Permalink
Switch branches/tags
Find file
Fetching contributors…
Cannot retrieve contributors at this time
550 lines (478 sloc) 14.2 KB
/*
* <signal.h> wrapper functions.
*
* Authors:
* Jonathan Pryor (jonpryor@vt.edu)
* Jonathan Pryor (jpryor@novell.com)
* Tim Jenks (tim.jenks@realtimeworlds.com)
*
* Copyright (C) 2004-2005 Jonathan Pryor
* Copyright (C) 2008 Novell, Inc.
*/
#include <signal.h>
#include "map.h"
#include "mph.h"
#ifndef HOST_WIN32
#include <sys/time.h>
#include <sys/types.h>
#if defined(HAVE_POLL_H)
#include <poll.h>
#elif defined(HAVE_SYS_POLL_H)
#include <sys/poll.h>
#endif
#include <unistd.h>
#include <stdlib.h>
#include <string.h>
#include <pthread.h>
#include <mono/utils/atomic.h>
#include <mono/metadata/appdomain.h>
#endif
G_BEGIN_DECLS
typedef void (*mph_sighandler_t)(int);
typedef struct Mono_Unix_UnixSignal_SignalInfo signal_info;
#ifndef HOST_WIN32
static int count_handlers (int signum);
#endif
void*
Mono_Posix_Stdlib_SIG_DFL (void)
{
return SIG_DFL;
}
void*
Mono_Posix_Stdlib_SIG_ERR (void)
{
return SIG_ERR;
}
void*
Mono_Posix_Stdlib_SIG_IGN (void)
{
return SIG_IGN;
}
void
Mono_Posix_Stdlib_InvokeSignalHandler (int signum, void *handler)
{
mph_sighandler_t _h = (mph_sighandler_t) handler;
_h (signum);
}
int Mono_Posix_SIGRTMIN (void)
{
#ifdef SIGRTMIN
return SIGRTMIN;
#else /* def SIGRTMIN */
return -1;
#endif /* ndef SIGRTMIN */
}
int Mono_Posix_SIGRTMAX (void)
{
#ifdef SIGRTMAX
return SIGRTMAX;
#else /* def SIGRTMAX */
return -1;
#endif /* ndef SIGRTMAX */
}
int Mono_Posix_FromRealTimeSignum (int offset, int *r)
{
if (NULL == r) {
errno = EINVAL;
return -1;
}
*r = 0;
#if defined (SIGRTMIN) && defined (SIGRTMAX)
if ((offset < 0) || (SIGRTMIN > SIGRTMAX - offset)) {
errno = EINVAL;
return -1;
}
*r = SIGRTMIN+offset;
return 0;
#else /* defined (SIGRTMIN) && defined (SIGRTMAX) */
# ifdef ENOSYS
errno = ENOSYS;
# endif /* ENOSYS */
return -1;
#endif /* defined (SIGRTMIN) && defined (SIGRTMAX) */
}
#ifndef HOST_WIN32
// Atomicity rules: Fields of signal_info read or written by the signal handler
// (see UnixSignal.cs) should be read and written using atomic functions.
// (For simplicity, we're protecting some things we don't strictly need to.)
// Because we are in MonoPosixHelper, we are banned from linking mono.
// We can still use atomic.h because that's all inline functions--
// unless WAPI_NO_ATOMIC_ASM is defined, in which case atomic.h calls linked functions.
#ifndef WAPI_NO_ATOMIC_ASM
#define mph_int_get(p) InterlockedExchangeAdd ((p), 0)
#define mph_int_inc(p) InterlockedIncrement ((p))
#define mph_int_dec_test(p) (InterlockedDecrement ((p)) == 0)
#define mph_int_set(p,n) InterlockedExchange ((p), (n))
// Pointer, original, new
#define mph_int_test_and_set(p,o,n) (o == InterlockedCompareExchange ((p), (n), (o)))
#elif GLIB_CHECK_VERSION(2,4,0)
#define mph_int_get(p) g_atomic_int_get ((p))
#define mph_int_inc(p) do {g_atomic_int_inc ((p));} while (0)
#define mph_int_dec_test(p) g_atomic_int_dec_and_test ((p))
#define mph_int_set(p,n) g_atomic_int_set ((p),(n))
#define mph_int_test_and_set(p,o,n) g_atomic_int_compare_and_exchange ((p), (o), (n))
#else
#error "GLIB 2.4 required because building without ASM atomics"
#endif
#if HAVE_PSIGNAL
int
Mono_Posix_Syscall_psignal (int sig, const char* s)
{
errno = 0;
psignal (sig, s);
return errno == 0 ? 0 : -1;
}
#endif /* def HAVE_PSIGNAL */
#define NUM_SIGNALS 64
static signal_info signals[NUM_SIGNALS];
static int acquire_mutex (pthread_mutex_t *mutex)
{
int mr;
while ((mr = pthread_mutex_lock (mutex)) == EAGAIN) {
/* try to acquire again */
}
if ((mr != 0) && (mr != EDEADLK)) {
errno = mr;
return -1;
}
return 0;
}
static void release_mutex (pthread_mutex_t *mutex)
{
int mr;
while ((mr = pthread_mutex_unlock (mutex)) == EAGAIN) {
/* try to release mutex again */
}
}
static inline int
keep_trying (int r)
{
return r == -1 && errno == EINTR;
}
// This tiny ad-hoc read/write lock is needed because of the very specific
// synchronization needed between default_handler and teardown_pipes:
// - Many default_handlers can be running at once
// - The signals_mutex already ensures only one teardown_pipes runs at once
// - If teardown_pipes starts while a default_handler is ongoing, it must block
// - If default_handler starts while a teardown_pipes is ongoing, it must *not* block
// Locks are implemented as ints.
// The lock is split into a teardown bit and a handler count (sign bit unused).
// There is a teardown running or waiting to run if the teardown bit is set.
// There is a handler running if the handler count is nonzero.
#define PIPELOCK_TEARDOWN_BIT ( (int)0x40000000 )
#define PIPELOCK_COUNT_MASK (~((int)0xC0000000))
#define PIPELOCK_GET_COUNT(x) ((x) & PIPELOCK_COUNT_MASK)
#define PIPELOCK_INCR_COUNT(x, by) (((x) & PIPELOCK_TEARDOWN_BIT) | (PIPELOCK_GET_COUNT (PIPELOCK_GET_COUNT (x) + (by))))
static inline void
acquire_pipelock_teardown (int *lock)
{
int lockvalue_draining;
// First mark that a teardown is occurring, so handlers will stop entering the lock.
while (1) {
int lockvalue = mph_int_get (lock);
lockvalue_draining = lockvalue | PIPELOCK_TEARDOWN_BIT;
if (mph_int_test_and_set (lock, lockvalue, lockvalue_draining))
break;
}
// Now wait for all handlers to complete.
while (1) {
if (0 == PIPELOCK_GET_COUNT (lockvalue_draining))
break; // We now hold the lock.
// Handler is still running, spin until it completes.
sched_yield (); // We can call this because !defined(HOST_WIN32)
lockvalue_draining = mph_int_get (lock);
}
}
static inline void
release_pipelock_teardown (int *lock)
{
while (1) {
int lockvalue = mph_int_get (lock);
int lockvalue_new = lockvalue & ~PIPELOCK_TEARDOWN_BIT;
// Technically this can't fail, because we hold both the pipelock and the mutex, but
if (mph_int_test_and_set (lock, lockvalue, lockvalue_new))
return;
}
}
// Return 1 for success
static inline int
acquire_pipelock_handler (int *lock)
{
while (1) {
int lockvalue = mph_int_get (lock);
if (lockvalue & PIPELOCK_TEARDOWN_BIT) // Final lock is being torn down
return 0;
int lockvalue_new = PIPELOCK_INCR_COUNT (lockvalue, 1);
if (mph_int_test_and_set (lock, lockvalue, lockvalue_new))
return 1;
}
}
static inline void
release_pipelock_handler (int *lock)
{
while (1) {
int lockvalue = mph_int_get (lock);
int lockvalue_new = PIPELOCK_INCR_COUNT (lockvalue, -1);
if (mph_int_test_and_set (lock, lockvalue, lockvalue_new))
return;
}
}
// This handler is registered once for each UnixSignal object. A pipe is maintained
// for each one; Wait users read at one end of this pipe, and default_handler sends
// a write on the pipe for each signal received while the Wait is ongoing.
//
// Notice a fairly unlikely race condition exists here: Because we synchronize with
// pipe teardown, but not install/uninstall (in other words, we are only trying to
// protect against writing on a closed pipe) it is technically possible a full
// uninstall and then an install could complete after signum is checked but before
// the remaining instructions execute. In this unlikely case count could be
// incremented or a byte written on the wrong signal handler.
static void
default_handler (int signum)
{
int i;
for (i = 0; i < NUM_SIGNALS; ++i) {
int fd;
signal_info* h = &signals [i];
if (mph_int_get (&h->signum) != signum)
continue;
mph_int_inc (&h->count);
if (!acquire_pipelock_handler (&h->pipelock))
continue; // Teardown is occurring on this object, no one to send to.
fd = mph_int_get (&h->write_fd);
if (fd > 0) { // If any listener exists to write to
int j,pipecounter;
char c = signum; // (Value is meaningless)
pipecounter = mph_int_get (&h->pipecnt); // Write one byte per pipe listener
for (j = 0; j < pipecounter; ++j) {
int r;
do { r = write (fd, &c, 1); } while (keep_trying (r));
}
}
release_pipelock_handler (&h->pipelock);
}
}
static pthread_mutex_t signals_mutex = PTHREAD_MUTEX_INITIALIZER;
// A UnixSignal object is being constructed
void*
Mono_Unix_UnixSignal_install (int sig)
{
#if defined(HAVE_SIGNAL)
int i;
signal_info* h = NULL; // signals[] slot to install to
int have_handler = 0; // Candidates for signal_info handler fields
void* handler = NULL;
if (acquire_mutex (&signals_mutex) == -1)
return NULL;
#if defined (SIGRTMIN) && defined (SIGRTMAX)
/*The runtime uses some rt signals for itself so it's important to not override them.*/
if (sig >= SIGRTMIN && sig <= SIGRTMAX && count_handlers (sig) == 0) {
struct sigaction sinfo;
sigaction (sig, NULL, &sinfo);
if (sinfo.sa_handler != SIG_DFL || (void*)sinfo.sa_sigaction != (void*)SIG_DFL) {
pthread_mutex_unlock (&signals_mutex);
errno = EADDRINUSE;
return NULL; // This is an rt signal with an existing handler. Bail out.
}
}
#endif /*defined (SIGRTMIN) && defined (SIGRTMAX)*/
// Scan through signals list looking for (1) an unused spot (2) a usable value for handler
for (i = 0; i < NUM_SIGNALS; ++i) {
int just_installed = 0;
// We're still looking for a signal_info spot, and this one is available:
if (h == NULL && mph_int_get (&signals [i].signum) == 0) {
h = &signals [i];
h->handler = signal (sig, default_handler);
if (h->handler == SIG_ERR) {
h->handler = NULL;
h = NULL;
break;
}
else {
just_installed = 1;
}
}
// Check if this slot has a "usable" (not installed by this file) handler-to-restore-later:
// (On the first signal to be installed, signals [i] will be == h when this happens.)
if (!have_handler && (just_installed || mph_int_get (&signals [i].signum) == sig) &&
signals [i].handler != default_handler) {
have_handler = 1;
handler = signals [i].handler;
}
if (h && have_handler) // We have everything we need
break;
}
if (h) {
// If we reached here without have_handler, this means that default_handler
// was set as the signal handler before the first UnixSignal object was installed.
g_assert (have_handler);
// Overwrite the tenative handler we set a moment ago with a known-usable one
h->handler = handler;
h->have_handler = 1;
mph_int_set (&h->count, 0);
mph_int_set (&h->pipecnt, 0);
mph_int_set (&h->signum, sig);
}
release_mutex (&signals_mutex);
return h;
#else
g_error ("signal() is not supported by this platform");
return 0;
#endif
}
static int
count_handlers (int signum)
{
int i;
int count = 0;
for (i = 0; i < NUM_SIGNALS; ++i) {
if (mph_int_get (&signals [i].signum) == signum)
++count;
}
return count;
}
// A UnixSignal object is being Disposed
int
Mono_Unix_UnixSignal_uninstall (void* info)
{
#if defined(HAVE_SIGNAL)
signal_info* h;
int r = -1;
if (acquire_mutex (&signals_mutex) == -1)
return -1;
h = info;
if (h == NULL || h < signals || h > &signals [NUM_SIGNALS])
errno = EINVAL;
else {
/* last UnixSignal -- we can unregister */
int signum = mph_int_get (&h->signum);
if (h->have_handler && count_handlers (signum) == 1) {
mph_sighandler_t p = signal (signum, h->handler);
if (p != SIG_ERR)
r = 0;
h->handler = NULL;
h->have_handler = 0;
}
mph_int_set (&h->signum, 0);
}
release_mutex (&signals_mutex);
return r;
#else
g_error ("signal() is not supported by this platform");
return 0;
#endif
}
// Set up a signal_info to begin waiting for signal
static int
setup_pipes (signal_info** signals, int count, struct pollfd *fd_structs, int *currfd)
{
int i;
int r = 0;
for (i = 0; i < count; ++i) {
signal_info* h;
int filedes[2];
h = signals [i];
if (mph_int_get (&h->pipecnt) == 0) { // First listener for this signal_info
if ((r = pipe (filedes)) != 0) {
break;
}
mph_int_set (&h->read_fd, filedes [0]);
mph_int_set (&h->write_fd, filedes [1]);
}
mph_int_inc (&h->pipecnt);
fd_structs[*currfd].fd = mph_int_get (&h->read_fd);
fd_structs[*currfd].events = POLLIN;
++(*currfd); // count is verified less than NUM_SIGNALS by caller
}
return r;
}
// Cleanup a signal_info after waiting for signal
static void
teardown_pipes (signal_info** signals, int count)
{
int i;
for (i = 0; i < count; ++i) {
signal_info* h = signals [i];
if (mph_int_dec_test (&h->pipecnt)) { // Final listener for this signal_info
acquire_pipelock_teardown (&h->pipelock);
int read_fd = mph_int_get (&h->read_fd);
int write_fd = mph_int_get (&h->write_fd);
if (read_fd != 0)
close (read_fd);
if (write_fd != 0)
close (write_fd);
mph_int_set (&h->read_fd, 0);
mph_int_set (&h->write_fd, 0);
release_pipelock_teardown (&h->pipelock);
}
}
}
// Given pipes set up, wait for a byte to arrive on one of them
static int
wait_for_any (signal_info** signals, int count, int *currfd, struct pollfd* fd_structs, int timeout, Mono_Posix_RuntimeIsShuttingDown shutting_down)
{
int r, idx;
// Poll until one of this signal_info's pipes is ready to read.
// Once a second, stop to check if the VM is shutting down.
do {
struct timeval tv;
struct timeval *ptv = NULL;
if (timeout != -1) {
tv.tv_sec = timeout / 1000;
tv.tv_usec = (timeout % 1000)*1000;
ptv = &tv;
}
r = poll (fd_structs, count, timeout);
} while (keep_trying (r) && !shutting_down ());
idx = -1;
if (r == 0)
idx = timeout;
else if (r > 0) { // The pipe[s] are ready to read.
int i;
for (i = 0; i < count; ++i) {
signal_info* h = signals [i];
if (fd_structs[i].revents & POLLIN) {
int r;
char c;
do {
r = read (mph_int_get (&h->read_fd), &c, 1);
} while (keep_trying (r) && !shutting_down ());
if (idx == -1)
idx = i;
}
}
}
return idx;
}
/*
* returns: -1 on error:
* timeout on timeout
* index into _signals array of signal that was generated on success
*/
int
Mono_Unix_UnixSignal_WaitAny (void** _signals, int count, int timeout /* milliseconds */, Mono_Posix_RuntimeIsShuttingDown shutting_down)
{
int r;
int currfd = 0;
struct pollfd fd_structs[NUM_SIGNALS];
signal_info** signals = (signal_info**) _signals;
if (count > NUM_SIGNALS)
return -1;
if (acquire_mutex (&signals_mutex) == -1)
return -1;
r = setup_pipes (signals, count, &fd_structs[0], &currfd);
release_mutex (&signals_mutex);
if (r == 0) {
r = wait_for_any (signals, count, &currfd, &fd_structs[0], timeout, shutting_down);
}
if (acquire_mutex (&signals_mutex) == -1)
return -1;
teardown_pipes (signals, count);
release_mutex (&signals_mutex);
return r;
}
#endif /* ndef HOST_WIN32 */
G_END_DECLS
/*
* vim: noexpandtab
*/