The mutex itself is stored in a shared list, so it should be valid. However, at random times, crash happens due to sigabrt with the trace below:
[ 4] 0x00007fff86b513d4 libsystem_pthread.dylib`__pthread_abort_reason + 179
[ 5] 0x00007fff86b4ddf0 libsystem_pthread.dylib`_pthread_mutex_unlock_drop + 160
Looking at the code it's not obvious if something can be causing the mutex value to be invalid, is there a good way to debug this?
> The mutex itself is stored in a shared list, so it should be valid.
What does this mean? What's a "shared list" in this context? Why does that imply the mutex should be valid?
How was the mutex created? What attributes? Have you tried pthread_mutexattr_settype() with PTHREAD_MUTEX_ERRORCHECK (if you're not already using PTHREAD_MUTEX_RECURSIVE)?
Is the unlock call happening on the same thread which locked the mutex?
Is there any chance you have a memory smashing bug that's overwritten parts of the mutex object? What's just before or after it in memory? Perhaps you're over/underflowing that?
