------- Additional Comments From schnetter at aei dot mpg dot de 2005-06-04 23:15 ------- I tried to reduce the failing source code, but failed to achieve much. Depending on what lines I remove, the error changes, sometimes turning into a bare "segmentation fault". Even renaming the source file can make the error go away. With a different source file (much smaller; attached) and such a bare segmentation fault, I get the following stack backtrace from valgrind: $ valgrind --log-file=gfortran --trace-children=yes ~/gcc/bin/gfortran -c transform_global.f90 /home/eschnett/Calpha/Cactus/arrangements/LSUDevelopment/HyperGR_MoL/src/transform_global.F90: In function 'local2global': /home/eschnett/Calpha/Cactus/arrangements/LSUDevelopment/HyperGR_MoL/src/transform_global.F90:112: internal compiler error: Segmentation fault ==28630== My PID = 28630, parent PID = 28629. Prog and args are: ==28630== /home/eschnett/gcc/libexec/gcc/i686-pc-linux-gnu/4.1.0/f951 ==28630== transform_global.f90 ==28630== -quiet ==28630== -dumpbase ==28630== transform_global.f90 ==28630== -mtune=pentiumpro ==28630== -auxbase ==28630== transform_global ==28630== -o ==28630== /tmp/ccT8xxXE.s ==28630== For more details, rerun with: -v ==28630== ==28630== Invalid read of size 4 ==28630== at 0x8099EDE: gfc_conv_scalarized_array_ref (trans-array.c:1578) ==28630== by 0x8099F9F: gfc_conv_array_ref (trans-array.c:1622) ==28630== by 0x80A6912: gfc_conv_variable (trans-expr.c:420) ==28630== by 0x80B20DC: set_string (trans-io.c:403) ==28630== by 0x80B38E2: build_dt (trans-io.c:1066) ==28630== by 0x80B3A22: gfc_trans_write (trans-io.c:1188) ==28630== by 0x80970E4: gfc_trans_code (trans.c:595) ==28630== by 0x80B4B43: gfc_trans_if_1 (trans-stmt.c:423) ==28630== by 0x809702E: gfc_trans_code (trans.c:547) ==28630== by 0x80B4B03: gfc_trans_if_1 (trans-stmt.c:413) ==28630== by 0x80B4B54: gfc_trans_if_1 (trans-stmt.c:427) ==28630== by 0x809702E: gfc_trans_code (trans.c:547) ==28630== Address 0xC4 is not stack'd, malloc'd or (recently) free'd
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21902
