? Oct 29 16:48:18 away from the nanl machine Oct 29 16:48:33 the nanl box is configured to reverse proxy to the new one Oct 29 16:50:28 mirko wants to shut it down in a few days when the dns converged Oct 29 16:59:56 jow: done Oct 29 17:05:30 3 beer tokens! Oct 29 17:32:52 jwh: cheater Oct 29 17:33:08 oh? :( Oct 29 17:33:16 :) Oct 29 17:33:30 should i not have 3? Oct 29 17:33:59 you should give me the extra two, i'll dispose them properly :) Oct 29 17:34:04 haha Oct 29 17:34:25 i look like a hobo, i need them more obviously Oct 29 17:34:35 lol Oct 29 18:02:16 Hi people is there a update on the ar71xx snapshots? Oct 29 18:21:45 probably not today Oct 29 19:41:45 Hi when you have used git to format a patch were does it save the file? Oct 29 19:41:52 does* Oct 29 19:46:07 typically in the directory you're in Oct 29 19:46:38 (except if you used -o or --output-directory) Oct 29 19:50:12 build #1077 of archs38/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/archs38%2Fgeneric/builds/1077 Oct 29 21:05:39 Is someone recording the summit this year? There are YouTube playlists for the past years Oct 29 21:29:40 i believe so Oct 29 22:02:18 I'm trying to use a backports package I created from my 4.16 ct kernel...it fails to compile with this..any idea what might be the issue? Oct 29 22:02:23 /home/greearb/git/openwrt-r7800-dev/build_dir/target-arm_cortex-a15+neon-vfpv4_musl_eabi/linux-ipq806x/backports-4.16-ct-oct-29-2018-1/include/linux/ath9k_platform.h /home/greearb/git/openwrt-r7800-dev/build_dir/target-arm_cortex-a15+neon-vfpv4_musl_eabi/linux-ipq806x/linux-4.14.78/include/linux/ath9k_platform.h differ: char 1437, line 49 Oct 29 22:07:00 greearb__: you need to make sure that the kernel's ath9k_platform.h and the backport's ath9k_platform.h are identical (by patching either one to match the other). this also applies to a few more header files Oct 29 22:07:54 ok, so copying the backports .h over the target/linux/generic.... file is proper way to go about it? Oct 29 22:13:13 greearb__: depending on the target you are building for, you might need to add the patches modifying the backports' version to match the kernel one Oct 29 22:14:07 ok, thanks for that info... Oct 29 22:14:20 https://github.com/openwrt/openwrt/blob/master/package/kernel/mac80211/Makefile#L480 these three files are checked, and ath5k_platform doesn't exist upstream Oct 29 22:21:08 got an image! Will test it tomorrow Oct 29 22:22:06 build #187 of at91/sama5d4 is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/at91%2Fsama5d4/builds/187 Oct 29 22:22:08 is there a prefix you can use for files in patches/ dir to skip them? Or some other minimal-change way to skip them? Oct 29 23:17:34 when are we moving to 4.19? Oct 29 23:18:19 next stable iirc Oct 29 23:18:26 internally don't ask me :) Oct 29 23:31:27 build #1027 of lantiq/falcon is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/lantiq%2Ffalcon/builds/1027 Oct 29 23:33:15 when is the ath79 source-only mask being lifted? Oct 29 23:45:29 philipp64: -ENOPATCH ;-) Oct 30 01:01:30 build #1029 of ixp4xx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ixp4xx%2Fgeneric/builds/1029 Oct 30 02:02:52 build #1000 of ar71xx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/ar71xx%2Fgeneric/builds/1000 Oct 30 02:36:29 build #1117 of brcm63xx/generic is complete: Success [build successful] Build details are at http://phase1.builds.lede-project.org/builders/brcm63xx%2Fgeneric/builds/1117 Oct 30 02:41:28 build #455 of octeontx/generic is complete: Failure [failed sourceupload] Build details are at http://phase1.builds.lede-project.org/builders/octeontx%2Fgeneric/builds/455 blamelist: Koen Vandeputte **** ENDING LOGGING AT Tue Oct 30 02:59:59 2018