fix: turn VPN off on sleep/resume and unload; harden teardown

- Detect resume (boottime-monotonic jump) and tear down the tunnel: on wake
  NetworkManager drops Mudfish's relay route and all traffic black-holes.
- Turn the tunnel off on plugin unload (Deck shutdown/restart).
- One verified teardown path for disconnect, connect failure/timeout and
  node-change failure; keep the route snapshot if tap0 stays up.
- Toggle can always turn off a stuck tunnel; status failure with tap0 up
  is reported as an error, not off.
- Re-sign in once when the cloud rejects an expired jwt.
- Expire cached pings after 10 minutes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

main.py

@@ -12,6 +12,8 @@ from mudfish_decky.client import MudfishClient  # noqa: E402
 from mudfish_decky.controller import Controller  # noqa: E402
 from mudfish_decky.store import Store  # noqa: E402
 
+RESUME_POLL_SECONDS = 2
+
 
 class Plugin:
     controller = None
@@ -44,9 +46,18 @@ class Plugin:
         log.setLevel(logging.INFO)
         for handler in decky.logger.handlers:
             log.addHandler(handler)
+        controller = self._get_controller()
         # Bring up the idle launcher and sign in so nodes are listed. Never starts the tunnel.
-        await asyncio.to_thread(self._get_controller().prepare)
+        await asyncio.to_thread(controller.prepare)
         decky.logger.info("Mudfish VPN plugin loaded")
+        # Watch for sleep/resume even while the panel is closed; a resumed tunnel is turned off.
+        while True:
+            try:
+                await asyncio.to_thread(controller.check_resume, system.sleep_offset())
+            except Exception as e:
+                decky.logger.warning("resume check failed: %s", e)
+            await asyncio.sleep(RESUME_POLL_SECONDS)
 
     async def _unload(self):
+        await asyncio.to_thread(self._get_controller().shutdown)
         decky.logger.info("Mudfish VPN plugin unloaded")

py_modules/mudfish_decky/client.py

@@ -27,6 +27,10 @@ class SigninError(MudfishError):
     pass
 
 
+class AuthError(MudfishError):
+    """The cloud API rejected our jwt (expired session); a forced re-signin may fix it."""
+
+
 # Decky's bundled Python ships without a CA store, so point it at SteamOS's system bundle.
 SYSTEM_CA_BUNDLES = ["/etc/ssl/certs/ca-certificates.crt", "/etc/ca-certificates/extracted/tls-ca-bundle.pem"]
 
@@ -84,18 +88,27 @@ class MudfishClient:
     def status(self) -> dict:
         return self._json("GET", self.base_url + "/do/mudrun/status")
 
-    def ensure_signed_in(self) -> dict:
-        """Replay the launcher's remember-me autologin when the daemon has no session."""
+    def ensure_signed_in(self, force: bool = False) -> dict:
+        """Replay the launcher's remember-me autologin when the daemon has no session (or force=True)."""
         st = self.status()
-        if st.get("isSignedIn"):
+        if st.get("isSignedIn") and not force:
             return st
-        if not st.get("autologin") and not st.get("token_v3"):
+        if not st.get("token_v3") or not st.get("username"):
             raise SigninError("no saved login")
         body, ctype = _multipart({
             "autologin": "ON", "username": st["username"], "password": st["token_v3"],
-            "item_conf_mode": st["item_conf_mode"], "version": st["version"],
+            "item_conf_mode": st.get("item_conf_mode", ""), "version": st.get("version", ""),
         })
-        signin = self._json("POST", f"https://{st['api_server']}/signin_ng_v3", body, {"Content-Type": ctype})
+        url = f"https://{st['api_server']}/signin_ng_v3"
+        code, data = self.transport("POST", url, body, {"Content-Type": ctype})
+        if code in (401, 403):
+            raise SigninError(f"signin HTTP {code}")
+        if not 200 <= code < 300:
+            raise MudfishError(f"HTTP {code} from {url}")
+        try:
+            signin = json.loads(data)
+        except ValueError:
+            raise MudfishError(f"invalid response from {url}")
         if signin.get("status") != 200:
             raise SigninError(f"signin status {signin.get('status')}")
         auth = self._post_json(self.base_url + "/do/auth", {
@@ -120,12 +133,24 @@ class MudfishClient:
     def rtts(self) -> dict:
         return rtts_from_response(self._json("GET", self.base_url + "/do/rtt/staticnodes"))
 
-    def _graphql_user(self, jwt, api, fields) -> dict:
-        resp = self._post_json(f"https://{api}/graphql", {"query": "{ user { %s } }" % fields}, {"Authorization": jwt})
+    def _cloud(self, jwt, url, payload) -> dict:
+        code, data = self.transport("POST", url, json.dumps(payload).encode(),
+                                    {"Content-Type": "application/json", "Authorization": jwt})
+        if code in (401, 403):
+            raise AuthError(f"HTTP {code} from {url}")
+        if not 200 <= code < 300:
+            raise MudfishError(f"HTTP {code} from {url}")
         try:
-            return resp["data"]["user"]
-        except (KeyError, TypeError):
-            raise MudfishError("graphql returned no user")
+            return json.loads(data)
+        except ValueError:
+            raise MudfishError(f"invalid response from {url}")
+
+    def _graphql_user(self, jwt, api, fields) -> dict:
+        resp = self._cloud(jwt, f"https://{api}/graphql", {"query": "{ user { %s } }" % fields})
+        user = (resp.get("data") or {}).get("user") if isinstance(resp, dict) else None
+        if not user:
+            raise AuthError("graphql returned no user")
+        return user
 
     def static_nodes(self, jwt, api) -> list:
         return self._graphql_user(jwt, api, "staticNodes { sid location }")["staticNodes"]
@@ -134,7 +159,7 @@ class MudfishClient:
         return self._graphql_user(jwt, api, "progConf { %s }" % " ".join(PROG_CONF_FIELDS))["progConf"]
 
     def save_prog_conf(self, jwt, api, conf: dict) -> dict:
-        resp = self._post_json(f"https://{api}/prog", conf, {"Authorization": jwt})
+        resp = self._cloud(jwt, f"https://{api}/prog", conf)
         if resp.get("status") != 200:
             raise MudfishError(resp.get("msg") or f"prog status {resp.get('status')}")
         return resp

py_modules/mudfish_decky/controller.py

@@ -1,7 +1,8 @@
 """Connect/disconnect orchestration and the state contract the panel renders.
 
 The core tunnel is only ever started from connect() (or re-applied by set_node() while already on);
-state() and plugin load must never start it, because Full VPN traffic is billed per GB.
+state(), prepare(), resume handling and shutdown must never start it, because Full VPN traffic is
+billed per GB. Every path that may leave a tunnel behind goes through _teardown().
 """
 import dataclasses
 import logging
@@ -9,7 +10,7 @@ import threading
 import time
 from typing import Optional
 
-from .client import SigninError
+from .client import AuthError, SigninError
 from .nodes import Node, japan_nodes, pick_auto
 from .routes import host_routes, leftover_routes
 
@@ -17,10 +18,15 @@ MSG_RELOGIN = "데스크탑 모드에서 미꾸라지(127.0.0.1:8282)에 '로그
 MSG_RTT = "노드 핑을 측정하는 중이에요. 잠시 후 다시 켜 주세요"
 MSG_TIMEOUT = "30초 안에 연결되지 않아 연결을 취소했어요"
 MSG_MISSING = "데스크탑 모드에서 미꾸라지를 설치하세요 (docs.mudfish.net Steam Deck 가이드)"
+MSG_STILL_UP = "터널이 아직 남아 있어요. 끄기를 다시 눌러 주세요"
+MSG_NO_RESPONSE = "미꾸라지가 응답하지 않아요. 터널이 남아 있을 수 있으니 끄기를 눌러 주세요"
+MSG_SLEEP = "절전에서 깨어나 VPN을 껐어요. 필요하면 다시 켜 주세요"
 
 CONNECT_TIMEOUT = 30
 DISCONNECT_TIMEOUT = 10
 PUBLIC_IP_TTL = 30
+RTT_TTL = 600          # cached pings older than this are ignored for the auto pick
+RESUME_JUMP = 3.0      # seconds of suspended time that count as a sleep/resume
 
 log = logging.getLogger("mudfish_decky")
 
@@ -34,8 +40,10 @@ class Controller:
         self.sleep = sleep
         self._lock = threading.Lock()
         self._error: Optional[str] = None
+        self._notice: Optional[str] = None
         self._static: Optional[list] = None
         self._ip_cache = (0.0, None)
+        self._sleep_offset: Optional[float] = None
 
     # ---- public API -------------------------------------------------------------------------
 
@@ -62,10 +70,11 @@ class Controller:
 
     def connect(self) -> dict:
         with self._lock:
-            self._error = None
+            self._error = self._notice = None
             installed = self.system.find_mudfish()
             if not installed:
                 return self._state()
+            started = False
             try:
                 st = self._signed_in(installed)
                 if self._is_on(st):
@@ -74,44 +83,35 @@ class Controller:
                 node = self._target_node()
                 if node is None:
                     return self._fail(MSG_RTT)
-                self._apply_sid(st, node.sid)
+                self._with_auth_retry(st, lambda s: self._apply_sid(s, node.sid))
                 self.store.update(routes_before=sorted(host_routes(self.system.route_text())), active_sid=node.sid)
+                started = True
                 self.client.start()
                 if not self._wait(lambda: self._is_on(self.client.status()), CONNECT_TIMEOUT):
-                    self.client.stop()
-                    return self._fail(MSG_TIMEOUT)
+                    return self._fail(MSG_TIMEOUT if self._teardown() else MSG_STILL_UP)
                 self.store.update(connected_since=self.clock())
             except SigninError:
-                return self._fail(MSG_RELOGIN)
+                return self._abort(started, MSG_RELOGIN)
             except Exception as e:
                 log.exception("connect failed")
-                return self._fail(f"연결 실패: {e}")
+                return self._abort(started, f"연결 실패: {e}")
             return self._state()
 
     def disconnect(self) -> dict:
         with self._lock:
-            self._error = None
-            try:
-                self.client.stop()
-                self._wait(lambda: not self.system.tap_up(), DISCONNECT_TIMEOUT)
-            except Exception as e:
-                log.exception("disconnect failed")
-                return self._fail(f"해제 실패: {e}")
-            before = self.store.load().get("routes_before")
-            lines = leftover_routes(set(before) if before is not None else None, self.system.route_text())
-            if lines:
-                log.info("removing %d leftover node routes", self.system.delete_routes(lines))
-            self.store.update(routes_before=None, connected_since=None, active_sid=None)
+            self._error = self._notice = None
+            if not self._teardown():
+                return self._fail(MSG_STILL_UP)
             return self._state()
 
     def set_node(self, choice: str) -> dict:
         with self._lock:
-            self._error = None
-            if choice == "auto":
-                self.store.update(node_mode="auto", manual_sid=None)
-            else:
-                self.store.update(node_mode="manual", manual_sid=int(choice))
+            self._error = self._notice = None
             try:
+                if choice == "auto":
+                    self.store.update(node_mode="auto", manual_sid=None)
+                else:
+                    self.store.update(node_mode="manual", manual_sid=int(choice))
                 st = self.client.status()
                 if self._is_on(st):
                     self._reapply(st)
@@ -122,12 +122,39 @@ class Controller:
                 return self._fail(f"노드 변경 실패: {e}")
             return self._state()
 
+    def check_resume(self, offset: float) -> bool:
+        """Feed boottime-minus-monotonic; a jump means the Deck slept. Then any tunnel is torn down,
+        because NetworkManager drops Mudfish's relay route on wake and the tunnel black-holes traffic."""
+        with self._lock:
+            prev, self._sleep_offset = self._sleep_offset, offset
+            if prev is None or offset - prev <= RESUME_JUMP:
+                return False
+            if self._tunnel_present():
+                log.info("resume detected; turning VPN off")
+                self._error = None
+                self._notice = MSG_SLEEP if self._teardown() else None
+                if self._notice is None:
+                    self._error = MSG_STILL_UP
+            return True
+
+    def shutdown(self) -> None:
+        """Plugin unload (Deck shutdown/restart, Decky restart): never leave a tunnel behind."""
+        with self._lock:
+            if self._tunnel_present():
+                log.info("unload; turning VPN off")
+                self._teardown()
+
     # ---- internals ----------------------------------------------------------------------------
 
     def _fail(self, message: str) -> dict:
         self._error = message
         return self._state()
 
+    def _abort(self, started: bool, message: str) -> dict:
+        if started and not self._teardown():
+            message = MSG_STILL_UP
+        return self._fail(message)
+
     def _wait(self, cond, seconds) -> bool:
         for _ in range(seconds):
             if cond():
@@ -135,9 +162,37 @@ class Controller:
             self.sleep(1)
         return cond()
 
+    def _teardown(self) -> bool:
+        """Stop the core, confirm tap0 is gone, remove the node routes Mudfish left, clear session state.
+        Returns False (keeping the route snapshot for a retry) when the tunnel is still up."""
+        try:
+            self.client.stop()
+        except Exception as e:
+            log.warning("stop failed: %s", e)
+        if not self._wait(lambda: not self.system.tap_up(), DISCONNECT_TIMEOUT):
+            return False
+        before = self.store.load().get("routes_before")
+        try:
+            lines = leftover_routes(set(before) if before is not None else None, self.system.route_text())
+            if lines:
+                log.info("removing %d leftover node routes", self.system.delete_routes(lines))
+        except Exception as e:
+            log.warning("route cleanup failed: %s", e)
+        self.store.update(routes_before=None, connected_since=None, active_sid=None)
+        self._ip_cache = (0.0, None)
+        return True
+
     def _is_on(self, st: dict) -> bool:
         return bool(st.get("core_running") and st.get("client_ready") and self.system.tap_up())
 
+    def _tunnel_present(self) -> bool:
+        if self.system.tap_up():
+            return True
+        try:
+            return bool(self.client.status().get("core_running"))
+        except Exception:
+            return False
+
     def _signed_in(self, installed) -> dict:
         self.system.ensure_daemon(installed[0], self.daemon_up)
         return self.client.ensure_signed_in()
@@ -149,18 +204,26 @@ class Controller:
         except Exception:
             return False
 
-    def _refresh_rtts(self) -> dict:
+    def _with_auth_retry(self, st: dict, fn):
+        """Run a cloud call; if the daemon's jwt was rejected, force a fresh signin once and retry."""
+        try:
+            return fn(st)
+        except AuthError:
+            return fn(self.client.ensure_signed_in(force=True))
+
+    def _refresh_rtts(self) -> None:
         try:
             rtts = self.client.rtts()
         except Exception:
             rtts = {}
         if rtts:
-            self.store.update(rtts={str(k): v for k, v in rtts.items()})
-            return rtts
-        return {int(k): v for k, v in (self.store.load().get("rtts") or {}).items()}
+            self.store.update(rtts={str(k): v for k, v in rtts.items()}, rtts_at=self.clock())
 
     def _cached_rtts(self) -> dict:
-        return {int(k): v for k, v in (self.store.load().get("rtts") or {}).items()}
+        data = self.store.load()
+        if not data.get("rtts") or self.clock() - (data.get("rtts_at") or 0) > RTT_TTL:
+            return {}
+        return {int(k): v for k, v in data["rtts"].items()}
 
     def _static_nodes(self, st: Optional[dict] = None) -> list:
         if self._static is None:
@@ -168,7 +231,7 @@ class Controller:
             if saved:
                 self._static = saved
             elif st and st.get("isSignedIn") and st.get("jwt"):
-                self._static = self.client.static_nodes(st["jwt"], st["api_server"])
+                self._static = self._with_auth_retry(st, lambda s: self.client.static_nodes(s["jwt"], s["api_server"]))
                 self.store.update(static_nodes=self._static)
         return self._static or []
 
@@ -204,7 +267,7 @@ class Controller:
         if node is None:
             return
         st = self.client.ensure_signed_in()
-        result = self._apply_sid(st, node.sid)
+        result = self._with_auth_retry(st, lambda s: self._apply_sid(s, node.sid))
         self.store.update(active_sid=node.sid)
         if not result["changed"]:
             return
@@ -213,8 +276,7 @@ class Controller:
             self._wait(lambda: not self.system.tap_up(), DISCONNECT_TIMEOUT)
             self.client.start()
             if not self._wait(lambda: self._is_on(self.client.status()), CONNECT_TIMEOUT):
-                self.client.stop()
-                raise RuntimeError(MSG_TIMEOUT)
+                raise RuntimeError(MSG_TIMEOUT if self._teardown() else MSG_STILL_UP)
         else:
             self.client.fullvpn_refresh(node.sid, result["conf"].get("mudfishFullvpnExceptions") or "")
         self._ip_cache = (0.0, None)
@@ -229,7 +291,7 @@ class Controller:
     def _state(self) -> dict:
         data = self.store.load()
         out = {"phase": "off", "message": None, "node_mode": data.get("node_mode") or "auto", "node": None,
-               "ip": None, "country": None, "connected_since": None}
+               "ip": None, "country": None, "connected_since": None, "tunnel": False}
         if not self.system.find_mudfish():
             out.update(phase="missing", message=MSG_MISSING)
             return out
@@ -237,9 +299,11 @@ class Controller:
             st = self.client.status()
         except Exception:
             st = None
+        tap = self.system.tap_up()
+        on = bool(st) and self._is_on(st)
+        out["tunnel"] = tap or bool(st and st.get("core_running"))
         if st:
             self._refresh_rtts()
-        on = bool(st) and self._is_on(st)
         if on:
             since = data.get("connected_since")
             if since is None:
@@ -247,14 +311,20 @@ class Controller:
                 self.store.update(connected_since=since)
             ip = self._public_ip() or {}
             out.update(phase="on", connected_since=since, ip=ip.get("ip"), country=ip.get("country"))
+        elif st is None and tap:
+            out.update(phase="error", message=MSG_NO_RESPONSE, connected_since=data.get("connected_since"))
         elif st and st.get("core_running"):
             out["phase"] = "connecting"
-        else:
+        elif not out["tunnel"]:
             if data.get("connected_since") is not None:
                 self.store.update(connected_since=None)
             self._ip_cache = (0.0, None)
+        if out["tunnel"] and not on and data.get("connected_since") is not None:
+            out["connected_since"] = data["connected_since"]
         if self._error and not on:
             out.update(phase="error", message=self._error)
+        elif self._notice and out["message"] is None:
+            out["message"] = self._notice
 
         nodes = self._jp_nodes(st)
         active = self.store.load().get("active_sid")

py_modules/mudfish_decky/system.py

@@ -57,6 +57,12 @@ def ensure_daemon(bin_path: str, is_up: Callable[[], bool], run=subprocess.run,
             return
 
 
+def sleep_offset(clock=time.clock_gettime, boot=getattr(time, "CLOCK_BOOTTIME", None),
+                 mono=time.CLOCK_MONOTONIC) -> float:
+    """Total time spent suspended since boot: BOOTTIME keeps counting during suspend, MONOTONIC doesn't."""
+    return clock(boot) - clock(mono)
+
+
 def tap_up() -> bool:
     return os.path.exists("/sys/class/net/tap0")
 

src/Panel.tsx

@@ -65,12 +65,17 @@ export function Panel() {
   const [advanced, setAdvanced] = useState(false);
   const [nodes, setNodes] = useState<NodeInfo[]>([]);
   const polling = useRef(false);
+  const acting = useRef(false);
 
   const refresh = async () => {
     if (polling.current) return;
     polling.current = true;
     try {
-      setState(await getState());
+      const next = await getState();
+      // A poll that started before a toggle press would overwrite the action's result.
+      if (!acting.current) setState(next);
+    } catch (e) {
+      console.error("mudfish: getState failed", e);
     } finally {
       polling.current = false;
     }
@@ -84,18 +89,28 @@ export function Panel() {
 
   useEffect(() => {
     if (!advanced) return;
-    const load = async () => setNodes(await listNodes());
+    const load = async () => {
+      try {
+        setNodes(await listNodes());
+      } catch (e) {
+        console.error("mudfish: listNodes failed", e);
+      }
+    };
     load();
     const id = setInterval(load, NODES_POLL_MS);
     return () => clearInterval(id);
   }, [advanced]);
 
   const run = async (action: () => Promise<VpnState>, pending?: Phase) => {
+    acting.current = true;
     setBusy(true);
     if (pending && state) setState({ ...state, phase: pending, message: null });
     try {
       setState(await action());
+    } catch (e) {
+      if (state) setState({ ...state, phase: "error", message: `요청 실패: ${e}` });
     } finally {
+      acting.current = false;
       setBusy(false);
     }
   };
@@ -110,7 +125,8 @@ export function Panel() {
     );
   }
 
-  const active = state.phase === "on" || state.phase === "connecting";
+  // Checked whenever any tunnel exists (even a stuck one), so the toggle can always turn it off.
+  const active = state.tunnel || state.phase === "on" || state.phase === "connecting";
   const options = [
     { data: "auto", label: "자동 (핑 최저)" },
     ...nodes.map((n) => ({ data: String(n.sid), label: `${n.label} · ${rttText(n.rtt)}` })),
@@ -128,7 +144,7 @@ export function Panel() {
         <ToggleField
           label="미꾸라지 Full VPN"
           checked={active}
-          disabled={busy || state.phase === "connecting" || state.phase === "missing"}
+          disabled={busy || state.phase === "missing"}
           onChange={(on) => run(on ? connect : disconnect, on ? "connecting" : undefined)}
         />
       </PanelSectionRow>

src/backend.ts

@@ -16,6 +16,7 @@ export interface VpnState {
   ip: string | null;
   country: string | null;
   connected_since: number | null;
+  tunnel: boolean;
 }
 
 export const getState = callable<[], VpnState>("get_state");

tests/test_client.py

@@ -1,7 +1,7 @@
 import json
 import unittest
 
-from mudfish_decky.client import MudfishClient, MudfishError, SigninError
+from mudfish_decky.client import AuthError, MudfishClient, MudfishError, SigninError
 
 BASE = "http://127.0.0.1:8282"
 SIGNED_OUT = {"status": 200, "isSignedIn": False, "username": "u", "token_v3": "123",
@@ -74,6 +74,44 @@ class ClientTest(unittest.TestCase):
         with self.assertRaises(SigninError):
             MudfishClient(BASE, t).ensure_signed_in()
 
+    def test_ensure_signed_in_force_skips_shortcut(self):
+        t = FakeTransport({
+            ("GET", BASE + "/do/mudrun/status"): [dict(SIGNED_OUT, isSignedIn=True)],
+            ("POST", "https://api.mudfish.net/signin_ng_v3"): [SIGNIN_OK],
+            ("POST", BASE + "/do/auth"): [{"status": 200, "msg": "OK"}],
+        })
+        MudfishClient(BASE, t).ensure_signed_in(force=True)
+        self.assertEqual(len(t.calls_to("POST", "https://api.mudfish.net/signin_ng_v3")), 1)
+
+    def test_ensure_signed_in_http_401_is_signin_error(self):
+        t = FakeTransport({("GET", BASE + "/do/mudrun/status"): [SIGNED_OUT]})
+        real = t.__call__
+
+        def transport(method, url, body, headers):
+            if "signin_ng_v3" in url:
+                return 401, b"{}"
+            return real(method, url, body, headers)
+
+        with self.assertRaises(SigninError):
+            MudfishClient(BASE, transport).ensure_signed_in()
+
+    def test_ensure_signed_in_without_token_is_signin_error(self):
+        t = FakeTransport({("GET", BASE + "/do/mudrun/status"): [dict(SIGNED_OUT, autologin=True, token_v3="")]})
+        with self.assertRaises(SigninError):
+            MudfishClient(BASE, t).ensure_signed_in()
+
+    def test_graphql_without_user_is_auth_error(self):
+        t = FakeTransport({("POST", "https://api.mudfish.net/graphql"): [{"errors": ["unauthorized"]}]})
+        with self.assertRaises(AuthError):
+            MudfishClient(BASE, t).prog_conf("JWT", "api.mudfish.net")
+
+    def test_prog_http_401_is_auth_error(self):
+        def transport(method, url, body, headers):
+            return 401, b"{}"
+
+        with self.assertRaises(AuthError):
+            MudfishClient(BASE, transport).save_prog_conf("JWT", "api.mudfish.net", {})
+
     def test_start_accepts_201(self):
         t = FakeTransport({("POST", BASE + "/do/mudfish/start"): [{"status": 201}]})
         MudfishClient(BASE, t).start()

tests/test_controller.py

@@ -2,8 +2,9 @@ import os
 import tempfile
 import unittest
 
-from mudfish_decky.client import SigninError
-from mudfish_decky.controller import MSG_MISSING, MSG_RELOGIN, MSG_RTT, MSG_TIMEOUT, Controller
+from mudfish_decky.client import AuthError, SigninError
+from mudfish_decky.controller import (MSG_MISSING, MSG_NO_RESPONSE, MSG_RELOGIN, MSG_RTT, MSG_SLEEP, MSG_STILL_UP,
+                                      MSG_TIMEOUT, Controller)
 from mudfish_decky.store import Store
 from tests.test_nodes import STATIC
 
@@ -58,7 +59,7 @@ class FakeClient:
     def status(self):
         return dict(self.st)
 
-    def ensure_signed_in(self):
+    def ensure_signed_in(self, force=False):
         if self.signin_error:
             raise self.signin_error
         return dict(self.st)
@@ -238,6 +239,146 @@ class ControllerTest(unittest.TestCase):
         self.assertEqual(st["node_mode"], "auto")
         self.assertEqual(st["node"]["sid"], 11)
 
+    # ---- final review fixes ------------------------------------------------------------------
+
+    def test_connect_status_error_after_start_tears_down(self):
+        real_start = self.client.start
+
+        def start():
+            real_start()
+            self.client.status = lambda: (_ for _ in ()).throw(OSError("launcher busy"))
+
+        self.client.start = start
+        st = self.make().connect()
+        self.assertEqual(self.client.stop_calls, 1)
+        self.assertEqual(st["phase"], "error")
+        self.assertEqual(len(self.system.deleted), 2)
+
+    def test_connect_timeout_cleans_routes_and_store(self):
+        self.client.ready_on_start = False
+        self.client.start = self._start_adding_routes_not_ready()
+        self.make().connect()
+        self.assertEqual(len(self.system.deleted), 2)
+        data = Store(self.path).load()
+        self.assertIsNone(data.get("routes_before"))
+        self.assertIsNone(data.get("active_sid"))
+
+    def _start_adding_routes_not_ready(self):
+        def start():
+            self.client.start_calls += 1
+            self.client.st["core_running"] = True
+            self.system.routes = BEFORE + NEW_ROUTES
+        return start
+
+    def test_disconnect_keeps_snapshot_when_tap_stays_up(self):
+        c = self.make()
+        c.connect()
+        real_stop = self.client.stop
+
+        def stop():
+            real_stop()
+            self.system.tap = True  # tunnel did not go away
+
+        self.client.stop = stop
+        st = c.disconnect()
+        self.assertEqual(st["phase"], "error")
+        self.assertEqual(st["message"], MSG_STILL_UP)
+        self.assertTrue(st["tunnel"])
+        data = Store(self.path).load()
+        self.assertIsNotNone(data.get("routes_before"))
+        self.assertEqual(data.get("connected_since"), 1000.0)
+        self.assertEqual(self.system.deleted, [])
+
+    def test_state_reports_tunnel_when_core_stuck(self):
+        self.client.st.update(core_running=True)
+        st = self.make().state()
+        self.assertEqual(st["phase"], "connecting")
+        self.assertTrue(st["tunnel"])
+
+    def test_state_tunnel_false_when_off(self):
+        self.assertFalse(self.make().state()["tunnel"])
+
+    def test_state_status_failure_with_tap_up_is_error_not_off(self):
+        Store(self.path).save({"connected_since": 100.0})
+        self.system.tap = True
+        self.client.status = lambda: (_ for _ in ()).throw(OSError("no answer"))
+        st = self.make().state()
+        self.assertEqual(st["phase"], "error")
+        self.assertEqual(st["message"], MSG_NO_RESPONSE)
+        self.assertTrue(st["tunnel"])
+        self.assertEqual(Store(self.path).load()["connected_since"], 100.0)
+
+    def test_connect_resigns_once_when_cloud_rejects_jwt(self):
+        calls = []
+        real_prog_conf = self.client.prog_conf
+
+        def prog_conf(jwt, api):
+            calls.append(jwt)
+            if len(calls) == 1:
+                raise AuthError("graphql returned no user")
+            return real_prog_conf(jwt, api)
+
+        forced = []
+        self.client.prog_conf = prog_conf
+        self.client.ensure_signed_in = lambda force=False: forced.append(force) or dict(self.client.st)
+        st = self.make().connect()
+        self.assertEqual(st["phase"], "on")
+        self.assertIn(True, forced)
+        self.assertEqual(len(calls), 2)
+
+    def test_connect_stale_rtt_cache_is_ignored(self):
+        Store(self.path).save({"rtts": {"11": 31.0}, "rtts_at": 1000.0 - 601})
+        self.client.rtt_map = {}
+        st = self.make().connect()
+        self.assertEqual(st["message"], MSG_RTT)
+        self.assertEqual(self.client.start_calls, 0)
+
+    # ---- sleep / shutdown --------------------------------------------------------------------
+
+    def test_resume_jump_turns_tunnel_off_with_notice(self):
+        c = self.make()
+        c.connect()
+        self.assertFalse(c.check_resume(5.0))   # first sample only sets the baseline
+        self.assertFalse(c.check_resume(5.5))   # small drift: not a resume
+        self.assertTrue(c.check_resume(65.5))   # suspended ~60s
+        st = c.state()
+        self.assertEqual(self.client.stop_calls, 1)
+        self.assertEqual(st["phase"], "off")
+        self.assertEqual(st["message"], MSG_SLEEP)
+        self.assertEqual(len(self.system.deleted), 2)
+
+    def test_resume_while_off_does_nothing(self):
+        c = self.make()
+        c.check_resume(5.0)
+        c.check_resume(100.0)
+        self.assertEqual(self.client.stop_calls, 0)
+        self.assertIsNone(c.state()["message"])
+
+    def test_resume_never_starts_core(self):
+        c = self.make()
+        c.check_resume(5.0)
+        c.check_resume(100.0)
+        self.assertEqual(self.client.start_calls, 0)
+
+    def test_notice_cleared_by_next_connect(self):
+        c = self.make()
+        c.connect()
+        c.check_resume(5.0)
+        c.check_resume(100.0)
+        self.assertIsNone(c.connect()["message"])
+
+    def test_shutdown_turns_tunnel_off(self):
+        c = self.make()
+        c.connect()
+        c.shutdown()
+        self.assertEqual(self.client.stop_calls, 1)
+        self.assertEqual(len(self.system.deleted), 2)
+        self.assertIsNone(Store(self.path).load().get("connected_since"))
+
+    def test_shutdown_while_off_does_not_call_stop(self):
+        self.make().shutdown()
+        self.assertEqual(self.client.stop_calls, 0)
+
     def test_prepare_signs_in_and_caches_nodes_without_starting(self):
         self.client.st["isSignedIn"] = False
         signed = []

tests/test_system.py

@@ -35,6 +35,12 @@ class SystemTest(unittest.TestCase):
         with tempfile.TemporaryDirectory() as var:
             self.assertEqual(api_base(var), "http://127.0.0.1:8282")
 
+    def test_sleep_offset_is_boottime_minus_monotonic(self):
+        from mudfish_decky.system import sleep_offset
+        clocks = {"boot": 500.0, "mono": 420.0}
+        self.assertEqual(sleep_offset(lambda c: clocks["boot"] if c == "boot" else clocks["mono"],
+                                      boot="boot", mono="mono"), 80.0)
+
     def test_ensure_daemon_noop_when_up(self):
         calls = []
         ensure_daemon("/bin/x", lambda: True, run=lambda cmd, **kw: calls.append(cmd))