sem_timedwait的用法

       #include <semaphore.h>

       int sem_timedwait(sem_t *sem, const struct timespec *abs_timeout);

       Link with -pthread.

　　对于这个函数，主要在于abs_timeout这个参数。一开始我以为是传入需要等待的时间。像这样：

struct timespec ts;

ts.tv_nsec = ;

ts.tv_sec   = ;

sem_timedwait(p_sem, &ts);

意思是我希望10秒1000纳秒才超时。结果，函数立即返回。网上查一下资料，才知道我错得多么离谱。这个abs_timeout竟然是UTC时间戳。看下面的代码http://linux.die.net/man/3/sem_timedwait：

#include <unistd.h>

#include <stdio.h>

#include <stdlib.h>

#include <semaphore.h>

#include <time.h>

#include <assert.h>

#include <errno.h>

#include <signal.h>

sem_t sem;

#define handle_error(msg) \

    do { perror(msg); exit(EXIT_FAILURE); } while ()

static void

handler(int sig)

{

    write(STDOUT_FILENO, "sem_post() from handler\n", );

    if (sem_post(&sem) == -) {

        write(STDERR_FILENO, "sem_post() failed\n", );

        _exit(EXIT_FAILURE);

    }

}

int

main(int argc, char *argv[])

{

    struct sigaction sa;

    struct timespec ts;

    int s;

   if (argc != ) {

        fprintf(stderr, "Usage: %s <alarm-secs> <wait-secs>\n",

                argv[]);

        exit(EXIT_FAILURE);

    }

   if (sem_init(&sem, , ) == -)

        handle_error("sem_init");

   /* Establish SIGALRM handler; set alarm timer using argv[1] */

   sa.sa_handler = handler;

    sigemptyset(&sa.sa_mask);

    sa.sa_flags = ;

    if (sigaction(SIGALRM, &sa, NULL) == -)

        handle_error("sigaction");

   alarm(atoi(argv[]));

   /* Calculate relative interval as current time plus

       number of seconds given argv[2] */

   if (clock_gettime(CLOCK_REALTIME, &ts) == -)

        handle_error("clock_gettime");

   ts.tv_sec += atoi(argv[]);

   printf("main() about to call sem_timedwait()\n");

    while ((s = sem_timedwait(&sem, &ts)) == - && errno == EINTR)

        continue;       /* Restart if interrupted by handler */

   /* Check what happened */

   if (s == -) {

        if (errno == ETIMEDOUT)

            printf("sem_timedwait() timed out\n");

        else

            perror("sem_timedwait");

    } else

        printf("sem_timedwait() succeeded\n");

   exit((s == ) ? EXIT_SUCCESS : EXIT_FAILURE);

}

在这段代码中，他没有处理溢出，下面是我的代码：

int32 CSeamphoreLock::time_lock( int32 nano_sec,int32 sec )

{

    struct timespec ts;

    if ( clock_gettime( CLOCK_REALTIME,&ts ) <  )

        return -;

    ts.tv_sec  += sec;

    ts.tv_nsec += nano_sec;

    //#define NSECTOSEC    1000000000

    ts.tv_sec += ts.tv_nsec/NSECTOSEC; //Nanoseconds [0 .. 999999999]

    ts.tv_nsec = ts.tv_nsec%NSECTOSEC;

    return sem_timedwait( m_psem,&ts );

}

　　PS：居然用的时间戳，如果正在等待的时候管理员调整时间会不会让某个程序出问题呢？？为什么不用clock_gettime的CLOCK_MONOTONIC来判断呢。

巴特西

sem_timedwait的用法

最新文章

热门文章