我正在linux上写一些TCP服务器代码,它正与带有buildroot工具链的Pi交叉编译。基本上,我接受一个套接字,然后创建一个线程来侦听数据并对其进行处理。为了进行测试,我使用telnet作为客户端。有时,套接字关闭时,线程清理期间程序崩溃,我无法弄清原因。我最好的猜测是文件描述符怪异,但这在目前主要是猜测。
代码如下:
#include <algorithm>
#include <iostream>
#include <fstream>
#include <string>
#include <iterator>
#include <map>
#include <set>
using namespace std;
int main(){
set<string> exclusionSet;
ifstream stopWords("stopwords.txt");
copy(istream_iterator<string>(stopWords),
istream_iterator<string>(),
inserter(exclusionSet, exclusionSet.end()));
stopWords.close();
map<string, int> wordMap;
ifstream words("sample_doc.txt");
string s;
while(words >> s) if (exclusionSet.count(s)==0) ++wordMap[s];
ofstream out("frequency.txt");
for_each(begin(wordMap),
end(wordMap),
[&out](const pair<const string, int>& e){ out<<e.first<<" "<<e.second<<endl;});
out.close();
return 0;
}
这是崩溃的回溯:
void NetworkControl::startListening()
{
mListenThread = std::thread(listener);
mListenThread.detach();
}
void NetworkControl::listener()
{
int sockfd, newsockfd, clilen;
struct sockaddr_in serv_addr, cli_addr;
int portno = 612;
std::thread mWorkThread;
sockfd = socket(AF_INET, SOCK_STREAM, 0);
if (sockfd < 0)
{} //print error
bzero((char *) &serv_addr, sizeof(serv_addr));
serv_addr.sin_family = AF_INET;
serv_addr.sin_addr.s_addr = INADDR_ANY;
serv_addr.sin_port = htons(portno);
if (bind(sockfd, (struct sockaddr *) &serv_addr,sizeof(serv_addr)) < 0)
{ }//print error
listen(sockfd,5);
clilen = sizeof(cli_addr);
while (1) {
newsockfd = accept(sockfd,
(struct sockaddr *) &cli_addr, (socklen_t*) &clilen);
if (newsockfd < 0) {
//print error
}else{
mWorkThread = std::thread(worker,newsockfd);
mWorkThread.detach();
}
}
}
void NetworkControl::worker(int sock)
{
uint32_t n;
char *buf = new char(200);
while(1){
n = read(sock,&buf[0],199);;
if (n <= 0){
goto exit;
}
buf[n] = '\x00';
printf("%s\n",buf);
}
exit:
free(buf);
close(sock);
}
控制台输出:
#0 0x767c4e0c in raise () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#1 0x767c60e4 in abort () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#2 0x767fe2c8 in __libc_message () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#3 0x76804790 in malloc_printerr () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#4 0x76805850 in _int_free () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#5 0x768a3c78 in tcache_thread_freeres () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#6 0x768a3cdc in __libc_thread_freeres () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
#7 0x76f01f3c in start_thread () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libpthread.so.0
#8 0x768683b8 in ?? () from /home/david/Desktop/BR-Pi2/output/host/arm-buildroot-linux-gnueabihf/sysroot/lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
如果有人有任何想法,我将不胜感激!
答案 0 :(得分:1)
正如已经告诉别人的那样,当您实际上需要new [] / delete []时,将new()与free()配对会导致崩溃。
但是,使用std::vector
,您甚至不必自己进行内存分配。 (示例代码未编译/未测试)
void NetworkControl::worker(int sock)
{
uint32_t n;
std::vector<char> buf(200);
while(1){
n = read(sock, buf.data(), 199);
if (n <= 0){
break;
}
buf[n] = '\x00';
printf("%s\n",buf.data());
}
close(sock);
}
对于buf
,您还可以使用C样式的数组或std::array<char>
,但随后必须确保没有堆栈溢出。
答案 1 :(得分:1)
free()函数释放ptr指向的内存空间,该空间必须由先前对malloc(),calloc()或realloc()的调用返回。其他- 明智的选择,或者如果以前已经调用过free(ptr),则会发生未定义的行为。如果ptr为NULL,则不执行任何操作。
但在您的代码上下文中:
字符* buf =新字符(200);
指针是由新函数获取的,它是C ++函数,这就是程序崩溃的原因。