[Messaging] Fatal IllegalStateException: Broadcast already finished in CloudMessagingReceiver

The crash is raised from BroadcastReceiver.PendingResult.finish() inside CloudMessagingReceiver.zza in the Play Services cloud-messaging library, so the report's stack trace is the main entry point, and the code in this repository is not shown in the payload. Start by checking which resolved play-services-cloud-messaging and Firebase Messaging versions the affected releases 1.52.0 and 1.53.0 use, and whether the BoM-pinned version matches. Done means identifying a known cause, a fixed version, or a reproducible case, and recording it in the issue; a code change would belong in the SDK's receiver handling only after that.

Environment

• Firebase component: Firebase Cloud Messaging for Android
• Firebase BoM configured in the currently checked source: 34.17.0
• Firebase Messaging version in the affected release: Not yet verified
• play-services-cloud-messaging version in the affected release: Not yet verified
• Affected app versions observed in Crashlytics: 1.52.0 and 1.53.0
• Android versions observed: Android 14 and Android 15
• Device models observed include Samsung SM-G7810, Samsung SM-F7110, and Xiaomi M2007J3SY

The configured BoM version above is from the current source tree. We have not yet verified the resolved dependency versions in the affected release artifacts.

Problem

We are seeing a fatal crash reported by Firebase Crashlytics:

java.lang.IllegalStateException: Broadcast already finished

The crash occurs while CloudMessagingReceiver is calling BroadcastReceiver.PendingResult.finish().

Recent Crashlytics samples were recorded while the app was in the background. One sample recorded a notification_receive event approximately two seconds before the crash, but we cannot determine from the available data whether that notification directly caused the crash.

In a report retrieved on November 3, 2026, Crashlytics showed four events affecting four users in its default seven-day reporting window. We have also observed this issue in earlier app versions.

We do not know what completed the broadcast before the call shown in the crash stack, or whether this is related to device-specific behavior.

Steps to reproduce

We do not currently have a reliable local reproduction. The crash was observed in production through Crashlytics.

The available samples suggest this context:

1. The app is running in the background.
2. Cloud Messaging handles a broadcast.
3. CloudMessagingReceiver calls PendingResult.finish().
4. Android throws IllegalStateException: Broadcast already finished.

These are observations from the reports, not confirmed reproduction steps.

Expected behavior

The app should not crash while Cloud Messaging finishes processing a broadcast.

Actual behavior

The process crashes with the following fatal exception:

Fatal Exception: java.lang.IllegalStateException: Broadcast already finished
       at android.content.BroadcastReceiver$PendingResult.sendFinished(BroadcastReceiver.java:303)
       at android.content.BroadcastReceiver$PendingResult.finish(BroadcastReceiver.java:288)
       at com.google.android.gms.cloudmessaging.CloudMessagingReceiver.zza(:179)
       at com.google.android.gms.cloudmessaging.zzg.run(:10)
       at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
       at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:644)
       at com.google.android.gms.common.util.concurrent.zza.run(:6)
       at java.lang.Thread.run(Thread.java:1019)

Other threads included in the supplied crash report:

process reaper:
       at jdk.internal.misc.Unsafe.park(Unsafe.java)
       at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:252)
       at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java:401)
       at java.util.concurrent.SynchronousQueue.poll(SynchronousQueue.java:903)
       at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1070)
       at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1131)
       at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:644)
       at java.lang.Thread.run(Thread.java:1019)

process reaper:
       at jdk.internal.misc.Unsafe.park(Unsafe.java)
       at java.util.concurrent.locks.LockSupport.parkNanos(LockSupport.java:252)
       at java.util.concurrent.SynchronousQueue$TransferStack.transfer(SynchronousQueue.java:401)
       at java.util.concurrent.SynchronousQueue.poll(SynchronousQueue.java:903)
       at java.util.concurrent.ThreadPoolExecutor.getTask(ThreadPoolExecutor.java:1070)
       at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1131)
       at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:644)
       at java.lang.Thread.run(Thread.java:1019)

FinalizerDaemon:
       at java.lang.Object.wait(Object.java)
       at java.lang.Object.wait(Object.java:386)
       at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:210)
       at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:231)
       at java.lang.Daemons$FinalizerDaemon.runInternal(Daemons.java:309)
       at java.lang.Daemons$Daemon.run(Daemons.java:145)
       at java.lang.Thread.run(Thread.java:1019)

ReferenceQueueDaemon:
       at java.lang.Object.wait(Object.java)
       at java.lang.Object.wait(Object.java:386)
       at java.lang.Object.wait(Object.java:524)
       at java.lang.Daemons$ReferenceQueueDaemon.runInternal(Daemons.java:239)
       at java.lang.Daemons$Daemon.run(Daemons.java:145)
       at java.lang.Thread.run(Thread.java:1019)

FinalizerWatchdogDaemon:
       at java.lang.Thread.sleep(Thread.java)
       at java.lang.Thread.sleep(Thread.java:455)
       at java.lang.Thread.sleep(Thread.java:360)
       at java.lang.Daemons$FinalizerWatchdogDaemon.sleepForNanos(Daemons.java:481)
       at java.lang.Daemons$FinalizerWatchdogDaemon.waitForProgress(Daemons.java:544)
       at java.lang.Daemons$FinalizerWatchdogDaemon.runInternal(Daemons.java:412)
       at java.lang.Daemons$Daemon.run(Daemons.java:145)
       at java.lang.Thread.run(Thread.java:1019)

Related reports

• AndroidX Glance catches IllegalStateException around its own PendingResult.finish() call. Its source comment mentions that some OEM devices can throw Broadcast already finished and references b/257513022:
  https://android.googlesource.com/platform/frameworks/support/+/e446b28b091a3a58813c2d169de4f629552cb3ac/glance/glance-appwidget/src/main/java/androidx/glance/appwidget/CoroutineBroadcastReceiver.kt

• Firebase issue #5553 also involves CloudMessagingReceiver and PendingResult.finish(), but reports a NullPointerException. We are including it as related context, not claiming it is the same issue:
  https://github.com/firebase/firebase-android-sdk/issues/5553

Questions

1. Is this a known issue in Firebase Cloud Messaging or play-services-cloud-messaging?
2. Could the device-specific behavior referenced by AndroidX issue b/257513022 also affect CloudMessagingReceiver?
3. Is there a recommended Firebase SDK version or supported workaround for this crash?
4. Would it be appropriate for CloudMessagingReceiver to handle this exception when calling PendingResult.finish()?
5. What additional logs, dependency information, or reproduction details would help investigate the first completion of the broadcast?

We can provide additional information as it becomes available.