freebsd-skq/lib/libkse/thread/thr_sig.c

905 lines
26 KiB
C
Raw Normal View History

/*
* Copyright (c) 1995-1998 John Birrell <jb@cimlogic.com.au>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by John Birrell.
* 4. Neither the name of the author nor the names of any co-contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY JOHN BIRRELL AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
1999-08-28 00:22:10 +00:00
* $FreeBSD$
*/
sigset_t change (part 5 of 5) ----------------------------- Most of the userland changes are in libc. For both the alpha and the i386 setjmp has been changed to accomodate for the new sigset_t. Internally, libc is mostly rewritten to use the new syscalls. The exception is in compat-43/sigcompat.c The POSIX thread library has also been rewritten to use the new sigset_t. Except, that it currently only handles NSIG signals instead of the maximum _SIG_MAXSIG. This should not be a problem because current applications don't use any signals higher than NSIG. There are version bumps for the following libraries: libdialog libreadline libc libc_r libedit libftpio libss These libraries either a) have one of the modified structures visible in the interface, or b) use sigset_t internally and may cause breakage if new binaries are used against libraries that don't have the sigset_t change. This not an immediate issue, but will be as soon as applications start using the new range to its fullest. NOTE: libncurses already had an version bump and has not been given one now. NOTE: doscmd is a real casualty and has been disconnected for the moment. Reconnection will eventually happen after doscmd has been fixed. I'm aware that being the last one to touch it, I'm automaticly promoted to being maintainer. According to good taste this means that I will receive a badge which either will be glued or mechanically stapled, drilled or otherwise violently forced onto me :-) NOTE: pcvt/vttest cannot be compiled with -traditional. The change cause sys/types to be included along the way which contains the const and volatile modifiers. I don't consider this a solution, but more a workaround.
1999-09-29 15:18:46 +00:00
#include <sys/param.h>
#include <sys/types.h>
#include <sys/signalvar.h>
#include <signal.h>
#include <errno.h>
#include <fcntl.h>
#include <unistd.h>
#include <string.h>
#include <pthread.h>
#include "thr_private.h"
#include "pthread_md.h"
/* Prototypes: */
static void build_siginfo(siginfo_t *info, int signo);
/* static void thr_sig_add(struct pthread *pthread, int sig, siginfo_t *info); */
static void thr_sig_check_state(struct pthread *pthread, int sig);
static struct pthread *thr_sig_find(struct kse *curkse, int sig,
siginfo_t *info);
static void handle_special_signals(struct kse *curkse, int sig);
static void thr_sigframe_add(struct pthread *thread, int sig,
siginfo_t *info);
static void thr_sigframe_restore(struct pthread *thread,
struct pthread_sigframe *psf);
static void thr_sigframe_save(struct pthread *thread,
struct pthread_sigframe *psf);
static void thr_sig_invoke_handler(struct pthread *, int sig,
siginfo_t *info, ucontext_t *ucp);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/* #define DEBUG_SIGNAL */
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
#ifdef DEBUG_SIGNAL
#define DBG_MSG stdout_debug
#else
#define DBG_MSG(x...)
#endif
/*
* Signal setup and delivery.
*
* 1) Delivering signals to threads in the same KSE.
* These signals are sent by upcall events and are set in the
* km_sigscaught field of the KSE mailbox. Since these signals
* are received while operating on the KSE stack, they can be
* delivered either by using signalcontext() to add a stack frame
* to the target thread's stack, or by adding them in the thread's
* pending set and having the thread run them down after it
* 2) Delivering signals to threads in other KSEs/KSEGs.
* 3) Delivering signals to threads in critical regions.
* 4) Delivering signals to threads after they change their signal masks.
*
* Methods of delivering signals.
*
* 1) Add a signal frame to the thread's saved context.
* 2) Add the signal to the thread structure, mark the thread as
* having signals to handle, and let the thread run them down
* after it resumes from the KSE scheduler.
*
* Problem with 1). You can't do this to a running thread or a
* thread in a critical region.
*
* Problem with 2). You can't do this to a thread that doesn't
* yield in some way (explicitly enters the scheduler). A thread
* blocked in the kernel or a CPU hungry thread will not see the
* signal without entering the scheduler.
*
* The solution is to use both 1) and 2) to deliver signals:
*
* o Thread in critical region - use 2). When the thread
* leaves the critical region it will check to see if it
* has pending signals and run them down.
*
* o Thread enters scheduler explicitly - use 2). The thread
* can check for pending signals after it returns from the
* the scheduler.
*
* o Thread is running and not current thread - use 2). When the
* thread hits a condition specified by one of the other bullets,
* the signal will be delivered.
*
* o Thread is running and is current thread (e.g., the thread
* has just changed its signal mask and now sees that it has
* pending signals) - just run down the pending signals.
*
* o Thread is swapped out due to quantum expiration - use 1)
*
* o Thread is blocked in kernel - kse_thr_wakeup() and then
* use 1)
*/
/*
* Rules for selecting threads for signals received:
*
* 1) If the signal is a sychronous signal, it is delivered to
* the generating (current thread). If the thread has the
* signal masked, it is added to the threads pending signal
* set until the thread unmasks it.
*
* 2) A thread in sigwait() where the signal is in the thread's
* waitset.
*
* 3) A thread in sigsuspend() where the signal is not in the
* thread's suspended signal mask.
*
* 4) Any thread (first found/easiest to deliver) that has the
* signal unmasked.
*/
/*
* This signal handler only delivers asynchronous signals.
* This must be called with upcalls disabled and without
* holding any locks.
*/
void
_thr_sig_dispatch(struct kse *curkse, int sig, siginfo_t *info)
{
struct pthread *thread;
DBG_MSG(">>> _thr_sig_dispatch(%d)\n", sig);
/* Some signals need special handling: */
handle_special_signals(curkse, sig);
stderr_debug("dispatch sig:%d\n", sig);
while ((thread = thr_sig_find(curkse, sig, info)) != NULL) {
/*
* Setup the target thread to receive the signal:
*/
DBG_MSG("Got signal %d, selecting thread %p\n", sig, thread);
KSE_SCHED_LOCK(curkse, thread->kseg);
if ((thread->state == PS_DEAD) ||
(thread->state == PS_DEADLOCK) ||
THR_IS_EXITING(thread) || THR_IS_SUSPENDED(thread)) {
KSE_SCHED_UNLOCK(curkse, thread->kseg);
_thr_ref_delete(NULL, thread);
} else {
_thr_sig_add(thread, sig, info);
KSE_SCHED_UNLOCK(curkse, thread->kseg);
_thr_ref_delete(NULL, thread);
break;
}
}
}
void
_thr_sig_handler(int sig, siginfo_t *info, ucontext_t *ucp)
{
void (*sigfunc)(int, siginfo_t *, void *);
struct kse *curkse;
curkse = _get_curkse();
if ((curkse == NULL) || ((curkse->k_flags & KF_STARTED) == 0)) {
/* Upcalls are not yet started; just call the handler. */
sigfunc = _thread_sigact[sig - 1].sa_sigaction;
ucp->uc_sigmask = _thr_proc_sigmask;
if (((__sighandler_t *)sigfunc != SIG_DFL) &&
((__sighandler_t *)sigfunc != SIG_IGN)) {
if (((_thread_sigact[sig - 1].sa_flags & SA_SIGINFO)
!= 0) || (info == NULL))
(*(sigfunc))(sig, info, ucp);
else
(*(sigfunc))(sig, (siginfo_t *)info->si_code,
ucp);
}
}
else {
/* Nothing. */
DBG_MSG("Got signal %d\n", sig);
sigaddset(&curkse->k_mbx.km_sigscaught, sig);
ucp->uc_sigmask = _thr_proc_sigmask;
}
}
static void
thr_sig_invoke_handler(struct pthread *curthread, int sig, siginfo_t *info,
ucontext_t *ucp)
{
void (*sigfunc)(int, siginfo_t *, void *);
sigset_t saved_mask;
int saved_seqno;
/* Invoke the signal handler without going through the scheduler:
*/
DBG_MSG("Got signal %d, calling handler for current thread %p\n",
sig, curthread);
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>
1999-06-20 08:28:48 +00:00
/*
* Setup the threads signal mask.
*
* The mask is changed in the thread's active signal mask
* (in the context) and not in the base signal mask because
* a thread is allowed to change its signal mask within a
* signal handler. If it does, the signal mask restored
* after the handler should be the same as that set by the
* thread during the handler, not the original mask from
* before calling the handler. The thread could also
* modify the signal mask in the context and expect this
* mask to be used.
*/
THR_SCHED_LOCK(curthread, curthread);
saved_mask = curthread->tmbx.tm_context.uc_sigmask;
saved_seqno = curthread->sigmask_seqno;
SIGSETOR(curthread->tmbx.tm_context.uc_sigmask,
_thread_sigact[sig - 1].sa_mask);
sigaddset(&curthread->tmbx.tm_context.uc_sigmask, sig);
THR_SCHED_UNLOCK(curthread, curthread);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Check that a custom handler is installed and if
* the signal is not blocked:
*/
sigfunc = _thread_sigact[sig - 1].sa_sigaction;
ucp->uc_sigmask = _thr_proc_sigmask;
if (((__sighandler_t *)sigfunc != SIG_DFL) &&
((__sighandler_t *)sigfunc != SIG_IGN)) {
if (((_thread_sigact[sig - 1].sa_flags & SA_SIGINFO) != 0) ||
(info == NULL))
(*(sigfunc))(sig, info, ucp);
else
(*(sigfunc))(sig, (siginfo_t *)info->si_code, ucp);
}
/*
* Restore the thread's signal mask.
*/
if (saved_seqno == curthread->sigmask_seqno)
curthread->tmbx.tm_context.uc_sigmask = saved_mask;
else
curthread->tmbx.tm_context.uc_sigmask = curthread->sigmask;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
/*
* Find a thread that can handle the signal. This must be called
* with upcalls disabled.
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
struct pthread *
thr_sig_find(struct kse *curkse, int sig, siginfo_t *info)
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>
1999-06-20 08:28:48 +00:00
{
int handler_installed;
struct pthread *pthread;
struct pthread *suspended_thread, *signaled_thread;
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>
1999-06-20 08:28:48 +00:00
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
DBG_MSG("Looking for thread to handle signal %d\n", sig);
handler_installed = (_thread_sigact[sig - 1].sa_handler != SIG_IGN) &&
(_thread_sigact[sig - 1].sa_handler != SIG_DFL);
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>
1999-06-20 08:28:48 +00:00
/* Check if the signal requires a dump of thread information: */
if (sig == SIGINFO) {
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>
1999-06-20 08:28:48 +00:00
/* Dump thread information to file: */
_thread_dump_info();
}
/*
* Enter a loop to look for threads that have the signal
* unmasked. POSIX specifies that a thread in a sigwait
* will get the signal over any other threads. Second
* preference will be threads in in a sigsuspend. Third
* preference will be the current thread. If none of the
* above, then the signal is delivered to the first thread
* that is found. Note that if a custom handler is not
* installed, the signal only affects threads in sigwait.
*/
suspended_thread = NULL;
signaled_thread = NULL;
KSE_LOCK_ACQUIRE(curkse, &_thread_list_lock);
TAILQ_FOREACH(pthread, &_thread_list, tle) {
/* Take the scheduling lock. */
KSE_SCHED_LOCK(curkse, pthread->kseg);
if ((pthread->state == PS_SIGWAIT) &&
sigismember(pthread->data.sigwait, sig)) {
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>
1999-06-20 08:28:48 +00:00
/*
* Return the signal number and make the
* thread runnable.
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>
1999-06-20 08:28:48 +00:00
*/
pthread->signo = sig;
_thr_setrunnable_unlocked(pthread);
KSE_SCHED_UNLOCK(curkse, pthread->kseg);
/*
* POSIX doesn't doesn't specify which thread
* will get the signal if there are multiple
* waiters, so we give it to the first thread
* we find.
*
* Do not attempt to deliver this signal
* to other threads and do not add the signal
* to the process pending set.
*/
KSE_LOCK_RELEASE(curkse, &_thread_list_lock);
DBG_MSG("Waking thread %p in sigwait with signal %d\n",
pthread, sig);
return (NULL);
}
else if ((pthread->state == PS_DEAD) ||
(pthread->state == PS_DEADLOCK) ||
THR_IS_EXITING(pthread) || THR_IS_SUSPENDED(pthread))
; /* Skip this thread. */
else if ((handler_installed != 0) &&
!sigismember(&pthread->tmbx.tm_context.uc_sigmask, sig)) {
if (pthread->state == PS_SIGSUSPEND) {
if (suspended_thread == NULL) {
suspended_thread = pthread;
suspended_thread->refcount++;
}
} else if (signaled_thread == NULL) {
signaled_thread = pthread;
signaled_thread->refcount++;
}
}
KSE_SCHED_UNLOCK(curkse, pthread->kseg);
}
KSE_LOCK_RELEASE(curkse, &_thread_list_lock);
/*
* Only perform wakeups and signal delivery if there is a
* custom handler installed:
*/
if (handler_installed == 0) {
/*
* There is no handler installed; nothing to do here.
*/
} else if (suspended_thread == NULL &&
signaled_thread == NULL) {
/*
* Add it to the set of signals pending
* on the process:
*/
KSE_LOCK_ACQUIRE(curkse, &_thread_signal_lock);
if (!sigismember(&_thr_proc_sigpending, sig)) {
sigaddset(&_thr_proc_sigpending, sig);
if (info == NULL)
build_siginfo(&_thr_proc_siginfo[sig], sig);
else
memcpy(&_thr_proc_siginfo[sig], info,
sizeof(*info));
}
KSE_LOCK_RELEASE(curkse, &_thread_signal_lock);
} else {
/*
* We only deliver the signal to one thread;
* give preference to the suspended thread:
*/
if (suspended_thread != NULL) {
pthread = suspended_thread;
_thr_ref_delete(NULL, signaled_thread);
} else
pthread = signaled_thread;
return (pthread);
}
return (NULL);
}
static void
build_siginfo(siginfo_t *info, int signo)
{
bzero(info, sizeof(*info));
info->si_signo = signo;
info->si_pid = _thr_pid;
}
/*
* This is called by a thread when it has pending signals to deliver.
* It should only be called from the context of the thread.
*/
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
void
_thr_sig_rundown(struct pthread *curthread, ucontext_t *ucp,
struct pthread_sigframe *psf)
{
struct pthread_sigframe psf_save;
sigset_t sigset;
int i;
THR_SCHED_LOCK(curthread, curthread);
memcpy(&sigset, &curthread->sigpend, sizeof(sigset));
sigemptyset(&curthread->sigpend);
if (psf != NULL) {
memcpy(&psf_save, psf, sizeof(*psf));
SIGSETOR(sigset, psf_save.psf_sigset);
sigemptyset(&psf->psf_sigset);
}
THR_SCHED_UNLOCK(curthread, curthread);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/* Check the threads previous state: */
if ((psf != NULL) && (psf_save.psf_state != PS_RUNNING)) {
/*
* Do a little cleanup handling for those threads in
* queues before calling the signal handler. Signals
* for these threads are temporarily blocked until
* after cleanup handling.
*/
switch (psf_save.psf_state) {
case PS_COND_WAIT:
_cond_wait_backout(curthread);
psf_save.psf_state = PS_RUNNING;
break;
case PS_MUTEX_WAIT:
_mutex_lock_backout(curthread);
psf_save.psf_state = PS_RUNNING;
break;
default:
break;
}
}
/*
* Lower the priority before calling the handler in case
* it never returns (longjmps back):
*/
curthread->active_priority &= ~THR_SIGNAL_PRIORITY;
for (i = 1; i < NSIG; i++) {
if (sigismember(&sigset, i) != 0) {
/* Call the handler: */
thr_sig_invoke_handler(curthread, i,
&curthread->siginfo[i], ucp);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
}
THR_SCHED_LOCK(curthread, curthread);
if (psf != NULL)
thr_sigframe_restore(curthread, &psf_save);
/* Restore the signal mask. */
curthread->tmbx.tm_context.uc_sigmask = curthread->sigmask;
THR_SCHED_UNLOCK(curthread, curthread);
_thr_sig_check_pending(curthread);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
/*
* This checks pending signals for the current thread. It should be
* called whenever a thread changes its signal mask. Note that this
* is called from a thread (using its stack).
*
* XXX - We might want to just check to see if there are pending
* signals for the thread here, but enter the UTS scheduler
* to actually install the signal handler(s).
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
void
_thr_sig_check_pending(struct pthread *curthread)
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
{
sigset_t sigset;
sigset_t pending_process;
sigset_t pending_thread;
kse_critical_t crit;
int i;
curthread->check_pending = 0;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Check if there are pending signals for the running
* thread or process that aren't blocked:
*/
crit = _kse_critical_enter();
KSE_LOCK_ACQUIRE(curthread->kse, &_thread_signal_lock);
sigset = _thr_proc_sigpending;
KSE_LOCK_RELEASE(curthread->kse, &_thread_signal_lock);
_kse_critical_leave(crit);
THR_SCHED_LOCK(curthread, curthread);
SIGSETOR(sigset, curthread->sigpend);
SIGSETNAND(sigset, curthread->tmbx.tm_context.uc_sigmask);
if (SIGNOTEMPTY(sigset)) {
ucontext_t uc;
volatile int once;
curthread->check_pending = 0;
THR_SCHED_UNLOCK(curthread, curthread);
/*
* Split the pending signals into those that were
* pending on the process and those that were pending
* on the thread.
*/
sigfillset(&pending_process);
sigfillset(&pending_thread);
for (i = 1; i < NSIG; i++) {
if (sigismember(&sigset, i) != 0) {
if (sigismember(&curthread->sigpend, i) != 0) {
build_siginfo(&curthread->siginfo[i], i);
sigdelset(&pending_thread, i);
} else {
memcpy(&curthread->siginfo[i],
&_thr_proc_siginfo[i],
sizeof(siginfo_t));
sigdelset(&pending_process, i);
}
}
}
/*
* Remove any process pending signals that were scheduled
* to be delivered from process' pending set.
*/
crit = _kse_critical_enter();
KSE_LOCK_ACQUIRE(curthread->kse, &_thread_signal_lock);
SIGSETAND(_thr_proc_sigpending, pending_process);
KSE_LOCK_RELEASE(curthread->kse, &_thread_signal_lock);
_kse_critical_leave(crit);
/*
* Remove any thread pending signals that were scheduled
* to be delivered from thread's pending set.
*/
THR_SCHED_LOCK(curthread, curthread);
SIGSETAND(curthread->sigpend, pending_thread);
THR_SCHED_UNLOCK(curthread, curthread);
once = 0;
THR_GETCONTEXT(&uc);
if (once == 0) {
once = 1;
for (i = 1; i < NSIG; i++) {
if (sigismember(&sigset, i) != 0) {
/* Call the handler: */
thr_sig_invoke_handler(curthread, i,
&curthread->siginfo[i], &uc);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
}
}
}
else
THR_SCHED_UNLOCK(curthread, curthread);
}
/*
* This must be called with upcalls disabled.
*/
static void
handle_special_signals(struct kse *curkse, int sig)
{
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
switch (sig) {
/*
* POSIX says that pending SIGCONT signals are
* discarded when one of these signals occurs.
*/
case SIGTSTP:
case SIGTTIN:
case SIGTTOU:
KSE_LOCK_ACQUIRE(curkse, &_thread_signal_lock);
sigdelset(&_thr_proc_sigpending, SIGCONT);
KSE_LOCK_RELEASE(curkse, &_thread_signal_lock);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
break;
default:
break;
}
}
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Perform thread specific actions in response to a signal.
* This function is only called if there is a handler installed
* for the signal, and if the target thread has the signal
* unmasked.
*
* This must be called with the thread's scheduling lock held.
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
void
_thr_sig_add(struct pthread *pthread, int sig, siginfo_t *info)
{
int restart;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
int suppress_handler = 0;
if (pthread->curframe == NULL) {
/*
* This thread is active. Just add it to the
* thread's pending set.
*/
sigaddset(&pthread->sigpend, sig);
pthread->check_pending = 1;
if (info == NULL)
build_siginfo(&pthread->siginfo[sig], sig);
else if (info != &pthread->siginfo[sig])
memcpy(&pthread->siginfo[sig], info,
sizeof(*info));
if ((pthread->blocked != 0) && !THR_IN_CRITICAL(pthread))
kse_thr_interrupt(&pthread->tmbx /* XXX - restart?!?! */);
}
else {
restart = _thread_sigact[sig - 1].sa_flags & SA_RESTART;
/* Make sure this signal isn't still in the pending set: */
sigdelset(&pthread->sigpend, sig);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Process according to thread state:
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
switch (pthread->state) {
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* States which do not change when a signal is trapped:
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
case PS_DEAD:
case PS_DEADLOCK:
case PS_LOCKWAIT:
case PS_SUSPENDED:
case PS_STATE_MAX:
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* You can't call a signal handler for threads in these
* states.
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
suppress_handler = 1;
break;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* States which do not need any cleanup handling when signals
* occur:
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
case PS_RUNNING:
/*
* Remove the thread from the queue before changing its
* priority:
*/
if ((pthread->flags & THR_FLAGS_IN_RUNQ) != 0)
THR_RUNQ_REMOVE(pthread);
break;
/*
* States which cannot be interrupted but still require the
* signal handler to run:
*/
case PS_COND_WAIT:
case PS_MUTEX_WAIT:
/*
* Remove the thread from the wait queue. It will
* be added back to the wait queue once all signal
* handlers have been invoked.
*/
KSE_WAITQ_REMOVE(pthread->kse, pthread);
break;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
case PS_SLEEP_WAIT:
/*
* Unmasked signals always cause sleep to terminate
* early regardless of SA_RESTART:
*/
pthread->interrupted = 1;
KSE_WAITQ_REMOVE(pthread->kse, pthread);
THR_SET_STATE(pthread, PS_RUNNING);
break;
case PS_JOIN:
case PS_SIGSUSPEND:
KSE_WAITQ_REMOVE(pthread->kse, pthread);
THR_SET_STATE(pthread, PS_RUNNING);
break;
case PS_SIGWAIT:
/*
* The signal handler is not called for threads in
* SIGWAIT.
*/
suppress_handler = 1;
/* Wake up the thread if the signal is blocked. */
if (sigismember(pthread->data.sigwait, sig)) {
/* Return the signal number: */
pthread->signo = sig;
/* Make the thread runnable: */
_thr_setrunnable_unlocked(pthread);
} else
/* Increment the pending signal count. */
sigaddset(&pthread->sigpend, sig);
break;
}
if (suppress_handler == 0) {
/*
* Setup a signal frame and save the current threads
* state:
*/
thr_sigframe_add(pthread, sig, info);
if (pthread->state != PS_RUNNING)
THR_SET_STATE(pthread, PS_RUNNING);
/*
* The thread should be removed from all scheduling
* queues at this point. Raise the priority and
* place the thread in the run queue. It is also
* possible for a signal to be sent to a suspended
* thread, mostly via pthread_kill(). If a thread
* is suspended, don't insert it into the priority
* queue; just set its state to suspended and it
* will run the signal handler when it is resumed.
*/
pthread->active_priority |= THR_SIGNAL_PRIORITY;
if ((pthread->flags & THR_FLAGS_IN_RUNQ) == 0)
THR_RUNQ_INSERT_TAIL(pthread);
}
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
}
static void
thr_sig_check_state(struct pthread *pthread, int sig)
{
/*
* Process according to thread state:
*/
switch (pthread->state) {
/*
* States which do not change when a signal is trapped:
*/
case PS_RUNNING:
case PS_LOCKWAIT:
case PS_MUTEX_WAIT:
case PS_COND_WAIT:
case PS_JOIN:
case PS_SUSPENDED:
case PS_DEAD:
case PS_DEADLOCK:
case PS_STATE_MAX:
break;
case PS_SIGWAIT:
/* Wake up the thread if the signal is blocked. */
if (sigismember(pthread->data.sigwait, sig)) {
/* Return the signal number: */
pthread->signo = sig;
/* Change the state of the thread to run: */
_thr_setrunnable_unlocked(pthread);
} else
/* Increment the pending signal count. */
sigaddset(&pthread->sigpend, sig);
break;
case PS_SIGSUSPEND:
case PS_SLEEP_WAIT:
/*
* Remove the thread from the wait queue and make it
* runnable:
*/
_thr_setrunnable_unlocked(pthread);
/* Flag the operation as interrupted: */
pthread->interrupted = 1;
break;
}
}
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Send a signal to a specific thread (ala pthread_kill):
*/
void
_thr_sig_send(struct pthread *pthread, int sig)
{
struct pthread *curthread = _get_curthread();
/* Lock the scheduling queue of the target thread. */
THR_SCHED_LOCK(curthread, pthread);
/* Check for signals whose actions are SIG_DFL: */
if (_thread_sigact[sig - 1].sa_handler == SIG_DFL) {
/*
* Check to see if a temporary signal handler is
* installed for sigwaiters:
*/
if (_thread_dfl_count[sig] == 0) {
/*
* Deliver the signal to the process if a handler
* is not installed:
*/
THR_SCHED_UNLOCK(curthread, pthread);
kill(getpid(), sig);
THR_SCHED_LOCK(curthread, pthread);
}
/*
* Assuming we're still running after the above kill(),
* make any necessary state changes to the thread:
*/
thr_sig_check_state(pthread, sig);
THR_SCHED_UNLOCK(curthread, pthread);
}
/*
* Check that the signal is not being ignored:
*/
else if (_thread_sigact[sig - 1].sa_handler != SIG_IGN) {
if (pthread->state == PS_SIGWAIT &&
sigismember(pthread->data.sigwait, sig)) {
/* Return the signal number: */
pthread->signo = sig;
/* Change the state of the thread to run: */
_thr_setrunnable_unlocked(pthread);
THR_SCHED_UNLOCK(curthread, pthread);
} else if (sigismember(&pthread->tmbx.tm_context.uc_sigmask, sig)) {
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/* Add the signal to the pending set: */
sigaddset(&pthread->sigpend, sig);
THR_SCHED_UNLOCK(curthread, pthread);
} else if (pthread == curthread) {
ucontext_t uc;
siginfo_t info;
volatile int once;
THR_SCHED_UNLOCK(curthread, pthread);
build_siginfo(&info, sig);
once = 0;
THR_GETCONTEXT(&uc);
if (once == 0) {
once = 1;
/*
* Call the signal handler for the current
* thread:
*/
thr_sig_invoke_handler(curthread, sig,
&info, &uc);
}
} else {
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Perform any state changes due to signal
* arrival:
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
_thr_sig_add(pthread, sig, NULL);
THR_SCHED_UNLOCK(curthread, pthread);
}
}
}
static void
thr_sigframe_add(struct pthread *thread, int sig, siginfo_t *info)
{
if (thread->curframe == NULL)
PANIC("Thread doesn't have signal frame ");
if (thread->have_signals == 0) {
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
/*
* Multiple signals can be added to the same signal
* frame. Only save the thread's state the first time.
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
*/
thr_sigframe_save(thread, thread->curframe);
thread->have_signals = 1;
thread->flags &= THR_FLAGS_PRIVATE;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
sigaddset(&thread->curframe->psf_sigset, sig);
if (info == NULL)
build_siginfo(&thread->siginfo[sig], sig);
else if (info != &thread->siginfo[sig])
memcpy(&thread->siginfo[sig], info, sizeof(*info));
/* Setup the new signal mask. */
SIGSETOR(thread->tmbx.tm_context.uc_sigmask,
_thread_sigact[sig - 1].sa_mask);
sigaddset(&thread->tmbx.tm_context.uc_sigmask, sig);
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
void
thr_sigframe_restore(struct pthread *thread, struct pthread_sigframe *psf)
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
{
thread->flags = psf->psf_flags;
thread->interrupted = psf->psf_interrupted;
thread->signo = psf->psf_signo;
thread->state = psf->psf_state;
thread->data = psf->psf_wait_data;
thread->wakeup_time = psf->psf_wakeup_time;
if (thread->sigmask_seqno == psf->psf_seqno)
thread->tmbx.tm_context.uc_sigmask = psf->psf_sigmask;
else
thread->tmbx.tm_context.uc_sigmask = thread->sigmask;
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
}
static void
thr_sigframe_save(struct pthread *thread, struct pthread_sigframe *psf)
Implement zero system call thread switching. Performance of thread switches should be on par with that under scheduler activations. o Timing is achieved through the use of a fixed interval timer (ITIMER_PROF) to count scheduling ticks instead of retrieving the time-of-day upon every thread switch and calculating elapsed real time. o Polling for I/O readiness is performed once for each scheduling tick instead of every thread switch. o The non-signal saving/restoring versions of setjmp/longjmp are used to save and restore thread contexts. This may allow the removal of _THREAD_SAFE macros from setjmp() and longjmp() - needs more investigation. Change signal handling so that signals are handled in the context of the thread that is receiving the signal. When signals are dispatched to a thread, a special signal handling frame is created on top of the target threads stack. The frame contains the threads saved state information and a new context in which the thread can run. The applications signal handler is invoked through a wrapper routine that knows how to restore the threads saved state and unwind to previous frames. Fix interruption of threads due to signals. Some states were being improperly interrupted while other states were not being interrupted. This should fix several PRs. Signal handlers, which are invoked as a result of a process signal (not by pthread_kill()), are now called with the code (or siginfo_t if SA_SIGINFO was set in sa_flags) and sigcontext_t as received from the process signal handler. Modify the search for a thread to which a signal is delivered. The search algorithm is now: o First thread found in sigwait() with signal in wait mask. o First thread found sigsuspend()'d on the signal. o Current thread if signal is unmasked. o First thread found with signal unmasked. Collapse machine dependent support into macros defined in pthread_private.h. These should probably eventually be moved into separate MD files. Change the range of settable priorities to be compliant with POSIX (0-31). The threads library uses higher priorities internally for real-time threads (not yet implemented) and threads executing signal handlers. Real-time threads and threads running signal handlers add 64 and 32, respectively, to a threads base priority. Some other small changes and cleanups. PR: 17757 18559 21943 Reviewed by: jasone
2000-10-13 22:12:32 +00:00
{
/* This has to initialize all members of the sigframe. */
psf->psf_flags = thread->flags & THR_FLAGS_PRIVATE;
psf->psf_interrupted = thread->interrupted;
psf->psf_signo = thread->signo;
psf->psf_state = thread->state;
psf->psf_wait_data = thread->data;
psf->psf_wakeup_time = thread->wakeup_time;
psf->psf_sigmask = thread->tmbx.tm_context.uc_sigmask;
psf->psf_seqno = thread->sigmask_seqno;
sigemptyset(&psf->psf_sigset);
}