std::future regression

[Quuxplusone]

Bugzilla Link	PR23099
Status	NEW
Importance	P normal
Reported by	dirk.ribbrock@mathematik.tu-dortmund.de
Reported on	2015-04-01 14:28:18 -0700
Last modified on	2015-04-06 15:22:43 -0700
Version	3.6
Hardware	PC Linux
CC	dgregor@apple.com, dimitry@andric.com, llvm-bugs@lists.llvm.org
Fixed by commit(s)
Attachments
Blocks
Blocked by
See also

Since using clang 3.6.0, the following code crashes with
"Illegal instruction (core dumped)"
Whereas clang 3.5.0  returns the correct output value "1"

//file future.cpp
#include <thread>
#include <future>
#include <iostream>

int bla()
{
  return 1;
}

int main()
{
  std::future<int> f = std::async(std::launch::async, bla);
  f.wait();
  std::cout<<f.get()<<std::endl;
  return 0;
}
//end file

Build it with
clang++ --std=c++11 -lpthread future.cpp

[Quuxplusone]

This works fine on FreeBSD 11 (with clang 3.6.0), but we're using a relatively recent libc++ there.

Which C++ library are you using in case of the crash? Is that libstdc++ on your particular Linux instance? And which version?

Also, it would be interesting to see a debug backtrace. Is it crashing on a 'ud2' instruction, by any chance?

[Quuxplusone]

Hi

I'am using Ubuntu 12.04 with my own gcc 4.8.3 compiler.
The libstdc++.so.6 used by clang and the repro code is build with this gcc
compiler.
The greatest version symbol i can find in there is
GLIBCXX_3.4.19.

gdb backtraces reads as follows:

Program received signal SIGILL, Illegal instruction.
bool
std::__future_base::_State_base::_S_check<std::__future_base::_State_base>(std::shared_ptr<std::__future_base::_State_base>
const&) ()
    at /sfw/gcc/4.7.3-static-gmp-mpfr-mpc-cloog/lib64/gcc/x86_64-unknown-linux-gnu/4.7.3/../../../../include/c++/4.7.3/future:462
462           if (!static_cast<bool>(__p))
(gdb) bt
#0  bool
std::__future_base::_State_base::_S_check<std::__future_base::_State_base>(std::shared_ptr<std::__future_base::_State_base>
const&) ()
    at /sfw/gcc/4.7.3-static-gmp-mpfr-mpc-cloog/lib64/gcc/x86_64-unknown-linux-gnu/4.7.3/../../../../include/c++/4.7.3/future:462
#1  0x0000000000402b1d in
std::__basic_future<int>::__basic_future(std::shared_ptr<std::__future_base::_State_base>
const&) ()
    at /sfw/gcc/4.7.3-static-gmp-mpfr-mpc-cloog/lib64/gcc/x86_64-unknown-linux-gnu/4.7.3/../../../../include/c++/4.7.3/future:613
#2  0x0000000000402ae8 in
std::future<int>::future(std::shared_ptr<std::__future_base::_State_base>
const&) ()
    at /sfw/gcc/4.7.3-static-gmp-mpfr-mpc-cloog/lib64/gcc/x86_64-unknown-linux-gnu/4.7.3/../../../../include/c++/4.7.3/future:654
#3  0x0000000000401f48 in
_ZSt5asyncIRFivEJEESt6futureINSt9result_ofIFT_DpT0_EE4typeEESt6launchOS4_DpOS5_
()
    at /sfw/gcc/4.7.3-static-gmp-mpfr-mpc-cloog/lib64/gcc/x86_64-unknown-linux-gnu/4.7.3/../../../../include/c++/4.7.3/future:1499
#4  0x0000000000401b1b in main () at future.cpp:12

line 12 is the line with std::async.

If i can give you any further information, i'm pleased to do so.

Thanks in advance
Dirk

[Quuxplusone]

My fault, its gcc 4.7.3.

[Quuxplusone]

Consequently, the greatest glibc version symbol is
GLIBCXX_3.4.17

[Quuxplusone]

I will try to rebuild clang with a newer libstdc++ to get rid of the problem :)

[Quuxplusone]

Works fine for me with newer version of libstdc++ :)

Thanks for your help!
Dirk