signal

Sets interrupt signal handling.

Syntax

void __cdecl *signal(int sig, int (*func)(int, int));

Parameters

sig
Signal value.

func
The second parameter is a pointer to the function to be executed. The first parameter is a signal value, and the second parameter is a subcode that can be used when the first parameter is SIGFPE.

Return value

signal returns the previous value of func that's associated with the given signal. For example, if the previous value of func was SIG_IGN, the return value is also SIG_IGN. A return value of SIG_ERR indicates an error; in that case, errno is set to EINVAL.

For more information about return codes, see errno, _doserrno, _sys_errlist, and _sys_nerr.

Remarks

The signal function enables a process to choose one of several ways to handle an interrupt signal from the operating system. The sig argument is the interrupt to which signal responds; it must be one of the following manifest constants, which are defined in SIGNAL.H.

If sig isn't one of the above values, the invalid parameter handler is invoked, as defined in Parameter validation . If execution is allowed to continue, this function sets errno to EINVAL and returns SIG_ERR.

By default, signal terminates the calling program with exit code 3, regardless of the value of sig.

The func argument is an address to a signal handler that you write, or to one of the predefined signal action constants SIG_DFL or SIG_IGN, which are also defined in SIGNAL.H. If func is a function, it's installed as the signal handler for the given signal. The signal handler's prototype requires one formal argument, sig, of type int. The operating system provides the actual argument through sig when an interrupt occurs; the argument is the signal that generated the interrupt. Therefore, you can use the six manifest constants (listed in the preceding table) in your signal handler to determine which interrupt occurred and take appropriate action. For example, you can call signal twice to assign the same handler to two different signals, and then test the sig argument in the handler to take different actions based on the signal received.

If you're testing for floating-point exceptions (SIGFPE), func points to a function that takes an optional second argument that is one of several manifest constants, defined in FLOAT.H, of the form FPE_xxx. When a SIGFPE signal occurs, you can test the value of the second argument to determine the kind of floating-point exception, and then take appropriate action. This argument and its possible values are Microsoft extensions.

For floating-point exceptions, the value of func isn't reset when the signal is received. To recover from floating-point exceptions, use try/except clauses to surround the floating point operations. It's also possible to recover by using setjmp with longjmp. In either case, the calling process resumes execution and leaves the floating-point state of the process undefined.

If the signal handler returns, the calling process resumes execution immediately following the point at which it received the interrupt signal, regardless of the kind of signal or operating mode.

Before the specified function is executed, the value of func is set to SIG_DFL. The next interrupt signal is treated as described for SIG_DFL, unless an intervening call to signal specifies otherwise. You can use this feature to reset signals in the called function.

Because signal-handler routines are often called asynchronously when an interrupt occurs, your signal-handler function may get control when a run-time operation is incomplete and in an unknown state. The following list summarizes the restrictions that determine which functions you can use in your signal-handler routine.

• Don't issue low-level or STDIO.H I/O routines (for example, printf or fread).

• Don't call heap routines or any routine that uses the heap routines (for example, malloc, _strdup, or _putenv). For more information, see malloc.

• Don't use any function that generates a system call (for example, _getcwd or time).

• Don't use longjmp unless the interrupt is caused by a floating-point exception (that is, sig is SIGFPE). In this case, first reinitialize the floating-point package by using a call to _fpreset.

• Don't use any overlay routines.

A program must contain floating-point code if it's to trap the SIGFPE exception by using the function. If your program doesn't have floating-point code and requires the run-time library's signal-handling code, just declare a volatile double and initialize it to zero:

volatile double d = 0.0f;

The SIGILL and SIGTERM signals aren't generated under Windows. They're included for ANSI compatibility. Therefore, you can set signal handlers for these signals by using signal, and you can also explicitly generate these signals by calling raise.

Signal settings aren't preserved in spawned processes that are created by calls to _exec or _spawn functions. The signal settings are reset to the default values in the new process.

Requirements

For more compatibility information, see Compatibility.

Example

The following example shows how to use signal to add some custom behavior to the SIGABRT signal. For more information about abort behavior, see _set_abort_behavior.

// crt_signal.c
// compile with: /EHsc /W4
// Use signal to attach a signal handler to the abort routine
#include <stdlib.h>
#include <signal.h>

void SignalHandler(int signal)
{
    if (signal == SIGABRT) {
        // abort signal handler code
    } else {
        // ...
    }
}

int main()
{
    typedef void (*SignalHandlerPointer)(int);

    SignalHandlerPointer previousHandler;
    previousHandler = signal(SIGABRT, SignalHandler);

    abort();
}

The output depends on the version of the runtime used, whether the app is a console or Windows app, and on Windows registry settings. For a console app, something like the following message may be sent to stderr:

Debug Error!

Program: c:\Projects\crt_signal\Debug\crt_signal.exe

R6010

- abort() has been called

See also

Process and environment control
abort
_exec, _wexec functions
exit, _Exit, _exit
_fpreset
_spawn, _wspawn functions