torlando-tech
0b10a08b93
fix(ble): Add connection race condition prevention and improve error handling
...
Implements comprehensive connection state tracking to prevent "Operation
already in progress" errors and connection retry storms.
BLE Interface changes:
- Record connection attempts before calling driver.connect()
- Add 5-second rate limiting between attempts to same peer
- Skip connections already in progress via _connecting_peers check
- Downgrade expected race conditions to DEBUG level
- Auto-blacklist MAC addresses on connection failures
- Add diagnostic logging for concurrent connection tracking
BLE Driver changes:
- Add _connecting_peers set to track in-progress connections
- Prevent concurrent connection attempts to same address
- Attach cleanup callbacks to connection Futures
- Add defense-in-depth cleanup in finally blocks
- Detailed logging for connection state debugging
Documentation updates:
- Add deployment workflow documentation to README.md
- Update .github/workflows/README.md with CD workflow details
- Document containerized runner SSH configuration
- Update reference documentation (CLAUDE.md, BLE_PROTOCOL, etc.)
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-11-07 22:32:00 -05:00
torlando-tech
4eaa9a4173
Refactor BLEInterface to driver-based architecture
...
Major architectural refactoring to separate high-level Reticulum protocol
logic from platform-specific Bluetooth operations. This enables code sharing
between pure Python and Android (Columba) implementations, improves
testability, and creates a clean boundary for future platform support.
ARCHITECTURE CHANGES:
1. **Driver Abstraction Layer**
- Created BLEDriverInterface (bluetooth_driver.py) defining the contract
for all platform-specific BLE drivers
- Abstraction includes 18 methods + 6 callbacks for complete BLE lifecycle
- Enhanced BLEDevice dataclass with service_uuids and manufacturer_data
- Added on_mtu_negotiated callback for delayed MTU reporting
- Added on_error callback for consistent platform error reporting
2. **Linux Driver Implementation**
- Created LinuxBluetoothDriver (linux_bluetooth_driver.py, 1534 lines)
- Moved ALL bleak/bluezero/D-Bus code from BLEInterface
- Preserves 5 critical platform workarounds:
* BlueZ ServicesResolved race condition patch
* D-Bus LE-only connection (ConnectDevice)
* BLE Agent registration for Just Works pairing
* MTU negotiation with 3-method fallback
* Service discovery delay for bluezero timing
- Role-aware send() automatically chooses GATT write vs notification
- Dedicated asyncio event loop management in separate thread
- Configuration via constructor (no Reticulum dependencies)
3. **Refactored BLEInterface**
- Removed 801 lines (32.3% reduction: 2479 → 1678 lines)
- Removed all platform-specific imports (bleak, bluezero, dbus_fast)
- Removed 9 async methods (moved to driver)
- Driver dependency injection via constructor
- Implemented 6 driver callbacks for event handling
- PRESERVED high-level logic:
* Peer scoring algorithm (RSSI + history + recency)
* Connection blacklist with exponential backoff
* MAC-based connection direction (prevents dual connections)
* Fragmentation/reassembly orchestration (identity-based keying)
* Interface spawning per peer
4. **Simplified BLEPeerInterface**
- Removed connection_type, client, mtu parameters
- Deleted _send_via_central() and _send_via_peripheral() methods
- Single send path via driver.send() (driver handles role routing)
- 77 lines removed from peer interface class
5. **Mock Driver for Testing**
- Created MockBLEDriver (tests/mock_ble_driver.py)
- Complete BLEDriverInterface implementation without hardware
- Bidirectional communication via link_drivers()
- Enables unit testing of BLEInterface logic (fragmentation, reassembly,
peer lifecycle, blacklist management)
CRITICAL FIXES:
1. **Restored Periodic Cleanup Task** (CRITICAL: prevents memory leaks)
- Converted from async (driver-owned loop) to threading.Timer
- Runs every 30 seconds to clean stale reassembly buffers
- Essential for long-running instances (Pi Zero with 512MB RAM)
- Properly cancelled in detach() for clean shutdown
2. **Fixed Naming Consistency**
- Renamed processOutgoing → process_outgoing (snake_case)
FILES MODIFIED:
- src/RNS/Interfaces/BLEInterface.py (refactored, -801 lines)
FILES ADDED:
- bluetooth_driver.py (driver abstraction interface)
- linux_bluetooth_driver.py (Linux/BlueZ implementation, 1534 lines)
- tests/mock_ble_driver.py (mock driver for unit tests)
- REFACTORING_GUIDE.md (comprehensive refactoring documentation)
- BLE_PROTOCOL_v2.2.md (protocol specification)
- tests/test_refactor_suite.py (initial test suite)
BENEFITS:
1. **Testability** - Mock driver enables hardware-free unit testing
2. **Portability** - Easy to create Android/Windows/macOS drivers
3. **Maintainability** - Platform quirks isolated in single driver file
4. **Code Sharing** - High-level logic shared across all platforms
5. **Clean Architecture** - Clear separation of concerns
TESTING REQUIRED:
- Tier 1 (Unit): Test with MockBLEDriver (fragmentation, reassembly, lifecycle)
- Tier 2 (Integration): Test on Raspberry Pi hardware (scanning, connecting,
dual mode, MTU negotiation, identity exchange)
- Tier 3 (Regression): Full Reticulum stack (announces, LXMF, multi-hop)
- Tier 4 (Edge Cases): MAC rotation, identity handshake, reconnection,
reassembly timeout, discovery cache pruning
BACKWARD COMPATIBILITY:
- Configuration: Fully backward compatible (same config parameters)
- Protocol: No changes to BLE wire protocol (v2.2)
- Interface API: Unchanged for Reticulum Transport integration
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com>
2025-11-03 23:15:22 -05:00
torlando-tech