Stability: WDT yield, BLE mutex fixes, time-based desync recovery

Reduces crash rate from every 60-85s to 1 reboot per 6+ minutes.
Zero WDT triggers in 10-minute stability test.

BLE mutex fixes (BLEInterface.cpp):
- Release _mutex before blocking GATT ops in onConnected() and
  onServicesDiscovered() — prevents 5-30s main-loop stalls during
  service discovery, notification subscribe, identity exchange
- Non-blocking try_lock() for peerCount(), getConnectedPeerSummaries(),
  get_stats() — returns empty/default if BLE task holds mutex
- Write-without-response in initiateHandshake()

WDT and persistence (main.cpp, sdkconfig.defaults, microReticulum):
- 30s WDT timeout (up from 10s) for SPIFFS flash I/O headroom
- Register Identity::set_persist_yield_callback() to feed WDT every
  5 entries during save_known_destinations() (70+ entries = 30-50s)
- WDT feeds between reticulum and identity persist calls

BLE host desync recovery (NimBLEPlatform):
- Time-based desync tracking instead of aggressive counter-based reboot
- 60s tolerance without connections, 5 minutes with active connections
  (data still flows over existing BLE mesh links)
- Remove immediate recoverBLEStack() from 574 handler and
  enterErrorRecovery() — let startScan() manage reboot decision
- Increase CONNECTION_COOLDOWN from 3s to 10s to reduce 574 risk
- Increase SCAN_FAIL_RECOVERY_THRESHOLD from 5 to 10

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src/main.cpp

@@ -1194,11 +1194,15 @@ void setup() {
     INFO("╚══════════════════════════════════════╝");
     INFO("");
 
-    // Subscribe main loop to Task Watchdog — detects hangs/deadlocks
-    // If loop() blocks for >10s (CONFIG_ESP_TASK_WDT_TIMEOUT_S), WDT fires
-    // with a backtrace showing exactly where the hang is
-    esp_task_wdt_add(NULL);  // NULL = current task (loopTask)
-    INFO("Task Watchdog: loopTask subscribed");
+    // Reconfigure Task Watchdog with 30s timeout (default 10s is too tight
+    // for SPIFFS flash I/O — identity persistence writes 40-50 entries and
+    // can take 5-15s with sector erases and garbage collection)
+    esp_task_wdt_init(30, true);  // 30s timeout, panic on trigger
+    esp_task_wdt_add(NULL);       // Subscribe loopTask
+    INFO("Task Watchdog: loopTask subscribed (30s timeout)");
+
+    // Feed WDT during long Identity persistence (71+ entries to SPIFFS can take >30s)
+    Identity::set_persist_yield_callback([]() { esp_task_wdt_reset(); });
 
     // Show startup message
     INFO("Press any key to start messaging");
@@ -1275,10 +1279,15 @@ void loop() {
     reticulum->loop();
 
     // Periodically persist identity/transport data (display names, paths, etc.)
+    // NOTE: Identity persistence writes 40-50 entries to SPIFFS flash, which
+    // involves sector erases (100ms each) and can take 5-15s total.
+    // WDT feeds between calls prevent timeout during heavy flash I/O.
     LOOP_STEP(5);  // persist data
     reticulum->should_persist_data();
+    esp_task_wdt_reset();
     // Fast-persist known destinations (5s after dirty) to survive crashes
     Identity::should_persist_data();
+    esp_task_wdt_reset();
 
     // Process TCP interface
     LOOP_STEP(6);  // TCP loop