Shulou(Shulou.com)06/02 Report--

What do you need to pay attention to in JVM memory? in view of this problem, this article introduces the corresponding analysis and solution in detail, hoping to help more partners who want to solve this problem to find a more simple and feasible way.

Preface

For programmers developed in the C language, in terms of memory management, they must be responsible for the life cycle of each object, from something to nothing.

For Java programmers, with the help of virtual machine memory management, there is no need to match free operations for every new object, and memory leaks and memory spills are not easy to occur, but it is precisely because memory management is handed over to the virtual machine, once the running program has a memory leak problem, it is very difficult to troubleshoot the process. Therefore, only when you understand the running mechanism of the Java virtual machine, can you manipulate all kinds of code. This article takes HotSpot as an example to talk about virtual machines.

The JAVA virtual machine divides the managed memory into several different data regions.

Java reactor

Java heap is a memory area shared by all threads, which is mainly used to store object instances. There is a description in the Java virtual machine specification that all object instances and data should be allocated on the heap. Allocating memory to an object is to separate a determined piece of memory from heap memory, and there are usually two ways to do this:

1. Pointer collision method

Assuming that the memory in the Java heap is complete, and the allocated memory and free memory are on different sides, through a pointer as a demarcation point, when you need to allocate memory, you only need to move the pointer to the idle end equal to the size of the object.

2. Free list method

In fact, the memory of the Java heap is not complete, and the allocated memory and free memory are interlaced with each other. JVM records the available memory blocks by maintaining a list, and when the allocation operation occurs, it finds a large enough memory block from the list to allocate to the object instance, and updates the records on the list.

Object creation is a very frequent behavior, and multithreading concurrency needs to be considered when allocating heap memory. It may occur that memory is being allocated to object A, the pointer or record has not been updated, and object B is allocated to the original memory at the same time. There are two solutions to this problem:

1. Using CAS to ensure the atomicity of data update operation.

2. The behavior of memory allocation is divided according to threads and carried out in different spaces. Each thread pre-allocates a memory block in the Java heap, which is called local thread allocation buffer (Thread Local Allocation Buffer, TLAB).

Java stack

Java stack is private to threads, and each thread corresponds to a Java stack. Each thread creates a corresponding stack frame (Stack Frame) when executing a method. The stack frame is responsible for storing information such as local variable table, Operand stack, dynamic link and method return address. The calling process of each method is equivalent to the process of loading and unloading stack frames on the Java stack.

The local variable table is used to store method parameters and local variables defined within the method, whose size has been determined during code compilation and will not change during the method run. The local variable table takes the variable slot (Slot) as the minimum storage unit, and each Slot can store 32-bit data of boolean, byte, char, shot, int, float, reference and returnAddress types. For 64-bit data types long and double, the virtual machine allocates two consecutive Slot spaces in a high-level alignment way.

When the method is executed, if it is an instance method, that is, a non-static method, the Slot of bit 0 in the local variable table stores the reference of the object instance by default, which can be accessed through the keyword this in the method. The parameters of the method are assigned from the first bit Slot according to the order of the parameter list, and the internal variables of the method are assigned the rest of the Slot according to the defined order.

Class test {public int calc (int a, int b, String operation) {operation = "+"; return a + b;} public void main (String args []) {calc (100,200, "+");}}

The corresponding local variables are as follows:

Use the javap-c command to view the bytecode of the method calc

Iload_1 and iload_2 loaded data from the first and second bits of the local variable scale, respectively.

Method area

The method area, like the Java heap, is a memory area shared by all threads, which is used to store data such as class information, constants, static variables and compiled code that have been loaded by the virtual machine.

The runtime pool is part of the method area that holds various literal constants and symbolic references generated during compilation.

Instruction counter

The instruction counter is private to the thread, and each thread has an independent instruction counter, which records the address of the bytecode instruction being executed by the virtual machine. branch, loop, jump, exception handling and thread recovery all rely on this counter. If the thread is executing the native method, this counter is empty.

Memory layout of the object

The layout of objects in memory can be divided into three areas: object headers, instance data, and alignment padding.

1. Object head

The object header contains two pieces of information: run-time data and a type pointer, and if the object is an array, you need a piece of data to record the length of the array.

1.1.Runtime data includes HashCode, GC generational age, lock status flag, lock held by thread, biased lock ID, biased timestamp and so on. The length of this data in 32-bit and 64-bit virtual machines is 32bit and 64bit respectively, officially known as "Mark Word". Mark Word is designed as a non-fixed data structure to store as much data as possible in limited space.

In a 32-bit virtual machine, when the object is not locked, the HashCode of the 25bit storage object, the generation age of the 4bit storage object, the 2bit storage lock flag bit, and the 1bit of the 25bit storage object in the 32bit of Mark Word are fixed to 0, as follows:

The storage contents of Mark Word under other states (lightweight lock, heavyweight lock, GC lock, biased lock) are as follows:

1.2. The type pointer of the object header points to the class metadata of the object, and the virtual machine can determine which class instance the object is through this pointer.

2. Instance data

Instance data are various types of fields defined in the program code, including those inherited from the parent class. The storage order of this part is affected by the virtual machine allocation policy and the order in which the fields are defined in the source code.

3. Align the fill

Because the automatic memory management of HotSpot requires that the starting address of the object must be an integer multiple of 8 bytes, that is, the size of the object must be an integral multiple of 8 bytes, and the data of the object header is exactly an integral multiple of 8 bytes, so when the instance data is less than 8 bytes, it needs to be filled by alignment.

This is the answer to the questions that need to be paid attention to in JVM memory. I hope the above content can be of some help to you. If you still have a lot of doubts to be solved, you can follow the industry information channel to learn more about it.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.