I'm using pipe() to split up a file by index, send that index to child processes, have the child process calculate the sum of the numbers in its designated block of the file, and return its sum to the parent.
My children seem to execute in order, where I would like them to execute at the same time to make this process more efficient.
Here's the code I'm working with:
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/wait.h>
#include <sys/types.h>
#include <sys/select.h>
#include <time.h>
int main(int argc, char *argv[])
{
int numchild;
struct timeval stop, start;
int i, j, len, ret, fpos=0, val, count=0, total=0, alltotal=0;
pid_t pid;
int nums = 1000;
FILE * file;
printf("How many children to use: ");
scanf("%d", &numchild);
printf("\nWill use %d child process(es).\n", numchild);
gettimeofday(&start, NULL);
int fd[numchild][2]; //parent to child. one for each
int results[2]; //all children to parent
pipe(results);
fd_set result_fd;
FD_ZERO(&result_fd);
FD_SET(results[0], &result_fd);
struct timeval tm = {.tv_sec=0, .tv_usec=1};
// create all pipes
for (i=0; i<numchild; i++)
{
pipe(fd[i]);
}
for (i=0; i<numchild; i++)
{
if((pid = fork()) == 0) // child process
{
pid = getpid();
// read from parent
len = read(fd[i][0], &fpos, sizeof(fpos));
if (len > 0)
{
file = fopen("file1.dat", "r");
fseek (file, fpos, SEEK_SET);
count = 0;
total = 0;
printf("Child(%d): Recieved position: %d\n", pid, fpos);
// read from file starting at fpos
// add values read to a total value
while (count < (nums/numchild))
{
fscanf(file, "%i", &val);
total += val;
count++;
}
//write to parent
write(results[1], &total, sizeof(total));
printf("Child(%d): Sent %d to parent.\n", pid, total);
}
else
{
printf("Child(%d): Error with len\n", pid);
}
_exit(0);
}
// parent process
pid = getpid();
fpos = ((i*nums*5)/numchild); // 5 is the offset of the file values
// write to child process
printf("Parent(%d): Sending file position to child\n", pid);
write(fd[i][1], &fpos, sizeof(fpos));
// wait for child responce
ret = select(FD_SETSIZE+1, &result_fd, NULL, NULL, NULL); //&tm
if (FD_ISSET(results[0], &result_fd))
{
ret = read(results[0], &total, sizeof(total));
// output total
printf("Parent(%d): Recieved %d from child.\n", pid, total);
alltotal += total;
//printf("\tParent(%d): Total: %d\n", pid, alltotal);
}
}
wait(0);
gettimeofday(&stop, NULL);
printf("\tTime elapsed: %lu microseconds\n", stop.tv_usec - start.tv_usec);
printf("\tParent(%d): Total: %d\n", pid, alltotal);
}
Please let me know what I need to change to have the child processes run simultaneously (not wait to run at the exact same time, but run as soon as the parent gives them their index, instead of waiting for the previous child to complete).
From the comments above, I conclude that:
1. this is an assignment of some type
2. it requires using fork and pipe
If I were doing something like this for real (and it's not clear that it would be worth doing), I would probably be using threads queues and semaphores.
Given the constraints, I'll try to answer your question.
The problem is that you are have the parent code inside the for loop. So what is happening is that each time through the loop, the parent is spawning a child, then sending the offset information, then waiting for the result. So that forces the child to complete before the parent goes through the next iteration of the loop.
The answer is to have more than one loop. In the first loop, spawn all of the children. In the second loop, send the children their offsets. In the third loop, collect the results. In the fourth loop wait for the children to terminate. It would probably be faster if you sent the children their offsets in the first loop.
Another way to do this is to set the offset in a variable prior to doing each fork. That would obviate the need to use pipes for the input. I believe that you could also have each child just exit with the sum. Then the return exit status of the child could be the sum. The parent could just total the sums and you avoid using a return pipe. That would be a better way to do this -- though it wouldn't follow your apparent rules.
Related
Good afternoon.
I am currently working on a C program that takes one and only one parameter which designates the number of "child generation"s to be created (the own father counts as 1 already). "wait()" system calls are not to be used for this exercise (the version with "wait" calls happens to work exactly as expected).
For instance, the call $program 4 should generate a hierarchy like this:
Process A creates B
Process B creates C
Process C creates D
The printed messages are not important, as they are merely orientative for the task. With the following code (which happens to work exactly how I want with a "wait()" call) states that all the child processes derive from the same father, which I don't understand why it's happening.
#include <unistd.h>
#include <stdlib.h>
#include <stdio.h>
int main(int argc, char *argv[]) {
int counter; pid_t result; int i;
/*
We are going to create as many processes as indicated in argv[1] taking into account that the main father already counts as 1!
*/
if (argc > 2 || argc == 1) {puts("IMPOSSIBLE EXECUTION\n"); exit(-1);}
int lim = atoi(argv[1]);
//We eliminate the impossible cases
if (lim < 1) {puts("IMPOSSIBLE EXECUTION\n"); exit(-1);}
if (lim == 1) {puts("The father himself constitutes a process all by his own, therefore:\n");
printf("Process%d, I'm %d and my father: %d\n", counter, getpid(), getppid());
}
else {
for (i = 0; i < lim; i++) {
result = fork();
if (result < 0) {
printf("Call%d \n", counter); perror("Has failed!");
exit(-1);
}
else if (result) {
break; //Father process
}
else {
counter++; //Child processes increment the counter
printf("Process%d, I am %d and my father: %d\n", counter, getpid(), getppid());
}
}
}
The hierarchy generated by the code above is not the one I expected...
All help is greatly appreciated.
Thank you
With the following code (which happens to work exactly how I want with
a "wait()" call) states that all the child processes derive from the
same father, which I don't understand why it's happening.
I don't see that in my tests, nor do I have any reason to expect that it's actually the case for you. HOWEVER, it might appear to be the case for you if what you see is some or all of the child processes reporting process 1 as their parent. That would happen if their original parent terminates before the child's getppid() call is handled. Processes that are orphaned in that way inherit process 1 as their parent. If the parent wait()s for the child to terminate first then that cannot happen, but if instead the parent terminates very soon after forking the child then that result is entirely plausible.
Here's a variation on your loop that will report the original parent process ID in every case:
pid_t my_pid = getpid();
for (i = 0; i < lim; i++) {
result = fork();
if (result < 0) {
printf("Call%d \n", counter); perror("Has failed!");
exit(-1);
} else if (result) {
break; //Father process
} else {
pid_t ppid = my_pid; // inherited from the parent
my_pid = getpid();
counter++; //Child processes increment the counter
printf("Process%d, I am %d and my father: %d\n", counter, (int) my_pid, (int) ppid);
}
}
You are missing a crucial function call.
for (i = 0; i < lim; i++) {
fflush(stdout); // <============== here
result = fork();
Without it, your fork duplicates parent's stdout buffer into the child process. This is why you are seeing parent process output repeated several times --- its children and grandchildren inherit the output buffer.
Live demo (with fixed formatting for your reading convenience).
I'm splitting a file, sending through pipe(), having children find the sum of their designated section of the file, returning the calculated sum to the parent through pipe(), and having the parent calculate the sum of the child sums.
I've got a working program. My issues is that it hangs after receiving showing the proper final value.
I'm not sure what I'm doing to have the parent expect more information, but I'd bet it has something to do with my for() loop containing my child code.
Here's my code:
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
int main(int argc, char *argv[])
{
int numchild;
int i, j, len, fpos=0, val, count=0, total=0, alltotal=0;
pid_t pid;
int nums = 1000;
FILE * file;
printf("How many children to use: ");
scanf("%d", &numchild);
printf("\nWill use %d child process(es).\n", numchild);
int fd[2*numchild][2]; //parent+child pipe
// create all pipes
for (i=0; i<2*numchild; i++)
{
pipe(fd[i]);
}
for (i=0; i<numchild; i++)
{
if((pid = fork()) == 0) // child process
{
pid = getpid();
// read from parent
len = read(fd[i][0], &fpos, sizeof(fpos));
if (len > 0)
{
file = fopen("file1.dat", "r");
fseek (file, fpos, SEEK_SET);
count = 0;
total = 0;
printf("Child(%d): Recieved position: %d\n", pid, fpos);
// read from file starting at fpos
// add values read to a total value
while (count < (nums/numchild))
{
fscanf(file, "%i", &val);
total += val;
count++;
}
//write to parent
write(fd[i+numchild][1], &total, sizeof(total));
printf("Child(%d): Sent %d to parent.\n", pid, total);
}
else
{
printf("Child(%d): Error with len\n", pid);
}
_exit;
}
// parent process
pid = getpid();
fpos = ((i*nums*5)/numchild); // 5 is the offset of the file values
// write to child process
printf("Parent(%d): Sending file position to child\n", pid);
write(fd[i][1], &fpos, sizeof(fpos));
// wait for child responce
len = read(fd[i+numchild][0], &total, sizeof(total));
if (len > 0)
{
printf("Parent(%d): Recieved %d from child.\n", pid, total);
alltotal += total;
printf("Parent(%d): Total: %d\n", pid, alltotal);
}
else
{
printf("Parent(%d): Error with len\n", pid);
}
}
}
I can't ask more questions, but if this is on Linux or Unix like systems (perhaps all posix):
You must do a wait (man 2 wait) for each of your child processes in your main program or you will create zombie processes.
Not knowing what environment you are running in makes it impossible for me to test this to determine if that is the cause of your not exiting properly.
Also (this is more like a comment), each cycle through the loop you are forking one child, then feeding it data, then getting a response, then printing the total. Is that really what you want to do? You don't need to create a bunch of pipes if you are only running one child at a time.
My guess is that you want to have some actual concurrency. You can do this by having a loop that creates all of the children followed by another loop that feeds them data, followed by a third loop that looks at the results and sums them, followed by a fourth loop that waits for each to terminate (to avoid zombies).
I would consider using poll or select to read the returns -- after all, there is no guarantee that the children will finish in order.
I won't select this as an answer until other people let me know this is probably what fixed it, but I believe that my program was hanging because the main() function wasn't actually returning anything. When I went to put in time collection (save begin time, end time, calculate difference, output) it had to be done outside of my for loop, so it was the last statement right inside my main() method. This seems to have stopped the hanging.
Right after calling fork() in your if-statement, you need to call wait(). This makes causes your parent to wait for your child to exit, and then continue he execution when the child exits.
Say I fork N children. I want to create pipes between 1 and 2, 2 and 3, 4 and 5, ... and so on. So I need some way to figure out which child is which. The code below is what I currently have. I just need some way to tell that child number n, is child number n.
int fd[5][2];
int i;
for(i=0; i<5; i++)
{
pipe(fd[i]);
}
int pid = fork();
if(pid == 0)
{
}
The following code will create a pipe for each child, fork the process as many times as it is needed and send from the parent to each child an int value (the id we want to give to the child), finally the children will read the value and terminate.
Note: since you are forking, the i variable will contain the iteration number, if the iteration number is the child id, then you do not need to use pipe.
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
int main(int argc, char *argv[]) {
int count = 3;
int fd[count][2];
int pid[count];
// create pipe descriptors
for (int i = 0; i < count; i++) {
pipe(fd[i]);
// fork() returns 0 for child process, child-pid for parent process.
pid[i] = fork();
if (pid[i] != 0) {
// parent: writing only, so close read-descriptor.
close(fd[i][0]);
// send the childID on the write-descriptor.
write(fd[i][1], &i, sizeof(i));
printf("Parent(%d) send childID: %d\n", getpid(), i);
// close the write descriptor
close(fd[i][1]);
} else {
// child: reading only, so close the write-descriptor
close(fd[i][1]);
// now read the data (will block)
int id;
read(fd[i][0], &id, sizeof(id));
// in case the id is just the iterator value, we can use that instead of reading data from the pipe
printf("%d Child(%d) received childID: %d\n", i, getpid(), id);
// close the read-descriptor
close(fd[i][0]);
//TODO cleanup fd that are not needed
break;
}
}
return 0;
}
I am creating multiple processes and I need to create two unnamed pipes for each process.
For each child, one pipe will be used to get int value from parent; one for sending to int arrays to parent. Parent will do some things while getting new data from childs.
The base code:
#include <sys/types.h>
#include <sys/wait.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h> // for reaching unix operations
int main(int argc, char *argv[]){
pid_t main = getpid();
int N = 30;
int i;
pid_t* children = (pid_t*) malloc(sizeof(pid_t) * N);
for(i = 0; i < N; i++){
pid_t child = fork();
if ( child == 0){
pid_t me = getpid();
printf("I'm a child and my pid is: %d\n", me);
sleep(1);
// exit(4);
return me * 2;
} else if ( child < 0){
// printf("Could not create child\n");
} else {
children[i] = child;
// printf("I have created a child and its pid %d\n", child);
}
}
// The child never reaches here
for(i = 0; i < N; i++){
int status;
waitpid(children[i], &status, 0);
printf("Process %d exited with return code %d\n", children[i], WEXITSTATUS(status));
}
return 0;
}
I tried many things with no success and I'm lost. Can you help me to continue?
Any help is appreciated! Thank you.
Here's how to set up one pipe for each child process so that each child writes to the parent:
Since you need two file descriptors for each child, declare:
int fd[2 * N];
Initialize them appropriately:
for (int i = 0; i < N; i++) {
pipe(&fd[2*i]);
}
Inside the i-th child process, use:
write(fd[2*i + 1], write_buffer, SIZE)
to write to the parent, and in the parent use:
read(fd[2*i], read_buffer, SIZE)
to read from the i-th child.
To close the pipes:
Inside the i-th child, you can use
close(fd[2*i])
right away, seeing as you're only writing. After you're done writing call
close(fd[2*i + 1])
to close the write end of the pipe.
The situation is parallel in the parent: when reading from the i-th child you can
close(fd[2*i + 1])
right away, since you're not writing, and after you're done reading call
close(fd[2*i])
to close the read end of the pipe.
Since you need two pipes per child process, create two arrays - one containing pipes for the children writing to the parent, and one containing pipes for the parent writing to the children.
How can I display the number of processes created?
(without using a formula)
for (i=0; i<3; i++)
fork();
count = count + 1;
printf("%d",count);
There are a number of ways to do this, and a good technique is to have each child write one byte into a file descriptor which the original process can read. Note that, for the sake of brevity, the following code contains absolutely no error checking. Also, we report only the number of spawned processes (7) rather than counting the original to get a count of 8:
int main(void) {
int fd[2];
int depth = 0; /* keep track of number of generations from original */
int i;
pipe(fd); /* create a pipe which will be inherited by all children */
for(i=0; i<3; i++) {
if(fork() == 0) { /* fork returns 0 in the child */
write(fd[1], &i, 1); /* write one byte into the pipe */
depth += 1;
}
}
close(fd[1]); /* exercise for the reader to learn why this is needed */
if( depth == 0 ) { /* original process */
i=0;
while(read(fd[0],&depth,1) != 0)
i += 1;
printf( "%d total processes spawned", i);
}
return 0;
}
Printing the count value out just once is the easy part. Because you can get the process pid before the for loop. And then get the pid again after the for loop and only print if the pids match. For the counting part, it depends on whether your child processes exit or not. If they exit the solution is easier. The below code demonstrates one possible solution if the child processes exit (for brevity have not done full error checking). The idea is that each child process counts its own children. Parent waits for each child to complete and adds in its count. Haven't had time to fully test/debug the program so there may be some errors. But hopefully gives you the general idea.
#include <stdio.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <unistd.h>
int main(void)
{
pid_t before_pid, after_pid;
pid_t forked_pid;
int count;
int i;
int status;
before_pid = getpid();
count = 1; /* count self */
for (i = 0; i < 3; i++) {
forked_pid = fork();
if (forked_pid > 0) {
waitpid(forked_pid, &status, 0);
/* parent process - count child and descendents */
count += WEXITSTATUS(status);
} else {
/* Child process - init with self count */
count = 1;
}
}
after_pid = getpid();
if (after_pid == before_pid) {
printf("%d processes created\n", count);
}
return (count);
}