<Android> Tensorflow crashs on Android15 ( with 16kb page size support)

[vegeta2102]

What happened?

https://developer.android.com/guide/practices/page-sizes
Tensorflow Android lite crashs on Android15 ( with 16kb page size support)

Reference: https://developer.android.com/guide/practices/page-sizes

Relevant code

I am using example project for testing on Android 15 (16kb page size)
https://github.com/tensorflow/examples/tree/master/lite/examples/image_segmentation/android

Relevant log output

06-17 15:34:10.241  9723  9723 F libc    : Fatal signal 11 (SIGSEGV), code 2 (SEGV_ACCERR), fault addr 0x70c48b9cf828 in tid 9723 (agesegmentation), pid 9723 (agesegmentation)
06-17 15:34:10.332  9744  9744 F DEBUG   : *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
06-17 15:34:10.332  9744  9744 F DEBUG   : Build fingerprint: 'google/sdk_gphone16k_arm64/emu64a16k:VanillaIceCream/AP31.240426.022.B1/11834390:userdebug/dev-keys'
06-17 15:34:10.332  9744  9744 F DEBUG   : Revision: '0'
06-17 15:34:10.332  9744  9744 F DEBUG   : ABI: 'arm64'
06-17 15:34:10.332  9744  9744 F DEBUG   : Timestamp: 2024-06-17 15:34:10.259545328+0900
06-17 15:34:10.332  9744  9744 F DEBUG   : Process uptime: 1s
06-17 15:34:10.332  9744  9744 F DEBUG   : Cmdline: org.tensorflow.lite.examples.imagesegmentation
06-17 15:34:10.332  9744  9744 F DEBUG   : pid: 9723, tid: 9723, name: agesegmentation  >>> org.tensorflow.lite.examples.imagesegmentation <<<
06-17 15:34:10.332  9744  9744 F DEBUG   : uid: 10193
06-17 15:34:10.332  9744  9744 F DEBUG   : tagged_addr_ctrl: 0000000000000001 (PR_TAGGED_ADDR_ENABLE)
06-17 15:34:10.332  9744  9744 F DEBUG   : pac_enabled_keys: 000000000000000f (PR_PAC_APIAKEY, PR_PAC_APIBKEY, PR_PAC_APDAKEY, PR_PAC_APDBKEY)
06-17 15:34:10.332  9744  9744 F DEBUG   : signal 11 (SIGSEGV), code 2 (SEGV_ACCERR), fault addr 0x000070c48b9cf828
06-17 15:34:10.332  9744  9744 F DEBUG   :     x0  000070c48b9cf828  x1  000070c48b9cf828  x2  00007fffe0f5df00  x3  000070c8659f8a80
06-17 15:34:10.332  9744  9744 F DEBUG   :     x4  000070c8659f8b00  x5  0000000000000004  x6  0000000001414d4c  x7  000070c8659f8004
06-17 15:34:10.332  9744  9744 F DEBUG   :     x8  0000000000000001  x9  000070c8659f8d00  x10 000000000000003f  x11 757274736e6f6320
06-17 15:34:10.332  9744  9744 F DEBUG   :     x12 676e696c6c61632b  x13 0000000000000000  x14 0000000000000000  x15 0000000000000000
06-17 15:34:10.332  9744  9744 F DEBUG   :     x16 000070c866d254a0  x17 000070c866c21c40  x18 000070c8663a8000  x19 000070c865a71710
06-17 15:34:10.332  9744  9744 F DEBUG   :     x20 000070c8659f8d00  x21 000070c48b9b8a10  x22 0000000000000000  x23 000070c48b372dcc
06-17 15:34:10.332  9744  9744 F DEBUG   :     x24 000070c866d53000  x25 00000000000000a4  x26 000070c866d4f000  x27 000070c866d4f000
06-17 15:34:10.332  9744  9744 F DEBUG   :     x28 000070c866d4f000  x29 00007fffe0f50940
06-17 15:34:10.332  9744  9744 F DEBUG   :     lr  000070c866c6f800  sp  00007fffe0f50920  pc  000070c48b4fa838  pst 0000000080001000
06-17 15:34:10.332  9744  9744 F DEBUG   : 11 total frames
06-17 15:34:10.332  9744  9744 F DEBUG   : backtrace:
06-17 15:34:10.332  9744  9744 F DEBUG   :       #00 pc 00000000001ce838  /data/app/~~Zpfpalz8lY94bs3oslVwBQ==/org.tensorflow.lite.examples.imagesegmentation-ZXFXOTBkRhWbOCD3F_qNfA==/lib/arm64/libtask_vision_jni.so
06-17 15:34:10.332  9744  9744 F DEBUG   :       #01 pc 00000000000b37fc  /apex/com.android.runtime/bin/linker64 (__dl__ZN6soinfo17call_constructorsEv+760) (BuildId: c08adba44c0fc120e06d56d862fce57e)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #02 pc 0000000000099a4c  /apex/com.android.runtime/bin/linker64 (__dl__Z9do_dlopenPKciPK17android_dlextinfoPKv+732) (BuildId: c08adba44c0fc120e06d56d862fce57e)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #03 pc 00000000000950a8  /apex/com.android.runtime/bin/linker64 (__dl__ZL10dlopen_extPKciPK17android_dlextinfoPKv.__uniq.234527301065430621646263515731762262959+72) (BuildId: c08adba44c0fc120e06d56d862fce57e)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #04 pc 00000000000040c8  /apex/com.android.runtime/lib64/bionic/libdl.so (android_dlopen_ext+16) (BuildId: 78ed0a0bc595ebbb171246d2fec02b93)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #05 pc 000000000001bf30  /apex/com.android.art/lib64/libnativeloader.so (android::NativeLoaderNamespace::Load(char const*) const+140) (BuildId: 876011460da2b12581e2c19c5303598a)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #06 pc 000000000000d1a0  /apex/com.android.art/lib64/libnativeloader.so (OpenNativeLibrary+1196) (BuildId: 876011460da2b12581e2c19c5303598a)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #07 pc 000000000066bbbc  /apex/com.android.art/lib64/libart.so (art::JavaVMExt::LoadNativeLibrary(_JNIEnv*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, _jobject*, _jclass*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >*)+620) (BuildId: 00a927538d6e55b7346c6547b4436b00)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #08 pc 00000000000058a8  /apex/com.android.art/lib64/libopenjdkjvm.so (JVM_NativeLoad+364) (BuildId: fee884f6680e39017670ed6bdbe02f38)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #09 pc 00000000003c5370  /apex/com.android.art/lib64/libart.so (art_quick_generic_jni_trampoline+144) (BuildId: 00a927538d6e55b7346c6547b4436b00)
06-17 15:34:10.332  9744  9744 F DEBUG   :       #10 pc 0000000000021438  [anon_shmem:dalvik-jit-code-cache] (offset 0x2000000)

tensorflow_hub Version

0.12.0 (latest stable release)

TensorFlow Version

2.8 (latest stable release)

Other libraries

No response

Python Version

3.x

OS

macOS

[bharatjetti]

Hi @vegeta2102  Thank you so much for showing interest! We truly appreciate your enthusiasm and support.
We regret to inform you that we are not accepting new implementation requests or feature additions at this time. Please feel free to fork the repository and submit pull requests if you have any suggestions for enhancements, problem fixes, or ideas. We’d be more than happy to review and collaborate! If you would like to contribute , please follow our CONTRIBUTION GUIDELINES here and submit a pull request.
Thank you again for your interest, and we look forward to seeing your contributions!