[Bug fortran/66347] New: Seg fault (ICE) on compile

[Bug fortran/66347] New: Seg fault (ICE) on compile

[wbrisken at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

            Bug ID: 66347
           Summary: Seg fault (ICE) on compile
           Product: gcc
           Version: 5.1.1
            Status: UNCONFIRMED
          Severity: major
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: wbrisken at gmail dot com
  Target Milestone: ---

Created attachment 35656
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=35656&action=edit
fortran test case causing ICE

I have an example of code that would compile without problem with gfortran
4.4.7 and gfortran 4.8.4 but causes ICE on the recent gfortran SVN 5.1 branch. 
This is not my own code and I know little of fortran programming so apologies
if the code is ugly or if it is not a minimal test case.

Compiler version causing crash:

wbrisken@shiitake sked> gfortran -v
Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/export/home/shiitake/gcc/libexec/gcc/x86_64-unknown-linux-gnu/5.1.1/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc/configure --prefix=/export/home/shiitake/gcc/
Thread model: posix
gcc version 5.1.1 20150523 (GCC)

System type:

Linux RHEL 6.6 running on x86-64 CPU

GCC build options:

../gcc/configure --prefix=/export/home/shiitake/gcc/

Compile command: 

gfortran -save-temps -finit-local-zero  -fno-automatic -fno-underscoring -C -w
-g -O2 -c -o crash.o crash3.f

Compiler output to screen:

f951: internal compiler error: Segmentation fault
0xa7f86f crash_signal
        ../../gcc/gcc/toplev.c:383
0x6a08c2 apply_default_init_local
        ../../gcc/gcc/fortran/resolve.c:10951
0x6a08c2 resolve_fl_variable
        ../../gcc/gcc/fortran/resolve.c:11213
0x6a08c2 resolve_symbol
        ../../gcc/gcc/fortran/resolve.c:13756
0x6b6f0b do_traverse_symtree
        ../../gcc/gcc/fortran/symbol.c:3646
0x6a16f2 resolve_types
        ../../gcc/gcc/fortran/resolve.c:14977
0x69d320 gfc_resolve(gfc_namespace*)
        ../../gcc/gcc/fortran/resolve.c:15081
0x688c46 gfc_parse_file()
        ../../gcc/gcc/fortran/parse.c:5476
0x6c83d5 gfc_be_parse_file
        ../../gcc/gcc/fortran/f95-lang.c:228
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.

Preprocessed output file:

Not generated.  Perhaps Seg Fault happened too early.



I hope this bug report is useful and complete.

Cheers,

Walter

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-06-05
                 CC|                            |kargl at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from kargl at gcc dot gnu.org ---
(In reply to Walter Brisken from comment #0)
>
> gfortran -save-temps -finit-local-zero  -fno-automatic -fno-underscoring -C
> -w -g -O2 -c -o crash.o crash3.f
> 

The problem is related to the -finit-local-zero option.
The obvious workaround is to not use that option.  I'll
also recommend not using the -fno-underscoring option
as the documentation warns about its use.

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

--- Comment #2 from kargl at gcc dot gnu.org ---
(In reply to kargl from comment #1)
> (In reply to Walter Brisken from comment #0)
> >
> > gfortran -save-temps -finit-local-zero  -fno-automatic -fno-underscoring -C
> > -w -g -O2 -c -o crash.o crash3.f
> > 
> 
> The problem is related to the -finit-local-zero option.
> The obvious workaround is to not use that option.  I'll
> also recommend not using the -fno-underscoring option
> as the documentation warns about its use.

Dang misread my terminal output.  -fno-automatic is the
guilty option.  Drop that option and things compile.

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|major                       |normal

--- Comment #3 from kargl at gcc dot gnu.org ---
(In reply to kargl from comment #2)
> (In reply to kargl from comment #1)
> > (In reply to Walter Brisken from comment #0)
> > >
> > > gfortran -save-temps -finit-local-zero  -fno-automatic -fno-underscoring -C
> > > -w -g -O2 -c -o crash.o crash3.f
> > > 
> > 
> > The problem is related to the -finit-local-zero option.
> > The obvious workaround is to not use that option.  I'll
> > also recommend not using the -fno-underscoring option
> > as the documentation warns about its use.
> 
> Dang misread my terminal output.  -fno-automatic is the
> guilty option.  Drop that option and things compile.

Clearly not my day!  One needs both  -finit-local-zero
and -fno-automatic to cause the ICE.  Time for more coffee!

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

--- Comment #4 from kargl at gcc dot gnu.org ---
Reduced testcase.

      block data

      integer n
      parameter (n=1)

      character*2 s1(n)
      character*8 s2(n)

      data (s1(i),s2(i),i=1,n)/"ab","12345678"/

      end

% gfc6 -c -fno-automatic -finit-local-zero block.f
f951: internal compiler error: Segmentation fault
0xb4971f crash_signal
        ../../gcc6/gcc/toplev.c:380
0x62abf3 apply_default_init_local
        ../../gcc6/gcc/fortran/resolve.c:10952
0x62abf3 resolve_fl_variable
        ../../gcc6/gcc/fortran/resolve.c:11214
0x62abf3 resolve_symbol
        ../../gcc6/gcc/fortran/resolve.c:13757
0x64155b do_traverse_symtree
        ../../gcc6/gcc/fortran/symbol.c:3646
0x62c002 resolve_types
        ../../gcc6/gcc/fortran/resolve.c:14978
0x627cc0 gfc_resolve(gfc_namespace*)
        ../../gcc6/gcc/fortran/resolve.c:15082
0x613889 gfc_parse_file()
        ../../gcc6/gcc/fortran/parse.c:5475
0x6536a5 gfc_be_parse_file
        ../../gcc6/gcc/fortran/f95-lang.c:225

[Bug fortran/66347] Seg fault (ICE) on compile

[sgk at troutmask dot apl.washington.edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

--- Comment #5 from Steve Kargl <sgk at troutmask dot apl.washington.edu> ---
On Fri, Jun 05, 2015 at 05:53:24PM +0000, kargl at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347
> 
> --- Comment #4 from kargl at gcc dot gnu.org ---
> Reduced testcase.
> 
>       block data
>       integer n
>       parameter (n=1)
>       character*2 s1(n)
>       character*8 s2(n)

        common /foo/ s1, s2
> 
>       data (s1(i),s2(i),i=1,n)/"ab","12345678"/
>       end

4.8.x and 4.9.x flag the missing common statement.

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
      Known to work|                            |4.8.4, 4.9.2
           Assignee|unassigned at gcc dot gnu.org      |kargl at gcc dot gnu.org
      Known to fail|                            |5.1.1, 6.0

--- Comment #6 from kargl at gcc dot gnu.org ---
I have a patch

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

--- Comment #7 from kargl at gcc dot gnu.org ---
Author: kargl
Date: Fri Jun  5 23:01:22 2015
New Revision: 224175

URL: https://gcc.gnu.org/viewcvs?rev=224175&root=gcc&view=rev
Log:
2015-06-05  Steven G. Kargl  <kargl@gcc.gnu.org>

        PR fortran/66347
        * resolve.c (apply_default_init_local): Do not dereference a NULL
        pointer.

2015-06-05  Steven G. Kargl  <kargl@gcc.gnu.org>

        PR fortran/66347
        * gfortran.dg/blockdata_9.f: New test.

Added:
    trunk/gcc/testsuite/gfortran.dg/blockdata_9.f
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/resolve.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

--- Comment #8 from kargl at gcc dot gnu.org ---
Author: kargl
Date: Fri Jun  5 23:16:57 2015
New Revision: 224176

URL: https://gcc.gnu.org/viewcvs?rev=224176&root=gcc&view=rev
Log:
2015-06-05  Steven G. Kargl  <kargl@gcc.gnu.org>

        PR fortran/66347
        * resolve.c (apply_default_init_local): Do not dereference a NULL
        pointer.

2015-06-05  Steven G. Kargl  <kargl@gcc.gnu.org>

        PR fortran/66347
        * gfortran.dg/blockdata_9.f: New test.

Added:
    branches/gcc-5-branch/gcc/testsuite/gfortran.dg/blockdata_9.f
Modified:
    branches/gcc-5-branch/gcc/fortran/ChangeLog
    branches/gcc-5-branch/gcc/fortran/resolve.c
    branches/gcc-5-branch/gcc/testsuite/ChangeLog

[Bug fortran/66347] Seg fault (ICE) on compile

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66347

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|---                         |5.2

--- Comment #9 from kargl at gcc dot gnu.org ---
Fixed on trunk and 5-branch.  Thanks for the bug report.