d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
507b03186a
freebsd-nq/lib/libc_r/uthread/pthread_private.h

1403 lines
38 KiB
C
Raw Normal View History

Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
Add a magic field to the pthread structure to help recognize valid threads from invalid ones. The pthread structure is opaque to the user so this change does not cause any incompatibilities. Hopefully this change will help code that was written for draft 4 fail gracefully if the programmer ignores the compiler warning about the change in the level of indirection for the argument passed to pthread_detach(). I got burnt, so I fixed then (expletive deleted) thing. These functions comply with the revised standard. That should shut Terry up!
1998-04-03 09:31:15 +00:00
* Copyright (c) 1995-1998 John Birrell <jb@cimlogic.com.au>.
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
* 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
Fix thread initialization to allow for the case where stdio file descriptors are not opened. PR: bin/12853 Reviewed by: jb
1999-08-05 12:08:10 +00:00
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
* 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.
*
* Private thread definitions for the uthread kernel.
*
$Id$ -> $FreeBSD$
1999-08-28 00:22:10 +00:00
* $FreeBSD$
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
*/
#ifndef _PTHREAD_PRIVATE_H
#define _PTHREAD_PRIVATE_H
/*
* Evaluate the storage class specifier.
*/
#ifdef GLOBAL_PTHREAD_PRIVATE
#define SCLASS
#else
#define SCLASS extern
#endif
/*
* Include files.
*/
#include <setjmp.h>
#include <signal.h>
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
#include <stdio.h>
Change the FILE locking to be by FILE, not by the underlying fd as it was. Add a FILE_WAIT state and queue threads waiting for a FILE lock. Start using the sys/queue.h macros instead of the way that MIT pthreads did it. Add a thread name to the private thread structure and a non-POSIX function to set this. This helps (me at least) when sending a SIGINFO to a threaded process to get a /tmp/uthread.dump to see what the <expletive deleted> threads are doing this time. It is nice to be able to recognise (yes, I spell that with an 's' too) which threads are which.
1998-04-11 07:47:22 +00:00
#include <sys/queue.h>
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#include <sys/types.h>
#include <sys/time.h>
Explicitly include sys/cdefs.h to get the definition of __strong_reference(), rather than getting lucky due to header dependencies.
2000-03-18 22:36:46 +00:00
#include <sys/cdefs.h>
Add sched_yield() witch is the draft 10 equivalent of pthread_yield() from draft 4. Move some of the schedule definitions to sched.h which is a POSIX header.
1998-03-08 02:37:27 +00:00
#include <sched.h>
Add compile time thread lock debug support. Add a thread specific flag to trap the case where pthread_exit() is called from a destructor in violation of the Posix standard.
1998-06-09 23:02:43 +00:00
#include <spinlock.h>
Don't rely on <sys/signal.h> to include <sys/ucontext.h>
2002-02-17 17:21:27 +00:00
#include <ucontext.h>
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
#include <pthread_np.h>
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +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
/*
* Define machine dependent macros to get and set the stack pointer
* from the supported contexts. Also define a macro to set the return
* address in a jmp_buf context.
*
* XXX - These need to be moved into architecture dependent support files.
Add MD definitions for sparc64. This is based on work by des.
2003-01-05 22:37:03 +00:00
* XXX - These need to be documented so porters know what's required.
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
*/
#if defined(__i386__)
#define GET_STACK_JB(jb) ((unsigned long)((jb)[0]._jb[2]))
#define GET_STACK_SJB(sjb) ((unsigned long)((sjb)[0]._sjb[2]))
#define GET_STACK_UC(ucp) ((unsigned long)((ucp)->uc_mcontext.mc_esp))
#define SET_STACK_JB(jb, stk) (jb)[0]._jb[2] = (int)(stk)
#define SET_STACK_SJB(sjb, stk) (sjb)[0]._sjb[2] = (int)(stk)
#define SET_STACK_UC(ucp, stk) (ucp)->uc_mcontext.mc_esp = (int)(stk)
#define FP_SAVE_UC(ucp) do { \
char *fdata; \
Add back the i386-specific hack to save and restore the FP state to/from a ucontext when a thread is interrupted by a signal. This will be removed when a proper fix is made in the kernel to save/restore the FP state without breaking the ABI.
2002-09-30 08:45:38 +00:00
fdata = (char *) (ucp)->uc_mcontext.mc_fpstate; \
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
__asm__("fnsave %0": :"m"(*fdata)); \
} while (0)
#define FP_RESTORE_UC(ucp) do { \
char *fdata; \
Add back the i386-specific hack to save and restore the FP state to/from a ucontext when a thread is interrupted by a signal. This will be removed when a proper fix is made in the kernel to save/restore the FP state without breaking the ABI.
2002-09-30 08:45:38 +00:00
fdata = (char *) (ucp)->uc_mcontext.mc_fpstate; \
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
__asm__("frstor %0": :"m"(*fdata)); \
} while (0)
Initialize the (i386) frame pointer when setting up a thread context. Submitted by: Marc Olzheim <marcolz@stack.nl> Tested by: Marc Olzheim <marcolz@stack.nl>
2004-01-22 19:15:08 +00:00
#define SET_RETURN_ADDR_JB(jb, ra) do { \
(jb)[0]._jb[0] = (int)(ra); \
(jb)[0]._jb[3] = 0; \
} while (0)
Port libc_r to amd64, and turn it back on for amd64. It passes all of the same src/lib/libc_r/test/* tests that the other platforms pass.
2003-06-02 22:22:52 +00:00
#elif defined(__amd64__)
#define GET_STACK_JB(jb) ((unsigned long)((jb)[0]._jb[2]))
#define GET_STACK_SJB(sjb) ((unsigned long)((sjb)[0]._sjb[2]))
#define GET_STACK_UC(ucp) ((unsigned long)((ucp)->uc_mcontext.mc_rsp))
#define SET_STACK_JB(jb, stk) (jb)[0]._jb[2] = (long)(stk)
#define SET_STACK_SJB(sjb, stk) (sjb)[0]._sjb[2] = (long)(stk)
#define SET_STACK_UC(ucp, stk) (ucp)->uc_mcontext.mc_rsp = (long)(stk)
#define FP_SAVE_UC(ucp) do { \
char *fdata; \
fdata = (char *) (ucp)->uc_mcontext.mc_fpstate; \
__asm__("fxsave %0": :"m"(*fdata)); \
} while (0)
#define FP_RESTORE_UC(ucp) do { \
char *fdata; \
fdata = (char *) (ucp)->uc_mcontext.mc_fpstate; \
__asm__("fxrstor %0": :"m"(*fdata)); \
} while (0)
#define SET_RETURN_ADDR_JB(jb, ra) (jb)[0]._jb[0] = (long)(ra)
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
#elif defined(__alpha__)
#include <machine/reg.h>
Try and get libc_r to compile again on the alpha after deischen's commit
2000-10-17 06:31:40 +00:00
#define GET_STACK_JB(jb) ((unsigned long)((jb)[0]._jb[R_SP + 4]))
#define GET_STACK_SJB(sjb) ((unsigned long)((sjb)[0]._sjb[R_SP + 4]))
#define GET_STACK_UC(ucp) ((ucp)->uc_mcontext.mc_regs[R_SP])
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 SET_STACK_JB(jb, stk) (jb)[0]._jb[R_SP + 4] = (long)(stk)
#define SET_STACK_SJB(sjb, stk) (sjb)[0]._sjb[R_SP + 4] = (long)(stk)
#define SET_STACK_UC(ucp, stk) (ucp)->uc_mcontext.mc_regs[R_SP] = (unsigned long)(stk)
#define FP_SAVE_UC(ucp)
#define FP_RESTORE_UC(ucp)
Fix MD macros to work for alpha. Without this fix, threads under alpha seem to be totally broke. MFC Candidate Submitted by: gallatin
2000-11-20 01:57:19 +00:00
#define SET_RETURN_ADDR_JB(jb, ra) do { \
Use the real function address (instead of function address + 8) for the return address when modifying a jmp_buf to create a new thread context. Also set t12 with the return address. This should fix libc_r on alpha. With much detective work by: Bernd Walter <ticso@cicely.de>
2001-12-22 06:11:06 +00:00
(jb)[0]._jb[2] = (long)(ra); \
(jb)[0]._jb[R_RA + 4] = (long)(ra); \
Fix MD macros to work for alpha. Without this fix, threads under alpha seem to be totally broke. MFC Candidate Submitted by: gallatin
2000-11-20 01:57:19 +00:00
(jb)[0]._jb[R_T12 + 4] = (long)(ra); \
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
} while (0)
Port libc_r to ia64. We need to do things slightly different because we have 2 stacks per thread: the regular downward memory stack and the irregular upward register stack. This implementation lets both stacks grow toward each other. An alternative scheme is to have them grow away from each other. The alternate scheme has the advantage that both stack grow toward guard pages. Since libc_r is virtually dead and we really want the *context stuff for thread switching, we don't try to be perfect, just functional.
2003-01-06 00:56:23 +00:00
#elif defined(__ia64__)
#define GET_BSP_JB(jb) (*((unsigned long*)JMPBUF_ADDR_OF(jb,J_BSP)))
#define GET_STACK_JB(jb) (*((unsigned long*)JMPBUF_ADDR_OF(jb,J_SP)))
#define GET_STACK_SJB(sjb) GET_STACK_JB(sjb)
#define SET_RETURN_ADDR_JB(jb, ra) \
do { \
*((unsigned long*)JMPBUF_ADDR_OF(jb,J_B0)) = ((long*)(ra))[0]; \
Fix threaded applications on ia64 that are linked dynamicly. We did not save (restore) the global pointer (GP) in the jmpbuf in setjmp (longjmp) because it's not needed in general. GP is considered a scratch register at callsites and hence is always restored after a call (when it's possible that the call resolves to a symbol in a different loadmodule; otherwise GP does not have to be saved and restored at all), including calls to setjmp/longjmp. There's just one problem with this now that we use setjmp/longjmp for context switching: A new context must have GP defined properly for the thread's entry point. This means that we need to put GP in the jmpbuf and consequently that we have to restore is in longjmp. This automaticly requires us to save it as well. When setjmp/longjmp isn't used for context switching, this can be reverted again.
2003-03-05 04:39:24 +00:00
*((unsigned long*)JMPBUF_ADDR_OF(jb,J_GP)) = ((long*)(ra))[1]; \
Port libc_r to ia64. We need to do things slightly different because we have 2 stacks per thread: the regular downward memory stack and the irregular upward register stack. This implementation lets both stacks grow toward each other. An alternative scheme is to have them grow away from each other. The alternate scheme has the advantage that both stack grow toward guard pages. Since libc_r is virtually dead and we really want the *context stuff for thread switching, we don't try to be perfect, just functional.
2003-01-06 00:56:23 +00:00
*((unsigned long*)JMPBUF_ADDR_OF(jb,J_PFS)) &= ~0x1FFFFFFFFFUL; \
} while (0)
#define SET_STACK_JB(jb, stk, sz) \
do { \
UPD_STACK_JB(jb, stk + sz - 16); \
GET_BSP_JB(jb) = (long)(stk); \
} while (0)
#define UPD_STACK_JB(jb, stk) GET_STACK_JB(jb) = (long)(stk)
Add MD definitions for sparc64. This is based on work by des.
2003-01-05 22:37:03 +00:00
#elif defined(__sparc64__)
#include <machine/frame.h>
#define CCFSZ sizeof (struct frame)
#define GET_STACK_JB(jb) \
((unsigned long)((jb)[0]._jb[_JB_SP]) + SPOFF)
#define GET_STACK_SJB(sjb) \
((unsigned long)((sjb)[0]._sjb[_JB_SP]) + SPOFF)
#define GET_STACK_UC(ucp) \
((ucp)->uc_mcontext.mc_sp + SPOFF)
/*
* XXX: sparc64 _longjmp() expects a register window on the stack
* at the given position, so we must make sure that the space where
* it is expected is readable. Subtracting the frame size here works
* because the SET_STACK macros are only used to set up new stacks
* or signal stacks, but it is a bit dirty.
*/
#define SET_STACK_JB(jb, stk) \
(jb)[0]._jb[_JB_SP] = (long)(stk) - SPOFF - CCFSZ
#define SET_STACK_SJB(sjb, stk) \
(sjb)[0]._sjb[_JB_SP] = (long)(stk) - SPOFF - CCFSZ
#define SET_STACK_UC(ucp, stk) \
(ucp)->uc_mcontext.mc_sp = (unsigned long)(stk) - SPOFF - CCFSZ
#define FP_SAVE_UC(ucp) /* XXX */
#define FP_RESTORE_UC(ucp) /* XXX */
#define SET_RETURN_ADDR_JB(jb, ra) \
(jb)[0]._jb[_JB_PC] = (long)(ra) - 8
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
#error "Don't recognize this architecture!"
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Kernel fatal error handler macro.
*/
#define PANIC(string) _thread_exit(__FILE__,__LINE__,string)
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
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
/* Output debug messages like this: */
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 stdout_debug(args...) do { \
char buf[128]; \
snprintf(buf, sizeof(buf), ##args); \
s/_thread_sys_write/__sys_write/ Submitted by: Mike Heffner <mheffner@vt.edu>
2001-01-31 02:16:57 +00:00
__sys_write(1, buf, strlen(buf)); \
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
} while (0)
#define stderr_debug(args...) do { \
char buf[128]; \
snprintf(buf, sizeof(buf), ##args); \
s/_thread_sys_write/__sys_write/ Submitted by: Mike Heffner <mheffner@vt.edu>
2001-01-31 02:16:57 +00:00
__sys_write(2, buf, strlen(buf)); \
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
} while (0)
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
/*
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
* Priority queue manipulation macros (using pqe link):
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
*/
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
#define PTHREAD_PRIOQ_INSERT_HEAD(thrd) _pq_insert_head(&_readyq,thrd)
#define PTHREAD_PRIOQ_INSERT_TAIL(thrd) _pq_insert_tail(&_readyq,thrd)
#define PTHREAD_PRIOQ_REMOVE(thrd) _pq_remove(&_readyq,thrd)
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
#define PTHREAD_PRIOQ_FIRST() _pq_first(&_readyq)
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
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
* Waiting queue manipulation macros (using pqe link):
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
*/
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
#define PTHREAD_WAITQ_REMOVE(thrd) _waitq_remove(thrd)
#define PTHREAD_WAITQ_INSERT(thrd) _waitq_insert(thrd)
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
#if defined(_PTHREADS_INVARIANTS)
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
#define PTHREAD_WAITQ_CLEARACTIVE() _waitq_clearactive()
#define PTHREAD_WAITQ_SETACTIVE() _waitq_setactive()
#else
#define PTHREAD_WAITQ_CLEARACTIVE()
#define PTHREAD_WAITQ_SETACTIVE()
#endif
/*
* Work queue manipulation macros (using qe link):
*/
#define PTHREAD_WORKQ_INSERT(thrd) do { \
TAILQ_INSERT_TAIL(&_workq,thrd,qe); \
(thrd)->flags |= PTHREAD_FLAGS_IN_WORKQ; \
} while (0)
#define PTHREAD_WORKQ_REMOVE(thrd) do { \
TAILQ_REMOVE(&_workq,thrd,qe); \
(thrd)->flags &= ~PTHREAD_FLAGS_IN_WORKQ; \
} while (0)
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* State change macro without scheduling queue change:
*/
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
#define PTHREAD_SET_STATE(thrd, newstate) do { \
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
(thrd)->state = newstate; \
(thrd)->fname = __FILE__; \
(thrd)->lineno = __LINE__; \
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
} while (0)
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* State change macro with scheduling queue change - This must be
* called with preemption deferred (see thread_kern_sched_[un]defer).
*/
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
#if defined(_PTHREADS_INVARIANTS)
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
#include <assert.h>
#define PTHREAD_ASSERT(cond, msg) do { \
if (!(cond)) \
PANIC(msg); \
} while (0)
#define PTHREAD_ASSERT_NOT_IN_SYNCQ(thrd) \
PTHREAD_ASSERT((((thrd)->flags & PTHREAD_FLAGS_IN_SYNCQ) == 0), \
"Illegal call from signal handler");
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
#define PTHREAD_NEW_STATE(thrd, newstate) do { \
if (_thread_kern_new_state != 0) \
PANIC("Recursive PTHREAD_NEW_STATE"); \
_thread_kern_new_state = 1; \
if ((thrd)->state != newstate) { \
if ((thrd)->state == PS_RUNNING) { \
PTHREAD_PRIOQ_REMOVE(thrd); \
Revamp suspend and resume. While I'm here add pthread_suspend_all_np() and pthread_resume_all_np(). These suspend and resume all threads except the current thread, respectively. The existing functions pthread_single_np() and pthread_multi_np(), which formerly had no effect, now exhibit the same behaviour and pthread_suspend_all_np() and pthread_resume_all_np(). These functions have been added mostly for the native java port. Don't allow the uthread kernel pipe to use the same descriptors as stdio. Mostily submitted by Oswald Buddenhagen <ossi@kde.org>. Correct some minor style nits.
2002-05-24 04:32:28 +00:00
PTHREAD_SET_STATE(thrd, newstate); \
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_WAITQ_INSERT(thrd); \
} else if (newstate == PS_RUNNING) { \
PTHREAD_WAITQ_REMOVE(thrd); \
Revamp suspend and resume. While I'm here add pthread_suspend_all_np() and pthread_resume_all_np(). These suspend and resume all threads except the current thread, respectively. The existing functions pthread_single_np() and pthread_multi_np(), which formerly had no effect, now exhibit the same behaviour and pthread_suspend_all_np() and pthread_resume_all_np(). These functions have been added mostly for the native java port. Don't allow the uthread kernel pipe to use the same descriptors as stdio. Mostily submitted by Oswald Buddenhagen <ossi@kde.org>. Correct some minor style nits.
2002-05-24 04:32:28 +00:00
PTHREAD_SET_STATE(thrd, newstate); \
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_PRIOQ_INSERT_TAIL(thrd); \
} \
} \
_thread_kern_new_state = 0; \
} while (0)
#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
#define PTHREAD_ASSERT(cond, msg)
#define PTHREAD_ASSERT_NOT_IN_SYNCQ(thrd)
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
#define PTHREAD_NEW_STATE(thrd, newstate) do { \
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
if ((thrd)->state != newstate) { \
if ((thrd)->state == PS_RUNNING) { \
PTHREAD_PRIOQ_REMOVE(thrd); \
PTHREAD_WAITQ_INSERT(thrd); \
} else if (newstate == PS_RUNNING) { \
PTHREAD_WAITQ_REMOVE(thrd); \
PTHREAD_PRIOQ_INSERT_TAIL(thrd); \
} \
} \
PTHREAD_SET_STATE(thrd, newstate); \
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
} while (0)
#endif
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* Define the signals to be used for scheduling.
*/
#if defined(_PTHREADS_COMPAT_SCHED)
#define _ITIMER_SCHED_TIMER ITIMER_VIRTUAL
#define _SCHED_SIGNAL SIGVTALRM
#else
#define _ITIMER_SCHED_TIMER ITIMER_PROF
#define _SCHED_SIGNAL SIGPROF
#endif
/*
* Priority queues.
*
* XXX It'd be nice if these were contained in uthread_priority_queue.[ch].
*/
typedef struct pq_list {
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_HEAD(, pthread) pl_head; /* list of threads at this priority */
TAILQ_ENTRY(pq_list) pl_link; /* link for queue of priority lists */
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int pl_prio; /* the priority of this list */
int pl_queued; /* is this in the priority queue */
} pq_list_t;
typedef struct pq_queue {
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_HEAD(, pq_list) pq_queue; /* queue of priority lists */
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
pq_list_t *pq_lists; /* array of all priority lists */
int pq_size; /* number of priority lists */
} pq_queue_t;
/*
* TailQ initialization values.
*/
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
#define TAILQ_INITIALIZER { NULL, NULL }
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
/*
* Mutex definitions.
*/
union pthread_mutex_data {
void *m_ptr;
int m_count;
};
struct pthread_mutex {
enum pthread_mutextype m_type;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int m_protocol;
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_HEAD(mutex_head, pthread) m_queue;
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
struct pthread *m_owner;
union pthread_mutex_data m_data;
long m_flags;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int m_refcount;
/*
* Used for priority inheritence and protection.
*
* m_prio - For priority inheritence, the highest active
* priority (threads locking the mutex inherit
* this priority). For priority protection, the
* ceiling priority of this mutex.
* m_saved_prio - mutex owners inherited priority before
* taking the mutex, restored when the owner
* unlocks the mutex.
*/
int m_prio;
int m_saved_prio;
/*
* Link for list of all mutexes a thread currently owns.
*/
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_ENTRY(pthread_mutex) m_qe;
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
/*
* Lock for accesses to this structure.
*/
Add compile time thread lock debug support. Add a thread specific flag to trap the case where pthread_exit() is called from a destructor in violation of the Posix standard.
1998-06-09 23:02:43 +00:00
spinlock_t lock;
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
};
/*
* Flags for mutexes.
*/
#define MUTEX_FLAGS_PRIVATE 0x01
#define MUTEX_FLAGS_INITED 0x02
#define MUTEX_FLAGS_BUSY 0x04
/*
* Static mutex initialization values.
*/
Rename static initializer defines for opaque structures so that the POSIX specified names can be declared in pthread.h.
1998-04-04 10:58:12 +00:00
#define PTHREAD_MUTEX_STATIC_INITIALIZER \
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
{ PTHREAD_MUTEX_DEFAULT, PTHREAD_PRIO_NONE, TAILQ_INITIALIZER, \
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
NULL, { NULL }, MUTEX_FLAGS_PRIVATE, 0, 0, 0, TAILQ_INITIALIZER, \
Fix a problem with static initialisation of mutexes and condition variables. Submitted by: Dan Eischen <eischen@vigrid.com>
1999-05-23 10:55:33 +00:00
_SPINLOCK_INITIALIZER }
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
struct pthread_mutex_attr {
enum pthread_mutextype m_type;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int m_protocol;
int m_ceiling;
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
long m_flags;
};
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
#define PTHREAD_MUTEXATTR_STATIC_INITIALIZER \
{ PTHREAD_MUTEX_DEFAULT, PTHREAD_PRIO_NONE, 0, MUTEX_FLAGS_PRIVATE }
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
/*
* Condition variable definitions.
*/
enum pthread_cond_type {
COND_TYPE_FAST,
COND_TYPE_MAX
};
struct pthread_cond {
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
enum pthread_cond_type c_type;
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_HEAD(cond_head, pthread) c_queue;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
pthread_mutex_t c_mutex;
void *c_data;
long c_flags;
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
int c_seqno;
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
/*
* Lock for accesses to this structure.
*/
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
spinlock_t lock;
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
};
struct pthread_cond_attr {
enum pthread_cond_type c_type;
long c_flags;
};
/*
* Flags for condition variables.
*/
#define COND_FLAGS_PRIVATE 0x01
#define COND_FLAGS_INITED 0x02
#define COND_FLAGS_BUSY 0x04
/*
* Static cond initialization values.
*/
Rename static initializer defines for opaque structures so that the POSIX specified names can be declared in pthread.h.
1998-04-04 10:58:12 +00:00
#define PTHREAD_COND_STATIC_INITIALIZER \
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
{ COND_TYPE_FAST, TAILQ_INITIALIZER, NULL, NULL, \
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
0, 0, _SPINLOCK_INITIALIZER }
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
Add sem_*() functions. Named semaphores and process-shared semaphores are not supported by this implementation, and the error return values from sem_init(), sem_open(), sem_close(), and sem_unlink() reflect this. Approved by: jkh
2000-01-20 07:54:49 +00:00
/*
* Semaphore definitions.
*/
struct sem {
#define SEM_MAGIC ((u_int32_t) 0x09fa4012)
u_int32_t magic;
pthread_mutex_t lock;
pthread_cond_t gtzero;
u_int32_t count;
u_int32_t nwaiters;
};
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
/*
* Cleanup definitions.
*/
struct pthread_cleanup {
struct pthread_cleanup *next;
void (*routine) ();
void *routine_arg;
};
struct pthread_attr {
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int sched_policy;
int sched_inherit;
int sched_interval;
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
int prio;
Add sched_yield() witch is the draft 10 equivalent of pthread_yield() from draft 4. Move some of the schedule definitions to sched.h which is a POSIX header.
1998-03-08 02:37:27 +00:00
int suspend;
int flags;
void *arg_attr;
void (*cleanup_attr) ();
void *stackaddr_attr;
size_t stacksize_attr;
Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to be malloc()ed, but they are now allocated using mmap(), just as the default-size stacks are. A separate cache of stacks is kept for non-default-size stacks. Collaboration with: deischen
2001-07-20 04:23:11 +00:00
size_t guardsize_attr;
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
};
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Thread creation state attributes.
*/
#define PTHREAD_CREATE_RUNNING 0
#define PTHREAD_CREATE_SUSPENDED 1
/*
* Miscellaneous definitions.
*/
Port libc_r to ia64. We need to do things slightly different because we have 2 stacks per thread: the regular downward memory stack and the irregular upward register stack. This implementation lets both stacks grow toward each other. An alternative scheme is to have them grow away from each other. The alternate scheme has the advantage that both stack grow toward guard pages. Since libc_r is virtually dead and we really want the *context stuff for thread switching, we don't try to be perfect, just functional.
2003-01-06 00:56:23 +00:00
#if !defined(__ia64__)
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#define PTHREAD_STACK_DEFAULT 65536
Port libc_r to ia64. We need to do things slightly different because we have 2 stacks per thread: the regular downward memory stack and the irregular upward register stack. This implementation lets both stacks grow toward each other. An alternative scheme is to have them grow away from each other. The alternate scheme has the advantage that both stack grow toward guard pages. Since libc_r is virtually dead and we really want the *context stuff for thread switching, we don't try to be perfect, just functional.
2003-01-06 00:56:23 +00:00
#else
#define PTHREAD_STACK_DEFAULT 0x40000
#endif
Don't explicitly mmap() red zones at the bottom of thread stacks (except the initial thread). Instead, just leave an unmapped gap between thread stacks and make sure that the thread stacks won't grow into these gaps, simply by limiting the size of the stacks with the 'len' argument to mmap(). This (if I understand correctly) reduces VM overhead considerably. Reviewed by: deischen
1999-12-29 15:44:59 +00:00
/*
Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to be malloc()ed, but they are now allocated using mmap(), just as the default-size stacks are. A separate cache of stacks is kept for non-default-size stacks. Collaboration with: deischen
2001-07-20 04:23:11 +00:00
* Size of default red zone at the end of each stack. In actuality, this "red
* zone" is merely an unmapped region, except in the case of the initial stack.
* Since mmap() makes it possible to specify the maximum growth of a MAP_STACK
* region, an unmapped gap between thread stacks achieves the same effect as
* explicitly mapped red zones.
Don't use PAGE_SIZE in userland, instead use getpagesize(), this is to allow running on other arches when the instructions are supported but the page size granularity is not. Glanced at by: peter
2002-05-13 07:58:15 +00:00
* This is declared and initialized in uthread_init.c.
Don't explicitly mmap() red zones at the bottom of thread stacks (except the initial thread). Instead, just leave an unmapped gap between thread stacks and make sure that the thread stacks won't grow into these gaps, simply by limiting the size of the stacks with the 'len' argument to mmap(). This (if I understand correctly) reduces VM overhead considerably. Reviewed by: deischen
1999-12-29 15:44:59 +00:00
*/
Undo namespace pollution by prefixing the globals pthread_guard_default and pthread_page_size. Fix a bunch line wrapping. Pointed out by: deischen
2002-05-15 05:37:48 +00:00
extern int _pthread_guard_default;
Don't use PAGE_SIZE in userland, instead use getpagesize(), this is to allow running on other arches when the instructions are supported but the page size granularity is not. Glanced at by: peter
2002-05-13 07:58:15 +00:00
Undo namespace pollution by prefixing the globals pthread_guard_default and pthread_page_size. Fix a bunch line wrapping. Pointed out by: deischen
2002-05-15 05:37:48 +00:00
extern int _pthread_page_size;
* Fix the stack allocation code so that it works for alpha. Change it to use mmap(..., MAP_STACK, ...) on alpha too since that should work now. * Add hooks to allow GDB to access the internals of pthreads without having to know the exact layout of struct pthread. Reviewed by: deischen
1999-11-28 19:47:43 +00:00
Modify previous changes to conform better to libc_r's coding style. Always use mmap() for default-size stack allocation. Use MAP_ANON instead of MAP_STACK on the alpha architecture. Reduce the amount of code executed while owning _gc_mutex during stack allocation.
1999-07-11 05:56:37 +00:00
/*
* Maximum size of initial thread's stack. This perhaps deserves to be larger
Always use growable thread stacks on the i386. The VM_STACK kernel option must be made default for the alpha before growable thread stacks are enabled for the alpha.
1999-07-06 00:25:38 +00:00
* than the stacks of other threads, since many applications are likely to run
Modify previous changes to conform better to libc_r's coding style. Always use mmap() for default-size stack allocation. Use MAP_ANON instead of MAP_STACK on the alpha architecture. Reduce the amount of code executed while owning _gc_mutex during stack allocation.
1999-07-11 05:56:37 +00:00
* almost entirely on this stack.
*/
Port libc_r to ia64. We need to do things slightly different because we have 2 stacks per thread: the regular downward memory stack and the irregular upward register stack. This implementation lets both stacks grow toward each other. An alternative scheme is to have them grow away from each other. The alternate scheme has the advantage that both stack grow toward guard pages. Since libc_r is virtually dead and we really want the *context stuff for thread switching, we don't try to be perfect, just functional.
2003-01-06 00:56:23 +00:00
#if !defined(__ia64__)
Use growable stacks for thread stacks that are the default stack size. Cache discarded default thread stacks for use in subsequent thread creations. Create a red zone at the end of each stack (including the initial thread stack), with the hope of causing a segfault if a stack overflows. To activate these modifications, add -D_PTHREAD_GSTACK to CFLAGS in src/lib/libc_r/Makefile. Since the modifications depend on the VM_STACK kernel option, I'm not sure how to safely use growable stacks by default. Testing, as well as algorithmic and stylistic comments are welcome.
1999-07-05 00:35:19 +00:00
#define PTHREAD_STACK_INITIAL 0x100000
Port libc_r to ia64. We need to do things slightly different because we have 2 stacks per thread: the regular downward memory stack and the irregular upward register stack. This implementation lets both stacks grow toward each other. An alternative scheme is to have them grow away from each other. The alternate scheme has the advantage that both stack grow toward guard pages. Since libc_r is virtually dead and we really want the *context stuff for thread switching, we don't try to be perfect, just functional.
2003-01-06 00:56:23 +00:00
#else
#define PTHREAD_STACK_INITIAL 0x400000
#endif
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 the different priority ranges. All applications have thread
* priorities constrained within 0-31. The threads library raises the
* priority when delivering signals in order to ensure that signal
* delivery happens (from the POSIX spec) "as soon as possible".
* In the future, the threads library will also be able to map specific
* threads into real-time (cooperating) processes or kernel threads.
* The RT and SIGNAL priorities will be used internally and added to
* thread base priorities so that the scheduling queue can handle both
* normal and RT priority threads with and without signal handling.
*
* The approach taken is that, within each class, signal delivery
* always has priority over thread execution.
*/
#define PTHREAD_DEFAULT_PRIORITY 15
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#define PTHREAD_MIN_PRIORITY 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
#define PTHREAD_MAX_PRIORITY 31 /* 0x1F */
#define PTHREAD_SIGNAL_PRIORITY 32 /* 0x20 */
#define PTHREAD_RT_PRIORITY 64 /* 0x40 */
#define PTHREAD_FIRST_PRIORITY PTHREAD_MIN_PRIORITY
#define PTHREAD_LAST_PRIORITY \
(PTHREAD_MAX_PRIORITY + PTHREAD_SIGNAL_PRIORITY + PTHREAD_RT_PRIORITY)
#define PTHREAD_BASE_PRIORITY(prio) ((prio) & PTHREAD_MAX_PRIORITY)
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +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
* Clock resolution in microseconds.
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +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
#define CLOCK_RES_USEC 10000
Limit threads clock resolution to no less than 1000usec (1000Hz). PR: 25300 Submitted by: Tom Pavel <pavel@alum.mit.edu> (in part)
2001-02-26 01:05:33 +00:00
#define CLOCK_RES_USEC_MIN 1000
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Time slice period in microseconds.
*/
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 TIMESLICE_USEC 20000
/*
* Define a thread-safe macro to get the current time of day
* which is updated at regular intervals by the scheduling signal
* handler.
*/
#define GET_CURRENT_TOD(tv) \
do { \
tv.tv_sec = _sched_tod.tv_sec; \
tv.tv_usec = _sched_tod.tv_usec; \
} while (tv.tv_sec != _sched_tod.tv_sec)
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
Implement pthread read/write locks as defined by Version 2 of the Single UNIX Specification. As with our standard mutexes, process shared locks are not supported at this time.
1998-09-07 19:01:43 +00:00
struct pthread_rwlockattr {
int pshared;
};
struct pthread_rwlock {
pthread_mutex_t lock; /* monitor lock */
int state; /* 0 = idle >0 = # of readers -1 = writer */
pthread_cond_t read_signal;
pthread_cond_t write_signal;
int blocked_writers;
};
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Thread states.
*/
enum pthread_state {
PS_RUNNING,
PS_SIGTHREAD,
PS_MUTEX_WAIT,
PS_COND_WAIT,
PS_FDLR_WAIT,
PS_FDLW_WAIT,
PS_FDR_WAIT,
PS_FDW_WAIT,
Change the FILE locking to be by FILE, not by the underlying fd as it was. Add a FILE_WAIT state and queue threads waiting for a FILE lock. Start using the sys/queue.h macros instead of the way that MIT pthreads did it. Add a thread name to the private thread structure and a non-POSIX function to set this. This helps (me at least) when sending a SIGINFO to a threaded process to get a /tmp/uthread.dump to see what the <expletive deleted> threads are doing this time. It is nice to be able to recognise (yes, I spell that with an 's' too) which threads are which.
1998-04-11 07:47:22 +00:00
PS_FILE_WAIT,
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
PS_POLL_WAIT,
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
PS_SELECT_WAIT,
PS_SLEEP_WAIT,
PS_WAIT_WAIT,
- Fix the debug macros. - Add support of a thread being listed in the dead thread list as well as the thread list. - Add a new thread state to make sigwait work properly. (Submitted by Daniel M. Eischen <eischen@vigrid.com>) - Add global variable for the garbage collector mutex and condition variable. - Delete a couple of prototypes that are no longer required. - Add a prototype for the garbage collector thread.
1998-09-30 06:22:07 +00:00
PS_SIGSUSPEND,
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
PS_SIGWAIT,
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
PS_SPINBLOCK,
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
PS_JOIN,
PS_SUSPENDED,
PS_DEAD,
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
PS_DEADLOCK,
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
PS_STATE_MAX
};
/*
* File descriptor locking definitions.
*/
#define FD_READ 0x1
#define FD_WRITE 0x2
#define FD_RDWR (FD_READ | FD_WRITE)
/*
* File descriptor table structure.
*/
struct fd_table_entry {
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
/*
* Lock for accesses to this file descriptor table
* entry. This is passed to _spinlock() to provide atomic
* access to this structure. It does *not* represent the
* state of the lock on the file descriptor.
*/
Add compile time thread lock debug support. Add a thread specific flag to trap the case where pthread_exit() is called from a destructor in violation of the Posix standard.
1998-06-09 23:02:43 +00:00
spinlock_t lock;
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_HEAD(, pthread) r_queue; /* Read queue. */
TAILQ_HEAD(, pthread) w_queue; /* Write queue. */
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
struct pthread *r_owner; /* Ptr to thread owning read lock. */
struct pthread *w_owner; /* Ptr to thread owning write lock. */
char *r_fname; /* Ptr to read lock source file name */
int r_lineno; /* Read lock source line number. */
char *w_fname; /* Ptr to write lock source file name */
int w_lineno; /* Write lock source line number. */
int r_lockcount; /* Count for FILE read locks. */
int w_lockcount; /* Count for FILE write locks. */
int flags; /* Flags used in open. */
};
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
struct pthread_poll_data {
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
int nfds;
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
struct pollfd *fds;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
};
union pthread_wait_data {
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
pthread_mutex_t mutex;
pthread_cond_t cond;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
const sigset_t *sigwait; /* Waiting on a signal in sigwait */
struct {
short fd; /* Used when thread waiting on fd */
short branch; /* Line number, for debugging. */
char *fname; /* Source file name for debugging.*/
} fd;
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
FILE *fp;
struct pthread_poll_data *poll_data;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
spinlock_t *spinlock;
Fix a bug in canceling joining threads. Do not detach canceled threads. Reported by: Arno Klaassen <arno@heho.snv.jussieu.fr> Collaboration with: deischen
2001-08-16 06:31:32 +00:00
struct pthread *thread;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
};
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
/*
* Define a continuation routine that can be used to perform a
* transfer of control:
*/
typedef void (*thread_continuation_t) (void *);
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
struct pthread_signal_frame;
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_state_data {
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
struct pthread_signal_frame *psd_curframe;
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 psd_sigmask;
struct timespec psd_wakeup_time;
union pthread_wait_data psd_wait_data;
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
enum pthread_state psd_state;
int psd_flags;
int psd_interrupted;
int psd_longjmp_val;
int psd_sigmask_seqno;
int psd_signo;
int psd_sig_defer_count;
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
/* XXX - What about thread->timeout and/or thread->error? */
};
Fix pthread_join so that it works if the target thread exits while the joining thread is in a signal handler. Reported by: Loren James Rittle <rittle@labs.mot.com> MFC after: 1 week
2001-11-17 14:28:39 +00:00
struct join_status {
struct pthread *thread;
This has been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
void *ret;
Fix pthread_join so that it works if the target thread exits while the joining thread is in a signal handler. Reported by: Loren James Rittle <rittle@labs.mot.com> MFC after: 1 week
2001-11-17 14:28:39 +00:00
int error;
};
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 been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
* The frame that is added to the top of a threads stack when setting up
* up the thread to run a signal handler.
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_signal_frame {
/*
* This stores the threads state before the signal.
*/
struct pthread_state_data saved_state;
/*
This has been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
* Threads return context; we use only jmp_buf's for now.
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
*/
union {
jmp_buf jb;
ucontext_t uc;
} ctx;
int signo; /* signal, arg 1 to sighandler */
int sig_has_args; /* use signal args if true */
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
ucontext_t uc;
siginfo_t siginfo;
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 ability to recognize old references to keys, and return NULL when old keys are referenced (after pthread_key_delete()) via pthread_getspecific().
2002-03-19 22:58:56 +00:00
struct pthread_specific_elem {
const void *data;
int seqno;
};
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Thread structure.
*/
struct pthread {
Add a magic field to the pthread structure to help recognize valid threads from invalid ones. The pthread structure is opaque to the user so this change does not cause any incompatibilities. Hopefully this change will help code that was written for draft 4 fail gracefully if the programmer ignores the compiler warning about the change in the level of indirection for the argument passed to pthread_detach(). I got burnt, so I fixed then (expletive deleted) thing. These functions comply with the revised standard. That should shut Terry up!
1998-04-03 09:31:15 +00:00
/*
* Magic value to help recognize a valid thread structure
* from an invalid one:
*/
#define PTHREAD_MAGIC ((u_int32_t) 0xd09ba115)
u_int32_t magic;
Change the FILE locking to be by FILE, not by the underlying fd as it was. Add a FILE_WAIT state and queue threads waiting for a FILE lock. Start using the sys/queue.h macros instead of the way that MIT pthreads did it. Add a thread name to the private thread structure and a non-POSIX function to set this. This helps (me at least) when sending a SIGINFO to a threaded process to get a /tmp/uthread.dump to see what the <expletive deleted> threads are doing this time. It is nice to be able to recognise (yes, I spell that with an 's' too) which threads are which.
1998-04-11 07:47:22 +00:00
char *name;
* Fix the stack allocation code so that it works for alpha. Change it to use mmap(..., MAP_STACK, ...) on alpha too since that should work now. * Add hooks to allow GDB to access the internals of pthreads without having to know the exact layout of struct pthread. Reviewed by: deischen
1999-11-28 19:47:43 +00:00
u_int64_t uniqueid; /* for gdb */
Add a magic field to the pthread structure to help recognize valid threads from invalid ones. The pthread structure is opaque to the user so this change does not cause any incompatibilities. Hopefully this change will help code that was written for draft 4 fail gracefully if the programmer ignores the compiler warning about the change in the level of indirection for the argument passed to pthread_detach(). I got burnt, so I fixed then (expletive deleted) thing. These functions comply with the revised standard. That should shut Terry up!
1998-04-03 09:31:15 +00:00
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
/*
* Lock for accesses to this thread structure.
*/
Add compile time thread lock debug support. Add a thread specific flag to trap the case where pthread_exit() is called from a destructor in violation of the Posix standard.
1998-06-09 23:02:43 +00:00
spinlock_t lock;
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
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
/* Queue entry for list of all threads: */
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_ENTRY(pthread) tle;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
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
/* Queue entry for list of dead threads: */
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_ENTRY(pthread) dle;
- Fix the debug macros. - Add support of a thread being listed in the dead thread list as well as the thread list. - Add a new thread state to make sigwait work properly. (Submitted by Daniel M. Eischen <eischen@vigrid.com>) - Add global variable for the garbage collector mutex and condition variable. - Delete a couple of prototypes that are no longer required. - Add a prototype for the garbage collector thread.
1998-09-30 06:22:07 +00:00
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Thread start routine, argument, stack pointer and thread
* attributes.
*/
Submitted by: John Birrell <cimaxp1!jb@werple.net.au> Here are the diffs for libc_r to get it one step closer to P1003.1c These make most of the thread/mutex/condvar structures opaque to the user. There are three functions which have been renamed with _np suffixes because they are extensions to P1003.1c (I did them for JAVA, which needs to suspend/resume threads and also start threads suspended). I've created a new header (pthread_np.h) for the non-POSIX stuff. The egrep tags stuff in /usr/src/lib/libc_r/Makefile that I uncommented doesn't work. I think its best to delete it. I don't think libc_r needs tags anyway, 'cause most of the source is in libc which does have tags. also: Here's the first batch of man pages for the thread functions. The diff to /usr/src/lib/libc_r/Makefile removes some stuff that was inherited from /usr/src/lib/libc/Makefile that should only be done with libc. also: I should have sent this diff with the pthread(3) man page. It allows people to type make -DWANT_LIBC_R world to get libc_r built with the rest of the world. I put this in the pthread(3) man page. The default is still not to build libc_r. also: The diff attached adds a pthread(3) man page to /usr/src/share/man/man3. The idea is that without libc_r installed, this man page will give people enough info to know that they have to build libc_r.
1996-08-20 08:22:01 +00:00
void *(*start_routine)(void *);
void *arg;
void *stack;
struct pthread_attr attr;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
/*
This has been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
* Threads return context; we use only jmp_buf's for now.
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
*/
union {
jmp_buf jb;
ucontext_t uc;
} ctx;
/*
* Used for tracking delivery of signal handlers.
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +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
struct pthread_signal_frame *curframe;
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
/*
* Cancelability flags - the lower 2 bits are used by cancel
* definitions in pthread.h
*/
#define PTHREAD_AT_CANCEL_POINT 0x0004
#define PTHREAD_CANCELLING 0x0008
#define PTHREAD_CANCEL_NEEDED 0x0010
int cancelflags;
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
thread_continuation_t continuation;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
* Current signal mask and pending signals.
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
*/
sigset_t sigmask;
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
sigset_t sigpend;
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
int sigmask_seqno;
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 check_pending;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* Thread state: */
enum pthread_state 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
/* Scheduling clock when this thread was last made active. */
long last_active;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +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
/* Scheduling clock when this thread was last made inactive. */
long last_inactive;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Number of microseconds accumulated by this thread when
* time slicing is active.
*/
long slice_usec;
/*
* Time to wake up thread. This is used for sleeping threads and
* for any operation which may time out (such as select).
*/
struct timespec wakeup_time;
/* TRUE if operation has timed out. */
int timeout;
/*
* Error variable used instead of errno. The function __error()
* returns a pointer to this.
*/
int error;
Fix pthread_join so that it works if the target thread exits while the joining thread is in a signal handler. Reported by: Loren James Rittle <rittle@labs.mot.com> MFC after: 1 week
2001-11-17 14:28:39 +00:00
/*
* The joiner is the thread that is joining to this thread. The
* join status keeps track of a join operation to another thread.
*/
struct pthread *joiner;
struct join_status join_status;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
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
* The current thread can belong to only one scheduling queue at
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
* a time (ready or waiting queue). It can also belong to:
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
*
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
* o A queue of threads waiting for a mutex
* o A queue of threads waiting for a condition variable
* o A queue of threads waiting for a file descriptor lock
* o A queue of threads needing work done by the kernel thread
* (waiting for a spinlock or file I/O)
Change the FILE locking to be by FILE, not by the underlying fd as it was. Add a FILE_WAIT state and queue threads waiting for a FILE lock. Start using the sys/queue.h macros instead of the way that MIT pthreads did it. Add a thread name to the private thread structure and a non-POSIX function to set this. This helps (me at least) when sending a SIGINFO to a threaded process to get a /tmp/uthread.dump to see what the <expletive deleted> threads are doing this time. It is nice to be able to recognise (yes, I spell that with an 's' too) which threads are which.
1998-04-11 07:47:22 +00:00
*
Instead of using a join queue for each thread, use a single pointer to keep track of a joiner. POSIX only supports a single joiner, so this simplification is acceptable. At the same time, make sure to mark a joined thread as detached so that its resources can be freed. Reviewed by: deischen PR: 24345
2001-05-20 23:08:33 +00:00
* A thread can also be joining a thread (the joiner field above).
*
* It must not be possible for a thread to belong to any of the
* above queues while it is handling a signal. Signal handlers
* may longjmp back to previous stack frames circumventing normal
* control flow. This could corrupt queue integrity if the thread
* retains membership in the queue. Therefore, if a thread is a
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
* member of one of these queues when a signal handler is invoked,
Instead of using a join queue for each thread, use a single pointer to keep track of a joiner. POSIX only supports a single joiner, so this simplification is acceptable. At the same time, make sure to mark a joined thread as detached so that its resources can be freed. Reviewed by: deischen PR: 24345
2001-05-20 23:08:33 +00:00
* it must remove itself from the queue before calling the signal
* handler and reinsert itself after normal return of the handler.
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
*
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
* Use pqe for the scheduling queue link (both ready and waiting),
Instead of using a join queue for each thread, use a single pointer to keep track of a joiner. POSIX only supports a single joiner, so this simplification is acceptable. At the same time, make sure to mark a joined thread as detached so that its resources can be freed. Reviewed by: deischen PR: 24345
2001-05-20 23:08:33 +00:00
* sqe for synchronization (mutex and condition variable) queue
* links, and qe for all other links.
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +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
TAILQ_ENTRY(pthread) pqe; /* priority queue link */
TAILQ_ENTRY(pthread) sqe; /* synchronization queue link */
TAILQ_ENTRY(pthread) qe; /* all other queues link */
Change the FILE locking to be by FILE, not by the underlying fd as it was. Add a FILE_WAIT state and queue threads waiting for a FILE lock. Start using the sys/queue.h macros instead of the way that MIT pthreads did it. Add a thread name to the private thread structure and a non-POSIX function to set this. This helps (me at least) when sending a SIGINFO to a threaded process to get a /tmp/uthread.dump to see what the <expletive deleted> threads are doing this time. It is nice to be able to recognise (yes, I spell that with an 's' too) which threads are which.
1998-04-11 07:47:22 +00:00
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* Wait data. */
union pthread_wait_data data;
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
/*
* Allocated for converting select into poll.
*/
struct pthread_poll_data poll_data;
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
/*
* Set to TRUE if a blocking operation was
* interrupted by a signal:
*/
int interrupted;
/* Signal number when in state PS_SIGWAIT: */
int signo;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
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
* Set to non-zero when this thread has deferred signals.
* We allow for recursive deferral.
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
*/
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 sig_defer_count;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* Set to TRUE if this thread should yield after undeferring
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
* signals.
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
*/
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 yield_on_sig_undefer;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
/* Miscellaneous flags; only set with signals deferred. */
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int flags;
#define PTHREAD_FLAGS_PRIVATE 0x0001
#define PTHREAD_EXITING 0x0002
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 PTHREAD_FLAGS_IN_WAITQ 0x0004 /* in waiting queue using pqe link */
#define PTHREAD_FLAGS_IN_PRIOQ 0x0008 /* in priority queue using pqe link */
#define PTHREAD_FLAGS_IN_WORKQ 0x0010 /* in work queue using qe link */
#define PTHREAD_FLAGS_IN_FILEQ 0x0020 /* in file lock queue using qe link */
#define PTHREAD_FLAGS_IN_FDQ 0x0040 /* in fd lock queue using qe link */
#define PTHREAD_FLAGS_IN_CONDQ 0x0080 /* in condition queue using sqe link*/
#define PTHREAD_FLAGS_IN_MUTEXQ 0x0100 /* in mutex queue using sqe link */
Revamp suspend and resume. While I'm here add pthread_suspend_all_np() and pthread_resume_all_np(). These suspend and resume all threads except the current thread, respectively. The existing functions pthread_single_np() and pthread_multi_np(), which formerly had no effect, now exhibit the same behaviour and pthread_suspend_all_np() and pthread_resume_all_np(). These functions have been added mostly for the native java port. Don't allow the uthread kernel pipe to use the same descriptors as stdio. Mostily submitted by Oswald Buddenhagen <ossi@kde.org>. Correct some minor style nits.
2002-05-24 04:32:28 +00:00
#define PTHREAD_FLAGS_SUSPENDED 0x0200 /* thread is suspended */
#define PTHREAD_FLAGS_TRACE 0x0400 /* for debugging purposes */
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 PTHREAD_FLAGS_IN_SYNCQ \
Instead of using a join queue for each thread, use a single pointer to keep track of a joiner. POSIX only supports a single joiner, so this simplification is acceptable. At the same time, make sure to mark a joined thread as detached so that its resources can be freed. Reviewed by: deischen PR: 24345
2001-05-20 23:08:33 +00:00
(PTHREAD_FLAGS_IN_CONDQ | PTHREAD_FLAGS_IN_MUTEXQ)
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* Base priority is the user setable and retrievable priority
* of the thread. It is only affected by explicit calls to
* set thread priority and upon thread creation via a thread
* attribute or default priority.
*/
char base_priority;
/*
* Inherited priority is the priority a thread inherits by
* taking a priority inheritence or protection mutex. It
* is not affected by base priority changes. Inherited
* priority defaults to and remains 0 until a mutex is taken
* that is being waited on by any other thread whose priority
* is non-zero.
*/
char inherited_priority;
/*
* Active priority is always the maximum of the threads base
* priority and inherited priority. When there is a change
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
* in either the base or inherited priority, the active
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
* priority must be recalculated.
*/
char active_priority;
/* Number of priority ceiling or protection mutexes owned. */
int priority_mutex_count;
MFlibpthread: Add a simple work-around for deadlocking on recursive readlocks on a rwlock while there are writers waiting.
2004-01-08 15:39:12 +00:00
/* Number rwlocks rdlocks held. */
int rdlock_count;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* Queue of currently owned mutexes.
*/
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
TAILQ_HEAD(, pthread_mutex) mutexq;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
Add the ability to recognize old references to keys, and return NULL when old keys are referenced (after pthread_key_delete()) via pthread_getspecific().
2002-03-19 22:58:56 +00:00
void *ret;
struct pthread_specific_elem *specific;
int specific_data_count;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* Cleanup handlers Link List */
struct pthread_cleanup *cleanup;
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
char *fname; /* Ptr to source file name */
int lineno; /* Source line number. */
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
};
/*
* Global variables for the uthread kernel.
*/
Make libc_r check the kern.usrstack sysctl instead of using internal kernel #defines to figure out where the stack is located. This stops libc_r from exploding when the kernel is compiled with a different KVM size. IMHO this is all kinda bogus, it would be better to just check %esp and work from that.
2001-10-26 21:19:22 +00:00
SCLASS void *_usrstack
#ifdef GLOBAL_PTHREAD_PRIVATE
= (void *) USRSTACK;
#else
;
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* Kernel thread structure used when there are no running threads: */
SCLASS struct pthread _thread_kern_thread;
/* Ptr to the thread structure for the running thread: */
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
SCLASS struct pthread * volatile _thread_run
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
= &_thread_kern_thread;
#else
;
#endif
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/* Ptr to the thread structure for the last user thread to run: */
SCLASS struct pthread * volatile _last_user_thread
#ifdef GLOBAL_PTHREAD_PRIVATE
= &_thread_kern_thread;
#else
;
#endif
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
/* List of all threads: */
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
SCLASS TAILQ_HEAD(, pthread) _thread_list
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
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
= TAILQ_HEAD_INITIALIZER(_thread_list);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#else
;
#endif
/*
* Array of kernel pipe file descriptors that are used to ensure that
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
* no signals are missed in calls to _select.
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
*/
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
SCLASS int _thread_kern_pipe[2]
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
= {
-1,
-1
};
#else
;
#endif
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
SCLASS int volatile _queue_signals
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0;
#else
;
#endif
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
SCLASS int _thread_kern_in_sched
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0;
#else
;
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +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
SCLASS int _sig_in_handler
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0;
#else
;
#endif
/* Time of day at last scheduling timer signal: */
SCLASS struct timeval volatile _sched_tod
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
= { 0, 0 };
#else
;
#endif
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
/*
* Current scheduling timer ticks; used as resource usage.
*/
SCLASS unsigned int volatile _sched_ticks
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0;
#else
;
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* Dead threads: */
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
SCLASS TAILQ_HEAD(, pthread) _dead_list
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
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
= TAILQ_HEAD_INITIALIZER(_dead_list);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#else
;
#endif
/* Initial thread: */
SCLASS struct pthread *_thread_initial
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL;
#else
;
#endif
/* Default thread attributes: */
At initialization, override the pthread stub routines in libc by filling in the jump table. Convert uses of pthread routines within libc_r to use the internal versions (_pthread_foo instead of pthread_foo). Remove a couple of globals from application namespace.
2002-11-13 18:13:26 +00:00
SCLASS struct pthread_attr _pthread_attr_default
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to be malloc()ed, but they are now allocated using mmap(), just as the default-size stacks are. A separate cache of stacks is kept for non-default-size stacks. Collaboration with: deischen
2001-07-20 04:23:11 +00:00
= { SCHED_RR, 0, TIMESLICE_USEC, PTHREAD_DEFAULT_PRIORITY,
PTHREAD_CREATE_RUNNING, PTHREAD_CREATE_JOINABLE, NULL, NULL, NULL,
Don't use PAGE_SIZE in userland, instead use getpagesize(), this is to allow running on other arches when the instructions are supported but the page size granularity is not. Glanced at by: peter
2002-05-13 07:58:15 +00:00
PTHREAD_STACK_DEFAULT, -1 };
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#else
;
#endif
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
/* Default mutex attributes: */
At initialization, override the pthread stub routines in libc by filling in the jump table. Convert uses of pthread routines within libc_r to use the internal versions (_pthread_foo instead of pthread_foo). Remove a couple of globals from application namespace.
2002-11-13 18:13:26 +00:00
#define PTHREAD_MUTEXATTR_DEFAULT \
{ PTHREAD_MUTEX_DEFAULT, PTHREAD_PRIO_NONE, 0, 0 }
Moved enum pthread_mutextype to pthread.h. Add pthread_mutexattr_default definition.
1996-11-11 09:07:05 +00:00
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
/* Default condition variable attributes: */
At initialization, override the pthread stub routines in libc by filling in the jump table. Convert uses of pthread routines within libc_r to use the internal versions (_pthread_foo instead of pthread_foo). Remove a couple of globals from application namespace.
2002-11-13 18:13:26 +00:00
#define PTHREAD_CONDATTR_DEFAULT { COND_TYPE_FAST, 0 }
Submitted by: John Birrell uthreads update from the author.
1997-02-05 23:26:09 +00:00
/*
* Standard I/O file descriptors need special flag treatment since
* setting one to non-blocking does all on *BSD. Sigh. This array
* is used to store the initial flag settings.
*/
SCLASS int _pthread_stdio_flags[3];
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* File table information: */
SCLASS struct fd_table_entry **_thread_fd_table
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL;
#else
;
#endif
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
/* Table for polling file descriptors: */
SCLASS struct pollfd *_thread_pfd_table
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL;
#else
;
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
SCLASS const int dtablecount
#ifdef GLOBAL_PTHREAD_PRIVATE
= 4096/sizeof(struct fd_table_entry);
#else
;
#endif
SCLASS int _thread_dtablesize /* Descriptor table size. */
#ifdef GLOBAL_PTHREAD_PRIVATE
Fix thread initialization to allow for the case where stdio file descriptors are not opened. PR: bin/12853 Reviewed by: jb
1999-08-05 12:08:10 +00:00
= 0;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#else
;
#endif
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
SCLASS int _clock_res_usec /* Clock resolution in usec. */
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
#ifdef GLOBAL_PTHREAD_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
= CLOCK_RES_USEC;
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
#else
;
#endif
- Fix the debug macros. - Add support of a thread being listed in the dead thread list as well as the thread list. - Add a new thread state to make sigwait work properly. (Submitted by Daniel M. Eischen <eischen@vigrid.com>) - Add global variable for the garbage collector mutex and condition variable. - Delete a couple of prototypes that are no longer required. - Add a prototype for the garbage collector thread.
1998-09-30 06:22:07 +00:00
/* Garbage collector mutex and condition variable. */
SCLASS pthread_mutex_t _gc_mutex
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL
#endif
;
SCLASS pthread_cond_t _gc_cond
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL
#endif
;
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
/*
* Array of signal actions for this process.
*/
Change signal handling to conform to POSIX specified semantics. Before this change, a signal was delivered to each thread that didn't have the signal masked. Signals also improperly woke up threads waiting on I/O. With this change, signals are now handled in the following way: o If a thread is waiting in a sigwait for the signal, then the thread is woken up. o If no threads are sigwait'ing on the signal and a thread is in a sigsuspend waiting for the signal, then the thread is woken up. o In the case that no threads are waiting or suspended on the signal, then the signal is delivered to the first thread we find that has the signal unmasked. o If no threads are waiting or suspended on the signal, and no threads have the signal unmasked, then the signal is added to the process wide pending signal set. The signal will be delivered to the first thread that unmasks the signal. If there is an installed signal handler, it is only invoked if the chosen thread was not in a sigwait. In the case that multiple threads are waiting or suspended on a signal, or multiple threads have the signal unmasked, we wake up/deliver the signal to the first thread we find. The above rules still apply. Reported by: Scott Hess <scott@avantgo.com> Reviewed by: jb, jasone
1999-12-04 22:55:59 +00:00
SCLASS struct sigaction _thread_sigact[NSIG];
If multiple threads are blocked in sigwait() for the same signal that does not have a user-supplied signal handler, when a signal is delivered, one thread will receive the signal, and then the code reverts to having no signal handler for the signal. This can leave the other sigwait()ing threads stranded permanently if the signal is later ignored, or can result in process termination when the process should have delivered the signal to one of the threads in sigwait(). To fix this problem, maintain a count of sigwait()ers for each signal that has no default signal handler. Use the count to correctly install/uninstall dummy signal handlers. Reviewed by: deischen
2000-06-27 21:30:16 +00:00
/*
* Array of counts of dummy handlers for SIG_DFL signals. This is used to
* assure that there is always a dummy signal handler installed while there is a
* thread sigwait()ing on the corresponding signal.
*/
SCLASS int _thread_dfl_count[NSIG];
Change signal handling to conform to POSIX specified semantics. Before this change, a signal was delivered to each thread that didn't have the signal masked. Signals also improperly woke up threads waiting on I/O. With this change, signals are now handled in the following way: o If a thread is waiting in a sigwait for the signal, then the thread is woken up. o If no threads are sigwait'ing on the signal and a thread is in a sigsuspend waiting for the signal, then the thread is woken up. o In the case that no threads are waiting or suspended on the signal, then the signal is delivered to the first thread we find that has the signal unmasked. o If no threads are waiting or suspended on the signal, and no threads have the signal unmasked, then the signal is added to the process wide pending signal set. The signal will be delivered to the first thread that unmasks the signal. If there is an installed signal handler, it is only invoked if the chosen thread was not in a sigwait. In the case that multiple threads are waiting or suspended on a signal, or multiple threads have the signal unmasked, we wake up/deliver the signal to the first thread we find. The above rules still apply. Reported by: Scott Hess <scott@avantgo.com> Reviewed by: jb, jasone
1999-12-04 22:55:59 +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
* Pending signals and mask for this process:
Change signal handling to conform to POSIX specified semantics. Before this change, a signal was delivered to each thread that didn't have the signal masked. Signals also improperly woke up threads waiting on I/O. With this change, signals are now handled in the following way: o If a thread is waiting in a sigwait for the signal, then the thread is woken up. o If no threads are sigwait'ing on the signal and a thread is in a sigsuspend waiting for the signal, then the thread is woken up. o In the case that no threads are waiting or suspended on the signal, then the signal is delivered to the first thread we find that has the signal unmasked. o If no threads are waiting or suspended on the signal, and no threads have the signal unmasked, then the signal is added to the process wide pending signal set. The signal will be delivered to the first thread that unmasks the signal. If there is an installed signal handler, it is only invoked if the chosen thread was not in a sigwait. In the case that multiple threads are waiting or suspended on a signal, or multiple threads have the signal unmasked, we wake up/deliver the signal to the first thread we find. The above rules still apply. Reported by: Scott Hess <scott@avantgo.com> Reviewed by: jb, jasone
1999-12-04 22:55:59 +00:00
*/
SCLASS sigset_t _process_sigpending;
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
SCLASS sigset_t _process_sigmask
#ifdef GLOBAL_PTHREAD_PRIVATE
= { {0, 0, 0, 0} }
#endif
;
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
/*
* Scheduling queues:
*/
SCLASS pq_queue_t _readyq;
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
SCLASS TAILQ_HEAD(, pthread) _waitingq;
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
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
/*
* Work queue:
*/
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
SCLASS TAILQ_HEAD(, pthread) _workq;
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
/* Tracks the number of threads blocked while waiting for a spinlock. */
SCLASS volatile int _spinblock_count
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0
#endif
;
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
/* Used to maintain pending and active signals: */
struct sigstatus {
int pending; /* Is this a pending signal? */
int blocked; /*
* A handler is currently active for
* this signal; ignore subsequent
* signals until the handler is done.
*/
int signo; /* arg 1 to signal handler */
siginfo_t siginfo; /* arg 2 to signal handler */
ucontext_t uc; /* arg 3 to signal handler */
};
SCLASS struct sigstatus _thread_sigq[NSIG];
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
/* Indicates that the signal queue needs to be checked. */
SCLASS volatile int _sigq_check_reqd
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0
#endif
;
/* Thread switch hook. */
SCLASS pthread_switch_routine_t _sched_switch_hook
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL
#endif
;
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
/*
* Declare the kernel scheduler jump buffer and stack:
*/
SCLASS jmp_buf _thread_kern_sched_jb;
SCLASS void * _thread_kern_sched_stack
#ifdef GLOBAL_PTHREAD_PRIVATE
= NULL
#endif
;
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
/* Used for _PTHREADS_INVARIANTS checking. */
SCLASS int _thread_kern_new_state
#ifdef GLOBAL_PTHREAD_PRIVATE
= 0
#endif
;
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/* Undefine the storage class specifier: */
#undef SCLASS
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
#ifdef _LOCK_DEBUG
Add compile time thread lock debug support. Add a thread specific flag to trap the case where pthread_exit() is called from a destructor in violation of the Posix standard.
1998-06-09 23:02:43 +00:00
#define _FD_LOCK(_fd,_type,_ts) _thread_fd_lock_debug(_fd, _type, \
_ts, __FILE__, __LINE__)
#define _FD_UNLOCK(_fd,_type) _thread_fd_unlock_debug(_fd, _type, \
__FILE__, __LINE__)
#else
#define _FD_LOCK(_fd,_type,_ts) _thread_fd_lock(_fd, _type, _ts)
#define _FD_UNLOCK(_fd,_type) _thread_fd_unlock(_fd, _type)
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
/*
* Function prototype definitions.
*/
__BEGIN_DECLS
char *__ttyname_basic(int);
char *__ttyname_r_basic(int, char *, size_t);
char *ttyname_r(int, char *, size_t);
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 _cond_wait_backout(pthread_t);
void _fd_lock_backout(pthread_t);
Change signal model to match POSIX (i.e. one set of signal handlers for the process, not a separate set for each thread). By default, the process now only has signal handlers installed for SIGVTALRM, SIGINFO and SIGCHLD. The thread kernel signal handler is installed for other signals on demand. This means that SIG_IGN and SIG_DFL processing is now left to the kernel, not the thread kernel. Change the signal dispatch to no longer use a signal thread, and call the signal handler using the stack of the thread that has the signal pending. Change the atomic lock method to use test-and-set asm code with a yield if blocked. This introduces separate locks for each type of object instead of blocking signals to prevent a context switch. It was this blocking of signals that caused the performance degradation the people have noted. This is a *big* change!
1998-04-29 09:59:34 +00:00
int _find_thread(pthread_t);
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
struct pthread *_get_curthread(void);
void _set_curthread(struct pthread *);
Implement pthread_attr_[gs]etguardsize(). Non-default-size stacks used to be malloc()ed, but they are now allocated using mmap(), just as the default-size stacks are. A separate cache of stacks is kept for non-default-size stacks. Collaboration with: deischen
2001-07-20 04:23:11 +00:00
void *_thread_stack_alloc(size_t, size_t);
void _thread_stack_free(void *, size_t, size_t);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
int _thread_create(pthread_t *,const pthread_attr_t *,void *(*start_routine)(void *),void *,pthread_t);
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
int _mutex_cv_lock(pthread_mutex_t *);
int _mutex_cv_unlock(pthread_mutex_t *);
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 _mutex_lock_backout(pthread_t);
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
void _mutex_notify_priochange(pthread_t);
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 _mutex_reinit(pthread_mutex_t *);
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
void _mutex_unlock_private(pthread_t);
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 _cond_reinit(pthread_cond_t *);
int _pq_alloc(struct pq_queue *, int, int);
int _pq_init(struct pq_queue *);
[ The author's description... ] o Runnable threads are now maintained in priority queues. The implementation requires two things: 1.) The priority queues must be protected during insertion and removal of threads. Since the kernel scheduler must modify the priority queues, a spinlock for protection cannot be used. The functions _thread_kern_sched_defer() and _thread_kern_sched_undefer() were added to {un}defer kernel scheduler activation. 2.) A thread (active) priority change can be performed only when the thread is removed from the priority queue. The implementation uses a threads active priority when inserting it into the queue. A by-product is that thread switches are much faster. A separate queue is used for waiting and/or blocked threads, and it is searched at most 2 times in the kernel scheduler when there are active threads. It should be possible to reduce this to once by combining polling of threads waiting on I/O with the loop that looks for timed out threads and the minimum timeout value. o Functions to defer kernel scheduler activation were added. These are _thread_kern_sched_defer() and _thread_kern_sched_undefer() and may be called recursively. These routines do not block the scheduling signal, but latch its occurrence. The signal handler will not call the kernel scheduler when the running thread has deferred scheduling, but it will be called when running thread undefers scheduling. o Added support for _POSIX_THREAD_PRIORITY_SCHEDULING. All the POSIX routines required by this should now be implemented. One note, SCHED_OTHER, SCHED_FIFO, and SCHED_RR are required to be defined by including pthread.h. These defines are currently in sched.h. I modified pthread.h to include sched.h but don't know if this is the proper thing to do. o Added support for priority protection and inheritence mutexes. This allows definition of _POSIX_THREAD_PRIO_PROTECT and _POSIX_THREAD_PRIO_INHERIT. o Added additional error checks required by POSIX for mutexes and condition variables. o Provided a wrapper for sigpending which is marked as a hidden syscall. o Added a non-portable function as a debugging aid to allow an application to monitor thread context switches. An application can install a routine that gets called everytime a thread (explicitly created by the application) gets context switched. The routine gets passed the pthread IDs of the threads that are being switched in and out. Submitted by: Dan Eischen <eischen@vigrid.com> Changes by me: o Added a PS_SPINBLOCK state to deal with the priority inversion problem most often (I think) seen by threads calling malloc/free/realloc. o Dispatch signals to the running thread directly rather than at a context switch to avoid the situation where the switch never occurs.
1999-03-23 05:07:56 +00:00
void _pq_remove(struct pq_queue *pq, struct pthread *);
void _pq_insert_head(struct pq_queue *pq, struct pthread *);
void _pq_insert_tail(struct pq_queue *pq, struct pthread *);
struct pthread *_pq_first(struct pq_queue *pq);
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
void *_pthread_getspecific(pthread_key_t);
int _pthread_key_create(pthread_key_t *, void (*) (void *));
int _pthread_key_delete(pthread_key_t);
int _pthread_mutex_destroy(pthread_mutex_t *);
int _pthread_mutex_init(pthread_mutex_t *, const pthread_mutexattr_t *);
int _pthread_mutex_lock(pthread_mutex_t *);
int _pthread_mutex_trylock(pthread_mutex_t *);
int _pthread_mutex_unlock(pthread_mutex_t *);
int _pthread_mutexattr_init(pthread_mutexattr_t *);
int _pthread_mutexattr_destroy(pthread_mutexattr_t *);
int _pthread_mutexattr_settype(pthread_mutexattr_t *, int);
int _pthread_once(pthread_once_t *, void (*) (void));
Remove (int) file descriptor locking. It should be up to the application to provide locking for I/O operations. This doesn't break any of my tests, but the old behavior can be restored by compiling with _FDLOCKS_ENABLED. This will eventually be removed when it is obvious it does not cause any problems. Remove most of flockfile implementation, with the exception of flockfile_debug. Make error messages more informational (submitted by Mike Heffner <spock@techfour.net>, who's now known as mikeh@FreeBSD.org).
2001-02-11 22:07:32 +00:00
pthread_t _pthread_self(void);
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
int _pthread_setspecific(pthread_key_t, const void *);
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
void _waitq_insert(pthread_t pthread);
void _waitq_remove(pthread_t 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
#if defined(_PTHREADS_INVARIANTS)
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
void _waitq_setactive(void);
void _waitq_clearactive(void);
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
void _thread_exit(char *, int, char *);
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
void _thread_exit_cleanup(void);
Remove much of the dereferencing of the fd table entries to look at file flags and replace it with functions that will avoid null pointer checks. MFC to be done by archie ;-) PR: 42100 Reviewed by: archie, robert MFC after: 3 days
2002-08-29 23:06:07 +00:00
int _thread_fd_getflags(int);
int _thread_fd_lock(int, int, struct timespec *);
int _thread_fd_lock_debug(int, int, struct timespec *,char *fname,int lineno);
void _thread_fd_setflags(int, int);
int _thread_fd_table_init(int fd);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
void _thread_fd_unlock(int, int);
Add compile time thread lock debug support. Add a thread specific flag to trap the case where pthread_exit() is called from a destructor in violation of the Posix standard.
1998-06-09 23:02:43 +00:00
void _thread_fd_unlock_debug(int, int, char *, int);
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
void _thread_fd_unlock_owned(pthread_t);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
void *_thread_cleanup(pthread_t);
void _thread_cleanupspecific(void);
void _thread_dump_info(void);
void _thread_init(void);
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
void _thread_kern_sched(ucontext_t *);
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 _thread_kern_scheduler(void);
Don't needlessly poll file descriptors when there are no file descriptors needing to be polled (Doh!). Reported by Dan Nelson <dnelson@emsphone.com>. Don't install and start the scheduling timer until the first thread is created. This prevents the overhead of having a periodic scheduling signal in a single threaded program. Reported by Dan Nelson <dnelson@emsphone.com>. Allow builtin longjmps out of application installed signal handlers without the need perform any post-handler cleanup: o Change signal handling to save the threads interrupted context on the stack. The threads current context is now always stored in the same place (in the pthread). If and when a signal handler returns, the interrupted context is copied back to the storage area in the pthread. o Before calling invoking a signal handler for a thread, back the thread out of any internal waiting queues (mutex, CV, join, etc) to which it belongs. Rework uthread_info.c a bit to make it easier to change the format of a thread dump. Use an alternal signal stack for the thread library's signal handler. This allows us to fiddle with the main threads stack without fear of it being in use. Reviewed by: jasone
2000-11-09 05:08:26 +00:00
void _thread_kern_sched_frame(struct pthread_signal_frame *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
void _thread_kern_sched_sig(void);
void _thread_kern_sched_state(enum pthread_state, char *fname, int lineno);
Close a window between unlocking a spinlock and changing the thread state.
1998-11-15 09:58:26 +00:00
void _thread_kern_sched_state_unlock(enum pthread_state state,
spinlock_t *lock, char *fname, int lineno);
Add wrapper for kevent() syscall Noted as missing by: nicolas.leonard@animaths.com
2000-08-07 16:51:56 +00:00
void _thread_kern_set_timeout(const struct timespec *);
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
void _thread_kern_sig_defer(void);
void _thread_kern_sig_undefer(void);
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 _thread_sig_handler(int, siginfo_t *, ucontext_t *);
This has been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
void _thread_sig_check_pending(struct pthread *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
void _thread_sig_handle_pending(void);
This has been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
void _thread_sig_send(struct pthread *pthread, 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
void _thread_sig_wrapper(void);
This has been sitting in my local tree long enough. Remove the use of an alternate signal stack for handling signals. Let the kernel send signals on the stack of the current thread and teach the threads signal handler how to deliver signals to the current thread if it needs to. Also, always store a threads context as a jmp_buf. Eventually this will change to be a ucontext_t or mcontext_t. Other small nits. Use struct pthread * instead of pthread_t in internal library routines. The threads code wants struct pthread *, and pthread_t doesn't necessarily have to be the same. Reviewed by: jasone
2002-02-09 19:58:41 +00:00
void _thread_sigframe_restore(struct pthread *thread,
struct pthread_signal_frame *psf);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
void _thread_start(void);
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 _thread_seterrno(pthread_t, int);
- Fix the debug macros. - Add support of a thread being listed in the dead thread list as well as the thread list. - Add a new thread state to make sigwait work properly. (Submitted by Daniel M. Eischen <eischen@vigrid.com>) - Add global variable for the garbage collector mutex and condition variable. - Delete a couple of prototypes that are no longer required. - Add a prototype for the garbage collector thread.
1998-09-30 06:22:07 +00:00
pthread_addr_t _thread_gc(pthread_addr_t);
add pthread_cancel, obtained from OpenBSD. eischen (Daniel Eischen) added wrappers to protect against cancled threads orphaning internal resources. the cancelability code is still a bit fuzzy but works for test programs of my own, OpenBSD's and some examples from ORA's books. add readdir_r to both libc and libc_r add some 'const' attributes to function parameters Reviewed by: eischen, jasone
1999-11-28 05:38:13 +00:00
void _thread_enter_cancellation_point(void);
void _thread_leave_cancellation_point(void);
void _thread_cancellation_point(void);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/acl.h> */
#ifdef _SYS_ACL_H
int __sys___acl_aclcheck_fd(int, acl_type_t, struct acl *);
int __sys___acl_delete_fd(int, acl_type_t);
int __sys___acl_get_fd(int, acl_type_t, struct acl *);
int __sys___acl_set_fd(int, acl_type_t, struct acl *);
#endif
_exit in libc is now __sys_exit not __sys__exit. Add another check for thread library initialization (jdp, we really need a way to get _thread_init called at program start before any constructors are run).
2001-01-29 03:24:23 +00:00
/* #include <sys/aio.h> */
#ifdef _SYS_AIO_H_
Unbreak world by correctly specifying the prototype for __sys_aio_suspend. A make buildworld was done but not with the committed pthread_private.h. Reported by: Manfred Antar <null@pozo.com>
2001-01-29 18:59:53 +00:00
int __sys_aio_suspend(const struct aiocb * const[], int, const struct timespec *);
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/event.h> */
#ifdef _SYS_EVENT_H_
int __sys_kevent(int, const struct kevent *, int, struct kevent *,
int, const struct timespec *);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/ioctl.h> */
#ifdef _SYS_IOCTL_H_
int __sys_ioctl(int, unsigned long, ...);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/mman.h> */
#ifdef _SYS_MMAN_H_
int __sys_msync(void *, size_t, int);
#endif
/* #include <sys/mount.h> */
#ifdef _SYS_MOUNT_H_
int __sys_fstatfs(int, struct statfs *);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
/* #include <sys/socket.h> */
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
#ifdef _SYS_SOCKET_H_
Removed: - uthread_signal.c; libc_r does not wrap signal() since 1998/04/29. - uthread_attr_setprio.c; it was never connected to the build, and pthread_attr_setprio() does not exist in POSIX. - uthread_sigblock.c and uthread_sigsetmask.c; these were no-ops bloating libc_r's space. pthread_private.h: - Removed prototypes of non-syscalls: send(). - Removed prototypes of unused syscalls: sigpending(), sigsuspend(), and select(). - Fixed prototype of fork(). - MFS: Fixed prototypes of <sys/socket.h> syscalls. Reviewed by: deischen Approved by: deischen, jasone
2001-10-26 17:46:36 +00:00
int __sys_accept(int, struct sockaddr *, socklen_t *);
int __sys_bind(int, const struct sockaddr *, socklen_t);
int __sys_connect(int, const struct sockaddr *, socklen_t);
int __sys_getpeername(int, struct sockaddr *, socklen_t *);
int __sys_getsockname(int, struct sockaddr *, socklen_t *);
int __sys_getsockopt(int, int, int, void *, socklen_t *);
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
int __sys_listen(int, int);
Removed: - uthread_signal.c; libc_r does not wrap signal() since 1998/04/29. - uthread_attr_setprio.c; it was never connected to the build, and pthread_attr_setprio() does not exist in POSIX. - uthread_sigblock.c and uthread_sigsetmask.c; these were no-ops bloating libc_r's space. pthread_private.h: - Removed prototypes of non-syscalls: send(). - Removed prototypes of unused syscalls: sigpending(), sigsuspend(), and select(). - Fixed prototype of fork(). - MFS: Fixed prototypes of <sys/socket.h> syscalls. Reviewed by: deischen Approved by: deischen, jasone
2001-10-26 17:46:36 +00:00
ssize_t __sys_recvfrom(int, void *, size_t, int, struct sockaddr *, socklen_t *);
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
ssize_t __sys_recvmsg(int, struct msghdr *, int);
Add weak definitions for wrapped system calls. In general: _foo - wrapped system call foo - weak definition to _foo and for cancellation points: _foo - wrapped system call __foo - enter cancellation point, call _foo(), leave cancellation point foo - weak definition to __foo Change use of global _thread_run to call a function to get the currently running thread. Make all pthread_foo functions weak definitions to _pthread_foo, where _pthread_foo is the implementation. This allows an application to provide its own pthread functions. Provide slightly different versions of pthread_mutex_lock and pthread_mutex_init so that we can tell the difference between a libc mutex and an application mutex. Threads holding mutexes internal to libc should never be allowed to exit, call signal handlers, or cancel. Approved by: -arch
2001-01-24 13:03:38 +00:00
int __sys_sendfile(int, int, off_t, size_t, struct sf_hdtr *, off_t *, int);
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
ssize_t __sys_sendmsg(int, const struct msghdr *, int);
Removed: - uthread_signal.c; libc_r does not wrap signal() since 1998/04/29. - uthread_attr_setprio.c; it was never connected to the build, and pthread_attr_setprio() does not exist in POSIX. - uthread_sigblock.c and uthread_sigsetmask.c; these were no-ops bloating libc_r's space. pthread_private.h: - Removed prototypes of non-syscalls: send(). - Removed prototypes of unused syscalls: sigpending(), sigsuspend(), and select(). - Fixed prototype of fork(). - MFS: Fixed prototypes of <sys/socket.h> syscalls. Reviewed by: deischen Approved by: deischen, jasone
2001-10-26 17:46:36 +00:00
ssize_t __sys_sendto(int, const void *,size_t, int, const struct sockaddr *, socklen_t);
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
int __sys_setsockopt(int, int, int, const void *, socklen_t);
int __sys_shutdown(int, int);
int __sys_socket(int, int, int);
int __sys_socketpair(int, int, int, int *);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/stat.h> */
#ifdef _SYS_STAT_H_
int __sys_fchflags(int, u_long);
int __sys_fchmod(int, mode_t);
int __sys_fstat(int, struct stat *);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/uio.h> */
#ifdef _SYS_UIO_H_
ssize_t __sys_readv(int, const struct iovec *, int);
ssize_t __sys_writev(int, const struct iovec *, int);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <sys/wait.h> */
#ifdef WNOHANG
pid_t __sys_wait4(pid_t, int *, int, struct rusage *);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
/* #include <dirent.h> */
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
#ifdef _DIRENT_H_
int __sys_getdirentries(int, char *, int, long *);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <fcntl.h> */
#ifdef _SYS_FCNTL_H_
int __sys_fcntl(int, int, ...);
int __sys_flock(int, int);
int __sys_open(const char *, int, ...);
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
#endif
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
/* #include <poll.h> */
#ifdef _SYS_POLL_H_
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
int __sys_poll(struct pollfd *, unsigned, int);
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
#endif
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
Style: sort __sys_foo() prototypes, tabs -> spaces, etc.
2001-10-26 18:45:02 +00:00
/* #include <signal.h> */
#ifdef _SIGNAL_H_
int __sys_sigaction(int, const struct sigaction *, struct sigaction *);
int __sys_sigprocmask(int, const sigset_t *, sigset_t *);
int __sys_sigreturn(ucontext_t *);
#endif
/* #include <unistd.h> */
#ifdef _UNISTD_H_
int __sys_close(int);
int __sys_dup(int);
int __sys_dup2(int, int);
int __sys_execve(const char *, char * const *, char * const *);
void __sys_exit(int);
int __sys_fchown(int, uid_t, gid_t);
pid_t __sys_fork(void);
long __sys_fpathconf(int, int);
int __sys_fsync(int);
int __sys_pipe(int *);
ssize_t __sys_read(int, void *, size_t);
ssize_t __sys_write(int, const void *, size_t);
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
#endif
/* #include <setjmp.h> */
#ifdef _SETJMP_H_
Minor *jmp() cleanups.
2000-01-20 21:53:59 +00:00
extern void __siglongjmp(sigjmp_buf, int) __dead2;
extern void __longjmp(jmp_buf, int) __dead2;
extern void ___longjmp(jmp_buf, int) __dead2;
Implement continuations to correctly handle [sig|_]longjmp() inside of a signal handler. Explicitly check for jumps to anywhere other than the current stack, since such jumps are undefined according to POSIX. While we're at it, convert thread cancellation to use continuations, since it's cleaner than the original cancellation code. Avoid delivering a signal to a thread twice. This was a pre-existing bug, but was likely unexposed until these other changes were made. Defer signals generated by pthread_kill() so that they can be delivered on the appropriate stack. deischen claims that this is unnecessary, which is likely true, but without this change, pthread_kill() can cause undefined priority queue states and/or PANICs in [sig|_]longjmp(), so I'm leaving this in for now. To compile this code out and exercise the bug, define the _NO_UNDISPATCH cpp macro. Defining _PTHREADS_INVARIANTS as well will cause earlier crashes. PR: kern/14685 Collaboration with: deischen
2000-01-19 07:04:50 +00:00
#endif
Reviewed by: julian Submitted by: john birrel One version of the pthreads library another will follow with differnt actions under some cases.. not QUITE complete
1996-01-22 00:23:58 +00:00
__END_DECLS
#endif /* !_PTHREAD_PRIVATE_H */
Reference in New Issue Copy Permalink