[Bug rtl-optimization/46248] [4.6 regression] crash+infinite recursion in combine

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46248

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ebotcazou at gcc dot
                   |                            |gnu.org, jakub at gcc dot
                   |                            |gnu.org

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-01 10:34:53 UTC ---
      if (added_sets_1)
        {
          rtx t = i1pat;
          if (i0_feeds_i1_n)
            t = subst (t, i0dest, i0src, 0, 0);

          XVECEXP (newpat, 0, --total_sets) = t;
        }
      if (added_sets_2)
        {
          rtx t = i2pat;
          if (i0_feeds_i2_n)
            t = subst (t, i0dest, i0src, 0, 0);
          if (i1_feeds_i2_n)
            t = subst (t, i1dest, i1src, 0, 0);
          if (i0_feeds_i1_n && i1_feeds_i2_n)   <==============
            t = subst (t, i0dest, i0src, 0, 0); <==============

          XVECEXP (newpat, 0, --total_sets) = t;
        }

The two lines above are problematic, because subst may clobber the first
argument's expression and
      newpat = subst (newpat, i0dest, i0src, 0,
                      i0_feeds_i1_n && i0dest_in_i0src); 
actually clobbered not just newpat, but also i1src which was in it.  Not sure
if we actually can rely it always happened say if substed_i0 is true and not do
the i0dest -> i0src replacement in i2pat here.