File Manager

003 File Manager

Current Path: /usr/src/sys/kern

📁 ..
📄 Make.tags.inc(2.13 KB)
📄 Makefile(302 B)
📄 bus_if.m(26.31 KB)
📄 capabilities.conf(13.67 KB)
📄 clock_if.m(1.7 KB)
📄 cpufreq_if.m(2.27 KB)
📄 device_if.m(10.41 KB)
📄 firmw.S(2.15 KB)
📄 genassym.sh(1.11 KB)
📄 genoffset.c(1.68 KB)
📄 genoffset.sh(3.58 KB)
📄 imgact_aout.c(9.45 KB)
📄 imgact_binmisc.c(18.64 KB)
📄 imgact_elf.c(76.32 KB)
📄 imgact_elf32.c(1.47 KB)
📄 imgact_elf64.c(1.47 KB)
📄 imgact_shell.c(8.41 KB)
📄 init_main.c(24.31 KB)
📄 init_sysent.c(95.3 KB)
📄 kern_acct.c(19.03 KB)
📄 kern_alq.c(24.97 KB)
📄 kern_clock.c(21.12 KB)
📄 kern_clocksource.c(23.34 KB)
📄 kern_condvar.c(11.28 KB)
📄 kern_conf.c(36.14 KB)
📄 kern_cons.c(15.75 KB)
📄 kern_context.c(3.59 KB)
📄 kern_cpu.c(30.77 KB)
📄 kern_cpuset.c(59.78 KB)
📄 kern_ctf.c(8.73 KB)
📄 kern_descrip.c(112.87 KB)
📄 kern_dtrace.c(2.94 KB)
📄 kern_dump.c(8.51 KB)
📄 kern_environment.c(22.75 KB)
📄 kern_et.c(7.1 KB)
📄 kern_event.c(62.49 KB)
📄 kern_exec.c(46.67 KB)
📄 kern_exit.c(34.61 KB)
📄 kern_fail.c(29.32 KB)
📄 kern_ffclock.c(12.66 KB)
📄 kern_fork.c(28.29 KB)
📄 kern_hhook.c(13.58 KB)
📄 kern_idle.c(2.74 KB)
📄 kern_intr.c(40.44 KB)
📄 kern_jail.c(112.67 KB)
📄 kern_kcov.c(15.32 KB)
📄 kern_khelp.c(9.45 KB)
📄 kern_kthread.c(11.8 KB)
📄 kern_ktr.c(11.93 KB)
📄 kern_ktrace.c(31.41 KB)
📄 kern_linker.c(54.3 KB)
📄 kern_lock.c(46.99 KB)
📄 kern_lockf.c(64.46 KB)
📄 kern_lockstat.c(3.8 KB)
📄 kern_loginclass.c(6.69 KB)
📄 kern_malloc.c(37.09 KB)
📄 kern_mbuf.c(43.16 KB)
📄 kern_mib.c(24.26 KB)
📄 kern_module.c(11.05 KB)
📄 kern_mtxpool.c(5.82 KB)
📄 kern_mutex.c(33.62 KB)
📄 kern_ntptime.c(32.49 KB)
📄 kern_osd.c(12.37 KB)
📄 kern_physio.c(5.74 KB)
📄 kern_pmc.c(8.89 KB)
📄 kern_poll.c(15.86 KB)
📄 kern_priv.c(9.14 KB)
📄 kern_proc.c(80.01 KB)
📄 kern_procctl.c(19.48 KB)
📄 kern_prot.c(57.94 KB)
📄 kern_racct.c(34.01 KB)
📄 kern_rangelock.c(8.67 KB)
📄 kern_rctl.c(53.87 KB)
📄 kern_resource.c(36.66 KB)
📄 kern_rmlock.c(28.27 KB)
📄 kern_rwlock.c(40.72 KB)
📄 kern_sdt.c(2.05 KB)
📄 kern_sema.c(4.85 KB)
📄 kern_sendfile.c(33.97 KB)
📄 kern_sharedpage.c(10.37 KB)
📄 kern_shutdown.c(43.34 KB)
📄 kern_sig.c(101.89 KB)
📄 kern_switch.c(13.85 KB)
📄 kern_sx.c(40.27 KB)
📄 kern_synch.c(18.17 KB)
📄 kern_syscalls.c(6.74 KB)
📄 kern_sysctl.c(67.24 KB)
📄 kern_tc.c(55.73 KB)
📄 kern_thr.c(14.14 KB)
📄 kern_thread.c(41.75 KB)
📄 kern_time.c(40.89 KB)
📄 kern_timeout.c(43.08 KB)
📄 kern_tslog.c(3.44 KB)
📄 kern_ubsan.c(50.74 KB)
📄 kern_umtx.c(107.14 KB)
📄 kern_uuid.c(11.68 KB)
📄 kern_xxx.c(10.44 KB)
📄 ksched.c(6.56 KB)
📄 link_elf.c(47.99 KB)
📄 link_elf_obj.c(44.41 KB)
📄 linker_if.m(3.96 KB)
📄 makesyscalls.sh(23.57 KB)
📄 md4c.c(7.89 KB)
📄 md5c.c(9.56 KB)
📄 msi_if.m(2.48 KB)
📄 p1003_1b.c(8.84 KB)
📄 pic_if.m(3.9 KB)
📄 posix4_mib.c(5.59 KB)
📄 sched_4bsd.c(45.03 KB)
📄 sched_ule.c(82.65 KB)
📄 serdev_if.m(3.49 KB)
📄 stack_protector.c(613 B)
📄 subr_acl_nfs4.c(37.42 KB)
📄 subr_acl_posix1e.c(17.71 KB)
📄 subr_atomic64.c(3.97 KB)
📄 subr_autoconf.c(7.7 KB)
📄 subr_blist.c(31.88 KB)
📄 subr_boot.c(5.8 KB)
📄 subr_bufring.c(2.21 KB)
📄 subr_bus.c(145.4 KB)
📄 subr_bus_dma.c(19.67 KB)
📄 subr_busdma_bufalloc.c(5.24 KB)
📄 subr_capability.c(11.93 KB)
📄 subr_clock.c(10.61 KB)
📄 subr_compressor.c(13.11 KB)
📄 subr_counter.c(4.44 KB)
📄 subr_coverage.c(6.17 KB)
📄 subr_csan.c(25.39 KB)
📄 subr_devmap.c(9.8 KB)
📄 subr_devstat.c(16.21 KB)
📄 subr_disk.c(8.54 KB)
📄 subr_dummy_vdso_tc.c(1.7 KB)
📄 subr_early.c(2.26 KB)
📄 subr_epoch.c(25.02 KB)
📄 subr_eventhandler.c(9.17 KB)
📄 subr_fattime.c(9.98 KB)
📄 subr_filter.c(12.2 KB)
📄 subr_firmware.c(13.88 KB)
📄 subr_gtaskqueue.c(20.19 KB)
📄 subr_hash.c(4.8 KB)
📄 subr_hints.c(12.87 KB)
📄 subr_intr.c(40.61 KB)
📄 subr_kdb.c(16.13 KB)
📄 subr_kobj.c(7.1 KB)
📄 subr_lock.c(18.81 KB)
📄 subr_log.c(7.64 KB)
📄 subr_mchain.c(11.06 KB)
📄 subr_module.c(12.98 KB)
📄 subr_msgbuf.c(10.6 KB)
📄 subr_param.c(10.93 KB)
📄 subr_pcpu.c(10.18 KB)
📄 subr_pctrie.c(20.99 KB)
📄 subr_physmem.c(11.52 KB)
📄 subr_pidctrl.c(5.43 KB)
📄 subr_power.c(3.13 KB)
📄 subr_prf.c(27.42 KB)
📄 subr_prng.c(3.36 KB)
📄 subr_prof.c(15.43 KB)
📄 subr_rangeset.c(8.5 KB)
📄 subr_rman.c(27.61 KB)
📄 subr_rtc.c(11.42 KB)
📄 subr_sbuf.c(20.53 KB)
📄 subr_scanf.c(15.59 KB)
📄 subr_sfbuf.c(6.17 KB)
📄 subr_sglist.c(22.83 KB)
📄 subr_sleepqueue.c(39.43 KB)
📄 subr_smp.c(31.62 KB)
📄 subr_smr.c(20.17 KB)
📄 subr_stack.c(6.47 KB)
📄 subr_stats.c(103.01 KB)
📄 subr_syscall.c(7.98 KB)
📄 subr_taskqueue.c(21.1 KB)
📄 subr_terminal.c(15.52 KB)
📄 subr_trap.c(10.87 KB)
📄 subr_turnstile.c(35.58 KB)
📄 subr_uio.c(11.38 KB)
📄 subr_unit.c(22.97 KB)
📄 subr_vmem.c(43.25 KB)
📄 subr_witness.c(84.59 KB)
📄 sys_capability.c(15.06 KB)
📄 sys_eventfd.c(8.42 KB)
📄 sys_generic.c(44.22 KB)
📄 sys_getrandom.c(4.21 KB)
📄 sys_pipe.c(45.14 KB)
📄 sys_procdesc.c(14.57 KB)
📄 sys_process.c(30.73 KB)
📄 sys_socket.c(20.11 KB)
📄 syscalls.c(22.73 KB)
📄 syscalls.master(60.26 KB)
📄 systrace_args.c(178.49 KB)
📄 sysv_ipc.c(6.53 KB)
📄 sysv_msg.c(48.65 KB)
📄 sysv_sem.c(49.85 KB)
📄 sysv_shm.c(43.93 KB)
📄 tty.c(55.14 KB)
📄 tty_compat.c(11.46 KB)
📄 tty_info.c(9.93 KB)
📄 tty_inq.c(12.22 KB)
📄 tty_outq.c(8.74 KB)
📄 tty_pts.c(19.74 KB)
📄 tty_tty.c(2.83 KB)
📄 tty_ttydisc.c(28.6 KB)
📄 uipc_accf.c(8.07 KB)
📄 uipc_debug.c(12.42 KB)
📄 uipc_domain.c(13.13 KB)
📄 uipc_ktls.c(55.7 KB)
📄 uipc_mbuf.c(52.45 KB)
📄 uipc_mbuf2.c(12.64 KB)
📄 uipc_mbufhash.c(4.9 KB)
📄 uipc_mqueue.c(64.64 KB)
📄 uipc_sem.c(25.18 KB)
📄 uipc_shm.c(50.47 KB)
📄 uipc_sockbuf.c(42.9 KB)
📄 uipc_socket.c(110.61 KB)
📄 uipc_syscalls.c(35.94 KB)
📄 uipc_usrreq.c(75.11 KB)
📄 vfs_acl.c(14.5 KB)
📄 vfs_aio.c(76.32 KB)
📄 vfs_bio.c(145.39 KB)
📄 vfs_cache.c(143.09 KB)
📄 vfs_cluster.c(28.36 KB)
📄 vfs_default.c(33.16 KB)
📄 vfs_export.c(14.55 KB)
📄 vfs_extattr.c(17.91 KB)
📄 vfs_hash.c(6 KB)
📄 vfs_init.c(15.86 KB)
📄 vfs_lookup.c(45.48 KB)
📄 vfs_mount.c(62.58 KB)
📄 vfs_mountroot.c(26.23 KB)
📄 vfs_subr.c(167.52 KB)
📄 vfs_syscalls.c(106.86 KB)
📄 vfs_vnops.c(86.28 KB)
📄 vnode_if.src(13.66 KB)

Editing: kern_clock.c

/*-
 * SPDX-License-Identifier: BSD-3-Clause
 *
 * 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
 *    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
 */

#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"

#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>
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/resource.h>
#include <sys/resourcevar.h>
#include <sys/sched.h>
#include <sys/sdt.h>
#include <sys/signalvar.h>
#include <sys/sleepqueue.h>
#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>

#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

#ifdef DEVICE_POLLING
extern void hardclock_device_poll(void);
#endif /* DEVICE_POLLING */

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

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 (%s), "
		    "blocked for %d ticks\n", __func__,
		    td, sched_tdname(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 (%s), "
		    "blocked for %d ticks\n", __func__,
		    td, sched_tdname(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 | CTLFLAG_MPSAFE, 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);
}

/*
 * Clock handling routines.
 *
 * This code is written to operate with two timers that run independently of
 * each other.
 *
 * The main timer, running hz times per second, is used to trigger interval
 * timers, timeouts and rescheduling as needed.
 *
 * 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.
 *
 * 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.)
 *
 * Time-of-day is maintained using a "timecounter", which may or may
 * not be related to the hardware generating the above mentioned
 * interrupts.
 */

int	stathz;
int	profhz;
int	profprocs;
volatile int	ticks;
int	psratio;

DPCPU_DEFINE_STATIC(int, pcputicks);	/* Per-CPU version of ticks. */
#ifdef DEVICE_POLLING
static int devpoll_run = 0;
#endif

/*
 * Initialize clock frequencies and start both clocks running.
 */
/* ARGSUSED*/
static void
initclocks(void *dummy)
{
	int i;

/*
	 * Set divisors to 1 (normal case) and let the machine-specific
	 * code do its bit.
	 */
	mtx_init(&time_lock, "time lock", NULL, MTX_DEF);
	cpu_initclocks();

/*
	 * 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;
#endif
}

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);
	}
}

void
hardclock(int cnt, int usermode)
{
	struct pstats *pstats;
	struct thread *td = curthread;
	struct proc *p = td->td_proc;
	int *t = DPCPU_PTR(pcputicks);
	int global, i, newticks;

/*
	 * Update per-CPU and possibly global ticks values.
	 */
	*t += cnt;
	global = ticks;
	do {
		newticks = *t - global;
		if (newticks <= 0) {
			if (newticks < -1)
				*t = global - 1;
			newticks = 0;
			break;
		}
	} while (!atomic_fcmpset_int(&ticks, &global, *t));

/*
	 * 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);

#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);
#endif
	/* We are in charge to handle this tick duty. */
	if (newticks > 0) {
		tc_ticktock(newticks);
#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 */
		if (watchdog_enabled > 0) {
			i = atomic_fetchadd_int(&watchdog_ticks, -newticks);
			if (i > 0 && i <= newticks)
				watchdog_fire();
		}
		intr_event_handle(clk_intr_event, NULL);
	}
	if (curcpu == CPU_FIRST())
		cpu_tick_calibration();
	if (__predict_false(DPCPU_GET(epoch_cb_count)))
		GROUPTASK_ENQUEUE(DPCPU_PTR(epoch_cb_task));
}

void
hardclock_sync(int cpu)
{
	int *t;
	KASSERT(!CPU_ABSENT(cpu), ("Absent CPU %d", cpu));
	t = DPCPU_ID_PTR(cpu, pcputicks);

*t = ticks;
}

/*
 * Compute number of ticks in the specified amount of time.
 */
int
tvtohz(struct timeval *tv)
{
	unsigned long ticks;
	long sec, usec;

/*
	 * 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.
	 *
	 * 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.
	 */
	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);
}

/*
 * 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)
{

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;
		mtx_lock(&time_lock);
		if (++profprocs == 1)
			cpu_startprofclock();
		mtx_unlock(&time_lock);
	}
}

/*
 * Stop profiling on a process.
 */
void
stopprofclock(struct proc *p)
{

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;
		mtx_lock(&time_lock);
		if (--profprocs == 0)
			cpu_stopprofclock();
		mtx_unlock(&time_lock);
	}
}

/*
 * Statistics clock.  Updates rusage information and calls the scheduler
 * to adjust priorities of the active thread.
 *
 * This should be called by all active processors.
 */
void
statclock(int cnt, int usermode)
{
	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);
	if (usermode) {
		/*
		 * Charge the time as appropriate.
		 */
		td->td_uticks += cnt;
		if (p->p_nice > NZERO)
			cp_time[CP_NICE] += cnt;
		else
			cp_time[CP_USER] += cnt;
	} 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.
		 */
		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;
		}
	}

/* 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
}

void
profclock(int cnt, int usermode, uintfptr_t pc)
{
	struct thread *td;
#ifdef GPROF
	struct gmonparam *g;
	uintfptr_t i;
#endif

td = curthread;
	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;
		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
}

/*
 * Return information about system clocks.
 */
static int
sysctl_kern_clockrate(SYSCTL_HANDLER_ARGS)
{
	struct clockinfo clkinfo;
	/*
	 * Construct clockinfo structure.
	 */
	bzero(&clkinfo, sizeof(clkinfo));
	clkinfo.hz = hz;
	clkinfo.tick = tick;
	clkinfo.profhz = profhz;
	clkinfo.stathz = stathz ? stathz : hz;
	return (sysctl_handle_opaque(oidp, &clkinfo, sizeof clkinfo, req));
}

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);

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
}

003 File Manager

Editing: kern_clock.c

Upload File

Create Folder