freebsd-dev/sys/kern/kern_clock.c

832 lines
21 KiB
C
Raw Normal View History

1994-05-24 10:09:53 +00:00
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
1994-05-24 10:09:53 +00:00
* Copyright (c) 1982, 1986, 1991, 1993
* The Regents of the University of California. All rights reserved.
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
1994-05-24 10:09:53 +00:00
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS 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 REGENTS 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.
*
* @(#)kern_clock.c 8.5 (Berkeley) 1/21/94
*/
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_kdb.h"
#include "opt_device_polling.h"
#include "opt_hwpmc_hooks.h"
#include "opt_ntp.h"
#include "opt_watchdog.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/callout.h>
#include <sys/epoch.h>
#include <sys/eventhandler.h>
#include <sys/gtaskqueue.h>
#include <sys/kdb.h>
1994-05-24 10:09:53 +00:00
#include <sys/kernel.h>
#include <sys/kthread.h>
2001-10-11 17:53:43 +00:00
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/mutex.h>
1994-05-24 10:09:53 +00:00
#include <sys/proc.h>
#include <sys/resource.h>
1994-05-24 10:09:53 +00:00
#include <sys/resourcevar.h>
#include <sys/sched.h>
#include <sys/sdt.h>
#include <sys/signalvar.h>
#include <sys/sleepqueue.h>
Overhaul of the SMP code. Several portions of the SMP kernel support have been made machine independent and various other adjustments have been made to support Alpha SMP. - It splits the per-process portions of hardclock() and statclock() off into hardclock_process() and statclock_process() respectively. hardclock() and statclock() call the *_process() functions for the current process so that UP systems will run as before. For SMP systems, it is simply necessary to ensure that all other processors execute the *_process() functions when the main clock functions are triggered on one CPU by an interrupt. For the alpha 4100, clock interrupts are delievered in a staggered broadcast fashion, so we simply call hardclock/statclock on the boot CPU and call the *_process() functions on the secondaries. For x86, we call statclock and hardclock as usual and then call forward_hardclock/statclock in the MD code to send an IPI to cause the AP's to execute forwared_hardclock/statclock which then call the *_process() functions. - forward_signal() and forward_roundrobin() have been reworked to be MI and to involve less hackery. Now the cpu doing the forward sets any flags, etc. and sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically return so that they can execute ast() and don't bother with setting the astpending or needresched flags themselves. This also removes the loop in forward_signal() as sched_lock closes the race condition that the loop worked around. - need_resched(), resched_wanted() and clear_resched() have been changed to take a process to act on rather than assuming curproc so that they can be used to implement forward_roundrobin() as described above. - Various other SMP variables have been moved to a MI subr_smp.c and a new header sys/smp.h declares MI SMP variables and API's. The IPI API's from machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h. - The globaldata_register() and globaldata_find() functions as well as the SLIST of globaldata structures has become MI and moved into subr_smp.c. Also, the globaldata list is only available if SMP support is compiled in. Reviewed by: jake, peter Looked over by: eivind
2001-04-27 19:28:25 +00:00
#include <sys/smp.h>
#include <vm/vm.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <sys/sysctl.h>
#include <sys/bus.h>
#include <sys/interrupt.h>
#include <sys/limits.h>
#include <sys/timetc.h>
1994-05-24 10:09:53 +00:00
#ifdef GPROF
#include <sys/gmon.h>
#endif
#ifdef HWPMC_HOOKS
#include <sys/pmckern.h>
PMC_SOFT_DEFINE( , , clock, hard);
PMC_SOFT_DEFINE( , , clock, stat);
PMC_SOFT_DEFINE_EX( , , clock, prof, \
cpu_startprofclock, cpu_stopprofclock);
#endif
Device Polling code for -current. Non-SMP, i386-only, no polling in the idle loop at the moment. To use this code you must compile a kernel with options DEVICE_POLLING and at runtime enable polling with sysctl kern.polling.enable=1 The percentage of CPU reserved to userland can be set with sysctl kern.polling.user_frac=NN (default is 50) while the remainder is used by polling device drivers and netisr's. These are the only two variables that you should need to touch. There are a few more parameters in kern.polling but the default values are adequate for all purposes. See the code in kern_poll.c for more details on them. Polling in the idle loop will be implemented shortly by introducing a kernel thread which does the job. Until then, the amount of CPU dedicated to polling will never exceed (100-user_frac). The equivalent (actually, better) code for -stable is at http://info.iet.unipi.it/~luigi/polling/ and also supports polling in the idle loop. NOTE to Alpha developers: There is really nothing in this code that is i386-specific. If you move the 2 lines supporting the new option from sys/conf/{files,options}.i386 to sys/conf/{files,options} I am pretty sure that this should work on the Alpha as well, just that I do not have a suitable test box to try it. If someone feels like trying it, I would appreciate it. NOTE to other developers: sure some things could be done better, and as always I am open to constructive criticism, which a few of you have already given and I greatly appreciated. However, before proposing radical architectural changes, please take some time to possibly try out this code, or at the very least read the comments in kern_poll.c, especially re. the reason why I am using a soft netisr and cannot (I believe) replace it with a simple timeout. Quick description of files touched by this commit: sys/conf/files.i386 new file kern/kern_poll.c sys/conf/options.i386 new option sys/i386/i386/trap.c poll in trap (disabled by default) sys/kern/kern_clock.c initialization and hardclock hooks. sys/kern/kern_intr.c minor swi_net changes sys/kern/kern_poll.c the bulk of the code. sys/net/if.h new flag sys/net/if_var.h declaration for functions used in device drivers. sys/net/netisr.h NETISR_POLL sys/dev/fxp/if_fxp.c sys/dev/fxp/if_fxpvar.h sys/pci/if_dc.c sys/pci/if_dcreg.h sys/pci/if_sis.c sys/pci/if_sisreg.h device driver modifications
2001-12-14 17:56:12 +00:00
#ifdef DEVICE_POLLING
extern void hardclock_device_poll(void);
#endif /* DEVICE_POLLING */
2002-03-19 21:25:46 +00:00
static void initclocks(void *dummy);
SYSINIT(clocks, SI_SUB_CLOCKS, SI_ORDER_FIRST, initclocks, NULL);
/* Spin-lock protecting profiling statistics. */
static struct mtx time_lock;
SDT_PROVIDER_DECLARE(sched);
SDT_PROBE_DEFINE2(sched, , , tick, "struct thread *", "struct proc *");
static int
sysctl_kern_cp_time(SYSCTL_HANDLER_ARGS)
{
int error;
long cp_time[CPUSTATES];
#ifdef SCTL_MASK32
int i;
unsigned int cp_time32[CPUSTATES];
#endif
2006-04-17 20:14:51 +00:00
read_cpu_time(cp_time);
#ifdef SCTL_MASK32
if (req->flags & SCTL_MASK32) {
if (!req->oldptr)
return SYSCTL_OUT(req, 0, sizeof(cp_time32));
for (i = 0; i < CPUSTATES; i++)
cp_time32[i] = (unsigned int)cp_time[i];
error = SYSCTL_OUT(req, cp_time32, sizeof(cp_time32));
} else
#endif
{
if (!req->oldptr)
return SYSCTL_OUT(req, 0, sizeof(cp_time));
error = SYSCTL_OUT(req, cp_time, sizeof(cp_time));
}
return error;
}
SYSCTL_PROC(_kern, OID_AUTO, cp_time, CTLTYPE_LONG|CTLFLAG_RD|CTLFLAG_MPSAFE,
0,0, sysctl_kern_cp_time, "LU", "CPU time statistics");
static long empty[CPUSTATES];
static int
sysctl_kern_cp_times(SYSCTL_HANDLER_ARGS)
{
struct pcpu *pcpu;
int error;
int c;
long *cp_time;
#ifdef SCTL_MASK32
unsigned int cp_time32[CPUSTATES];
int i;
#endif
if (!req->oldptr) {
#ifdef SCTL_MASK32
if (req->flags & SCTL_MASK32)
return SYSCTL_OUT(req, 0, sizeof(cp_time32) * (mp_maxid + 1));
else
#endif
return SYSCTL_OUT(req, 0, sizeof(long) * CPUSTATES * (mp_maxid + 1));
}
for (error = 0, c = 0; error == 0 && c <= mp_maxid; c++) {
if (!CPU_ABSENT(c)) {
pcpu = pcpu_find(c);
cp_time = pcpu->pc_cp_time;
} else {
cp_time = empty;
}
#ifdef SCTL_MASK32
if (req->flags & SCTL_MASK32) {
for (i = 0; i < CPUSTATES; i++)
cp_time32[i] = (unsigned int)cp_time[i];
error = SYSCTL_OUT(req, cp_time32, sizeof(cp_time32));
} else
#endif
error = SYSCTL_OUT(req, cp_time, sizeof(long) * CPUSTATES);
}
return error;
}
SYSCTL_PROC(_kern, OID_AUTO, cp_times, CTLTYPE_LONG|CTLFLAG_RD|CTLFLAG_MPSAFE,
0,0, sysctl_kern_cp_times, "LU", "per-CPU time statistics");
#ifdef DEADLKRES
static const char *blessed[] = {
"getblk",
"so_snd_sx",
"so_rcv_sx",
NULL
};
static int slptime_threshold = 1800;
static int blktime_threshold = 900;
static int sleepfreq = 3;
static void
deadlres_td_on_lock(struct proc *p, struct thread *td, int blkticks)
{
int tticks;
sx_assert(&allproc_lock, SX_LOCKED);
PROC_LOCK_ASSERT(p, MA_OWNED);
THREAD_LOCK_ASSERT(td, MA_OWNED);
/*
* The thread should be blocked on a turnstile, simply check
* if the turnstile channel is in good state.
*/
MPASS(td->td_blocked != NULL);
tticks = ticks - td->td_blktick;
if (tticks > blkticks)
/*
* Accordingly with provided thresholds, this thread is stuck
* for too long on a turnstile.
*/
panic("%s: possible deadlock detected for %p, "
"blocked for %d ticks\n", __func__, td, tticks);
}
static void
deadlres_td_sleep_q(struct proc *p, struct thread *td, int slpticks)
{
const void *wchan;
int i, slptype, tticks;
sx_assert(&allproc_lock, SX_LOCKED);
PROC_LOCK_ASSERT(p, MA_OWNED);
THREAD_LOCK_ASSERT(td, MA_OWNED);
/*
* Check if the thread is sleeping on a lock, otherwise skip the check.
* Drop the thread lock in order to avoid a LOR with the sleepqueue
* spinlock.
*/
wchan = td->td_wchan;
tticks = ticks - td->td_slptick;
slptype = sleepq_type(wchan);
if ((slptype == SLEEPQ_SX || slptype == SLEEPQ_LK) &&
tticks > slpticks) {
/*
* Accordingly with provided thresholds, this thread is stuck
* for too long on a sleepqueue.
* However, being on a sleepqueue, we might still check for the
* blessed list.
*/
for (i = 0; blessed[i] != NULL; i++)
if (!strcmp(blessed[i], td->td_wmesg))
return;
panic("%s: possible deadlock detected for %p, "
"blocked for %d ticks\n", __func__, td, tticks);
}
}
static void
deadlkres(void)
{
struct proc *p;
struct thread *td;
int blkticks, slpticks, tryl;
tryl = 0;
for (;;) {
blkticks = blktime_threshold * hz;
slpticks = slptime_threshold * hz;
/*
* Avoid to sleep on the sx_lock in order to avoid a
* possible priority inversion problem leading to
* starvation.
* If the lock can't be held after 100 tries, panic.
*/
if (!sx_try_slock(&allproc_lock)) {
if (tryl > 100)
panic("%s: possible deadlock detected "
"on allproc_lock\n", __func__);
tryl++;
pause("allproc", sleepfreq * hz);
continue;
}
tryl = 0;
FOREACH_PROC_IN_SYSTEM(p) {
PROC_LOCK(p);
if (p->p_state == PRS_NEW) {
PROC_UNLOCK(p);
continue;
}
FOREACH_THREAD_IN_PROC(p, td) {
thread_lock(td);
if (TD_ON_LOCK(td))
deadlres_td_on_lock(p, td,
blkticks);
else if (TD_IS_SLEEPING(td))
deadlres_td_sleep_q(p, td,
slpticks);
thread_unlock(td);
}
PROC_UNLOCK(p);
}
sx_sunlock(&allproc_lock);
/* Sleep for sleepfreq seconds. */
pause("-", sleepfreq * hz);
}
}
static struct kthread_desc deadlkres_kd = {
"deadlkres",
deadlkres,
(struct thread **)NULL
};
SYSINIT(deadlkres, SI_SUB_CLOCKS, SI_ORDER_ANY, kthread_start, &deadlkres_kd);
static SYSCTL_NODE(_debug, OID_AUTO, deadlkres, CTLFLAG_RW, 0,
"Deadlock resolver");
SYSCTL_INT(_debug_deadlkres, OID_AUTO, slptime_threshold, CTLFLAG_RW,
&slptime_threshold, 0,
"Number of seconds within is valid to sleep on a sleepqueue");
SYSCTL_INT(_debug_deadlkres, OID_AUTO, blktime_threshold, CTLFLAG_RW,
&blktime_threshold, 0,
"Number of seconds within is valid to block on a turnstile");
SYSCTL_INT(_debug_deadlkres, OID_AUTO, sleepfreq, CTLFLAG_RW, &sleepfreq, 0,
"Number of seconds between any deadlock resolver thread run");
#endif /* DEADLKRES */
void
read_cpu_time(long *cp_time)
{
struct pcpu *pc;
int i, j;
/* Sum up global cp_time[]. */
bzero(cp_time, sizeof(long) * CPUSTATES);
CPU_FOREACH(i) {
pc = pcpu_find(i);
for (j = 0; j < CPUSTATES; j++)
cp_time[j] += pc->pc_cp_time[j];
}
}
#include <sys/watchdog.h>
static int watchdog_ticks;
static int watchdog_enabled;
static void watchdog_fire(void);
static void watchdog_config(void *, u_int, int *);
static void
watchdog_attach(void)
{
EVENTHANDLER_REGISTER(watchdog_list, watchdog_config, NULL, 0);
}
1994-05-24 10:09:53 +00:00
/*
* Clock handling routines.
*
* This code is written to operate with two timers that run independently of
* each other.
1994-05-24 10:09:53 +00:00
*
* The main timer, running hz times per second, is used to trigger interval
* timers, timeouts and rescheduling as needed.
1994-05-24 10:09:53 +00:00
*
* The second timer handles kernel and user profiling,
* and does resource use estimation. If the second timer is programmable,
* it is randomized to avoid aliasing between the two clocks. For example,
* the randomization prevents an adversary from always giving up the cpu
* just before its quantum expires. Otherwise, it would never accumulate
* cpu ticks. The mean frequency of the second timer is stathz.
*
* If no second timer exists, stathz will be zero; in this case we drive
* profiling and statistics off the main clock. This WILL NOT be accurate;
* do not do it unless absolutely necessary.
*
1994-05-24 10:09:53 +00:00
* The statistics clock may (or may not) be run at a higher rate while
* profiling. This profile clock runs at profhz. We require that profhz
* be an integral multiple of stathz.
*
* If the statistics clock is running fast, it must be divided by the ratio
* profhz/stathz for statistics. (For profiling, every tick counts.)
1994-05-24 10:09:53 +00:00
*
* Time-of-day is maintained using a "timecounter", which may or may
* not be related to the hardware generating the above mentioned
* interrupts.
1994-05-24 10:09:53 +00:00
*/
int stathz;
int profhz;
int profprocs;
volatile int ticks;
int psratio;
1994-05-24 10:09:53 +00:00
DPCPU_DEFINE_STATIC(int, pcputicks); /* Per-CPU version of ticks. */
#ifdef DEVICE_POLLING
static int devpoll_run = 0;
#endif
1994-05-24 10:09:53 +00:00
/*
* Initialize clock frequencies and start both clocks running.
*/
/* ARGSUSED*/
static void
initclocks(void *dummy)
1994-05-24 10:09:53 +00:00
{
int i;
1994-05-24 10:09:53 +00:00
/*
* Set divisors to 1 (normal case) and let the machine-specific
* code do its bit.
*/
Implement new event timers infrastructure. It provides unified APIs for writing event timer drivers, for choosing best possible drivers by machine independent code and for operating them to supply kernel with hardclock(), statclock() and profclock() events in unified fashion on various hardware. Infrastructure provides support for both per-CPU (independent for every CPU core) and global timers in periodic and one-shot modes. MI management code at this moment uses only periodic mode, but one-shot mode use planned for later, as part of tickless kernel project. For this moment infrastructure used on i386 and amd64 architectures. Other archs are welcome to follow, while their current operation should not be affected. This patch updates existing drivers (i8254, RTC and LAPIC) for the new order, and adds event timers support into the HPET driver. These drivers have different capabilities: LAPIC - per-CPU timer, supports periodic and one-shot operation, may freeze in C3 state, calibrated on first use, so may be not exactly precise. HPET - depending on hardware can work as per-CPU or global, supports periodic and one-shot operation, usually provides several event timers. i8254 - global, limited to periodic mode, because same hardware used also as time counter. RTC - global, supports only periodic mode, set of frequencies in Hz limited by powers of 2. Depending on hardware capabilities, drivers preferred in following orders, either LAPIC, HPETs, i8254, RTC or HPETs, LAPIC, i8254, RTC. User may explicitly specify wanted timers via loader tunables or sysctls: kern.eventtimer.timer1 and kern.eventtimer.timer2. If requested driver is unavailable or unoperational, system will try to replace it. If no more timers available or "NONE" specified for second, system will operate using only one timer, multiplying it's frequency by few times and uing respective dividers to honor hz, stathz and profhz values, set during initial setup.
2010-06-20 21:33:29 +00:00
mtx_init(&time_lock, "time lock", NULL, MTX_DEF);
cpu_initclocks();
1994-05-24 10:09:53 +00:00
/*
* Compute profhz/stathz, and fix profhz if needed.
*/
i = stathz ? stathz : hz;
if (profhz == 0)
profhz = i;
psratio = profhz / i;
#ifdef SW_WATCHDOG
/* Enable hardclock watchdog now, even if a hardware watchdog exists. */
watchdog_attach();
#else
/* Volunteer to run a software watchdog. */
if (wdog_software_attach == NULL)
wdog_software_attach = watchdog_attach;
Add an EARLY_AP_STARTUP option to start APs earlier during boot. Currently, Application Processors (non-boot CPUs) are started by MD code at SI_SUB_CPU, but they are kept waiting in a "pen" until SI_SUB_SMP at which point they are released to run kernel threads. SI_SUB_SMP is one of the last SYSINIT levels, so APs don't enter the scheduler and start running threads until fairly late in the boot. This change moves SI_SUB_SMP up to just before software interrupt threads are created allowing the APs to start executing kernel threads much sooner (before any devices are probed). This allows several initialization routines that need to perform initialization on all CPUs to now perform that initialization in one step rather than having to defer the AP initialization to a second SYSINIT run at SI_SUB_SMP. It also permits all CPUs to be available for handling interrupts before any devices are probed. This last feature fixes a problem on with interrupt vector exhaustion. Specifically, in the old model all device interrupts were routed onto the boot CPU during boot. Later after the APs were released at SI_SUB_SMP, interrupts were redistributed across all CPUs. However, several drivers for multiqueue hardware allocate N interrupts per CPU in the system. In a system with many CPUs, just a few drivers doing this could exhaust the available pool of interrupt vectors on the boot CPU as each driver was allocating N * mp_ncpu vectors on the boot CPU. Now, drivers will allocate interrupts on their desired CPUs during boot meaning that only N interrupts are allocated from the boot CPU instead of N * mp_ncpu. Some other bits of code can also be simplified as smp_started is now true much earlier and will now always be true for these bits of code. This removes the need to treat the single-CPU boot environment as a special case. As a transition aid, the new behavior is available under a new kernel option (EARLY_AP_STARTUP). This will allow the option to be turned off if need be during initial testing. I plan to enable this on x86 by default in a followup commit in the next few days and to have all platforms moved over before 11.0. Once the transition is complete, the option will be removed along with the !EARLY_AP_STARTUP code. These changes have only been tested on x86. Other platform maintainers are encouraged to port their architectures over as well. The main things to check for are any uses of smp_started in MD code that can be simplified and SI_SUB_SMP SYSINITs in MD code that can be removed in the EARLY_AP_STARTUP case (e.g. the interrupt shuffling). PR: kern/199321 Reviewed by: markj, gnn, kib Sponsored by: Netflix
2016-05-14 18:22:52 +00:00
#endif
1994-05-24 10:09:53 +00:00
}
static __noinline void
hardclock_itimer(struct thread *td, struct pstats *pstats, int cnt, int usermode)
{
struct proc *p;
int flags;
flags = 0;
p = td->td_proc;
if (usermode &&
timevalisset(&pstats->p_timer[ITIMER_VIRTUAL].it_value)) {
PROC_ITIMLOCK(p);
if (itimerdecr(&pstats->p_timer[ITIMER_VIRTUAL],
tick * cnt) == 0)
flags |= TDF_ALRMPEND | TDF_ASTPENDING;
PROC_ITIMUNLOCK(p);
}
if (timevalisset(&pstats->p_timer[ITIMER_PROF].it_value)) {
PROC_ITIMLOCK(p);
if (itimerdecr(&pstats->p_timer[ITIMER_PROF],
tick * cnt) == 0)
flags |= TDF_PROFPEND | TDF_ASTPENDING;
PROC_ITIMUNLOCK(p);
}
if (flags != 0) {
thread_lock(td);
td->td_flags |= flags;
thread_unlock(td);
}
}
Overhaul of the SMP code. Several portions of the SMP kernel support have been made machine independent and various other adjustments have been made to support Alpha SMP. - It splits the per-process portions of hardclock() and statclock() off into hardclock_process() and statclock_process() respectively. hardclock() and statclock() call the *_process() functions for the current process so that UP systems will run as before. For SMP systems, it is simply necessary to ensure that all other processors execute the *_process() functions when the main clock functions are triggered on one CPU by an interrupt. For the alpha 4100, clock interrupts are delievered in a staggered broadcast fashion, so we simply call hardclock/statclock on the boot CPU and call the *_process() functions on the secondaries. For x86, we call statclock and hardclock as usual and then call forward_hardclock/statclock in the MD code to send an IPI to cause the AP's to execute forwared_hardclock/statclock which then call the *_process() functions. - forward_signal() and forward_roundrobin() have been reworked to be MI and to involve less hackery. Now the cpu doing the forward sets any flags, etc. and sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically return so that they can execute ast() and don't bother with setting the astpending or needresched flags themselves. This also removes the loop in forward_signal() as sched_lock closes the race condition that the loop worked around. - need_resched(), resched_wanted() and clear_resched() have been changed to take a process to act on rather than assuming curproc so that they can be used to implement forward_roundrobin() as described above. - Various other SMP variables have been moved to a MI subr_smp.c and a new header sys/smp.h declares MI SMP variables and API's. The IPI API's from machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h. - The globaldata_register() and globaldata_find() functions as well as the SLIST of globaldata structures has become MI and moved into subr_smp.c. Also, the globaldata list is only available if SMP support is compiled in. Reviewed by: jake, peter Looked over by: eivind
2001-04-27 19:28:25 +00:00
void
hardclock(int cnt, int usermode)
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
{
struct pstats *pstats;
struct thread *td = curthread;
struct proc *p = td->td_proc;
int *t = DPCPU_PTR(pcputicks);
int global, i, newticks;
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
/*
* Update per-CPU and possibly global ticks values.
*/
*t += cnt;
global = ticks;
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
do {
newticks = *t - global;
if (newticks <= 0) {
if (newticks < -1)
*t = global - 1;
newticks = 0;
break;
}
} while (!atomic_fcmpset_int(&ticks, &global, *t));
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
/*
* Run current process's virtual and profile time, as needed.
*/
pstats = p->p_stats;
if (__predict_false(
timevalisset(&pstats->p_timer[ITIMER_VIRTUAL].it_value) ||
timevalisset(&pstats->p_timer[ITIMER_PROF].it_value)))
hardclock_itimer(td, pstats, cnt, usermode);
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
#ifdef HWPMC_HOOKS
if (PMC_CPU_HAS_SAMPLES(PCPU_GET(cpuid)))
PMC_CALL_HOOK_UNLOCKED(curthread, PMC_FN_DO_SAMPLES, NULL);
if (td->td_intr_frame != NULL)
PMC_SOFT_CALL_TF( , , clock, hard, td->td_intr_frame);
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
#endif
/* We are in charge to handle this tick duty. */
if (newticks > 0) {
tc_ticktock(newticks);
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
#ifdef DEVICE_POLLING
/* Dangerous and no need to call these things concurrently. */
if (atomic_cmpset_acq_int(&devpoll_run, 0, 1)) {
/* This is very short and quick. */
hardclock_device_poll();
atomic_store_rel_int(&devpoll_run, 0);
}
#endif /* DEVICE_POLLING */
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
if (watchdog_enabled > 0) {
i = atomic_fetchadd_int(&watchdog_ticks, -newticks);
if (i > 0 && i <= newticks)
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
watchdog_fire();
}
}
if (curcpu == CPU_FIRST())
cpu_tick_calibration();
if (__predict_false(DPCPU_GET(epoch_cb_count)))
GROUPTASK_ENQUEUE(DPCPU_PTR(epoch_cb_task));
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
}
void
hardclock_sync(int cpu)
{
int *t;
KASSERT(!CPU_ABSENT(cpu), ("Absent CPU %d", cpu));
t = DPCPU_ID_PTR(cpu, pcputicks);
Refactor timer management code with priority to one-shot operation mode. The main goal of this is to generate timer interrupts only when there is some work to do. When CPU is busy interrupts are generating at full rate of hz + stathz to fullfill scheduler and timekeeping requirements. But when CPU is idle, only minimum set of interrupts (down to 8 interrupts per second per CPU now), needed to handle scheduled callouts is executed. This allows significantly increase idle CPU sleep time, increasing effect of static power-saving technologies. Also it should reduce host CPU load on virtualized systems, when guest system is idle. There is set of tunables, also available as writable sysctls, allowing to control wanted event timer subsystem behavior: kern.eventtimer.timer - allows to choose event timer hardware to use. On x86 there is up to 4 different kinds of timers. Depending on whether chosen timer is per-CPU, behavior of other options slightly differs. kern.eventtimer.periodic - allows to choose periodic and one-shot operation mode. In periodic mode, current timer hardware taken as the only source of time for time events. This mode is quite alike to previous kernel behavior. One-shot mode instead uses currently selected time counter hardware to schedule all needed events one by one and program timer to generate interrupt exactly in specified time. Default value depends of chosen timer capabilities, but one-shot mode is preferred, until other is forced by user or hardware. kern.eventtimer.singlemul - in periodic mode specifies how much times higher timer frequency should be, to not strictly alias hardclock() and statclock() events. Default values are 2 and 4, but could be reduced to 1 if extra interrupts are unwanted. kern.eventtimer.idletick - makes each CPU to receive every timer interrupt independently of whether they busy or not. By default this options is disabled. If chosen timer is per-CPU and runs in periodic mode, this option has no effect - all interrupts are generating. As soon as this patch modifies cpu_idle() on some platforms, I have also refactored one on x86. Now it makes use of MONITOR/MWAIT instrunctions (if supported) under high sleep/wakeup rate, as fast alternative to other methods. It allows SMP scheduler to wake up sleeping CPUs much faster without using IPI, significantly increasing performance on some highly task-switching loads. Tested by: many (on i386, amd64, sparc64 and powerc) H/W donated by: Gheorghe Ardelean Sponsored by: iXsystems, Inc.
2010-09-13 07:25:35 +00:00
*t = ticks;
}
1994-05-24 10:09:53 +00:00
/*
* Compute number of ticks in the specified amount of time.
1994-05-24 10:09:53 +00:00
*/
int
tvtohz(struct timeval *tv)
1994-05-24 10:09:53 +00:00
{
unsigned long ticks;
long sec, usec;
1994-05-24 10:09:53 +00:00
/*
* If the number of usecs in the whole seconds part of the time
* difference fits in a long, then the total number of usecs will
* fit in an unsigned long. Compute the total and convert it to
* ticks, rounding up and adding 1 to allow for the current tick
* to expire. Rounding also depends on unsigned long arithmetic
* to avoid overflow.
1994-05-24 10:09:53 +00:00
*
* Otherwise, if the number of ticks in the whole seconds part of
* the time difference fits in a long, then convert the parts to
* ticks separately and add, using similar rounding methods and
* overflow avoidance. This method would work in the previous
* case but it is slightly slower and assumes that hz is integral.
*
* Otherwise, round the time difference down to the maximum
* representable value.
*
* If ints have 32 bits, then the maximum value for any timeout in
* 10ms ticks is 248 days.
1994-05-24 10:09:53 +00:00
*/
sec = tv->tv_sec;
usec = tv->tv_usec;
if (usec < 0) {
sec--;
usec += 1000000;
}
if (sec < 0) {
#ifdef DIAGNOSTIC
if (usec > 0) {
sec++;
usec -= 1000000;
}
printf("tvotohz: negative time difference %ld sec %ld usec\n",
sec, usec);
#endif
ticks = 1;
} else if (sec <= LONG_MAX / 1000000)
ticks = howmany(sec * 1000000 + (unsigned long)usec, tick) + 1;
else if (sec <= LONG_MAX / hz)
ticks = sec * hz
+ howmany((unsigned long)usec, tick) + 1;
else
ticks = LONG_MAX;
if (ticks > INT_MAX)
ticks = INT_MAX;
return ((int)ticks);
1994-05-24 10:09:53 +00:00
}
/*
* Start profiling on a process.
*
* Kernel profiling passes proc0 which never exits and hence
* keeps the profile clock running constantly.
*/
void
startprofclock(struct proc *p)
1994-05-24 10:09:53 +00:00
{
PROC_LOCK_ASSERT(p, MA_OWNED);
if (p->p_flag & P_STOPPROF)
return;
if ((p->p_flag & P_PROFIL) == 0) {
p->p_flag |= P_PROFIL;
Implement new event timers infrastructure. It provides unified APIs for writing event timer drivers, for choosing best possible drivers by machine independent code and for operating them to supply kernel with hardclock(), statclock() and profclock() events in unified fashion on various hardware. Infrastructure provides support for both per-CPU (independent for every CPU core) and global timers in periodic and one-shot modes. MI management code at this moment uses only periodic mode, but one-shot mode use planned for later, as part of tickless kernel project. For this moment infrastructure used on i386 and amd64 architectures. Other archs are welcome to follow, while their current operation should not be affected. This patch updates existing drivers (i8254, RTC and LAPIC) for the new order, and adds event timers support into the HPET driver. These drivers have different capabilities: LAPIC - per-CPU timer, supports periodic and one-shot operation, may freeze in C3 state, calibrated on first use, so may be not exactly precise. HPET - depending on hardware can work as per-CPU or global, supports periodic and one-shot operation, usually provides several event timers. i8254 - global, limited to periodic mode, because same hardware used also as time counter. RTC - global, supports only periodic mode, set of frequencies in Hz limited by powers of 2. Depending on hardware capabilities, drivers preferred in following orders, either LAPIC, HPETs, i8254, RTC or HPETs, LAPIC, i8254, RTC. User may explicitly specify wanted timers via loader tunables or sysctls: kern.eventtimer.timer1 and kern.eventtimer.timer2. If requested driver is unavailable or unoperational, system will try to replace it. If no more timers available or "NONE" specified for second, system will operate using only one timer, multiplying it's frequency by few times and uing respective dividers to honor hz, stathz and profhz values, set during initial setup.
2010-06-20 21:33:29 +00:00
mtx_lock(&time_lock);
if (++profprocs == 1)
cpu_startprofclock();
Implement new event timers infrastructure. It provides unified APIs for writing event timer drivers, for choosing best possible drivers by machine independent code and for operating them to supply kernel with hardclock(), statclock() and profclock() events in unified fashion on various hardware. Infrastructure provides support for both per-CPU (independent for every CPU core) and global timers in periodic and one-shot modes. MI management code at this moment uses only periodic mode, but one-shot mode use planned for later, as part of tickless kernel project. For this moment infrastructure used on i386 and amd64 architectures. Other archs are welcome to follow, while their current operation should not be affected. This patch updates existing drivers (i8254, RTC and LAPIC) for the new order, and adds event timers support into the HPET driver. These drivers have different capabilities: LAPIC - per-CPU timer, supports periodic and one-shot operation, may freeze in C3 state, calibrated on first use, so may be not exactly precise. HPET - depending on hardware can work as per-CPU or global, supports periodic and one-shot operation, usually provides several event timers. i8254 - global, limited to periodic mode, because same hardware used also as time counter. RTC - global, supports only periodic mode, set of frequencies in Hz limited by powers of 2. Depending on hardware capabilities, drivers preferred in following orders, either LAPIC, HPETs, i8254, RTC or HPETs, LAPIC, i8254, RTC. User may explicitly specify wanted timers via loader tunables or sysctls: kern.eventtimer.timer1 and kern.eventtimer.timer2. If requested driver is unavailable or unoperational, system will try to replace it. If no more timers available or "NONE" specified for second, system will operate using only one timer, multiplying it's frequency by few times and uing respective dividers to honor hz, stathz and profhz values, set during initial setup.
2010-06-20 21:33:29 +00:00
mtx_unlock(&time_lock);
1994-05-24 10:09:53 +00:00
}
}
/*
* Stop profiling on a process.
*/
void
stopprofclock(struct proc *p)
1994-05-24 10:09:53 +00:00
{
PROC_LOCK_ASSERT(p, MA_OWNED);
if (p->p_flag & P_PROFIL) {
if (p->p_profthreads != 0) {
while (p->p_profthreads != 0) {
p->p_flag |= P_STOPPROF;
msleep(&p->p_profthreads, &p->p_mtx, PPAUSE,
"stopprof", 0);
}
}
if ((p->p_flag & P_PROFIL) == 0)
return;
p->p_flag &= ~P_PROFIL;
Implement new event timers infrastructure. It provides unified APIs for writing event timer drivers, for choosing best possible drivers by machine independent code and for operating them to supply kernel with hardclock(), statclock() and profclock() events in unified fashion on various hardware. Infrastructure provides support for both per-CPU (independent for every CPU core) and global timers in periodic and one-shot modes. MI management code at this moment uses only periodic mode, but one-shot mode use planned for later, as part of tickless kernel project. For this moment infrastructure used on i386 and amd64 architectures. Other archs are welcome to follow, while their current operation should not be affected. This patch updates existing drivers (i8254, RTC and LAPIC) for the new order, and adds event timers support into the HPET driver. These drivers have different capabilities: LAPIC - per-CPU timer, supports periodic and one-shot operation, may freeze in C3 state, calibrated on first use, so may be not exactly precise. HPET - depending on hardware can work as per-CPU or global, supports periodic and one-shot operation, usually provides several event timers. i8254 - global, limited to periodic mode, because same hardware used also as time counter. RTC - global, supports only periodic mode, set of frequencies in Hz limited by powers of 2. Depending on hardware capabilities, drivers preferred in following orders, either LAPIC, HPETs, i8254, RTC or HPETs, LAPIC, i8254, RTC. User may explicitly specify wanted timers via loader tunables or sysctls: kern.eventtimer.timer1 and kern.eventtimer.timer2. If requested driver is unavailable or unoperational, system will try to replace it. If no more timers available or "NONE" specified for second, system will operate using only one timer, multiplying it's frequency by few times and uing respective dividers to honor hz, stathz and profhz values, set during initial setup.
2010-06-20 21:33:29 +00:00
mtx_lock(&time_lock);
if (--profprocs == 0)
cpu_stopprofclock();
Implement new event timers infrastructure. It provides unified APIs for writing event timer drivers, for choosing best possible drivers by machine independent code and for operating them to supply kernel with hardclock(), statclock() and profclock() events in unified fashion on various hardware. Infrastructure provides support for both per-CPU (independent for every CPU core) and global timers in periodic and one-shot modes. MI management code at this moment uses only periodic mode, but one-shot mode use planned for later, as part of tickless kernel project. For this moment infrastructure used on i386 and amd64 architectures. Other archs are welcome to follow, while their current operation should not be affected. This patch updates existing drivers (i8254, RTC and LAPIC) for the new order, and adds event timers support into the HPET driver. These drivers have different capabilities: LAPIC - per-CPU timer, supports periodic and one-shot operation, may freeze in C3 state, calibrated on first use, so may be not exactly precise. HPET - depending on hardware can work as per-CPU or global, supports periodic and one-shot operation, usually provides several event timers. i8254 - global, limited to periodic mode, because same hardware used also as time counter. RTC - global, supports only periodic mode, set of frequencies in Hz limited by powers of 2. Depending on hardware capabilities, drivers preferred in following orders, either LAPIC, HPETs, i8254, RTC or HPETs, LAPIC, i8254, RTC. User may explicitly specify wanted timers via loader tunables or sysctls: kern.eventtimer.timer1 and kern.eventtimer.timer2. If requested driver is unavailable or unoperational, system will try to replace it. If no more timers available or "NONE" specified for second, system will operate using only one timer, multiplying it's frequency by few times and uing respective dividers to honor hz, stathz and profhz values, set during initial setup.
2010-06-20 21:33:29 +00:00
mtx_unlock(&time_lock);
1994-05-24 10:09:53 +00:00
}
}
/*
* Statistics clock. Updates rusage information and calls the scheduler
* to adjust priorities of the active thread.
*
* This should be called by all active processors.
1994-05-24 10:09:53 +00:00
*/
void
statclock(int cnt, int usermode)
1994-05-24 10:09:53 +00:00
{
struct rusage *ru;
struct vmspace *vm;
struct thread *td;
struct proc *p;
long rss;
long *cp_time;
uint64_t runtime, new_switchtime;
td = curthread;
p = td->td_proc;
cp_time = (long *)PCPU_PTR(cp_time);
Tweak how the MD code calls the fooclock() methods some. Instead of passing a pointer to an opaque clockframe structure and requiring the MD code to supply CLKF_FOO() macros to extract needed values out of the opaque structure, just pass the needed values directly. In practice this means passing the pair (usermode, pc) to hardclock() and profclock() and passing the boolean (usermode) to hardclock_cpu() and hardclock_process(). Other details: - Axe clockframe and CLKF_FOO() macros on all architectures. Basically, all the archs were taking a trapframe and converting it into a clockframe one way or another. Now they can just extract the PC and usermode values directly out of the trapframe and pass it to fooclock(). - Renamed hardclock_process() to hardclock_cpu() as the latter is more accurate. - On Alpha, we now run profclock() at hz (profhz == hz) rather than at the slower stathz. - On Alpha, for the TurboLaser machines that don't have an 8254 timecounter, call hardclock() directly. This removes an extra conditional check from every clock interrupt on Alpha on the BSP. There is probably room for even further pruning here by changing Alpha to use the simplified timecounter we use on x86 with the lapic timer since we don't get interrupts from the 8254 on Alpha anyway. - On x86, clkintr() shouldn't ever be called now unless using_lapic_timer is false, so add a KASSERT() to that affect and remove a condition to slightly optimize the non-lapic case. - Change prototypeof arm_handler_execute() so that it's first arg is a trapframe pointer rather than a void pointer for clarity. - Use KCOUNT macro in profclock() to lookup the kernel profiling bucket. Tested on: alpha, amd64, arm, i386, ia64, sparc64 Reviewed by: bde (mostly)
2005-12-22 22:16:09 +00:00
if (usermode) {
1994-05-24 10:09:53 +00:00
/*
* Charge the time as appropriate.
1994-05-24 10:09:53 +00:00
*/
td->td_uticks += cnt;
if (p->p_nice > NZERO)
cp_time[CP_NICE] += cnt;
1994-05-24 10:09:53 +00:00
else
cp_time[CP_USER] += cnt;
1994-05-24 10:09:53 +00:00
} else {
/*
* Came from kernel mode, so we were:
* - handling an interrupt,
* - doing syscall or trap work on behalf of the current
* user process, or
* - spinning in the idle loop.
* Whichever it is, charge the time as appropriate.
* Note that we charge interrupts to the current process,
* regardless of whether they are ``for'' that process,
* so that we know how much of its real time was spent
* in ``non-process'' (i.e., interrupt) work.
*/
Reorganize the interrupt handling code a bit to make a few things cleaner and increase flexibility to allow various different approaches to be tried in the future. - Split struct ithd up into two pieces. struct intr_event holds the list of interrupt handlers associated with interrupt sources. struct intr_thread contains the data relative to an interrupt thread. Currently we still provide a 1:1 relationship of events to threads with the exception that events only have an associated thread if there is at least one threaded interrupt handler attached to the event. This means that on x86 we no longer have 4 bazillion interrupt threads with no handlers. It also means that interrupt events with only INTR_FAST handlers no longer have an associated thread either. - Renamed struct intrhand to struct intr_handler to follow the struct intr_foo naming convention. This did require renaming the powerpc MD struct intr_handler to struct ppc_intr_handler. - INTR_FAST no longer implies INTR_EXCL on all architectures except for powerpc. This means that multiple INTR_FAST handlers can attach to the same interrupt and that INTR_FAST and non-INTR_FAST handlers can attach to the same interrupt. Sharing INTR_FAST handlers may not always be desirable, but having sio(4) and uhci(4) fight over an IRQ isn't fun either. Drivers can always still use INTR_EXCL to ask for an interrupt exclusively. The way this sharing works is that when an interrupt comes in, all the INTR_FAST handlers are executed first, and if any threaded handlers exist, the interrupt thread is scheduled afterwards. This type of layout also makes it possible to investigate using interrupt filters ala OS X where the filter determines whether or not its companion threaded handler should run. - Aside from the INTR_FAST changes above, the impact on MD interrupt code is mostly just 's/ithread/intr_event/'. - A new MI ddb command 'show intrs' walks the list of interrupt events dumping their state. It also has a '/v' verbose switch which dumps info about all of the handlers attached to each event. - We currently don't destroy an interrupt thread when the last threaded handler is removed because it would suck for things like ppbus(8)'s braindead behavior. The code is present, though, it is just under #if 0 for now. - Move the code to actually execute the threaded handlers for an interrrupt event into a separate function so that ithread_loop() becomes more readable. Previously this code was all in the middle of ithread_loop() and indented halfway across the screen. - Made struct intr_thread private to kern_intr.c and replaced td_ithd with a thread private flag TDP_ITHREAD. - In statclock, check curthread against idlethread directly rather than curthread's proc against idlethread's proc. (Not really related to intr changes) Tested on: alpha, amd64, i386, sparc64 Tested on: arm, ia64 (older version of patch by cognet and marcel)
2005-10-25 19:48:48 +00:00
if ((td->td_pflags & TDP_ITHREAD) ||
td->td_intr_nesting_level >= 2) {
td->td_iticks += cnt;
cp_time[CP_INTR] += cnt;
} else {
td->td_pticks += cnt;
td->td_sticks += cnt;
if (!TD_IS_IDLETHREAD(td))
cp_time[CP_SYS] += cnt;
else
cp_time[CP_IDLE] += cnt;
}
1994-05-24 10:09:53 +00:00
}
/* Update resource usage integrals and maximums. */
MPASS(p->p_vmspace != NULL);
vm = p->p_vmspace;
ru = &td->td_ru;
ru->ru_ixrss += pgtok(vm->vm_tsize) * cnt;
ru->ru_idrss += pgtok(vm->vm_dsize) * cnt;
ru->ru_isrss += pgtok(vm->vm_ssize) * cnt;
rss = pgtok(vmspace_resident_count(vm));
if (ru->ru_maxrss < rss)
ru->ru_maxrss = rss;
KTR_POINT2(KTR_SCHED, "thread", sched_tdname(td), "statclock",
"prio:%d", td->td_priority, "stathz:%d", (stathz)?stathz:hz);
SDT_PROBE2(sched, , , tick, td, td->td_proc);
thread_lock_flags(td, MTX_QUIET);
/*
* Compute the amount of time during which the current
* thread was running, and add that to its total so far.
*/
new_switchtime = cpu_ticks();
runtime = new_switchtime - PCPU_GET(switchtime);
td->td_runtime += runtime;
td->td_incruntime += runtime;
PCPU_SET(switchtime, new_switchtime);
sched_clock(td, cnt);
thread_unlock(td);
#ifdef HWPMC_HOOKS
if (td->td_intr_frame != NULL)
PMC_SOFT_CALL_TF( , , clock, stat, td->td_intr_frame);
#endif
Overhaul of the SMP code. Several portions of the SMP kernel support have been made machine independent and various other adjustments have been made to support Alpha SMP. - It splits the per-process portions of hardclock() and statclock() off into hardclock_process() and statclock_process() respectively. hardclock() and statclock() call the *_process() functions for the current process so that UP systems will run as before. For SMP systems, it is simply necessary to ensure that all other processors execute the *_process() functions when the main clock functions are triggered on one CPU by an interrupt. For the alpha 4100, clock interrupts are delievered in a staggered broadcast fashion, so we simply call hardclock/statclock on the boot CPU and call the *_process() functions on the secondaries. For x86, we call statclock and hardclock as usual and then call forward_hardclock/statclock in the MD code to send an IPI to cause the AP's to execute forwared_hardclock/statclock which then call the *_process() functions. - forward_signal() and forward_roundrobin() have been reworked to be MI and to involve less hackery. Now the cpu doing the forward sets any flags, etc. and sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically return so that they can execute ast() and don't bother with setting the astpending or needresched flags themselves. This also removes the loop in forward_signal() as sched_lock closes the race condition that the loop worked around. - need_resched(), resched_wanted() and clear_resched() have been changed to take a process to act on rather than assuming curproc so that they can be used to implement forward_roundrobin() as described above. - Various other SMP variables have been moved to a MI subr_smp.c and a new header sys/smp.h declares MI SMP variables and API's. The IPI API's from machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h. - The globaldata_register() and globaldata_find() functions as well as the SLIST of globaldata structures has become MI and moved into subr_smp.c. Also, the globaldata list is only available if SMP support is compiled in. Reviewed by: jake, peter Looked over by: eivind
2001-04-27 19:28:25 +00:00
}
- Change fast interrupts on x86 to push a full interrupt frame and to return through doreti to handle ast's. This is necessary for the clock interrupts to work properly. - Change the clock interrupts on the x86 to be fast instead of threaded. This is needed because both hardclock() and statclock() need to run in the context of the current process, not in a separate thread context. - Kill the prevproc hack as it is no longer needed. - We really need Giant when we call psignal(), but we don't want to block during the clock interrupt. Instead, use two p_flag's in the proc struct to mark the current process as having a pending SIGVTALRM or a SIGPROF and let them be delivered during ast() when hardclock() has finished running. - Remove CLKF_BASEPRI, which was #ifdef'd out on the x86 anyways. It was broken on the x86 if it was turned on since cpl is gone. It's only use was to bogusly run softclock() directly during hardclock() rather than scheduling an SWI. - Remove the COM_LOCK simplelock and replace it with a clock_lock spin mutex. Since the spin mutex already handles disabling/restoring interrupts appropriately, this also lets us axe all the *_intr() fu. - Back out the hacks in the APIC_IO x86 cpu_initclocks() code to use temporary fast interrupts for the APIC trial. - Add two new process flags P_ALRMPEND and P_PROFPEND to mark the pending signals in hardclock() that are to be delivered in ast(). Submitted by: jakeb (making statclock safe in a fast interrupt) Submitted by: cp (concept of delaying signals until ast())
2000-10-06 02:20:21 +00:00
Overhaul of the SMP code. Several portions of the SMP kernel support have been made machine independent and various other adjustments have been made to support Alpha SMP. - It splits the per-process portions of hardclock() and statclock() off into hardclock_process() and statclock_process() respectively. hardclock() and statclock() call the *_process() functions for the current process so that UP systems will run as before. For SMP systems, it is simply necessary to ensure that all other processors execute the *_process() functions when the main clock functions are triggered on one CPU by an interrupt. For the alpha 4100, clock interrupts are delievered in a staggered broadcast fashion, so we simply call hardclock/statclock on the boot CPU and call the *_process() functions on the secondaries. For x86, we call statclock and hardclock as usual and then call forward_hardclock/statclock in the MD code to send an IPI to cause the AP's to execute forwared_hardclock/statclock which then call the *_process() functions. - forward_signal() and forward_roundrobin() have been reworked to be MI and to involve less hackery. Now the cpu doing the forward sets any flags, etc. and sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically return so that they can execute ast() and don't bother with setting the astpending or needresched flags themselves. This also removes the loop in forward_signal() as sched_lock closes the race condition that the loop worked around. - need_resched(), resched_wanted() and clear_resched() have been changed to take a process to act on rather than assuming curproc so that they can be used to implement forward_roundrobin() as described above. - Various other SMP variables have been moved to a MI subr_smp.c and a new header sys/smp.h declares MI SMP variables and API's. The IPI API's from machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h. - The globaldata_register() and globaldata_find() functions as well as the SLIST of globaldata structures has become MI and moved into subr_smp.c. Also, the globaldata list is only available if SMP support is compiled in. Reviewed by: jake, peter Looked over by: eivind
2001-04-27 19:28:25 +00:00
void
profclock(int cnt, int usermode, uintfptr_t pc)
Overhaul of the SMP code. Several portions of the SMP kernel support have been made machine independent and various other adjustments have been made to support Alpha SMP. - It splits the per-process portions of hardclock() and statclock() off into hardclock_process() and statclock_process() respectively. hardclock() and statclock() call the *_process() functions for the current process so that UP systems will run as before. For SMP systems, it is simply necessary to ensure that all other processors execute the *_process() functions when the main clock functions are triggered on one CPU by an interrupt. For the alpha 4100, clock interrupts are delievered in a staggered broadcast fashion, so we simply call hardclock/statclock on the boot CPU and call the *_process() functions on the secondaries. For x86, we call statclock and hardclock as usual and then call forward_hardclock/statclock in the MD code to send an IPI to cause the AP's to execute forwared_hardclock/statclock which then call the *_process() functions. - forward_signal() and forward_roundrobin() have been reworked to be MI and to involve less hackery. Now the cpu doing the forward sets any flags, etc. and sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically return so that they can execute ast() and don't bother with setting the astpending or needresched flags themselves. This also removes the loop in forward_signal() as sched_lock closes the race condition that the loop worked around. - need_resched(), resched_wanted() and clear_resched() have been changed to take a process to act on rather than assuming curproc so that they can be used to implement forward_roundrobin() as described above. - Various other SMP variables have been moved to a MI subr_smp.c and a new header sys/smp.h declares MI SMP variables and API's. The IPI API's from machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h. - The globaldata_register() and globaldata_find() functions as well as the SLIST of globaldata structures has become MI and moved into subr_smp.c. Also, the globaldata list is only available if SMP support is compiled in. Reviewed by: jake, peter Looked over by: eivind
2001-04-27 19:28:25 +00:00
{
struct thread *td;
#ifdef GPROF
struct gmonparam *g;
uintfptr_t i;
#endif
Overhaul of the SMP code. Several portions of the SMP kernel support have been made machine independent and various other adjustments have been made to support Alpha SMP. - It splits the per-process portions of hardclock() and statclock() off into hardclock_process() and statclock_process() respectively. hardclock() and statclock() call the *_process() functions for the current process so that UP systems will run as before. For SMP systems, it is simply necessary to ensure that all other processors execute the *_process() functions when the main clock functions are triggered on one CPU by an interrupt. For the alpha 4100, clock interrupts are delievered in a staggered broadcast fashion, so we simply call hardclock/statclock on the boot CPU and call the *_process() functions on the secondaries. For x86, we call statclock and hardclock as usual and then call forward_hardclock/statclock in the MD code to send an IPI to cause the AP's to execute forwared_hardclock/statclock which then call the *_process() functions. - forward_signal() and forward_roundrobin() have been reworked to be MI and to involve less hackery. Now the cpu doing the forward sets any flags, etc. and sends a very simple IPI_AST to the other cpu(s). AST IPIs now just basically return so that they can execute ast() and don't bother with setting the astpending or needresched flags themselves. This also removes the loop in forward_signal() as sched_lock closes the race condition that the loop worked around. - need_resched(), resched_wanted() and clear_resched() have been changed to take a process to act on rather than assuming curproc so that they can be used to implement forward_roundrobin() as described above. - Various other SMP variables have been moved to a MI subr_smp.c and a new header sys/smp.h declares MI SMP variables and API's. The IPI API's from machine/ipl.h have moved to machine/smp.h which is included by sys/smp.h. - The globaldata_register() and globaldata_find() functions as well as the SLIST of globaldata structures has become MI and moved into subr_smp.c. Also, the globaldata list is only available if SMP support is compiled in. Reviewed by: jake, peter Looked over by: eivind
2001-04-27 19:28:25 +00:00
td = curthread;
Tweak how the MD code calls the fooclock() methods some. Instead of passing a pointer to an opaque clockframe structure and requiring the MD code to supply CLKF_FOO() macros to extract needed values out of the opaque structure, just pass the needed values directly. In practice this means passing the pair (usermode, pc) to hardclock() and profclock() and passing the boolean (usermode) to hardclock_cpu() and hardclock_process(). Other details: - Axe clockframe and CLKF_FOO() macros on all architectures. Basically, all the archs were taking a trapframe and converting it into a clockframe one way or another. Now they can just extract the PC and usermode values directly out of the trapframe and pass it to fooclock(). - Renamed hardclock_process() to hardclock_cpu() as the latter is more accurate. - On Alpha, we now run profclock() at hz (profhz == hz) rather than at the slower stathz. - On Alpha, for the TurboLaser machines that don't have an 8254 timecounter, call hardclock() directly. This removes an extra conditional check from every clock interrupt on Alpha on the BSP. There is probably room for even further pruning here by changing Alpha to use the simplified timecounter we use on x86 with the lapic timer since we don't get interrupts from the 8254 on Alpha anyway. - On x86, clkintr() shouldn't ever be called now unless using_lapic_timer is false, so add a KASSERT() to that affect and remove a condition to slightly optimize the non-lapic case. - Change prototypeof arm_handler_execute() so that it's first arg is a trapframe pointer rather than a void pointer for clarity. - Use KCOUNT macro in profclock() to lookup the kernel profiling bucket. Tested on: alpha, amd64, arm, i386, ia64, sparc64 Reviewed by: bde (mostly)
2005-12-22 22:16:09 +00:00
if (usermode) {
/*
* Came from user mode; CPU was in user state.
* If this process is being profiled, record the tick.
* if there is no related user location yet, don't
* bother trying to count it.
*/
if (td->td_proc->p_flag & P_PROFIL)
addupc_intr(td, pc, cnt);
}
#ifdef GPROF
else {
/*
* Kernel statistics are just like addupc_intr, only easier.
*/
g = &_gmonparam;
Tweak how the MD code calls the fooclock() methods some. Instead of passing a pointer to an opaque clockframe structure and requiring the MD code to supply CLKF_FOO() macros to extract needed values out of the opaque structure, just pass the needed values directly. In practice this means passing the pair (usermode, pc) to hardclock() and profclock() and passing the boolean (usermode) to hardclock_cpu() and hardclock_process(). Other details: - Axe clockframe and CLKF_FOO() macros on all architectures. Basically, all the archs were taking a trapframe and converting it into a clockframe one way or another. Now they can just extract the PC and usermode values directly out of the trapframe and pass it to fooclock(). - Renamed hardclock_process() to hardclock_cpu() as the latter is more accurate. - On Alpha, we now run profclock() at hz (profhz == hz) rather than at the slower stathz. - On Alpha, for the TurboLaser machines that don't have an 8254 timecounter, call hardclock() directly. This removes an extra conditional check from every clock interrupt on Alpha on the BSP. There is probably room for even further pruning here by changing Alpha to use the simplified timecounter we use on x86 with the lapic timer since we don't get interrupts from the 8254 on Alpha anyway. - On x86, clkintr() shouldn't ever be called now unless using_lapic_timer is false, so add a KASSERT() to that affect and remove a condition to slightly optimize the non-lapic case. - Change prototypeof arm_handler_execute() so that it's first arg is a trapframe pointer rather than a void pointer for clarity. - Use KCOUNT macro in profclock() to lookup the kernel profiling bucket. Tested on: alpha, amd64, arm, i386, ia64, sparc64 Reviewed by: bde (mostly)
2005-12-22 22:16:09 +00:00
if (g->state == GMON_PROF_ON && pc >= g->lowpc) {
i = PC_TO_I(g, pc);
if (i < g->textsize) {
KCOUNT(g, i) += cnt;
}
}
}
#endif
#ifdef HWPMC_HOOKS
if (td->td_intr_frame != NULL)
PMC_SOFT_CALL_TF( , , clock, prof, td->td_intr_frame);
#endif
1994-05-24 10:09:53 +00:00
}
/*
* Return information about system clocks.
*/
static int
sysctl_kern_clockrate(SYSCTL_HANDLER_ARGS)
1994-05-24 10:09:53 +00:00
{
struct clockinfo clkinfo;
/*
* Construct clockinfo structure.
*/
bzero(&clkinfo, sizeof(clkinfo));
1994-05-24 10:09:53 +00:00
clkinfo.hz = hz;
clkinfo.tick = tick;
clkinfo.profhz = profhz;
clkinfo.stathz = stathz ? stathz : hz;
return (sysctl_handle_opaque(oidp, &clkinfo, sizeof clkinfo, req));
1994-05-24 10:09:53 +00:00
}
SYSCTL_PROC(_kern, KERN_CLOCKRATE, clockrate,
CTLTYPE_STRUCT|CTLFLAG_RD|CTLFLAG_MPSAFE,
0, 0, sysctl_kern_clockrate, "S,clockinfo",
"Rate and period of various kernel clocks");
static void
watchdog_config(void *unused __unused, u_int cmd, int *error)
{
u_int u;
u = cmd & WD_INTERVAL;
if (u >= WD_TO_1SEC) {
watchdog_ticks = (1 << (u - WD_TO_1SEC)) * hz;
watchdog_enabled = 1;
*error = 0;
} else {
watchdog_enabled = 0;
}
}
/*
* Handle a watchdog timeout by dumping interrupt information and
* then either dropping to DDB or panicking.
*/
static void
watchdog_fire(void)
{
int nintr;
uint64_t inttotal;
u_long *curintr;
char *curname;
curintr = intrcnt;
curname = intrnames;
inttotal = 0;
nintr = sintrcnt / sizeof(u_long);
2006-04-17 20:14:51 +00:00
printf("interrupt total\n");
while (--nintr >= 0) {
if (*curintr)
printf("%-12s %20lu\n", curname, *curintr);
curname += strlen(curname) + 1;
inttotal += *curintr++;
}
printf("Total %20ju\n", (uintmax_t)inttotal);
#if defined(KDB) && !defined(KDB_UNATTENDED)
kdb_backtrace();
kdb_enter(KDB_WHY_WATCHDOG, "watchdog timeout");
#else
panic("watchdog timeout");
#endif
}