Linux下system_uptime函数

[ReMeppo]

看这个函数有点头晕, 哪位热心的大哥大姐帮解释解释... 感激不尽

#define WRAPAROUND_VALUE (0xffffffffUL / HZ + 1) /* HZ = frequency of ticks
                                                    per second. */

/* Static function to get current sec and usec.  */
static int
system_uptime (struct pal_timeval *tv, struct pal_tzval *tz)
{
  struct sysinfo info;
  static unsigned long prev = 0;
  static unsigned long wraparound_count = 0;
  unsigned long uptime;
  static long base = 0;
  static long offset = 0;
  long leap;
  long diff;
  
  /* Get sysinfo.  */
  if (sysinfo (&info) < 0)   //得到系统信息包括启动时间
    return RESULT_ERROR;

  /* Check for wraparound. */
  if (prev > info.uptime)
    wraparound_count++;	//这个有什么用? 这个count是代表什么, 这里表达意思是什么? 怎么样count才++
  
  /* System uptime.  */
  uptime = wraparound_count * WRAPAROUND_VALUE + info.uptime;  //这里算的是什么? 
  prev = info.uptime;      
  
  /* Get tv_sec and tv_usec.  */
  gettimeofday (tv, tz);

//下面一连串的调整 都不知道是什么意思了. 几个静态变量和临时变量混和在一起... 
  /* Deffernce between gettimeofday sec and uptime.  */
  leap = tv->tv_sec - uptime;

  /* Remember base diff for adjustment.  */
  if (! base)
    base = leap;

  /* Basically we use gettimeofday's return value because it is the
     only way to get required granularity.  But when diff is very
     different we adjust the value using base value.  */
  diff = (leap - base) + offset;

  /* When system time go forward than 2 sec.  */
  if (diff > 2 || diff < -2)
    offset -= diff;

  /* Adjust second.  */
  tv->tv_sec += offset;

  return RESULT_OK;
}

[Liv2005]

先提出具体问题，再要求解答吧....

不然就好像在说，数学咱不懂，谁给讲下？

[ReMeppo]

[Quote=引用 1 楼 liv2005 的回复:]

先提出具体问题，再要求解答吧....

不然就好像在说，数学咱不懂，谁给讲下？
[/Quote] 问题不在代码里的注释里面了么...

[ReMeppo]

没人??? 自己顶个

[ReMeppo]

[帅得不敢出门]

WRAPAROUND_VALUE (0xffffffffUL / HZ + 1) 溢出一次需要的时间

 /* Check for wraparound. */
  if (prev > info.uptime)
    wraparound_count++;    //溢出一次 wraparound_count++ 
  
  /* System uptime.  */
  uptime = wraparound_count * WRAPAROUND_VALUE + info.uptime;  所有溢出时间相加+info.uptime就是总uptime
  prev = info.uptime;      

// 溢出的例子    比如总量为10 ,此时prev = 10  那么info.uptime为 1而不是11,此时prev > info.uptime
那么此时wraparound_count=1  ，然后再算出现在的uptime

下面的代码都有英文注释，我就不解释了。

[ReMeppo]