Ralph,
Sorry, but no change - still failing to load libmyriexpress.so and still no
rpath at link:
/bin/sh ../../libtool --tag=CC --mode=link gcc -std=gnu99
-fno-strict-aliasing -pthread -g -o dlopen_test dlopen_test.o ../../ompi/
libmpi.la ../../opal/libopen-pal.la -lrt -lm -lutil
libtool: link: gcc -std=gnu99 -fno-strict-aliasing -pthread -g -o
dlopen_test dlopen_test.o ../../ompi/.libs/libmpi.a -L/opt/mx2g/lib
-libverbs -lmyriexpress
/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/orte/.libs/libopen-rte.a
/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/opal/.libs/libopen-pal.a
../../opal/.libs/libopen-pal.a -ldl -lrt -lm -lutil -pthread
I am not surprised, since predefined_gap_test already had those LDFLAGS and
had also failed to locate the library at run time.
Don't misunderstand - I agree the change you made is correct (and
necessary), but it is apparently not sufficient.
It does look like configure did its job:
$ grep ,-rpath config.status
S["OMPI_PKG_CONFIG_LDFLAGS"]="-L/opt/mx2g/lib -Wl,-rpath
-Wl,/opt/mx2g/lib -Wl,-rpath -Wl,${libdir} -Wl,--enable-new-dtags"
S["OMPI_WRAPPER_EXTRA_LDFLAGS"]="-L/opt/mx2g/lib -Wl,-rpath
-Wl,/opt/mx2g/lib -Wl,-rpath -Wl,@{libdir} -Wl,--enable-new-dtags"
S["ORTE_PKG_CONFIG_LDFLAGS"]=" -Wl,-rpath -Wl,${libdir}
-Wl,--enable-new-dtags"
S["ORTE_WRAPPER_EXTRA_LDFLAGS"]=" -Wl,-rpath -Wl,@{libdir}
-Wl,--enable-new-dtags"
S["OPAL_PKG_CONFIG_LDFLAGS"]=" -Wl,-rpath -Wl,${libdir}
-Wl,--enable-new-dtags"
S["OPAL_WRAPPER_EXTRA_LDFLAGS"]=" -Wl,-rpath -Wl,@{libdir}
-Wl,--enable-new-dtags"
D["WRAPPER_EXTRA_LDFLAGS"]=" \"-L/opt/mx2g/lib -Wl,-rpath
-Wl,/opt/mx2g/lib -Wl,-rpath -Wl,@{libdir} -Wl,--enable-new-dtags\""
However, WRAPPER_EXTRA_LDFLAGS isn't defined in any Makefile, despite being
used in several:
$ find . -name Makefile | xargs grep -w WRAPPER_EXTRA_LDFLAGS
./test/event/Makefile:signal_test_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/event/Makefile:time_test_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/event/Makefile:event_test_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:ddt_test_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:ddt_raw_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:ddt_pack_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:checksum_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:position_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:position_noncontig_LDFLAGS =
$(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:to_self_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./test/datatype/Makefile:opal_datatype_test_LDFLAGS =
$(WRAPPER_EXTRA_LDFLAGS)
./ompi/debuggers/Makefile:dlopen_test_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
./ompi/debuggers/Makefile:predefined_gap_test_LDFLAGS =
$(WRAPPER_EXTRA_LDFLAGS)
So, the following result is probably not too surprising (unless due to the
tests NOT appearing):
$ make -k check 2>&1 | grep ^FAIL
FAIL: predefined_gap_test
FAIL: dlopen_test
FAIL: checksum
FAIL: position
FAIL: position_noncontig
FAIL: ddt_test
FAIL: ddt_raw
-Paul
On Sun, Aug 23, 2015 at 7:46 PM, Paul Hargrove <phhargr...@lbl.gov> wrote:
> Ralph,
>
> I will try the requested change and let you know.
>
> -Paul
>
> On Sun, Aug 23, 2015 at 7:27 PM, Ralph Castain <r...@open-mpi.org> wrote:
>
>> Rather than generating another rc right away, could you please apply the
>> following patch and see if it fixes the problem?
>>
>> *diff --git a/ompi/debuggers/Makefile.am b/ompi/debuggers/Makefile.am*
>> *index 3e48af8..343a0c4 100644*
>> *--- a/ompi/debuggers/Makefile.am*
>> *+++ b/ompi/debuggers/Makefile.am*
>> @@ -10,6 +10,7 @@
>> # Copyright (c) 2004-2005 The Regents of the University of California.
>> # All rights reserved.
>> # Copyright (c) 2007-2015 Cisco Systems, Inc. All rights reserved.
>> +# Copyright (c) 2015 Intel, Inc. All rights reserved.
>> # $COPYRIGHT$
>> #
>> # Additional copyrights may follow
>> @@ -43,6 +44,7 @@ headers = \
>> # Simple checks to ensure that the DSOs are functional
>>
>>
>> dlopen_test_SOURCES = dlopen_test.c
>> +dlopen_test_LDFLAGS = $(WRAPPER_EXTRA_LDFLAGS)
>> dlopen_test_LDADD = \
>> $(top_builddir)/ompi/libmpi.la \
>> $(top_builddir)/opal/libopen-pal.la
>>
>>
>> Thanks
>> Ralph
>>
>>
>> On Aug 23, 2015, at 4:58 PM, Paul Hargrove <phhargr...@lbl.gov> wrote:
>>
>> I regret to say that in my endless search for perfection (which is a
>> journey, not a destination) I believe I found another issue:
>>
>> I had mx2g sources sitting around, which I compiled and installed on two
>> systems (x86 and x86-64).
>> These provide only compile/link tests, since my systems lack the hardware
>> (actually do have some in a box somewhere).
>>
>> Things worked fine on both systems when configure with --with-mx=[path].
>> However, on both systems adding "--enable-static --disable-shared"
>> results in failure from "make check":
>>
>> make[3]: Entering directory
>> `/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/ompi/debuggers'
>> CC predefined_gap_test.o
>> CCLD predefined_gap_test
>> CC dlopen_test.o
>> CCLD dlopen_test
>> make[3]: Leaving directory
>> `/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/ompi/debuggers'
>> make check-TESTS
>> make[3]: Entering directory
>> `/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/ompi/debuggers'
>> ./predefined_gap_test: error while loading shared libraries:
>> libmyriexpress.so: cannot open shared object file: No such file or directory
>> FAIL: predefined_gap_test
>> ./dlopen_test: error while loading shared libraries: libmyriexpress.so:
>> cannot open shared object file: No such file or directory
>> FAIL: dlopen_test
>>
>>
>> The following *does* show mpicc using a correct rpath argument for the
>> mx library:
>>
>> $ ./INST/bin/mpicc --show openmpi-1.10.0rc5/examples/ring_c.c
>> gcc openmpi-1.10.0rc5/examples/ring_c.c
>> -I/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/INST/include
>> -pthread -L/opt/mx2g/lib -Wl,-rpath -Wl,/opt/mx2g/lib -Wl,-rpath
>> -Wl,/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/INST/lib
>> -Wl,--enable-new-dtags
>> -L/home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/INST/lib -lmpi
>> -lopen-rte -lopen-pal -lm -ldl -lrt -lmyriexpress -libverbs -lutil
>> $ ls /opt/mx2g/lib
>> debug libmyriexpress.a libmyriexpress.so
>>
>>
>> However, the *test* is built without an rpath:
>>
>> /bin/sh ../../libtool --tag=CC --mode=link gcc -std=gnu99
>> -fno-strict-aliasing -pthread -g -o dlopen_test dlopen_test.o ../../ompi/
>> libmpi.la ../../opal/libopen-pal.la -lrt -lm -lutil
>> libtool: link: gcc -std=gnu99 -fno-strict-aliasing -pthread -g -o
>> dlopen_test dlopen_test.o ../../ompi/.libs/libmpi.a -L/opt/mx2g/lib
>> -libverbs -lmyriexpress
>> /home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/orte/.libs/libopen-rte.a
>> /home/phargrov/OMPI/openmpi-1.10.0rc5-linux-x86-mx-static/BLD/opal/.libs/libopen-pal.a
>> ../../opal/.libs/libopen-pal.a -ldl -lrt -lm -lutil -pthread
>>
>>
>> I have no reason to suspect this is not also present in 1.8.x, but have
>> not checked.
>> Sometime after 1.10.0 is out to door, I expect to test the v1.8 nightly
>> tarball.
>>
>> -Paul [who enjoyed breaking his toys as a child and now breaks software
>> with the same enthusiasm]
>>
>>
>> --
>> Paul H. Hargrove phhargr...@lbl.gov
>> Computer Languages & Systems Software (CLaSS) Group
>> Computer Science Department Tel: +1-510-495-2352
>> Lawrence Berkeley National Laboratory Fax: +1-510-486-6900
>> _______________________________________________
>> 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/2015/08/17803.php
>>
>>
>>
>> _______________________________________________
>> 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/2015/08/17804.php
>>
>
>
>
> --
> Paul H. Hargrove phhargr...@lbl.gov
> Computer Languages & Systems Software (CLaSS) Group
> Computer Science Department Tel: +1-510-495-2352
> Lawrence Berkeley National Laboratory Fax: +1-510-486-6900
>
--
Paul H. Hargrove phhargr...@lbl.gov
Computer Languages & Systems Software (CLaSS) Group
Computer Science Department Tel: +1-510-495-2352
Lawrence Berkeley National Laboratory Fax: +1-510-486-6900
