What is the difference between a system call and a function call? Is fopen() a system call or a function call?
The question has excellent answers already, but I think I can add something (one segment from ostepthat isn't already in other answers
Sometimes system call and function call have the same signature, for example, open()
:
open()
-system call--- ~/Documents » man open(2)
OPEN(2) Linux Programmer's Manual OPEN(2)
int open(const char *pathname, int flags);
int open(const char *pathname, int flags, mode_t mode);
...
open()
-function call$ man open(3)
--- ~/Documents »
OPEN(3P) POSIX Programmer's Manual OPEN(3P)
...
int open(const char *path, int oflag, ...);
...
You may wonder why a call to a system call, such as
open()
orread()
, looks exactly like a typical procedure call in C; that is, if it looks just like a procedure call, how does the system know it’s a system call, and do all the right stuff? The simple reason: it is a procedure call, but hidden inside that procedure call is the famous trap instruction. More specifically, when you callopen()
(for example), you are executing a procedure call into the C library. Therein, whether foropen()
or any of the other system calls provided, the library uses an agreed-upon calling convention with the kernel to put the arguments to open in well-known locations(e.g., on the stack, or in specific registers), puts the system-call number into a well-known location as well (again, onto the stack or a register), and then executes the aforementioned trap instruction. The code in the library after the trap unpacks return values and returns control to the program that issued the system call. Thus, the parts of the C library that make system calls are hand-coded in assembly, as they need to carefully follow convention in order to process arguments and return values correctly, as well as execute the hardware-specific trap instruction. And now you know why you personally don’t have to write assembly code to trap into an OS; somebody has already written that assembly for you.
A point of view to add to this discussion is that a function call generally in the most optimistic case has overhead of a a few 8-bit instructions (4-10 on average)in x86.
A system call has the following properties.
For these three primitive reasons (there are probably more), one should reduce the amount of system calls where possible -- e.g., networked system software keeps socket handles (and other application specific internal data structures used by a connection) around to assign to new connection, why bother the kernel ?
Remember that software is built like a upside down pyramid. System calls are at the base.
System call is executed at kernet level and not in user spce because it requires some prievilege to access the hardware.
Therfore when programming in user space and making some ordinary function call like fopen
in C language the libc generally wrap this function to specific code code where an interrupt is generated to switch from user space to kernel space , then in kernel space the required system call to perform the functionality of the function call at hardware level will be executed in kernel space .
Actually, the system call is not related to function call. The only common of these two mechanism is that they both provides services to the caller.
From view of thread execution to see system call:
A system call is function for application mode program to request services provided by underline OS. The system call will bring the running thread from user mode into kernel mode, execute the system call handler function, then return back to user mode.
Syscall Parameters:
The parameter of a system call is (syscall number, params...). The meaning and format of params depends on syscall number.
From view of syscall library provided to userland program:
The user mode program usually calls glibc's library to call system call. For example, the open() function in glibc:
fopen is a function call.
A system call interacts with the underlying OS, which manages resources. Its orders of magnitud more expensive than a function call, because many steps have to be taken to preserve the state of the process that made the syscall.
On *nix systems, fopen wraps open, which makes the system call (open is the C - wrapper for the syscall). The same happens with fread /read, fwrite / write , etc..
Here there's a nice description of the tasks executed by a unix syscall.
System call actually calls out to an API executed by the kernel space. With all the associated costs this assumes (see Wiki, or this link for details)
A function call is a call to a piece of code in user space.
However, please note that a function call MIGHT be to a function which in the process of its execution does system calls - "fopen" being one of such examples. So while the call to fopen itself is a call to a function, doesn't mean that the system call will not happen to handle the actual IO.