[Bug tree-optimization/55079] [4.8 regression] false positive -Warray-bounds (also seen at -O3 bootstrap)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55079

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
         AssignedTo|unassigned at gcc dot       |rguenth at gcc dot gnu.org
                   |gnu.org                     |

--- Comment #11 from Richard Biener <rguenth at gcc dot gnu.org> 2012-12-10 13:23:17 UTC ---
Created attachment 28911
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=28911
prototype patch

The pattern we have is

  <bb 6>:
  _36 = i_33 + 1;
  _37 = a[i_33];
  a[_36] = _37;
  i_39 = i_33 + 4294967295;
  if (i_33 != 0)
    goto <bb 7>;
  else
    goto <bb 11>;

  <bb 7>:
  _42 = i_39 + 1;
  _43 = a[i_39];
  a[_42] = _43;

and eventually adding an assert in bb7 that i_39 != 1 would help.  But
the only thing we try to add extra asserts for is stuff in the definition
chain of comparison operands ... this OTOH is for stuff that uses
comparison operands and live on the edge.

Prototype patch attached, fixes comment#8 at least.