<colemickens> idk where I left off. Last thing I was trying was laying down rpi4uefi and then also triggering the regular systemd-boot bootloader installer. I think I was close, but I am jumping around a lot.
orivej has quit [Ping timeout: 260 seconds]
s1ng0c has quit [Quit: WeeChat 2.9]
veleiro has quit [Remote host closed the connection]
veleiro has joined #nixos-aarch64
veleiro has quit [Ping timeout: 265 seconds]
veleiro has joined #nixos-aarch64
Darkmatter66_ has quit [Ping timeout: 240 seconds]
Darkmatter66 has joined #nixos-aarch64
cole-h has quit [Ping timeout: 272 seconds]
veleiro has quit [Ping timeout: 256 seconds]
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 272 seconds]
tilpner_ is now known as tilpner
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
h0m1 has quit [Ping timeout: 264 seconds]
h0m1 has joined #nixos-aarch64
<clever> Gaelan: maybe also move here, since the issue is arm based, and -dev is getting busy
<Gaelan> sure
<clever> i'm currently building master without that patch
<clever> binutils
<Gaelan> I'm building xpwcd9s984flhrfdz0wc8hrr8mgiayrq-bootstrap-stage2-gcc-wrapper-.drv, right after the patch
<clever> Gaelan: lines 42 and 172
Darkmatter66 has quit [Ping timeout: 256 seconds]
<clever> this is AVR code without a kernel, but its a rough example of what libc is probably doing internally
<Gaelan> right
<clever> i can also see it in here
<clever> ive already commented it out in default.nix
<clever> the 1st half of the patch, the if statement was previously checking if a section was readonly and nothing else
<clever> but it has been modified to also check that nocopyreloc==0, the alloc flag is set, and the size is non-zero
<clever> and the 2nd part, only adds checks that its readonly
<Gaelan> presumably a FILE isn't readonly, and now I guess it's not getting copied?
<clever> its the same FILE from `fopen`, and it likely needs to be modified, to deal with things like `setbuf`
<clever> there is both the memory for the FILE struct, and the FILE *stderr pointer
<clever> binutils> Making installcheck in po
<Gaelan> ugh this is going to involve me reading glibc code isn't it
<clever> bootstrap-stage2-gcc-wrapper-> copying path '/nix/store/r2wqr177dagxl13n382sghafqq0i81wc-bootstrap-stage0-stdenv-linux' from 'file:///mnt/'...
<clever> linux-headers> unpacking source archive /nix/store/cai15i0m74ihfajdqfz56xgayc9bf9m3-linux-5.9.8.tar.xz
<clever> still unpacking, lol
<Gaelan> boot.tmpOnTmpfs = true
red[evilred] has joined #nixos-aarch64
<red[evilred]> like me after my last move
<red[evilred]> complete with the target block house^H^H^H^H^Hdevice becoming full
<clever> Gaelan: nix-daemon respects $TMPDIR
<clever> so i can just point it anywhere
<Gaelan> oh neat, that's good to know
<Gaelan> but yeah, linux unpacks way more quickly on a ramfs
<clever> and in single-user mode, all nix tools respect TMPDIR
<Gaelan> tmpfs*
<clever> linux-headers> tar: linux-5.9.8/lib: Cannot mkdir: No space left on device
<Gaelan> oooof
<Gaelan> but you shouldn't need linux-headers, just the cc it uses
<clever> was more, confirming that commenting the patch out fixed it
<clever> Dec 30 23:36:15 nas hydra-queue-runner[5136]: possibly transient failure building ‘/nix/store/0l60safsi9kr6rk7gr00r4mxrbvn7b0l-linux-5.4.86.drv’ on ‘pi@pi400’:
<clever> hydra failed an unrelated aarch64 build on the same box
<clever> for the same reason
<Gaelan> ahh
<clever> and that 2nd failure already deleted things, freeing up space, lol
<clever> root@pi400:~# mount -t tmpfs none /mnt/tmp/
<clever> Gaelan: and i already used TMPDIR to put things on a usb stick, because the SD card was full
<Gaelan> ah
<clever> already unpacked, lol
<clever> linux-headers> installing
<clever> yep, commenting that out fixes things, on to the test...
rajivr has joined #nixos-aarch64
<Gaelan> ok, I've got bad and good binaries
<clever> [clever@amd-nixos:~/apps/nixpkgs-hasura]$ nix build -f boom-segfault.nix test2 --option builders 'pi@pi400 armv7l-linux,aarch64-linux /etc/nixos/keys/distro 4 1 big-parallel' --keep-going -L
<clever> [clever@amd-nixos:~/apps/nixpkgs-hasura]$ gist -p boom-segfault.nix
<clever> i have a good binary, and an expr to build it
<Gaelan> nice, that's more elegant than me
<clever> i just had to spam __bootPackages until it stopped trying to build autoconf
<clever> nice site
<clever> the size of a LOAD section changed, 8 bytes larger
<clever> .rel.dyn is 8 bytes larger?
<Gaelan> it's missing an R_ARM_COPY reloc for stderr
<Gaelan> (good is on the right btw)
<Gaelan> which seems consistent with what the patch does
<clever> why is stdout and stderr diff ...
<clever> without the patch, stderr and stdout both work
<Gaelan> stdout also fails
<clever> patch is back in, lets break it!
<clever> stdout works with the patch... ?
<clever> stderr also works...
<Gaelan> fprintf(stdout, "foo") fails with the patch. printf("foo") works with the patch
<Gaelan> stderr works with the patch for you?
<clever> yeah
<Gaelan> well that's bizzare
<clever> with the expr i gave
<clever> what about fwrite on stderr?
<clever> [clever@amd-nixos:~/apps/nixpkgs-hasura]$ git pull
<clever> Updating 1017eb06b19..70a061bd6c5
<clever> Gaelan: which nixpkgs rev did you test on?
<Gaelan> b3640e024f0
<Gaelan> the problematic commit itself
<Gaelan> i'm also using armv6l, if it matters
<Gaelan> although I thought I reproduced it on armv7l
<Gaelan> and all of my testing is on a 64-bit kernel on an RPi4
<clever> same, aarch64 pi400, with builds targetting armv7l
<clever> linuxHeaders fails, but test2 doesnt
justanotheruser has joined #nixos-aarch64
<Gaelan> …now I can't reproduce it
<Gaelan> what is going on
Darkmatter66 has joined #nixos-aarch64
<clever> gist updated
<clever> /nix/store/53bwzq4jpqz9im31gvf8iz3qzl24f13y-test1/bin/test1: ELF 32-bit LSB executable, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /nix/store/f2s8qh3f4crglqyqbwxdi9ijfa3kgysq-bootstrap-stage0-glibc/lib/ld-linux-armhf.so.3, for GNU/Linux 2.6.32, with debug_info, not stripped
<clever> this test1 runs successfully
<clever> [3/0/13 built, 0.0 MiB DL] building xz-5.2.5 on ssh://pi@pi400 (round 1/2)
<clever> ~13 builds until i can test it on b3640e024f0
orivej has joined #nixos-aarch64
Darkmatter66 has quit [Ping timeout: 272 seconds]
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
orivej has quit [Ping timeout: 240 seconds]
cole-h has joined #nixos-aarch64
orivej has joined #nixos-aarch64
<sphalerite> andi-++ that would have bitten me sooner or later
<{^_^}> andi-'s karma got increased to 53
<sphalerite> red[evilred]: experiments! :D All I know of is thefl*weringash's work, but I'd quite like to get a mainline kernel running
orivej has quit [Ping timeout: 260 seconds]
orivej has joined #nixos-aarch64
<sphalerite> red[evilred]: I need to go into quarantine when I'm home though, so I guess I'll have lots of time for experiments :p
<sphalerite> … if I can actually get the device :|
<sphalerite> UPS is pissing me off majorly right now
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-aarch64
cole-h has quit [Ping timeout: 260 seconds]
alpernebbi has joined #nixos-aarch64
noonien has joined #nixos-aarch64
<noonien> Perhaps this is a better place to ask
<noonien> I've managed to get a nixos installer image running on rpi from https://github.com/Robertof/nixos-docker-sd-image-builder However when trying to `nixos-rebuild` with this configuration.nix: http://ix.io/2KoA the pi doesn't seem to fully boot anymore (the green light stops flashing). Plugging it into a monitor shows that it seems to be stuck at
<noonien> "Starting kernel...", however, this is the case when booting the installation image as well, and that keeps working. The pi doesn't respond over wifi, nor over ethernet (to pings, ssh, nmap). Any tips?
<noonien> i managed to get the rpi3 to boot. turns out i bumped into https://github.com/NixOS/nixpkgs/issues/97064
<{^_^}> #97064 (by tobiasBora, 16 weeks ago, open): Raspberry Pi 3 B+ can't boot with latest kernel 5.7.19
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-aarch64
rajivr has quit [Quit: Connection closed for inactivity]
<samueldr> this issue seems like a catch-all issue AFAICT, where it's unlikely the initial issue is related
<samueldr> (I hate those catch-all issues)
<clever> samueldr: any idea on the linuxHeaders issue me and Gaelan are having?
<{^_^}> #107386 (by Gaelan, 1 week ago, open): Can't build linux-headers on armv6l
<samueldr> none at all
<samueldr> I'm not that good with compiler issues :)
veleiro has joined #nixos-aarch64
mschwaig1 has quit [Quit: WeeChat 2.7.1]
mschwaig has joined #nixos-aarch64
<simpson> I got an update from https://www.crowdsupply.com/eoma68/micro-desktop and it sounds like production might start soon. Finally, after years, another sunxi for my collection.
Darkmatter66 has joined #nixos-aarch64
alpernebbi has quit [Remote host closed the connection]
<samueldr> oh wow, EOMA68, been a while since I read that name
<samueldr> A20
Darkmatter66 has quit [Ping timeout: 265 seconds]
Darkmatter66 has joined #nixos-aarch64
Darkmatter66 has quit [Read error: Connection reset by peer]
Darkmatter66 has joined #nixos-aarch64
cole-h has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 260 seconds]
ryantrinkle has quit [Ping timeout: 272 seconds]
Darkmatter66 has quit [Ping timeout: 256 seconds]
elvishjerricco has quit [Ping timeout: 260 seconds]
ashkitten has quit [Ping timeout: 260 seconds]
elvishjerricco has joined #nixos-aarch64
ashkitten has joined #nixos-aarch64
justanotheruser has joined #nixos-aarch64
<Ke> so finally installed nixos on pbp
<Ke> what I do now is I nixos install to a subvol on mcbin, then send that subvol to pbp and rsync from the subvol to /nix
<Ke> isn't this ho nix is supposed to work?
<samueldr> I'm not sure I follow
<Ke> I could just omit the metadata and use only /nix/store directly
<samueldr> probably not a good idea, wouldn't things not be gc-rooted?
<samueldr> though it may work if you assume no use of nix on the other side?
<Ke> might break indeed, but if that happens, I have the config and I can just rebuild
<samueldr> this would be an off-label use of a nix build I guess
<samueldr> all warranties (none) all void :)
<Ke> in general, I thought I might want to do this with no nix-daemon and have rootfs in dm-verity
<Ke> if I had some secure boot platform
<samueldr> that probably would work
<samueldr> basically the nix store would be an implementation detail
<Ke> obviously when I say rootfs, I mean only /nix most other stuff is on tmpfs
<Ke> I have some minor files on /mnt/persist, but mostly nothing writable anyway, so I could have it on verity as well
<Ke> I also have the system of the pbp as a package linked to /etc on the mcbin, so reinstalling pbp system at any time is possible
<Ke> anyway, the 5.10 with the stripped down manjaro patches has stability issue for me
<samueldr> I still haven't booted the pinebook pro since whatever patch I last updated to on the PR
justanotheruser has quit [Ping timeout: 264 seconds]
<samueldr> pbb++ thanks for rebasing the u-boot raspberry pi 4 PR
<{^_^}> pbb's karma got increased to 3
justanotheruser has joined #nixos-aarch64
orivej has quit [Ping timeout: 260 seconds]
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-aarch64
ryantrinkle has joined #nixos-aarch64
justanotheruser has quit [Quit: WeeChat 2.9]
Acou_Bass has quit [Quit: ZNC 1.8.1 - https://znc.in]
Acou_Bass has joined #nixos-aarch64
fooker has quit [Remote host closed the connection]
fooker has joined #nixos-aarch64
veleiro has quit [Ping timeout: 272 seconds]