destroying a vector of threads segmentation fault

Question

I warped up the p_thread inside a class Worker :

class Worker { 
public:
    Worker(int worker_num ,void* (*function)(void*),void * params);
    virtual ~Worker();

private:
    pthread_t thread;
    int worker_num;
};

Worker::Worker(int worker_num, void* (*function)(void*), void * params) {
    pthread_create(&thread, NULL, function, (void *) params);
    logger.log(DEBUG_LEVEL_DEBUG, "Worker id=%u was created\n", worker_num);
    this->worker_num = worker_num;
    return;
}

Worker::~Worker() {
    pthread_cancel(thread);
    logger.log(DEBUG_LEVEL_DEBUG, "Worker id=%u was canceled\n", worker_num);
    return;
}

than I created a vector of workers on the constructor of another class THREAD_POOL:

for (int i = 0; i < worker_num; i++) {
    this->workers.push_back(new Worker(i, thread_exe, ((void*) this)));
}

(the function thread_exe is an endless loop). When I am trying to destroy those workers on the destroctor of THREAD_POOL class like so :

ThreadPool::~ThreadPool() {

    for(int worker_index = 0 ; worker_index < workers.size();worker_index++){
        delete workers[worker_index];
    }

I my programs crashes:

Worker id=0 was canceled * glibc detected * /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect: double free or corruption (out): 0x0000000000629590 * ======= Backtrace: ========= /lib64/libc.so.6[0x3300476166] /lib64/libc.so.6[0x3300478c93] /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect[0x413ae2] /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect[0x412a04] /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect[0x412b70] /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect[0x40e527] /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect[0x4083fe] /lib64/libc.so.6(__libc_start_main+0xfd)[0x330041ed1d] /.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect[0x403a69] ======= Memory map: ======== 00400000-00426000 r-xp 00000000 00:21 19570593
/.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect 00625000-00626000 rw-p 00025000 00:21 19570593
/.autodirect/fwgwork/yehudas/workspace/dt_utils/code_coverage/Debug/fcollect 00626000-00648000 rw-p 00000000 00:00 0
[heap] 32ffc00000-32ffc20000 r-xp 00000000 fc:01 670751
/lib64/ld-2.12.so 32ffe1f000-32ffe20000 r--p 0001f000 fc:01 670751
/lib64/ld-2.12.so 32ffe20000-32ffe21000 rw-p 00020000 fc:01 670751
/lib64/ld-2.12.so 32ffe21000-32ffe22000 rw-p 00000000 00:00 0 3300000000-3300083000 r-xp 00000000 fc:01 670759
/lib64/libm-2.12.so 3300083000-3300282000 ---p 00083000 fc:01 670759
/lib64/libm-2.12.so 3300282000-3300283000 r--p 00082000 fc:01 670759
/lib64/libm-2.12.so 3300283000-3300284000 rw-p 00083000 fc:01 670759
/lib64/libm-2.12.so 3300400000-330058b000 r-xp 00000000 fc:01 670752
/lib64/libc-2.12.so 330058b000-330078a000 ---p 0018b000 fc:01 670752
/lib64/libc-2.12.so 330078a000-330078e000 r--p 0018a000 fc:01 670752
/lib64/libc-2.12.so 330078e000-330078f000 rw-p 0018e000 fc:01 670752
/lib64/libc-2.12.so 330078f000-3300794000 rw-p 00000000 00:00 0 3300800000-3300802000 r-xp 00000000 fc:01 670755
/lib64/libdl-2.12.so 3300802000-3300a02000 ---p 00002000 fc:01 670755 /lib64/libdl-2.12.so 3300a02000-3300a03000 r--p 00002000 fc:01 670755 /lib64/libdl-2.12.so 3300a03000-3300a04000 rw-p 00003000 fc:01 670755 /lib64/libdl-2.12.so 3300c00000-3300c17000 r-xp 00000000 fc:01 670753 /lib64/libpthread-2.12.so 3300c17000-3300e17000 ---p 00017000 fc:01 670753 /lib64/libpthread-2.12.so 3300e17000-3300e18000 r--p 00017000 fc:01 670753
/lib64/libpthread-2.12.so 3300e18000-3300e19000 rw-p 00018000 fc:01 670753 /lib64/libpthread-2.12.so 3300e19000-3300e1d000 rw-p 00000000 00:00 0 3303c00000-3303c16000 r-xp 00000000 fc:01 670760
/lib64/libgcc_s-4.4.7-20120601.so.1 3303c16000-3303e15000 ---p 00016000 fc:01 670760
/lib64/libgcc_s-4.4.7-20120601.so.1 3303e15000-3303e16000 rw-p 00015000 fc:01 670760
/lib64/libgcc_s-4.4.7-20120601.so.1 3304800000-33048e8000 r-xp 00000000 fc:01 918134
/usr/lib64/libstdc++.so.6.0.13 33048e8000-3304ae8000 ---p 000e8000 fc:01 918134 /usr/lib64/libstdc++.so.6.0.13 3304ae8000-3304aef000 r--p 000e8000 fc:01 918134
/usr/lib64/libstdc++.so.6.0.13 3304aef000-3304af1000 rw-p 000ef000 fc:01 918134 /usr/lib64/libstdc++.so.6.0.13 3304af1000-3304b06000 rw-p 00000000 00:00 0 7fffe8000000-7fffe8021000 rw-p 00000000 00:00 0 7fffe8021000-7fffec000000 ---p 00000000 00:00 0 7ffff0000000-7ffff0021000 rw-p 00000000 00:00 0 7ffff0021000-7ffff4000000 ---p 00000000 00:00 0 7ffff6bdb000-7ffff6bdc000 ---p 00000000 00:00 0 7ffff6bdc000-7ffff75dc000 rw-p 00000000 00:00 0 7ffff75dc000-7ffff75dd000 ---p 00000000 00:00 0 7ffff75dd000-7ffff7fe3000 rw-p 00000000 00:00 0 7ffff7ffa000-7ffff7ffe000 rw-p 00000000 00:00 0 7ffff7ffe000-7ffff7fff000 r-xp 00000000 00:00 0
[vdso] 7ffffffe9000-7ffffffff000 rw-p 00000000 00:00 0
[stack] ffffffffff600000-ffffffffff601000 r-xp 00000000 00:00 0
[vsyscall]

need advise please...

I think you still have to pthread_join() a cancelled thread. But why not using std::thread?? — πάντα ῥεῖ
Google 'rule of three' and block your copy assign and construct. — Yakk - Adam Nevraumont
I am using p_thread cause I want it to be compatible to posix standarts — yehudahs
Yakk - implemented all the copy constructors and assign with no luck - still getting the error. — yehudahs

AlexT AlexT · Accepted Answer · 2014-02-03T07:52:28

You forgot to increment cycle counter (++worker_index)

ThreadPool::~ThreadPool() {

for(int worker_index = 0 ; worker_index < workers.size(); ++worker_index){
    delete workers[worker_index];
}

destroying a vector of threads segmentation fault

1 Answers