**** BEGIN LOGGING AT Wed May 08 02:59:56 2019 May 08 09:32:05 I don't understand what this means from Dave Miller https://lore.kernel.org/netdev/20190507.172000.384528161562920463.davem@davemloft.net/ May 08 09:32:48 When is net-next 'open'? And on the re-submit do I start at v1 again? May 08 09:50:39 ldir: net-next will open again when the kernel merge window is over May 08 09:50:47 i.e. when 5.2-rc1 is released May 08 09:55:38 ldir: you can use http://vger.kernel.org/~davem/net-next.html to check if net-next is open May 08 09:58:04 ldir: and you stay at whatever V you were when last submitting it when just resubmitting May 08 10:02:39 ok, thanks! May 08 10:28:38 this applies to the new features or to the fixes as well? May 08 10:35:15 answering to myself: yes as in https://www.kernel.org/doc/Documentation/networking/netdev-FAQ.txt May 08 12:17:13 ynezz ping May 08 13:35:11 I've got to say that as boring as it is to watch, seeing a new device banging out [r/w] for sysupgrade is putting a smile on my face May 08 13:38:29 There isn't a neat trick to get the LED setup done through the DTS for netdev triggers the way there is for USB, is there? May 08 13:43:39 Anyone know why my RAVpower FileHub Plus no longer detects the battery status LEDs as an i2c device (i2c-0) with newer snapshot images? May 08 13:44:36 Oops just meant battery status measuring chip May 08 13:45:15 I used to be able to check the battery status thanks to wd03-utils (see GitHub) May 08 13:58:24 I think I just found a bug in the python-twisted package. I would like to confirm it here before filing a bug report. May 08 13:59:35 The problem is that python-twisted is missing the twisted.web.test module. All other distributions of the python twisted package include it. May 08 14:00:26 I don't think python-twisted has been divided into multiple packages that contain the missing modules. May 08 14:00:41 That's why I am asking here before filing a bug. May 08 14:42:46 What sets `PCI_SUPPORT` in the build system? May 08 14:43:28 FEATURES=pci May 08 14:43:31 afaik May 08 14:44:12 OK, I'll give it a shot, thanks! May 08 14:44:20 (Then figure out why it's missing!) May 08 14:56:13 is it possible to get the LAN bridge interface name via UCI? I presume not as netifd will generate it. I'm trying to figure out a one-liner for getting the name of the LAN bridge. May 08 14:56:18 any suggestions please? May 08 14:56:49 shell-script one line, using ubus call network.interface.lan status is a pain because of JSON output. Thanks in advance May 08 14:59:50 jsonfilter --help May 08 15:00:17 "Patterns are JsonPath" May 08 16:01:58 Neighbor11111118: ifname=$(ifstatus lan | jsonfilter -e '@.l3_device') May 08 16:02:17 Neighbor11111118: or alternatively, eval $(ifstatus lan | jsonfilter -e 'ifname=@.l3_device'); echo $ifname May 08 16:02:44 note that ifstatus X in turn is just an alias for "ubus call network.interface.$X status" May 08 16:04:26 thanks jow :) May 08 16:09:10 https://promwad.com/case-studies/openwrt-realtek … May 08 16:10:01 I just came across that link on twitter May 08 16:11:05 are realtek  wifi radios any good? May 08 16:13:11 ynezz please merge https://github.com/openwrt/openwrt/pull/2042 May 08 16:33:39 Does anyone know why I'm seeing " daemon.warn netifd: Failed to set IPv6 mtu to 1500on interface 'wg0'" on a regular basis? May 08 16:34:20 and I've pushed a fix to put a space between the reported mtu and "on interface" May 08 16:35:59 and attempting to set an mtu of 1500 on a wireguard interface is unlikely to work...which it doesn't. Usual mtu is 1420. May 08 16:53:25 Is there a simple explanation of how `config-4.19` and `/config-default` get pieced together for `make menuconfig` and the kernel's `.config`? May 08 17:00:57 anyone know if there is a modern chromebook with a recent qca chipset (or that will take full-height mini-pcie)? May 08 17:21:05 dedeckeh: ping May 08 17:21:55 pong May 08 18:19:40 brucethompson: patch/002-omit-tests.patch is excluding it from the build 'packages'. May 08 18:34:24 ldir:pong May 08 18:42:15 dedeckeh: just having food at the mo - let me know when is a good time to catch you - it's ipv6 mtu setting warnings May 08 18:43:02 ldir:later in the evening is fine; enjoy your food May 08 18:57:24 dedeckeh: ok - food was nice! I'm getting some fallout related to https://git.openwrt.org/?p=project/netifd.git;a=commit;h=115a694ab9a44cd52a8ab18e07909c753de9fb94 May 08 18:58:05 I'm seeing regular warnings from netifd for both of my wireguard interfaces "daemon.warn netifd: Failed to set IPv6 mtu to 1500on interface 'wg0'" May 08 18:58:48 I've not seen this before and suspect it's related to a recent change that I did to my APU2 network config. May 08 18:59:32 Normally, openwrt runs the lan 'interface' as a bridge - br-lan. In my case it was a bridge composed of one single ethernet interface - eth1. May 08 19:00:37 so I dropped the bridge definition from my config and said that lan uses 'eth1' as the ifname. May 08 19:01:37 ldir:I would not expect that to be a problem May 08 19:02:03 but you get the warning on the wg0 interface May 08 19:02:16 I get the warning on wg0 & wg1 May 08 19:02:55 trying to set an mtu > 1420 on wireguard isn't going to fly, hence the warning. May 08 19:03:00 This logic tries to update the IPv6 mtu of the downstream IPv6 interfaces May 08 19:03:19 to make PMTU work May 08 19:04:00 I guess the the IPv6 mtu of the upstream interface is 1500 ? May 08 19:04:42 correct - I'm not sure if the upstream interface in this case is regarded as the 'lan' interface (eth1) or 'wan' (eth0) but either way mtu is 1500 May 08 19:04:53 it's wan May 08 19:06:01 would you like to see my /e/c/network in case I've done something really stupid May 08 19:06:24 yes May 08 21:13:53 fascinating... libiconv-stub does not even work properly May 08 21:35:12 Hello! May 08 21:35:12 Found a serious bug with the ethernet driver on the ath79 architecture. May 08 21:35:23 https://img-host.ru/FwNj.jpg May 08 21:36:15 Reproduced on tp link 710, 3500, 4300 May 08 21:37:06 At 710 there is a loss of packets. May 08 21:39:35 There were no such bugs on the fixed drivers for the ar71xx architecture. Even 710 calmly tested in both directions at the same time passed with a result of ~ 192 Mbps for 24 hours. May 08 21:40:11 I just sent email with info on latest ath10k-ct firmware/driver. If someone could add this to OpenWRT I'd appreciate it. May 08 21:44:29 there was still a bug with the switch off, now no traffic is passed without vlan. May 09 02:21:15 Halp May 09 02:21:26 build failed.... again May 09 02:21:31 but this time i got logs **** ENDING LOGGING AT Thu May 09 02:59:56 2019