This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Segmentation fault in openmp simple routine from libgomp testsuite.
- From: Diego Novillo <dnovillo at redhat dot com>
- To: FX Coudert <fxcoudert at gmail dot com>
- Cc: Jakub Jelinek <jakub at redhat dot com>, Benjamin Réveill é <benjamin dot reveille at gmail dot com>, Fortran List <fortran at gcc dot gnu dot org>, GCC Development <gcc at gcc dot gnu dot org>, Richard Henderson <rth at redhat dot com>
- Date: Tue, 07 Mar 2006 12:34:05 -0500
- Subject: Re: Segmentation fault in openmp simple routine from libgomp testsuite.
- Organization: Red Hat Canada
- References: <200602280858.23231.benjamin.reveille@gmail.com> <19c433eb0602280508k6c1ef204x645c1fc263fa77eb@mail.gmail.com> <9e363c0d0602280813y54cee9ebk392dc2b195fd9b7a@mail.gmail.com> <9e363c0d0602280815r33761e56l681b8567730e8a9b@mail.gmail.com> <d0602e7eba0bd4d0884d935e145b4f3a@gmail.com>
On 02/28/06 18:42, FX Coudert wrote:
> Jakuk, Diego? Is this a bug or a feature? :)
>
Looks like a bug, but I'm not really sure what is causing it. I can
reproduce it with one of the tests in libgomp
(libgomp.fortran/appendix-a/a.16.1.f90), but I get a different trace:
---------------------------------------------------------------------------
$ <bld>/gcc/xgcc -B<bld>/gcc/
<src>/libgomp/testsuite/libgomp.fortran/appendix-a/a.16.1.f90
-B<bld>/i686-pc-linux-gnu/./libgomp/ -I<bld>/i686-pc-linux-gnu/./libgomp
-I<src>/libgomp/testsuite/.. -march=i486 -fmessage-length=0 -O0
-L<bld>/i686-pc-linux-gnu/./libgomp/.libs -lgomp
-L<bld>/i686-pc-linux-gnu/./libgomp/../libgfortran/.libs -lgfortranbegin
-lgfortran -lm -o ./a.16.1.exe -static
$ ./a.16.1.exe
Segmentation fault (core dumped)
$ gdb ./a.16.1.exe core.23910
[ ... ]
Reading symbols from shared object read from target memory...done.
Loaded system supplied DSO at 0x589000
Core was generated by `./a.16.1.exe'.
Program terminated with signal 11, Segmentation fault.
#0 0x00000000 in ?? ()
(gdb) bt
#0 0x00000000 in ?? ()
#1 0x0804d112 in find_unit_1 (n=6, do_create=1)
at /home/dnovillo/gomp/src/libgfortran/../gcc/gthr-posix.h:528
#2 0x0804be69 in data_transfer_init (dtp=0xbfffeebc, read_flag=0)
at /home/dnovillo/gomp/src/libgfortran/io/transfer.c:1333
#3 0x0804857b in MAIN__ ()
#4 0x080493e7 in main (argc=Cannot access memory at address 0x0)
at /home/dnovillo/gomp/src/libgfortran/fmain.c:18
----------------------------------------------------------------------
Richard mentioned similar problems with broken libc versions that
wouldn't initialize TLS properly, but this particular one doesn't seem
related. Richard, any ideas?