This program finds the path of a file in any current working directory as well as its subdirectories by utilizing a concurrent programming design and implementation so that the user can run many simultaneous finding processes.

This project was my final project for my systems programming class. It applies all the different concepts that I learned in the class, including how to use dup2(), pipe(), and signals.

To complete this project, I had to have a solid grip on how to establish a pipeline between the child and the parent processes. I also had to figure out how to use <dirent.h> and <sys/stat.h> so that I could parse specific files.

My Approach

I extensively researched all capabilities of pipe(), signal(), kill(), <dirent.h>, and <stat/sys.h>, and I parsed out what functionalities I thought I needed for the program. I then wrote procedures on how to use all these functionalities on sticky notes for my own reference.

This snippet of code from find.c demonstrates how I used a pipe in the program. (This is code in my child process, both the parent and the child have a copy of the pipe)

close(fd[0]); //close read
.
.
.
strcat(childreport,"\0");//null terminator is important, because pipe!
int length = strlen(filesrch.report);
write(fd[1],filesrch.report, length);
close(fd[1]); //close write
kill(parentPid,SIGUSR1);

On the parent's side, there is an interrupt function that overrides stdin:

void myfct(int y){
    dup2(fd[0],STDIN_FILENO); //Overwrite userinput
    overridemode=1;
}

Through this project, I established my foundations in concurrent programming, including how to pipe between a child and parent, how to use signals to override stdin, and how redirect the 'read' end of a pipe to stdout using dup and dup2. I have applied this knowledge to other parallelization projects, like matrix multiplication parallelization, which can be seen here.