[Bug fortran/114467] New: f951: internal compiler error: Segmentation fault

[Bug fortran/114467] New: f951: internal compiler error: Segmentation fault

[thomas.kalscheuer at geo dot uu.se]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114467

            Bug ID: 114467
           Summary: f951: internal compiler error: Segmentation fault
           Product: gcc
           Version: 13.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: thomas.kalscheuer at geo dot uu.se
  Target Milestone: ---

I use a system with dual Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz cpus, CentOS
Linux 7 and gfortran v. 13.2.0

According to gcc -v, gcc was configured with the options "Configured with:
../gcc-13.2.0/configure --enable-languages=c,c++,fortran"

By compiling a fortran code with 
gfortran -O2 -march=native -mtune=native -malign-data=cacheline
-finline-functions -std=f2018 -ffree-form -ffree-line-length-none -pedantic
-Wsurprising -Wopenmp-simd -Wvector-operation-performance -ftree-vectorize
-ftree-vectorizer-verbose=2 -fopt-info -fopenmp -I/usr/include
-I/usr/include/hdf5/serial/ -freport-bug -save-temps -x f95-cpp-input -DRMT -c
code.f90 -o code.o
I get the following error message:
f951: internal compiler error: Segmentation fault
0xd29d7f crash_signal
        ../../gcc-13.2.0/gcc/toplev.cc:314
0x7c2a65 gfc_expression_rank(gfc_expr*)
        ../../gcc-13.2.0/gcc/fortran/resolve.cc:5643
0x7c2c14 gfc_op_rank_conformable(gfc_expr*, gfc_expr*)
        ../../gcc-13.2.0/gcc/fortran/resolve.cc:5676
0x709e9c eval_intrinsic
        ../../gcc-13.2.0/gcc/fortran/arith.cc:1674
0x781cfc match_level_4
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:636
0x781cfc match_and_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:693
0x781d66 match_or_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:722
0x781e76 match_equiv_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:765
0x781f88 match_level_5
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:811
0x7810e4 gfc_match_expr(gfc_expr**)
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:870
0x78132b match_primary
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:165
0x78132b match_level_1
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:211
0x78132b match_mult_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:267
0x7815c8 match_add_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:356
0x78189c match_level_2
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:480
0x781a56 match_level_3
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:551
0x781b90 match_level_4
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:599
0x781b90 match_and_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:693
0x781d66 match_or_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:722
0x781e76 match_equiv_operand
        ../../gcc-13.2.0/gcc/fortran/matchexp.cc:765
Please submit a full bug report, with preprocessed source.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
make: *** [code.o] Error 1

Compilation with
gfortran -g -std=f2018 -ffree-form -ffree-line-length-none -pedantic -Wall
-I/usr/include -I/usr/include/hdf5/serial/ -freport-bug -save-temps -x
f95-cpp-input -DRMT -c code.f90 -o code.o
leads to the same error message.

Note, that inclusion of the -save-temps option has not led to any *.* file
being generated by the compiler.

[Bug fortran/114467] f951: internal compiler error: Segmentation fault

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114467

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2024-03-25
     Ever confirmed|0                           |1

--- Comment #1 from anlauf at gcc dot gnu.org ---
Can you attached a self-contained reproducer?

The traceback looks familiar.  Are you by chance using an associate
construct?  There has been a fix backported from 14-mainline to 13-branch
after the release of 13.2, so it might have been fixed in the meantime.

(Likely a dup of pr109948).

[Bug fortran/114467] f951: internal compiler error: Segmentation fault

[thomas.kalscheuer at geo dot uu.se]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114467

--- Comment #2 from thomas <thomas.kalscheuer at geo dot uu.se> ---
Yes, I use associate constructs at multiple places in the code. Unfortunately,
I won't easily be able to provide a self-contained reproducer, as the error
occurs in the middle of a larger compilation cascade involving about 30 or so
modules. I will talk to our admin to install a more recent version of gcc 13.2.

[Bug fortran/114467] f951: internal compiler error: Segmentation fault

[thomas.kalscheuer at geo dot uu.se]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114467

--- Comment #3 from thomas <thomas.kalscheuer at geo dot uu.se> ---
(In reply to anlauf from comment #1)
> Can you attached a self-contained reproducer?
> 
> The traceback looks familiar.  Are you by chance using an associate
> construct?  There has been a fix backported from 14-mainline to 13-branch
> after the release of 13.2, so it might have been fixed in the meantime.
> 
> (Likely a dup of pr109948).

Just to be on the safe side, would gcc 13.2.0 from 2023-07-27 be the right
version to test?

[Bug fortran/114467] f951: internal compiler error: Segmentation fault

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114467

--- Comment #4 from anlauf at gcc dot gnu.org ---
(In reply to thomas from comment #3)
> (In reply to anlauf from comment #1)
> > Can you attached a self-contained reproducer?
> > 
> > The traceback looks familiar.  Are you by chance using an associate
> > construct?  There has been a fix backported from 14-mainline to 13-branch
> > after the release of 13.2, so it might have been fixed in the meantime.
> > 
> > (Likely a dup of pr109948).
> 
> Just to be on the safe side, would gcc 13.2.0 from 2023-07-27 be the right
> version to test?

No.  The backport was applied to 13-branch on 2023-08-27.

I'd recommend to use a more recent snapshot of 13-branch.

[Bug fortran/114467] f951: internal compiler error: Segmentation fault

[thomas.kalscheuer at geo dot uu.se]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114467

thomas <thomas.kalscheuer at geo dot uu.se> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |WORKSFORME
             Status|WAITING                     |RESOLVED

--- Comment #5 from thomas <thomas.kalscheuer at geo dot uu.se> ---
(In reply to anlauf from comment #4)

I have downloaded a recent snapshot, which turned out to be v. 14.0.1 rather
than the latest 13-branch. But I can confirm that the problem does not occur
anymore when using gfortran v. 14.0.1.

Many thanks for your kind help!