feat: add audio input and output support

Author: SaicharanKandukuri

README.md

@@ -30,6 +30,7 @@ the app's **About** page.
 | Embedded X11 | Use the bundled Termux:X11 server and uDroid display surface without installing the separate Termux:X11 application. Keyboard, direct-pointer, trackpad-style, native multi-touch, and display settings are available inside uDroid. |
 | Desktop environments | Detect installed X11 sessions from standard `xsessions` directories. Start, stop, or restart the selected desktop while uDroid reports which rootfs and desktop own `DISPLAY :0`. |
 | Compositing | Enable or disable compositing for supported XFCE, Plasma, and MATE sessions, then apply the setting through a supervised desktop restart. GNOME is treated as compositor-required. |
+| Audio | Play Linux audio through Android with an app-owned PulseAudio/OpenSL ES bridge. Enable microphone input per Linux system with Android runtime consent; guest traffic stays on authenticated device-local loopback. |
 
 ### Current limitations
 
@@ -48,13 +49,15 @@ the app's **About** page.
 - The embedded display is functional, but general-purpose GPU acceleration is
   not yet a supported app contract. Desktop responsiveness depends on the
   device, resolution, applications, and compositor choice.
-- Audio integration, background desktop recovery, multiple displays, and
-  production-level desktop compatibility remain incomplete.
+- Background desktop recovery, multiple displays, and production-level desktop
+  compatibility remain incomplete. Audio currently targets PulseAudio clients;
+  direct ALSA-only applications may require a PulseAudio compatibility plugin.
 
 See the [distribution catalogue](docs/DISTRIBUTION_CATALOGUE.md),
 [OCI image architecture](docs/OCI_IMAGES.md),
 [Linux application launcher](docs/LINUX_APPLICATION_LAUNCHER.md),
 [X11 runtime architecture](docs/X11_RUNTIME_ARCHITECTURE.md),
+[audio runtime](docs/AUDIO_RUNTIME.md),
 [performance notes](docs/PERFORMANCE.md), and
 [app updates](docs/APP_UPDATES.md) for current behavior and technical details.
 
@@ -104,6 +107,7 @@ export ANDROID_NDK_HOME="$ANDROID_HOME/ndk/28.2.13676358"
 ./app/src/main/cpp/build-runtime-probe.sh
 ./tools/build-proot-assets.sh
 ./tools/build-gnu-tar-assets.sh
+./tools/build-pulseaudio-assets.sh
 ./gradlew :app:assembleRelease
 ```
 

THIRD_PARTY_NOTICES.md

@@ -75,6 +75,23 @@ retain their existing MIT grants and may also be distributed under GPLv3 as
 part of the combined application. Complete corresponding source includes this
 repository and all recursively pinned Termux:X11 submodules.
 
+## PulseAudio and Termux runtime libraries
+
+- PulseAudio project: <https://www.freedesktop.org/wiki/Software/PulseAudio/>
+- PulseAudio version: `17.0`, Termux package revision `3`
+- Binary package repository: <https://packages.termux.dev/apt/termux-main/>
+- Package versions and SHA-256 values: `tools/pulseaudio-packages.tsv`
+- Repackaging script: `tools/build-pulseaudio-assets.sh`
+- PulseAudio license: GNU Lesser General Public License version 2.1 or later
+
+The APK contains a compact relocation of the official Termux PulseAudio
+runtime. It includes only the daemon, native TCP protocol, OpenSL ES
+sink/source modules, and their required shared libraries. uDroid does not
+expose PulseAudio beyond authenticated device-local loopback. The pinned
+manifest records every input package, version, repository path, and archive
+checksum used for all supported ABIs.
+The individual dependency packages retain their upstream licenses.
+
 ## PRoot-Distro rootfs recipes
 
 - Project: <https://github.com/termux/proot-distro>

app/src/androidTest/java/org/randomcoder/udroid/audio/AudioRuntimeProbeTest.kt

@@ -0,0 +1,48 @@
+package org.randomcoder.udroid.audio
+
+import androidx.test.core.app.ApplicationProvider
+import androidx.test.ext.junit.runners.AndroidJUnit4
+import org.junit.Assert.assertTrue
+import org.junit.Test
+import org.junit.runner.RunWith
+import org.randomcoder.udroid.runtime.EventJournal
+import java.net.InetAddress
+import java.net.Socket
+import java.util.concurrent.Executors
+
+@RunWith(AndroidJUnit4::class)
+class AudioRuntimeProbeTest {
+    @Test
+    fun packagedRuntimeStartsOutputOnAuthenticatedLoopback() {
+        val context = ApplicationProvider.getApplicationContext<android.content.Context>()
+        val executor = Executors.newCachedThreadPool()
+        val controller = AudioServerController(context, EventJournal(context), executor)
+        try {
+            val started =
+                controller.apply(
+                    AudioConfiguration(outputEnabled = true, microphoneEnabled = false),
+                    bootId = "audio-device-probe",
+                )
+            assertTrue(started.running)
+            val cookie =
+                controller.endpoint().hostAuthDirectory.resolve(AudioEndpoint.COOKIE_NAME)
+            val deadline = System.nanoTime() + 5_000_000_000L
+            var acceptingConnections = false
+            while (!acceptingConnections && controller.current().running && System.nanoTime() < deadline) {
+                acceptingConnections =
+                    runCatching {
+                        Socket(InetAddress.getLoopbackAddress(), AudioEndpoint.PORT).use { }
+                        true
+                    }.getOrDefault(false)
+                Thread.sleep(20)
+            }
+            assertTrue("PulseAudio did not create its authentication cookie", cookie.isFile)
+            assertTrue("PulseAudio did not listen on device-local loopback", acceptingConnections)
+            assertTrue(controller.current().outputEnabled)
+            assertTrue(!controller.current().microphoneEnabled)
+        } finally {
+            controller.stop("audio-device-probe")
+            executor.shutdownNow()
+        }
+    }
+}

app/src/main/AndroidManifest.xml

@@ -6,8 +6,11 @@
     <uses-permission android:name="android.permission.ACCESS_NETWORK_STATE" />
     <uses-permission android:name="android.permission.FOREGROUND_SERVICE" />
     <uses-permission android:name="android.permission.FOREGROUND_SERVICE_DATA_SYNC" />
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE_MEDIA_PLAYBACK" />
+    <uses-permission android:name="android.permission.FOREGROUND_SERVICE_MICROPHONE" />
     <uses-permission android:name="android.permission.FOREGROUND_SERVICE_SPECIAL_USE" />
     <uses-permission android:name="android.permission.POST_NOTIFICATIONS" />
+    <uses-permission android:name="android.permission.RECORD_AUDIO" />
     <uses-permission android:name="android.permission.REQUEST_INSTALL_PACKAGES" />
     <uses-permission android:name="android.permission.RECEIVE_BOOT_COMPLETED" />
 
@@ -86,7 +89,7 @@
             android:name=".runtime.RuntimeSupervisorService"
             android:description="@string/supervisor_description"
             android:exported="false"
-            android:foregroundServiceType="specialUse"
+            android:foregroundServiceType="specialUse|mediaPlayback|microphone"
             android:stopWithTask="false">
             <property
                 android:name="android.app.PROPERTY_SPECIAL_USE_FGS_SUBTYPE"

app/src/main/assets/runtime/arm64-v8a/pulseaudio-runtime.zip

@@ -16,6 +16,7 @@ import androidx.activity.ComponentActivity
 import androidx.activity.SystemBarStyle
 import androidx.activity.compose.setContent
 import androidx.activity.enableEdgeToEdge
+import androidx.activity.result.contract.ActivityResultContracts
 import androidx.compose.runtime.getValue
 import androidx.compose.runtime.mutableStateOf
 import androidx.compose.runtime.setValue
@@ -25,6 +26,8 @@ import androidx.lifecycle.lifecycleScope
 import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.launch
 import kotlinx.coroutines.withContext
+import org.randomcoder.udroid.audio.AudioConfiguration
+import org.randomcoder.udroid.audio.AudioConfigurationStore
 import org.randomcoder.udroid.catalog.DistroCatalogRepository
 import org.randomcoder.udroid.catalog.DistroCatalogState
 import org.randomcoder.udroid.catalog.DistroVariant
@@ -111,6 +114,8 @@ class MainActivity : ComponentActivity() {
         mutableStateOf(DesktopConfiguration(null, compositingEnabled = false, touchScaleEnabled = true))
     private var desktopScanLoading by mutableStateOf(false)
     private var desktopScanMessage by mutableStateOf<String?>(null)
+    private var audioConfiguration by mutableStateOf(AudioConfiguration())
+    private var audioConfigurationMessage by mutableStateOf<String?>(null)
     private var linuxApplicationsState by
         mutableStateOf<LinuxApplicationsState>(LinuxApplicationsState.Loading)
     private var linuxApplicationMessage by mutableStateOf<String?>(null)
@@ -119,6 +124,7 @@ class MainActivity : ComponentActivity() {
     private var pendingDesktopStart: PendingDesktopStart? = null
     private var pendingShortcutRootfsName: String? = null
     private var pendingTerminalRootfsName: String? = null
+    private var pendingMicrophoneRootfsName: String? = null
     private var linuxApplicationsLoadGeneration = 0L
     private var desktopScanGeneration = 0L
     private var ociCatalogueLoadGeneration = 0L
@@ -128,9 +134,33 @@ class MainActivity : ComponentActivity() {
     private var runtimeService by mutableStateOf<RuntimeSupervisorService?>(null)
     private var runtimeServiceBound = false
     private val desktopConfigurationStore by lazy { DesktopConfigurationStore(this) }
+    private val audioConfigurationStore by lazy { AudioConfigurationStore(this) }
     private val ociHubCatalogueRepository by lazy { OciHubCatalogRepository(this) }
     private val ociHubTagRepository by lazy { OciHubTagRepository(this) }
 
+    private val microphonePermissionLauncher =
+        registerForActivityResult(ActivityResultContracts.RequestPermission()) { granted ->
+            val rootfsName = pendingMicrophoneRootfsName
+            pendingMicrophoneRootfsName = null
+            if (rootfsName == null) return@registerForActivityResult
+            if (granted) {
+                val current = audioConfigurationStore.load(rootfsName)
+                saveAudioConfiguration(
+                    rootfsName,
+                    current.copy(microphoneEnabled = true),
+                )
+            } else {
+                if (selectedSystemRootfsName == rootfsName) {
+                    audioConfiguration =
+                        audioConfigurationStore
+                            .load(rootfsName)
+                            .copy(microphoneEnabled = false)
+                    audioConfigurationMessage =
+                        "Microphone access was not granted. Linux input remains off."
+                }
+            }
+        }
+
     private val runtimeServiceConnection =
         object : ServiceConnection {
             override fun onServiceConnected(
@@ -193,6 +223,8 @@ class MainActivity : ComponentActivity() {
                     desktopConfiguration = desktopConfiguration,
                     desktopScanLoading = desktopScanLoading,
                     desktopScanMessage = desktopScanMessage,
+                    audioConfiguration = audioConfiguration,
+                    audioConfigurationMessage = audioConfigurationMessage,
                     linuxApplicationsState = linuxApplicationsState,
                     linuxApplicationMessage = linuxApplicationMessage,
                     showInstallTerminal = showInstallTerminal,
@@ -226,6 +258,8 @@ class MainActivity : ComponentActivity() {
                     onSelectDesktopEnvironment = { selectDesktopEnvironment(it) },
                     onCompositingChanged = { updateCompositing(it) },
                     onTouchScaleChanged = { updateTouchScale(it) },
+                    onAudioOutputChanged = { updateAudioOutput(it) },
+                    onMicrophoneChanged = { updateMicrophone(it) },
                     onStartDesktop = { startSelectedDesktop() },
                     onStopDesktop = { runtimeService?.stopDesktop() },
                     onRestartDesktop = { restartSelectedDesktop() },
@@ -606,6 +640,7 @@ class MainActivity : ComponentActivity() {
         if (installedRootfses.none { it.name == rootfsName }) return
         selectedSystemRootfsName = rootfsName
         rootfsMaintenanceMessage = null
+        loadAudioConfiguration(rootfsName)
         selectDestination(UdroidDestination.SYSTEM)
     }
 
@@ -679,6 +714,9 @@ class MainActivity : ComponentActivity() {
             runCatching { desktopConfigurationStore.remove(rootfsName) }
                 .exceptionOrNull()
                 ?.let { cleanupWarnings += it.message ?: "desktop settings" }
+            runCatching { audioConfigurationStore.remove(rootfsName) }
+                .exceptionOrNull()
+                ?.let { cleanupWarnings += it.message ?: "audio settings" }
             runCatching {
                 LinuxApplicationShortcutPublisher(this@MainActivity)
                     .disableForRootfs(rootfsName)
@@ -850,6 +888,86 @@ class MainActivity : ComponentActivity() {
         )
     }
 
+    private fun loadAudioConfiguration(rootfsName: String) {
+        runCatching { audioConfigurationStore.load(rootfsName) }
+            .onSuccess {
+                audioConfiguration = it
+                audioConfigurationMessage = null
+            }.onFailure {
+                audioConfiguration = AudioConfiguration()
+                audioConfigurationMessage = it.message ?: "Could not load audio settings"
+            }
+    }
+
+    private fun updateAudioOutput(enabled: Boolean) {
+        val rootfsName = selectedSystemRootfsName ?: return
+        saveAudioConfiguration(
+            rootfsName,
+            audioConfiguration.copy(outputEnabled = enabled),
+        )
+    }
+
+    private fun updateMicrophone(enabled: Boolean) {
+        val rootfsName = selectedSystemRootfsName ?: return
+        if (!enabled) {
+            pendingMicrophoneRootfsName = null
+            saveAudioConfiguration(
+                rootfsName,
+                audioConfiguration.copy(microphoneEnabled = false),
+            )
+            return
+        }
+        if (
+            ContextCompat.checkSelfPermission(this, Manifest.permission.RECORD_AUDIO) ==
+            PackageManager.PERMISSION_GRANTED
+        ) {
+            saveAudioConfiguration(
+                rootfsName,
+                audioConfiguration.copy(microphoneEnabled = true),
+            )
+            return
+        }
+        pendingMicrophoneRootfsName = rootfsName
+        audioConfigurationMessage = "Allow microphone access to send input to Linux."
+        microphonePermissionLauncher.launch(Manifest.permission.RECORD_AUDIO)
+    }
+
+    private fun saveAudioConfiguration(
+        rootfsName: String,
+        configuration: AudioConfiguration,
+    ) {
+        runCatching { audioConfigurationStore.save(rootfsName, configuration) }
+            .onSuccess { saved ->
+                val service = runtimeService
+                val runningThisRootfs =
+                    service?.currentTerminalSession()?.isRunning == true &&
+                        service.currentTerminalSession()?.mSessionName == rootfsName
+                if (selectedSystemRootfsName == rootfsName) {
+                    audioConfiguration = saved
+                    audioConfigurationMessage =
+                        if (runningThisRootfs) {
+                            "Applying audio settings to the running Linux system…"
+                        } else {
+                            "Audio settings will apply when this Linux system starts."
+                        }
+                }
+                if (runningThisRootfs) {
+                    service.applyAudioConfiguration(rootfsName, saved) { result ->
+                        if (selectedSystemRootfsName != rootfsName) return@applyAudioConfiguration
+                        audioConfigurationMessage =
+                            result.fold(
+                                onSuccess = { it.message },
+                                onFailure = {
+                                    it.message ?: "Could not apply audio settings"
+                                },
+                            )
+                    }
+                }
+            }.onFailure {
+                audioConfigurationMessage = it.message ?: "Could not save audio settings"
+            }
+    }
+
     private fun saveDesktopConfiguration(
         rootfsName: String,
         configuration: DesktopConfiguration,

app/src/main/assets/runtime/armeabi-v7a/pulseaudio-runtime.zip

@@ -0,0 +1,62 @@
+package org.randomcoder.udroid.audio
+
+import android.content.Context
+
+data class AudioConfiguration(
+    val outputEnabled: Boolean = true,
+    val microphoneEnabled: Boolean = false,
+)
+
+class AudioConfigurationStore(context: Context) {
+    private val preferences =
+        context.applicationContext
+            .getSharedPreferences(PREFERENCES_NAME, Context.MODE_PRIVATE)
+
+    fun load(rootfsName: String): AudioConfiguration =
+        AudioConfiguration(
+            outputEnabled =
+                preferences.getBoolean(key(rootfsName, KEY_OUTPUT), true),
+            microphoneEnabled =
+                preferences.getBoolean(key(rootfsName, KEY_MICROPHONE), false),
+        )
+
+    fun save(
+        rootfsName: String,
+        configuration: AudioConfiguration,
+    ): AudioConfiguration {
+        check(
+            preferences
+                .edit()
+                .putBoolean(key(rootfsName, KEY_OUTPUT), configuration.outputEnabled)
+                .putBoolean(
+                    key(rootfsName, KEY_MICROPHONE),
+                    configuration.microphoneEnabled,
+                ).commit(),
+        ) {
+            "Could not save audio settings for $rootfsName"
+        }
+        return configuration
+    }
+
+    fun remove(rootfsName: String) {
+        val prefix = "$rootfsName:"
+        val editor = preferences.edit()
+        preferences.all.keys
+            .filter { it.startsWith(prefix) }
+            .forEach(editor::remove)
+        check(editor.commit()) {
+            "Could not clear audio settings for $rootfsName"
+        }
+    }
+
+    private fun key(
+        rootfsName: String,
+        setting: String,
+    ): String = "$rootfsName:$setting"
+
+    private companion object {
+        const val PREFERENCES_NAME = "audio-configuration"
+        const val KEY_OUTPUT = "output"
+        const val KEY_MICROPHONE = "microphone"
+    }
+}