time.c

来自「LINUX 2.6.17.4的源码」· C语言 代码 · 共 1,200 行 · 第 1/3 页

/*
 * Common time routines among all ppc machines.
 *
 * Written by Cort Dougan ([e-mail]) to merge
 * Paul Mackerras' version and mine for PReP and Pmac.
 * MPC8xx/MBX changes by Dan Malek ([e-mail]).
 * Converted for 64-bit by Mike Corrigan ([e-mail])
 *
 * First round of bugfixes by Gabriel Paubert ([e-mail])
 * to make clock more stable (2.4.0-test5). The only thing
 * that this code assumes is that the timebases have been synchronized
 * by firmware on SMP and are never stopped (never do sleep
 * on SMP then, nap and doze are OK).
 * 
 * Speeded up do_gettimeofday by getting rid of references to
 * xtime (which required locks for consistency). ([e-mail])
 *
 * TODO (not necessarily in this file):
 * - improve precision and reproducibility of timebase frequency
 * measurement at boot time. (for iSeries, we calibrate the timebase
 * against the Titan chip's clock.)
 * - for astronomical applications: add a new function to get
 * non ambiguous timestamps even around leap seconds. This needs
 * a new timestamp format and a good name.
 *
 * 1997-09-10  Updated NTP code according to technical memorandum Jan '96
 *             "A Kernel Model for Precision Timekeeping" by Dave Mills
 *
 *      This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/config.h>
#include <linux/errno.h>
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/param.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/timex.h>
#include <linux/kernel_stat.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/profile.h>
#include <linux/cpu.h>
#include <linux/security.h>
#include <linux/percpu.h>
#include <linux/rtc.h>
#include <linux/jiffies.h>
#include <linux/posix-timers.h>

#include <asm/io.h>
#include <asm/processor.h>
#include <asm/nvram.h>
#include <asm/cache.h>
#include <asm/machdep.h>
#include <asm/uaccess.h>
#include <asm/time.h>
#include <asm/prom.h>
#include <asm/irq.h>
#include <asm/div64.h>
#include <asm/smp.h>
#include <asm/vdso_datapage.h>
#ifdef CONFIG_PPC64
#include <asm/firmware.h>
#endif
#ifdef CONFIG_PPC_ISERIES
#include <asm/iseries/it_lp_queue.h>
#include <asm/iseries/hv_call_xm.h>
#endif
#include <asm/smp.h>

/* keep track of when we need to update the rtc */
time_t last_rtc_update;
extern int piranha_simulator;
#ifdef CONFIG_PPC_ISERIES
unsigned long iSeries_recal_titan = 0;
unsigned long iSeries_recal_tb = 0; 
static unsigned long first_settimeofday = 1;
#endif

/* The decrementer counts down by 128 every 128ns on a 601. */
#define DECREMENTER_COUNT_601	(1000000000 / HZ)

#define XSEC_PER_SEC (1024*1024)

#ifdef CONFIG_PPC64
#define SCALE_XSEC(xsec, max)	(((xsec) * max) / XSEC_PER_SEC)
#else
/* compute ((xsec << 12) * max) >> 32 */
#define SCALE_XSEC(xsec, max)	mulhwu((xsec) << 12, max)
#endif

unsigned long tb_ticks_per_jiffy;
unsigned long tb_ticks_per_usec = 100; /* sane default */
EXPORT_SYMBOL(tb_ticks_per_usec);
unsigned long tb_ticks_per_sec;
EXPORT_SYMBOL(tb_ticks_per_sec);	/* for cputime_t conversions */
u64 tb_to_xs;
unsigned tb_to_us;

#define TICKLEN_SCALE	(SHIFT_SCALE - 10)
u64 last_tick_len;	/* units are ns / 2^TICKLEN_SCALE */
u64 ticklen_to_xs;	/* 0.64 fraction */

/* If last_tick_len corresponds to about 1/HZ seconds, then
   last_tick_len << TICKLEN_SHIFT will be about 2^63. */
#define TICKLEN_SHIFT	(63 - 30 - TICKLEN_SCALE + SHIFT_HZ)

DEFINE_SPINLOCK(rtc_lock);
EXPORT_SYMBOL_GPL(rtc_lock);

u64 tb_to_ns_scale;
unsigned tb_to_ns_shift;

struct gettimeofday_struct do_gtod;

extern unsigned long wall_jiffies;

extern struct timezone sys_tz;
static long timezone_offset;

unsigned long ppc_proc_freq;
unsigned long ppc_tb_freq;

u64 tb_last_jiffy __cacheline_aligned_in_smp;
unsigned long tb_last_stamp;

/*
 * Note that on ppc32 this only stores the bottom 32 bits of
 * the timebase value, but that's enough to tell when a jiffy
 * has passed.
 */
DEFINE_PER_CPU(unsigned long, last_jiffy);

#ifdef CONFIG_VIRT_CPU_ACCOUNTING
/*
 * Factors for converting from cputime_t (timebase ticks) to
 * jiffies, milliseconds, seconds, and clock_t (1/USER_HZ seconds).
 * These are all stored as 0.64 fixed-point binary fractions.
 */
u64 __cputime_jiffies_factor;
EXPORT_SYMBOL(__cputime_jiffies_factor);
u64 __cputime_msec_factor;
EXPORT_SYMBOL(__cputime_msec_factor);
u64 __cputime_sec_factor;
EXPORT_SYMBOL(__cputime_sec_factor);
u64 __cputime_clockt_factor;
EXPORT_SYMBOL(__cputime_clockt_factor);

static void calc_cputime_factors(void)
{
	struct div_result res;

	div128_by_32(HZ, 0, tb_ticks_per_sec, &res);
	__cputime_jiffies_factor = res.result_low;
	div128_by_32(1000, 0, tb_ticks_per_sec, &res);
	__cputime_msec_factor = res.result_low;
	div128_by_32(1, 0, tb_ticks_per_sec, &res);
	__cputime_sec_factor = res.result_low;
	div128_by_32(USER_HZ, 0, tb_ticks_per_sec, &res);
	__cputime_clockt_factor = res.result_low;
}

/*
 * Read the PURR on systems that have it, otherwise the timebase.
 */
static u64 read_purr(void)
{
	if (cpu_has_feature(CPU_FTR_PURR))
		return mfspr(SPRN_PURR);
	return mftb();
}

/*
 * Account time for a transition between system, hard irq
 * or soft irq state.
 */
void account_system_vtime(struct task_struct *tsk)
{
	u64 now, delta;
	unsigned long flags;

	local_irq_save(flags);
	now = read_purr();
	delta = now - get_paca()->startpurr;
	get_paca()->startpurr = now;
	if (!in_interrupt()) {
		delta += get_paca()->system_time;
		get_paca()->system_time = 0;
	}
	account_system_time(tsk, 0, delta);
	local_irq_restore(flags);
}

/*
 * Transfer the user and system times accumulated in the paca
 * by the exception entry and exit code to the generic process
 * user and system time records.
 * Must be called with interrupts disabled.
 */
void account_process_vtime(struct task_struct *tsk)
{
	cputime_t utime;

	utime = get_paca()->user_time;
	get_paca()->user_time = 0;
	account_user_time(tsk, utime);
}

static void account_process_time(struct pt_regs *regs)
{
	int cpu = smp_processor_id();

	account_process_vtime(current);
	run_local_timers();
	if (rcu_pending(cpu))
		rcu_check_callbacks(cpu, user_mode(regs));
	scheduler_tick();
 	run_posix_cpu_timers(current);
}

#ifdef CONFIG_PPC_SPLPAR
/*
 * Stuff for accounting stolen time.
 */
struct cpu_purr_data {
	int	initialized;			/* thread is running */
	u64	tb0;			/* timebase at origin time */
	u64	purr0;			/* PURR at origin time */
	u64	tb;			/* last TB value read */
	u64	purr;			/* last PURR value read */
	u64	stolen;			/* stolen time so far */
	spinlock_t lock;
};

static DEFINE_PER_CPU(struct cpu_purr_data, cpu_purr_data);

static void snapshot_tb_and_purr(void *data)
{
	struct cpu_purr_data *p = &__get_cpu_var(cpu_purr_data);

	p->tb0 = mftb();
	p->purr0 = mfspr(SPRN_PURR);
	p->tb = p->tb0;
	p->purr = 0;
	wmb();
	p->initialized = 1;
}

/*
 * Called during boot when all cpus have come up.
 */
void snapshot_timebases(void)
{
	int cpu;

	if (!cpu_has_feature(CPU_FTR_PURR))
		return;
	for_each_possible_cpu(cpu)
		spin_lock_init(&per_cpu(cpu_purr_data, cpu).lock);
	on_each_cpu(snapshot_tb_and_purr, NULL, 0, 1);
}

void calculate_steal_time(void)
{
	u64 tb, purr, t0;
	s64 stolen;
	struct cpu_purr_data *p0, *pme, *phim;
	int cpu;

	if (!cpu_has_feature(CPU_FTR_PURR))
		return;
	cpu = smp_processor_id();
	pme = &per_cpu(cpu_purr_data, cpu);
	if (!pme->initialized)
		return;		/* this can happen in early boot */
	p0 = &per_cpu(cpu_purr_data, cpu & ~1);
	phim = &per_cpu(cpu_purr_data, cpu ^ 1);
	spin_lock(&p0->lock);
	tb = mftb();
	purr = mfspr(SPRN_PURR) - pme->purr0;
	if (!phim->initialized || !cpu_online(cpu ^ 1)) {
		stolen = (tb - pme->tb) - (purr - pme->purr);
	} else {
		t0 = pme->tb0;
		if (phim->tb0 < t0)
			t0 = phim->tb0;
		stolen = phim->tb - t0 - phim->purr - purr - p0->stolen;
	}
	if (stolen > 0) {
		account_steal_time(current, stolen);
		p0->stolen += stolen;
	}
	pme->tb = tb;
	pme->purr = purr;
	spin_unlock(&p0->lock);
}

/*
 * Must be called before the cpu is added to the online map when
 * a cpu is being brought up at runtime.
 */
static void snapshot_purr(void)
{
	int cpu;
	u64 purr;
	struct cpu_purr_data *p0, *pme, *phim;
	unsigned long flags;

	if (!cpu_has_feature(CPU_FTR_PURR))
		return;
	cpu = smp_processor_id();
	pme = &per_cpu(cpu_purr_data, cpu);
	p0 = &per_cpu(cpu_purr_data, cpu & ~1);
	phim = &per_cpu(cpu_purr_data, cpu ^ 1);
	spin_lock_irqsave(&p0->lock, flags);
	pme->tb = pme->tb0 = mftb();
	purr = mfspr(SPRN_PURR);
	if (!phim->initialized) {
		pme->purr = 0;
		pme->purr0 = purr;
	} else {
		/* set p->purr and p->purr0 for no change in p0->stolen */
		pme->purr = phim->tb - phim->tb0 - phim->purr - p0->stolen;
		pme->purr0 = purr - pme->purr;
	}
	pme->initialized = 1;
	spin_unlock_irqrestore(&p0->lock, flags);
}

#endif /* CONFIG_PPC_SPLPAR */

#else /* ! CONFIG_VIRT_CPU_ACCOUNTING */
#define calc_cputime_factors()
#define account_process_time(regs)	update_process_times(user_mode(regs))
#define calculate_steal_time()		do { } while (0)
#endif

#if !(defined(CONFIG_VIRT_CPU_ACCOUNTING) && defined(CONFIG_PPC_SPLPAR))
#define snapshot_purr()			do { } while (0)
#endif

/*
 * Called when a cpu comes up after the system has finished booting,
 * i.e. as a result of a hotplug cpu action.
 */
void snapshot_timebase(void)
{
	__get_cpu_var(last_jiffy) = get_tb();
	snapshot_purr();
}

void __delay(unsigned long loops)
{
	unsigned long start;
	int diff;

	if (__USE_RTC()) {
		start = get_rtcl();
		do {
			/* the RTCL register wraps at 1000000000 */
			diff = get_rtcl() - start;
			if (diff < 0)
				diff += 1000000000;
		} while (diff < loops);
	} else {
		start = get_tbl();
		while (get_tbl() - start < loops)
			HMT_low();
		HMT_medium();
	}
}
EXPORT_SYMBOL(__delay);

void udelay(unsigned long usecs)
{
	__delay(tb_ticks_per_usec * usecs);
}
EXPORT_SYMBOL(udelay);

static __inline__ void timer_check_rtc(void)
{
        /*
         * update the rtc when needed, this should be performed on the
         * right fraction of a second. Half or full second ?
         * Full second works on mk48t59 clocks, others need testing.
         * Note that this update is basically only used through 
         * the adjtimex system calls. Setting the HW clock in
         * any other way is a /dev/rtc and userland business.
         * This is still wrong by -0.5/+1.5 jiffies because of the
         * timer interrupt resolution and possible delay, but here we 
         * hit a quantization limit which can only be solved by higher
         * resolution timers and decoupling time management from timer
         * interrupts. This is also wrong on the clocks
         * which require being written at the half second boundary.