cudaSDR/Source/NEXT_STEPS.md

Next Steps

Priority

1. Investigate Protocol 2 stop handling.
2. Validate Protocol 2 wideband behavior on real hardware.
3. Complete Protocol 2 Alex support for boards with Alex1.

Immediate Issues To Debug

Protocol 2 stop crash

• Re-test why the application crashes when the user presses Stop while working with a Protocol 2 device.
• Re-verify why the Protocol 2 stop command may still not be correctly sent to the transceiver.
• Check the shutdown path across:
  • src/DataEngine/cusdr_dataEngine.cpp
  • src/DataEngine/cusdr_dataIO.cpp
  • src/DataEngine/cusdr_protocol2_io.cpp
• Already fixed one likely cause in src/DataEngine/cusdr_protocol2_io.cpp:
  • duplicate P2 stop send,
  • stop without m_running,
  • timer/socket signal disconnect before deletion.
• Already fixed a second class of issues in src/DataEngine/cusdr_dataEngine.cpp:
  • DataIO control calls now go through QMetaObject::invokeMethod(..., Qt::BlockingQueuedConnection),
  • so start/stop/init execute in the DataIO thread instead of the caller thread.
• Already fixed a concrete Protocol 2 destructor bug in src/DataEngine/cusdr_dataIO.cpp:
  • m_dataIOSocketOn == true with m_dataIOSocket == 0 could dereference a null pointer during shutdown.
• Already moved CSoundOut deletion out of DataIO::stop() and into the destructor to reduce timer/thread teardown issues.
• Pay special attention to:
  • socket lifetime,
  • timer lifetime,
  • QObject ownership and deletion thread,
  • repeated stop calls,
  • networkDeviceStartStop(0),
  • Protocol2DataPath::stop(),
  • sendHighPriorityPacket(false),
  • whether the command socket is still valid when stop is triggered,
  • whether DataIO / Protocol2DataPath should be destroyed via deleteLater() in their own thread instead of direct delete from DataEngine,
  • whether the target P2 board expects additional shutdown packets or repeated run = 0 frames.

Protocol 2 Functional Work

Protocol 2 wideband runtime verification

• Re-test Protocol 2 wideband on real hardware after the 2026-04-14 integration.

• Verify that pressing the UI Wideband button now really causes the transceiver to:

  • start sending wideband UDP data,
  • stop sending wideband UDP data.

• Verify that this works while the device is already running, without requiring a full stop/start cycle.

• Verify the incoming packet format and pacing match what the implementation now assumes:

  • UDP base port 1027,
  • sequence number in bytes 0..3,
  • 16-bit raw ADC samples starting at byte 4,
  • one full spectrum block formed from Settings::getWidebandBuffers() packets.

• Pay special attention to boards that may not follow the default 32 packets per frame behavior.

• If a board reports or requires fixed wideband constraints, use the protocol document/device profile to specialize the current defaults.

• Implement proper Alex1/BPF2/RX2 handling for boards that support a second Alex register:

  • Orion MkII
  • SATURN / ANAN-G2

• Use Appendix D from doc/openHPSDR Ethernet Protocol v4.3.docx to map the second Alex register bits correctly.

• Consider extending Protocol 2 capability handling beyond board defaults where discovery returns more precise information.

• Consider support for XML/full hardware description discovery replies for board 254/255.

UI Work

• Review UI areas that still assume Metis/Hermes semantics internally.
• Confirm the Wideband button behavior is now correct from the user point of view:
  • button state,
  • dock visibility,
  • widebandData setting,
  • real hardware stream enable/disable.
• Make Protocol 2 model/capability presentation more explicit where useful:
  • selected model,
  • protocol version,
  • ADC/DDC count,
  • Alex availability.

Reference

• Detailed log of completed work:
  • WORKLOG_PROTOCOL2_2026-04-13.md