OutOfMemory only on brand new Samsung models

Mobile
#1

Hi, In google console I notice traces only on Samsung phones.
my original issue for OkHttp Brand new Samsung Devices with OutOfMemory · Issue #6553 · square/okhttp · GitHub but was instantly closed.

java.lang.OutOfMemoryError:
at java.util.ArrayList.iterator (ArrayList.java:834)
at okhttp3.internal.concurrent.TaskRunner.awaitTaskToRun (TaskRunner.java:167)
at okhttp3.internal.concurrent.TaskRunner$runnable$1.run (TaskRunner.java:58)
at java.util.concurrent.ThreadPoolExecutor.runWorker (ThreadPoolExecutor.java:1167)
at java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:641)
at java.lang.Thread.run (Thread.java:919)

java.lang.OutOfMemoryError:
at java.util.concurrent.SynchronousQueue$TransferStack.snode (SynchronousQueue.java:318)
at java.util.concurrent.SynchronousQueue$TransferStack.transfer (SynchronousQueue.java:361)
at java.util.concurrent.SynchronousQueue.poll (SynchronousQueue.java:937)
at java.util.concurrent.ThreadPoolExecutor.getTask (ThreadPoolExecutor.java:1091)
at java.util.concurrent.ThreadPoolExecutor.runWorker (ThreadPoolExecutor.java:1152)
at java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:641)
at java.lang.Thread.run (Thread.java:923)

107267922-8f5f8400-6a58-11eb-87d1-09536bb3ec51
107267922-8f5f8400-6a58-11eb-87d1-09536bb3ec51900×1571 206 KB

#2

Hi,
Would you please help me to reproduce the issue?
As I did not face the issue, so I would like to reproduce it first before analyzing it.

Thanks.

#3

I can’t reproduce it, I catch this error in Crashlytics on my project, also I don’t have new Samsung device from the list of above

There are some of new exceptions only from Samsung devices:

Fatal Exception: java.lang.OutOfMemoryError
Failed to allocate a 16 byte allocation with 1283056 free bytes and 1252KB until OOM, target footprint 201326592, growth limit 201326592; failed due to fragmentation (largest possible contiguous allocation 55836672 bytes)

  1. android.os.ThreadLocalWorkSource.setUid (ThreadLocalWorkSource.java:68)
    android.os.Binder.execTransact (Binder.java:1027)

  2. java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.addConditionWaiter (AbstractQueuedSynchronizer.java:1882)
    java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await (AbstractQueuedSynchronizer.java:2063)
    java.util.concurrent.LinkedBlockingQueue.take (LinkedBlockingQueue.java:442)
    java.util.concurrent.ThreadPoolExecutor.getTask (ThreadPoolExecutor.java:1092)
    java.util.concurrent.ThreadPoolExecutor.runWorker (ThreadPoolExecutor.java:1152)
    java.util.concurrent.ThreadPoolExecutor$Worker.run (ThreadPoolExecutor.java:641)
    java.lang.Thread.run (Thread.java:923)

It looks as internal error of firmware, could you check the difference with original implementation with AOSP from java.util.concurrent package?

#4

This also seem to happen when running updates within F-Droid client.

Would you please help me to reproduce the issue?
As I did not face the issue, so I would like to reproduce it first before analyzing it.

Sure:

  1. download the open source app store F-Droid
  2. install some outdated application (at least ten) versions and properly close F-Droid
  3. start F-Droid and try to update the apps

You can also see videos of the responsiveness and stack traces at this thread

#5

Any update on this issue?
We are also seeing this issue happening only on Samsung devices in our crash monitoring report.

#6

I see the same OOM issue in Crashlytics, mostly on Galaxy S21+ runnin g Android 11, however I can’t reporoduce it by myself. All devices have a lot of free memory, like 1.5-2GB. All stacktraces look the same:

Fatal Exception: java.lang.OutOfMemoryError: Failed to allocate a 16 byte allocation with 246104 free bytes and 240KB until OOM, target footprint 536870912, growth limit 536870912; failed due to fragmentation (largest possible contiguous allocation 394526720 bytes)
at android.os.ThreadLocalWorkSource.setUid(ThreadLocalWorkSource.java:68)
at android.os.Binder.execTransact(Binder.java:1157)

#7

I see a similar OOM issue in Crashlytics, generally on Galaxy S21+ runnin g Android 11, anyway I can’t reporoduce it without help from anyone else. All gadgets have a ton of free memory, as 1.5-2GB. All stacktraces appear to be identical:

Lethal Exception: java.lang.OutOfMemoryError: Failed to distribute a 16 byte portion with 246104 free bytes and 240KB until OOM, target impression 536870912, development limit 536870912; flopped because of discontinuity (biggest conceivable coterminous allotment 394526720 bytes)

at android.os.ThreadLocalWorkSource.setUid(ThreadLocalWorkSource.java:68)

at android.os.Binder.execTransact(Binder.java:1157)

#8

I see a similar OOM issue in Crashlytics, generally on Galaxy S21+ runnin g Android 11, anyway I can’t reporoduce it without anyone else. All gadgets have a ton of free memory, as 1.5-2GB. All stacktraces appear to be identical:

Lethal Exception: java.lang.OutOfMemoryError: Failed to apportion a 16 byte designation with 246104 free bytes and 240KB until OOM, target impression 536870912, development limit 536870912; bombed because of discontinuity (biggest conceivable adjoining assignment 394526720 bytes)

at android.os.ThreadLocalWorkSource.setUid(ThreadLocalWorkSource.java:68)

at android.os.Binder.execTransact(Binder.java:1157)

#9

I see a comparative OOM issue in Crash lytics, for the most part on Galaxy S21+ running Android 11, anyway I can’t reproduce it without anyone else. All contraptions have a huge load of free memory, as 1.5-2GB. All stacktraces seem, by all accounts, to be indistinguishable:

Deadly Exception: java.lang.OutOfMemoryError: Failed to distribute a 16 byte assignment with 246104 free bytes and 240KB until OOM, target impression 536870912, advancement limit 536870912; besieged as a result of brokenness (greatest possible connecting task 394526720 bytes)

at android.os.ThreadLocalWorkSource.setUid(ThreadLocalWorkSource.java:68)

#10

It seems that this is solved after upgrading to Android 12 / One UI 4.
So this issues is definitely on Samsung’s side and should also be backported to supported devices.