samueldr changed the topic of #nixos-dev to: #nixos-dev NixOS Development (#nixos for questions) | NixOS 20.03 Feature Freeze Feb 10 https://discourse.nixos.org/t/nixos-20-03-feature-freeze/5655 | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | https://r13y.com | 19.09 RMs: disasm, sphalerite; 20.03: worldofpeace, disasm | https://logs.nix.samueldr.com/nixos-dev
orivej has quit [Ping timeout: 240 seconds]
ris has quit [Ping timeout: 240 seconds]
drakonis has quit [Quit: WeeChat 2.6]
lovesegfault has joined #nixos-dev
bhipple has joined #nixos-dev
lovesegfault has quit [Ping timeout: 256 seconds]
lovesegfault has joined #nixos-dev
zarel_ has joined #nixos-dev
zarel has quit [Ping timeout: 245 seconds]
justanotheruser has joined #nixos-dev
bhipple has quit [Remote host closed the connection]
ixxie has joined #nixos-dev
orivej has joined #nixos-dev
lassulus_ has joined #nixos-dev
lassulus has quit [Ping timeout: 260 seconds]
lassulus_ is now known as lassulus
ris has joined #nixos-dev
ixxie has quit [Ping timeout: 240 seconds]
<flokli> clever: regarding networkd killing you network link, take a look at the KeepConfiguration= directive for networkd: https://systemd.network/systemd.network.html#KeepConfiguration=
__monty__ has joined #nixos-dev
ixxie has joined #nixos-dev
janneke has quit [Ping timeout: 268 seconds]
janneke has joined #nixos-dev
FRidh2 has joined #nixos-dev
ixxie has quit [Ping timeout: 265 seconds]
_ris has joined #nixos-dev
ris has quit [Ping timeout: 272 seconds]
ixxie has joined #nixos-dev
zarel_ has quit [Quit: ZNC 1.7.4 - https://znc.in]
zarel has joined #nixos-dev
drakonis has joined #nixos-dev
<flokli> https://github.com/NixOS/nixpkgs/pull/77507#issuecomment-583744823 I hope I'm speaking for you here as well, gchristensen
<gchristensen> yeah, I'm all for it, but we need to be building from upstream's source, not some random fork on github
<gchristensen> building from a strange source is shady, the tor launcher bit is shady, restoring old versions is shady
<flokli> yes
<flokli> and as written in there, communication the way it currently happens isn't productive either, but just tiring me out.
<gchristensen> aye
<gchristensen> flokli: maybe reiterate the points of what you'd like to see instead... like, you'd like a tor-browser in nixpkgs but ... probably not disable the tor launcher? I'm not the expert here, but that feels shady to me ... build from upstream's sources ... not include old or out of date versions ... if patches are needed, include them with `fetchpatch`, not by pointing to an alternative source --
<gchristensen> ideally with a reference to an upstream ticket. for example
<gchristensen> that will help by giving direction for their next attempt and how to participate better
<gchristensen> (it may not work but it is good to try)
<flokli> Well, most of these questions have already been raised, but just ignored.
<gchristensen> yes, but spelling it out once more may help them understand better and more clearly
<gchristensen> and will hopefully help them learn
<flokli> I'm very exhausted currently. Will do some sports to get some endorphines ;-)
<gchristensen> okay, I will write that :)
<flokli> Thanks <3
<flokli> gchristensen++
<{^_^}> gchristensen's karma got increased to 205
<worldofpeace> Jan Tojnar: holy PR https://github.com/NixOS/nixpkgs/pull/79416 🤣
<{^_^}> #79416 (by jtojnar, 1 day ago, open): flatpak: 1.4.2 → 1.6.1
<worldofpeace> jtojnar++
<{^_^}> jtojnar's karma got increased to 27
<worldofpeace> jtojnar++
<{^_^}> jtojnar's karma got increased to 28
<worldofpeace> jtojnar++
<{^_^}> jtojnar's karma got increased to 29
<gchristensen> haha wow that is quite the PR
<worldofpeace> thanks so much
drakonis has quit [Quit: WeeChat 2.6]
<worldofpeace> gchristensen: reviewing it would be noteworthy
<gchristensen> only 4,444,378 rows left... joy ;)
FRidh2 has quit [Remote host closed the connection]
<clever> flokli: ah, looks like i would need to pair that up with networking.useNetworkd
<flokli> clever: no guarantees about any bash scripts and how they handle your networking ;-)
<flokli> also, our stage 1 and 2 are a bit funky, so not sure if we don't accidentially disconnect things there (I hope we don't)
<clever> flokli: booting is done, using the initrd networking
<clever> flokli: shutdown is the problem, since the root disk basically vanishes mid shutdown
<flokli> clever: unmounted, or link disconnected?
<clever> flokli: link disconnected
<flokli> and you're not using networkd currently
<clever> yeah
<clever> just nixos defaults
<flokli> clever: we have a bunch of "ip link del" in scripted networking postStop scripts
<flokli> nixos/modules/tasks/network-interfaces-scripted.nix
<clever> yeah, i was wondering how to stop those, but switching to networkd might be simpler
<flokli> :+1:
<clever> flokli: weird, we have both systemd.network.enable and networking.useNetworkd
<flokli> clever: yes
<clever> another problem i need to fix, is the mac address
<clever> currently, it has a random mac on every boot
<flokli> That's a problem of the bootloader or kernel not properly generating a stable one?
<clever> i know what it should be, but not how to set it from the bootloader
<clever> and due to some PLL issues, the bootloader cant even contact the usb bus, linux finishes that config
orivej has quit [Ping timeout: 272 seconds]
<flokli> but not sure if that works at all. IIRC, .link files are read by udev, so they might need to be around in initrd already
<flokli> (you need to match by something else than the mac address obviously
<flokli> Path= maybe
<clever> the official firmware, used device-tree to describe the usb devices
* clever looks
<clever> they are fully describing the usb devices hard-wired to the usb controller
<clever> [ 3.197519] smsc95xx 1-1.1:1.0 eth0: register 'smsc95xx' at usb-3f980000.usb-1.1, smsc95xx USB 2.0 Ethernet, b8:27:eb:77:df:95
<clever> and linux prints this on bootup
<clever> root@raspberrypi:/proc/device-tree/soc/usb@7e980000/usb1@1/usbether@1# ls -l
<clever> flokli: thats not in the sources!
<clever> -r--r--r-- 1 root root 6 Feb 8 03:55 local-mac-address
<flokli> clever: I don't understand the question ;-)
<flokli> oh, wat
<flokli> why doesn't it apply that mac address? sheesh
<clever> the official bootloader, is adding a local-mac-address to the devicetree, on boot
<clever> based on the hardware serial#
<clever> i'm using custom dtb files with a custom bootloader
<clever> so i now need to re-implement that
<clever> Bus 001 Device 003: ID 0424:ec00 Standard Microsystems Corp. SMSC9512/9514 Fast Ethernet Adapter
<clever> yeah, thats vid:pid pairs in DT
<clever> and local-mac-address is 6 raw bytes
<clever> ok, it compiles, now will it boot...
<clever> flokli: i just realized, i never compiled the ethernet driver into the kernel
<clever> flokli: how is it booting? lol
<clever> drivers/net/usb/smsc95xx.c: .description = "smsc95xx USB 2.0 Ethernet",
<clever> CONFIG_USB_NET_SMSC95XX
<clever> ah, already =y somehow
<ma27[m]> niksnut: is there anything I can do to make https://github.com/NixOS/nix/pull/3216 (Add an optional `hash` parameter to `builtins.fetchGit`)? :)
<clever> [ OK ] Reached target Power-Off.
<clever> [ 858.808369] Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000000
<clever> [ 858.715682] reboot: System halted
<{^_^}> nix#3216 (by Ma27, 13 weeks ago, open): Add an optional `hash` parameter to `builtins.fetchGit`
<clever> flokli: it does look like networkd fixed things
<clever> flokli: now i just have to implement actual shutdown, lol
<flokli> clever: nice to hear :-)
<clever> flokli: the next major problem i have to deal with, is the arm clock freq
<clever> flokli: its running at 1/50th of the normal speed
<ma27[m]> s/make/make mergable/ btw %)
bhipple has joined #nixos-dev
<clever> local-mac-address = [ b8 27 eb 77 df 95 ];
<clever> flokli: i can confirm, this sets the mac!
<clever> now i just need to fill that in from the bootloader, and i'm all set
<flokli> Coolio
drakonis has joined #nixos-dev
<flokli> clever: I still don't fully understand it - so why does the kernel not just initialize the NIC with the proper mac address?
<clever> flokli: the ethernet adapter lacks a flash chip to hold such config
<clever> flokli: the official firmware, will use the 3 byte broadcom id, and 3 bytes of the cpu serial# to generate a 6 byte mac, and fill in the DT
<clever> flokli: if i fill the same part of DT in, then linux does correctly obey it, and fills in the mac
<flokli> clever: yeah, but still - wouldn't it then be the kernels job to configure that from the DT?
<flokli> ah
<flokli> so, the .link stuff is a workaround, till you fill the DT properly
<flokli> gotcha
<clever> the .link stuff, is to make linux keep eth0 up at shutdown
<clever> because the rootfs is on the network
<clever> the DT is more for a static mac addr, so the ip is predictable
<flokli> no, .link is not about keeping the interface up
<flokli> .link is udev stuff, and can be used to change mac addresses
<flokli> .network is to tell networkd to not randomly remove addresses ;-)
<clever> 38 systemd.network.links.eth0 = {
<clever> 40 extraConfig = ''
<clever> 41 KeepConfiguration=dhcp
<clever> that is what ive done
<flokli> yeah, that's in .network
<clever> not sure which config file it lands in, ah
<clever> [root@nixos:~]# cat /etc/systemd/network/eth0.link
<clever> [Link]
<clever> KeepConfiguration=dhcp
<clever> flokli: it landed in a .link file
<flokli> that sounds wrong
<flokli> there's also .link files, which can change mac addresses and other low-level stuff
<flokli> but those are udev land
<clever> yeah, man page says its in the wrong file
<clever> maybe just the act of using networkd fixed things?
<flokli> I assume so, yes.
<flokli> But you should still put it in systemd.network.networks.<name>.networkConfig
<flokli> I assume this keeps it working when you restart networkd, or your lease expires
<flokli> systemd ignores config options it doesn't know. you might want to run `systemd-analyze log-level debug` and then restart networkd to see how things are parsed
<flokli> if you go via the attrset, there should also be some assertions, making sure you don't configure stuff in the wrong places
<flokli> for extraConfig = ''…'', we obviously can't do that
<clever> yeah
<{^_^}> firing: BuildsStuckOverTwoDays: https://status.nixos.org/prometheus/alerts
ixxie has quit [Ping timeout: 260 seconds]
<jtojnar> looks like the test driver is just re-using the old /tmp/vm-state-machine/machine.qcow2 without resizing it, even if I change virtualisation.diskSize 😠️
orivej has joined #nixos-dev
<flokli> yes - it also provides network access :-/
drakonis has quit [Remote host closed the connection]
<flokli> Which is why I wrap it in nixosTests.…
drakonis has joined #nixos-dev
evils has quit [Quit: Lost terminal]
evils has joined #nixos-dev
<jtojnar> flokli where does that put the image?
phreedom has quit [Ping timeout: 240 seconds]
phreedom has joined #nixos-dev
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-dev
phreedom has quit [Ping timeout: 240 seconds]
phreedom has joined #nixos-dev
justanotheruser has quit [Ping timeout: 272 seconds]
orivej has quit [Ping timeout: 265 seconds]
orivej has joined #nixos-dev
<flokli> In a derivation
<flokli> I mean, it's sandboxed and gone
justanotheruser has joined #nixos-dev
__monty__ has quit [Quit: leaving]
justanotheruser has quit [Ping timeout: 256 seconds]
justanotheruser has joined #nixos-dev