jyapayne/switch-pico
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Preserve analog triggers in generic HID modes

Browse source
This commit is contained in:
Joey Yakimowich-Payne 2026-09-03 11:51:30 -06:00
commit ecd8ca2b40
9 changed files with 252 additions and 130 deletions
Download patch file Download diff file Expand all files Collapse all files

2
README.md
View file

@ -183,7 +183,7 @@ On macOS, inspect the firmware's raw Game Pad values before GameController or br
swift tools/Test-SwitchPicoMac.swift
```
The diagnostic matches only `CAFE:4021`, identifies each of the four interfaces by interface and location, and prints changed axes, hats, and buttons with their HID usage and logical range. Move each analog trigger slowly and confirm that `Ry`/`Rz` report intermediate values across `0...65535`, rather than only the endpoints. It continues through hot-plug events until Ctrl-C. If opening a device fails, allow the terminal (or the app launching Swift) under **System Settings → Privacy & Security → Input Monitoring**, then rerun it.
The diagnostic matches only `CAFE:4021`, identifies each of the four interfaces by interface and location, and prints changed axes, hats, and buttons with their HID usage and logical range. The four signed stick axes remain `X`/`Y`/`Z`/`Rx`. Move each analog trigger slowly and confirm output such as `LeftBrake page=0x02 usage=0xC5 logical=0...65535 value=32768` and `RightAccelerator page=0x02 usage=0xC4 logical=0...65535 value=32768`; each trigger should traverse intermediate values across `0...65535`, not only the endpoints. The diagnostic continues through hot-plug events until Ctrl-C. If opening a device fails, allow the terminal (or the app launching Swift) under **System Settings → Privacy & Security → Input Monitoring**, then rerun it.
On the tested Linux host, all four HID interfaces enumerated, but `hid-nintendo` timed out (`-110`) while requesting controller information from the composite device and removed the transient hidraw nodes. This is an observed, undiagnosed composite interoperability limitation; its root cause has not been established. The timeout was not observed on the Switch, so successful `hid-nintendo` binding is not the release criterion for the four-interface AIO firmware. The pairing CLI uses vendor control transfers on endpoint 0 and does not depend on those hidraw nodes.