diff --git a/tcg/optimize.c b/tcg/optimize.c index bda727d5ed..dbb03ef96b 100644 --- a/tcg/optimize.c +++ b/tcg/optimize.c @@ -1120,11 +1120,10 @@ void tcg_optimize(TCGContext *s) case INDEX_op_dup2_vec: assert(TCG_TARGET_REG_BITS == 32); if (arg_is_const(op->args[1]) && arg_is_const(op->args[2])) { - tmp = arg_info(op->args[1])->val; - if (tmp == arg_info(op->args[2])->val) { - tcg_opt_gen_movi(s, &temps_used, op, op->args[0], tmp); - break; - } + tcg_opt_gen_movi(s, &temps_used, op, op->args[0], + deposit64(arg_info(op->args[1])->val, 32, 32, + arg_info(op->args[2])->val)); + break; } else if (args_are_copies(op->args[1], op->args[2])) { op->opc = INDEX_op_dup_vec; TCGOP_VECE(op) = MO_32; diff --git a/tcg/tcg-op-vec.c b/tcg/tcg-op-vec.c index cdbf11c573..9fbed1366c 100644 --- a/tcg/tcg-op-vec.c +++ b/tcg/tcg-op-vec.c @@ -216,25 +216,17 @@ void tcg_gen_mov_vec(TCGv_vec r, TCGv_vec a) } } -#define MO_REG (TCG_TARGET_REG_BITS == 64 ? MO_64 : MO_32) - -static void do_dupi_vec(TCGv_vec r, unsigned vece, TCGArg a) -{ - TCGTemp *rt = tcgv_vec_temp(r); - vec_gen_2(INDEX_op_dupi_vec, rt->base_type, vece, temp_arg(rt), a); -} - TCGv_vec tcg_const_zeros_vec(TCGType type) { TCGv_vec ret = tcg_temp_new_vec(type); - do_dupi_vec(ret, MO_REG, 0); + tcg_gen_dupi_vec(MO_64, ret, 0); return ret; } TCGv_vec tcg_const_ones_vec(TCGType type) { TCGv_vec ret = tcg_temp_new_vec(type); - do_dupi_vec(ret, MO_REG, -1); + tcg_gen_dupi_vec(MO_64, ret, -1); return ret; } @@ -252,39 +244,28 @@ TCGv_vec tcg_const_ones_vec_matching(TCGv_vec m) void tcg_gen_dup64i_vec(TCGv_vec r, uint64_t a) { - if (TCG_TARGET_REG_BITS == 64) { - do_dupi_vec(r, MO_64, a); - } else if (a == dup_const(MO_32, a)) { - do_dupi_vec(r, MO_32, a); - } else { - TCGv_i64 c = tcg_const_i64(a); - tcg_gen_dup_i64_vec(MO_64, r, c); - tcg_temp_free_i64(c); - } + tcg_gen_dupi_vec(MO_64, r, a); } void tcg_gen_dup32i_vec(TCGv_vec r, uint32_t a) { - do_dupi_vec(r, MO_REG, dup_const(MO_32, a)); + tcg_gen_dupi_vec(MO_32, r, a); } void tcg_gen_dup16i_vec(TCGv_vec r, uint32_t a) { - do_dupi_vec(r, MO_REG, dup_const(MO_16, a)); + tcg_gen_dupi_vec(MO_16, r, a); } void tcg_gen_dup8i_vec(TCGv_vec r, uint32_t a) { - do_dupi_vec(r, MO_REG, dup_const(MO_8, a)); + tcg_gen_dupi_vec(MO_8, r, a); } void tcg_gen_dupi_vec(unsigned vece, TCGv_vec r, uint64_t a) { - if (vece == MO_64) { - tcg_gen_dup64i_vec(r, a); - } else { - do_dupi_vec(r, MO_REG, dup_const(vece, a)); - } + TCGTemp *rt = tcgv_vec_temp(r); + tcg_gen_mov_vec(r, tcg_constant_vec(rt->base_type, vece, a)); } void tcg_gen_dup_i64_vec(unsigned vece, TCGv_vec r, TCGv_i64 a) @@ -489,8 +470,8 @@ void tcg_gen_abs_vec(unsigned vece, TCGv_vec r, TCGv_vec a) if (tcg_can_emit_vec_op(INDEX_op_sari_vec, type, vece) > 0) { tcg_gen_sari_vec(vece, t, a, (8 << vece) - 1); } else { - do_dupi_vec(t, MO_REG, 0); - tcg_gen_cmp_vec(TCG_COND_LT, vece, t, a, t); + tcg_gen_cmp_vec(TCG_COND_LT, vece, t, a, + tcg_constant_vec(type, vece, 0)); } tcg_gen_xor_vec(vece, r, a, t); tcg_gen_sub_vec(vece, r, r, t); diff --git a/tcg/tcg.c b/tcg/tcg.c index 802f0b8a32..ad1348d811 100644 --- a/tcg/tcg.c +++ b/tcg/tcg.c @@ -3539,16 +3539,11 @@ static void temp_load(TCGContext *s, TCGTemp *ts, TCGRegSet desired_regs, * The targets will, in general, have to do this search anyway, * do this generically. */ - if (TCG_TARGET_REG_BITS == 32) { - val = dup_const(MO_32, val); - vece = MO_32; - } if (val == dup_const(MO_8, val)) { vece = MO_8; } else if (val == dup_const(MO_16, val)) { vece = MO_16; - } else if (TCG_TARGET_REG_BITS == 64 && - val == dup_const(MO_32, val)) { + } else if (val == dup_const(MO_32, val)) { vece = MO_32; }