Due to the independence between processes , Let two unrelated processes communicate , The first problem to be solved is to let different processes see the same resource （ Memory files 、 Memory 、 queue ）.

Anonymous pipeline It's through “ The child process will inherit the resources of the parent process ” This feature realizes the communication between parent and child processes , In this way, the parent-child process can see the same resource

name pipes It's through   route + file name   See the same resource . The most typical is like this , After knowing the path , A process writes data to the disk file first , Another process reads data from the disk file , But such reading and writing efficiency is relatively low

Let's follow this idea to understand how the named pipeline acts as a bridge between two processes

---

Catalog

One 、 How named pipes work

Two 、 Named pipe creation function mkfifo

3、 ... and 、 Pipeline file creation and use

1、 Create a pipeline in the server process

2、 The server reads data from the pipeline

3、  The client writes data into the pipeline  

4、 Process communication test

---

One 、 How named pipes work

A file can be opened by multiple processes , The most typical is ours stdout . After the file is opened , Will be loaded into memory , We let the process A Open in write only mode , process A When writing , The disk cannot be refreshed ; process B Open as read-only

  The process then A Write data to this memory , process B Read data from this memory , This enables basic communication between unrelated processes

Two 、 Named pipe creation function mkfifo

mkfifo After being called , Will create a pipeline file ( The essence of pipeline is file ), Equivalent to the disk file in the figure above

The first parameter ： Indicates the creation path of the pipeline file , Where to create this file

The second parameter ： Indicates the initial permission of the pipeline file , When a general file is created , Permissions include readable / Can write / Executable

Return value ： When the pipeline is created successfully , return 0; Otherwise return to -1

3、 ... and 、 Pipeline file creation and use

Next, we will try to let the two processes communicate through the pipeline , A process is the server , A process is a client . The server is responsible for reading , The client is responsible for writing

1、 Create a pipeline in the server process

Generally speaking , The server creates a pipeline , Then the client can communicate directly with the pipeline created by the server

You can write one here Makefile Test whether it can be created successfully

Makefile The contents and test results are as follows , If there is a pipeline file like this , That means the creation is successful

 

2、 The server reads data from the pipeline

After creating the pipeline file , The server can open the file , Then read the data from it

Be careful ： 

Because the client wants to write data to the pipeline file , If the client closes the file , The server cannot read the data , That is, the number of bytes of the file read bytes = 0, There is no need to continue reading at this time , So you need to be in while Loop plus one close(fd)

3、  The client writes data into the pipeline  

After the pipe is created , That is, the disk file has , A process ( Client side ) With Just write To open a file , Write data to this file

But how does the client get the pipeline path ？？ We can move the macro definition of the pipeline path in the server to a header file common.h in , Let the client and server include this header file at the same time

======================== common.h ========================

because read Functions and write The following header files are required for all functions , So we put the following header files into common.h in , And then let server.c and client.c Both contain common.h

#include <sys/stat.h>
#include <sys/types.h>
#include <unistd.h>    //close function 

 ======================== server.c ========================

 ======================== client.c ========================

Let's read the content from the keyboard first , Then write the read content to the pipeline file , The specific operation steps are shown in the figure below

 

4、 Process communication test