Relationship between fork and pipeline

First of all, let's understand fork After the implementation , Two processes will appear ：

 

problem :fork Is used to create processes , The child process copies the same code as the parent process , But with the pipe Continuous time , Also copy pipe Do you ？ It's a self-directed thing , Read something , How to realize interprocess communication ？
answer :
        fork() Not only create a child process with the same code as the parent process , And the parent process is fork All context scenarios of the execution point are also automatically copied into the subprocess , Include ：（1） Global and local variables （2） Open file handle （3） Shared memory 、 Synchronization objects such as messages .
         because pipe Calling is equivalent to creating 2 File handles , So in fork The latter two handles are also automatically copied to the child process , The operation effect of these two handles is consistent with that of the main process , This makes it easy for parent-child processes to exchange data through this mechanism , for instance ：
         hypothesis pipe The resulting handle is P[0],P[1], stay fork After the father and son processes have P[0],P[1] Handle , that ：
— The parent process can send its own P[1] Middle write data , Subprocesses from P[0] The data can be read out in ; Remember that the parent process cannot also start from P[0] Reading data , Otherwise, the data written by yourself may be read by yourself first
— vice versa , Subprocess to P[1] After data is written in , The parent process starts from P[0] This data can be read in ; Remember that the subprocess should not start from P[0] Read data .
         You may not understand why the data read and write inside the process will be transferred to another process , But don't forget ,pipe Anonymous pipes and files ,socket And so on are the management objects of the operating system , It is read and written by OS Kernel code to control . Call... In the parent process pipe After creating the pipe , The pipeline object is stored in OS Inside , Obtained by the parent process P[0] and P[1] Are just references to this object （ Equivalent to pointer ）; stay fork After the subprocess , The child process copies only the same pointer , It's still pointing to OS The same object in , This is why parent-child processes can communicate with each other through it .
————————————————
Link to the original text ：https://blog.csdn.net/yizhou35/article/details/12288023