feat(reliability): risky-op preflight snapshots + detached rollback watchdog

payload/user/remote_access/pager-webui/server.py

@@ -3526,6 +3526,8 @@ def _apply_radio1_ap(openap, wpa):
 
 
 def h_pineap_wifi_set_ap(ctx):
+    import mk8_gate
+    mk8_gate.enter('ap_change')
     body = ctx.body or {}
     wpa = body.get('wpa') or {}
     openap = body.get('open') or {}
@@ -4350,6 +4352,8 @@ def _enterprise_boot_recover():
 
 
 def h_attacks_deploy(ctx):
+    import mk8_gate
+    mk8_gate.enter('attack_deploy')
     body = ctx.body or {}
     kind = (body.get('kind') or '').strip().lower()
     if kind not in ('wpa', 'open', 'enterprise'):
@@ -4439,6 +4443,8 @@ def _radio1_ap_active():
 
 
 def h_attacks_stop(ctx):
+    import mk8_gate
+    mk8_gate.enter('attack_stop')
     kind = ((ctx.body or {}).get('kind') or '')
     stopped = []
     if kind in ('wpa', 'open'):
@@ -4890,6 +4896,8 @@ def _disable_sta_uplink():
     keeps it off across reboots/wifi reloads; taking wlan0 down immediately
     frees phy0's channel for wlan0mon. Never runs `wifi reload` here — that
     tears down live APs and drops the monitors mid-assessment."""
+    import mk8_gate
+    mk8_gate.enter('uplink_disable')
     device_run(['uci', 'set', 'wireless.dummy_radio0.disabled=1'])
     device_run(['uci', 'commit', 'wireless'])
     for iface in ('wlan0',):
@@ -7189,6 +7197,8 @@ ROUTER.add('POST', r'/api/mode/release', h_mode_release)
 
 
 def serve():
+    import mk8_gate
+    mk8_gate.ENABLED = True
     LIVE_STOP.clear()
     HEALTH_STOP.clear()
     _recon_hopper_stop.set()
@@ -7255,6 +7265,26 @@ if __name__ == '__main__':
         except Exception as exc:  # boot must never fail here
             print(json.dumps({'error': str(exc)}))
         sys.exit(0)
+    if '--rollback-snapshot' in sys.argv:
+        import mk8_gate
+        mk8_gate.ENABLED = True
+        name = sys.argv[sys.argv.index('--rollback-snapshot') + 1]
+        import mk8_profiles
+        result = mk8_profiles.restore(name)
+        device_run(['wifi', 'reload'], timeout=90)
+        import mk8_events
+        mk8_events.log_event('rollback', sev='warn',
+                             msg='watchdog restored %s' % name,
+                             meta=result)
+        print(json.dumps(result))
+        sys.exit(0)
+    if '--promote-snapshot' in sys.argv:
+        import mk8_gate
+        mk8_gate.ENABLED = True
+        name = sys.argv[sys.argv.index('--promote-snapshot') + 1]
+        import mk8_profiles
+        print(json.dumps({'promoted': mk8_profiles.promote_lastknown_good()}))
+        sys.exit(0)
     signal.signal(signal.SIGTERM, _request_shutdown)
     signal.signal(signal.SIGINT, _request_shutdown)
     serve()