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]
Other format: [Raw text]

Re: [patch] s390 support for tpf os


On Tue, May 27, 2003 at 01:17:34PM -0700, Eric Christopher wrote:
> +      addr = gen_rtx_CONST_INT (HImode, 0xfe0);
> +      unkn = gen_rtx_CONST_INT (SImode, 0);

GEN_INT and const0_rtx respectively.

> @@ -5513,6 +5541,33 @@ s390_emit_epilogue ()
> +      if (!TARGET_64BIT)
> +	emit_insn (gen_bas_31 (addr, unkn, link));
> +      else
> +	emit_insn (gen_bas_64 (addr, unkn, link));
> +
> +      emit_insn (gen_blockage ());

Um, shouldn't the blockage go before the bas in the epilogue?

> +  emit_insn (gen_zero_extendsidi2 (operands[0], operands[1]));
> +  emit_insn (gen_anddi3 (operands[0], operands[0], GEN_INT (0x7fffffff)));

Does it work as well to emit

	gen_anddi3 (operands[0],
		    gen_lowpart (DImode, operands[1]),
		    GEN_INT (0x7fffffff));

(i.e. with a paradoxical subreg), or does that just confuse
the optimizers?


r~


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