当前位置：网站首页>The NDK portal: C

前两天在看AndroidSystem from boot up toSystemServerProcesses and services to create this part of knowledge,有很多C++的代码,So is going to take a one or two daysC和C++Under this part of the things a little preview,Haha I didn't dare say review,Let oneself can understand a little.So there was the birth of this content,Then put it in theNDK专栏中,Crap I also not much said,开搞！

一、Linux内存布局原理

①、内存定义

硬件角度：内存是计算机必不可少的一个组成部分,是与CPU沟通的桥梁,计算机中所有的程序都是运行在内存中的.
逻辑角度：内存是一块具备随机访问能力,支持读、写操作,用来存放程序及程序运行中产生的数据的区域.

②、内存单位

位（bit）：是计算机中最小的数据单位,每一位的状态只能是0或1
字节：1Byte=8bit,Is a basic unit of measurement in memory
KB：1KB=1024Byte,也就是1024个字节
MB：1MB=1024KB,类似的还有GB、TB

③、内存编址

In the computer memory in bytes addressing,Each storage unit can store a byte address（8个bit）的数据,CPU通过内存地址获取指令和数据,Don't care about the space the address represents specific in what position、怎么分布,Because hardware design to ensure that an address corresponds to a fixed space,所以说：内存地址和地址指向的空间共同构成了一个内存单元.

④、Memory address and distribution rules of

内存地址：内存地址通常用十六进制的数据表示,指向内存中某一块区域.
Memory address allocation rules：Memory allocation rule is continuous,一个挨着一个.When the object to apply for a memory,This object is assigned a code first to,This code is the memory address

比如：The above is a memory address map,In each byte is an object to,Don't worry as to where the object is on,We need to know is this a series of encoding is continuous.

⑤、内存对象

Pointer to the area of memory storage is the memory object,Memory object can be of different types,Specific what are the types of？

Int：4字节
Short：2字节
long：4字节
char：1字节
double：8字节
long double：10字节
引用类型

问题： C里面longWhy only type4个字节？

long和int在早期16When a computerint 2个字节,long 4个字节,The computer to now,一般32位和64位机器下long和int一样.而Java中的long类型是8个字节

⑥、内存组成

首先来看Android的内存组成：

Android APPIs based on the virtual machine to load,Memory allocation as shown in the above,The memory in the graph are in the virtual machine,虚拟机是由C和C++进行开发的,These memory out of the virtual machine is not exist.下面要说的CMemory allocation is based on the local hardware environment,Compared with android memory is much simple：

栈：栈又称堆栈,存放程序的局部变量（但不包括static声明的变量,static 意味着在数据段中存放变量）.除此以外,在函数被调用时,栈用来传递参数和返回值.由于栈的先进先出特点,所以栈特别方便用来保存/恢复调用现场.储动态内存分配,Developers need to manually assign,手工释放

堆：堆是用于存放进程运行中被动态分配的内存段,它的大小并不固定,可动态扩张或缩减.当进程调用malloc/free等函数分配内存时,新分配的内存就被动态添加到堆上（堆被扩张）/释放的内存从堆中被剔除（堆被缩减）

BSS段：通常是指用来存放程序中未初始化的全局变量的一块内存区域.BSS是英文Block Started by Symbol的简称.BSS段属于静态内存分配.

数据段：数据段（data segment）Usually refers to used to store the global variable has been initialized in the program of a memory area.数据段属于静态内存分配.

代码段：代码段就是程序中的可执行部分,Intuitive understanding of the code segment is the function of.

二、函数

2.1、函数定义

函数和JavaThe method is a concept,C是面向过程的,It is the core carrier of function.

定义：函数是一组一起执行一个任务的语句.每个C程序都至少有一个函数,即主函数main(),所有简单的程序都可以定义其他额外的函数.

//返回值 函数名 参数列表
return_type function_name(parameter list){
    body of the function //函数体
}

举个最简单的例子：

#include <iostream>
//定义函数func,返回值类型为int,入参类型为void,Meaning no parameters
int func(void) {
	printf("这是一个简单的函数"); //打印一句话
	return -1; //Give it a return value
}
int main()
{
	func(); //在main函数中进行调用
}

结果很显然,Print out the correct word：

注意：这里不能把func()函数定义在main()函数的下面,比如这样：

这样写编译器会报错：

原因：在CThe implementation of language function is top-down to load,So when it executes tomain()函数的时候,它没有发现func()函数,所以会报错.这一点跟Java是不一样的,因为Java是面向对象的,它会把整个class文件加载到内存中,Up and down so it won't have this kind of concept.CThe language of this way of execution is more similar to the front ofDOM解析,Is also a top-down parsing.

解决方案：Based on the above this kind of situation has two solutions,The first is like the above the beginning,把func()函数定义在main()函数的上面,The second way is also define a header file.h,然后在.c文件中引入：

//类似于Java中的接口,Only a function declaration,没有函数的实现
int func(void);
int func(int a, int b);

#include <iostream> //System header files with<>引进来
#include "CTest.h" //自己的头文件用""引进来
int main()
{
	func(); //在main函数中进行调用
	func(1, 2);
}
//定义函数func,返回值类型为int,入参类型为void,Meaning no parameters
int func(void) {
	printf("这是一个简单的函数"); //打印一句话
	return -1; //Give it a return value
}
int func(int a,int b) {
	printf("\nThis is a function of overloaded"); //打印一句话
	return -1; //Give it a return value
}

So to perform don't have any problem：

2.2、指针函数

定义：指针函数是指带指针的函数,即本质是一个函数.函数返回类型是某一类型的指针.比如：

举个例子：

//返回值类型为int*,入参void*For unsigned type pointer,相当于Java中的Object,So as long as pointer types are hereOK
int* pointerFuc(void* param) { 
	printf("指针函数");
	int a = 100;
	int *b = &a;
	return b; //The return value for pointer type in here,Use address operator to obtain variablea的地址
}
int main()
{
	int a = 10;
	pointerFuc(&a); //Pass a pointer type
}

2.3、函数指针

指针函数是一个函数,Its return value is a pointer type,The following function pointer it is not a function,而是一个变量.

定义：函数指针是指向函数的指针变量,即本质是一个指针变量.比如：

举个例子：

CTest.h

//声明了一个变量,This variable is only function
void(*funcp)(int a); //（*func）Parentheses represent its priority,一定要写
void point_func(int a);

CTest.c

int main()
{
	funcp = point_func; //The functions assigned to the above definition of pointer variable,In the process of assignment does not produce call
	funcp(4); //手动调用funcp(),Only here can make call
}
void point_func(int a) {
	printf("函数指针\n");
}

执行结果：

三、指针数组

3.1、数组

数组是一个数据的集合,Reflected in the memory is a continuity of memory area,比如,我们可以声明一个int类型的数组,长度为3,然后使用一个指针*p指向数组的首地址,This can operate it by pointer array：

#include <iostream>;
int main() {
	//定义int类型的数组
	int arr[] = {10,20,30};
	for (int i = 0; i < 3; i++){
		printf("数组元素 %d\n", arr[i]);
	}
	int*p = arr; //将arr赋值给指针p,指向数组的第一个元素
	*p = 88; //The first element array is modified to10
	for (int i = 0; i < 3; i++)
	{
		printf("The modified array elements %d\n", arr[i]);
	}
	return 0;
}

执行结果为：

The above example pointer is the default to the first address of the array,Below we through the pointer arithmetic to modify the second element of an array of,此时我们只需要修改 *p=88; 这一行代码即可,Add the pointer1修改为*(p+1) = 88; 这样一来,The results also became like this：

3.2、指针数组

Pointer to a pointer array as the name suggests is the each element in the array,比如有一个数组arr,Its each element to store an address,Each address all point to a variable,Such arrays we call it a pointer array.

Let's define a pointer array：

#include <iostream>;
int main() {
	int arr[] = {10,20,30};//定义int类型的数组
	int *p[3]; //指针数组
	//补充：指针优先级 () > [] > *
	for (int i = 0; i < 3; i++)
	{
		p[i] = &arr[i]; //For each element of the pointer array assignment,&是取地址符
		printf("Array elements after operation%d\n", *p[i]); //*On behalf of pointing to the variable
	}
	return 0;
}

注意：在指针数组中 int *p[3]; This expression is a priority,Define the priority order for：()>[ ]>*

3.3、数组指针（行指针）

定义：int (*p)[n]

优先级高,首先说明p是一个指针,指向一个整型的一维数组,长度为n,也可以说是p的步长.也就是说执行p+1时,p要跨过n个整型数据的长度.

指针数组：代表的是一个数组,Each element deposit is an address

数组指针：不是一个数组,只是一个变量,Storage is a pointer,Its step changed,It is the address of the first point to a two dimensional array,If the current pointer+1操作,It will point to the second row first address,It forms and pointer array are not the same.举个例子如下：

四、结构体

4.1、结构体定义

结构体类似于Java中的Java Bean,Some variables can be encapsulated as an object,同样的在C语言中,Some variables can be encapsulated into a struct,The following view structure is how to say？Here are three kinds of definition：

第一种：

#include <iostream>;
//Student相当于类名
struct Student
{
	int height;
	int age;
};
int main() {
	struct Student stu; //声明一个struct类型的变量,类似于Java中的private Person p;
	stu.height= 175; //赋值
	stu.age = 18;
	return 0;
}

第二种：

#include <iostream>;
//h和a表示结构变量,也就是Student类型的变量
struct Student
{
	int height;
	int age;
}h,a;
int main() {
	h.height = 175;
	h.age = 18;
	return 0;
}

第三种：

#include <iostream>;
//使用typedef定义
typedef struct {
	int height;
	int age;
}Student;
int main() {
	Student stu;
	stu.height = 175;
	stu.age = 18;
	return 0;
}

4.2、内存对齐

在介绍这个概念之前,Us to define the structure again and how much to print the size of the structure is：

#include <iostream>;
//使用typedef定义
typedef struct {
	int height;
	short age;
}Student;

int main() {
	Student stu;
	stu.height = 175;
	stu.age = 18;
	printf("结构体大小%d", sizeof(stu));
	return 0;
}

嗯？一个int类型,一个short类型,怎么变成8了呢？在C语言中,To the object structure allocates memory,It will follow the memory alignment rules.

定义：对齐跟数据在内存中的位置有关.如果一个变量的内存地址正好位于它长度的整数倍,It is known as a natural alignment.比如在32位cpu下,假设一个整型变量的地址为0x00000004,It is naturally aligned.

When structures need too much memory,使用动态内存申请.结构体占用字节数和结构体内字段有关,指针占用内存就是4/8字节,So the pointer is higher than by value efficiency.

Structure principle of storage：

结构体变量中成员的偏移量必须是成员大小的整数倍（0被认为是任何数的整数倍）
结构体大小必须是所有成员大小的整数倍,也即所有成员大小的公倍数

在32位操作系统中,数据总线是32位,地址总线是32位.Because the address bus is32位,Means addressing space is according to the4递增的,数据总线32A a read/write4byte.

Then to see such a struct：

According to the common sense should be the total takes up6个字节,But in fact it was：

五、共用体

共用体和结构体类似,Said is a collection of data type,Structure in the body just save cost more.

定义：共用体是一种特殊的数据类型,允许在相同的内存位置存储不同的数据类型.可以定义一个带有多成员的共用体,但是任何时候只能有一个成员带有值.共用体提供了一种使用相同的内存位置的有效方式.举个例子如下：

union Data
{
	int i;
	float f;
	char str[20];
}data;

共用体占用的内存应足够存储共用体中最大的成员.比如在上面的案例中,In each member variables,charTake up the space is one of the biggest,所以整个Data将占用20个字节的内存空间.

The following to define a common body together：

#include <iostream>;
union MyStudent
{
	int i;
	float f;
};
int main() {
	union MyStudent myStu;
	myStu.i = 10;
	myStu.f = 11;
	printf("i的地址   %#x\n", &myStu.i);
	printf("f的地址   %#x\n", &myStu.f);
	return 0;
}

执行结果为：Can see the starting address of each member of the share in the body is the same,Only in memory, respectively is a member of the different variables

And then to modify the code：

#include <iostream>;
union MyStudent
{
	int i;
	int j;
};

int main() {
	union MyStudent myStu;
	myStu.i = 10;
	printf("之前 %d\n", myStu.i);
	myStu.j = 11;
	printf("i的地址   %#x\n", &myStu.i);
	printf("f的地址   %#x\n", &myStu.j);
	printf("之后 %d\n", myStu.i);
	printf("之后 %d\n", myStu.j);
	return 0;
}

执行结果如下：

We found a strange results at this time,给j赋值之后,i和j的值都变成了11,这是为什么呢？

因为：It order to save the open space on the memory size,将之前iValue area of memory to reusej值,This reuse scene in general use in common body has a lot of variable declaration,最开始初始化的时候iValues in the code used to,但是在下面的代码中iValues do not need to,This saves memory performance is very high.

六、动态库

6.1、库的概念

在windows平台和LinuxPlatform has a large number of library,Android中同样也是.Library as the name implies,Refers to a container file,Inside is a function,由于windows和Linux的平台不同（Mainly is the compiler、The compiler and the connector is different）,因此二者库的二进制是不兼容的.

库的种类：库分为动态库和静态库,Below are（前者为Linux平台,后者为Windows平台）Different forms of：