[Bug middle-end/55943] [4.6/4.7/4.8/4.9 Regression] ICE in gen_reg_rtx

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55943

John David Anglin <danglin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|ia64-*-*                    |ia64-*-*,hppa*64*-*-*
                 CC|                            |danglin at gcc dot gnu.org

--- Comment #4 from John David Anglin <danglin at gcc dot gnu.org> 2013-03-16 13:17:17 UTC ---
I believe the problem is in this hunk of code:

  /* See if we can coerce the target into moving both values at once.  */

  /* Move floating point as parts.  */
  if (GET_MODE_CLASS (mode) == MODE_COMPLEX_FLOAT
      && can_create_pseudo_p ()
      && optab_handler (mov_optab, GET_MODE_INNER (mode)) != CODE_FOR_nothing)
    try_int = false;
  /* Not possible if the values are inherently not adjacent.  */
  else if (GET_CODE (x) == CONCAT || GET_CODE (y) == CONCAT)
    try_int = false;
  /* Is possible if both are registers (or subregs of registers).  */
  else if (register_operand (x, mode) && register_operand (y, mode))
    try_int = true;
  /* If one of the operands is a memory, and alignment constraints
     are friendly enough, we may be able to do combined memory operations.
     We do not attempt this if Y is a constant because that combination is
     usually better with the by-parts thing below.  */
  else if ((MEM_P (x) ? !CONSTANT_P (y) : MEM_P (y))
           && (!STRICT_ALIGNMENT
               || get_mode_alignment (mode) == BIGGEST_ALIGNMENT))
    try_int = true;
  else
    try_int = false;

When the inner mode is smaller than a word, a pseudo is needed to extract
the value.  The asm causes this to occur after reload starts, causing the
ICE.