守護進程：看門狗watchdog的添加

本文爲項目開發總結的原創文檔。

 

本項目，添加一個watchdog守護進程，用來監控環境的三大進程mozart、bitbox、mplayer，任何一個進程出現故障，整個環境進行重啓。

 

首先有經過版級驅動/arch/mips/xburst/soc-x1000/common# vim reset.c 

找到與看門狗有關的code[同事發現，牛！]；所以主要是將核心代碼從內核空間搬移到用戶空間，及如何監控應用層的進程。

 

總體實現思路：

1.建立一個進程做爲守護進程：watchdog

  

進程的添加：

在configs下添加watchdog.mak;

在src下添加watchdog包，用於加入watchdog相關的code；Makefile;

 

進程的啓動：在app.c中的startall中調用mozart_system("watchdog -b");

 

2.調整進程優先級：

1）如何查看進程的優先級

2）如何修改進程的優先級；

 

3.watchdog守護進程如何監控mozart和bitbox和mplayer

咱們知道內核會經過/proc虛擬文件系統導出系統中正在運行的進程信息，每一個進程都有一個/proc/<pid>目錄。所以咱們能夠將檢測進程是否存在轉換爲檢測/proc/<pid>目錄是否存在，這樣就簡單多了。

以下文詳細代碼中的processExists；

 

實現難點突破：

用戶空間和內核空間操做的都是虛擬地址。

1）若是是拿到的是物理地址，用戶空間能夠經過mmap的方式將物理地址轉成虛擬地址（每一次的地址值都不同），能夠直接對這個虛擬地址賦值。

以下：

static int dev_fd;

    dev_fd = open("/dev/mem", O_RDWR | O_NDELAY);

 

    if (dev_fd < 0) {

        printf("open(/dev/mem) failed.");

        return 0;

    }

 

    unsigned char *map_base=(unsigned char * ) mmap(NULL, MAP_SIZE, PROT_READ | PROT_WRITE, MAP_SHARED, dev_fd,  WDT_IOBASE （這個地址是物理地址！）);

 

*( unsigned long*)(map_base + TCU_TSCR) = (1 << 16);

 

close(dev_fd);

 

2）而內核空間，從物理地址轉成虛擬地址，通常是固定的。0x1000200---0xb000200；

 

3）同一個物理地址轉成虛擬地址，用戶空間和內核空間是不相同的。

關於用戶空間和內核空間：

http://blog.csdn.net/f22jay/article/details/7925531 

物理地址在內核空間和用戶空間映射地址不同~~~

 

看門狗實際上就是一個定時器，其硬件內部維護了一個計數的寄存器。每當時鐘信號到來時，計數寄存器減掉1,。若是減到0，則重啓系統。

若是減到0以前，系統又設置計數寄存器到一個較大的值，則系統永遠不會重啓。 

 

watchdog的基本實現原理是：

用戶空間程序打開 /dev/mem設備（俗稱「開門放狗」），

就會致使在內核中啓動一個定時器（本項目mdt_start_count的入參是20000ms即20s），此後，用戶空間程序須要保證在20分鐘以內向這個設備寫入數據（俗稱「按期喂狗」），每次寫操做會致使從新設定定時器（本項目是每sleep 10s從新去設定）。若是用戶空間程序在20分鐘以內沒有寫操做，定時器到期會致使一次系統Reboot操做（「狗咬人了」）。 

 

watchdog.c 內容以下：

#include <string.h>
#include <stdlib.h>
#include <signal.h>
#include <stdbool.h>
#include <unistd.h>
#include <pthread.h>
#include <errno.h>
#include <time.h>
#include <sys/types.h>
#include <pwd.h>
#include <sys/stat.h>
#include <linux/input.h>
#include <fcntl.h>
#include <execinfo.h>
#include <sys/mman.h>

#define WDT_IOBASE (0x10002000)
#define MAP_SIZE        0xFF



#define JZ_EXTAL_RTC      32768     /* RTC extal freq: 32.768 KHz */
#define TCU_IOBASE      0x10002000
#define TCU_TSCR   (0x3C)   /* Timer Stop Clear Register */

#define WDT_TCSR                (0x0c)  /* rw, 32, 0x???????? */
#define WDT_TCER                (0x04)  /* rw, 32, 0x???????? */
#define WDT_TDR                 (0x00)  /* rw, 32, 0x???????? */
#define WDT_TCNT                (0x08)  /* rw, 32, 0x???????? */
#define TCU_TSSR   (0x2C)   /* Timer Stop Set Register */

static void wdt_start_count(int msecs)
{
    static int dev_fd;
    dev_fd = open("/dev/mem", O_RDWR | O_NDELAY);

    if (dev_fd < 0) {
        printf("open(/dev/mem) failed.");
        return 0;
    }

    unsigned char *map_base=(unsigned char * )mmap(NULL, MAP_SIZE, PROT_READ | PROT_WRITE, MAP_SHARED, dev_fd, WDT_IOBASE );

        int time = JZ_EXTAL_RTC / 64 * msecs / 1000;
        if(time > 65535)
                time = 65535;

#if 0
        outl(1 << 16,TCU_IOBASE + TCU_TSCR);

        outl(0,WDT_IOBASE + WDT_TCNT);          //counter
        outl(time,WDT_IOBASE + WDT_TDR);        //data
        outl((3<<3 | 1<<1),WDT_IOBASE + WDT_TCSR);
        outl(0,WDT_IOBASE + WDT_TCER);
        outl(1,WDT_IOBASE + WDT_TCER);
#endif


/*上文屏蔽部分是內核空間對寄存器的操做，修改爲用戶空間對寄存器的操做，關鍵是物理地址在用戶空間須要經過mmap進行轉換*/


//    printf("wdt_start_count  begin~~~. map_base = %p,time=%d\n",map_base,time);
    *( unsigned long*)(map_base + TCU_TSCR) = (1 << 16);
    *( unsigned long*)(map_base + WDT_TCNT) = 0;//counter
    *( unsigned long*)(map_base + WDT_TDR) = time;//data
    *( unsigned long*)(map_base + WDT_TCSR) = (3<<3 | 1<<1);
    *( unsigned long*)(map_base + WDT_TCER) = 0;
    *( unsigned long*)(map_base + WDT_TCER) = 1;

    close(dev_fd);
//    printf("wdt_start_count  end.\n");

}

static void wdt_stop_count(void)
{
    static int dev_fd;
    dev_fd = open("/dev/mem", O_RDWR | O_NDELAY);

    if (dev_fd < 0) {
        printf("open(/dev/mem) failed.");
        return 0;
    }

    unsigned char *map_base=(unsigned char * )mmap(NULL, MAP_SIZE, PROT_READ | PROT_WRITE, MAP_SHARED, dev_fd, WDT_IOBASE );
    printf("\033[1;33mwdt_stop_count  begin~~~. map_base = %p\n\033[m",map_base);
#if 0
        outl(1 << 16,TCU_IOBASE + TCU_TSCR);
        outl(0,WDT_IOBASE + WDT_TCNT);          //counter
        outl(65535,WDT_IOBASE + WDT_TDR);       //data
        outl(1 << 16,TCU_IOBASE + TCU_TSSR);
#endif

    *( unsigned long*)(map_base + TCU_TSCR) = (1 << 16);
    *( unsigned long*)(map_base + WDT_TCNT) = 0;//counter
    *( unsigned long*)(map_base + WDT_TDR) = 65535;
    *( unsigned long*)(map_base + TCU_TSSR) = (1 << 16);
    close(dev_fd);
    printf("wdt_stop_count  end.\n");
}


void jz_wdt_restart()
{
        printf("Restarting after 4 ms\n");

     while(1)
     {
        wdt_start_count(20000);     
        sleep(10);
     }
     
        while(1)
                printf("check wdt.\n");
}
 /*判 斷 進 程 是 否 存 在*/ bool processExists(char * process_name) {  
        FILE *ptr;  
        int RE_BUF_SIZE = 32;  
        char rebuff[RE_BUF_SIZE];  
        char ps[128];  
        snprintf(ps, sizeof(ps), "ps | grep %s |grep -v grep| wc -l", process_name);  
        if((ptr = popen(ps, "r")) != NULL) {  
            int count = 0;  
            fgets(rebuff, RE_BUF_SIZE, ptr);  
            if(rebuff != NULL) {  
                count = atoi(rebuff);  
            }  
            pclose(ptr);  
            return count >= 1;  
        }  

     printf("Current process %s is not Exist!!!!\n",process_name);
        return false;  
    }  



static char *signal_str[] = {
    [1] = "SIGHUP",       [2] = "SIGINT",       [3] = "SIGQUIT",      [4] = "SIGILL",      [5] = "SIGTRAP",
    [6] = "SIGABRT",      [7] = "SIGBUS",       [8] = "SIGFPE",       [9] = "SIGKILL",     [10] = "SIGUSR1",
    [11] = "SIGSEGV",     [12] = "SIGUSR2",     [13] = "SIGPIPE",     [14] = "SIGALRM",    [15] = "SIGTERM",
    [16] = "SIGSTKFLT",   [17] = "SIGCHLD",     [18] = "SIGCONT",     [19] = "SIGSTOP",    [20] = "SIGTSTP",
    [21] = "SIGTTIN",     [22] = "SIGTTOU",     [23] = "SIGURG",      [24] = "SIGXCPU",    [25] = "SIGXFSZ",
    [26] = "SIGVTALRM",   [27] = "SIGPROF",     [28] = "SIGWINCH",    [29] = "SIGIO",      [30] = "SIGPWR",
    [31] = "SIGSYS",      [34] = "SIGRTMIN",    [35] = "SIGRTMIN+1",  [36] = "SIGRTMIN+2", [37] = "SIGRTMIN+3",
    [38] = "SIGRTMIN+4",  [39] = "SIGRTMIN+5",  [40] = "SIGRTMIN+6",  [41] = "SIGRTMIN+7", [42] = "SIGRTMIN+8",
    [43] = "SIGRTMIN+9",  [44] = "SIGRTMIN+10", [45] = "SIGRTMIN+11", [46] = "SIGRTMIN+12", [47] = "SIGRTMIN+13",
    [48] = "SIGRTMIN+14", [49] = "SIGRTMIN+15", [50] = "SIGRTMAX-14", [51] = "SIGRTMAX-13", [52] = "SIGRTMAX-12",
    [53] = "SIGRTMAX-11", [54] = "SIGRTMAX-10", [55] = "SIGRTMAX-9",  [56] = "SIGRTMAX-8", [57] = "SIGRTMAX-7",
    [58] = "SIGRTMAX-6",  [59] = "SIGRTMAX-5",  [60] = "SIGRTMAX-4",  [61] = "SIGRTMAX-3", [62] = "SIGRTMAX-2",
    [63] = "SIGRTMAX-1",  [64] = "SIGRTMAX",
};


static void usage(const char *app_name)
{
    printf("%s [-f file] -h\n"
           " -h help (show this usage text)\n"
           " -f file\n", app_name);

    return;
}


void sig_handler(int signo)
{
    char cmd[64] = {};
    void *array[10];
    int size = 0;
    char **strings = NULL;
    int i = 0;

#if 0
    printf("\n\n[%s: %d] bitbox crashed by signal %s.\n", __func__, __LINE__, signal_str[signo]);

    printf("Call Trace:\n");
    size = backtrace(array, 10);
    strings = backtrace_symbols(array, size);
    if (strings) {
        for (i = 0; i < size; i++)
            printf ("  %s\n", strings[i]);
        free (strings);
    } else {
        printf("Not Found\n\n");
    }

    if (signo == SIGSEGV || signo == SIGBUS ||
        signo == SIGTRAP || signo == SIGABRT) {
        sprintf(cmd, "cat /proc/%d/maps", getpid());
        printf("Process maps:\n");
        system(cmd);
    }
#else
    wdt_stop_count();
#endif

    exit(-1);
}

int main(int argc, char **argv)
{
    int c = -1;
    int daemonize = 0;
    printf("watchdog V1.7 start!!!!@_@\n");

    signal(SIGPIPE, SIG_IGN);
    signal(SIGINT, sig_handler);
    signal(SIGTERM, sig_handler);
    signal(SIGBUS, sig_handler);
    signal(SIGSEGV, sig_handler);
    signal(SIGABRT, sig_handler);
    
    while (1) {
        c = getopt(argc, argv, "bBf:h");
        if (c < 0)
            break;
        switch (c) {
            case 'b':
            case 'B':
                daemonize = 1;
                break;
            case 'f':
                break;
            case 'h':

                return 0;
            default:
                return -1;
        }
    }

    /* run in the background */
    if (daemonize) {
        if (daemon(0, 1)) {
            perror("daemon");
            return -1;
        }
    }

    while(1)
    {
        if(processExists("mozart")==true && processExists("bitbox")==true && processExists("mplayer")==true)
        {
//            printf("Both mozart and bitbox and mplayer are exists!!!!!\n");
            wdt_start_count(20000);     
                sleep(10);
        }
        else
        {
            printf(" mozart or  bitbox or mplayer is  not exist, Reboot!!!!! \n");
            printf("Mozart process  exist ???: %d\n",processExists("mozart"));
            printf("BitBox process  exist ???: %d\n",processExists("bitbox"));
            printf("Mplayer process  exist ???: %d\n",processExists("mplayer"));
            
            break;
        }
    }

}

 

 

程序運行起來後，經過ps可查看到：

S 0 241 1 7868 716 0:0 13:16 00:00:01 watchdog -b