In the words of the author:

  o The polling mechanism for I/O readiness was changed from
    select() to poll().  In additon, a wrapped version of poll()
    is now provided.

  o The wrapped select routine now converts each fd_set to a
    poll array so that the thread scheduler doesn't have to
    perform a bitwise search for selected fds each time file
    descriptors are polled for I/O readiness.

  o The thread scheduler was modified to use a new queue (_workq)
    for threads that need work.  Threads waiting for I/O readiness
    and spinblocks are added to the work queue in addition to the
    waiting queue.  This reduces the time spent forming/searching
    the array of file descriptors being polled.

  o The waiting queue (_waitingq) is now maintained in order of
    thread wakeup time.  This allows the thread scheduler to
    find the nearest wakeup time by looking at the first thread
    in the queue instead of searching the entire queue.

  o Removed file descriptor locking for select/poll routines.  An
    application should not rely on the threads library for providing
    this locking; if necessary, the application should use mutexes
    to protect selecting/polling of file descriptors.

  o Retrieve and use the kernel clock rate/resolution at startup
    instead of hardcoding the clock resolution to 10 msec (tested
    with kernel running at 1000 HZ).

  o All queues have been changed to use queue.h macros.  These
    include the queues of all threads, dead threads, and threads
    waiting for file descriptor locks.

  o Added reinitialization of the GC mutex and condition variable
    after a fork.  Also prevented reallocation of the ready queue
    after a fork.

  o Prevented the wrapped close routine from closing the thread
    kernel pipes.

  o Initialized file descriptor table for stdio entries at thread
    init.

  o Provided additional flags to indicate to what queues threads
    belong.

  o Moved TAILQ initialization for statically allocated mutex and
    condition variables to after the spinlock.

  o Added dispatching of signals to pthread_kill.  Removing the
    dispatching of signals from thread activation broke sigsuspend
    when pthread_kill was used to send a signal to a thread.

  o Temporarily set the state of a thread to PS_SUSPENDED when it
    is first created and placed in the list of threads so that it
    will not be accidentally scheduled before becoming a member
    of one of the scheduling queues.

  o Change the signal handler to queue signals to the thread kernel
    pipe if the scheduling queues are protected.  When scheduling
    queues are unprotected, signals are then dequeued and handled.

  o Ensured that all installed signal handlers block the scheduling
    signal and that the scheduling signal handler blocks all
    other signals.  This ensures that the signal handler is only
    interruptible for and by non-scheduling signals.  An atomic
    lock is used to decide which instance of the signal handler
    will handle pending signals.

  o Removed _lock_thread_list and _unlock_thread_list as they are
    no longer used to protect the thread list.

  o Added missing RCS IDs to modified files.

  o Added checks for appropriate queue membership and activity when
    adding, removing, and searching the scheduling queues.  These
    checks add very little overhead and are enabled when compiled
    with _PTHREADS_INVARIANTS defined.  Suggested and implemented
    by Tor Egge with some modification by me.

  o Close a race condition in uthread_close.  (Tor Egge)

  o Protect the scheduling queues while modifying them in
    pthread_cond_signal and _thread_fd_unlock.  (Tor Egge)

  o Ensure that when a thread gets a mutex, the mutex is on that
    threads list of owned mutexes.  (Tor Egge)

  o Set the kernel-in-scheduler flag in _thread_kern_sched_state
    and _thread_kern_sched_state_unlock to prevent a scheduling
    signal from calling the scheduler again.  (Tor Egge)

  o Don't use TAILQ_FOREACH macro while searching the waiting
    queue for threads in a sigwait state, because a change of
    state destroys the TAILQ link.  It is actually safe to do
    so, though, because once a sigwaiting thread is found, the
    loop ends and the function returns.  (Tor Egge)

  o When dispatching signals to threads, make the thread inherit
    the signal deferral flag of the currently running thread.
    (Tor Egge)

Submitted by: Daniel Eischen <eischen@vigrid.com> and
              Tor Egge <Tor.Egge@fast.no>

1 files changed, 119 insertions, 91 deletions

diff --git a/lib/libkse/thread/thr_sig.c b/lib/libkse/thread/thr_sig.c
index e51d949..ac475a4 100644
--- a/lib/libkse/thread/thr_sig.c
+++ b/lib/libkse/thread/thr_sig.c
@@ -29,6 +29,7 @@
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
  *
+ * $Id$
  */
 #include <signal.h>
 #include <fcntl.h>
@@ -38,107 +39,51 @@
 #include <pthread.h>
 #include "pthread_private.h"
 
-/*
- * State change macro for signal handler:
- */
-#define PTHREAD_SIG_NEW_STATE(thrd, newstate) {				\
-	if ((_thread_run->sched_defer_count == 0) &&			\
-	    (_thread_kern_in_sched == 0)) { 				\
-		PTHREAD_NEW_STATE(thrd, newstate);			\
-	} else {							\
-		_waitingq_check_reqd = 1;				\
-		PTHREAD_SET_STATE(thrd, newstate);			\
-	}								\
-}
-
 /* Static variables: */
-static int		volatile yield_on_unlock_thread	= 0;
-static spinlock_t	thread_link_list_lock	= _SPINLOCK_INITIALIZER;
-
-/* Lock the thread list: */
-void
-_lock_thread_list()
-{
-	/* Lock the thread list: */
-	_SPINLOCK(&thread_link_list_lock);
-}
+static spinlock_t	signal_lock = _SPINLOCK_INITIALIZER;
+unsigned int		pending_sigs[NSIG];
+unsigned int		handled_sigs[NSIG];
+int			volatile check_pending = 0;
 
-/* Lock the thread list: */
+/* Initialize signal handling facility: */
 void
-_unlock_thread_list()
+_thread_sig_init(void)
 {
-	/* Unlock the thread list: */
-	_SPINUNLOCK(&thread_link_list_lock);
-
-	/*
-	 * Check if a scheduler interrupt occurred while the thread
-	 * list was locked:
-	 */
-	if (yield_on_unlock_thread) {
-		/* Reset the interrupt flag: */
-		yield_on_unlock_thread = 0;
+	int i;
 
-		/* This thread has overstayed it's welcome: */
-		sched_yield();
+	/* Clear pending and handled signal counts: */
+	for (i = 1; i < NSIG; i++) {
+		pending_sigs[i - 1] = 0;
+		handled_sigs[i - 1] = 0;
 	}
+
+	/* Clear the lock: */
+	signal_lock.access_lock = 0;
 }
 
 void
 _thread_sig_handler(int sig, int code, struct sigcontext * scp)
 {
-	char            c;
-	int             i;
-	pthread_t       pthread;
-
-	/*
-	 * Check if the pthread kernel has unblocked signals (or is about to)
-	 * and was on its way into a _select when the current
-	 * signal interrupted it: 
-	 */
-	if (_thread_kern_in_select) {
-		/* Cast the signal number to a character variable: */
-		c = sig;
-
-		/*
-		 * Write the signal number to the kernel pipe so that it will
-		 * be ready to read when this signal handler returns. This
-		 * means that the _select call will complete
-		 * immediately. 
-		 */
-		_thread_sys_write(_thread_kern_pipe[1], &c, 1);
-	}
-	/* Check if the signal requires a dump of thread information: */
-	if (sig == SIGINFO)
-		/* Dump thread information to file: */
-		_thread_dump_info();
+	char	c;
+	int	i;
 
 	/* Check if an interval timer signal: */
-	else if (sig == _SCHED_SIGNAL) {
-		/* Check if the scheduler interrupt has come at an
-		 * unfortunate time which one of the threads is
-		 * modifying the thread list:
-		 */
-		if (thread_link_list_lock.access_lock)
+	if (sig == _SCHED_SIGNAL) {
+		if (_thread_kern_in_sched != 0) {
 			/*
-			 * Set a flag so that the thread that has
-			 * the lock yields when it unlocks the
-			 * thread list:
+			 * The scheduler is already running; ignore this
+			 * signal.
 			 */
-			yield_on_unlock_thread = 1;
-
+		}
 		/*
 		 * Check if the scheduler interrupt has come when
 		 * the currently running thread has deferred thread
-		 * scheduling.
+		 * signals.
 		 */
-		else if (_thread_run->sched_defer_count)
-			_thread_run->yield_on_sched_undefer = 1;
+		else if (_thread_run->sig_defer_count > 0)
+			_thread_run->yield_on_sig_undefer = 1;
 
-		/*
-		 * Check if the kernel has not been interrupted while
-		 * executing scheduler code:
-		 */
-		else if (!_thread_kern_in_sched) {
+		else {
 			/*
 			 * Schedule the next thread. This function is not
 			 * expected to return because it will do a longjmp
@@ -152,6 +97,72 @@ _thread_sig_handler(int sig, int code, struct sigcontext * scp)
 			 */
 			PANIC("Returned to signal function from scheduler");
 		}
+	}
+	/*
+	 * Check if the kernel has been interrupted while the scheduler
+	 * is accessing the scheduling queues or if there is a currently
+	 * running thread that has deferred signals.
+	 */
+	else if ((_queue_signals != 0) || ((_thread_kern_in_sched == 0) &&
+	    (_thread_run->sig_defer_count > 0))) {
+		/* Cast the signal number to a character variable: */
+		c = sig;
+
+		/*
+		 * Write the signal number to the kernel pipe so that it will
+		 * be ready to read when this signal handler returns.
+		 */
+		_thread_sys_write(_thread_kern_pipe[1], &c, 1);
+
+		/* Indicate that there are queued signals in the pipe. */
+		_sigq_check_reqd = 1;
+	}
+	else {
+		if (_atomic_lock(&signal_lock.access_lock)) {
+			/* There is another signal handler running: */
+			pending_sigs[sig - 1]++;
+			check_pending = 1;
+		}
+		else {
+			/* It's safe to handle the signal now. */
+			_thread_sig_handle(sig, scp);
+
+			/* Reset the pending and handled count back to 0: */
+			pending_sigs[sig - 1] = 0;
+			handled_sigs[sig - 1] = 0;
+
+			signal_lock.access_lock = 0;
+		}
+
+		/* Enter a loop to process pending signals: */
+		while ((check_pending != 0) &&
+		    (_atomic_lock(&signal_lock.access_lock) == 0)) {
+			check_pending = 0;
+			for (i = 1; i < NSIG; i++) {
+				if (pending_sigs[i - 1] > handled_sigs[i - 1])
+					_thread_sig_handle(i, scp);
+			}
+			signal_lock.access_lock = 0;
+		}
+	}
+}
+
+void
+_thread_sig_handle(int sig, struct sigcontext * scp)
+{
+	int		i;
+	pthread_t	pthread, pthread_next;
+
+	/* Check if the signal requires a dump of thread information: */
+	if (sig == SIGINFO)
+		/* Dump thread information to file: */
+		_thread_dump_info();
+
+	/* Check if an interval timer signal: */
+	else if (sig == _SCHED_SIGNAL) {
+		/*
+		 * This shouldn't ever occur (should this panic?).
+		 */
 	} else {
 		/* Check if a child has terminated: */
 		if (sig == SIGCHLD) {
@@ -183,10 +194,9 @@ _thread_sig_handler(int sig, int code, struct sigcontext * scp)
 			 * Enter a loop to discard pending SIGCONT
 			 * signals:
 			 */
-			for (pthread = _thread_link_list;
-			    pthread != NULL;
-			    pthread = pthread->nxt)
+			TAILQ_FOREACH(pthread, &_thread_list, tle) {
 				sigdelset(&pthread->sigpend,SIGCONT);
+			}
 		}
 
 		/*
@@ -196,11 +206,18 @@ _thread_sig_handler(int sig, int code, struct sigcontext * scp)
 		 * if there are multiple waiters, we'll give it to the
 		 * first one we find.
 		 */
-		TAILQ_FOREACH(pthread, &_waitingq, pqe) {
+		for (pthread = TAILQ_FIRST(&_waitingq);
+		    pthread != NULL; pthread = pthread_next) {
+			/*
+			 * Grab the next thread before possibly destroying
+			 * the link entry.
+			 */
+			pthread_next = TAILQ_NEXT(pthread, pqe);
+
 			if ((pthread->state == PS_SIGWAIT) &&
 			    sigismember(pthread->data.sigwait, sig)) {
 				/* Change the state of the thread to run: */
-				PTHREAD_SIG_NEW_STATE(pthread,PS_RUNNING);
+				PTHREAD_NEW_STATE(pthread,PS_RUNNING);
 
 				/* Return the signal number: */
 				pthread->signo = sig;
@@ -219,10 +236,13 @@ _thread_sig_handler(int sig, int code, struct sigcontext * scp)
 			 * Enter a loop to process each thread in the linked
 			 * list: 
 			 */
-			for (pthread = _thread_link_list; pthread != NULL;
-			     pthread = pthread->nxt) {
+			TAILQ_FOREACH(pthread, &_thread_list, tle) {
 				pthread_t pthread_saved = _thread_run;
 
+				/* Current thread inside critical region? */
+				if (_thread_run->sig_defer_count > 0)
+					pthread->sig_defer_count++;
+
 				_thread_run = pthread;
 				_thread_signal(pthread,sig);
 
@@ -232,6 +252,10 @@ _thread_sig_handler(int sig, int code, struct sigcontext * scp)
 				 */
 				_dispatch_signals();
 				_thread_run = pthread_saved;
+
+				/* Current thread inside critical region? */
+				if (_thread_run->sig_defer_count > 0)
+					pthread->sig_defer_count--;
 			}
 	}
 
@@ -284,7 +308,7 @@ _thread_signal(pthread_t pthread, int sig)
 			pthread->interrupted = 1;
 
 		/* Change the state of the thread to run: */
-		PTHREAD_SIG_NEW_STATE(pthread,PS_RUNNING);
+		PTHREAD_NEW_STATE(pthread,PS_RUNNING);
 
 		/* Return the signal number: */
 		pthread->signo = sig;
@@ -296,6 +320,7 @@ _thread_signal(pthread_t pthread, int sig)
 	 */
 	case PS_FDR_WAIT:
 	case PS_FDW_WAIT:
+	case PS_POLL_WAIT:
 	case PS_SLEEP_WAIT:
 	case PS_SELECT_WAIT:
 		if (sig != SIGCHLD ||
@@ -303,8 +328,11 @@ _thread_signal(pthread_t pthread, int sig)
 			/* Flag the operation as interrupted: */
 			pthread->interrupted = 1;
 
+			if (pthread->flags & PTHREAD_FLAGS_IN_WORKQ)
+				PTHREAD_WORKQ_REMOVE(pthread);
+
 			/* Change the state of the thread to run: */
-			PTHREAD_SIG_NEW_STATE(pthread,PS_RUNNING);
+			PTHREAD_NEW_STATE(pthread,PS_RUNNING);
 
 			/* Return the signal number: */
 			pthread->signo = sig;
@@ -319,7 +347,7 @@ _thread_signal(pthread_t pthread, int sig)
 		if (!sigismember(&pthread->sigmask, sig) &&
 		    _thread_sigact[sig - 1].sa_handler != SIG_DFL) {
 			/* Change the state of the thread to run: */
-			PTHREAD_SIG_NEW_STATE(pthread,PS_RUNNING);
+			PTHREAD_NEW_STATE(pthread,PS_RUNNING);
 
 			/* Return the signal number: */
 			pthread->signo = sig;