crash when pthread_mutex_lock is called on a destroyed mutex

[juha-h]

Could this crash be caused by bug in baresip library:

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

backtrace:
  #00  pc 0x00000000000832a8  /apex/com.android.runtime/lib64/bionic/libc.so (abort+160)
  #01  pc 0x00000000000e86ac  /apex/com.android.runtime/lib64/bionic/libc.so (__fortify_fatal(char const*, ...)+120)
  #02  pc 0x00000000000e7c60  /apex/com.android.runtime/lib64/bionic/libc.so (HandleUsingDestroyedMutex(pthread_mutex_t*, char const*)+52)
  #03  pc 0x00000000000e7b10  /apex/com.android.runtime/lib64/bionic/libc.so (pthread_mutex_lock+268)
  #04  pc 0x000000000033de7c  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (mtx_lock+12) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #05  pc 0x00000000002c0938  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (aurecv_codec+28) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #06  pc 0x00000000002d0de4  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (rtpstat_print+120) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #07  pc 0x00000000002f8c08  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (vhprintf+32768) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #08  pc 0x00000000002f9cf4  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (re_vsdprintf+120) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #09  pc 0x00000000003385e4  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (sipsess_set_close_headers+148) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #10  pc 0x00000000002c506c  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (call_hangupf+124) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #11  pc 0x00000000002d949c  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (ua_hangupf+140) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)
  #12  pc 0x00000000002b7db0  /data/app/com.tutpro.baresip-Qkf_UzhtT4bHdGyrzEO9KQ==/split_config.arm64_v8a.apk!libbaresip.so (Java_com_tutpro_baresip_Api_ua_1hangup+132) (BuildId: 23d1a726d5930ecdd89e0b4db7cfa3cd)

It happens when ua_hangup is called like this:

JNIEXPORT void JNICALL Java_com_tutpro_baresip_Api_ua_1hangup(
        JNIEnv *env, jobject obj, jlong ua, jlong call, jint code, jstring reason)
{
    (void)obj;
    const uint16_t native_code = code;
    const char *native_reason = (*env)->GetStringUTFChars(env, reason, 0);
    LOGD("hanging up call %ld/%ld\n", (long)ua, (long)call);
    re_thread_enter();
    if (strlen(native_reason) == 0)
        ua_hangup((struct ua *)ua, (struct call *)call, native_code, NULL);
    else
        ua_hangup((struct ua *)ua, (struct call *)call, native_code, native_reason);
    re_thread_leave();
    (*env)->ReleaseStringUTFChars(env, reason, native_reason);
}

[juha-h]

Looks like the crash happens when user touches hangup button more than once rapidly. I don't know if it would make sense to add a test in the lib to check if ua_hangup has already called for the call.