From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugs-return-409395-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 22301 invoked by alias); 10 Dec 2012 15:56:08 -0000
Received: (qmail 22170 invoked by uid 48); 10 Dec 2012 15:55:51 -0000
From: "Joost.VandeVondele at mat dot ethz.ch" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug sanitizer/55561] TSAN crashes for Fortran
Date: Mon, 10 Dec 2012 15:56:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: sanitizer
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: Joost.VandeVondele at mat dot ethz.ch
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Changed-Fields:
Message-ID: <bug-55561-4-hyHzQtVtKM@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-55561-4@http.gcc.gnu.org/bugzilla/>
References: <bug-55561-4@http.gcc.gnu.org/bugzilla/>
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
X-SW-Source: 2012-12/txt/msg01003.txt.bz2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55561

--- Comment #13 from Joost VandeVondele <Joost.VandeVondele at mat dot ethz.ch> 2012-12-10 15:55:50 UTC ---
(In reply to comment #12)
> That's great that gcc tsan works for Fortran/OpenMP out of the box!

I'm afraid it yields false positives. Something like this is supposed to be OK
OMP-wise (and gives the right result in 10000 repetitions) but produces
warnings.

> cat test.f90 
INTEGER, PARAMETER :: N=10
INTEGER :: data(N)

!$OMP PARALLEL DO
DO j=1,N
  data(j)=j
ENDDO

!$OMP PARALLEL DO
DO j=N,1,-1
  data(j)=data(j)-j
ENDDO

IF (ANY(data.NE.0)) CALL abort()

END

> gfortran -fsanitize=thread -O1 -g -fopenmp -fPIC -pie test.f90 ; ./a.out
==================
WARNING: ThreadSanitizer: data race (pid=18871)


> However, most likely Fortran/OpenMP will require some special handling to
> produce better reports and catch more bugs (ThreadSanitizer reasons about
> synchronization on pthread level, and not on OpenMP level). We won't have spare
> cycles for that in near future.
> I am working on ThreadSanitizer tasking API, I would appreciate if you review
> it and comment on whether it will fit Fortran/OpenMP model (the work is
> currently in very early "what if" stage).

I'm sure there are knowledgeable people reading the gcc list... posting this
text there seems like a good idea.