Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

Subject: [OMPI devel] ud oob is borked
From: Jeff Squyres (jsquyres_at_[hidden])
Date: 2012-07-31 17:00:37


There's a compile error in the ud oob right now. I tried a few different ways to fix it, but I'm still consistently getting segv's.

-----
[svbu-mpi046:02934] wc status = 2
[svbu-mpi046:02934] *** Process received signal ***
[svbu-mpi046:02934] Signal: Segmentation fault (11)
[svbu-mpi046:02934] Signal code: Address not mapped (1)
[svbu-mpi046:02934] Failing at address: 0x128
[svbu-mpi046:02934] [ 0] /lib64/libpthread.so.0() [0x3d5940f4a0]
[svbu-mpi046:02934] [ 1] /home/jsquyres/bogus/lib/libopen-rte.so.0(mca_oob_ud_msg_post_send+0x1ce) [0x7ffff7c686d7]
[svbu-mpi046:02934] [ 2] /home/jsquyres/bogus/lib/libopen-rte.so.0(mca_oob_ud_send_nb+0x5d1) [0x7ffff7c6a851]
[svbu-mpi046:02934] [ 3] /home/jsquyres/bogus/lib/libopen-rte.so.0(orte_rml_oob_send_buffer_nb+0x5bd) [0x7ffff7cb70f3]
[svbu-mpi046:02934] [ 4] /home/jsquyres/bogus/lib/libopen-rte.so.0(orte_daemon+0x17de) [0x7ffff7c1c701]
[svbu-mpi046:02934] [ 5] /home/jsquyres/bogus/bin/orted() [0x40082a]
[svbu-mpi046:02934] [ 6] /lib64/libc.so.6(__libc_start_main+0xfd) [0x3d5901ecdd]
[svbu-mpi046:02934] [ 7] /home/jsquyres/bogus/bin/orted() [0x4006e9]
[svbu-mpi046:02934] *** End of error message ***
Segmentation fault (core dumped)
-----

So that we don't get another night of 161K MTT failures at Cisco (before I killed it), I'm going to .ompi_ignore the ud oob on the trunk.

Nathan: feel free to un-ompi-ignore it when you have it fixed. Thanks.

-- 
Jeff Squyres
jsquyres_at_[hidden]
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/