diff mbox series

[v2,22/67] target/arm: Expand vfp neg and abs inline

Message ID 20240524232121.284515-23-richard.henderson@linaro.org
State New
Headers show
Series target/arm: Convert a64 advsimd to decodetree (part 1) | expand

Commit Message

Richard Henderson May 24, 2024, 11:20 p.m. UTC
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 target/arm/helper.h            |  6 ----
 target/arm/tcg/translate.h     | 30 +++++++++++++++++++
 target/arm/tcg/translate-a64.c | 44 +++++++++++++--------------
 target/arm/tcg/translate-vfp.c | 54 +++++++++++++++++-----------------
 target/arm/vfp_helper.c        | 30 -------------------
 5 files changed, 79 insertions(+), 85 deletions(-)
diff mbox series

Patch

diff --git a/target/arm/helper.h b/target/arm/helper.h
index 7ee15b9651..0fd01c9c52 100644
--- a/target/arm/helper.h
+++ b/target/arm/helper.h
@@ -132,12 +132,6 @@  DEF_HELPER_3(vfp_maxnumd, f64, f64, f64, ptr)
 DEF_HELPER_3(vfp_minnumh, f16, f16, f16, ptr)
 DEF_HELPER_3(vfp_minnums, f32, f32, f32, ptr)
 DEF_HELPER_3(vfp_minnumd, f64, f64, f64, ptr)
-DEF_HELPER_1(vfp_negh, f16, f16)
-DEF_HELPER_1(vfp_negs, f32, f32)
-DEF_HELPER_1(vfp_negd, f64, f64)
-DEF_HELPER_1(vfp_absh, f16, f16)
-DEF_HELPER_1(vfp_abss, f32, f32)
-DEF_HELPER_1(vfp_absd, f64, f64)
 DEF_HELPER_2(vfp_sqrth, f16, f16, env)
 DEF_HELPER_2(vfp_sqrts, f32, f32, env)
 DEF_HELPER_2(vfp_sqrtd, f64, f64, env)
diff --git a/target/arm/tcg/translate.h b/target/arm/tcg/translate.h
index ecfa242eef..b05a9eb668 100644
--- a/target/arm/tcg/translate.h
+++ b/target/arm/tcg/translate.h
@@ -406,6 +406,36 @@  static inline void gen_swstep_exception(DisasContext *s, int isv, int ex)
  */
 uint64_t vfp_expand_imm(int size, uint8_t imm8);
 
+static inline void gen_vfp_absh(TCGv_i32 d, TCGv_i32 s)
+{
+    tcg_gen_andi_i32(d, s, INT16_MAX);
+}
+
+static inline void gen_vfp_abss(TCGv_i32 d, TCGv_i32 s)
+{
+    tcg_gen_andi_i32(d, s, INT32_MAX);
+}
+
+static inline void gen_vfp_absd(TCGv_i64 d, TCGv_i64 s)
+{
+    tcg_gen_andi_i64(d, s, INT64_MAX);
+}
+
+static inline void gen_vfp_negh(TCGv_i32 d, TCGv_i32 s)
+{
+    tcg_gen_xori_i32(d, s, 1u << 15);
+}
+
+static inline void gen_vfp_negs(TCGv_i32 d, TCGv_i32 s)
+{
+    tcg_gen_xori_i32(d, s, 1u << 31);
+}
+
+static inline void gen_vfp_negd(TCGv_i64 d, TCGv_i64 s)
+{
+    tcg_gen_xori_i64(d, s, 1ull << 63);
+}
+
 /* Vector operations shared between ARM and AArch64.  */
 void gen_gvec_ceq0(unsigned vece, uint32_t rd_ofs, uint32_t rm_ofs,
                    uint32_t opr_sz, uint32_t max_sz);
diff --git a/target/arm/tcg/translate-a64.c b/target/arm/tcg/translate-a64.c
index 6f8207d842..878f83298f 100644
--- a/target/arm/tcg/translate-a64.c
+++ b/target/arm/tcg/translate-a64.c
@@ -6591,10 +6591,10 @@  static void handle_fp_1src_half(DisasContext *s, int opcode, int rd, int rn)
         tcg_gen_mov_i32(tcg_res, tcg_op);
         break;
     case 0x1: /* FABS */
-        tcg_gen_andi_i32(tcg_res, tcg_op, 0x7fff);
+        gen_vfp_absh(tcg_res, tcg_op);
         break;
     case 0x2: /* FNEG */
-        tcg_gen_xori_i32(tcg_res, tcg_op, 0x8000);
+        gen_vfp_negh(tcg_res, tcg_op);
         break;
     case 0x3: /* FSQRT */
         fpst = fpstatus_ptr(FPST_FPCR_F16);
@@ -6645,10 +6645,10 @@  static void handle_fp_1src_single(DisasContext *s, int opcode, int rd, int rn)
         tcg_gen_mov_i32(tcg_res, tcg_op);
         goto done;
     case 0x1: /* FABS */
-        gen_helper_vfp_abss(tcg_res, tcg_op);
+        gen_vfp_abss(tcg_res, tcg_op);
         goto done;
     case 0x2: /* FNEG */
-        gen_helper_vfp_negs(tcg_res, tcg_op);
+        gen_vfp_negs(tcg_res, tcg_op);
         goto done;
     case 0x3: /* FSQRT */
         gen_helper_vfp_sqrts(tcg_res, tcg_op, tcg_env);
@@ -6720,10 +6720,10 @@  static void handle_fp_1src_double(DisasContext *s, int opcode, int rd, int rn)
 
     switch (opcode) {
     case 0x1: /* FABS */
-        gen_helper_vfp_absd(tcg_res, tcg_op);
+        gen_vfp_absd(tcg_res, tcg_op);
         goto done;
     case 0x2: /* FNEG */
-        gen_helper_vfp_negd(tcg_res, tcg_op);
+        gen_vfp_negd(tcg_res, tcg_op);
         goto done;
     case 0x3: /* FSQRT */
         gen_helper_vfp_sqrtd(tcg_res, tcg_op, tcg_env);
@@ -6949,7 +6949,7 @@  static void handle_fp_2src_single(DisasContext *s, int opcode,
     switch (opcode) {
     case 0x8: /* FNMUL */
         gen_helper_vfp_muls(tcg_res, tcg_op1, tcg_op2, fpst);
-        gen_helper_vfp_negs(tcg_res, tcg_res);
+        gen_vfp_negs(tcg_res, tcg_res);
         break;
     default:
     case 0x0: /* FMUL */
@@ -6983,7 +6983,7 @@  static void handle_fp_2src_double(DisasContext *s, int opcode,
     switch (opcode) {
     case 0x8: /* FNMUL */
         gen_helper_vfp_muld(tcg_res, tcg_op1, tcg_op2, fpst);
-        gen_helper_vfp_negd(tcg_res, tcg_res);
+        gen_vfp_negd(tcg_res, tcg_res);
         break;
     default:
     case 0x0: /* FMUL */
@@ -7017,7 +7017,7 @@  static void handle_fp_2src_half(DisasContext *s, int opcode,
     switch (opcode) {
     case 0x8: /* FNMUL */
         gen_helper_advsimd_mulh(tcg_res, tcg_op1, tcg_op2, fpst);
-        tcg_gen_xori_i32(tcg_res, tcg_res, 0x8000);
+        gen_vfp_negh(tcg_res, tcg_res);
         break;
     default:
     case 0x0: /* FMUL */
@@ -7102,11 +7102,11 @@  static void handle_fp_3src_single(DisasContext *s, bool o0, bool o1,
      * flipped if it is a negated-input.
      */
     if (o1 == true) {
-        gen_helper_vfp_negs(tcg_op3, tcg_op3);
+        gen_vfp_negs(tcg_op3, tcg_op3);
     }
 
     if (o0 != o1) {
-        gen_helper_vfp_negs(tcg_op1, tcg_op1);
+        gen_vfp_negs(tcg_op1, tcg_op1);
     }
 
     gen_helper_vfp_muladds(tcg_res, tcg_op1, tcg_op2, tcg_op3, fpst);
@@ -7134,11 +7134,11 @@  static void handle_fp_3src_double(DisasContext *s, bool o0, bool o1,
      * flipped if it is a negated-input.
      */
     if (o1 == true) {
-        gen_helper_vfp_negd(tcg_op3, tcg_op3);
+        gen_vfp_negd(tcg_op3, tcg_op3);
     }
 
     if (o0 != o1) {
-        gen_helper_vfp_negd(tcg_op1, tcg_op1);
+        gen_vfp_negd(tcg_op1, tcg_op1);
     }
 
     gen_helper_vfp_muladdd(tcg_res, tcg_op1, tcg_op2, tcg_op3, fpst);
@@ -9246,7 +9246,7 @@  static void handle_3same_float(DisasContext *s, int size, int elements,
             switch (fpopcode) {
             case 0x39: /* FMLS */
                 /* As usual for ARM, separate negation for fused multiply-add */
-                gen_helper_vfp_negd(tcg_op1, tcg_op1);
+                gen_vfp_negd(tcg_op1, tcg_op1);
                 /* fall through */
             case 0x19: /* FMLA */
                 read_vec_element(s, tcg_res, rd, pass, MO_64);
@@ -9270,7 +9270,7 @@  static void handle_3same_float(DisasContext *s, int size, int elements,
                 break;
             case 0x7a: /* FABD */
                 gen_helper_vfp_subd(tcg_res, tcg_op1, tcg_op2, fpst);
-                gen_helper_vfp_absd(tcg_res, tcg_res);
+                gen_vfp_absd(tcg_res, tcg_res);
                 break;
             case 0x7c: /* FCMGT */
                 gen_helper_neon_cgt_f64(tcg_res, tcg_op1, tcg_op2, fpst);
@@ -9304,7 +9304,7 @@  static void handle_3same_float(DisasContext *s, int size, int elements,
             switch (fpopcode) {
             case 0x39: /* FMLS */
                 /* As usual for ARM, separate negation for fused multiply-add */
-                gen_helper_vfp_negs(tcg_op1, tcg_op1);
+                gen_vfp_negs(tcg_op1, tcg_op1);
                 /* fall through */
             case 0x19: /* FMLA */
                 read_vec_element_i32(s, tcg_res, rd, pass, MO_32);
@@ -9328,7 +9328,7 @@  static void handle_3same_float(DisasContext *s, int size, int elements,
                 break;
             case 0x7a: /* FABD */
                 gen_helper_vfp_subs(tcg_res, tcg_op1, tcg_op2, fpst);
-                gen_helper_vfp_abss(tcg_res, tcg_res);
+                gen_vfp_abss(tcg_res, tcg_res);
                 break;
             case 0x7c: /* FCMGT */
                 gen_helper_neon_cgt_f32(tcg_res, tcg_op1, tcg_op2, fpst);
@@ -9741,10 +9741,10 @@  static void handle_2misc_64(DisasContext *s, int opcode, bool u,
         }
         break;
     case 0x2f: /* FABS */
-        gen_helper_vfp_absd(tcg_rd, tcg_rn);
+        gen_vfp_absd(tcg_rd, tcg_rn);
         break;
     case 0x6f: /* FNEG */
-        gen_helper_vfp_negd(tcg_rd, tcg_rn);
+        gen_vfp_negd(tcg_rd, tcg_rn);
         break;
     case 0x7f: /* FSQRT */
         gen_helper_vfp_sqrtd(tcg_rd, tcg_rn, tcg_env);
@@ -12567,10 +12567,10 @@  static void disas_simd_two_reg_misc(DisasContext *s, uint32_t insn)
                     }
                     break;
                 case 0x2f: /* FABS */
-                    gen_helper_vfp_abss(tcg_res, tcg_op);
+                    gen_vfp_abss(tcg_res, tcg_op);
                     break;
                 case 0x6f: /* FNEG */
-                    gen_helper_vfp_negs(tcg_res, tcg_op);
+                    gen_vfp_negs(tcg_res, tcg_op);
                     break;
                 case 0x7f: /* FSQRT */
                     gen_helper_vfp_sqrts(tcg_res, tcg_op, tcg_env);
@@ -13291,7 +13291,7 @@  static void disas_simd_indexed(DisasContext *s, uint32_t insn)
             switch (16 * u + opcode) {
             case 0x05: /* FMLS */
                 /* As usual for ARM, separate negation for fused multiply-add */
-                gen_helper_vfp_negd(tcg_op, tcg_op);
+                gen_vfp_negd(tcg_op, tcg_op);
                 /* fall through */
             case 0x01: /* FMLA */
                 read_vec_element(s, tcg_res, rd, pass, MO_64);
diff --git a/target/arm/tcg/translate-vfp.c b/target/arm/tcg/translate-vfp.c
index 8e755fcde8..39ec971ff7 100644
--- a/target/arm/tcg/translate-vfp.c
+++ b/target/arm/tcg/translate-vfp.c
@@ -1768,7 +1768,7 @@  static void gen_VMLS_hp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
     TCGv_i32 tmp = tcg_temp_new_i32();
 
     gen_helper_vfp_mulh(tmp, vn, vm, fpst);
-    gen_helper_vfp_negh(tmp, tmp);
+    gen_vfp_negh(tmp, tmp);
     gen_helper_vfp_addh(vd, vd, tmp, fpst);
 }
 
@@ -1786,7 +1786,7 @@  static void gen_VMLS_sp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
     TCGv_i32 tmp = tcg_temp_new_i32();
 
     gen_helper_vfp_muls(tmp, vn, vm, fpst);
-    gen_helper_vfp_negs(tmp, tmp);
+    gen_vfp_negs(tmp, tmp);
     gen_helper_vfp_adds(vd, vd, tmp, fpst);
 }
 
@@ -1804,7 +1804,7 @@  static void gen_VMLS_dp(TCGv_i64 vd, TCGv_i64 vn, TCGv_i64 vm, TCGv_ptr fpst)
     TCGv_i64 tmp = tcg_temp_new_i64();
 
     gen_helper_vfp_muld(tmp, vn, vm, fpst);
-    gen_helper_vfp_negd(tmp, tmp);
+    gen_vfp_negd(tmp, tmp);
     gen_helper_vfp_addd(vd, vd, tmp, fpst);
 }
 
@@ -1824,7 +1824,7 @@  static void gen_VNMLS_hp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
     TCGv_i32 tmp = tcg_temp_new_i32();
 
     gen_helper_vfp_mulh(tmp, vn, vm, fpst);
-    gen_helper_vfp_negh(vd, vd);
+    gen_vfp_negh(vd, vd);
     gen_helper_vfp_addh(vd, vd, tmp, fpst);
 }
 
@@ -1844,7 +1844,7 @@  static void gen_VNMLS_sp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
     TCGv_i32 tmp = tcg_temp_new_i32();
 
     gen_helper_vfp_muls(tmp, vn, vm, fpst);
-    gen_helper_vfp_negs(vd, vd);
+    gen_vfp_negs(vd, vd);
     gen_helper_vfp_adds(vd, vd, tmp, fpst);
 }
 
@@ -1864,7 +1864,7 @@  static void gen_VNMLS_dp(TCGv_i64 vd, TCGv_i64 vn, TCGv_i64 vm, TCGv_ptr fpst)
     TCGv_i64 tmp = tcg_temp_new_i64();
 
     gen_helper_vfp_muld(tmp, vn, vm, fpst);
-    gen_helper_vfp_negd(vd, vd);
+    gen_vfp_negd(vd, vd);
     gen_helper_vfp_addd(vd, vd, tmp, fpst);
 }
 
@@ -1879,8 +1879,8 @@  static void gen_VNMLA_hp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
     TCGv_i32 tmp = tcg_temp_new_i32();
 
     gen_helper_vfp_mulh(tmp, vn, vm, fpst);
-    gen_helper_vfp_negh(tmp, tmp);
-    gen_helper_vfp_negh(vd, vd);
+    gen_vfp_negh(tmp, tmp);
+    gen_vfp_negh(vd, vd);
     gen_helper_vfp_addh(vd, vd, tmp, fpst);
 }
 
@@ -1895,8 +1895,8 @@  static void gen_VNMLA_sp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
     TCGv_i32 tmp = tcg_temp_new_i32();
 
     gen_helper_vfp_muls(tmp, vn, vm, fpst);
-    gen_helper_vfp_negs(tmp, tmp);
-    gen_helper_vfp_negs(vd, vd);
+    gen_vfp_negs(tmp, tmp);
+    gen_vfp_negs(vd, vd);
     gen_helper_vfp_adds(vd, vd, tmp, fpst);
 }
 
@@ -1911,8 +1911,8 @@  static void gen_VNMLA_dp(TCGv_i64 vd, TCGv_i64 vn, TCGv_i64 vm, TCGv_ptr fpst)
     TCGv_i64 tmp = tcg_temp_new_i64();
 
     gen_helper_vfp_muld(tmp, vn, vm, fpst);
-    gen_helper_vfp_negd(tmp, tmp);
-    gen_helper_vfp_negd(vd, vd);
+    gen_vfp_negd(tmp, tmp);
+    gen_vfp_negd(vd, vd);
     gen_helper_vfp_addd(vd, vd, tmp, fpst);
 }
 
@@ -1940,7 +1940,7 @@  static void gen_VNMUL_hp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
 {
     /* VNMUL: -(fn * fm) */
     gen_helper_vfp_mulh(vd, vn, vm, fpst);
-    gen_helper_vfp_negh(vd, vd);
+    gen_vfp_negh(vd, vd);
 }
 
 static bool trans_VNMUL_hp(DisasContext *s, arg_VNMUL_sp *a)
@@ -1952,7 +1952,7 @@  static void gen_VNMUL_sp(TCGv_i32 vd, TCGv_i32 vn, TCGv_i32 vm, TCGv_ptr fpst)
 {
     /* VNMUL: -(fn * fm) */
     gen_helper_vfp_muls(vd, vn, vm, fpst);
-    gen_helper_vfp_negs(vd, vd);
+    gen_vfp_negs(vd, vd);
 }
 
 static bool trans_VNMUL_sp(DisasContext *s, arg_VNMUL_sp *a)
@@ -1964,7 +1964,7 @@  static void gen_VNMUL_dp(TCGv_i64 vd, TCGv_i64 vn, TCGv_i64 vm, TCGv_ptr fpst)
 {
     /* VNMUL: -(fn * fm) */
     gen_helper_vfp_muld(vd, vn, vm, fpst);
-    gen_helper_vfp_negd(vd, vd);
+    gen_vfp_negd(vd, vd);
 }
 
 static bool trans_VNMUL_dp(DisasContext *s, arg_VNMUL_dp *a)
@@ -2115,12 +2115,12 @@  static bool do_vfm_hp(DisasContext *s, arg_VFMA_sp *a, bool neg_n, bool neg_d)
     vfp_load_reg16(vm, a->vm);
     if (neg_n) {
         /* VFNMS, VFMS */
-        gen_helper_vfp_negh(vn, vn);
+        gen_vfp_negh(vn, vn);
     }
     vfp_load_reg16(vd, a->vd);
     if (neg_d) {
         /* VFNMA, VFNMS */
-        gen_helper_vfp_negh(vd, vd);
+        gen_vfp_negh(vd, vd);
     }
     fpst = fpstatus_ptr(FPST_FPCR_F16);
     gen_helper_vfp_muladdh(vd, vn, vm, vd, fpst);
@@ -2174,12 +2174,12 @@  static bool do_vfm_sp(DisasContext *s, arg_VFMA_sp *a, bool neg_n, bool neg_d)
     vfp_load_reg32(vm, a->vm);
     if (neg_n) {
         /* VFNMS, VFMS */
-        gen_helper_vfp_negs(vn, vn);
+        gen_vfp_negs(vn, vn);
     }
     vfp_load_reg32(vd, a->vd);
     if (neg_d) {
         /* VFNMA, VFNMS */
-        gen_helper_vfp_negs(vd, vd);
+        gen_vfp_negs(vd, vd);
     }
     fpst = fpstatus_ptr(FPST_FPCR);
     gen_helper_vfp_muladds(vd, vn, vm, vd, fpst);
@@ -2239,12 +2239,12 @@  static bool do_vfm_dp(DisasContext *s, arg_VFMA_dp *a, bool neg_n, bool neg_d)
     vfp_load_reg64(vm, a->vm);
     if (neg_n) {
         /* VFNMS, VFMS */
-        gen_helper_vfp_negd(vn, vn);
+        gen_vfp_negd(vn, vn);
     }
     vfp_load_reg64(vd, a->vd);
     if (neg_d) {
         /* VFNMA, VFNMS */
-        gen_helper_vfp_negd(vd, vd);
+        gen_vfp_negd(vd, vd);
     }
     fpst = fpstatus_ptr(FPST_FPCR);
     gen_helper_vfp_muladdd(vd, vn, vm, vd, fpst);
@@ -2414,13 +2414,13 @@  static bool trans_VMOV_imm_dp(DisasContext *s, arg_VMOV_imm_dp *a)
 DO_VFP_VMOV(VMOV_reg, sp, tcg_gen_mov_i32)
 DO_VFP_VMOV(VMOV_reg, dp, tcg_gen_mov_i64)
 
-DO_VFP_2OP(VABS, hp, gen_helper_vfp_absh, aa32_fp16_arith)
-DO_VFP_2OP(VABS, sp, gen_helper_vfp_abss, aa32_fpsp_v2)
-DO_VFP_2OP(VABS, dp, gen_helper_vfp_absd, aa32_fpdp_v2)
+DO_VFP_2OP(VABS, hp, gen_vfp_absh, aa32_fp16_arith)
+DO_VFP_2OP(VABS, sp, gen_vfp_abss, aa32_fpsp_v2)
+DO_VFP_2OP(VABS, dp, gen_vfp_absd, aa32_fpdp_v2)
 
-DO_VFP_2OP(VNEG, hp, gen_helper_vfp_negh, aa32_fp16_arith)
-DO_VFP_2OP(VNEG, sp, gen_helper_vfp_negs, aa32_fpsp_v2)
-DO_VFP_2OP(VNEG, dp, gen_helper_vfp_negd, aa32_fpdp_v2)
+DO_VFP_2OP(VNEG, hp, gen_vfp_negh, aa32_fp16_arith)
+DO_VFP_2OP(VNEG, sp, gen_vfp_negs, aa32_fpsp_v2)
+DO_VFP_2OP(VNEG, dp, gen_vfp_negd, aa32_fpdp_v2)
 
 static void gen_VSQRT_hp(TCGv_i32 vd, TCGv_i32 vm)
 {
diff --git a/target/arm/vfp_helper.c b/target/arm/vfp_helper.c
index 3e5e37abbe..ce26b8a71a 100644
--- a/target/arm/vfp_helper.c
+++ b/target/arm/vfp_helper.c
@@ -281,36 +281,6 @@  VFP_BINOP(minnum)
 VFP_BINOP(maxnum)
 #undef VFP_BINOP
 
-dh_ctype_f16 VFP_HELPER(neg, h)(dh_ctype_f16 a)
-{
-    return float16_chs(a);
-}
-
-float32 VFP_HELPER(neg, s)(float32 a)
-{
-    return float32_chs(a);
-}
-
-float64 VFP_HELPER(neg, d)(float64 a)
-{
-    return float64_chs(a);
-}
-
-dh_ctype_f16 VFP_HELPER(abs, h)(dh_ctype_f16 a)
-{
-    return float16_abs(a);
-}
-
-float32 VFP_HELPER(abs, s)(float32 a)
-{
-    return float32_abs(a);
-}
-
-float64 VFP_HELPER(abs, d)(float64 a)
-{
-    return float64_abs(a);
-}
-
 dh_ctype_f16 VFP_HELPER(sqrt, h)(dh_ctype_f16 a, CPUARMState *env)
 {
     return float16_sqrt(a, &env->vfp.fp_status_f16);