mirror/freebsd
1
0
Fork 0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-23 11:18:54 +00:00
Code Issues Releases Activity
aa785996fd
freebsd/sys/kern/kern_switch.c

634 lines
17 KiB
C
Raw Normal View History

/* -> /*- for copyright notices, minor format tweaks as necessary
2005-01-06 23:35:40 +00:00
/*-
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
* Copyright (c) 2001 Jake Burkholder <jake@FreeBSD.org>
* All rights reserved.
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
*
* 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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
Use __FBSDID().
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
#include "opt_sched.h"
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
#ifndef KERN_SWITCH_INCLUDE
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
Update for the KDB framework: o Make debugging code conditional upon KDB instead of DDB. o Call kdb_enter() instead of Debugger(). o Call kdb_backtrace() instead of db_print_backtrace() or backtrace(). kern_mutex.c: o Replace checks for db_active with checks for kdb_active and make them unconditional. kern_shutdown.c: o s/DDB_UNATTENDED/KDB_UNATTENDED/g o s/DDB_TRACE/KDB_TRACE/g o Save the TID of the thread doing the kernel dump so the debugger knows which thread to select as the current when debugging the kernel core file. o Clear kdb_active instead of db_active and do so unconditionally. o Remove backtrace() implementation. kern_synch.c: o Call kdb_reenter() instead of db_error().
2004-07-10 21:36:01 +00:00
#include <sys/kdb.h>
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
#include <sys/kernel.h>
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#include <sys/ktr.h>
Catch up to header include changes: - <sys/mutex.h> now requires <sys/systm.h> - <sys/mutex.h> and <sys/sx.h> now require <sys/lock.h>
2001-03-28 09:17:56 +00:00
#include <sys/lock.h>
Catch up to moving headers: - machine/ipl.h -> sys/ipl.h - machine/mutex.h -> sys/mutex.h
2000-10-20 07:58:15 +00:00
#include <sys/mutex.h>
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
#include <sys/proc.h>
#include <sys/queue.h>
- Create a new scheduler api that is defined in sys/sched.h - Begin moving scheduler specific functionality into sched_4bsd.c - Replace direct manipulation of scheduler data with hooks provided by the new api. - Remove KSE specific state modifications and single runq assumptions from kern_switch.c Reviewed by: -arch
2002-10-12 05:32:24 +00:00
#include <sys/sched.h>
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
#else /* KERN_SWITCH_INCLUDE */
Initial landing of SMP support for FreeBSD/amd64. - This is heavily derived from John Baldwin's apic/pci cleanup on i386. - I have completely rewritten or drastically cleaned up some other parts. (in particular, bootstrap) - This is still a WIP. It seems that there are some highly bogus bioses on nVidia nForce3-150 boards. I can't stress how broken these boards are. I have a workaround in mind, but right now the Asus SK8N is broken. The Gigabyte K8NPro (nVidia based) is also mind-numbingly hosed. - Most of my testing has been with SCHED_ULE. SCHED_4BSD works. - the apic and acpi components are 'standard'. - If you have an nVidia nForce3-150 board, you are stuck with 'device atpic' in addition, because they somehow managed to forget to connect the 8254 timer to the apic, even though its in the same silicon! ARGH! This directly violates the ACPI spec.
2003-11-17 08:58:16 +00:00
#if defined(SMP) && (defined(__i386__) || defined(__amd64__))
Commit a partial lazy thread switch mechanism for i386. it isn't as lazy as it could be and can do with some more cleanup. Currently its under options LAZY_SWITCH. What this does is avoid %cr3 reloads for short context switches that do not involve another user process. ie: we can take an interrupt, switch to a kthread and return to the user without explicitly flushing the tlb. However, this isn't as exciting as it could be, the interrupt overhead is still high and too much blocks on Giant still. There are some debug sysctls, for stats and for an on/off switch. The main problem with doing this has been "what if the process that you're running on exits while we're borrowing its address space?" - in this case we use an IPI to give it a kick when we're about to reclaim the pmap. Its not compiled in unless you add the LAZY_SWITCH option. I want to fix a few more things and get some more feedback before turning it on by default. This is NOT a replacement for Bosko's lazy interrupt stuff. This was more meant for the kthread case, while his was for interrupts. Mine helps a little for interrupts, but his helps a lot more. The stats are enabled with options SWTCH_OPTIM_STATS - this has been a pseudo-option for years, I just added a bunch of stuff to it. One non-trivial change was to select a new thread before calling cpu_switch() in the first place. This allows us to catch the silly case of doing a cpu_switch() to the current process. This happens uncomfortably often. This simplifies a bit of the asm code in cpu_switch (no longer have to call choosethread() in the middle). This has been implemented on i386 and (thanks to jake) sparc64. The others will come soon. This is actually seperate to the lazy switch stuff. Glanced at by: jake, jhb
2003-04-02 23:53:30 +00:00
#include <sys/smp.h>
#endif
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
#if defined(SMP) && defined(SCHED_4BSD)
#include <sys/sysctl.h>
#endif
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
#include <machine/cpu.h>
Restore KTR_CRITICAL but conditionally compile it in as KTR_SCHED. Requested by: scottl, jhb
2005-12-18 18:10:57 +00:00
/* Uncomment this to enable logging of critical_enter/exit. */
#if 0
#define KTR_CRITICAL KTR_SCHED
#else
#define KTR_CRITICAL 0
#endif
Turn PREEMPTION into a kernel option. Make sure that it's defined if FULL_PREEMPTION is defined. Add a runtime warning to ULE if PREEMPTION is enabled (code inspired by the PREEMPTION warning in kern_switch.c). This is a possible MT5 candidate.
2004-09-02 18:59:15 +00:00
#ifdef FULL_PREEMPTION
#ifndef PREEMPTION
#error "The FULL_PREEMPTION option requires the PREEMPTION option"
#endif
#endif
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
Make the run queue parameters machine dependent. Optimize 64 bit architectures by using a 64 bit word for the bit array which keeps track of non-empty queues. Reviewed by: peter
2002-05-25 01:12:23 +00:00
CTASSERT((RQB_BPW * RQB_LEN) == RQ_NQS);
Add a read-only kern.sched.preemption sysctl so that user space can tell if "options PREEMPTION" is compiled into the kernel.
2005-03-20 17:05:12 +00:00
/*
* kern.sched.preemption allows user space to determine if preemption support
* is compiled in or not. It is not currently a boot or runtime flag that
* can be changed.
*/
#ifdef PREEMPTION
static int kern_sched_preemption = 1;
#else
static int kern_sched_preemption = 0;
#endif
SYSCTL_INT(_kern_sched, OID_AUTO, preemption, CTLFLAG_RD,
&kern_sched_preemption, 0, "Kernel preemption enabled");
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
#ifdef SCHED_STATS
long switch_preempt;
long switch_owepreempt;
long switch_turnstile;
long switch_sleepq;
long switch_sleepqtimo;
long switch_relinquish;
long switch_needresched;
static SYSCTL_NODE(_kern_sched, OID_AUTO, stats, CTLFLAG_RW, 0, "switch stats");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, preempt, CTLFLAG_RD, &switch_preempt, 0, "");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, owepreempt, CTLFLAG_RD, &switch_owepreempt, 0, "");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, turnstile, CTLFLAG_RD, &switch_turnstile, 0, "");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, sleepq, CTLFLAG_RD, &switch_sleepq, 0, "");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, sleepqtimo, CTLFLAG_RD, &switch_sleepqtimo, 0, "");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, relinquish, CTLFLAG_RD, &switch_relinquish, 0, "");
SYSCTL_INT(_kern_sched_stats, OID_AUTO, needresched, CTLFLAG_RD, &switch_needresched, 0, "");
static int
sysctl_stats_reset(SYSCTL_HANDLER_ARGS)
{
int error;
int val;
val = 0;
error = sysctl_handle_int(oidp, &val, 0, req);
if (error != 0 || req->newptr == NULL)
return (error);
if (val == 0)
return (0);
switch_preempt = 0;
switch_owepreempt = 0;
switch_turnstile = 0;
switch_sleepq = 0;
switch_sleepqtimo = 0;
switch_relinquish = 0;
switch_needresched = 0;
return (0);
}
SYSCTL_PROC(_kern_sched_stats, OID_AUTO, reset, CTLTYPE_INT | CTLFLAG_WR, NULL,
0, sysctl_stats_reset, "I", "Reset scheduler statistics");
#endif
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
/************************************************************************
* Functions that manipulate runnability from a thread perspective. *
************************************************************************/
Make KSE a kernel option, turned on by default in all GENERIC kernel configs except sun4v (which doesn't process signals properly with KSE). Reviewed by: davidxu@
2006-10-26 21:42:22 +00:00
/*
* Select the thread that will be run next.
*/
KSE Milestone 2 Note ALL MODULES MUST BE RECOMPILED make the kernel aware that there are smaller units of scheduling than the process. (but only allow one thread per process at this time). This is functionally equivalent to teh previousl -current except that there is a thread associated with each process. Sorry john! (your next MFC will be a doosie!) Reviewed by: peter@freebsd.org, dillon@freebsd.org X-MFC after: ha ha ha ha
2001-09-12 08:38:13 +00:00
struct thread *
choosethread(void)
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
{
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
struct thread *td;
Initial landing of SMP support for FreeBSD/amd64. - This is heavily derived from John Baldwin's apic/pci cleanup on i386. - I have completely rewritten or drastically cleaned up some other parts. (in particular, bootstrap) - This is still a WIP. It seems that there are some highly bogus bioses on nVidia nForce3-150 boards. I can't stress how broken these boards are. I have a workaround in mind, but right now the Asus SK8N is broken. The Gigabyte K8NPro (nVidia based) is also mind-numbingly hosed. - Most of my testing has been with SCHED_ULE. SCHED_4BSD works. - the apic and acpi components are 'standard'. - If you have an nVidia nForce3-150 board, you are stuck with 'device atpic' in addition, because they somehow managed to forget to connect the 8254 timer to the apic, even though its in the same silicon! ARGH! This directly violates the ACPI spec.
2003-11-17 08:58:16 +00:00
#if defined(SMP) && (defined(__i386__) || defined(__amd64__))
Commit a partial lazy thread switch mechanism for i386. it isn't as lazy as it could be and can do with some more cleanup. Currently its under options LAZY_SWITCH. What this does is avoid %cr3 reloads for short context switches that do not involve another user process. ie: we can take an interrupt, switch to a kthread and return to the user without explicitly flushing the tlb. However, this isn't as exciting as it could be, the interrupt overhead is still high and too much blocks on Giant still. There are some debug sysctls, for stats and for an on/off switch. The main problem with doing this has been "what if the process that you're running on exits while we're borrowing its address space?" - in this case we use an IPI to give it a kick when we're about to reclaim the pmap. Its not compiled in unless you add the LAZY_SWITCH option. I want to fix a few more things and get some more feedback before turning it on by default. This is NOT a replacement for Bosko's lazy interrupt stuff. This was more meant for the kthread case, while his was for interrupts. Mine helps a little for interrupts, but his helps a lot more. The stats are enabled with options SWTCH_OPTIM_STATS - this has been a pseudo-option for years, I just added a bunch of stuff to it. One non-trivial change was to select a new thread before calling cpu_switch() in the first place. This allows us to catch the silly case of doing a cpu_switch() to the current process. This happens uncomfortably often. This simplifies a bit of the asm code in cpu_switch (no longer have to call choosethread() in the middle). This has been implemented on i386 and (thanks to jake) sparc64. The others will come soon. This is actually seperate to the lazy switch stuff. Glanced at by: jake, jhb
2003-04-02 23:53:30 +00:00
if (smp_active == 0 && PCPU_GET(cpuid) != 0) {
/* Shutting down, run idlethread on AP's */
td = PCPU_GET(idlethread);
CTR1(KTR_RUNQ, "choosethread: td=%p (idle)", td);
TD_SET_RUNNING(td);
return (td);
}
#endif
Allow alphas to do crashdumps: Refuse to run anything in choosethread() after a panic which is not an interrupt thread, or the thread which caused the panic. Also, remove panicstr checks from msleep() and from cv_wait() in order to allow threads to go to sleep and yeild the cpu to the panicing thread, or to an interrupt thread which might be doing the crashdump. Reviewed by: jhb (and it was mostly his idea too)
2002-07-17 02:23:44 +00:00
retry:
- Remove setrunqueue and replace it with direct calls to sched_add(). setrunqueue() was mostly empty. The few asserts and thread state setting were moved to the individual schedulers. sched_add() was chosen to displace it for naming consistency reasons. - Remove adjustrunqueue, it was 4 lines of code that was ifdef'd to be different on all three schedulers where it was only called in one place each. - Remove the long ifdef'd out remrunqueue code. - Remove the now redundant ts_state. Inspect the thread state directly. - Don't set TSF_* flags from kern_switch.c, we were only doing this to support a feature in one scheduler. - Change sched_choose() to return a thread rather than a td_sched. Also, rely on the schedulers to return the idlethread. This simplifies the logic in choosethread(). Aside from the run queue links kern_switch.c mostly does not care about the contents of td_sched. Discussed with: julian - Move the idle thread loop into the per scheduler area. ULE wants to do something different from the other schedulers. Suggested by: jhb Tested on: x86/amd64 sched_{4BSD, ULE, CORE}.
2007-01-23 08:46:51 +00:00
td = sched_choose();
Add code to ddb to allow backtracing an arbitrary thread. (show thread {address}) Remove the IDLE kse state and replace it with a change in the way threads sahre KSEs. Every KSE now has a thread, which is considered its "owner" however a KSE may also be lent to other threads in the same group to allow completion of in-kernel work. n this case the owner remains the same and the KSE will revert to the owner when the other work has been completed. All creations of upcalls etc. is now done from kse_reassign() which in turn is called from mi_switch or thread_exit(). This means that special code can be removed from msleep() and cv_wait(). kse_release() does not leave a KSE with no thread any more but converts the existing thread into teh KSE's owner, and sets it up for doing an upcall. It is just inhibitted from being scheduled until there is some reason to do an upcall. Remove all trace of the kse_idle queue since it is no-longer needed. "Idle" KSEs are now on the loanable queue.
2002-12-28 01:23:07 +00:00
/*
When we are spilling threads out of the run queue during panic, make sure we keep the thread state variable consistent with its real state. i.e. Don't say it's on the run queue when it isn't. Also clarify the associated comment. Turns a double panic back to a single panic :-/ Approved by: re@ (jhb)
2003-05-21 18:53:25 +00:00
* If we are in panic, only allow system threads,
* plus the one we are running in, to be run.
Add code to ddb to allow backtracing an arbitrary thread. (show thread {address}) Remove the IDLE kse state and replace it with a change in the way threads sahre KSEs. Every KSE now has a thread, which is considered its "owner" however a KSE may also be lent to other threads in the same group to allow completion of in-kernel work. n this case the owner remains the same and the KSE will revert to the owner when the other work has been completed. All creations of upcalls etc. is now done from kse_reassign() which in turn is called from mi_switch or thread_exit(). This means that special code can be removed from msleep() and cv_wait(). kse_release() does not leave a KSE with no thread any more but converts the existing thread into teh KSE's owner, and sets it up for doing an upcall. It is just inhibitted from being scheduled until there is some reason to do an upcall. Remove all trace of the kse_idle queue since it is no-longer needed. "Idle" KSEs are now on the loanable queue.
2002-12-28 01:23:07 +00:00
*/
Allow alphas to do crashdumps: Refuse to run anything in choosethread() after a panic which is not an interrupt thread, or the thread which caused the panic. Also, remove panicstr checks from msleep() and from cv_wait() in order to allow threads to go to sleep and yeild the cpu to the panicing thread, or to an interrupt thread which might be doing the crashdump. Reviewed by: jhb (and it was mostly his idea too)
2002-07-17 02:23:44 +00:00
if (panicstr && ((td->td_proc->p_flag & P_SYSTEM) == 0 &&
When we are spilling threads out of the run queue during panic, make sure we keep the thread state variable consistent with its real state. i.e. Don't say it's on the run queue when it isn't. Also clarify the associated comment. Turns a double panic back to a single panic :-/ Approved by: re@ (jhb)
2003-05-21 18:53:25 +00:00
(td->td_flags & TDF_INPANIC) == 0)) {
/* note that it is no longer on the run queue */
TD_SET_CAN_RUN(td);
Allow alphas to do crashdumps: Refuse to run anything in choosethread() after a panic which is not an interrupt thread, or the thread which caused the panic. Also, remove panicstr checks from msleep() and from cv_wait() in order to allow threads to go to sleep and yeild the cpu to the panicing thread, or to an interrupt thread which might be doing the crashdump. Reviewed by: jhb (and it was mostly his idea too)
2002-07-17 02:23:44 +00:00
goto retry;
When we are spilling threads out of the run queue during panic, make sure we keep the thread state variable consistent with its real state. i.e. Don't say it's on the run queue when it isn't. Also clarify the associated comment. Turns a double panic back to a single panic :-/ Approved by: re@ (jhb)
2003-05-21 18:53:25 +00:00
}
Add code to ddb to allow backtracing an arbitrary thread. (show thread {address}) Remove the IDLE kse state and replace it with a change in the way threads sahre KSEs. Every KSE now has a thread, which is considered its "owner" however a KSE may also be lent to other threads in the same group to allow completion of in-kernel work. n this case the owner remains the same and the KSE will revert to the owner when the other work has been completed. All creations of upcalls etc. is now done from kse_reassign() which in turn is called from mi_switch or thread_exit(). This means that special code can be removed from msleep() and cv_wait(). kse_release() does not leave a KSE with no thread any more but converts the existing thread into teh KSE's owner, and sets it up for doing an upcall. It is just inhibitted from being scheduled until there is some reason to do an upcall. Remove all trace of the kse_idle queue since it is no-longer needed. "Idle" KSEs are now on the loanable queue.
2002-12-28 01:23:07 +00:00
Completely redo thread states. Reviewed by: davidxu@freebsd.org
2002-09-11 08:13:56 +00:00
TD_SET_RUNNING(td);
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
return (td);
}
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
/*
* Kernel thread preemption implementation. Critical sections mark
* regions of code in which preemptions are not allowed.
*/
Modify the critical section API as follows: - The MD functions critical_enter/exit are renamed to start with a cpu_ prefix. - MI wrapper functions critical_enter/exit maintain a per-thread nesting count and a per-thread critical section saved state set when entering a critical section while at nesting level 0 and restored when exiting to nesting level 0. This moves the saved state out of spin mutexes so that interlocking spin mutexes works properly. - Most low-level MD code that used critical_enter/exit now use cpu_critical_enter/exit. MI code such as device drivers and spin mutexes use the MI wrappers. Note that since the MI wrappers store the state in the current thread, they do not have any return values or arguments. - mtx_intr_enable() is replaced with a constant CRITICAL_FORK which is assigned to curthread->td_savecrit during fork_exit(). Tested on: i386, alpha
2001-12-18 00:27:18 +00:00
void
critical_enter(void)
{
struct thread *td;
td = curthread;
td->td_critnest++;
Restore KTR_CRITICAL but conditionally compile it in as KTR_SCHED. Requested by: scottl, jhb
2005-12-18 18:10:57 +00:00
CTR4(KTR_CRITICAL, "critical_enter by thread %p (%ld, %s) to %d", td,
Add basic critical section tracing to KTR using event type KTR_CRITICAL. This generates a KTR event for each critical section entered and exited. It would be desirable to also log the filename and line number of the source entering or exiting the critical section, but this requires hacking up the critical section API, so I've not done that yet.
2004-11-07 23:11:32 +00:00
(long)td->td_proc->p_pid, td->td_proc->p_comm, td->td_critnest);
Modify the critical section API as follows: - The MD functions critical_enter/exit are renamed to start with a cpu_ prefix. - MI wrapper functions critical_enter/exit maintain a per-thread nesting count and a per-thread critical section saved state set when entering a critical section while at nesting level 0 and restored when exiting to nesting level 0. This moves the saved state out of spin mutexes so that interlocking spin mutexes works properly. - Most low-level MD code that used critical_enter/exit now use cpu_critical_enter/exit. MI code such as device drivers and spin mutexes use the MI wrappers. Note that since the MI wrappers store the state in the current thread, they do not have any return values or arguments. - mtx_intr_enable() is replaced with a constant CRITICAL_FORK which is assigned to curthread->td_savecrit during fork_exit(). Tested on: i386, alpha
2001-12-18 00:27:18 +00:00
}
void
critical_exit(void)
{
struct thread *td;
td = curthread;
- style fixes to the critical_exit() KASSERT(). Submitted by: bde
2004-02-02 08:13:27 +00:00
KASSERT(td->td_critnest != 0,
("critical_exit: td_critnest == 0"));
Use low level constructs borrowed from interrupt threads to wait for work in proc0. Remove the TDP_WAKEPROC0 workaround.
2005-05-23 23:01:53 +00:00
#ifdef PREEMPTION
Modify the critical section API as follows: - The MD functions critical_enter/exit are renamed to start with a cpu_ prefix. - MI wrapper functions critical_enter/exit maintain a per-thread nesting count and a per-thread critical section saved state set when entering a critical section while at nesting level 0 and restored when exiting to nesting level 0. This moves the saved state out of spin mutexes so that interlocking spin mutexes works properly. - Most low-level MD code that used critical_enter/exit now use cpu_critical_enter/exit. MI code such as device drivers and spin mutexes use the MI wrappers. Note that since the MI wrappers store the state in the current thread, they do not have any return values or arguments. - mtx_intr_enable() is replaced with a constant CRITICAL_FORK which is assigned to curthread->td_savecrit during fork_exit(). Tested on: i386, alpha
2001-12-18 00:27:18 +00:00
if (td->td_critnest == 1) {
Sprinkle some volatile magic and rearrange things a bit to avoid race conditions in critical_exit now that it no longer blocks interrupts. Reviewed by: jhb
2005-04-08 03:37:53 +00:00
td->td_critnest = 0;
if (td->td_owepreempt) {
td->td_critnest = 1;
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
thread_lock(td);
Sprinkle some volatile magic and rearrange things a bit to avoid race conditions in critical_exit now that it no longer blocks interrupts. Reviewed by: jhb
2005-04-08 03:37:53 +00:00
td->td_critnest--;
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
SCHED_STAT_INC(switch_owepreempt);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
mi_switch(SW_INVOL, NULL);
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
thread_unlock(td);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
}
Trim trailing whitespace.
2005-12-28 17:13:31 +00:00
} else
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
#endif
Modify the critical section API as follows: - The MD functions critical_enter/exit are renamed to start with a cpu_ prefix. - MI wrapper functions critical_enter/exit maintain a per-thread nesting count and a per-thread critical section saved state set when entering a critical section while at nesting level 0 and restored when exiting to nesting level 0. This moves the saved state out of spin mutexes so that interlocking spin mutexes works properly. - Most low-level MD code that used critical_enter/exit now use cpu_critical_enter/exit. MI code such as device drivers and spin mutexes use the MI wrappers. Note that since the MI wrappers store the state in the current thread, they do not have any return values or arguments. - mtx_intr_enable() is replaced with a constant CRITICAL_FORK which is assigned to curthread->td_savecrit during fork_exit(). Tested on: i386, alpha
2001-12-18 00:27:18 +00:00
td->td_critnest--;
Trim trailing whitespace.
2005-12-28 17:13:31 +00:00
Restore KTR_CRITICAL but conditionally compile it in as KTR_SCHED. Requested by: scottl, jhb
2005-12-18 18:10:57 +00:00
CTR4(KTR_CRITICAL, "critical_exit by thread %p (%ld, %s) to %d", td,
Add basic critical section tracing to KTR using event type KTR_CRITICAL. This generates a KTR event for each critical section entered and exited. It would be desirable to also log the filename and line number of the source entering or exiting the critical section, but this requires hacking up the critical section API, so I've not done that yet.
2004-11-07 23:11:32 +00:00
(long)td->td_proc->p_pid, td->td_proc->p_comm, td->td_critnest);
Modify the critical section API as follows: - The MD functions critical_enter/exit are renamed to start with a cpu_ prefix. - MI wrapper functions critical_enter/exit maintain a per-thread nesting count and a per-thread critical section saved state set when entering a critical section while at nesting level 0 and restored when exiting to nesting level 0. This moves the saved state out of spin mutexes so that interlocking spin mutexes works properly. - Most low-level MD code that used critical_enter/exit now use cpu_critical_enter/exit. MI code such as device drivers and spin mutexes use the MI wrappers. Note that since the MI wrappers store the state in the current thread, they do not have any return values or arguments. - mtx_intr_enable() is replaced with a constant CRITICAL_FORK which is assigned to curthread->td_savecrit during fork_exit(). Tested on: i386, alpha
2001-12-18 00:27:18 +00:00
}
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
/*
* This function is called when a thread is about to be put on run queue
* because it has been made runnable or its priority has been adjusted. It
* determines if the new thread should be immediately preempted to. If so,
* it switches to it and eventually returns true. If not, it returns false
* so that the caller may place the thread on an appropriate run queue.
*/
int
maybe_preempt(struct thread *td)
{
Unbreak build for the the !PREEMPTION case: don't define variables that aren't used in that case.
2004-07-03 00:57:43 +00:00
#ifdef PREEMPTION
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
struct thread *ctd;
int cpri, pri;
Unbreak build for the the !PREEMPTION case: don't define variables that aren't used in that case.
2004-07-03 00:57:43 +00:00
#endif
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
#ifdef PREEMPTION
/*
* The new thread should not preempt the current thread if any of the
* following conditions are true:
*
A further step on the journey of meaking panics and debugging more reliable: in the window between the beginning of panic() and entering the debugger, it's possible to receive interrupts. If we receive an interrupt, don't preempt if panicstr != NULL, as the system is in the process of failing, and the preempting thread is likely to stumble over the failure. The typical scenario is during the printf() in panic() prior to entering the debugger, but when running with a slower console type such as serial console. It could be that the panic string should be passed to the debugger to print, so that it can run from the debugger's environment rather than a regular kernel printf. Glanced at by: jhb
2005-03-17 15:18:01 +00:00
* - The kernel is in the throes of crashing (panicstr).
- Move TDF_OWEPREEMPT, TDF_OWEUPC, and TDF_USTATCLOCK over to td_pflags since they are only accessed by curthread and thus do not need any locking. - Move pr_addr and pr_ticks out of struct uprof (which is per-process) and directly into struct thread as td_profil_addr and td_profil_ticks as these variables are really per-thread. (They are used to defer an addupc_intr() that was too "hard" until ast()).
2004-07-16 21:04:55 +00:00
* - The current thread has a higher (numerically lower) or
* equivalent priority. Note that this prevents curthread from
* trying to preempt to itself.
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
* - It is too early in the boot for context switches (cold is set).
* - The current thread has an inhibitor set or is in the process of
* exiting. In this case, the current thread is about to switch
* out anyways, so there's no point in preempting. If we did,
* the current thread would not be properly resumed as well, so
* just avoid that whole landmine.
* - If the new thread's priority is not a realtime priority and
* the current thread's priority is not an idle priority and
* FULL_PREEMPTION is disabled.
*
* If all of these conditions are false, but the current thread is in
* a nested critical section, then we have to defer the preemption
* until we exit the critical section. Otherwise, switch immediately
* to the new thread.
*/
ctd = curthread;
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
THREAD_LOCK_ASSERT(td, MA_OWNED);
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
KASSERT ((ctd->td_sched != NULL && ctd->td_sched->ts_thread == ctd),
Don't do IPIs on behalf of interrupt threads. just punt straight on through to teh preemption code. Make a KASSSERT out of a condition that can no longer occur. MFC after: 1 week
2004-09-06 07:23:14 +00:00
("thread has no (or wrong) sched-private part."));
Add some kasserts
2004-09-13 23:02:52 +00:00
KASSERT((td->td_inhibitors == 0),
Prefer a more traditional spelling of inhibited in comments and panic messages.
2006-12-31 15:56:04 +00:00
("maybe_preempt: trying to run inhibited thread"));
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
pri = td->td_priority;
cpri = ctd->td_priority;
A further step on the journey of meaking panics and debugging more reliable: in the window between the beginning of panic() and entering the debugger, it's possible to receive interrupts. If we receive an interrupt, don't preempt if panicstr != NULL, as the system is in the process of failing, and the preempting thread is likely to stumble over the failure. The typical scenario is during the printf() in panic() prior to entering the debugger, but when running with a slower console type such as serial console. It could be that the panic string should be passed to the debugger to print, so that it can run from the debugger's environment rather than a regular kernel printf. Glanced at by: jhb
2005-03-17 15:18:01 +00:00
if (panicstr != NULL || pri >= cpri || cold /* || dumping */ ||
- Remove setrunqueue and replace it with direct calls to sched_add(). setrunqueue() was mostly empty. The few asserts and thread state setting were moved to the individual schedulers. sched_add() was chosen to displace it for naming consistency reasons. - Remove adjustrunqueue, it was 4 lines of code that was ifdef'd to be different on all three schedulers where it was only called in one place each. - Remove the long ifdef'd out remrunqueue code. - Remove the now redundant ts_state. Inspect the thread state directly. - Don't set TSF_* flags from kern_switch.c, we were only doing this to support a feature in one scheduler. - Change sched_choose() to return a thread rather than a td_sched. Also, rely on the schedulers to return the idlethread. This simplifies the logic in choosethread(). Aside from the run queue links kern_switch.c mostly does not care about the contents of td_sched. Discussed with: julian - Move the idle thread loop into the per scheduler area. ULE wants to do something different from the other schedulers. Suggested by: jhb Tested on: x86/amd64 sched_{4BSD, ULE, CORE}.
2007-01-23 08:46:51 +00:00
TD_IS_INHIBITED(ctd))
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
return (0);
#ifndef FULL_PREEMPTION
Restore preemption of idle threads. Submitted by: jhb
2005-06-10 03:00:29 +00:00
if (pri > PRI_MAX_ITHD && cpri < PRI_MIN_IDLE)
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
return (0);
#endif
Lots of whitespace cleanup. Fix for broken if condition. Submitted by: nate@
2005-06-09 19:43:08 +00:00
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
if (ctd->td_critnest > 1) {
CTR1(KTR_PROC, "maybe_preempt: in critical section %d",
ctd->td_critnest);
Sprinkle some volatile magic and rearrange things a bit to avoid race conditions in critical_exit now that it no longer blocks interrupts. Reviewed by: jhb
2005-04-08 03:37:53 +00:00
ctd->td_owepreempt = 1;
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
return (0);
}
/*
When preempting a thread, put it back on the HEAD of its run queue. (Only really implemented in 4bsd) MFC after: 4 days
2004-10-05 22:03:10 +00:00
* Thread is runnable but not yet put on system run queue.
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
*/
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
MPASS(ctd->td_lock == &sched_lock);
MPASS(td->td_lock == &sched_lock);
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
MPASS(TD_ON_RUNQ(td));
TD_SET_RUNNING(td);
CTR3(KTR_PROC, "preempting to thread %p (pid %d, %s)\n", td,
td->td_proc->p_pid, td->td_proc->p_comm);
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
SCHED_STAT_INC(switch_preempt);
When preempting a thread, put it back on the HEAD of its run queue. (Only really implemented in 4bsd) MFC after: 4 days
2004-10-05 22:03:10 +00:00
mi_switch(SW_INVOL|SW_PREEMPT, td);
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
/*
* td's lock pointer may have changed. We have to return with it
* locked.
*/
spinlock_enter();
thread_unlock(ctd);
thread_lock(td);
spinlock_exit();
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
return (1);
#else
return (0);
#endif
}
Don't scare users with a warning about preemption being off when it isn't yet safe to have on by default.
2004-08-06 15:49:44 +00:00
#if 0
Implement preemption of kernel threads natively in the scheduler rather than as one-off hacks in various other parts of the kernel: - Add a function maybe_preempt() that is called from sched_add() to determine if a thread about to be added to a run queue should be preempted to directly. If it is not safe to preempt or if the new thread does not have a high enough priority, then the function returns false and sched_add() adds the thread to the run queue. If the thread should be preempted to but the current thread is in a nested critical section, then the flag TDF_OWEPREEMPT is set and the thread is added to the run queue. Otherwise, mi_switch() is called immediately and the thread is never added to the run queue since it is switch to directly. When exiting an outermost critical section, if TDF_OWEPREEMPT is set, then clear it and call mi_switch() to perform the deferred preemption. - Remove explicit preemption from ithread_schedule() as calling setrunqueue() now does all the correct work. This also removes the do_switch argument from ithread_schedule(). - Do not use the manual preemption code in mtx_unlock if the architecture supports native preemption. - Don't call mi_switch() in a loop during shutdown to give ithreads a chance to run if the architecture supports native preemption since the ithreads will just preempt DELAY(). - Don't call mi_switch() from the page zeroing idle thread for architectures that support native preemption as it is unnecessary. - Native preemption is enabled on the same archs that supported ithread preemption, namely alpha, i386, and amd64. This change should largely be a NOP for the default case as committed except that we will do fewer context switches in a few cases and will avoid the run queues completely when preempting. Approved by: scottl (with his re@ hat)
2004-07-02 20:21:44 +00:00
#ifndef PREEMPTION
/* XXX: There should be a non-static version of this. */
static void
printf_caddr_t(void *data)
{
printf("%s", (char *)data);
}
static char preempt_warning[] =
"WARNING: Kernel preemption is disabled, expect reduced performance.\n";
SYSINIT(preempt_warning, SI_SUB_COPYRIGHT, SI_ORDER_ANY, printf_caddr_t,
preempt_warning)
#endif
Don't scare users with a warning about preemption being off when it isn't yet safe to have on by default.
2004-08-06 15:49:44 +00:00
#endif
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
/************************************************************************
* SYSTEM RUN QUEUE manipulations and tests *
************************************************************************/
/*
* Initialize a run structure.
*/
void
runq_init(struct runq *rq)
{
int i;
bzero(rq, sizeof *rq);
for (i = 0; i < RQ_NQS; i++)
TAILQ_INIT(&rq->rq_queues[i]);
}
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
/*
* Clear the status bit of the queue corresponding to priority level pri,
* indicating that it is empty.
*/
static __inline void
runq_clrbit(struct runq *rq, int pri)
{
struct rqbits *rqb;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
rqb = &rq->rq_status;
CTR4(KTR_RUNQ, "runq_clrbit: bits=%#x %#x bit=%#x word=%d",
rqb->rqb_bits[RQB_WORD(pri)],
rqb->rqb_bits[RQB_WORD(pri)] & ~RQB_BIT(pri),
RQB_BIT(pri), RQB_WORD(pri));
rqb->rqb_bits[RQB_WORD(pri)] &= ~RQB_BIT(pri);
}
/*
* Find the index of the first non-empty run queue. This is done by
* scanning the status bits, a set bit indicates a non-empty queue.
*/
static __inline int
runq_findbit(struct runq *rq)
{
struct rqbits *rqb;
int pri;
int i;
rqb = &rq->rq_status;
for (i = 0; i < RQB_LEN; i++)
if (rqb->rqb_bits[i]) {
Move the "- 1" into the RQB_FFS(mask) macro itself so that implementations can provide a base zero ffs function if they wish. This changes #define RQB_FFS(mask) (ffs64(mask)) foo = RQB_FFS(mask) - 1; to #define RQB_FFS(mask) (ffs64(mask) - 1) foo = RQB_FFS(mask); On some platforms we can get the "- 1" for free, eg: those that use the C code for ffs64(). Reviewed by: jake (in principle)
2002-06-20 06:21:20 +00:00
pri = RQB_FFS(rqb->rqb_bits[i]) + (i << RQB_L2BPW);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
CTR3(KTR_RUNQ, "runq_findbit: bits=%#x i=%d pri=%d",
rqb->rqb_bits[i], i, pri);
return (pri);
}
return (-1);
}
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
static __inline int
- Change types for necent runq additions to u_char rather than int. - Fix these types in ULE as well. This fixes bugs in priority index calculations in certain edge cases. (int)-1 % 64 != (uint)-1 % 64. Reported by: kkenn using pho's stress2.
2007-02-08 01:52:25 +00:00
runq_findbit_from(struct runq *rq, u_char start)
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
{
struct rqbits *rqb;
int bit;
int pri;
int i;
rqb = &rq->rq_status;
bit = start & (RQB_BPW -1);
pri = 0;
CTR1(KTR_RUNQ, "runq_findbit_from: start %d", start);
again:
for (i = RQB_WORD(start); i < RQB_LEN; i++) {
CTR3(KTR_RUNQ, "runq_findbit_from: bits %d = %#x bit = %d",
i, rqb->rqb_bits[i], bit);
if (rqb->rqb_bits[i]) {
if (bit != 0) {
for (pri = bit; pri < RQB_BPW; pri++)
if (rqb->rqb_bits[i] & (1ul << pri))
break;
bit = 0;
if (pri >= RQB_BPW)
continue;
} else
pri = RQB_FFS(rqb->rqb_bits[i]);
pri += (i << RQB_L2BPW);
CTR3(KTR_RUNQ, "runq_findbit_from: bits=%#x i=%d pri=%d",
rqb->rqb_bits[i], i, pri);
return (pri);
}
bit = 0;
}
if (start != 0) {
CTR0(KTR_RUNQ, "runq_findbit_from: restarting");
start = 0;
goto again;
}
return (-1);
}
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
/*
* Set the status bit of the queue corresponding to priority level pri,
* indicating that it is non-empty.
*/
static __inline void
runq_setbit(struct runq *rq, int pri)
{
struct rqbits *rqb;
rqb = &rq->rq_status;
CTR4(KTR_RUNQ, "runq_setbit: bits=%#x %#x bit=%#x word=%d",
rqb->rqb_bits[RQB_WORD(pri)],
rqb->rqb_bits[RQB_WORD(pri)] | RQB_BIT(pri),
RQB_BIT(pri), RQB_WORD(pri));
rqb->rqb_bits[RQB_WORD(pri)] |= RQB_BIT(pri);
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
}
/*
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
* Add the thread to the queue specified by its priority, and set the
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
* corresponding status bit.
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
*/
void
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
runq_add(struct runq *rq, struct td_sched *ts, int flags)
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
{
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
struct rqhead *rqh;
int pri;
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
pri = ts->ts_thread->td_priority / RQ_PPQ;
ts->ts_rqindex = pri;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
runq_setbit(rq, pri);
rqh = &rq->rq_queues[pri];
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
CTR5(KTR_RUNQ, "runq_add: td=%p ts=%p pri=%d %d rqh=%p",
ts->ts_thread, ts, ts->ts_thread->td_priority, pri, rqh);
When preempting a thread, put it back on the HEAD of its run queue. (Only really implemented in 4bsd) MFC after: 4 days
2004-10-05 22:03:10 +00:00
if (flags & SRQ_PREEMPTED) {
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
TAILQ_INSERT_HEAD(rqh, ts, ts_procq);
When preempting a thread, put it back on the HEAD of its run queue. (Only really implemented in 4bsd) MFC after: 4 days
2004-10-05 22:03:10 +00:00
} else {
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
TAILQ_INSERT_TAIL(rqh, ts, ts_procq);
When preempting a thread, put it back on the HEAD of its run queue. (Only really implemented in 4bsd) MFC after: 4 days
2004-10-05 22:03:10 +00:00
}
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
}
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
void
- Change types for necent runq additions to u_char rather than int. - Fix these types in ULE as well. This fixes bugs in priority index calculations in certain edge cases. (int)-1 % 64 != (uint)-1 % 64. Reported by: kkenn using pho's stress2.
2007-02-08 01:52:25 +00:00
runq_add_pri(struct runq *rq, struct td_sched *ts, u_char pri, int flags)
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
{
struct rqhead *rqh;
KASSERT(pri < RQ_NQS, ("runq_add_pri: %d out of range", pri));
ts->ts_rqindex = pri;
runq_setbit(rq, pri);
rqh = &rq->rq_queues[pri];
CTR5(KTR_RUNQ, "runq_add_pri: td=%p ke=%p pri=%d idx=%d rqh=%p",
ts->ts_thread, ts, ts->ts_thread->td_priority, pri, rqh);
if (flags & SRQ_PREEMPTED) {
TAILQ_INSERT_HEAD(rqh, ts, ts_procq);
} else {
TAILQ_INSERT_TAIL(rqh, ts, ts_procq);
}
}
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
/*
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
* Return true if there are runnable processes of any priority on the run
* queue, false otherwise. Has no side effects, does not modify the run
* queue structure.
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
*/
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
int
runq_check(struct runq *rq)
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
{
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
struct rqbits *rqb;
int i;
rqb = &rq->rq_status;
for (i = 0; i < RQB_LEN; i++)
if (rqb->rqb_bits[i]) {
CTR2(KTR_RUNQ, "runq_check: bits=%#x i=%d",
rqb->rqb_bits[i], i);
return (1);
}
CTR0(KTR_RUNQ, "runq_check: empty");
return (0);
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
}
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
#if defined(SMP) && defined(SCHED_4BSD)
int runq_fuzz = 1;
SYSCTL_INT(_kern_sched, OID_AUTO, runq_fuzz, CTLFLAG_RW, &runq_fuzz, 0, "");
#endif
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
/*
- Create a new scheduler api that is defined in sys/sched.h - Begin moving scheduler specific functionality into sched_4bsd.c - Replace direct manipulation of scheduler data with hooks provided by the new api. - Remove KSE specific state modifications and single runq assumptions from kern_switch.c Reviewed by: -arch
2002-10-12 05:32:24 +00:00
* Find the highest priority process on the run queue.
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
*/
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
struct td_sched *
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
runq_choose(struct runq *rq)
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
{
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
struct rqhead *rqh;
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
struct td_sched *ts;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
int pri;
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
while ((pri = runq_findbit(rq)) != -1) {
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
rqh = &rq->rq_queues[pri];
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
#if defined(SMP) && defined(SCHED_4BSD)
/* fuzz == 1 is normal.. 0 or less are ignored */
if (runq_fuzz > 1) {
/*
* In the first couple of entries, check if
* there is one for our CPU as a preference.
*/
int count = runq_fuzz;
int cpu = PCPU_GET(cpuid);
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
struct td_sched *ts2;
ts2 = ts = TAILQ_FIRST(rqh);
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
while (count-- && ts2) {
if (ts->ts_thread->td_lastcpu == cpu) {
ts = ts2;
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
break;
}
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
ts2 = TAILQ_NEXT(ts2, ts_procq);
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
}
Trim trailing whitespace.
2005-12-28 17:13:31 +00:00
} else
Give the 4bsd scheduler the ability to wake up idle processors when there is new work to be done. MFC after: 5 days
2004-09-01 06:42:02 +00:00
#endif
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
ts = TAILQ_FIRST(rqh);
KASSERT(ts != NULL, ("runq_choose: no proc on busy queue"));
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
CTR3(KTR_RUNQ,
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
"runq_choose: pri=%d td_sched=%p rqh=%p", pri, ts, rqh);
return (ts);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
}
CTR1(KTR_RUNQ, "runq_choose: idleproc pri=%d", pri);
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
return (NULL);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
}
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
struct td_sched *
- Change types for necent runq additions to u_char rather than int. - Fix these types in ULE as well. This fixes bugs in priority index calculations in certain edge cases. (int)-1 % 64 != (uint)-1 % 64. Reported by: kkenn using pho's stress2.
2007-02-08 01:52:25 +00:00
runq_choose_from(struct runq *rq, u_char idx)
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
{
struct rqhead *rqh;
struct td_sched *ts;
int pri;
- Don't pass a pointer into runq_choose_from(). The caller can adjust the index if it chooses to.
2007-01-04 12:10:58 +00:00
if ((pri = runq_findbit_from(rq, idx)) != -1) {
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
rqh = &rq->rq_queues[pri];
ts = TAILQ_FIRST(rqh);
KASSERT(ts != NULL, ("runq_choose: no proc on busy queue"));
CTR4(KTR_RUNQ,
"runq_choose_from: pri=%d kse=%p idx=%d rqh=%p",
pri, ts, ts->ts_rqindex, rqh);
return (ts);
}
CTR1(KTR_RUNQ, "runq_choose_from: idleproc pri=%d", pri);
return (NULL);
}
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
/*
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
* Remove the thread from the queue specified by its priority, and clear the
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
* corresponding status bit if the queue becomes empty.
- Remove setrunqueue and replace it with direct calls to sched_add(). setrunqueue() was mostly empty. The few asserts and thread state setting were moved to the individual schedulers. sched_add() was chosen to displace it for naming consistency reasons. - Remove adjustrunqueue, it was 4 lines of code that was ifdef'd to be different on all three schedulers where it was only called in one place each. - Remove the long ifdef'd out remrunqueue code. - Remove the now redundant ts_state. Inspect the thread state directly. - Don't set TSF_* flags from kern_switch.c, we were only doing this to support a feature in one scheduler. - Change sched_choose() to return a thread rather than a td_sched. Also, rely on the schedulers to return the idlethread. This simplifies the logic in choosethread(). Aside from the run queue links kern_switch.c mostly does not care about the contents of td_sched. Discussed with: julian - Move the idle thread loop into the per scheduler area. ULE wants to do something different from the other schedulers. Suggested by: jhb Tested on: x86/amd64 sched_{4BSD, ULE, CORE}.
2007-01-23 08:46:51 +00:00
* Caller must set state afterwards.
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
*/
void
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
runq_remove(struct runq *rq, struct td_sched *ts)
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
{
runq_remove_idx(rq, ts, NULL);
}
void
- Change types for necent runq additions to u_char rather than int. - Fix these types in ULE as well. This fixes bugs in priority index calculations in certain edge cases. (int)-1 % 64 != (uint)-1 % 64. Reported by: kkenn using pho's stress2.
2007-02-08 01:52:25 +00:00
runq_remove_idx(struct runq *rq, struct td_sched *ts, u_char *idx)
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
{
struct rqhead *rqh;
- Change types for necent runq additions to u_char rather than int. - Fix these types in ULE as well. This fixes bugs in priority index calculations in certain edge cases. (int)-1 % 64 != (uint)-1 % 64. Reported by: kkenn using pho's stress2.
2007-02-08 01:52:25 +00:00
u_char pri;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
KASSERT(ts->ts_thread->td_proc->p_sflag & PS_INMEM,
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
("runq_remove_idx: process swapped out"));
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
pri = ts->ts_rqindex;
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
KASSERT(pri < RQ_NQS, ("runq_remove_idx: Invalid index %d\n", pri));
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
rqh = &rq->rq_queues[pri];
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
CTR5(KTR_RUNQ, "runq_remove_idx: td=%p, ts=%p pri=%d %d rqh=%p",
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
ts->ts_thread, ts, ts->ts_thread->td_priority, pri, rqh);
Commit 1/14 of sched_lock decomposition. - Move all scheduler locking into the schedulers utilizing a technique similar to solaris's container locking. - A per-process spinlock is now used to protect the queue of threads, thread count, suspension count, p_sflags, and other process related scheduling fields. - The new thread lock is actually a pointer to a spinlock for the container that the thread is currently owned by. The container may be a turnstile, sleepqueue, or run queue. - thread_lock() is now used to protect access to thread related scheduling fields. thread_unlock() unlocks the lock and thread_set_lock() implements the transition from one lock to another. - A new "blocked_lock" is used in cases where it is not safe to hold the actual thread's lock yet we must prevent access to the thread. - sched_throw() and sched_fork_exit() are introduced to allow the schedulers to fix-up locking at these points. - Add some minor infrastructure for optionally exporting scheduler statistics that were invaluable in solving performance problems with this patch. Generally these statistics allow you to differentiate between different causes of context switches. Tested by: kris, current@ Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc. Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
2007-06-04 23:50:30 +00:00
{
struct td_sched *nts;
TAILQ_FOREACH(nts, rqh, ts_procq)
if (nts == ts)
break;
if (ts != nts)
panic("runq_remove_idx: ts %p not on rqindex %d",
ts, pri);
}
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
TAILQ_REMOVE(rqh, ts, ts_procq);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if (TAILQ_EMPTY(rqh)) {
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
CTR0(KTR_RUNQ, "runq_remove_idx: empty");
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
runq_clrbit(rq, pri);
- Add three new functions to support circular run queues. - runq_add_pri allows the caller to position the thread at any rqindex regardless of priority. - runq_choose_from() chooses the lowest priority thread starting from a given index. The index is updated with the rqindex of the chosen thread. This routine is used to pick the lowest priority relative to a given index. - runq_remove_idx() updates the index if the run queue that held the removed thread is now empty.
2007-01-04 08:39:58 +00:00
if (idx != NULL && *idx == pri)
*idx = (pri + 1) % RQ_NQS;
Extract the next runnable process selection out of cpu_switch() into a fairly machine independent C routine. gcc actually does a pretty good job of this. Reviewed by: msmith (in principle)
1999-08-19 00:06:53 +00:00
}
Part 1 of KSE-III The ability to schedule multiple threads per process (one one cpu) by making ALL system calls optionally asynchronous. to come: ia64 and power-pc patches, patches for gdb, test program (in tools) Reviewed by: Almost everyone who counts (at various times, peter, jhb, matt, alfred, mini, bernd, and a cast of thousands) NOTE: this is still Beta code, and contains lots of debugging stuff. expect slight instability in signals..
2002-06-29 17:26:22 +00:00
}
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
/****** functions that are temporarily here ***********/
#include <vm/uma.h>
extern struct mtx kse_zombie_lock;
/*
* Allocate scheduler specific per-process resources.
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
* The thread and proc have already been linked in.
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
*
* Called from:
* proc_init() (UMA init method)
*/
void
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
sched_newproc(struct proc *p, struct thread *td)
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
{
}
/*
* thread is being either created or recycled.
* Fix up the per-scheduler resources associated with it.
* Called from:
* sched_fork_thread()
* thread_dtor() (*may go away)
* thread_init() (*may go away)
*/
void
sched_newthread(struct thread *td)
{
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
struct td_sched *ts;
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
Threading cleanup.. part 2 of several. Make part of John Birrell's KSE patch permanent.. Specifically, remove: Any reference of the ksegrp structure. This feature was never fully utilised and made things overly complicated. All code in the scheduler that tried to make threaded programs fair to unthreaded programs. Libpthread processes will already do this to some extent and libthr processes already disable it. Also: Since this makes such a big change to the scheduler(s), take the opportunity to rename some structures and elements that had to be moved anyhow. This makes the code a lot more readable. The ULE scheduler compiles again but I have no idea if it works. The 4bsd scheduler still reqires a little cleaning and some functions that now do ALMOST nothing will go away, but I thought I'd do that as a separate commit. Tested by David Xu, and Dan Eischen using libthr and libpthread.
2006-12-06 06:34:57 +00:00
ts = (struct td_sched *) (td + 1);
bzero(ts, sizeof(*ts));
td->td_sched = ts;
ts->ts_thread = td;
Refactor a bunch of scheduler code to give basically the same behaviour but with slightly cleaned up interfaces. The KSE structure has become the same as the "per thread scheduler private data" structure. In order to not make the diffs too great one is #defined as the other at this time. The KSE (or td_sched) structure is now allocated per thread and has no allocation code of its own. Concurrency for a KSEGRP is now kept track of via a simple pair of counters rather than using KSE structures as tokens. Since the KSE structure is different in each scheduler, kern_switch.c is now included at the end of each scheduler. Nothing outside the scheduler knows the contents of the KSE (aka td_sched) structure. The fields in the ksegrp structure that are to do with the scheduler's queueing mechanisms are now moved to the kg_sched structure. (per ksegrp scheduler private data structure). In other words how the scheduler queues and keeps track of threads is no-one's business except the scheduler's. This should allow people to write experimental schedulers with completely different internal structuring. A scheduler call sched_set_concurrency(kg, N) has been added that notifies teh scheduler that no more than N threads from that ksegrp should be allowed to be on concurrently scheduled. This is also used to enforce 'fainess' at this time so that a ksegrp with 10000 threads can not swamp a the run queue and force out a process with 1 thread, since the current code will not set the concurrency above NCPU, and both schedulers will not allow more than that many onto the system run queue at a time. Each scheduler should eventualy develop their own methods to do this now that they are effectively separated. Rejig libthr's kernel interface to follow the same code paths as linkse for scope system threads. This has slightly hurt libthr's performance but I will work to recover as much of it as I can. Thread exit code has been cleaned up greatly. exit and exec code now transitions a process back to 'standard non-threaded mode' before taking the next step. Reviewed by: scottl, peter MFC after: 1 week
2004-09-05 02:09:54 +00:00
}
#endif /* KERN_SWITCH_INCLUDE */
Reference in New Issue Copy Permalink