[Bug libstdc++/49204] New: remaining issues in <future>

[Bug libstdc++/49204] New: remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

           Summary: remaining issues in <future>
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
        AssignedTo: redi@gcc.gnu.org
        ReportedBy: redi@gcc.gnu.org


(this is a placeholder and reminder to myself)

We don't return future_status from the timed waiting functions.


The requirements for async say:

 If the implementation chooses the launch::async policy,
   — a call to a waiting function on an asynchronous return object that
     shares the shared state created by this async call shall block until
     the associated thread has completed, as if joined (30.3.1.5);

The current implementation doesn't meet this requirement,
_Async_state::_M_do_run could make the state ready and unblock waiters
before the thread completes.

Proposed fix: _State_base::_M_wait() already calls _M_run_deferred()
to give a deferred function a chance to run, we could add override it
as _Async_state::_M_run_deferred() and join the thread.


@@ -1325,7 +1325,7 @@
   _M_thread(mem_fn(&_Async_state::_M_do_run), this)
      { }

-      ~_Async_state() { _M_thread.join(); }
+      ~_Async_state() { _M_join(); }

    private:
      void _M_do_run()
@@ -1334,11 +1334,18 @@
        _M_set_result(std::move(__setter));
      }

+      void _M_run_deferred()
+      { _M_join(); }
+
+      void _M_join()
+      { std::call_once(_M_once, &thread::join, ref(_M_thread)); }
+
      template<typename, typename> friend class _Task_setter;
      typedef typename __future_base::_Ptr<_Result<_Res>>::type _Ptr_type;
      _Ptr_type _M_result;
      std::function<_Res()> _M_fn;
      thread _M_thread;
+      once_flag _M_once;
    };

  /// async


That would ensure non-timed waiting functions block as if joined.  It
doesn't handle non-timed waiting functions, but I'm not convinced they
should block.

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2011.05.28 00:31:49
   Target Milestone|---                         |4.7.0
            Summary|remaining issues in         |[C++0x] remaining issues in
                   |<future>                    |<future>
     Ever Confirmed|0                           |1

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-05-31 10:52:17 UTC ---
also http://lwg.github.com/issues/lwg-active.html#2056

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-05-31 13:40:03 UTC ---
We should also make std::async check the system load when deciding whether to
run asynchronously or deferred.  We should be able to reuse GNU Make's code for
deciding whether to run a new job or not, see load_too_high() in job.c

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-23 16:10:58 UTC ---
Author: redi
Date: Fri Dec 23 16:10:48 2011
New Revision: 182658

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=182658
Log:
    PR libstdc++/49204
    * include/std/future (future_errc): Implement LWG 2056.


Modified:
    trunk/libstdc++-v3/ChangeLog
    trunk/libstdc++-v3/include/std/future

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-30 13:10:28 UTC ---
(In reply to comment #0)
> doesn't handle non-timed waiting functions, but I'm not convinced they
> should block.

Reported as http://cplusplus.github.com/LWG/lwg-active.html#2100

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-01-02 17:56:19 UTC ---
(In reply to comment #2)
> We should also make std::async check the system load when deciding whether to
> run asynchronously or deferred.  We should be able to reuse GNU Make's code for
> deciding whether to run a new job or not, see load_too_high() in job.c

It turns out that calling get_nprocs() (via
std::thread::hardware_concurrency()) and getloadavg() from several threads
concurrently is a really bad idea, so I'm working on a simple implementation
that just compares the number of active async jobs to the number of processors.

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-02-01 00:20:27 UTC ---
Author: redi
Date: Wed Feb  1 00:20:08 2012
New Revision: 183788

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=183788
Log:
    PR libstdc++/49204
    * include/std/future (__future_base::_State_base::wait()): Use lambda
    expression for predicate and remove redundant test.
    (__future_base::_State_base::wait_for()): Return future_status and
    use lambda expression for predicate.
    (__future_base::_State_base::wait_until()): Likewise.
    (__basic_future::wait_for(), __basic_future::wait_until()): Likewise.
    (__future_base::_Async_state): Replace with _Async_state_common
    class for non-dependent functionality and _Async_state_impl class
    template for dependent functionality.
    (__future_base::_Async_state_common::_M_join): Serialize attempts to
    join thread.
    (__future_base::_Async_state_common::_M_run_deferred): Join.
    (__future_base::_Async_state::_M_do_run): Replace with lambda.
    * src/c++11/future.cc (__future_base::_Async_state_common): Define
    destructor, so key function is in the library.
    * config/abi/pre/gnu.ver: Add exports for ~_Async_state_common.
    * testsuite/30_threads/packaged_task/members/get_future.cc: Expect
    future_status return instead of bool.
    * testsuite/30_threads/shared_future/members/wait_until.cc: Likewise.
    * testsuite/30_threads/shared_future/members/wait_for.cc: Likewise.
    * testsuite/30_threads/future/members/wait_until.cc: Likewise.
    * testsuite/30_threads/future/members/wait_for.cc: Likewise.
    * testsuite/30_threads/promise/members/set_value2.cc: Likewise.
    * testsuite/30_threads/promise/members/set_value3.cc: Likewise.
    * testsuite/30_threads/promise/members/swap.cc: Likewise.

Modified:
    trunk/libstdc++-v3/ChangeLog
    trunk/libstdc++-v3/config/abi/pre/gnu.ver
    trunk/libstdc++-v3/include/std/future
    trunk/libstdc++-v3/src/c++11/future.cc
    trunk/libstdc++-v3/testsuite/30_threads/future/members/wait_for.cc
    trunk/libstdc++-v3/testsuite/30_threads/future/members/wait_until.cc
    trunk/libstdc++-v3/testsuite/30_threads/packaged_task/members/get_future.cc
    trunk/libstdc++-v3/testsuite/30_threads/promise/members/set_value2.cc
    trunk/libstdc++-v3/testsuite/30_threads/promise/members/set_value3.cc
    trunk/libstdc++-v3/testsuite/30_threads/promise/members/swap.cc
    trunk/libstdc++-v3/testsuite/30_threads/shared_future/members/wait_for.cc
    trunk/libstdc++-v3/testsuite/30_threads/shared_future/members/wait_until.cc

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-02-01 00:25:31 UTC ---
Still outstanding tasks:

timed waiting functions do nto return future_status::deferred if the shared
state contains a deferred function (requires an additional virtual function to
be added to the vtable)

check system load when deciding whether to use async or deferred policy

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.7.0                       |4.8.0

--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-02-03 09:22:18 UTC ---
Not planning more work before 4.7.0

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ABI

--- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-08-22 19:20:19 UTC ---
(In reply to comment #7)
> timed waiting functions do nto return future_status::deferred if the shared
> state contains a deferred function (requires an additional virtual function to
> be added to the vtable)

setting ABI keyword as this can't be properly fixed without an incompatible
change

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.0                       |4.8.1

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-03-22 14:41:50 UTC ---
GCC 4.8.0 is being released, adjusting target milestone.

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.1                       |4.9.0

--- Comment #11 from Jonathan Wakely <redi at gcc dot gnu.org> 2013-03-23 15:01:56 UTC ---
I didn't get this done for 4.8 and it would be too invasive for 4.8.1 so
adjusting milestone

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

--- Comment #12 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Author: redi
Date: Wed Nov 20 20:59:19 2013
New Revision: 205144

URL: http://gcc.gnu.org/viewcvs?rev=205144&root=gcc&view=rev
Log:
    PR libstdc++/49204
    * include/std/future (__future_base::_State_base): Rename to
    __future_base::_State_baseV2.
    (__future_base::_State_baseV2::~_State_baseV2): Define as defaulted.
    (__future_base::_State_baseV2::_M_run_deferred): Rename to
    _M_complete_async.
    (__future_base::_State_baseV2::_M_has_deferred): Add new virtual.
    (__future_base::_State_baseV2::wait_for): Call _M_has_deferred() to
    test for a deferred function, or call _M_complete_async() to join an
    async thread that has made the shared state ready.
    (__future_base::_State_baseV2::wait_until): Likewise.
    (__future_base::_Async_state_common): Rename to _Async_state_commonV2.
    (__future_base::_Async_state_commonV2::_M_run_deferred): Rename to
    _M_complete_async.
    * src/c++11/compatibility-thread-c++0x.cc (__future_base::_State_base):
    Export old definition.
    (__future_base::_Async_state_common): Likewise.
    * src/c++11/future.cc (__future_base::_State_base::~_State_base):
    Remove.
    * doc/xml/manual/status_cxx2011.xml: Update status.
    * testsuite/30_threads/async/async.cc: Test future_status::timeout
    and future_status::ready.
    * testsuite/30_threads/async/sync.cc: Test future_status::deferred.

Modified:
    trunk/libstdc++-v3/ChangeLog
    trunk/libstdc++-v3/doc/xml/manual/status_cxx2011.xml
    trunk/libstdc++-v3/include/std/future
    trunk/libstdc++-v3/src/c++11/compatibility-thread-c++0x.cc
    trunk/libstdc++-v3/src/c++11/future.cc
    trunk/libstdc++-v3/testsuite/30_threads/async/async.cc
    trunk/libstdc++-v3/testsuite/30_threads/async/sync.cc

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|ABI                         |
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #13 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jonathan Wakely from comment #7)
> Still outstanding tasks:
> 
> timed waiting functions do nto return future_status::deferred if the shared
> state contains a deferred function (requires an additional virtual function
> to be added to the vtable)

This part is fixed for 4.9.0

> check system load when deciding whether to use async or deferred policy

PR 51617 covers that.

We're also missing all the "_at_thread_exit" throughout Clause 30, but I'm
going to resolve this PR as fixed.

[Bug libstdc++/49204] [C++0x] remaining issues in <future>

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49204
Bug 49204 depends on bug 51617, which changed state.

Bug 51617 Summary: [C++0x] async(f) isn't.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51617

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED