std::thread segfault

[cyberfish]

I just decided to start playing with my shiny new GCC-4.4's std::thread, and... well... it kept segfaulting on me .

Admittedly I didn't RTFM, since TFM for C++0x is somewhat hard to find currently, and I assumed it's going to be very similar to boost::thread.

#include <thread>

#include <unistd.h>

void t() {
        sleep(5);
}

int main() {
        std::thread t1(&t);
        std::thread t2(&t);

        t1.join();
        t2.join();
}

Anyone wants to venture a guess at what might be wrong?

cyberfish@cyberfish-desktop:/tmp$ g++-4.4 -g -std=gnu++0x a.cpp
cyberfish@cyberfish-desktop:/tmp$ gdb a.out
GNU gdb 6.8-debian
Copyright (C) 2008 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu"...
(gdb) r
Starting program: /tmp/a.out 

Program received signal SIGSEGV, Segmentation fault.
0x0000000000000000 in ?? ()
(gdb) bt
#0  0x0000000000000000 in ?? ()
#1  0x00007fa9de939fa7 in std::thread::_M_start_thread (this=0x7fffe6da40e0, 
    __b=DWARF-2 expression error: DW_OP_reg operations must be used either alone or in conjuction with DW_OP_piece.
)
    at /home/cyberfish/gcc_svn/gcc/build/x86_64-linux-gnu/libstdc++-v3/include/x86_64-linux-gnu/bits/gthr-default.h:682
#2  0x0000000000400e6c in thread<void (*)()> (this=0x7fffe6da40e0, 
    __f=0x400c0c <t()>)
    at /opt/gcc/lib/gcc/x86_64-linux-gnu/4.4.0/../../../../include/c++/4.4.0/thread:134
#3  0x0000000000400c38 in main () at a.cpp:10
(gdb) l
1	#include <thread>
2	
3	#include <unistd.h>
4	
5	void t() {
6		sleep(5);
7	}
8	
9	int main() {
10		std::thread t1(&t);
(gdb)

Thanks

[laserlight]

Admittedly I didn't RTFM, since TFM for C++0x is somewhat hard to find currently, and I assumed it's going to be very similar to boost::thread.

Check out the standards committee website for the current working draft of C++0x.

[cyberfish]

Thanks!

I read the part on std::thread, and still don't see anything wrong.

[cyberfish]

Solution found on GCC mailing list -
Timothy Brownawell - Re: std::thread segfault

add "-pthread"

Working great now!

#include <thread>
#include <mutex>

#include <iostream>

#include <unistd.h>

std::mutex io_mutex;

void f(int id) {
	for (int i = 0; i < 5; ++i) {
		sleep(1);
		{
			std::lock_guard<std::mutex> lock(io_mutex); //scoped lock
			std::cout << "T" << id << ":" << i << std::endl;
		}
	}
}

int main() {
	std::thread t1(f, 0);
	std::thread t2(f, 1);

	t1.join();
	t2.join();

	return 0;
}

cyberfish@cyberfish-desktop:/tmp$ g++-4.4 -std=gnu++0x a.cpp -pthread
cyberfish@cyberfish-desktop:/tmp$ ./a.out
T0:0
T1:0
T0:1
T1:1
T0:2
T1:2
T0:3
T1:3
T0:4
T1:4

[brewbuck]

I hate that "-pthread" flag. A lot of times, everything links and seems to work just fine without it. Other times it explodes. I can find no rhyme or reason to it.

In this day and age, I'd prefer it if we had to explicitly specify "-no-pthread" instead of the other way around...