[Bug tree-optimization/71857] [7 Regression] gcc.dg/tree-ssa/ivopt_mult_4.c FAILs

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71857

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |amker.cheng at gmail dot com

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
The difference is caused by increased average number of loop iterations (from 5
to 10), which caused a different selection of candidates:

before:
Selected IV set for loop 1 at
/home/marxin/Programming/gcc/gcc/testsuite/gcc.dg/tree-ssa/ivopt_mult_4.c:19, 5
avg niters, 0 expressions, 2 IVs:
Candidate 7:
  Var befor: p2_27
  Var after: p2_14
  Incr POS: orig biv
  IV struct:
    Type:       long int *
    Base:       p2_11(D) + 8
    Step:       8
    Object:     (void *) p2_11(D)
    Biv:        N
    Overflowness wrto loop niter:       Overflow
Candidate 9:
  Var befor: i_28
  Var after: i_15
  Incr POS: orig biv
  IV struct:
    Type:       long unsigned int
    Base:       16
    Step:       16
    Biv:        N
    Overflowness wrto loop niter:       No-overflow

after:
Selected IV set for loop 1 at
/home/marxin/Programming/gcc/gcc/testsuite/gcc.dg/tree-ssa/ivopt_mult_4.c:19,
10 avg niters, 0 expressions, 2 IVs:
Candidate 7:
  Var befor: p2_27
  Var after: p2_14
  Incr POS: orig biv
  IV struct:
    Type:       long int *
    Base:       p2_11(D) + 8
    Step:       8
    Object:     (void *) p2_11(D)
    Biv:        N
    Overflowness wrto loop niter:       Overflow
Candidate 10:
  Var befor: ivtmp.11_21
  Var after: ivtmp.11_29
  Incr POS: before exit test
  IV struct:
    Type:       unsigned long
    Base:       (unsigned long) (p_16(D) + 128)
    Step:       128
    Object:     (void *) p_16(D)
    Biv:        N
    Overflowness wrto loop niter:       Overflow

Replacing exit test: if (_22 > i_28)

Should I just amend the dump scan to cover the 'Replacing exit test' ?