man2/sigreturn.2

.\" Copyright (C) 2008, 2014, Michael Kerrisk <mtk.manpages@gmail.com>
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.\" Created   Sat Aug 21 1995     Thomas K. Dyas <tdyas@eden.rutgers.edu>
.\" Modified Tue Oct 22 22:09:03 1996 by Eric S. Raymond <esr@thyrsus.com>
.\" 2008-06-26, mtk, added some more detail on the work done by sigreturn()
.\" 2014-12-05, mtk, rewrote all of the rest of the original page
.\"
.TH SIGRETURN 2 2017-09-15 "Linux" "Linux Programmer's Manual"
.SH NAME
sigreturn, rt_sigreturn \- return from signal handler and cleanup stack frame
.SH SYNOPSIS
.BI "int sigreturn(...);"
.SH DESCRIPTION
If the Linux kernel determines that an unblocked
signal is pending for a process, then,
at the next transition back to user mode in that process
(e.g., upon return from a system call or
when the process is rescheduled onto the CPU),
it creates a new frame on the user-space stack where it
saves various pieces of process context
(processor status word, registers, signal mask, and signal stack settings).
.\" See arch/x86/kernel/signal.c::__setup_frame() [in 3.17 source code]
.PP
The kernel also arranges that, during the transition back to user mode,
the signal handler is called, and that, upon return from the handler,
control passes to a piece of user-space code commonly called
the "signal trampoline".
The signal trampoline code in turn calls
.BR sigreturn ().
.PP
This
.BR sigreturn ()
call undoes everything that was
done\(emchanging the process's signal mask, switching signal stacks (see
.BR sigaltstack "(2))\(emin "
order to invoke the signal handler.
Using the information that was earlier saved on the user-space stack
.BR sigreturn ()
restores the process's signal mask, switches stacks,
and restores the process's context
(processor flags and registers,
including the stack pointer and instruction pointer),
so that the process resumes execution
at the point where it was interrupted by the signal.
.SH RETURN VALUE
.BR sigreturn ()
never returns.
.SH CONFORMING TO
Many UNIX-type systems have a
.BR sigreturn ()
system call or near equivalent.
However, this call is not specified in POSIX,
and details of its behavior vary across systems.
.SH NOTES
.BR sigreturn ()
exists only to allow the implementation of signal handlers.
It should
.B never
be called directly.
(Indeed, a simple
.BR sigreturn ()
.\" See sysdeps/unix/sysv/linux/sigreturn.c and
.\" signal/sigreturn.c in the glibc source
wrapper in the GNU C library simply returns -1, with
.I errno
set to
.BR ENOSYS .)
Details of the arguments (if any) passed to
.BR sigreturn ()
vary depending on the architecture.
(On some architectures, such as x86-64,
.BR sigreturn ()
takes no arguments, since all of the information that it requires
is available in the stack frame that was previously created by the
kernel on the user-space stack.)
.PP
Once upon a time, UNIX systems placed the signal trampoline code
onto the user stack.
Nowadays, pages of the user stack are protected so as to
disallow code execution.
Thus, on contemporary Linux systems, depending on the architecture,
the signal trampoline code lives either in the
.BR vdso (7)
or in the C library.
In the latter case,
.\" See, for example, sysdeps/unix/sysv/linux/i386/sigaction.c and
.\" sysdeps/unix/sysv/linux/x86_64/sigaction.c in the glibc (2.20) source.
the C library's
.BR sigaction (2)
wrapper function informs the kernel of the location of the trampoline code
by placing its address in the
.I sa_restorer
field of the
.I sigaction
structure,
and sets the
.BR SA_RESTORER
flag in the
.IR sa_flags
field.
.PP
The saved process context information is placed in a
.I ucontext_t
structure (see
.IR <sys/ucontext.h> ).
That structure is visible within the signal handler
as the third argument of a handler established via
.BR sigaction (2)
with the
.BR SA_SIGINFO
flag.
.PP
On some other UNIX systems,
the operation of the signal trampoline differs a little.
In particular, on some systems, upon transitioning back to user mode,
the kernel passes control to the trampoline (rather than the signal handler),
and the trampoline code calls the signal handler (and then calls
.BR sigreturn ()
once the handler returns).
.\"
.SS C library/kernel differences
The original Linux system call was named
.BR sigreturn ().
However, with the addition of real-time signals in Linux 2.2,
a new system call,
.BR rt_sigreturn ()
was added to support an enlarged
.IR sigset_t
type.
The GNU C library
hides these details from us, transparently employing
.BR rt_sigreturn ()
when the kernel provides it.
.\"
.SH SEE ALSO
.BR kill (2),
.BR restart_syscall (2),
.BR sigaltstack (2),
.BR signal (2),
.BR getcontext (3),
.BR signal (7),
.BR vdso (7)