sched.c

Linux 1.0 内核C源代码 Linux最早版本代码 由Linus Torvalds亲自书写的

			break;
		}
		timer->expires -= (*p)->expires;
		p = &(*p)->next;
	}
	*p = timer;
	restore_flags(flags);
}

int del_timer(struct timer_list * timer)
{
	unsigned long flags;
	unsigned long expires = 0;
	struct timer_list **p;

	p = &next_timer;
	save_flags(flags);
	cli();
	while (*p) {
		if (*p == timer) {
			if ((*p = timer->next) != NULL)
				(*p)->expires += timer->expires;
			timer->expires += expires;
			restore_flags(flags);
			return 1;
		}
		expires += (*p)->expires;
		p = &(*p)->next;
	}
	restore_flags(flags);
	return 0;
}

unsigned long timer_active = 0;
struct timer_struct timer_table[32];

/*
 * Hmm.. Changed this, as the GNU make sources (load.c) seems to
 * imply that avenrun[] is the standard name for this kind of thing.
 * Nothing else seems to be standardized: the fractional size etc
 * all seem to differ on different machines.
 */
unsigned long avenrun[3] = { 0,0,0 };

/*
 * Nr of active tasks - counted in fixed-point numbers
 */
static unsigned long count_active_tasks(void)
{
	struct task_struct **p;
	unsigned long nr = 0;

	for(p = &LAST_TASK; p > &FIRST_TASK; --p)
		if (*p && ((*p)->state == TASK_RUNNING ||
			   (*p)->state == TASK_UNINTERRUPTIBLE ||
			   (*p)->state == TASK_SWAPPING))
			nr += FIXED_1;
	return nr;
}

static inline void calc_load(void)
{
	unsigned long active_tasks; /* fixed-point */
	static int count = LOAD_FREQ;

	if (count-- > 0)
		return;
	count = LOAD_FREQ;
	active_tasks = count_active_tasks();
	CALC_LOAD(avenrun[0], EXP_1, active_tasks);
	CALC_LOAD(avenrun[1], EXP_5, active_tasks);
	CALC_LOAD(avenrun[2], EXP_15, active_tasks);
}

/*
 * this routine handles the overflow of the microsecond field
 *
 * The tricky bits of code to handle the accurate clock support
 * were provided by Dave Mills ([e-mail]) of NTP fame.
 * They were originally developed for SUN and DEC kernels.
 * All the kudos should go to Dave for this stuff.
 *
 * These were ported to Linux by Philip Gladstone.
 */
static void second_overflow(void)
{
	long ltemp;
	/* last time the cmos clock got updated */
	static long last_rtc_update=0;
	extern int set_rtc_mmss(unsigned long);

	/* Bump the maxerror field */
	time_maxerror = (0x70000000-time_maxerror < time_tolerance) ?
	  0x70000000 : (time_maxerror + time_tolerance);

	/* Run the PLL */
	if (time_offset < 0) {
		ltemp = (-(time_offset+1) >> (SHIFT_KG + time_constant)) + 1;
		time_adj = ltemp << (SHIFT_SCALE - SHIFT_HZ - SHIFT_UPDATE);
		time_offset += (time_adj * HZ) >> (SHIFT_SCALE - SHIFT_UPDATE);
		time_adj = - time_adj;
	} else if (time_offset > 0) {
		ltemp = ((time_offset-1) >> (SHIFT_KG + time_constant)) + 1;
		time_adj = ltemp << (SHIFT_SCALE - SHIFT_HZ - SHIFT_UPDATE);
		time_offset -= (time_adj * HZ) >> (SHIFT_SCALE - SHIFT_UPDATE);
	} else {
		time_adj = 0;
	}

	time_adj += (time_freq >> (SHIFT_KF + SHIFT_HZ - SHIFT_SCALE))
	    + FINETUNE;

	/* Handle the leap second stuff */
	switch (time_status) {
		case TIME_INS:
		/* ugly divide should be replaced */
		if (xtime.tv_sec % 86400 == 0) {
			xtime.tv_sec--; /* !! */
			time_status = TIME_OOP;
			printk("Clock: inserting leap second 23:59:60 GMT\n");
		}
		break;

		case TIME_DEL:
		/* ugly divide should be replaced */
		if (xtime.tv_sec % 86400 == 86399) {
			xtime.tv_sec++;
			time_status = TIME_OK;
			printk("Clock: deleting leap second 23:59:59 GMT\n");
		}
		break;

		case TIME_OOP:
		time_status = TIME_OK;
		break;
	}
	if (xtime.tv_sec > last_rtc_update + 660)
	  if (set_rtc_mmss(xtime.tv_sec) == 0)
	    last_rtc_update = xtime.tv_sec;
}

/*
 * disregard lost ticks for now.. We don't care enough.
 */
static void timer_bh(void * unused)
{
	unsigned long mask;
	struct timer_struct *tp;

	cli();
	while (next_timer && next_timer->expires == 0) {
		void (*fn)(unsigned long) = next_timer->function;
		unsigned long data = next_timer->data;
		next_timer = next_timer->next;
		sti();
		fn(data);
		cli();
	}
	sti();
	
	for (mask = 1, tp = timer_table+0 ; mask ; tp++,mask += mask) {
		if (mask > timer_active)
			break;
		if (!(mask & timer_active))
			continue;
		if (tp->expires > jiffies)
			continue;
		timer_active &= ~mask;
		tp->fn();
		sti();
	}
}

/*
 * The int argument is really a (struct pt_regs *), in case the
 * interrupt wants to know from where it was called. The timer
 * irq uses this to decide if it should update the user or system
 * times.
 */
static void do_timer(struct pt_regs * regs)
{
	unsigned long mask;
	struct timer_struct *tp;

	long ltemp;

	/* Advance the phase, once it gets to one microsecond, then
	 * advance the tick more.
	 */
	time_phase += time_adj;
	if (time_phase < -FINEUSEC) {
		ltemp = -time_phase >> SHIFT_SCALE;
		time_phase += ltemp << SHIFT_SCALE;
		xtime.tv_usec += tick + time_adjust_step - ltemp;
	}
	else if (time_phase > FINEUSEC) {
		ltemp = time_phase >> SHIFT_SCALE;
		time_phase -= ltemp << SHIFT_SCALE;
		xtime.tv_usec += tick + time_adjust_step + ltemp;
	} else
		xtime.tv_usec += tick + time_adjust_step;

	if (time_adjust)
	{
	    /* We are doing an adjtime thing. 
	     *
	     * Modify the value of the tick for next time.
	     * Note that a positive delta means we want the clock
	     * to run fast. This means that the tick should be bigger
	     *
	     * Limit the amount of the step for *next* tick to be
	     * in the range -tickadj .. +tickadj
	     */
	     if (time_adjust > tickadj)
	       time_adjust_step = tickadj;
	     else if (time_adjust < -tickadj)
	       time_adjust_step = -tickadj;
	     else
	       time_adjust_step = time_adjust;
	     
	    /* Reduce by this step the amount of time left  */
	    time_adjust -= time_adjust_step;
	}
	else
	    time_adjust_step = 0;

	if (xtime.tv_usec >= 1000000) {
	    xtime.tv_usec -= 1000000;
	    xtime.tv_sec++;
	    second_overflow();
	}

	jiffies++;
	calc_load();
	if ((VM_MASK & regs->eflags) || (3 & regs->cs)) {
		current->utime++;
		if (current != task[0]) {
			if (current->priority < 15)
				kstat.cpu_nice++;
			else
				kstat.cpu_user++;
		}
		/* Update ITIMER_VIRT for current task if not in a system call */
		if (current->it_virt_value && !(--current->it_virt_value)) {
			current->it_virt_value = current->it_virt_incr;
			send_sig(SIGVTALRM,current,1);
		}
	} else {
		current->stime++;
		if(current != task[0])
			kstat.cpu_system++;
#ifdef CONFIG_PROFILE
		if (prof_buffer && current != task[0]) {
			unsigned long eip = regs->eip;
			eip >>= 2;
			if (eip < prof_len)
				prof_buffer[eip]++;
		}
#endif
	}
	if (current == task[0] || (--current->counter)<=0) {
		current->counter=0;
		need_resched = 1;
	}
	/* Update ITIMER_PROF for the current task */
	if (current->it_prof_value && !(--current->it_prof_value)) {
		current->it_prof_value = current->it_prof_incr;
		send_sig(SIGPROF,current,1);
	}
	for (mask = 1, tp = timer_table+0 ; mask ; tp++,mask += mask) {
		if (mask > timer_active)
			break;
		if (!(mask & timer_active))
			continue;
		if (tp->expires > jiffies)
			continue;
		mark_bh(TIMER_BH);
	}
	cli();
	itimer_ticks++;
	if (itimer_ticks > itimer_next)
		need_resched = 1;
	if (next_timer) {
		if (next_timer->expires) {
			next_timer->expires--;
			if (!next_timer->expires)
				mark_bh(TIMER_BH);
		} else {
			lost_ticks++;
			mark_bh(TIMER_BH);
		}
	}
	sti();
}

asmlinkage int sys_alarm(long seconds)
{
	struct itimerval it_new, it_old;

	it_new.it_interval.tv_sec = it_new.it_interval.tv_usec = 0;
	it_new.it_value.tv_sec = seconds;
	it_new.it_value.tv_usec = 0;
	_setitimer(ITIMER_REAL, &it_new, &it_old);
	return(it_old.it_value.tv_sec + (it_old.it_value.tv_usec / 1000000));
}

asmlinkage int sys_getpid(void)
{
	return current->pid;
}

asmlinkage int sys_getppid(void)
{
	return current->p_opptr->pid;
}

asmlinkage int sys_getuid(void)
{
	return current->uid;
}

asmlinkage int sys_geteuid(void)
{
	return current->euid;
}

asmlinkage int sys_getgid(void)
{
	return current->gid;
}

asmlinkage int sys_getegid(void)
{
	return current->egid;
}

asmlinkage int sys_nice(long increment)
{
	int newprio;

	if (increment < 0 && !suser())
		return -EPERM;
	newprio = current->priority - increment;
	if (newprio < 1)
		newprio = 1;
	if (newprio > 35)
		newprio = 35;
	current->priority = newprio;
	return 0;
}

static void show_task(int nr,struct task_struct * p)
{
	static char * stat_nam[] = { "R", "S", "D", "Z", "T", "W" };

	printk("%-8s %3d ", p->comm, (p == current) ? -nr : nr);
	if (((unsigned) p->state) < sizeof(stat_nam)/sizeof(char *))
		printk(stat_nam[p->state]);
	else
		printk(" ");
	if (p == current)
		printk(" current  ");
	else
		printk(" %08lX ", ((unsigned long *)p->tss.esp)[3]);
	printk("%5lu %5d %6d ",
		p->tss.esp - p->kernel_stack_page, p->pid, p->p_pptr->pid);
	if (p->p_cptr)
		printk("%5d ", p->p_cptr->pid);
	else
		printk("      ");
	if (p->p_ysptr)
		printk("%7d", p->p_ysptr->pid);
	else
		printk("       ");
	if (p->p_osptr)
		printk(" %5d\n", p->p_osptr->pid);
	else
		printk("\n");
}

void show_state(void)
{
	int i;

	printk("                         free                        sibling\n");
	printk("  task             PC    stack   pid father child younger older\n");
	for (i=0 ; i<NR_TASKS ; i++)
		if (task[i])
			show_task(i,task[i]);
}

void sched_init(void)
{
	int i;
	struct desc_struct * p;

	bh_base[TIMER_BH].routine = timer_bh;
	if (sizeof(struct sigaction) != 16)
		panic("Struct sigaction MUST be 16 bytes");
	set_tss_desc(gdt+FIRST_TSS_ENTRY,&init_task.tss);
	set_ldt_desc(gdt+FIRST_LDT_ENTRY,&default_ldt,1);
	set_system_gate(0x80,&system_call);
	p = gdt+2+FIRST_TSS_ENTRY;
	for(i=1 ; i<NR_TASKS ; i++) {
		task[i] = NULL;
		p->a=p->b=0;
		p++;
		p->a=p->b=0;
		p++;
	}
/* Clear NT, so that we won't have troubles with that later on */
	__asm__("pushfl ; andl $0xffffbfff,(%esp) ; popfl");
	load_TR(0);
	load_ldt(0);
	outb_p(0x34,0x43);		/* binary, mode 2, LSB/MSB, ch 0 */
	outb_p(LATCH & 0xff , 0x40);	/* LSB */
	outb(LATCH >> 8 , 0x40);	/* MSB */
	if (request_irq(TIMER_IRQ,(void (*)(int)) do_timer)!=0)
		panic("Could not allocate timer IRQ!");
}