I guess you have an old glibc there ? and HAVE_OPENAT isn't defined in
config.h ? I am fixing this.
Brice
Le 16/07/2010 15:34, Jeff Squyres a écrit :
> (gdb) bt
> #0 0x0060bcb8 in strcmp () from /lib/tls/libc.so.6
> #1 0x00498d7d in hwloc_look_linux (topology=0x8807008)
> at topology-linux.c:1805
> #2 0x0048dd9c in hwloc_discover (topology=0x8807008) at topology.c:1423
> #3 0x0048ed8e in hwloc_topology_load (topology=0x8807008) at topology.c:2005
> #4 0x080489a4 in main () at hwloc_is_thissystem.c:34
> (gdb) up
> #1 0x00498d7d in hwloc_look_linux (topology=0x8807008)
> at topology-linux.c:1805
> 1805 if (!strcmp(topology->backend_params.sysfs.root_path, "/"))
> (gdb) p topology->backend_params.sysfs
> $1 = {root_path = 0x0, root_fd = -1}
> (gdb)
>
> Need any other info?
>
>
> On Jul 16, 2010, at 12:59 AM, Brice Goglin wrote:
>
>
>> Le 16/07/2010 03:02, MPI Team a écrit :
>>
>>> /bin/sh: line 1: 1024 Segmentation fault ${dir}$tst
>>> FAIL: hwloc_bind
>>> /bin/sh: line 1: 1048 Segmentation fault ${dir}$tst
>>> FAIL: hwloc_object_userdata
>>> PASS: hwloc_synthetic
>>> /bin/sh: line 1: 1094 Segmentation fault ${dir}$tst
>>> FAIL: hwloc_is_thissystem
>>> /bin/sh: line 1: 1118 Segmentation fault ${dir}$tst
>>> FAIL: hwloc_insert_misc
>>> /bin/sh: line 1: 1142 Segmentation fault ${dir}$tst
>>> FAIL: glibc-sched
>>>
>>>
>> I can't reproduce this on my machines. Can somebody send the
>> corresponding gdb backtrace? Take trunk, ./configure, make, make check,
>> and it fails
>> libtool --mode=execute gdb <test program>
>>
>> thanks
>> Brice
>>
>> _______________________________________________
>> hwloc-devel mailing list
>> hwloc-de...@open-mpi.org
>> http://www.open-mpi.org/mailman/listinfo.cgi/hwloc-devel
>>
>>
>
>
