I want to do some work after all C++ thread_local destructors called.
This is platform specific - Android, so I have access to pthreads.
The question is, when pthread_key_created destructors should be called, before or after C++ thread_local destructors? Or they can be interleaved?
I tested On Linux Mint and pthread destructors called after C++ 's.
bionic/pthread_exit.cpp currently has the same order:
void pthread_exit(void* return_value) {
// Call dtors for thread_local objects first.
__cxa_thread_finalize();
// Call the TLS destructors.
pthread_key_clean_all();
However, this is not documented behavior and you should not build something relying on it.
libstdc++ from GCC uses pthread_key_create in case the platform does not provide __cxa_thread_atexit_impl. In this case, C++ destructors run somewhere in the middle of the POSIX destructors.
To my knowledge, there is no standard which requires any particular behavior here because C++ does not know about POSIX and POSIX does not know about C++, so neither standard says what happens here. There are also some corner cases involving the resurrection of thread-local data during thread destruction which will vary among implementations. (A typical example is a per-thread logger object which is used to log from destructors of thread-local variables.)
Related
I have a C library which I'm cross-compiling to use in Android & iOS apps.
It makes use of memcpy() and mktime() so I want to know if these functions are implicitly thread-safe when used in multi-threaded environments.
iOS apps compiled with modern Xcode and Android libraries compiled with modern Android NDK use a clang compiler which is LLVM-based.
I've reviewed the following questions, but have been unable to find a definitive answer:
Is memcpy process-safe?
Are functions in the C standard library thread safe?
POSIX requires of conforming implementations that all functions it standardizes be thread safe, with the exception of a relatively short list of functions. memcpy() and mktime() are both covered by POSIX, and neither is on the list of exceptions, so POSIX requires them to be thread safe (but read on).
Note well, however, that this is not a matter of the compiler used, but rather of the C library that supports your application. I recall Apple's C libraries being non-conforming in some areas. Nevertheless, there's nothing in particular about memcpy() and mktime() that makes them inherently risky from a thread safety perspective. That is, there's no reason to expect that they access any shared data, except any provided to them via their arguments.
And there's the rub. You can rely on memcpy() and mktime() not to, say, rely internally on static data, but POSIX's requirement for thread safety does not extend to working as documented in the face of data races you create through choice of arguments. Thus, for example, if two different threads call memcpy(), and the target region of one call overlaps either the source or target region of the other, then you need some flavor of synchronization between the threads.
The question if memcpy() is thread-safe might be discussible.
I would say that memcpy() is indeed thread-safe. It doesn't rely on a (global) state, which could be mangled up by multiple instances of memcpy() running. This, however, doesn't mean, that there is some magic preventing a memory area, which is concurrently the copy destination of multiple threads doing memcpy() gets badly mangled up, i.e. the copy process as a whole is not atomic. You would have to care yourself using mutexes to ensure atomicity.
mktime() is trivially threadsafe, since it doesn't use static buffers, use a global state or similar. The manpage mentions a few functions from that family being not threadsafe (those have corresponding *_r functions), but mktime() is not amongst those.
I cannot find any references to a detailed explanation about how JNI works on Android in detail, so:
Since every Android application runs in its own process, with its own instance of the Dalvik/ART virtual machine, I think that the native code will be executed in the same process, am I right?
I read that when the VM invokes a function, it passes a JNIEnv pointer, a jobject pointer, and any Java arguments declared by the Java method.
But how is this made at assembly level (under the hood)?
I read that you can instantiate objects, call methods, and so on, like Reflection, using the functions provided by the JNIEnv. Therefore, my question is: have I a "direct" memory access to the VM or I have always to use the JNIEnv's functions?
The Android JVM is under Apache license, so the best detailed and precise description can be found in the form of source code. Note that there are two different JVMs: dalvik and art. Under the hood they are very different, to the extent that a user of JNI may consider special adaptations.
the native code will be executed in the same process
Exactly. Note that an Android app can run in more than one process, and also it can spawn child processes (normal Unix behavior). But JNI is not IPC.
how is this made at assembly level?
More or less, this is described in a related
question: What does a JVM have to do when calling a native method?
have I a "direct" memory access to the VM?
Yes, you have. There is no security barrier between your C code and the JVM. You can reverse engineer the data structures, and do whatever you like. The exact implementations of the JVM not only depend on the Android version, but may be modified without notice by the vendor, as long as the public API of the JVM (including JNI) is compatible. The chances that you will do something useful with direct memory access to JVM are minimal, but the risk that it will crash is very high.
Note that this is not a security issue: your C code is running in a separate process (with your Java code), and is subject to the same permissions restrictions as the Java code. It has no access to the private memory of other apps or procsses. Whatever you change in your instance of JVM will not effect VM that runs other apps.
I have some confusion about the life cycle of native code in Android aps. I have seen references that say that the native code is executed inside the Dalvik VM, but is that true? I was under the impression that the VM only runs Dalvik bytecode. On the otherhand, the native code uses JNI which is be called from Java inside the VM. Lastly, does the use of NativeActivity make any difference?
I thought I was understanding the NDK fairly well, until I sat down and tried to explain it to myself. I'm not even sure that I'm asking the question in a sensible manner.
I have seen references that say that the native code is executed inside the Dalvik VM, but is that true?
It executes inside a process that contains a Dalvik VM. Personally, I would not describe it as executing inside the VM -- as you say, Dalvik bytecode executes inside the VM. "Under the control of the Dalvik VM" would be better phrasing, IMHO. Of course, it boils down to your definition of "in", I suppose.
Lastly, does the use of NativeActivity make any difference?
Not really, insofar as NativeActivity is implemented in Java. While you may not have any Java, Java is still lightly involved in the act of running your native code.
I am very interrogative and perplexed by this commit on android's dalvik platform pushed a year ago.
File extensions were changed to C++ extensions in order to "move the interpreter into C++" - use the compiler's C++ front-end.
What could be the benefits of this change ? Dalvik Platform is a 100% C & asm project and not any C++ feature is used.
I can only speculate, but considering how the Android system has grown in complexity, the scoping features of C++ (classes and namespaces) might make the code base more manageable.
EDIT
Even if the project doesn't currently make use of any C++ features, they may simply be planning ahead.
Apart from some minor differences (namely some parameter conventions most people avoid anyway), C source code compiles as C++ without modification. That being said, in some areas C++ syntax is stricter than C (C allows you to assign a void pointer to another pointer type without a cast; in C++, this is an error), and enforcing this strictness avoids problems down the road.
*
*) (That's an overly simplistic view, see comment)
One further reason for the change may be that because most modern development favors C++ over C, a richer set of tools is available.
Speculating again, but at the birth of Android C may have been the only viable option for embedded device development, and now that restriction is no longer an issue.
This is about Android. The situation:
C++ library and java wrapper classes plus native functions (JNI) for working with C++ classes from the library. When common java code needs C++ object, it creates corresponding java wrapper object which creates C++ object through native function and remembers the pointer to the native object in 'long' variable. In all next actions the wrapper gives this pointer to the native functions etc.
The problem:
How to release all allocated C++ objects at the end? Currently every wrapper class has 'finalize' method where it calls native function for releasing of the C++ object, but Android doesn't guarantee the calling of 'finalize'! At the other side, normally the C++ library has no idea how many and what types of C++ objects are allocated by java code.
What will happens with the remaining allocated memory when our java application terminates, will Android release automatically the whole heap, used from the native library, when the OS unloads the library?
At the end of the process lifetime, all process memory (both Java and C++ heap) will be freed and reclaimed by the system. One thing is though, Android activity closing does not necessarily end the process. I'm not sure what's the process shutdown policy there.
On the other hand, relying on the garbage collection and finalize() sounds like solid design to me. You claim - "Android does not guarantee finalize()". Do you have a cite for that? 'Cause if it comes with a disclaimer of "when the object is freed as a part of process shutdown...", then we're still good.
And if you're super-paranoid, you can write your own malloc()/free()/realloc() wrapper, store a list of all allocated objects, and introduce a cleanup function that walks the list and frees them all. The containing Java objects, however, might end in a weird zombie state where the memory has been freed from under them. This is a tricky proposition that is very easy to get wrong. So I'd still say - have faith in the garbage collector. Lack thereof would be... disturbing.
Due to the difference in paradigms, you have to incorporate explicit destruction into your Java objects that are implemented under the hood using C++ resources. So a close() or other such method. The same issue comes up with the JNI, so answers to those questions will apply to you:
Force Java to call my C++ destructor (JNI)
As for the memory issue on closing, it's generally best in my opinion to not rely on this. If you get to a clean state, valgrind and such can make sure you weren't leaking.
But from a technical standpoint--since Android is based on Linux, I'd imagine it does the usual thing and will free all the memory when the process closes. Taking advantage of that can make your program exit faster than explicitly freeing memory (for experts only who use other methods to ensure this maintains program correctness and they aren't leaking at runtime).
We are using JNIs and we had a problem like that
Actually, the problem resided in the fact that we were overloading finalize() to do the clean up. We solved our problems by removing our finalize() and creating a clean() instead. This clean() calls the JNI function that does the appropriate deletes (and set the C++ pointers to null, just in case). We call clean() just as you would in C++ with delete (e.g. when the variable goes out of scope).
That worked for us. I hope it works for you. Good luck!