<red[evilred]>
could just about fit a few intel NUCs in there
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
srk has quit [Remote host closed the connection]
srk has joined #nixos-aarch64
<samueldr>
a comodore 64 is small comparatively
sigtrm has quit [Ping timeout: 256 seconds]
sigtrm has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
orivej has joined #nixos-aarch64
<angerman>
anyone put nixos on their kobol helios64 yet?
<samueldr>
haven't had the time yet, but I expect in the next few weeks to look at it
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
mschwaig1 has quit [Ping timeout: 240 seconds]
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
mschwaig1 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
<colemickens>
If someone has root on the aarch64 machine, could they possibly get a coredump backtrace for me? The exact same scenario does not repro on my Pi.
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
<samueldr>
colemickens: niche?
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
<samueldr>
colemickens: moved it in your home
<colemickens>
thanks.
orivej has quit [Ping timeout: 256 seconds]
<colemickens>
(enabling dontStrip=true; to see if I can get any better info)
<colemickens>
and of course it works now
<colemickens>
(not really, it crashed again, hopefully this time with a more informational dump.)
<samueldr>
colemickens: copied
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
<samueldr>
enable = true; version = 4; and don't enable uboot
<red[evilred]>
I was able to get it to work
<red[evilred]>
able to do a nixos-rebuil;d
<red[evilred]>
but on reboot - it would always go back to the single user nixos automatcally logging in with none of my changes implemented
<red[evilred]>
sure - I can just nixos-rebuild on boot
<samueldr>
check that the boot partition mounts
<red[evilred]>
but that's just... sub-optimal
<samueldr>
there's a bug in the temporary configuration
<red[evilred]>
yup it's unmounted
<samueldr>
also note that the partition, being relatively "small", means you'll have fun not being able to rebuild too many concurrent generations
<samueldr>
though since the u-boot build breaks the raspberry pi fancy stuff, I mostly have thrown the towel for the time being
<samueldr>
it's so exhausting
<samueldr>
everyone tugging on the covers
<samueldr>
it is _seriously_ less annoying to deal with OEM shenanigans on android than dealing with the "open source community friendly" facade of the raspberry pi foundation
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
adamzivcak has left #nixos-aarch64 [#nixos-aarch64]
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
FRidh has quit [Ping timeout: 260 seconds]
FRidh has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
adamzivcak1 has joined #nixos-aarch64
{^_^} has quit [Remote host closed the connection]
{^_^} has joined #nixos-aarch64
zupo has joined #nixos-aarch64
<clever>
sphalerite: i think pigpio (the c library) has functions for software pwm, but they will consume a decent amount of cpu time
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite>
clever: I want the kernel to do it though.
<sphalerite>
it also looks like it's only for the pi?
<sds2>
patagonicus: I've tried again with larger swap space (8Gb) but it didn't work either. I partitioned the sdd with two partitions, one is labled "nixos" and the other as "swap". I also did a "swapon /dev/sda2" before starting. I'm not sure if the build/install used the swap at all. I occasionally watched it during the long build process but it was never filled with anything. Here is the dmesg output: https://pastebin.pl/view/b5a62e8f The interesting part st
<sds2>
art at line 480. Any ideas on this?
zupo has joined #nixos-aarch64
<colemickens>
Does the community box have something weird configured for libeatmydata system-wide?
<colemickens>
I can't figure out how that's showing up in a stack trace for a fairly simple Go app I wrote?
<Ke>
normally this would make sense for nixos builds, single sync should do after build
<colemickens>
pretty sure I know who the blame on that is.
<sphalerite>
oh wow
lickingball has quit [Ping timeout: 268 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<patagonicus>
sds2: weird. I don't really know much about this, but it looks like nix-build requested GFP_ATOMIC memory, which I'm guessing can't be put in swap, and that failed.
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
<patagonicus>
My pinephone accessories arrived. :)
<patagonicus>
Labelled as "Single board computer $20" for customs, for some reason.
alp has quit [Ping timeout: 260 seconds]
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
<patagonicus>
And, uh, steps 1 and 4 for the screen protector are in English while steps 2 and 3 are in Italien I think. Wat.
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
alp has joined #nixos-aarch64
alpernebbi has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
FRidh has quit [Ping timeout: 240 seconds]
FRidh has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
s1ng0c has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
zupo has joined #nixos-aarch64
adamzivcak has joined #nixos-aarch64
adamzivcak has left #nixos-aarch64 [#nixos-aarch64]
<NickHu>
Does hydra still build the armv6l image for raspberry pi?
sds2 has quit [Quit: sds2]
<NickHu>
The wiki isn't very clear
sds2 has joined #nixos-aarch64
<NickHu>
Also is there any way to see what's in a particular cachix?
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
<hexa->
I was wondering the same recently, haven't found a way
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
<patagonicus>
As far as I know Hydra doesn't build anything 32-bit ARM (ARM 7 or lower).
<hexa->
yeah, pretty sure graham was looking for plausible builders, but fast and pre-armv8 is tough
<patagonicus>
Building the whole world certainly isn't fun. I tried setting up hydra on my armv7 machines, but hydra itself failed to install …
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #nixos-aarch64
alp has quit [Remote host closed the connection]
alp has joined #nixos-aarch64
<LinuxHackerman>
hydra's quite a piece of work. Though you don't need to run your hydra on the arm builders.
<Lordcirth>
Can you build 32bit arm on 64bit arm? Or is that still cross-compiling enough to be troublesome?
<simpson>
I think that the community can probably scrape together enough compute capacity to build snapshots of the -small package set, maybe three or four times per year.
<srk>
lordcirth: yup, with qemu in 32 bit mode so not-cross
<LinuxHackerman>
lordcirth: it's… complicated.
<LinuxHackerman>
you can cross-build, or you can build natively, using kvm as srk says to hide the fact that it's actually aarch64 hardware from the toolchain
<Lordcirth>
And running 32bit QEMU on 64bit ARM is still a lot more efficient than running it on x86_64, presumably?
<srk>
correct
<LinuxHackerman>
with KVM, yes
<sphalerite>
Not all hardware supports KVM though.
<srk>
there are other issues with it
<hexa->
and not all thunderx cpus have 32-bit arm support
<sphalerite>
thunderx in general doens't support it, no?
<sphalerite>
but yeah definitely not all aarch64 stuff supports it :D
<Lordcirth>
So, if we were to get one of those aarch64 servers with tons of cores, we could build all the arm versions in hydra?
<Lordcirth>
If it was a supporting CPU?
<sphalerite>
The community aarch64 box for example does support it
<sphalerite>
but there are issues, last time I tried there were weird hangs and stuff
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sphalerite>
though I'm not the person who's experimented the most with this
<Lordcirth>
Is fixing that the kind of work that could be supported by the opencollective page?
<Lordcirth>
"Pay for projects that improve the open source nixos ecosystem (security patching, a designer, etc.)" - so it should be possible
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
s1ng0c has quit [Quit: Connection closed for inactivity]
alp has quit [Ping timeout: 272 seconds]
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
adamzivcak has joined #nixos-aarch64
adamzivcak has left #nixos-aarch64 [#nixos-aarch64]
lickingball has joined #nixos-aarch64
alp has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
cole-h has joined #nixos-aarch64
red[evilred] has joined #nixos-aarch64
<red[evilred]>
there's always a danger of paying people to work on projects like that - two things immediately spring to mind:
zupo has joined #nixos-aarch64
<red[evilred]>
a. Appearence of a conflict of interest
<red[evilred]>
b. The "monetization" of volunteer work discourages people from volunteering
rajivr has quit [Quit: Connection closed for inactivity]
<simpson>
There is a basic pattern which fixes those, but it does require that the paid contributors be paid to work on specific projects. https://www.x.org/wiki/XorgEVoC/ is an example of another community organization doing this relatively well.
<red[evilred]>
in one way, GSoC solves those problems by a: Having the decisions made by google. b: The person aplying is entirely responsible for the process end-to-end.
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
<simpson>
(There's no way that we can be more dysfunctional than X.org, BTW.)
<red[evilred]>
I'm not saying it's a bad idea or an unsolvable problem. Just saying that there are things that need to be considered
alp has quit [Ping timeout: 272 seconds]
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
lickingball has quit [K-Lined]
<colemickens>
gchristensen: I don't want to bother, but I have a fairly simple Golang app that segfaults in libeatmydata on the aarch64 community box. I noticed you added it to LD_PRELOAD for nix-daemon. Have you seen something like this before? https://gist.github.com/colemickens/25b3ec2e44a2fd89c8054a425603db30
<colemickens>
I wonder if I'm specifically being affected because the Go app is executed as a post-build-hook running under the daemon?
<gchristensen>
ehh weird
<gchristensen>
never
<gchristensen>
we could delete libeatmydata
<Lordcirth>
How much perf improvment does it get?
<gchristensen>
no idea
<gchristensen>
never measured :)
<sphalerite>
gchristensen: manny.metal and I are finally overlapping :D what are the machines that 5.4 works on without any issues?
<gchristensen>
nice
<gchristensen>
c2.large.arm
<sphalerite>
and is the community builder one of those too?
<gchristensen>
hhhno
<gchristensen>
that one is like
<gchristensen>
a .xda machine I think, he'll know what that means
<sphalerite>
could you maybe set me up with the two thunderx's again, but with ubuntu instead of alpine? :')
<gchristensen>
yea
<gchristensen>
can you ping in like 3h to do that?
<sphalerite>
nah, I want to be asleep in 3h :D maybe at the weekend then
<gchristensen>
aye
<gchristensen>
thanks, sorry
alp has joined #nixos-aarch64
veleiro has joined #nixos-aarch64
zupo has joined #nixos-aarch64
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
alpernebbi has quit [Quit: alpernebbi]
FRidh has quit [Quit: Konversation terminated!]
alp has quit [Ping timeout: 272 seconds]
orivej has quit [Ping timeout: 264 seconds]
Darkmatter66 has joined #nixos-aarch64
alp has joined #nixos-aarch64
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
cole-h has quit [Quit: Goodbye]
<colemickens>
oh no, we're still stuck on meson.build:30:4: ERROR: Problem encountered: Could not find bcm_host.h.
<samueldr>
disable the raspberry pi specific flag I guess?
<samueldr>
anyway it's part of an unfree package IIRC
<samueldr>
not like having an unfree gstreamer plugin package would be better
adamzivcak1 has joined #nixos-aarch64
adamzivcak1 has left #nixos-aarch64 [#nixos-aarch64]
Darkmatter66 has quit [Remote host closed the connection]
Darkmatter66 has joined #nixos-aarch64
cole-h has joined #nixos-aarch64
<hexa->
[root@dione:~]# dmesg
<hexa->
dmesg: command not found
<hexa->
btrfs on sata-ssd over usb3 uas on a pi4
<hexa->
my home-assistant still works
<hexa->
though it can't fetch graph data from pgsql
sds2 has quit [Quit: sds2]
sds2 has joined #nixos-aarch64
sds2 has quit [Client Quit]
sds2 has joined #nixos-aarch64
Darkmatter66 has quit [Ping timeout: 246 seconds]
Darkmatter66 has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]