qemu-timer.c

/*
 * QEMU System Emulator
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */

#include "sysemu.h"
#include "net.h"
#include "monitor.h"
#include "console.h"

#include "hw/hw.h"

#include <unistd.h>
#include <fcntl.h>
#include <time.h>
#include <errno.h>
#include <sys/time.h>
#include <signal.h>
#ifdef __FreeBSD__
#include <sys/param.h>
#endif

#ifdef __linux__
#include <sys/ioctl.h>
#include <linux/rtc.h>
/* For the benefit of older linux systems which don't supply it,
   we use a local copy of hpet.h. */
/* #include <linux/hpet.h> */
#include "hpet.h"
#endif

#ifdef _WIN32
#include <windows.h>
#include <mmsystem.h>
#endif

#include "qemu-timer.h"

/* Conversion factor from emulated instructions to virtual clock ticks.  */
int icount_time_shift;
/* Arbitrarily pick 1MIPS as the minimum allowable speed.  */
#define MAX_ICOUNT_SHIFT 10
/* Compensate for varying guest execution speed.  */
int64_t qemu_icount_bias;
static QEMUTimer *icount_rt_timer;
static QEMUTimer *icount_vm_timer;


/***********************************************************/
/* real time host monotonic timer */


static int64_t get_clock_realtime(void)
{
    struct timeval tv;

    gettimeofday(&tv, NULL);
    return tv.tv_sec * 1000000000LL + (tv.tv_usec * 1000);
}

#ifdef WIN32

static int64_t clock_freq;

static void init_get_clock(void)
{
    LARGE_INTEGER freq;
    int ret;
    ret = QueryPerformanceFrequency(&freq);
    if (ret == 0) {
        fprintf(stderr, "Could not calibrate ticks\n");
        exit(1);
    }
    clock_freq = freq.QuadPart;
}

static int64_t get_clock(void)
{
    LARGE_INTEGER ti;
    QueryPerformanceCounter(&ti);
    return muldiv64(ti.QuadPart, get_ticks_per_sec(), clock_freq);
}

#else

static int use_rt_clock;

static void init_get_clock(void)
{
    use_rt_clock = 0;
#if defined(__linux__) || (defined(__FreeBSD__) && __FreeBSD_version >= 500000) \
    || defined(__DragonFly__) || defined(__FreeBSD_kernel__)
    {
        struct timespec ts;
        if (clock_gettime(CLOCK_MONOTONIC, &ts) == 0) {
            use_rt_clock = 1;
        }
    }
#endif
}

static int64_t get_clock(void)
{
#if defined(__linux__) || (defined(__FreeBSD__) && __FreeBSD_version >= 500000) \
	|| defined(__DragonFly__) || defined(__FreeBSD_kernel__)
    if (use_rt_clock) {
        struct timespec ts;
        clock_gettime(CLOCK_MONOTONIC, &ts);
        return ts.tv_sec * 1000000000LL + ts.tv_nsec;
    } else
#endif
    {
        /* XXX: using gettimeofday leads to problems if the date
           changes, so it should be avoided. */
        return get_clock_realtime();
    }
}
#endif

/***********************************************************/
/* guest cycle counter */

typedef struct TimersState {
    int64_t cpu_ticks_prev;
    int64_t cpu_ticks_offset;
    int64_t cpu_clock_offset;
    int32_t cpu_ticks_enabled;
    int64_t dummy;
} TimersState;

static void timer_save(QEMUFile *f, void *opaque)
{
    TimersState *s = opaque;

    if (s->cpu_ticks_enabled) {
        hw_error("cannot save state if virtual timers are running");
    }
    qemu_put_be64(f, s->cpu_ticks_prev);
    qemu_put_be64(f, s->cpu_ticks_offset);
    qemu_put_be64(f, s->cpu_clock_offset);
 }

static int timer_load(QEMUFile *f, void *opaque, int version_id)
{
    TimersState *s = opaque;

    if (version_id != 1 && version_id != 2)
        return -EINVAL;
    if (s->cpu_ticks_enabled) {
        return -EINVAL;
    }
    s->cpu_ticks_prev   = qemu_get_sbe64(f);
    s->cpu_ticks_offset = qemu_get_sbe64(f);
    if (version_id == 2) {
        s->cpu_clock_offset = qemu_get_sbe64(f);
    }
    return 0;
}


TimersState timers_state;

/* return the host CPU cycle counter and handle stop/restart */
int64_t cpu_get_ticks(void)
{
    if (use_icount) {
        return cpu_get_icount();
    }
    if (!timers_state.cpu_ticks_enabled) {
        return timers_state.cpu_ticks_offset;
    } else {
        int64_t ticks;
        ticks = cpu_get_real_ticks();
        if (timers_state.cpu_ticks_prev > ticks) {
            /* Note: non increasing ticks may happen if the host uses
               software suspend */
            timers_state.cpu_ticks_offset += timers_state.cpu_ticks_prev - ticks;
        }
        timers_state.cpu_ticks_prev = ticks;
        return ticks + timers_state.cpu_ticks_offset;
    }
}

/* return the host CPU monotonic timer and handle stop/restart */
static int64_t cpu_get_clock(void)
{
    int64_t ti;
    if (!timers_state.cpu_ticks_enabled) {
        return timers_state.cpu_clock_offset;
    } else {
        ti = get_clock();
        return ti + timers_state.cpu_clock_offset;
    }
}

#ifndef CONFIG_IOTHREAD
static int64_t qemu_icount_delta(void)
{
    if (!use_icount) {
        return 5000 * (int64_t) 1000000;
    } else if (use_icount == 1) {
        /* When not using an adaptive execution frequency
           we tend to get badly out of sync with real time,
           so just delay for a reasonable amount of time.  */
        return 0;
    } else {
        return cpu_get_icount() - cpu_get_clock();
    }
}
#endif

/* enable cpu_get_ticks() */
void cpu_enable_ticks(void)
{
    if (!timers_state.cpu_ticks_enabled) {
        timers_state.cpu_ticks_offset -= cpu_get_real_ticks();
        timers_state.cpu_clock_offset -= get_clock();
        timers_state.cpu_ticks_enabled = 1;
    }
}

/* disable cpu_get_ticks() : the clock is stopped. You must not call
   cpu_get_ticks() after that.  */
void cpu_disable_ticks(void)
{
    if (timers_state.cpu_ticks_enabled) {
        timers_state.cpu_ticks_offset = cpu_get_ticks();
        timers_state.cpu_clock_offset = cpu_get_clock();
        timers_state.cpu_ticks_enabled = 0;
    }
}

/***********************************************************/
/* timers */

#define QEMU_CLOCK_REALTIME 0
#define QEMU_CLOCK_VIRTUAL  1
#define QEMU_CLOCK_HOST     2

struct QEMUClock {
    int type;
    int enabled;
    /* XXX: add frequency */
};

struct QEMUTimer {
    QEMUClock *clock;
    int64_t expire_time;
    QEMUTimerCB *cb;
    void *opaque;
    struct QEMUTimer *next;
};

struct qemu_alarm_timer {
    char const *name;
    int (*start)(struct qemu_alarm_timer *t);
    void (*stop)(struct qemu_alarm_timer *t);
    void (*rearm)(struct qemu_alarm_timer *t);
    void *priv;

    char expired;
    char pending;
};

static struct qemu_alarm_timer *alarm_timer;

int qemu_alarm_pending(void)
{
    return alarm_timer->pending;
}

static inline int alarm_has_dynticks(struct qemu_alarm_timer *t)
{
    return !!t->rearm;
}

static void qemu_rearm_alarm_timer(struct qemu_alarm_timer *t)
{
    if (!alarm_has_dynticks(t))
        return;

    t->rearm(t);
}

/* TODO: MIN_TIMER_REARM_US should be optimized */
#define MIN_TIMER_REARM_US 250

#ifdef _WIN32

struct qemu_alarm_win32 {
    MMRESULT timerId;
    unsigned int period;
} alarm_win32_data = {0, 0};

static int win32_start_timer(struct qemu_alarm_timer *t);
static void win32_stop_timer(struct qemu_alarm_timer *t);
static void win32_rearm_timer(struct qemu_alarm_timer *t);

#else

static int unix_start_timer(struct qemu_alarm_timer *t);
static void unix_stop_timer(struct qemu_alarm_timer *t);

#ifdef __linux__

static int dynticks_start_timer(struct qemu_alarm_timer *t);
static void dynticks_stop_timer(struct qemu_alarm_timer *t);
static void dynticks_rearm_timer(struct qemu_alarm_timer *t);

static int hpet_start_timer(struct qemu_alarm_timer *t);
static void hpet_stop_timer(struct qemu_alarm_timer *t);

static int rtc_start_timer(struct qemu_alarm_timer *t);
static void rtc_stop_timer(struct qemu_alarm_timer *t);

#endif /* __linux__ */

#endif /* _WIN32 */

/* Correlation between real and virtual time is always going to be
   fairly approximate, so ignore small variation.
   When the guest is idle real and virtual time will be aligned in
   the IO wait loop.  */
#define ICOUNT_WOBBLE (get_ticks_per_sec() / 10)

static void icount_adjust(void)
{
    int64_t cur_time;
    int64_t cur_icount;
    int64_t delta;
    static int64_t last_delta;
    /* If the VM is not running, then do nothing.  */
    if (!vm_running)
        return;

    cur_time = cpu_get_clock();
    cur_icount = qemu_get_clock(vm_clock);
    delta = cur_icount - cur_time;
    /* FIXME: This is a very crude algorithm, somewhat prone to oscillation.  */
    if (delta > 0
        && last_delta + ICOUNT_WOBBLE < delta * 2
        && icount_time_shift > 0) {
        /* The guest is getting too far ahead.  Slow time down.  */
        icount_time_shift--;
    }
    if (delta < 0
        && last_delta - ICOUNT_WOBBLE > delta * 2
        && icount_time_shift < MAX_ICOUNT_SHIFT) {
        /* The guest is getting too far behind.  Speed time up.  */
        icount_time_shift++;
    }
    last_delta = delta;
    qemu_icount_bias = cur_icount - (qemu_icount << icount_time_shift);
}

static void icount_adjust_rt(void * opaque)
{
    qemu_mod_timer(icount_rt_timer,
                   qemu_get_clock(rt_clock) + 1000);
    icount_adjust();
}

static void icount_adjust_vm(void * opaque)
{
    qemu_mod_timer(icount_vm_timer,
                   qemu_get_clock(vm_clock) + get_ticks_per_sec() / 10);
    icount_adjust();
}

int64_t qemu_icount_round(int64_t count)
{
    return (count + (1 << icount_time_shift) - 1) >> icount_time_shift;
}

static struct qemu_alarm_timer alarm_timers[] = {
#ifndef _WIN32
#ifdef __linux__
    /* HPET - if available - is preferred */
    {"hpet", hpet_start_timer, hpet_stop_timer, NULL, NULL},
    /* ...otherwise try RTC */
    {"rtc", rtc_start_timer, rtc_stop_timer, NULL, NULL},
#endif
    {"unix", unix_start_timer, unix_stop_timer, NULL, NULL},
#ifdef __linux__
    /* on Linux, the 'dynticks' clock sometimes doesn't work
     * properly. this results in the UI freezing while emulation
     * continues, for several seconds... So move it to the end
     * of the list. */
    {"dynticks", dynticks_start_timer,
     dynticks_stop_timer, dynticks_rearm_timer, NULL},
#endif
#else
    {"dynticks", win32_start_timer,
     win32_stop_timer, win32_rearm_timer, &alarm_win32_data},
    {"win32", win32_start_timer,
     win32_stop_timer, NULL, &alarm_win32_data},
#endif
    {NULL, }
};

static void show_available_alarms(void)
{
    int i;

    printf("Available alarm timers, in order of precedence:\n");
    for (i = 0; alarm_timers[i].name; i++)
        printf("%s\n", alarm_timers[i].name);
}

void configure_alarms(char const *opt)
{
    int i;
    int cur = 0;
    int count = ARRAY_SIZE(alarm_timers) - 1;
    char *arg;
    char *name;
    struct qemu_alarm_timer tmp;

    if (!strcmp(opt, "?")) {
        show_available_alarms();
        exit(0);
    }

    arg = qemu_strdup(opt);

    /* Reorder the array */
    name = strtok(arg, ",");
    while (name) {
        for (i = 0; i < count && alarm_timers[i].name; i++) {
            if (!strcmp(alarm_timers[i].name, name))
                break;
        }

        if (i == count) {
            fprintf(stderr, "Unknown clock %s\n", name);
            goto next;
        }

        if (i < cur)
            /* Ignore */
            goto next;

	/* Swap */
        tmp = alarm_timers[i];
        alarm_timers[i] = alarm_timers[cur];
        alarm_timers[cur] = tmp;

        cur++;
next:
        name = strtok(NULL, ",");
    }

    qemu_free(arg);

    if (cur) {
        /* Disable remaining timers */
        for (i = cur; i < count; i++)
            alarm_timers[i].name = NULL;
    } else {
        show_available_alarms();
        exit(1);
    }
}

#define QEMU_NUM_CLOCKS 3

QEMUClock *rt_clock;
QEMUClock *vm_clock;
QEMUClock *host_clock;

static QEMUTimer *active_timers[QEMU_NUM_CLOCKS];

static QEMUClock *qemu_new_clock(int type)
{
    QEMUClock *clock;
    clock = qemu_mallocz(sizeof(QEMUClock));
    clock->type = type;
    clock->enabled = 1;
    return clock;
}

void qemu_clock_enable(QEMUClock *clock, int enabled)
{
    clock->enabled = enabled;
}

QEMUTimer *qemu_new_timer(QEMUClock *clock, QEMUTimerCB *cb, void *opaque)
{
    QEMUTimer *ts;

    ts = qemu_mallocz(sizeof(QEMUTimer));
    ts->clock = clock;
    ts->cb = cb;
    ts->opaque = opaque;
    return ts;
}

void qemu_free_timer(QEMUTimer *ts)
{
    qemu_free(ts);
}

/* stop a timer, but do not dealloc it */
void qemu_del_timer(QEMUTimer *ts)
{
    QEMUTimer **pt, *t;

    /* NOTE: this code must be signal safe because
       qemu_timer_expired() can be called from a signal. */
    pt = &active_timers[ts->clock->type];
    for(;;) {
        t = *pt;
        if (!t)
            break;
        if (t == ts) {
            *pt = t->next;
            break;
        }
        pt = &t->next;
    }
}

/* modify the current timer so that it will be fired when current_time
   >= expire_time. The corresponding callback will be called. */
void qemu_mod_timer(QEMUTimer *ts, int64_t expire_time)
{
    QEMUTimer **pt, *t;

    qemu_del_timer(ts);

    /* add the timer in the sorted list */
    /* NOTE: this code must be signal safe because
       qemu_timer_expired() can be called from a signal. */
    pt = &active_timers[ts->clock->type];
    for(;;) {
        t = *pt;
        if (!t)
            break;
        if (t->expire_time > expire_time)
            break;
        pt = &t->next;
    }
    ts->expire_time = expire_time;
    ts->next = *pt;
    *pt = ts;

    /* Rearm if necessary  */
    if (pt == &active_timers[ts->clock->type]) {
        if (!alarm_timer->pending) {
            qemu_rearm_alarm_timer(alarm_timer);
        }
        /* Interrupt execution to force deadline recalculation.  */
        if (use_icount)
            qemu_notify_event();
    }
}

int qemu_timer_pending(QEMUTimer *ts)
{
    QEMUTimer *t;
    for(t = active_timers[ts->clock->type]; t != NULL; t = t->next) {
        if (t == ts)
            return 1;
    }
    return 0;
}

int qemu_timer_expired(QEMUTimer *timer_head, int64_t current_time)
{
    if (!timer_head)
        return 0;
    return (timer_head->expire_time <= current_time);
}

static void qemu_run_timers(QEMUClock *clock)
{
    QEMUTimer **ptimer_head, *ts;
    int64_t current_time;

    if (!clock->enabled)
        return;

    current_time = qemu_get_clock (clock);
    ptimer_head = &active_timers[clock->type];
    for(;;) {
        ts = *ptimer_head;
        if (!ts || ts->expire_time > current_time)
            break;
        /* remove timer from the list before calling the callback */
        *ptimer_head = ts->next;
        ts->next = NULL;

        /* run the callback (the timer list can be modified) */
        ts->cb(ts->opaque);
    }
}

int64_t qemu_get_clock(QEMUClock *clock)
{
    switch(clock->type) {
    case QEMU_CLOCK_REALTIME:
        return get_clock() / 1000000;
    default:
    case QEMU_CLOCK_VIRTUAL:
        if (use_icount) {
            return cpu_get_icount();
        } else {
            return cpu_get_clock();
        }
    case QEMU_CLOCK_HOST:
        return get_clock_realtime();
    }
}

int64_t qemu_get_clock_ns(QEMUClock *clock)
{
    switch(clock->type) {
    case QEMU_CLOCK_REALTIME:
        return get_clock();
    default:
    case QEMU_CLOCK_VIRTUAL:
        if (use_icount) {
            return cpu_get_icount();
        } else {
            return cpu_get_clock();
        }
    case QEMU_CLOCK_HOST:
        return get_clock_realtime();
    }
}

void init_clocks(void)
{
    init_get_clock();
    rt_clock = qemu_new_clock(QEMU_CLOCK_REALTIME);
    vm_clock = qemu_new_clock(QEMU_CLOCK_VIRTUAL);
    host_clock = qemu_new_clock(QEMU_CLOCK_HOST);

    rtc_clock = host_clock;
}

/* save a timer */
void qemu_put_timer(QEMUFile *f, QEMUTimer *ts)
{
    uint64_t expire_time;

    if (qemu_timer_pending(ts)) {
        expire_time = ts->expire_time;
    } else {
        expire_time = -1;
    }
    qemu_put_be64(f, expire_time);
}

void qemu_get_timer(QEMUFile *f, QEMUTimer *ts)
{
    uint64_t expire_time;

    expire_time = qemu_get_be64(f);
    if (expire_time != -1) {
        qemu_mod_timer(ts, expire_time);
    } else {
        qemu_del_timer(ts);
    }
}

#if 0
static const VMStateDescription vmstate_timers = {
    .name = "timer",
    .version_id = 2,
    .minimum_version_id = 1,
    .minimum_version_id_old = 1,
    .fields      = (VMStateField []) {
        VMSTATE_INT64(cpu_ticks_offset, TimersState),
        VMSTATE_INT64(dummy, TimersState),
        VMSTATE_INT64_V(cpu_clock_offset, TimersState, 2),
        VMSTATE_END_OF_LIST()
    }
};
#endif

void configure_icount(const char *option)
{
    register_savevm("timer", 0, 2, timer_save, timer_load, &timers_state);

    if (!option)
        return;

    if (strcmp(option, "auto") != 0) {
        icount_time_shift = strtol(option, NULL, 0);
        use_icount = 1;
        return;
    }

    use_icount = 2;

    /* 125MIPS seems a reasonable initial guess at the guest speed.
       It will be corrected fairly quickly anyway.  */
    icount_time_shift = 3;

    /* Have both realtime and virtual time triggers for speed adjustment.
       The realtime trigger catches emulated time passing too slowly,
       the virtual time trigger catches emulated time passing too fast.
       Realtime triggers occur even when idle, so use them less frequently
       than VM triggers.  */
    icount_rt_timer = qemu_new_timer(rt_clock, icount_adjust_rt, NULL);
    qemu_mod_timer(icount_rt_timer,
                   qemu_get_clock(rt_clock) + 1000);
    icount_vm_timer = qemu_new_timer(vm_clock, icount_adjust_vm, NULL);
    qemu_mod_timer(icount_vm_timer,
                   qemu_get_clock(vm_clock) + get_ticks_per_sec() / 10);
}

void qemu_run_all_timers(void)
{
    alarm_timer->pending = 0;

    /* rearm timer, if not periodic */
    if (alarm_timer->expired) {
        alarm_timer->expired = 0;
        qemu_rearm_alarm_timer(alarm_timer);
    }

    /* vm time timers */
    if (vm_running) {
        qemu_run_timers(vm_clock);
    }

    qemu_run_timers(rt_clock);
    qemu_run_timers(host_clock);
}

static int timer_alarm_pending = 1;

int qemu_timer_alarm_pending(void)
{
    int ret = timer_alarm_pending;
    timer_alarm_pending = 0;
    return ret;
}

#ifdef _WIN32
static void CALLBACK host_alarm_handler(UINT uTimerID, UINT uMsg,
                                        DWORD_PTR dwUser, DWORD_PTR dw1,
                                        DWORD_PTR dw2)
#else
static void host_alarm_handler(int host_signum)
#endif
{
    struct qemu_alarm_timer *t = alarm_timer;
    if (!t)
	return;

#if 0
#define DISP_FREQ 1000
    {
        static int64_t delta_min = INT64_MAX;
        static int64_t delta_max, delta_cum, last_clock, delta, ti;
        static int count;
        ti = qemu_get_clock(vm_clock);
        if (last_clock != 0) {
            delta = ti - last_clock;
            if (delta < delta_min)
                delta_min = delta;
            if (delta > delta_max)
                delta_max = delta;
            delta_cum += delta;
            if (++count == DISP_FREQ) {
                printf("timer: min=%" PRId64 " us max=%" PRId64 " us avg=%" PRId64 " us avg_freq=%0.3f Hz\n",
                       muldiv64(delta_min, 1000000, get_ticks_per_sec()),
                       muldiv64(delta_max, 1000000, get_ticks_per_sec()),
                       muldiv64(delta_cum, 1000000 / DISP_FREQ, get_ticks_per_sec()),
                       (double)get_ticks_per_sec() / ((double)delta_cum / DISP_FREQ));
                count = 0;
                delta_min = INT64_MAX;
                delta_max = 0;
                delta_cum = 0;
            }
        }
        last_clock = ti;
    }
#endif
    if (alarm_has_dynticks(t) ||
        (!use_icount &&
            qemu_timer_expired(active_timers[QEMU_CLOCK_VIRTUAL],
                               qemu_get_clock(vm_clock))) ||
        qemu_timer_expired(active_timers[QEMU_CLOCK_REALTIME],
                           qemu_get_clock(rt_clock)) ||
        qemu_timer_expired(active_timers[QEMU_CLOCK_HOST],
                           qemu_get_clock(host_clock))) {

        t->expired = alarm_has_dynticks(t);
        t->pending = 1;
        timer_alarm_pending = 1;
        qemu_notify_event();
    }
}

int64_t qemu_next_deadline(void)
{
    /* To avoid problems with overflow limit this to 2^32.  */
    int64_t delta = INT32_MAX;

    if (active_timers[QEMU_CLOCK_VIRTUAL]) {
        delta = active_timers[QEMU_CLOCK_VIRTUAL]->expire_time -
                     qemu_get_clock(vm_clock);
    }
    if (active_timers[QEMU_CLOCK_HOST]) {
        int64_t hdelta = active_timers[QEMU_CLOCK_HOST]->expire_time -
                 qemu_get_clock(host_clock);
        if (hdelta < delta)
            delta = hdelta;
    }

    if (delta < 0)
        delta = 0;

    return delta;
}

#ifndef _WIN32

#if defined(__linux__)

#define RTC_FREQ 1024

static uint64_t qemu_next_deadline_dyntick(void)
{
    int64_t delta;
    int64_t rtdelta;

    if (use_icount)
        delta = INT32_MAX;
    else
        delta = (qemu_next_deadline() + 999) / 1000;

    if (active_timers[QEMU_CLOCK_REALTIME]) {
        rtdelta = (active_timers[QEMU_CLOCK_REALTIME]->expire_time -
                 qemu_get_clock(rt_clock))*1000;
        if (rtdelta < delta)
            delta = rtdelta;
    }

    if (delta < MIN_TIMER_REARM_US)
        delta = MIN_TIMER_REARM_US;

    return delta;
}

static void enable_sigio_timer(int fd)
{
    struct sigaction act;

    /* timer signal */
    sigfillset(&act.sa_mask);
    act.sa_flags = 0;
    act.sa_handler = host_alarm_handler;

    sigaction(SIGIO, &act, NULL);
    fcntl_setfl(fd, O_ASYNC);
    fcntl(fd, F_SETOWN, getpid());
}

static int hpet_start_timer(struct qemu_alarm_timer *t)
{
    struct hpet_info info;
    int r, fd;

    fd = open("/dev/hpet", O_RDONLY);
    if (fd < 0)
        return -1;

    /* Set frequency */
    r = ioctl(fd, HPET_IRQFREQ, RTC_FREQ);
    if (r < 0) {
        fprintf(stderr, "Could not configure '/dev/hpet' to have a 1024Hz timer. This is not a fatal\n"
                "error, but for better emulation accuracy type:\n"
                "'echo 1024 > /proc/sys/dev/hpet/max-user-freq' as root.\n");
        goto fail;
    }

    /* Check capabilities */
    r = ioctl(fd, HPET_INFO, &info);
    if (r < 0)
        goto fail;

    /* Enable periodic mode */
    r = ioctl(fd, HPET_EPI, 0);
    if (info.hi_flags && (r < 0))
        goto fail;

    /* Enable interrupt */
    r = ioctl(fd, HPET_IE_ON, 0);
    if (r < 0)
        goto fail;

    enable_sigio_timer(fd);
    t->priv = (void *)(long)fd;

    return 0;
fail:
    close(fd);
    return -1;
}

static void hpet_stop_timer(struct qemu_alarm_timer *t)
{
    int fd = (long)t->priv;

    close(fd);
}

static int rtc_start_timer(struct qemu_alarm_timer *t)
{
    int rtc_fd;
    unsigned long current_rtc_freq = 0;

    TFR(rtc_fd = open("/dev/rtc", O_RDONLY));
    if (rtc_fd < 0)
        return -1;
    ioctl(rtc_fd, RTC_IRQP_READ, &current_rtc_freq);
    if (current_rtc_freq != RTC_FREQ &&
        ioctl(rtc_fd, RTC_IRQP_SET, RTC_FREQ) < 0) {
        fprintf(stderr, "Could not configure '/dev/rtc' to have a 1024 Hz timer. This is not a fatal\n"
                "error, but for better emulation accuracy either use a 2.6 host Linux kernel or\n"
                "type 'echo 1024 > /proc/sys/dev/rtc/max-user-freq' as root.\n");
        goto fail;
    }
    if (ioctl(rtc_fd, RTC_PIE_ON, 0) < 0) {
    fail:
        close(rtc_fd);
        return -1;
    }

    enable_sigio_timer(rtc_fd);

    t->priv = (void *)(long)rtc_fd;

    return 0;
}

static void rtc_stop_timer(struct qemu_alarm_timer *t)
{
    int rtc_fd = (long)t->priv;

    close(rtc_fd);
}

static int dynticks_start_timer(struct qemu_alarm_timer *t)
{
    struct sigevent ev;
    timer_t host_timer;
    struct sigaction act;

    sigfillset(&act.sa_mask);
    act.sa_flags = 0;
    act.sa_handler = host_alarm_handler;

    sigaction(SIGALRM, &act, NULL);

    /*
     * Initialize ev struct to 0 to avoid valgrind complaining
     * about uninitialized data in timer_create call
     */
    memset(&ev, 0, sizeof(ev));
    ev.sigev_value.sival_int = 0;
    ev.sigev_notify = SIGEV_SIGNAL;
    ev.sigev_signo = SIGALRM;

    if (timer_create(CLOCK_REALTIME, &ev, &host_timer)) {
        perror("timer_create");

        /* disable dynticks */
        fprintf(stderr, "Dynamic Ticks disabled\n");

        return -1;
    }

    t->priv = (void *)(long)host_timer;

    return 0;
}

static void dynticks_stop_timer(struct qemu_alarm_timer *t)
{
    timer_t host_timer = (timer_t)(long)t->priv;

    timer_delete(host_timer);
}

static void dynticks_rearm_timer(struct qemu_alarm_timer *t)
{
    timer_t host_timer = (timer_t)(long)t->priv;
    struct itimerspec timeout;
    int64_t nearest_delta_us = INT64_MAX;
    int64_t current_us;

    assert(alarm_has_dynticks(t));
    if (!active_timers[QEMU_CLOCK_REALTIME] &&
        !active_timers[QEMU_CLOCK_VIRTUAL] &&
        !active_timers[QEMU_CLOCK_HOST])
        return;

    nearest_delta_us = qemu_next_deadline_dyntick();

    /* check whether a timer is already running */
    if (timer_gettime(host_timer, &timeout)) {
        perror("gettime");
        fprintf(stderr, "Internal timer error: aborting\n");
        exit(1);
    }
    current_us = timeout.it_value.tv_sec * 1000000 + timeout.it_value.tv_nsec/1000;
    if (current_us && current_us <= nearest_delta_us)
        return;

    timeout.it_interval.tv_sec = 0;
    timeout.it_interval.tv_nsec = 0; /* 0 for one-shot timer */
    timeout.it_value.tv_sec =  nearest_delta_us / 1000000;
    timeout.it_value.tv_nsec = (nearest_delta_us % 1000000) * 1000;
    if (timer_settime(host_timer, 0 /* RELATIVE */, &timeout, NULL)) {
        perror("settime");
        fprintf(stderr, "Internal timer error: aborting\n");
        exit(1);
    }
}

#endif /* defined(__linux__) */

static int unix_start_timer(struct qemu_alarm_timer *t)
{
    struct sigaction act;
    struct itimerval itv;
    int err;

    /* timer signal */
    sigfillset(&act.sa_mask);
    act.sa_flags = 0;
    act.sa_handler = host_alarm_handler;

    sigaction(SIGALRM, &act, NULL);

    itv.it_interval.tv_sec = 0;
    /* for i386 kernel 2.6 to get 1 ms */
    itv.it_interval.tv_usec = 999;
    itv.it_value.tv_sec = 0;
    itv.it_value.tv_usec = 10 * 1000;

    err = setitimer(ITIMER_REAL, &itv, NULL);
    if (err)
        return -1;

    return 0;
}

static void unix_stop_timer(struct qemu_alarm_timer *t)
{
    struct itimerval itv;

    memset(&itv, 0, sizeof(itv));
    setitimer(ITIMER_REAL, &itv, NULL);
}

#endif /* !defined(_WIN32) */


#ifdef _WIN32

static int win32_start_timer(struct qemu_alarm_timer *t)
{
    TIMECAPS tc;
    struct qemu_alarm_win32 *data = t->priv;
    UINT flags;

    memset(&tc, 0, sizeof(tc));
    timeGetDevCaps(&tc, sizeof(tc));

    data->period = tc.wPeriodMin;
    timeBeginPeriod(data->period);

    flags = TIME_CALLBACK_FUNCTION;
    if (alarm_has_dynticks(t))
        flags |= TIME_ONESHOT;
    else
        flags |= TIME_PERIODIC;

    data->timerId = timeSetEvent(1,         // interval (ms)
                        data->period,       // resolution
                        host_alarm_handler, // function
                        (DWORD)t,           // parameter
                        flags);

    if (!data->timerId) {
        fprintf(stderr, "Failed to initialize win32 alarm timer: %ld\n",
                GetLastError());
        timeEndPeriod(data->period);
        return -1;
    }

    return 0;
}

static void win32_stop_timer(struct qemu_alarm_timer *t)
{
    struct qemu_alarm_win32 *data = t->priv;

    timeKillEvent(data->timerId);
    timeEndPeriod(data->period);
}

static void win32_rearm_timer(struct qemu_alarm_timer *t)
{
    struct qemu_alarm_win32 *data = t->priv;

    assert(alarm_has_dynticks(t));
    if (!active_timers[QEMU_CLOCK_REALTIME] &&
        !active_timers[QEMU_CLOCK_VIRTUAL] &&
        !active_timers[QEMU_CLOCK_HOST])
        return;

    timeKillEvent(data->timerId);

    data->timerId = timeSetEvent(1,
                        data->period,
                        host_alarm_handler,
                        (DWORD)t,
                        TIME_ONESHOT | TIME_CALLBACK_FUNCTION);

    if (!data->timerId) {
        fprintf(stderr, "Failed to re-arm win32 alarm timer %ld\n",
                GetLastError());

        timeEndPeriod(data->period);
        exit(1);
    }
}

#endif /* _WIN32 */

static void alarm_timer_on_change_state_rearm(void *opaque, int running, int reason)
{
    if (running)
        qemu_rearm_alarm_timer((struct qemu_alarm_timer *) opaque);
}

int init_timer_alarm(void)
{
    struct qemu_alarm_timer *t = NULL;
    int i, err = -1;

    for (i = 0; alarm_timers[i].name; i++) {
        t = &alarm_timers[i];

        err = t->start(t);
        if (!err)
            break;
    }

    if (err) {
        err = -ENOENT;
        goto fail;
    }

    /* first event is at time 0 */
    t->pending = 1;
    alarm_timer = t;
    qemu_add_vm_change_state_handler(alarm_timer_on_change_state_rearm, t);

    return 0;

fail:
    return err;
}

void quit_timers(void)
{
    struct qemu_alarm_timer *t = alarm_timer;
    alarm_timer = NULL;
    t->stop(t);
}

extern int tcg_has_work(void);

int qemu_calculate_timeout(void)
{
#ifndef CONFIG_IOTHREAD
    int timeout;

    if (!vm_running)
        timeout = 5000;
    else if (tcg_has_work())
        timeout = 0;
    else if (!use_icount) {
#ifdef WIN32
        /* This corresponds to the case where the emulated system is
         * totally idle and waiting for i/o. The problem is that on
         * Windows, the default value will prevent Windows user events
         * to be delivered in less than 5 seconds.
         *
         * Upstream contains a different way to handle this, for now
         * this hack should be sufficient until we integrate it into
         * our tree.
         */
        timeout = 1000/15;  /* deliver user events every 15/th of second */
#else
        timeout = 5000;
#endif
    } else {
     /* XXX: use timeout computed from timers */
        int64_t add;
        int64_t delta;
        /* Advance virtual time to the next event.  */
	delta = qemu_icount_delta();
        if (delta > 0) {
            /* If virtual time is ahead of real time then just
               wait for IO.  */
            timeout = (delta + 999999) / 1000000;
        } else {
            /* Wait for either IO to occur or the next
               timer event.  */
            add = qemu_next_deadline();
            /* We advance the timer before checking for IO.
               Limit the amount we advance so that early IO
               activity won't get the guest too far ahead.  */
            if (add > 10000000)
                add = 10000000;
            delta += add;
            qemu_icount += qemu_icount_round (add);
            timeout = delta / 1000000;
            if (timeout < 0)
                timeout = 0;
        }
    }

    return timeout;
#else /* CONFIG_IOTHREAD */
    return 1000;
#endif
}

/* Return the virtual CPU time, based on the instruction counter.  */
int64_t cpu_get_icount(void)
{
    int64_t icount;
    CPUState *env = cpu_single_env;;

    icount = qemu_icount;
    if (env) {
        if (!can_do_io(env)) {
            fprintf(stderr, "Bad clock read\n");
        }
        icount -= (env->icount_decr.u16.low + env->icount_extra);
    }
    return qemu_icount_bias + (icount << icount_time_shift);
}