[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH 08/14] i386: do not compute eflags multiple time
From: |
Blue Swirl |
Subject: |
Re: [Qemu-devel] [PATCH 08/14] i386: do not compute eflags multiple times consecutively |
Date: |
Sun, 7 Oct 2012 19:09:30 +0000 |
On Sat, Oct 6, 2012 at 12:30 PM, Paolo Bonzini <address@hidden> wrote:
> After calling gen_compute_eflags, leave the computed value in cc_reg_src
> and set cc_op to CC_OP_EFLAGS. The next few patches will remove anyway
> most calls to gen_compute_eflags.
>
> As a result of this change it is more natural to remove the register
> argument from gen_compute_eflags and change all the callers.
>
> Signed-off-by: Paolo Bonzini <address@hidden>
Reviewed-by: Blue Swirl <address@hidden>
> ---
> target-i386/translate.c | 73
> ++++++++++++++++++++++++-------------------------
> 1 file modificato, 36 inserzioni(+), 37 rimozioni(-)
>
> diff --git a/target-i386/translate.c b/target-i386/translate.c
> index 8f22119..09512c3 100644
> --- a/target-i386/translate.c
> +++ b/target-i386/translate.c
> @@ -834,48 +834,49 @@ static void gen_compute_eflags_c(DisasContext *s, TCGv
> reg)
> }
>
> /* compute all eflags to cc_src */
> -static void gen_compute_eflags(DisasContext *s, TCGv reg)
> +static void gen_compute_eflags(DisasContext *s)
> {
> if (s->cc_op != CC_OP_DYNAMIC) {
> gen_op_set_cc_op(s->cc_op);
> }
> - gen_helper_cc_compute_all(cpu_tmp2_i32, cpu_env, cpu_cc_op);
> - if (reg == cpu_cc_src) {
> - tcg_gen_discard_tl(cpu_cc_dst);
> - s->cc_op = CC_OP_EFLAGS;
> + if (s->cc_op == CC_OP_EFLAGS) {
> + return;
> }
> - tcg_gen_extu_i32_tl(reg, cpu_tmp2_i32);
> + gen_helper_cc_compute_all(cpu_tmp2_i32, cpu_env, cpu_cc_op);
> + tcg_gen_discard_tl(cpu_cc_dst);
> + s->cc_op = CC_OP_EFLAGS;
> + tcg_gen_extu_i32_tl(cpu_cc_src, cpu_tmp2_i32);
> }
>
> /* compute eflags.P to reg */
> static void gen_compute_eflags_p(DisasContext *s, TCGv reg)
> {
> - gen_compute_eflags(s, reg);
> - tcg_gen_shri_tl(reg, reg, 2);
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(reg, cpu_cc_src, 2);
> tcg_gen_andi_tl(reg, reg, 1);
> }
>
> /* compute eflags.S to reg */
> static void gen_compute_eflags_s(DisasContext *s, TCGv reg)
> {
> - gen_compute_eflags(s, reg);
> - tcg_gen_shri_tl(reg, reg, 7);
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(reg, cpu_cc_src, 7);
> tcg_gen_andi_tl(reg, reg, 1);
> }
>
> /* compute eflags.O to reg */
> static void gen_compute_eflags_o(DisasContext *s, TCGv reg)
> {
> - gen_compute_eflags(s, reg);
> - tcg_gen_shri_tl(reg, reg, 11);
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(reg, cpu_cc_src, 11);
> tcg_gen_andi_tl(reg, reg, 1);
> }
>
> /* compute eflags.Z to reg */
> static void gen_compute_eflags_z(DisasContext *s, TCGv reg)
> {
> - gen_compute_eflags(s, reg);
> - tcg_gen_shri_tl(reg, reg, 6);
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(reg, cpu_cc_src, 6);
> tcg_gen_andi_tl(reg, reg, 1);
> }
>
> @@ -892,9 +893,9 @@ static inline void gen_setcc_slow_T0(DisasContext *s, int
> jcc_op)
> gen_compute_eflags_z(s, cpu_T[0]);
> break;
> case JCC_BE:
> - gen_compute_eflags(s, cpu_tmp0);
> - tcg_gen_shri_tl(cpu_T[0], cpu_tmp0, 6);
> - tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(cpu_T[0], cpu_cc_src, 6);
> + tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_cc_src);
> tcg_gen_andi_tl(cpu_T[0], cpu_T[0], 1);
> break;
> case JCC_S:
> @@ -904,18 +905,18 @@ static inline void gen_setcc_slow_T0(DisasContext *s,
> int jcc_op)
> gen_compute_eflags_p(s, cpu_T[0]);
> break;
> case JCC_L:
> - gen_compute_eflags(s, cpu_tmp0);
> - tcg_gen_shri_tl(cpu_T[0], cpu_tmp0, 11); /* CC_O */
> - tcg_gen_shri_tl(cpu_tmp0, cpu_tmp0, 7); /* CC_S */
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(cpu_T[0], cpu_cc_src, 11); /* CC_O */
> + tcg_gen_shri_tl(cpu_tmp0, cpu_cc_src, 7); /* CC_S */
> tcg_gen_xor_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
> tcg_gen_andi_tl(cpu_T[0], cpu_T[0], 1);
> break;
> default:
> case JCC_LE:
> - gen_compute_eflags(s, cpu_tmp0);
> - tcg_gen_shri_tl(cpu_T[0], cpu_tmp0, 11); /* CC_O */
> - tcg_gen_shri_tl(cpu_tmp4, cpu_tmp0, 7); /* CC_S */
> - tcg_gen_shri_tl(cpu_tmp0, cpu_tmp0, 6); /* CC_Z */
> + gen_compute_eflags(s);
> + tcg_gen_shri_tl(cpu_T[0], cpu_cc_src, 11); /* CC_O */
> + tcg_gen_shri_tl(cpu_tmp4, cpu_cc_src, 7); /* CC_S */
> + tcg_gen_shri_tl(cpu_tmp0, cpu_cc_src, 6); /* CC_Z */
> tcg_gen_xor_tl(cpu_T[0], cpu_T[0], cpu_tmp4);
> tcg_gen_or_tl(cpu_T[0], cpu_T[0], cpu_tmp0);
> tcg_gen_andi_tl(cpu_T[0], cpu_T[0], 1);
> @@ -1614,7 +1615,7 @@ static void gen_rot_rm_T1(DisasContext *s, int ot, int
> op1,
> }
>
> /* update eflags. It is needed anyway most of the time, do it always.
> */
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> assert(s->cc_op == CC_OP_EFLAGS);
>
> label2 = gen_new_label();
> @@ -1691,7 +1692,7 @@ static void gen_rot_rm_im(DisasContext *s, int ot, int
> op1, int op2,
>
> if (op2 != 0) {
> /* update eflags */
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> assert(s->cc_op == CC_OP_EFLAGS);
>
> tcg_gen_andi_tl(cpu_cc_src, cpu_cc_src, ~(CC_O | CC_C));
> @@ -1717,9 +1718,7 @@ static void gen_rotc_rm_T1(DisasContext *s, int ot, int
> op1,
> {
> int label1;
>
> - if (s->cc_op != CC_OP_DYNAMIC)
> - gen_op_set_cc_op(s->cc_op);
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
>
> /* load */
> if (op1 == OR_TMP0)
> @@ -6512,7 +6511,7 @@ static target_ulong disas_insn(DisasContext *s,
> target_ulong pc_start)
> if (CODE64(s) && !(s->cpuid_ext3_features & CPUID_EXT3_LAHF_LM))
> goto illegal_op;
> gen_op_mov_TN_reg(OT_BYTE, 0, R_AH);
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> tcg_gen_andi_tl(cpu_cc_src, cpu_cc_src, CC_O);
> tcg_gen_andi_tl(cpu_T[0], cpu_T[0], CC_S | CC_Z | CC_A | CC_P |
> CC_C);
> tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, cpu_T[0]);
> @@ -6520,21 +6519,21 @@ static target_ulong disas_insn(DisasContext *s,
> target_ulong pc_start)
> case 0x9f: /* lahf */
> if (CODE64(s) && !(s->cpuid_ext3_features & CPUID_EXT3_LAHF_LM))
> goto illegal_op;
> - gen_compute_eflags(s, cpu_T[0]);
> + gen_compute_eflags(s);
> /* Note: gen_compute_eflags() only gives the condition codes */
> - tcg_gen_ori_tl(cpu_T[0], cpu_T[0], 0x02);
> + tcg_gen_ori_tl(cpu_T[0], cpu_cc_src, 0x02);
> gen_op_mov_reg_T0(OT_BYTE, R_AH);
> break;
> case 0xf5: /* cmc */
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> tcg_gen_xori_tl(cpu_cc_src, cpu_cc_src, CC_C);
> break;
> case 0xf8: /* clc */
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> tcg_gen_andi_tl(cpu_cc_src, cpu_cc_src, ~CC_C);
> break;
> case 0xf9: /* stc */
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> tcg_gen_ori_tl(cpu_cc_src, cpu_cc_src, CC_C);
> break;
> case 0xfc: /* cld */
> @@ -6889,7 +6888,7 @@ static target_ulong disas_insn(DisasContext *s,
> target_ulong pc_start)
> case 1: /* loopz */
> gen_op_add_reg_im(s->aflag, R_ECX, -1);
> gen_op_jz_ecx(s->aflag, l3);
> - gen_compute_eflags(s, cpu_tmp0);
> + gen_compute_eflags(s);
> tcg_gen_andi_tl(cpu_tmp0, cpu_tmp0, CC_Z);
> if (b == 0) {
> tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_tmp0, 0, l1);
> @@ -7431,7 +7430,7 @@ static target_ulong disas_insn(DisasContext *s,
> target_ulong pc_start)
> } else {
> gen_op_mov_reg_v(ot, rm, t0);
> }
> - gen_compute_eflags(s, cpu_cc_src);
> + gen_compute_eflags(s);
> tcg_gen_andi_tl(cpu_cc_src, cpu_cc_src, ~CC_Z);
> tcg_gen_or_tl(cpu_cc_src, cpu_cc_src, t2);
> tcg_temp_free(t0);
> --
> 1.7.12.1
>
>
>
- [Qemu-devel] [PATCH 07/14] i386: add helper functions to get other flags, (continued)
- [Qemu-devel] [PATCH 07/14] i386: add helper functions to get other flags, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 06/14] i386: factor gen_op_set_cc_op/tcg_gen_discard_tl around computing flags, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 09/14] i386: do not call helper to compute ZF/SF, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 08/14] i386: do not compute eflags multiple times consecutively, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 05/14] i386: move eflags computation closer to gen_op_set_cc_op, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 10/14] i386: use inverted setcond when computing NS or NZ, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 13/14] i386: optimize setbe, Paolo Bonzini, 2012/10/06
- [Qemu-devel] [PATCH 12/14] i386: change gen_setcc_slow_T0 to gen_setcc_slow, Paolo Bonzini, 2012/10/06