Automatically discover and verify ACE 2 on generic USB adapters

[Tareku99]

Summary

• Automatically discover ACE 2 units connected through generic USB-RS485 adapters, without entering adapter VID/PID or enabling a discovery option.
• Enable ACE 2 support and verified generic USB probing by default. Honor existing explicit false settings, and keep v2_probe_generic_usb: false as an optional disable switch.
• Enumerate stable USB serial links and direct USB tty paths, then verify candidates using bounded DISCOVER_DEVICE and GET_INFO exchanges. Require valid CRC, command, sequence, non-zero UID, and firmware identity.
• Exclude Klipper-configured serial paths and ports reported open by visible Linux processes. Suspend generic probes when the configured-port inventory or idle printer state cannot be established.
• Probe at most one new port per scan, rotate attempts across candidates, and back off after failures.
• Preserve genuine-cable, V1, and explicit v2_extra_usb_ids selection behavior.

ACELink context

ACELink informed the use of GET_INFO to confirm device identity. Its first-matching-adapter heuristic is not copied as the multiACE scanner. Automatic discovery and verification are handled by multiACE's existing V2 protocol code.

Validation and remaining work

• After rebasing and updating the package-default assertions: 7 managed package tests and 10 discovery tests passed locally. The workflow now runs the discovery tests as well.
• Python syntax parsing and git diff --check passed.
• Hardware validation remains for generic and genuine adapters, V1, multiple units, and reconnect behavior.
• Ownership checks are best-effort and exclusivity is advisory. Eligible USB serial ports receive identification bytes; behavior on unrelated hardware requires review and hardware validation before merging.
• Host permissions remain the platform's responsibility. Explicit extra USB IDs continue to bypass generic verification for compatibility.

Refs #154

[Tareku99]

Whipped something out really fast. Did not test this yet but I took the code from my own project (https://github.com/Tareku99/ACELink) and it worked fine there. I am out of time, but I wanted to make something before I left.

[decay71]

Thanks Nicola, that was quick. I'd like to stay with the opt-in from your first commit: please set the defaults of v2_probe_generic_usb and enable_ace_v2 back to false, but keep the improvements from 648c58a (suspend when the configured-port list is unknown, the probe rotation). A stock setup should never send bytes to a serial port it doesn't know.

Please also take out the exclusive=True in open_transport (that's the reconnect path we tested on hardware, unrelated to discovery) and the README change. Opening a port also toggles DTR, which resets some CH340 boards; could the probe open with DTR/RTS off?

No rush, this can wait until you're back. Safe travels.

Dirk

[Tareku99]

Thanks Nicola, that was quick. I'd like to stay with the opt-in from your first commit: please set the defaults of v2_probe_generic_usb and enable_ace_v2 back to false, but keep the improvements from 648c58a (suspend when the configured-port list is unknown, the probe rotation). A stock setup should never send bytes to a serial port it doesn't know.

Please also take out the exclusive=True in open_transport (that's the reconnect path we tested on hardware, unrelated to discovery) and the README change. Opening a port also toggles DTR, which resets some CH340 boards; could the probe open with DTR/RTS off?

No rush, this can wait until you're back. Safe travels.

Dirk

Sounds good. Thank you for the feedback, shall take care of this when back in 2 weeks or so.
Thank you!