can someone help me about how to create multiple child processes which have the same parent in order to do "some" part of particular job?
for example, an external sorting algorithm which is applied with child processes; each child process sorts a part of data and finally the parent merges them..
EDIT: Maybe I should mention the forking multiple child processes with loop..
Here is how to fork 10 children and wait for them to finish:
pid_t pids[10];
int i;
int n = 10;
/* Start children. */
for (i = 0; i < n; ++i) {
if ((pids[i] = fork()) < 0) {
perror("fork");
abort();
} else if (pids[i] == 0) {
DoWorkInChild();
exit(0);
}
}
/* Wait for children to exit. */
int status;
pid_t pid;
while (n > 0) {
pid = wait(&status);
printf("Child with PID %ld exited with status 0x%x.\n", (long)pid, status);
--n; // TODO(pts): Remove pid from the pids array.
}
If you want to launch several forks, you should do it recursively. This is because you must call fork from the parent process. Otherwise, if you launch a second fork, you will duplicate both parent and first child process. Here's an example:
void forker(int nprocesses)
{
pid_t pid;
if(nprocesses > 0)
{
if ((pid = fork()) < 0)
{
perror("fork");
}
else if (pid == 0)
{
//Child stuff here
printf("Child %d end\n", nprocesses);
}
else if(pid > 0)
{
//parent
forker(nprocesses - 1);
}
}
}
I think it would be worth pointing out why threads are more appropriate here:
As you are trying to do a "part" of the job in parallel i assume that your program needs to know about the result of the computation. fork()s of a process don't share more then the initial information after fork(). Every change in one process is unknow to the other and you would need to pass the information as a message (e.g. through a pipe, see "man pipe").
Threads in a process share the same adress space and therefor are able to manipulate data and have them visible toeach other "immediatly". Also adding the benefits of being more lightweight, I'd go with pthreads().
After all: You will learn all you need to know about fork() if you use pthreads anyway.
You can do this with fork. A given parent can fork as may times as it wants. However, I agree with AviD pthreads may be more appropriate.
pid_t firstChild, secondChild;
firstChild = fork();
if(firstChild > 0)
{
// In parent
secondChild = fork();
if(secondChild > 0)
{
// In parent
}
else if(secondChild < 0)
{
// Error
}
else
{
// In secondChild
}
}
else if(firstChild < 0 )
{
// Error
}
else
{
// In firstChild
}
Related
I want to create a program in C, where I use fork() to create multiple children, then wait for all of them to finish and do parent code (only once).
I tried using for loop and two forks but I there is a problem: either parent code isn't running at the end or children are not running parallel.
//Number of processes I want to create
int processes = 6;
pid_t *main_fork = fork();
if(main_fork ==0){
for(int i=0;i<processes;i++){
pid_t *child_fork = fork();
if(child_fork ==0){
// child code
exit(0);
}
else if(child_fork >0){
//And here is the problem, with the wait: children don't
//run parallel and if I delete it, the main parent code doesn't run
wait(NULL);
}else{
// Child fork failed
printf("fork() failed!\n");
return 1;
}
}
}else if(main_fork >0){
wait(NULL);
//Main parent code - here I want to do something only once after all
//children are done
}else{
// Main fork failed
printf("fork() failed!\n");
return 1;
}
If somebody could could fix my code, or write a better solution to this problem I would be so grateful!
If you want all the children to run in parallel, you have to do the wait after all the children has been started. Otherwise you start a child, wait for it to finish, start a new one, wait for that to finish, start a third one, wait for the third one to finish, and so on...
So what you typically want to do is to start all the children and put all the pid_t in an array, and when you are finished you may call wait() for each pid_t
This is the simple, and good enough solution for your case.
Here is a sample code that you can fit to your problem:
pid_t children[processes];
for(int i=0; i<processes; i++)
{
pid_t child = fork();
if(child == 0)
{
// child code
....
// We call _exit() rather than exit() since we don't want to clean up
// data structures inherited from parent
_exit(0);
}
else if (child == -1)
{
// Child fork failed
fprintf(stderr, "myprog: fork failed, %s", strerror(errno));
// Do real cleanup on failure is to complicated for this example, so we
// just exit
exit(EXIT_FAILURE);
}
children[i] = child;
}
// Do something if you want to do something before you expect the children to exit
....
for(int i=0; i<processes; i++)
{
pid_t child = children[i];
int status;
waitpid(child, &status, );
// Do something with status
}
Naturally this is not a complete example that fits any situation. Sometimes you have to tell the children when they should exit. Other times the children are not started/stopped in one go, and you have to play with asynchronous events, and so on...
I'm struggling to understand this concept. Let's say I want to run 3 concurrent processes (threads are not an option, this is an assignment).
I would do:
int main(){
fork();
if (getPid() == -1) { //this is the parent
fork(); //make a third process
} else if (getPid() == 0) { //child process
//do child things
}
So from what I've learned the parent pid is -1. And there's two children, both with PID 0?
So then the "parent" can spawn as many children as possible, correct? They will all do child things.
What if I want to do 3 different things? How do I track the PID's so that I have 3 unique
ones?
as per the comments - is this how it's done?
pid_t child2Pid;
pid_t child1Pid = fork();
switch /*if getPid is wrong what do I put here?*/ {
case -1: //parent
child2Pid = fork(); //create another child
case child1Pid :
//do what child1 would do
case child2Pid :
//do what child2 would do
pid_t child1, child2, child3;
if ((child1 = fork()) == 0)
{
// first child stuff goes here
_exit(0);
}
if ((child2 = fork()) == 0)
{
// second child stuff goes here
_exit(0);
}
if ((child3 = fork()) == 0)
{
// third child stuff goes here
_exit(0);
}
// we are in the parent, we have the PIDs of our three
// children in child1, child2, and child3
The whole idea is that you enter fork() once, but leave it twice - once in the parent, once in the child. In the parent, fork() returns the child's PID, in the child, fork() returns 0. -1 means error and there is no child.
So when you call fork(), look at the return value. Like this:
if(fork() > 0) //We're in the parent, and a child was spawned
fork(); //Spawn another.
This will create three processes. Error handling omitted.
**
int main()
{
pid_t pid;
int i;
for(i = 0;i < 3;i++)
{
pid = fork();
}
if(pid == 0)
{
printf("child\n");
}
else if(pid > 0)
{
printf("parent\n");
}
return 0;
}
you can use pid_t = fork() to record the value,and select what you do by the value.**
I am trying to make a processor farm in C. I start with opening message queues, and afterwards try to make worker processes: (note that NROF_WORKERS is 5)
static void
makechildren (void) {
// Only the parent should fork. Any children created will become workers.
pid_t processID;
pid_t farmerPID = getpid(); // To identify who the farmer is
// Loop creating processes, indexed by NROF_WORKERS
int i = 0;
while (i < NROF_WORKERS){
if (getpid() == farmerPID){
i++;
printf ("Parent is creating a child!%d\n", getpid());
processID = fork();
}
}
if (processID < 0){
perror("fork() failed");
exit(1);
}
else {
// If parent, start farming
if (processID == farmerPID) {
printf("Parent reporting in!%d\n");
}
// If child, become a worker
if (processID == 0) {
printf("Child reporting in!%d\n", getpid());
join();
}
}
}
As you can see, I want the parent to report any time a child is created, and afterwards I want the parent and all children to report. However, this is all I get:
Parent is creating a child!11909
Parent is creating a child!11909
Parent is creating a child!11909
Parent is creating a child!11909
Parent is creating a child!11909
Child reporting in!11914
Now, I do notice the difference in 11909 and 11914 is 5. So my question: are the other processes created? If so, how come they don't report? And if not, what am I doing wrong? Also, the parent is not reporting at all, how is this caused?
All of the children are created, but will loop forever in the while loop, as i is incremented only for the parent:
int i = 0;
while (i < NROF_WORKERS){
if (getpid() == farmerPID){
i++; // <---- This is happening for the parent process only.
printf ("Parent is creating a child!%d\n", getpid());
processID = fork();
}
}
The only child to terminate is the last one, for which the i is equal to NROF_WORKERS.
Also parent is "not reporting" since the processID you are checking to be equal to the parent PID is never equal to it, as it is equal to the latest fork result, i.e. the latest created child PID:
.........
processID = fork();
.........
.........
if (processID == farmerPID) {
printf("Parent reporting in!%d\n");
}
You always print the farmerPid! But as the message is printed 5 times, you effectively created 5 processes:
while (i < NROF_WORKERS){
if (getpid() == farmerPID){
i++;
printf ("Parent is creating a child!%d\n", getpid());
processID = fork();
}
}
If you want to print the children pids then your code must makes a difference in between parent and child, as in:
while (i < NROF_WORKERS){
if (getpid() == farmerPID){
i++;
printf ("Parent is creating a child!\n");
processID = fork();
if (processID==0) { // child
printf("I am the child %d\n",getpid());
} else { // parent
printf("Parent just created child %d\n",processID);
}
}
}
I have the following code in which I'm trying to create sub processes by forking. I want that exactly 3 sub processes are made. However, when I run the code I seem to be getting more, probably because of the children processes forking grandchildren. What am I missing here, how can I prevent this.
Code:
for(j = 0; j < 3 ; j++){
if((pid = fork()) == 0){ // child process
dosomething();
exit(0); // terminate child process
}
else if((pid = fork()) > 0){
printf("I'm in parent of the client spawn loop\n");
// exit(0);
}
}
Output:
I'm in parent of the client spawn loop
I'm in parent of the client spawn loop
I'm in parent of the client spawn loop
I'm in parent of the client spawn loop
I'm in parent of the client spawn loop
I'm in parent of the client spawn loop
I'm in parent of the client spawn loop
Don't do the second fork call as it will create a new child. The first is enough:
for (j = 0; j < 3; ++j)
{
pid_t pid = fork();
if (pid == 0)
{
printf("In child (j = %d)\n", j);
exit(0);
}
else if (pid > 0)
{
printf("In parent (j = %d)\n", j);
}
}
Will print "In child" three times, with j equal to 0, 1 and 2. The same for the parent printing.
In your real code you should check for errors though.
Don't call fork() more than once in the loop.
The parent shouldn't call fork() again, that will create yet another child and introduce another child-parent split point.
You should have, in the loop:
const int pid = fork();
if(pid == 0)
{
doSomething();
exit();
}
can someone help me about how to create multiple child processes which have the same parent in order to do "some" part of particular job?
for example, an external sorting algorithm which is applied with child processes; each child process sorts a part of data and finally the parent merges them..
EDIT: Maybe I should mention the forking multiple child processes with loop..
Here is how to fork 10 children and wait for them to finish:
pid_t pids[10];
int i;
int n = 10;
/* Start children. */
for (i = 0; i < n; ++i) {
if ((pids[i] = fork()) < 0) {
perror("fork");
abort();
} else if (pids[i] == 0) {
DoWorkInChild();
exit(0);
}
}
/* Wait for children to exit. */
int status;
pid_t pid;
while (n > 0) {
pid = wait(&status);
printf("Child with PID %ld exited with status 0x%x.\n", (long)pid, status);
--n; // TODO(pts): Remove pid from the pids array.
}
If you want to launch several forks, you should do it recursively. This is because you must call fork from the parent process. Otherwise, if you launch a second fork, you will duplicate both parent and first child process. Here's an example:
void forker(int nprocesses)
{
pid_t pid;
if(nprocesses > 0)
{
if ((pid = fork()) < 0)
{
perror("fork");
}
else if (pid == 0)
{
//Child stuff here
printf("Child %d end\n", nprocesses);
}
else if(pid > 0)
{
//parent
forker(nprocesses - 1);
}
}
}
I think it would be worth pointing out why threads are more appropriate here:
As you are trying to do a "part" of the job in parallel i assume that your program needs to know about the result of the computation. fork()s of a process don't share more then the initial information after fork(). Every change in one process is unknow to the other and you would need to pass the information as a message (e.g. through a pipe, see "man pipe").
Threads in a process share the same adress space and therefor are able to manipulate data and have them visible toeach other "immediatly". Also adding the benefits of being more lightweight, I'd go with pthreads().
After all: You will learn all you need to know about fork() if you use pthreads anyway.
You can do this with fork. A given parent can fork as may times as it wants. However, I agree with AviD pthreads may be more appropriate.
pid_t firstChild, secondChild;
firstChild = fork();
if(firstChild > 0)
{
// In parent
secondChild = fork();
if(secondChild > 0)
{
// In parent
}
else if(secondChild < 0)
{
// Error
}
else
{
// In secondChild
}
}
else if(firstChild < 0 )
{
// Error
}
else
{
// In firstChild
}