From nobody Tue Feb 10 17:45:12 2026 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 1624534017476249.99423240739316; Thu, 24 Jun 2021 04:26:57 -0700 (PDT) Received: from localhost ([::1]:35628 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1lwNVY-0005v4-C2 for importer@patchew.org; Thu, 24 Jun 2021 07:26:56 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:42884) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1lwN6c-0000jw-T7; Thu, 24 Jun 2021 07:01:12 -0400 Received: from out28-122.mail.aliyun.com ([115.124.28.122]:57864) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1lwN6Z-0007YW-9f; Thu, 24 Jun 2021 07:01:10 -0400 Received: from roman-VirtualBox.hz.ali.com(mailfrom:zhiwei_liu@c-sky.com fp:SMTPD_---.KXK24HG_1624532462) by smtp.aliyun-inc.com(10.147.40.233); Thu, 24 Jun 2021 19:01:02 +0800 X-Alimail-AntiSpam: AC=CONTINUE; BC=0.07436699|-1; CH=green; DM=|CONTINUE|false|; DS=CONTINUE|ham_system_inform|0.469349-9.41521e-05-0.530557; FP=0|0|0|0|0|-1|-1|-1; HT=ay29a033018047193; MF=zhiwei_liu@c-sky.com; NM=1; PH=DS; RN=6; RT=6; SR=0; TI=SMTPD_---.KXK24HG_1624532462; From: LIU Zhiwei To: qemu-devel@nongnu.org, qemu-riscv@nongnu.org Subject: [PATCH v3 09/37] target/riscv: SIMD 16-bit Multiply Instructions Date: Thu, 24 Jun 2021 18:54:53 +0800 Message-Id: <20210624105521.3964-10-zhiwei_liu@c-sky.com> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20210624105521.3964-1-zhiwei_liu@c-sky.com> References: <20210624105521.3964-1-zhiwei_liu@c-sky.com> Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: none client-ip=115.124.28.122; envelope-from=zhiwei_liu@c-sky.com; helo=out28-122.mail.aliyun.com X-Spam_score_int: -18 X-Spam_score: -1.9 X-Spam_bar: - X-Spam_report: (-1.9 / 5.0 requ) BAYES_00=-1.9, RCVD_IN_DNSWL_NONE=-0.0001, RCVD_IN_MSPIKE_H4=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_HELO_NONE=0.001, SPF_NONE=0.001, UNPARSEABLE_RELAY=0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: palmer@dabbelt.com, bin.meng@windriver.com, Alistair.Francis@wdc.com, LIU Zhiwei Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" There are 6 instructions, including 16-bit signed or unsigned multiply, 16-bit signed or unsigned crossed multiply, Q15 signed or signed crossed saturating multiply. Signed-off-by: LIU Zhiwei --- target/riscv/helper.h | 7 ++ target/riscv/insn32.decode | 7 ++ target/riscv/insn_trans/trans_rvp.c.inc | 69 ++++++++++++++++ target/riscv/packed_helper.c | 104 ++++++++++++++++++++++++ 4 files changed, 187 insertions(+) diff --git a/target/riscv/helper.h b/target/riscv/helper.h index c424e45fe5..d13b84f165 100644 --- a/target/riscv/helper.h +++ b/target/riscv/helper.h @@ -1218,3 +1218,10 @@ DEF_HELPER_3(scmplt8, tl, env, tl, tl) DEF_HELPER_3(scmple8, tl, env, tl, tl) DEF_HELPER_3(ucmplt8, tl, env, tl, tl) DEF_HELPER_3(ucmple8, tl, env, tl, tl) + +DEF_HELPER_3(smul16, i64, env, tl, tl) +DEF_HELPER_3(smulx16, i64, env, tl, tl) +DEF_HELPER_3(umul16, i64, env, tl, tl) +DEF_HELPER_3(umulx16, i64, env, tl, tl) +DEF_HELPER_3(khm16, tl, env, tl, tl) +DEF_HELPER_3(khmx16, tl, env, tl, tl) diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode index fdbf3798c7..cbee995229 100644 --- a/target/riscv/insn32.decode +++ b/target/riscv/insn32.decode @@ -821,3 +821,10 @@ scmplt8 0000111 ..... ..... 000 ..... 1110111 @r scmple8 0001111 ..... ..... 000 ..... 1110111 @r ucmplt8 0010111 ..... ..... 000 ..... 1110111 @r ucmple8 0011111 ..... ..... 000 ..... 1110111 @r + +smul16 1010000 ..... ..... 000 ..... 1110111 @r +smulx16 1010001 ..... ..... 000 ..... 1110111 @r +umul16 1011000 ..... ..... 000 ..... 1110111 @r +umulx16 1011001 ..... ..... 000 ..... 1110111 @r +khm16 1000011 ..... ..... 000 ..... 1110111 @r +khmx16 1001011 ..... ..... 000 ..... 1110111 @r diff --git a/target/riscv/insn_trans/trans_rvp.c.inc b/target/riscv/insn_tr= ans/trans_rvp.c.inc index aa432701c8..b93ba63dd8 100644 --- a/target/riscv/insn_trans/trans_rvp.c.inc +++ b/target/riscv/insn_trans/trans_rvp.c.inc @@ -217,3 +217,72 @@ GEN_RVP_R_OOL(scmplt8); GEN_RVP_R_OOL(scmple8); GEN_RVP_R_OOL(ucmplt8); GEN_RVP_R_OOL(ucmple8); + +/* SIMD 16-bit Multiply Instructions */ +static void set_pair_regs(DisasContext *ctx, TCGv_i64 dst, int rd) +{ + TCGv t1, t2; + + t1 =3D tcg_temp_new(); + t2 =3D tcg_temp_new(); + + if (is_32bit(ctx)) { + TCGv_i32 lo, hi; + + lo =3D tcg_temp_new_i32(); + hi =3D tcg_temp_new_i32(); + tcg_gen_extr_i64_i32(lo, hi, dst); + + tcg_gen_ext_i32_tl(t1, lo); + tcg_gen_ext_i32_tl(t2, hi); + + gen_set_gpr(rd, t1); + gen_set_gpr(rd + 1, t2); + tcg_temp_free_i32(lo); + tcg_temp_free_i32(hi); + } else { + tcg_gen_trunc_i64_tl(t1, dst); + gen_set_gpr(rd, t1); + } + tcg_temp_free(t1); + tcg_temp_free(t2); +} + +static inline bool +r_d64_ool(DisasContext *ctx, arg_r *a, + void (* fn)(TCGv_i64, TCGv_ptr, TCGv, TCGv)) +{ + TCGv t1, t2; + TCGv_i64 t3; + + if (!has_ext(ctx, RVP) || !ctx->ext_psfoperand) { + return false; + } + + t1 =3D tcg_temp_new(); + t2 =3D tcg_temp_new(); + t3 =3D tcg_temp_new_i64(); + + gen_get_gpr(t1, a->rs1); + gen_get_gpr(t2, a->rs2); + fn(t3, cpu_env, t1, t2); + set_pair_regs(ctx, t3, a->rd); + + tcg_temp_free(t1); + tcg_temp_free(t2); + tcg_temp_free_i64(t3); + return true; +} + +#define GEN_RVP_R_D64_OOL(NAME) \ +static bool trans_##NAME(DisasContext *s, arg_r *a) \ +{ \ + return r_d64_ool(s, a, gen_helper_##NAME); \ +} + +GEN_RVP_R_D64_OOL(smul16); +GEN_RVP_R_D64_OOL(smulx16); +GEN_RVP_R_D64_OOL(umul16); +GEN_RVP_R_D64_OOL(umulx16); +GEN_RVP_R_OOL(khm16); +GEN_RVP_R_OOL(khmx16); diff --git a/target/riscv/packed_helper.c b/target/riscv/packed_helper.c index ff86e015e4..13fed2c4d1 100644 --- a/target/riscv/packed_helper.c +++ b/target/riscv/packed_helper.c @@ -723,3 +723,107 @@ static inline void do_ucmple8(CPURISCVState *env, voi= d *vd, void *va, } =20 RVPR(ucmple8, 1, 1); + +/* SIMD 16-bit Multiply Instructions */ +typedef void PackedFn3(CPURISCVState *, void *, void *, void *); +static inline uint64_t rvpr64(CPURISCVState *env, target_ulong a, + target_ulong b, PackedFn3 *fn) +{ + uint64_t result; + + fn(env, &result, &a, &b); + return result; +} + +#define RVPR64(NAME) \ +uint64_t HELPER(NAME)(CPURISCVState *env, target_ulong a, \ + target_ulong b) \ +{ \ + return rvpr64(env, a, b, (PackedFn3 *)do_##NAME); \ +} + +static inline void do_smul16(CPURISCVState *env, void *vd, void *va, void = *vb) +{ + int32_t *d =3D vd; + int16_t *a =3D va, *b =3D vb; + d[H4(0)] =3D (int32_t)a[H2(0)] * b[H2(0)]; + d[H4(1)] =3D (int32_t)a[H2(1)] * b[H2(1)]; +} + +RVPR64(smul16); + +static inline void do_smulx16(CPURISCVState *env, void *vd, void *va, void= *vb) +{ + int32_t *d =3D vd; + int16_t *a =3D va, *b =3D vb; + d[H4(0)] =3D (int32_t)a[H2(0)] * b[H2(1)]; + d[H4(1)] =3D (int32_t)a[H2(1)] * b[H2(0)]; +} + +RVPR64(smulx16); + +static inline void do_umul16(CPURISCVState *env, void *vd, void *va, void = *vb, + uint8_t i) +{ + uint32_t *d =3D vd; + uint16_t *a =3D va, *b =3D vb; + d[H4(0)] =3D (uint32_t)a[H2(0)] * b[H2(0)]; + d[H4(1)] =3D (uint32_t)a[H2(1)] * b[H2(1)]; +} + +RVPR64(umul16); + +static inline void do_umulx16(CPURISCVState *env, void *vd, void *va, void= *vb, + uint8_t i) +{ + uint32_t *d =3D vd; + uint16_t *a =3D va, *b =3D vb; + d[H4(0)] =3D (uint32_t)a[H2(0)] * b[H2(1)]; + d[H4(1)] =3D (uint32_t)a[H2(1)] * b[H2(0)]; +} + +RVPR64(umulx16); + +static inline void do_khm16(CPURISCVState *env, void *vd, void *va, + void *vb, uint8_t i) +{ + int16_t *d =3D vd, *a =3D va, *b =3D vb; + + if (a[i] =3D=3D INT16_MIN && b[i] =3D=3D INT16_MIN) { + env->vxsat =3D 1; + d[i] =3D INT16_MAX; + } else { + d[i] =3D (int32_t)a[i] * b[i] >> 15; + } +} + +RVPR(khm16, 1, 2); + +static inline void do_khmx16(CPURISCVState *env, void *vd, void *va, + void *vb, uint8_t i) +{ + int16_t *d =3D vd, *a =3D va, *b =3D vb; + + /* + * t[x] =3D ra.H[x] s* rb.H[y]; + * rt.H[x] =3D SAT.Q15(t[x] s>> 15); + * + * (RV32: (x,y)=3D(1,0),(0,1), + * RV64: (x,y)=3D(3,2),(2,3), + * (1,0),(0,1) + */ + if (a[H2(i)] =3D=3D INT16_MIN && b[H2(i + 1)] =3D=3D INT16_MIN) { + env->vxsat =3D 1; + d[H2(i)] =3D INT16_MAX; + } else { + d[H2(i)] =3D (int32_t)a[H2(i)] * b[H2(i + 1)] >> 15; + } + if (a[H2(i + 1)] =3D=3D INT16_MIN && b[H2(i)] =3D=3D INT16_MIN) { + env->vxsat =3D 1; + d[H2(i + 1)] =3D INT16_MAX; + } else { + d[H2(i + 1)] =3D (int32_t)a[H2(i + 1)] * b[H2(i)] >> 15; + } +} + +RVPR(khmx16, 2, 2); --=20 2.17.1