Hi
I have a server which creates a new detached pthread every time it
accepts a new connection.
Inside such a thread an object is created which modifies a std:vector.
At the end of the program, long after the connection has terminated,
and very probably after the thread has finished, said object is deleted.
Now helgrind makes messages such as
==18078== Possible data race during read of size 4 at 0x4330b98 by thread #1
==18078== at 0x804C9A4: std::_Deque_iterator<std::string,
std::string&, std::string*>::_Deque_iterator(std::_Deque_iterator<std::string,
std::string&, std::string*> const&) (stl_deque.h:131)
==18078== by 0x804D15B: std::deque<std::string,
std::allocator<std::string> >::~deque() (stl_deque.h:725)
==18078== by 0x804C64D: LineBuffer::~LineBuffer() (LineBuffer.cpp:21)
==18078== by 0x8049FC2: OutBuffer::~OutBuffer() (OutBuffer.cpp:28)
==18078== by 0x8049205: main (in /home/jody/progs/outputredir/OCTest)
==18078== This conflicts with a previous write of size 4 by thread #3
==18078== at 0x804D83F: std::deque<std::string,
std::allocator<std::string> >::push_back(std::string const&)
(stl_deque.h:1042)
==18078== by 0x804C744: LineBuffer::addLine(char*) (LineBuffer.cpp:28)
==18078== by 0x80495F2: OutputCollector::threadStart(void*)
(OutputCollector.cpp:197)
==18078== by 0x40099BA: mythread_wrapper (hg_intercepts.c:194)
==18078== by 0xB0950A: start_thread (in /lib/libpthread-2.7.so)
==18078== by 0xA4AB2D: clone (in /lib/libc-2.7.so)
I have fixed that by inserting mutexes:
pthread_mutex_t lxlx;
LineBuffer::~LineBuffer() {
// (helgrind) this mutex is necessary and must be the same as the one below
pthread_mutex_lock(&lxlx);
pthread_mutex_unlock(&lxlx);
}
void LineBuffer::addLine(char *pLine) {
// (helgrind) this mutex is necessary and must be the same as the one above
pthread_mutex_lock(&lxlx);
if (m_vLines.size() >= m_iMaxLines) {
m_vLines.pop_front();
}
m_vLines.push_back(pLine);
m_pObs->handleNewLine(m_iIndex, pLine);
pthread_mutex_unlock(&lxlx);
// notify observer
}
My question is: has there really been a data race, or did valgrind
only decide that if the timing is right, there could be a data race?
>From the way i see it, there could not have been a real data race...
Thank You
Jody
------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Valgrind-users mailing list
Valgrind-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/valgrind-users
