fix: waterfall device claim fails on frequency change due to USB release lag

When restarting capture for a new frequency, the USB handle from the
just-killed process may not be released by the kernel in time for the
rtl_test probe inside claim_sdr_device. Add retry logic (up to 4
attempts with 0.4s backoff) matching the pattern already used by the
audio start endpoint.

Also clean up stale shared-monitor state in the frontend error handler
so the monitor button is not left disabled when the capture restart
fails.

routes/waterfall_websocket.py

@@ -462,8 +462,16 @@ def init_waterfall_websocket(app: Flask):
                     start_freq = center_freq_mhz - effective_span_mhz / 2
                     end_freq = center_freq_mhz + effective_span_mhz / 2
 
-                    # Claim the device
+                    # Claim the device (retry when restarting to allow
-                    claim_err = app_module.claim_sdr_device(device_index, 'waterfall')
+                    # the kernel time to release the USB handle).
+                    max_claim_attempts = 4 if was_restarting else 1
+                    claim_err = None
+                    for _claim_attempt in range(max_claim_attempts):
+                        claim_err = app_module.claim_sdr_device(device_index, 'waterfall')
+                        if not claim_err:
+                            break
+                        if _claim_attempt < max_claim_attempts - 1:
+                            time.sleep(0.4)
                     if claim_err:
                         ws.send(json.dumps({
                             'status': 'error',

static/js/modes/waterfall.js

@@ -2488,6 +2488,17 @@ const Waterfall = (function () {
                 } else if (msg.status === 'error') {
                     _running = false;
                     _scanStartPending = false;
+                    _pendingSharedMonitorRearm = false;
+                    // If the monitor was using the shared IQ stream that
+                    // just failed, tear down the stale monitor state so
+                    // the button becomes clickable again after restart.
+                    if (_monitoring && _monitorSource === 'waterfall') {
+                        clearTimeout(_monitorRetuneTimer);
+                        _monitoring = false;
+                        _monitorSource = 'process';
+                        _syncMonitorButtons();
+                        _setMonitorState('Monitor stopped (waterfall error)');
+                    }
                     if (_scanRunning) {
                         _scanAwaitingCapture = true;
                         _setScanState(msg.message || 'Waterfall retune error, retrying...', true);