nixos: Support fileSystems.<name>.depends with fstab-generator

[ElvishJerricco]

Description of changes

I have barely tested this, so I'm not entirely sure it's correct. Closes #217179

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/)
• 23.05 Release Notes (or backporting 22.11 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.

[RaitoBezarius]

FYI, in the diskless VM PR, I will introduce this definitely and remove any manual mount.

[astro]

What is the diskless VM PR? It got referenced below.

[ElvishJerricco]

This would fix #217179

[ElvishJerricco]

@nikstur I updated this with a little bit of tweaking to the overlay fs stuff. The depends functionality should be the same on systemd initrd, with slightly better behavior in scripted initrd. While testing with scripted, I noticed it doesn't work with the ro overlays, so I just added an assertion to disallow that, rather than trying to implement it.

(I also nixfmt'd some of this, so forgive the reformatting).

[philiptaron]

I'll note that ofBorg is unhappy for the same reason that I created #320846: the value of the message depends on the assertion failing, and as written it doesn't protect against that being so.

[ElvishJerricco]

the value of the message depends on the assertion failing, and as written it doesn't protect against that being so.

@philiptaron No, that's not what's causing ofborg to fail. I apparently committed the assertion without testing that it works correctly. It should be fs.overlay.upperdir, not fs.upperdir. My bad. The message only depends on fs.mountPoint, which should always exist.

[philiptaron]

This looks great. I'm going to give it a whirl on my home machine and report back.

[philiptaron]

Tested and works.