]> git.donarmstrong.com Git - kiibohd-controller.git/commit
Major USB update, fixes most (if not all) known issues
authorJacob Alexander <haata@kiibohd.com>
Tue, 31 May 2016 07:19:45 +0000 (00:19 -0700)
committerJacob Alexander <haata@kiibohd.com>
Tue, 31 May 2016 07:40:10 +0000 (00:40 -0700)
commit358d480ea9862106f851cbb68c1d68da7369ddd5
tree9ccd3ed547f2283fbc53a13a1febe949874a42dd
parent8a96e70b4bedf23625fad1d66502775159b67d87
Major USB update, fixes most (if not all) known issues

USB - General
- Refactored descriptors
- Enabled/Disable USB endpoints
- Added debug flags for special features
- Code cleanup
- Interface count calculation based off of enabled endpoints
- Delayed wTotalLength calculation to simplify descriptor offsets
- Re-ordered endpoints and interfaces
- Added more debug output
- Added usbInitTime to show how long keyboard initialization took
  (Useful when debugging bad init sequences)
- Added function for usb_resume() which takes care of the resume sequence
  * Resume is now only called if packets are starting to timeout

USB - Special Options
- Added enableDeviceRestartOnUSBTimeout
  * A last resort hammer for bad USB Chipsets/OSs, don't use if you can help it
  * Disabled
- Added enableUSBResume
  * Enables host resume wake-up signalling, required to wake a computer from sleep
  * Enabled
- Added enableUSBLowPowerNegotiation
  * Enables power negotiation hack
  * Required to use firmware with an IPad and other hard-limit low-power USB hosts
  * Hasn't been tested with the recent changes
  * Disabled
- Added enableUSBSuspend
  * Enables power down events on host USB bus suspend
  * Enabled

USB - Keyboard
- Attempted to cleanup HID SET_REPORT
  * Works much better
  * Still has an issue under Linux which generates *a lot* of NAKs (initializes quickly regardless)
    + Not present on other keyboards
    + SETUP -> OUT -> IN : This sequence is the problem
    + Specifically during the OUT phase
- Enabled

USB - CDC Virtual Serial Port
- Code cleanup
- Added convenience struct USBCDCLineCoding for easier debugging
- Attempted to cleanup CDC_SET_LING_CODING
  * Works much better
  * Still has an issue under Linux which generates *a lot* of NAKs (initializes quickly regardless)
    + SETUP -> OUT -> IN : This sequence is the problem
    + Specifically during the OUT phase
    + Likely the same issues as HID SET_REPORT
- Enabled

USB - Mouse
- Enabled

USB - Joystick
- Disabled

USB - RawIO
- Initial code, API not used yet
- Disabled

DFU
- Updated load script, now faster
18 files changed:
Debug/print/print.c
LoadFile/load.dfu
Output/pjrcUSB/arm/usb_desc.c
Output/pjrcUSB/arm/usb_desc.h
Output/pjrcUSB/arm/usb_dev.c
Output/pjrcUSB/arm/usb_dev.h
Output/pjrcUSB/arm/usb_joystick.c
Output/pjrcUSB/arm/usb_keyboard.c
Output/pjrcUSB/arm/usb_mouse.c
Output/pjrcUSB/arm/usb_rawio.c [new file with mode: 0644]
Output/pjrcUSB/arm/usb_rawio.h [new file with mode: 0644]
Output/pjrcUSB/arm/usb_serial.c
Output/pjrcUSB/arm/usb_serial.h
Output/pjrcUSB/capabilities.kll
Output/pjrcUSB/output_com.c
Output/pjrcUSB/output_com.h
Output/pjrcUSB/setup.cmake
Output/usbMuxUart/output_com.c