这是什么分段故障原因是什么？ [英] What's the reason of this segmentation fault?

问题描述

这并不总是发生，但是当服务器应用程序已经运行了相当一段时间不会发生。

是什么原因，我该如何解决这个问题？

code如下：

 结构hostent *主机;
Result-＆GT; sin_family = AF_INET;
Result-＆GT; sin_port = htons（（无符号短）端口）;主机=的gethostbyname（主机名）;如果（！主机）
{
    unsigned long int类型地址= inet_addr（主机名）;
    如果（地址！=  -  1）
        HOST = gethostbyaddr（安培;地址，的sizeof（地址），AF_INET）;    如果（！主机）
    {
        如果（错误！= ETIMEDOUT）
            错误号= -1; / *使用h_errno * /
        的printf（未知主机服务器[％S]。，主机名）;
        返回（0）;
    }
}的memcpy（（字符*）及Result-＆GT; sin_addr，属于（的char *）＆主机的GT; h_addr，sizeof的（Result-＆GT; sin_addr））;
 

核心转储：

 ＃0 0x0000000000401913在proxy_getaddr在proxy.c（主机名= 0x7ae30be0stackoverflow.com，端口= 80，结果= 0x7ae30bd0）：529529的memcpy（（字符*）及Result-＆GT; sin_addr，属于（的char *）＆主机的GT; h_addr，sizeof的（Result-＆GT; sin_addr））;
（GDB）P *主机
$ 4 = {h_name = 0xc4ee048，h_aliases = 0xc4ee030，h_addrtype = 2，长度h_length = 4，h_addr_list = 0xc4ee038}
（GDB）打印结果
$ 5 =（结构SOCKADDR_IN *）0x7ae30bd0
（GDB）的打印结果*
$ 6 = {sin_family = 2，sin_port = 20480，sin_addr = {s_addr = 0} =的sin_zero\\ 000 \\ 000 \\ 000 \\ 000 \\ 000 \\ 000 \\ 000}（GDB）p主机的＆GT; h_addr_list [0]
$ 1 =为0x0
（GDB）p主机的＆GT; h_addr_list
$ 2 =（字符**）0x1bd9d050
 

解决方案

由于主机和结果变量都指向内存的合法块，最可能的原因是， HOST-＆GT; h_addr 是 NULL 。这将是情况下，如果通过返回的地址列表中的gethostbyname（）或 gethostbyaddr（）是空的。

我不知道怎么会引起（我的OS X系统上的文档意味着这两个函数应返回NULL，如果没有地址可以找到）。不过，我会检查主机的方式＆gt; h_addr_list [0] 在调试器中，以确认

修改

调试信息的更新显示问题出在哪里： HOST-＆GT; h_addr 为NULL。 h_addr其实就是这样的＃定义：

 的#define h_addr h_addr_list [0]
 

其中的一个功能返回一个结构hostent 与空地址列表。

It doesn't happen always,but will happen when the server application has been running for quite a while.

What's the reason and how can I fix it?

Code as follows:

struct hostent*     Host;
Result->sin_family=AF_INET;
Result->sin_port=htons((unsigned short)Port);

Host=gethostbyname(HostName);

if(!Host)
{
    unsigned long int addr=inet_addr(HostName);
    if(addr!=-1)
        Host=gethostbyaddr(&addr,sizeof(addr),AF_INET);

    if(!Host)
    {
        if(errno!=ETIMEDOUT)
            errno=-1; /* use h_errno */
        printf("Unknown host for server [%s].", HostName);
        return(0);
    }
}

memcpy((char*)&Result->sin_addr,(char*)Host->h_addr,sizeof(Result->sin_addr));

core dump:

#0  0x0000000000401913 in proxy_getaddr (HostName=0x7ae30be0 "stackoverflow.com", Port=80, Result=0x7ae30bd0) at proxy.c:529

529     memcpy((char*)&Result->sin_addr,(char*)Host->h_addr,sizeof(Result->sin_addr));
(gdb) p *Host
$4 = {h_name = 0xc4ee048 "", h_aliases = 0xc4ee030, h_addrtype = 2, h_length = 4, h_addr_list = 0xc4ee038}
(gdb) print Result 
$5 = (struct sockaddr_in *) 0x7ae30bd0
(gdb) print *Result 
$6 = {sin_family = 2, sin_port = 20480, sin_addr = {s_addr = 0}, sin_zero = "\000\000\000\000\000\000\000"}



(gdb) p Host->h_addr_list[0]
$1 = 0x0
(gdb) p Host->h_addr_list
$2 = (char **) 0x1bd9d050

解决方案

Given that the Host and Result variables both point to legitimate blocks of memory, the most likely cause is that Host->h_addr is NULL. This would be the case if the list of addresses returned by gethostbyname() or gethostbyaddr() were empty.

I don't know how that could be caused (the documentation on my OS X system implies that both functions should return NULL if no addresses can be found). However, I would check Host->h_addr_list[0] in the debugger to confirm.

Edit

The update of the debug info shows where the problem is: Host->h_addr is NULL. h_addr is actually a #define like this:

#define h_addr h_addr_list[0]

One of the functions is returning a struct hostent with an empty address list.

这篇关于这是什么分段故障原因是什么？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！