Looking directly at the patch you are right, the lock is always taken after the
atomic. The merge on my branch left both the try lock and lock out of the code
without any markers for conflict (<<< … === … >>>). Strange.
Sorry for the noise.
George.
On Oct 22, 2013, at 17:46 , Ralph Castain w
I may be missing it, but it looks to me like the lock is taken right before the
list manipulation. All he did was add a check for re-entry before taking the
lock.
On Oct 22, 2013, at 8:39 AM, George Bosilca wrote:
> This patch is not correct, the list manipulation should remain protected
> b
This patch is not correct, the list manipulation should remain protected behind
the mutex.
George.
On Oct 22, 2013, at 17:33 , svn-commit-mai...@open-mpi.org wrote:
> Author: hjelmn (Nathan Hjelm)
> Date: 2013-10-22 11:33:39 EDT (Tue, 22 Oct 2013)
> New Revision: 29465
> URL: https://svn.open
