Exit (system call)

How it works

Under Unix and Unix-like operating systems, a process is started when its parent process executes a fork system call. The parent process may then wait for the child process to terminate, or may continue execution (possibly forking off other child processes). When the child process terminates ("dies"), either normally by calling exit, or abnormally due to a fatal error or signal (e.g., SIGTERM, SIGINT, SIGKILL), an exit status is returned to the operating system and a SIGCHLD signal is sent to the parent process. The exit status can then be retrieved by the parent process via the wait system call.

Most operating systems allow the terminating process to provide a specific exit status to the system, which is made available to the parent process. Typically this is an integer value, although some operating systems (e.g., Plan 9 from Bell Labs) allow a character string to be returned. Systems returning an integer value commonly use a zero value to indicate successful execution and non-zero values to indicate error conditions. Other systems (e.g., OpenVMS) use even-numbered values for success and odd values for errors. Still other systems (e.g., IBM z/OS and its predecessors) use ranges of integer values to indicate success, warning, and error completion results.

Clean up

The exit operation typically performs clean-up operations within the process space before returning control back to the operating system. Some systems and programming languages allow user subroutines to be registered so that they are invoked at program termination before the process actually terminates for good. As the final step of termination, a primitive system exit call is invoked, informing the operating system that the process has terminated and allows it to reclaim the resources used by the process.

It is sometimes possible to bypass the usual cleanup; C99 offers the _exit() function which terminates the current process without any extra program clean-up. This may be used, for example, in a fork-exec routine when the exec call fails to replace the child process; calling atexit routines would erroneously release resources belonging to the parent.

Orphans and zombies

Some operating systems handle a child process whose parent process has terminated in a special manner. Such an orphan process becomes a child of a special root process, which then waits for the child process to terminate. Likewise, a similar strategy is used to deal with a zombie process, which is a child process that has terminated but whose exit status is ignored by its parent process. Such a process becomes the child of a special parent process, which retrieves the child's exit status and allows the operating system to complete the termination of the dead process. Dealing with these special cases keeps the system process table in a consistent state.

Examples

The following programs terminate and return a success exit status to the system.

COBOL:

Java:

PC DOS Batch file:

Perl:

PHP:

Python:

Unix shell:

Pascal:

PC DOS Assembly:

; For MASM/TASM .MODEL SMALL .STACK .CODE main PROC NEAR MOV AH, 4Ch ; Service 4Ch - Terminate with Error Code MOV AL, 0 ; Error code INT 21h ; Interrupt 21h - DOS General Interrupts main ENDP END main ; Starts at main

Some programmers may prepare everything for INT 21h at once:

MOV AX, 4C00h ; replace the 00 with your error code in HEX

Linux 32-bit x86 Assembly:

; For NASM MOV AL, 1 ; Function 1: exit() MOV EBX, 0 ; Return code INT 80h ; The only interrupt Linux uses! # For GAS .text .global _start _start: movl $1, %eax # System call number 1: exit() movl $0, %ebx # Exits with exit status 0 int $0x80 # Passes control to interrupt vector # invokes system call—in this case system call # number 1 with argument 0

Linux 64-bit x86 64 Assembly: for FASM

Windows

On Windows, a program can terminate itself by calling ExitProcess or RtlExitUserProcess function.