Nathan, r31829 caused many sigsegv :-( /* i am testing on a RHEL6.3 like VM with --mca btl,tcp */
this is now fixed in r31830, i think i get the intent of the code and i believe we are all set now. bottom line : - we agree on line 487 (e.g. use btl_send) - your update of line 485 is correct (e.g. use btl_send) - my suggested update of line 498 (e.g. use btl_send) was correct. Cheers, Gilles On 2014/05/20 4:06, Nathan Hjelm wrote: > On Mon, May 19, 2014 at 02:14:57PM +0900, Gilles Gouaillardet wrote: >> Nathan, >> >> do you mean the bug/typo was not at line 487 >> (e.g. btl_send was ok) >> but at line 498 ? >> (e.g. btl_send must be used instead of btl_eager) > Yup. If you look at the next loop (L497) it looks through the btl_send > list and then calls del_procs if it finds the btl in that list. > >> at first sight, that make sense. >> >> i'd rather let the author/maintainer of this part comment on that > I don't know if the original author still works on Open MPI. I think we > will have to guess the intent of the code. Let me take a closer look and > see if I can determine for sure what was intended. If I can determine > for sure I will include this change with another bml fix that needs to > go in. > > -Nathan > > > _______________________________________________ > devel mailing list > de...@open-mpi.org > Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/devel > Link to this post: > http://www.open-mpi.org/community/lists/devel/2014/05/14828.php