I haven't mentioned that I am using gcc 4.6 latest version.
To generalize the question. If I use an operand like lc_operand
(below) and leave the constraint open, is it guaranteed that the
register that would be chosen would be of class lc?

2010/8/3 roy rosen <roy.1ro...@gmail.com>:
> Hi All,
>
> If I don't use a constraint, is it possible that during ira I get a
> register which is not acceptable by the predicate?
> In my port I have the following to support HW loops:
>
> (define_predicate "lc_operand"
>    (match_operand 0 "register_operand")
> {
>        unsigned int regno;
>        if (GET_CODE (op) == SUBREG)
>            op = SUBREG_REG (op);
>        regno = REGNO (op);
>        return (regno >= FIRST_PSEUDO_REGISTER ||
> REGNO_REG_CLASS(regno) == LC_REGS);
> }
> )
>
> (define_expand "doloop_end"
>  [(use (match_operand 0 "" ""))        ; loop pseudo
>   (use (match_operand 1 "" ""))        ; iterations; zero if unknown
>   (use (match_operand 2 "" ""))        ; max iterations
>   (use (match_operand 3 "" ""))        ; loop level
>   (use (match_operand 4 "" ""))]       ; label
>  ""
>  {
>    if (INTVAL (operands[3]) > 4
>    )
>        FAIL;
>    emit_jump_insn (gen_doloop_end_internal (operands[4], operands[0]));
>    DONE;
>  }
> )
>
> (define_insn "doloop_end_internal"
>  [(set (pc)
>        (if_then_else (ne (match_operand:SI 1 "lc_operand" "")
>                          (const_int 0))
>                      (label_ref (match_operand 0 "" ""))
>                      (pc)))
>   (set (match_dup 1) (plus:SI (match_dup 1) (const_int -1)))]
>  ""
>  "doloop_end%b1 %!"
> )
>
> I see that during ira operand 1 in doloop_end_internal is getting
> "best GENERAL_REGS" and eventually ends up as a register of a class
> other than LC_REGS.
> I thought that if I have predicate lc_operand then I don't need a
> constraint since all other register classes would be rejected in all
> stages for this operand.
> What might be the problem here?
>
> Thanks, Roy.
>

Reply via email to