Code indexing in gitaly is broken and leads to code not being visible to the user. We work on the issue with highest priority.

Skip to content

OPAL (master) gives a segfault when ending

Summary

The most recent master version compiles fine on fedora fc35 but it gives a segfault upon finishing A checkout of the OPAL-2011-1 compile on the same distro works just fine.

Steps to reproduce the failure

git check out git checkout master

cd ../master CC=mpicc CXX=mpicxx cmake ../src/ -DHDF5_C_COMPILER_EXECUTABLE=/usr/lib64/openmpi/bin/h5pcc -D CMAKE_BUILD_TYPE=Release -D CMAKE_CXX_FLAGS="-Wno-error=cast-function-type -Wno-cast-function-type -Wl,--copy-dt-needed-entries"

make -j 24

cd ../example

mpirun -np 8 ../master/src/opal awaDrive_all.in

What is the current bug behavior?

Timings{0}> WakeField........... Wall max =          0, CPU max =          0
Timings{0}>                      Wall avg =          0, CPU avg =          0
Timings{0}>                      Wall min =          0, CPU min =          0
Timings{0}> 
Timings{0}> Write H5-File....... Wall max =    2.27603, CPU max =       1.23
Timings{0}>                      Wall avg =    2.27519, CPU avg =      1.115
Timings{0}>                      Wall min =    2.27071, CPU min =       1.03
Timings{0}> 
Timings{0}> Write Stat.......... Wall max =   0.105094, CPU max =       0.07
Timings{0}>                      Wall avg =  0.0356317, CPU avg =    0.02625
Timings{0}>                      Wall min =  0.0245887, CPU min =          0
Timings{0}> 
Timings{0}> -----------------------------------------------------------------
[localhost:356319:0:356319] Caught signal 11 (Segmentation fault: address not mapped to object at address (nil))
[localhost:356320:0:356320] Caught signal 11 (Segmentation fault: address not mapped to object at address (nil))
[localhost:356318:0:356318] Caught signal 11 (Segmentation fault: address not mapped to object at address (nil))
[localhost:356321:0:356321] Caught signal 11 (Segmentation fault: address not mapped to object at address (nil))
==== backtrace (tid: 356320) ====
 0  /lib64/libucs.so.0(ucs_handle_error+0x2a4) [0x7f1d48c29864]
 1  /lib64/libucs.so.0(+0x2a41d) [0x7f1d48c2d41d]
 2  /lib64/libucs.so.0(+0x2a5fa) [0x7f1d48c2d5fa]
 3  opal() [0xc8f6c0]
 4  opal() [0x910701]
 5  opal() [0x7eb18c]
 6  opal() [0xb935d9]
 7  opal() [0x918acf]
 8  opal() [0xa1a7af]
 9  opal() [0x72c16a]
10  opal() [0x72c189]
11  opal() [0x6f368a]
12  /lib64/libc.so.6(+0x57a15) [0x7f1d4d257a15]
13  /lib64/libc.so.6(on_exit+0) [0x7f1d4d257b90]
14  /lib64/libc.so.6(+0x40447) [0x7f1d4d240447]
15  /lib64/libc.so.6(__libc_start_main+0x80) [0x7f1d4d2404f0]
16  opal() [0x4abbd5]
=================================
==== backtrace (tid: 356319) ====
 0  /lib64/libucs.so.0(ucs_handle_error+0x2a4) [0x7fa60d660864]
 1  /lib64/libucs.so.0(+0x2a41d) [0x7fa60d66441d]
 2  /lib64/libucs.so.0(+0x2a5fa) [0x7fa60d6645fa]
 3  opal() [0xc8f6c0]
 4  opal() [0x910701]
 5  opal() [0x7eb18c]
 6  opal() [0xb935d9]
 7  opal() [0x918acf]

What is the expected correct behavior?

git check out git checkout OPAL-2011-1

cd ../OPAL-2011-1

CC=mpicc CXX=mpicxx cmake ../src/ -DHDF5_C_COMPILER_EXECUTABLE=/usr/lib64/openmpi/bin/h5pcc -D CMAKE_BUILD_TYPE=Release -D CMAKE_CXX_FLAGS="-Wno-error=cast-function-type -Wno-cast-function-type -Wl,--copy-dt-needed-entries"

make -j 24

cd ../example

mpirun -np 8 ../OPAL-2011-1/src/opal awaDrive_all.in

works fine

Relevant logs and/or screenshots

the input file I use are the AWA photoinjector; the specific version is available at https://xgitlab.cels.anl.gov/awa/awa-opal-lattices

Edited by ext-piot_p