[Bug middle-end/66630] New: Missing ubsan/ftrapv error

[Bug middle-end/66630] New: Missing ubsan/ftrapv error

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66630

            Bug ID: 66630
           Summary: Missing ubsan/ftrapv error
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: mpolacek at gcc dot gnu.org
  Target Milestone: ---

int
main (void)
{
  int i = -__INT_MAX__ - 1;
  int j = -i - 1;
  return j;
}

$ xgcc -fsanitize=undefined a.c; ./a.out

says nothing :(.  We should report the overflow when computing -i.

[Bug middle-end/66630] Missing ubsan/ftrapv error

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66630

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This might be because we convert -i - 1 to be ~i.

[Bug middle-end/66630] Missing ubsan/ftrapv error

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66630

--- Comment #2 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
That's one thing.  But there also something else going on.  I hope it's just
missing TYPE_OVERFLOW_SANITIZED in some match.pd patterns.

[Bug middle-end/66630] Missing ubsan/ftrapv error

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66630

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
For -ftrapv I never got around to commit some of the fixes I posted last
year...
(guard CCP properly for example).  Eventually at least CCP should be more
aggressive and simplify trapping stuff to __builtin_trap () (but the optab
implementations in libgcc use abort ()).

[Bug middle-end/66630] Missing ubsan/ftrapv error

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66630

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vincent-gcc at vinc17 dot net

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 87311 has been marked as a duplicate of this bug. ***

[Bug middle-end/66630] Missing ubsan/ftrapv error

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66630

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Marek Polacek from comment #2)
> That's one thing.  But there also something else going on.  I hope it's just
> missing TYPE_OVERFLOW_SANITIZED in some match.pd patterns.

```
  /* ~A + A -> -1 */
  (simplify
   (plus:c (convert? (bit_not @0)) (convert? @0))
   (if (!TYPE_OVERFLOW_TRAPS (type))
    (convert { build_all_ones_cst (TREE_TYPE (@0)); })))

  /* ~A + 1 -> -A */
  (simplify
   (plus (convert? (bit_not @0)) integer_each_onep)
   (if (tree_nop_conversion_p (type, TREE_TYPE (@0)))
    (negate (convert @0))))

  /* -A - 1 -> ~A */
  (simplify
   (minus (convert? (negate @0)) integer_each_onep)
   (if (!TYPE_OVERFLOW_TRAPS (type)
        && TREE_CODE (type) != COMPLEX_TYPE
        && tree_nop_conversion_p (type, TREE_TYPE (@0)))
    (bit_not (convert @0))))

  /* -1 - A -> ~A */
  (simplify
   (minus integer_all_onesp @0)
   (if (TREE_CODE (type) != COMPLEX_TYPE)
    (bit_not @0)))

```

I am not sure which one of these patterns need the check for
TYPE_OVERFLOW_SANITIZED though.