fix(native/session): make the stop flag enforceable so a session can't outlive its client

Native sessions could survive long after the client was gone, in two
independent ways.

HOST: `stop` was only advisory. The one thing that ends a session is the
stream thread returning — every teardown (conn.close, the joins, and the
RAII drops of the session permit, admission entry and stream marker)
sits after that await, and nothing forced the issue. The encode loop
checks `stop` between iterations, so any unbounded call INSIDE one never
reaches the check, and one stuck syscall became a permanent zombie: it
held its semaphore slot (four of those and the host stops accepting
QUIC entirely), its admission entry (a later client gets "host busy"
forever), and not even the console's Stop button could clear it — that
button sets this same flag.

  * Bound the wait: once the session has been told to stop, the thread
    gets STREAM_STOP_GRACE (90 s, well past the 40 s capture-rebuild
    budget) to return, then teardown runs anyway. The thread is detached,
    not killed — Rust can't cancel a blocking thread — so it keeps its
    capturer/encoder until the stuck call returns, but the session's slot
    and admission entry come back and the host keeps serving. It logs at
    ERROR as the host wedge it is.
  * Bound the audio/input joins too — the last unbounded await in
    teardown.
  * Take the session permit AFTER the QUIC handshake instead of before
    `accept()`, so a host at its concurrency cap still accepts and the
    waiting client sees a live path instead of a silent dial timeout.
  * Bound the compositor helpers that caused the wedge in the first
    place: new pf-vdisplay `proc::{status_within, output_within}` kill a
    child that outlives its budget. `kscreen-doctor` is a Wayland client
    of the very compositor it configures, so against a wedged KWin it
    never returned; same for systemctl/dbus against a stuck session bus.

CLIENTS: the connection was never closed, so the host was right to keep
the session — it still had a live, keep-alive-answering peer.

  * Android: backgrounding did no teardown at all, and Android doesn't
    suspend the process, so the worker kept answering keep-alives until
    the OS reclaimed it (on a TV box, never). End the session on ON_STOP,
    via the existing onDispose path; a plain close, not a quit, so the
    host lingers the display for a fast return.
  * Apple: the .background arm was iOS-only AND gated on an opt-in that
    defaults off, so backgrounding did nothing — while the `audio`
    background mode kept the app (and its connection) alive indefinitely.
    Act unconditionally, and cover tvOS.
  * Core: `conn.close()` only queues the frame, and run_pump is the body
    of a block_on whose runtime is dropped the instant it returns, so the
    driver could never put it on the wire — a deliberate quit reached the
    host as silence (8 s idle timeout, no quit code, and the linger meant
    for an unwanted disconnect). Carry the endpoint out of the handshake
    and flush with wait_idle(), the same discipline the pairing and probe
    paths already use.

Linux check/clippy/tests green: 262 host, 71 pf-vdisplay (incl. new
bounded-process tests), 231 core.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

clients/android/app/src/main/kotlin/io/unom/punktfunk/StreamScreen.kt

@@ -50,6 +50,9 @@ import androidx.core.content.ContextCompat
 import androidx.core.view.WindowCompat
 import androidx.core.view.WindowInsetsCompat
 import androidx.core.view.WindowInsetsControllerCompat
+import androidx.lifecycle.Lifecycle
+import androidx.lifecycle.LifecycleEventObserver
+import androidx.lifecycle.LifecycleOwner
 import io.unom.punktfunk.kit.GamepadFeedback
 import io.unom.punktfunk.kit.GamepadRouter
 import io.unom.punktfunk.kit.deviceBodyVibrator
@@ -384,6 +387,26 @@ fun StreamScreen(handle: Long, micEnabled: Boolean, onDisconnect: () -> Unit) {
     // Back gesture = a deliberate exit → signal the quit so the host tears down now (no linger).
     BackHandler { NativeBridge.nativeDisconnectQuit(handle); onDisconnect() }
 
+    // Leaving the app (Home, task switch, screen off) MUST end the session. Android does not
+    // suspend a process for going to background, so without this the native worker kept running and
+    // its QUIC connection kept answering the host's keep-alives — the user was long gone but the
+    // host still saw a live client and held the session (and its display + encoder) open until the
+    // OS eventually reclaimed the process, which on a TV box is effectively never.
+    //
+    // Route it through `onDisconnect()` so the composable's `onDispose` above runs the one real
+    // teardown path. Deliberately NOT a `nativeDisconnectQuit`: backgrounding isn't a user "quit",
+    // so the host should linger the display and make coming straight back a fast reconnect.
+    DisposableEffect(handle) {
+        val lifecycle = (context as? LifecycleOwner)?.lifecycle
+        val obs = LifecycleEventObserver { _, event ->
+            if (event == Lifecycle.Event.ON_STOP) {
+                onDisconnect()
+            }
+        }
+        lifecycle?.addObserver(obs)
+        onDispose { lifecycle?.removeObserver(obs) }
+    }
+
     // Auto-engage pointer capture at stream start (setting on + a mouse actually present).
     // Delayed a beat: the grab needs window focus and the capture view attached.
     LaunchedEffect(handle) {

clients/apple/Sources/PunktfunkClient/ContentView.swift

@@ -144,14 +144,26 @@ struct ContentView: View {
         // tap uses, so trust policy / WoL / the approval sheet all come along. Never starts a
         // parallel session — this drives the one `model` ContentView owns.
         .onOpenURL { handleDeepLink($0) }
-        #if os(iOS)
-        // Background keep-alive driver (opt-in). Only .background/.active matter; .inactive (a
-        // transient peek) is ignored so the disconnect timer never starts for a Control-Center pull.
+        #if os(iOS) || os(tvOS)
+        // Backgrounding driver. Only .background/.active matter; .inactive (a transient peek) is
+        // ignored so neither branch fires for a Control-Center pull.
+        //
+        // Backgrounding MUST end the session one way or the other: the app keeps running while
+        // streaming (the `audio` background mode plus a live audio session), so its QUIC connection
+        // keeps answering the host's keep-alives with the user long gone — the host has no way to
+        // tell that apart from someone watching, and the session survived indefinitely. Either hold
+        // it under the opt-in keep-alive (bounded by that path's own auto-disconnect timer) or end
+        // it here.
         .onChange(of: scenePhase) { _, phase in
             switch phase {
             case .background:
-                if backgroundKeepAlive, model.phase == .streaming {
+                guard model.phase == .streaming else { break }
+                if backgroundKeepAlive {
                     model.enterBackground(timeoutMinutes: backgroundTimeoutMinutes)
+                } else {
+                    // Not deliberate: the user may come straight back, so let the host linger the
+                    // display for a fast reconnect instead of tearing it down.
+                    model.disconnect(deliberate: false)
                 }
             case .active:
                 model.exitBackground()
@@ -159,7 +171,11 @@ struct ContentView: View {
                 break
             }
         }
-        // Live Activity lifecycle, driven from the model's published state.
+        #endif
+        #if os(iOS)
+        // Live Activity lifecycle, driven from the model's published state. iPhone/iPad only —
+        // ActivityKit (and so `liveActivity`) does not exist on tvOS, which is why this stays in its
+        // own os(iOS) block rather than riding the backgrounding driver's.
         .onChange(of: model.phase) { _, phase in
             switch phase {
             case .streaming: