This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: MIPS fix testcase 20001226-1.c branch shortening.


On Thu, Jul 26, 2001 at 09:01:57PM +0100, Graham Stott wrote:
> +   else if (INSN_P (insn) && GET_CODE (PATTERN (insn)) == SET)

single_set?

> +       register rtx next_insn = NEXT_INSN (insn);
> +
> +      while (next_insn != 0 && ! INSN_P (next_insn))
> +        next_insn = NEXT_INSN (next_insn);

You want next_active_insn.

> +      if (next_insn && set)

Why would set be null?

> +        while (GET_CODE (set) == SUBREG)
> +          set = SUBREG_REG (set);

There shouldn't be any subregs after reload.

> +                 && GET_CODE (XEXP (PATTERN (insn), 1)) == MEM)

This is clearer as get_attr_type (insn) == TYPE_LOAD.



r~


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]