subject:"\[Bug libfortran\/66756\] libgfortran\: ThreadSanitizer\: lock\-order\-inversion"

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-28 Thread tkoenig at gcc dot gnu.org

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 Thomas Koenig changed: What|Removed |Added Status|REOPENED|ASSIGNED Assignee|unassigned

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-27 Thread tkoenig at gcc dot gnu.org

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 --- Comment #12 from Thomas Koenig --- Correction... the patch does not work with a simple example such as program main !$OMP PARALLEL NUM_THREADS(4) print *,"Hello, world" !$OMP END PARALLEL end program main Some more digging to do...

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-27 Thread tkoenig at gcc dot gnu.org

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 --- Comment #11 from Thomas Koenig --- Created attachment 42250 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=42250=edit Proposed patch This patch is an attempt at getting rid of the lock-order inversion. It seems to do the right thing,

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-27 Thread dominiq at lps dot ens.fr

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 --- Comment #10 from Dominique d'Humieres --- > Could this still be fixed / filtered out in the ThreadSanitizer somehow? Should it be moved to the sanitizer component?

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-27 Thread tkoenig at gcc dot gnu.org

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 Thomas Koenig changed: What|Removed |Added Status|RESOLVED|REOPENED Resolution|INVALID

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-26 Thread Melven.Roehrig-Zoellner at DLR dot de

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 --- Comment #8 from Melven.Roehrig-Zoellner at DLR dot de --- Could this still be fixed / filtered out in the ThreadSanitizer somehow? Otherwise the problem persists that you get huge amounts of false positives with gfortran. (I'm sorry, but I

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-26 Thread tkoenig at gcc dot gnu.org

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 Thomas Koenig changed: What|Removed |Added Status|WAITING |RESOLVED Resolution|---

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2017-09-25 Thread tkoenig at gcc dot gnu.org

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 Thomas Koenig changed: What|Removed |Added Status|NEW |WAITING --- Comment #6 from Thomas

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2016-10-18 Thread dominiq at lps dot ens.fr

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 Dominique d'Humieres changed: What|Removed |Added Status|WAITING |NEW --- Comment #5 from

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2016-10-17 Thread Melven.Roehrig-Zoellner at DLR dot de

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 Melven.Roehrig-Zoellner at DLR dot de changed: What|Removed |Added CC|

[Bug libfortran/66756] libgfortran: ThreadSanitizer: lock-order-inversion

2015-11-11 Thread dominiq at lps dot ens.fr

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756 --- Comment #3 from Dominique d'Humieres --- > but why do set the status on waiting ? Is there any question implied ? Yes. In a perfect world, someone else would have to confirm it. In GCC land, you can change the status to NEW if you still see