nixos/qemu-vm: use cfg.host.pkgs

[nikstur]

Description of changes

Hopefully fixes at least part of the issue described here: #236656 (comment)

I'm not sure if all these instances need to be exchanged for cfg.host.pkgs but it seems right to me.

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 23.11 Release Notes (or backporting 23.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

[emilazy]

Thank you! This looks right (dunno about the closureInfo but I'll defer to you). I'll look into e2fsprogs myself soon.

[RaitoBezarius]

This looks good indeed! Thank you

[nixos-discourse]

This pull request has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/darwin-builder-on-an-m1-mac-problems-with-mkfs-ext4/29523/3

[RaitoBezarius]

We need a Darwin person to confirm this is fixing darwin.builder.

[Gabriella439]

I confirmed locally that this PR fixes darwin.builder on aarch64-darwin. Specifically, I first tested that darwin.builder failed locally on master for the reason described in #236656 (comment) and then worked when I switched to this branch.

HOWEVER, I had to also wipe the old nixos.qcow2 image before it worked again.

[nixos-discourse]

This pull request has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/darwin-builder-on-an-m1-mac-problems-with-mkfs-ext4/29523/5

[nikstur]

@Gabriella439 thank you for testing

HOWEVER, I had to also wipe the old nixos.qcow2 image before it worked again.

Sadly, I don't think there is anything we can do about that.

[emilazy]

We need a Darwin person to confirm this is fixing darwin.builder.

It's still broken on x86_64 because of CoreFoundation rpath hook problems breaking e2fsprogs. I have a workaround for that but I'm waiting for the Darwin stdenv rework to see if it's necessary to PR it.

In other words, LGTM; this is more correct even if it doesn't fix the package on all platforms immediately.

[Atemu]

I was not able to run the created machine because it prints this error but creation works again which is what this PR's purpose is.

qemu-system-aarch64: -netdev user,id=user.0,hostfwd=tcp::22-:22,: Could not set up host forwarding rule 'tcp::22-:22'

[lilyball]

HOWEVER, I had to also wipe the old nixos.qcow2 image before it worked again.

I noticed recently that you have to do this when changing the disk space option too. I really wish the qemu-vm stuff would recognize when something changes that requires wiping the image, but that seems orthogonal to this particular issue.

[emilazy]

With #239473 maybe we can eliminate the persistent image altogether? But agreed that this PR should land as-is.

[nikstur]

With #239473 maybe we can eliminate the persistent image altogether?

It's a little more complicated than this. There are multiple and different kinds of persistent images used in the QEMU module. The erofs PR would only lay the groundwork to get rid of one of them (the Nix Store image built with make-disk-image.nix).

[Atemu]

LGTM.

Nit: The commit message could explain itself a bit better.

[Atemu]

Should this be backported?

[emilazy]

I think the changes that broke darwin.builder aren't in 23.05 so it's probably fine?