refactor: introduce Transport abstraction and sysfs backend

[hagz0r]

This PR refactors the cd321x device logic to be transport-agnostic by introducing
a Transport trait, and moves the existing sysfs-based implementation into a
dedicated backend.

Motivation:

• decouple protocol/device logic from the concrete I/O mechanism
• make it possible to add alternative transports in the future
  (e.g. mock/testing, userspace bridges, other host OSes)
• reduce cd321x-specific filesystem handling in higher-level code

Changes

• introduce a Transport trait defining read/write semantics
• rework cd321x::Device to operate on a generic Transport
• move current sysfs implementation into transport::sysfs
• no functional behavior changes intended

This is a pure refactor; behavior and command formats are unchanged.

Tested by building and running the tool against the existing sysfs backend.

[jannau]

Hej @hagz0r,
I was unfortunately not watching the repo and missed this PR. I will merge #4 first to drop the sysfs interface immediately. This will cause conflicts. If you're not motivated to fix them I can do it myself as penance for missing the PR. Please sign you commit with Developer Certificate of Origin so I can base it on your commit. The refactoring itself looks good, thanks I'm sorry for noticing it too late.

[hagz0r]

Hi Janne, Thank you for the update. I am motivated to continue maintaining this and will handle the conflicts resulting from merging #4. I will also make sure to sign my commits with the Developer Certificate of Origin as requested. Best regards, Hagz0r

…

On Tue 20. 1. 2026 at 23:17, Janne Grunau ***@***.***> wrote: *jannau* left a comment (AsahiLinux/tuxvdmtool#3) <#3 (comment)> Hej @hagz0r <https://github.com/hagz0r>, I was unfortunately not watching the repo and missed this PR. I will merge #4 <#4> first to drop the sysfs interface immediately. This will cause conflicts. If you're not motivated to fix them I can do it myself as penance for missing the PR. Please sign you commit with Developer Certificate of Origin <https://github.com/AsahiLinux/tuxvdmtool/pull/3/checks?check_run_id=59409538864> so I can base it on your commit. The refactoring itself looks good, thanks I'm sorry for noticing it too late. — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ALRGCA4YZUSYSULG3QVXWWL4HZWJZAVCNFSM6AAAAACQU3B3OSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTONZUGMYDANZRGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[jannau]

thanks, please note that @bjarne-dietrich has already started looking into converting macvdmtool. I hope the work is mostly orthogonal so far and it can integrate with this transport refactoring.

[bjarne-dietrich]

Hey Hagz0r,

Sorry, I didnt see your PR before. I will continue working on the macOS implementation meanwhile and integrate it once you are done.
Bjarne

[hagz0r]

I wanted to clarify the testing situation for this PR.

At the moment I’m working on macOS and I’m also tied up with work commitments, so I can’t realistically run full Linux-side testing right now. Because of that, I’m currently limited to compile-time checks only. I should be able to properly test the changes on Linux hardware after the 26th.

If full testing is required sooner, I’d appreciate it if the maintainers could verify the changes on their side (e.g. via CI or a local Linux setup).