combine: Don't mess with subregs of floating point (PR78590)

PR78590 shows a problem in change_zero_ext, where we change a zero_extend
of a subreg to a logical and.  We should only do this if the thing we are
taking the subreg of is a scalar integer, otherwise we will take a subreg
of (e.g.) a float in a different size, which is nonsensical and hits an
assert.


	PR rtl-optimization/78590
	* combine.c (change_zero_ext): Transform zero_extend of subregs only
	if the subreg_reg is a scalar integer mode.

From-SVN: r242999

gcc/ChangeLog

+2016-11-30  Segher Boessenkool  <segher@kernel.crashing.org>
+
+	PR rtl-optimization/78590
+	* combine.c (change_zero_ext): Transform zero_extend of subregs only
+	if the subreg_reg is a scalar integer mode.
+
 2016-11-30  Jakub Jelinek  <jakub@redhat.com>
 
 	PR tree-optimization/78586

gcc/combine.c

@@ -11251,6 +11251,7 @@ change_zero_ext (rtx pat)
       else if (GET_CODE (x) == ZERO_EXTEND
 	       && SCALAR_INT_MODE_P (mode)
 	       && GET_CODE (XEXP (x, 0)) == SUBREG
+	       && SCALAR_INT_MODE_P (GET_MODE (SUBREG_REG (XEXP (x, 0))))
 	       && !paradoxical_subreg_p (XEXP (x, 0))
 	       && subreg_lowpart_p (XEXP (x, 0)))
 	{