http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55876
--- Comment #1 from vries at gcc dot gnu.org 2013-01-04 15:23:25 UTC --- while expanding the xor, we try to widen an DImode reg from QImode to SImode: ... #6 0x085d7da5 in widen_operand (op=0xf7cd2ec0, mode=SImode, oldmode=QImode, unsignedp=1, no_extend=1) at src/gcc-mainline/gcc/optabs.c:333 333 return gen_rtx_SUBREG (mode, force_reg (GET_MODE (op), op), 0); (gdb) l 328 return convert_modes (mode, oldmode, op, unsignedp); 329 330 /* If MODE is no wider than a single word, we return a paradoxical 331 SUBREG. */ 332 if (GET_MODE_SIZE (mode) <= UNITS_PER_WORD) 333 return gen_rtx_SUBREG (mode, force_reg (GET_MODE (op), op), 0); 334 335 /* Otherwise, get an object of MODE, clobber it, and set the low-order 336 part to OP. */ 337 (gdb) call debug_rtx (op) (reg:DI 222 [ used+-7 ]) ... Just before expand, the representation looks like: ... foo () { unsigned charD.13 usedD.1407; charD.2 _7; unsigned charD.13 _13; charD.2 _19; charD.2 _28; charD.2 _37; ;; basic block 2, loop depth 0 ;; pred: ENTRY # VUSE <.MEM_1(D)> _19 = MEM[(const charD.2 *)&bufferD.1387]; used_20 = _19 != 70 ? 1 : 0; # VUSE <.MEM_1(D)> _28 = MEM[(const charD.2 *)&bufferD.1387 + 1B]; used_29 = _28 == 70 ? used_20 : 1; # VUSE <.MEM_1(D)> _37 = MEM[(const charD.2 *)&bufferD.1387 + 2B]; used_38 = _37 == 70 ? used_29 : 1; # VUSE <.MEM_1(D)> _7 = MEM[(const charD.2 *)&bufferD.1387 + 3B]; used_10 = _7 == 70 ? used_38 : 1; _13 = used_10 ^ 1; # VUSE <.MEM_1(D)> return _13; ;; succ: EXIT } ... The reason that used_10 is in a DImode reg is because used_10 = _7 == 70 ? used_38 : 1 and the all the other conditional moves before are promoted by expand_cond_expr_using_cmove.