| 12 Jun 2025 |
 adamcstephens | Doesn't systemd-networkd need access to secrets? | 14:19:27 |
 emily | the current state is that pppd succeeds and networkd considers the resulting device to have a carrier:
Dec 20 12:04:39 bpi-r4 pppd[2265]: Connect: ppp-wan <--> wan
Dec 20 12:04:39 bpi-r4 systemd-timesyncd[430]: Network configuration changed, trying to establish connection.
Dec 20 12:04:39 bpi-r4 systemd-networkd[471]: ppp-wan: Configuring with /etc/systemd/network/10-ppp-wan.network.
Dec 20 12:04:39 bpi-r4 systemd-networkd[471]: ppp-wan: Link UP
Dec 20 12:04:39 bpi-r4 systemd-networkd[471]: ppp-wan: Gained carrier
but it remains in the configuring state
19 ppp-wan ppp carrier configuring
and doesn't get any address
17: ppp-wan: <POINTOPOINT,MULTICAST,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN group default qlen 3
link/ppp
| 14:19:41 |
| emily | (that's after I manually networkctl up wan and with
systemd.network.networks."10-ppp-wan" = {
matchConfig.Name = "ppp-wan";
networkConfig.KeepConfiguration = true;
linkConfig.RequiredForOnline = "routable";
};
for the created interface)
| 14:20:00 |
| emily | (wait, in what context?) | 14:20:12 |
| adamcstephens | I know when working with networkd and wireguard, it's necessary to give systemd-networkd permissions to read the keys | 14:20:36 |
| emily | pppd is a separate daemon. networkd has no native support for PPPoE | 14:20:58 |
| adamcstephens | ahh | 14:21:01 |
| emily | (or this would be a lot less of a headache) | 14:21:05 |
| adamcstephens | sorry, i've not used ppp since the 90s | 14:21:09 |
| emily | what, your gigabit fibre doesn't use PPPoE for no reason whatsoever? 🫠| 14:21:38 |
| adamcstephens | the one fiber connection i've used did not (AT&T) | 14:22:00 |
| adamcstephens | There are a couple smaller providers who are apparently rolling out fiber in my city, so we'll see if and when they get to my house. But I don't think using ppp is as popular here in the US as it is across the pond | 14:27:46 |
| emily | there are some providers that do IPoE. but sadly it is not very common. | 14:30:11 |
| 13 Jun 2025 |
|  @magic_rb:numtide.com joined the room. | 14:11:38 |
| @magic_rb:numtide.com | hello from another account, so im running ifstate to setup the network on my BPI-R4 and im running into a very weird, uh, bug? i genuinely don't know what's happening. When ifstate runs for the first time it does some processing and then eventually runs the following command:
/nix/store/gwx71g0s855x2ln1l17c720mh0rvvr3c-ethtool-6.14/bin/ethtool --change end2 autoneg off duplex full speed 1000
which fails with
netlink error: link settings update failed
netlink error: Invalid argument
which is weird, but then if i run ifstate again, it works, if i run that ethtool command in a shell, it works. It just doesnt work when ifstate is doing it during boot up. Is there some way to get more out of the kernel than "invalid argument"? because that is not descriptive enough
| 14:14:20 |
| @magic_rb:numtide.com | its probably being "fixed" by something ifstate does later, i just have no fucking clue what | 14:15:14 |
| @magic_rb:numtide.com | ah, progress:
Jun 13 16:08:26 hela kernel: sfp sfp1: Unable to ascertain link mode
Jun 13 16:08:26 hela kernel: mtk_soc_eth 15100000.ethernet end2: selection of interface failed, advertisement 0000000,00000000,00000000,00000000
| 14:22:12 |
| @magic_rb:numtide.com | but i still dont know how to fix this | 14:22:18 |
| @magic_rb:numtide.com | right, so if i ip link set end2 up before it works, weird that you have to pull it up to be able to set the ethtool settings | 14:41:25 |
| @magic_rb:numtide.com | will open a bug report in ifstate | 14:41:30 |
 K900 | Yeah that depends on the MAC I think | 14:46:19 |
| @magic_rb:numtide.com | okay so ifstate bug then | 14:47:20 |
| @magic_rb:numtide.com | on a related note, my wifi broke | 14:47:25 |
| @magic_rb:numtide.com | [ 11.874819] mt7996e 0000:01:00.0: vgaarb: pci_notify
[ 11.874869] mt7996e 0000:01:00.0: assign IRQ: got 120
[ 11.874886] mt7996e 0000:01:00.0: enabling device (0000 -> 0002)
[ 11.874930] mt7996e 0000:01:00.0: enabling bus mastering
[ 11.952294] mtk-pcie-gen3 11300000.pcie: msi#0x1 address_hi 0x0 address_lo 0x11300c00 data 1
[ 11.973753] mt7996e 0000:01:00.0: Direct firmware load for mediatek/mt7996/mt7996_rom_patch_233.bin failed with error -2
[ 11.973773] mt7996e 0000:01:00.0: Falling back to sysfs fallback for: mediatek/mt7996/mt7996_rom_patch_233.bin
| 14:47:43 |
| @magic_rb:numtide.com | main@hela:~$ ls /run/current-system/firmware/mediatek/mt7996/mt7996_rom_patch.bin.zst
/run/current-system/firmware/mediatek/mt7996/mt7996_rom_patch.bin.zst
```3
| 14:47:58 |
| @magic_rb:numtide.com | * main@hela:~$ ls /run/current-system/firmware/mediatek/mt7996/mt7996_rom_patch.bin.zst
/run/current-system/firmware/mediatek/mt7996/mt7996_rom_patch.bin.zst
| 14:48:01 |
| @magic_rb:numtide.com | * main@hela:~$ ls /run/current-system/firmware/mediatek/mt7996/mt7996_rom_patch_233.bin.zst
/run/current-system/firmware/mediatek/mt7996/mt7996_rom_patch_233.bin.zst
| 14:48:21 |
| K900 | Probably need to reboot it | 14:50:22 |
| K900 | Sometimes it gets stuck in a weird state after a soft reboot | 14:50:45 |
| K900 | Cutting the power helps | 14:50:45 |
