[Bug fortran/95939] New: ice with -O3 in compute_live_loop_exits

[Bug fortran/95939] New: ice with -O3 in compute_live_loop_exits

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

            Bug ID: 95939
           Summary: ice with -O3 in compute_live_loop_exits
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

Recent gcc trunk gfortran compiler does this:

$ /home/dcb/gcc/results/bin/gfortran -c -O3 ./gfortran.dg/graphite/pr42285.f90
during GIMPLE pass: vect
./gfortran.dg/graphite/pr42285.f90:3:0:

    3 | SUBROUTINE EFGRDM(NCF,NFRG,G,RTRMS,GM,IOPT,K1)
      | 
internal compiler error: in compute_live_loop_exits, at
tree-ssa-loop-manip.c:247
0x6d94fd compute_live_loop_exits
        ../../trunk.git/gcc/tree-ssa-loop-manip.c:247
0x6d94fd add_exit_phis_var
        ../../trunk.git/gcc/tree-ssa-loop-manip.c:334
0x6d94fd add_exit_phis
        ../../trunk.git/gcc/tree-ssa-loop-manip.c:356
0x6d94fd rewrite_into_loop_closed_ssa_1(bitmap_head*, unsigned int, int, loop*)
        ../../trunk.git/gcc/tree-ssa-loop-manip.c:678

This seems to have gone wrong sometime between date 20200624
and 20200625.

It is often worthwhile to run all the Fortran code in the testsuite 
through the compiler with -O3 or -O3 -march=native.

[Bug fortran/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

--- Comment #1 from David Binderman <dcb314 at hotmail dot com> ---
Same error in testsuite files gcc.dg/tree-ssa/pr23433.c
and gcc.target/aarch64/sve/reduc_strict_8.c.

I'll have a go at reducing the first.

[Bug fortran/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

--- Comment #2 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to David Binderman from comment #1)
> I'll have a go at reducing the first.

Pointless. It is only a few lines of code anyway.

[Bug middle-end/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |kargl at gcc dot gnu.org
          Component|fortran                     |middle-end

--- Comment #3 from kargl at gcc dot gnu.org ---
Change component as this appears to not be Fortran specific.

[Bug tree-optimization/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
   Target Milestone|---                         |11.0
          Component|middle-end                  |tree-optimization

[Bug tree-optimization/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2020-06-29
             Status|UNCONFIRMED                 |WAITING
     Ever confirmed|0                           |1
                 CC|                            |marxin at gcc dot gnu.org

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
I can't reproduce it with current master.

[Bug tree-optimization/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

--- Comment #5 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to Martin Liška from comment #4)
> I can't reproduce it with current master.

Agreed. It seems to have gone away by 20200628.

[Bug tree-optimization/95939] [11 Regression] ice with -O3 in compute_live_loop_exits

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95939

Richard Biener <rguenth at gcc dot gnu.org> changed:

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