Rive crashes the app randomly on lowend devices.

[dawidpodolak]

We have noticed crashes reported in Google Play and Sentry that appear to be caused by Rive. Here is the stack trace. We weren’t able to reproduce these crashes in our debug build. We also observed that they mainly occur on low-end devices.

Rive version: 10.5.1 (we are about to release version with 10.5.3). It occurs for earlier versions too.
Android version: from 7 to 16

FORTIFY: pthread_mutex_lock called on a destroyed mutex (0x<sanitized>)

*** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
pid: 0, tid: 12269 >>> com.zen.application <<<

backtrace:
  #00  pc 0x00000000000530f4  /apex/com.android.runtime/lib64/bionic/libc.so (abort+164)
  #01  pc 0x0000000000054f7c  /apex/com.android.runtime/lib64/bionic/libc.so (__fortify_fatal(char const*, ...)+124)
  #02  pc 0x00000000000c29dc  /apex/com.android.runtime/lib64/bionic/libc.so (HandleUsingDestroyedMutex(pthread_mutex_t*, char const*)+60)
  #03  pc 0x00000000000c2870  /apex/com.android.runtime/lib64/bionic/libc.so (pthread_mutex_lock+208)
  #04  pc 0x00000000017dd044  /vendor/lib64/egl/libGLES_mali.so
  #05  pc 0x00000000017dd51c  /vendor/lib64/egl/libGLES_mali.so
  #06  pc 0x00000000008987c8  /vendor/lib64/egl/libGLES_mali.so
  #07  pc 0x00000000008993fc  /vendor/lib64/egl/libGLES_mali.so
  #08  pc 0x00000000007d88bc  /vendor/lib64/egl/libGLES_mali.so
  #09  pc 0x00000000007f2540  /vendor/lib64/egl/libGLES_mali.so
  #10  pc 0x00000000007f2f84  /vendor/lib64/egl/libGLES_mali.so
  #11  pc 0x00000000007e4874  /vendor/lib64/egl/libGLES_mali.so
  #12  pc 0x0000000000792e28  /vendor/lib64/egl/libGLES_mali.so (glTexSubImage2D+88)
  #13  pc 0x000000000035329c  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #14  pc 0x0000000000350f18  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (rive::gpu::RenderContextGLImpl::flush(rive::gpu::FlushDescriptor const&)+140) (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #15  pc 0x000000000035d684  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (rive::gpu::RenderContext::flush(rive::gpu::RenderContext::FlushResources const&)+2752) (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #16  pc 0x00000000001b1c98  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #17  pc 0x0000000000213f10  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (rive::CommandServer::processCommands()+40692) (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #18  pc 0x0000000000209fa4  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (rive::CommandServer::serveUntilDisconnect()+56) (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #19  pc 0x00000000001b1760  /data/app/~~NjEMQmaIW_4JInzlXZnFcA==/com.zen.application-X0uDAQreOJip39hS7AEtGg==/split_config.arm64_v8a.apk!librive-android.so (BuildId: d5c158c73dcd67af197045cfbf6b7f6220fd429f)
  #20  pc 0x00000000000c163c  /apex/com.android.runtime/lib64/bionic/libc.so (__pthread_start(void*)+204)
  #21  pc 0x0000000000054930  /apex/com.android.runtime/lib64/bionic/libc.so (__start_thread+64)

[nikolai-krasilnikov-perplexity]

We're experiencing a similar crash that appears to have the same root cause - a race condition in CommandServer during cleanup.

Our Stack Trace

Fatal signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0x10 in tid (Rive CmdServer)

#00 pc 00000000003422c0  librive-android.so
#01 pc 000000000033da64  librive-android.so  rive::gpu::RenderContextGLImpl::flush(rive::gpu::FlushDescriptor const&)+2972
#02 pc 0000000000348f30  librive-android.so  rive::gpu::RenderContext::flush(rive::gpu::RenderContext::FlushResources const&)+1776
#03 pc 00000000001a66a4  librive-android.so
#04 pc 0000000000205fa4  librive-android.so  rive::CommandServer::processCommands()+39368
#05 pc 00000000001fc564  librive-android.so  rive::CommandServer::serveUntilDisconnect()+56

Key Difference

• Your crash: pthread_mutex_lock on destroyed mutex → FORTIFY abort (Mali GPU)
• Our crash: SIGSEGV at address 0x10 → null pointer dereference (various GPUs)

Both crashes occur in the same code path: CommandServer::processCommands() → RenderContext::flush()

Environment

• rive-android version: 10.4.4
• OS distribution:
  • Android 16: 52%
  • Android 15: 34%
  • Android 14: 7%
• Devices: Samsung Galaxy S24 Ultra, Pixel, Xiaomi, Infinix (80% high-end)
• Device class: 80% high-end (opposite of your low-end pattern)

Reproduction

The crash occurs when users navigate away from a screen containing a Rive animation (using Compose RiveUI). It's rare and we cannot reproduce it in development - only captured via Sentry from production users.

Conclusion

This suggests the race condition affects both low-end (Mali) and high-end devices, just manifesting differently depending on GPU driver behavior. The underlying issue appears to be improper synchronization during resource cleanup in CommandServer.

[OlehHavryshV]

We have exactly the same issue as @nikolai-krasilnikov-perplexity

Rive runtime version is 10.5.3, ComposeUI. So the issue exists in the latest version as well