UNIX for Advanced & Expert Users

what is diffenrence between spawn and fork ?

"fork() system call spawns the processess" what is mean by this sentence.

When UNIX creates a new process, for example when you login:

1. the kernel has to take an existing process and then copy all of it into another part of memory.

2. Next it creates kernel entries to support the new process.

3. Finally it gives control of the cpu to user space in the new process

4. (optional) load an executable image into the process overwriting the old process memory. exec() does this, fork() does not do this.

Steps 1-3 is what fork() does, it is commonly called spawning a new process. Step 4 is performed by exec(). They are very different. So when you type

Code:

awk '{print $3}' somefile

You create a child process using 1,2, 3, then you exec() awk, step 4.

Where the problems come in:

1. all modern UNIX systems also support posix_spawn(), which does not copy all of memory. Its function is to do a "lightweight" version of fork().

2. Linux has something like posix_spawn() called clone().

3. Some other oddball systems also have a spawn() system call which acts more or less like fork()/exec().

An example of fork()/exec() is the C system() function.

So in effect you can say:
fork() spawns a new process.
posix_spawn() spawns a new process.
clone() spawns a new process.
system() spawns a new process.

So, the word spawn does not have to refer to a specific UNIX function, just what happens to make a new process.