The system defines a set of signals that may be delivered to a process. Signal delivery resembles the occurrence of a hardware interrupt: the signal is blocked from further occurrence, the current process context is saved, and a new one is built. A process may specify a handler to which a signal is delivered, or specify that a signal is to be blocked or ignored. A process may also specify that a default action is to be taken by the system when a signal occurs. Normally, signal handlers execute on the current stack of the process. This may be changed, on a per-handler basis, so that signals are taken on a special signal stack.

All signals have the same priority. Signal routines execute with the signal that caused their invocation to be blocked, but other signals may yet occur. A global signal mask defines the set of signals currently blocked from delivery to a process. The signal mask for a process is initialized from that of its parent (normally 0). It may be changed with a sigblock() or sigsetmask() call, or when a signal is delivered to the process.

A process may also specify a set of flags for a signal that affect the delivery of that signal.

When a signal condition arises for a process, the signal is added to a set of signals pending for the process. If the signal is not currently blocked by the process then it is delivered to the process. When a signal is delivered, the current state of the process is saved, a new signal mask is calculated (as described below), and the signal handler is invoked. The call to the handler is arranged so that if the signal handling routine returns normally the process will resume execution in the context from before the signal's delivery. If the process wishes to resume in a different context, then it must arrange to restore the previous context itself.

When a signal is delivered to a process a new signal mask is installed for the duration of the process' signal handler (or until a sigblock() or sigsetmask() call is made). This mask is formed by taking the current signal mask, adding the signal to be delivered, and ORing in the signal mask associated with the handler to be invoked.

The action to be taken when the signal is delivered is specified by a sigvec() structure, which includes the following members:

void (*sv_handler)(); /* signal handler */ int sv_mask; /* signal mask to apply */ int sv_flags; /* see signal options */ #define SV_ONSTACK /* take signal on signal stack */ #define SV_INTERRUPT /* do not restart system on signal return */ #define SV_RESETHAND /* reset handler to SIG_DFL when signal taken*/

If the SV_ONSTACK bit is set in the flags for that signal, the system will deliver the signal to the process on the signal stack specified with sigstack(3UCB) rather than delivering the signal on the current stack.

If nvec is not a NULL pointer, sigvec() assigns the handler specified by sv_handler(), the mask specified by sv_mask(), and the flags specified by sv_flags() to the specified signal. If nvec is a NULL pointer, sigvec() does not change the handler, mask, or flags for the specified signal.

The mask specified in nvec is not allowed to block SIGKILL, SIGSTOP, or SIGCONT. The system enforces this restriction silently.

If ovec is not a NULL pointer, the handler, mask, and flags in effect for the signal before the call to sigvec() are returned to the user. A call to sigvec() with nvec a NULL pointer and ovec not a NULL pointer can be used to determine the handling information currently in effect for a signal without changing that information.

The following is a list of all signals with names as in the include file <signal.h>:

SIGHUP

hangup

SIGINT

interrupt

SIGQUIT*

quit

SIGILL*

illegal instruction

SIGTRAP*

trace trap

SIGABRT*

abort (generated by abort(3C) routine)

SIGEMT*

emulator trap

SIGFPE*

arithmetic exception

SIGKILL

kill (cannot be caught, blocked, or ignored)

SIGBUS*

bus error

SIGSEGV*

segmentation violation

SIGSYS*

bad argument to function

SIGPIPE

write on a pipe or other socket with no one to read it

SIGALRM

alarm clock

SIGTERM

software termination signal

SIGURG*

urgent condition present on socket

SIGSTOP**

stop (cannot be caught, blocked, or ignored)

SIGTSTP**

stop signal generated from keyboard

SIGCONT*

continue after stop (cannot be blocked)

SIGCHLD*

child status has changed

SIGTTIN**

background read attempted from control terminal

SIGTTOU**

background write attempted to control terminal

SIGIO*

I/O is possible on a descriptor (see fcntl(2))

SIGXCPU

cpu time limit exceeded (see getrlimit(2))

SIGXFSZ

file size limit exceeded (see getrlimit(2))

SIGVTALRM

virtual time alarm; see setitimer() on getitimer(2)

SIGPROF

profiling timer alarm; see setitimer() on getitimer(2)

SIGWINCH*

window changed (see termio(7I))

SIGLOST

resource lost (see lockd(1M))

SIGUSR1

user-defined signal 1

SIGUSR2

user-defined signal 2

The starred signals in the list above cause a core image if not caught or ignored.

Once a signal handler is installed, it remains installed until another sigvec() call is made, or an execve(2) is performed, unless the SV_RESETHAND bit is set in the flags for that signal. In that case, the value of the handler for the caught signal will be set to SIG_DFL before entering the signal-catching function, unless the signal is SIGILL, SIGPWR, or SIGTRAP. Also, if this bit is set, the bit for that signal in the signal mask will not be set; unless the signal mask associated with that signal blocks that signal, further occurrences of that signal will not be blocked. The SV_RESETHAND flag is not available in 4.2BSD, hence it should not be used if backward compatibility is needed.

The default action for a signal may be reinstated by setting the signal's handler to SIG_DFL; this default is termination except for signals marked with * or **. Signals marked with * are discarded if the action is SIG_DFL; signals marked with ** cause the process to stop. If the process is terminated, a "core image" will be made in the current working directory of the receiving process if the signal is one for which an asterisk appears in the above list (see core(4)).

If the handler for that signal is SIG_IGN, the signal is subsequently ignored, and pending instances of the signal are discarded.

If a caught signal occurs during certain functions, the call is normally restarted. The call can be forced to terminate prematurely with an EINTR error return by setting the SV_INTERRUPT bit in the flags for that signal. The SV_INTERRUPT flag is not available in 4.2BSD, hence it should not be used if backward compatibility is needed. The affected functions are read(2) or write(2) on a slow device (such as a terminal or pipe or other socket, but not a file) and during a wait(2).

After a fork(2) or vfork(2) the child inherits all signals, the signal mask, the signal stack, and the restart/interrupt and reset-signal-handler flags.

The execve(2) call resets all caught signals to default action and resets all signals to be caught on the user stack. Ignored signals remain ignored; the signal mask remains the same; signals that interrupt functions continue to do so.

The accuracy of addr is machine dependent. For example, certain machines may supply an address that is on the same page as the address that caused the fault. If an appropriate addr cannot be computed it will be set to SIG_NOADDR.

A 0 value indicates that the call succeeded. A -1 return value indicates that an error occurred and errno is set to indicate the reason.

sigvec() will fail and no new signal handler will be installed if one of the following occurs:

EFAULT

Either nvec or ovec is not a NULL pointer and points to memory that is not a valid part of the process address space.

EINVAL

sig is not a valid signal number, or, SIGKILL, or SIGSTOP.

intro(2), exec(2), fcntl(2), fork(2), getitimer(2), getrlimit(2), ioctl(2), kill(2), ptrace(2), read(2), umask(2), vfork(2), wait(2), write(2), setjmp(3C) sigblock(3UCB), sigstack(3UCB), signal(3UCB), wait(3UCB), signal(3C), core(4), streamio(7I), termio(7I)

Use of these interfaces should be restricted to only applications written on BSD platforms. Use of these interfaces with any of the system libraries or in multi-thread applications is unsupported.

SIGPOLL is a synonym for SIGIO. A SIGIO will be issued when a file descriptor corresponding to a STREAMS (see intro(2)) file has a "selectable" event pending. Unless that descriptor has been put into asynchronous mode (see fcntl(2)), a process may specifically request that this signal be sent using the I_SETSIG ioctl(2) call (see streamio(7I)). Otherwise, the process will never receive SIGPOLLs0.

The handler routine can be declared:

void handler(int sig, int code, struct sigcontext *scp, char *addr);

Here sig is the signal number; code is a parameter of certain signals that provides additional detail; scp is a pointer to the sigcontext structure (defined in signal.h), used to restore the context from before the signal; and addr is additional address information.

The signals SIGKILL, SIGSTOP, and SIGCONT cannot be ignored.