[04/13] target/arm: Convert SQRDMLAH, SQRDMLSH to decodetree

Message ID	20240625050810.1475643-5-richard.henderson@linaro.org (mailing list archive)
State	New, archived
Headers	show Return-Path: <qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org> From: Richard Henderson <richard.henderson@linaro.org> To: qemu-devel@nongnu.org Cc: qemu-arm@nongnu.org Subject: [PATCH 04/13] target/arm: Convert SQRDMLAH, SQRDMLSH to decodetree Date: Mon, 24 Jun 2024 22:08:01 -0700 Message-Id: <20240625050810.1475643-5-richard.henderson@linaro.org> In-Reply-To: <20240625050810.1475643-1-richard.henderson@linaro.org> References: <20240625050810.1475643-1-richard.henderson@linaro.org> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Received-SPF: pass client-ip=2607:f8b0:4864:20::635; envelope-from=richard.henderson@linaro.org; helo=mail-pl1-x635.google.com X-Spam_score_int: -20 X-Spam_score: -2.1 X-Spam_bar: -- X-Spam_report: (-2.1 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action Precedence: list Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org
Series	target/arm: AdvSIMD conversion, part 2 \| expand [00/13] target/arm: AdvSIMD conversion, part 2 [01/13] target/arm: Fix VCMLA Dd, Dn, Dm[idx] [02/13] target/arm: Fix SQDMULH (by element) with Q=0 [03/13] target/arm: Fix FJCVTZS vs flush-to-zero [04/13] target/arm: Convert SQRDMLAH, SQRDMLSH to decodetree [05/13] target/arm: Convert SDOT, UDOT to decodetree [06/13] target/arm: Convert SUDOT, USDOT to decodetree [07/13] target/arm: Convert BFDOT to decodetree [08/13] target/arm: Convert BFMLALB, BFMLALT to decodetree [09/13] target/arm: Convert BFMMLA, SMMLA, UMMLA, USMMLA to decodetree [10/13] target/arm: Add data argument to do_fp3_vector [11/13] target/arm: Convert FCADD to decodetree [12/13] target/arm: Convert FCMLA to decodetree [13/13] target/arm: Delete dead code from disas_simd_indexed

diff --git a/target/arm/helper.h b/target/arm/helper.h index eca2043fc2..970d059dec 100644 --- a/target/arm/helper.h +++ b/target/arm/helper.h @@ -979,6 +979,16 @@ DEF_HELPER_FLAGS_5(neon_sqrdmulh_idx_h, TCG_CALL_NO_RWG, DEF_HELPER_FLAGS_5(neon_sqrdmulh_idx_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(neon_sqrdmlah_idx_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(neon_sqrdmlah_idx_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_5(neon_sqrdmlsh_idx_h, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_5(neon_sqrdmlsh_idx_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, i32) + DEF_HELPER_FLAGS_4(sve2_sqdmulh_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve2_sqdmulh_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve2_sqdmulh_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) diff --git a/target/arm/tcg/a64.decode b/target/arm/tcg/a64.decode index 2b7a3254a0..613cc9365c 100644 --- a/target/arm/tcg/a64.decode +++ b/target/arm/tcg/a64.decode @@ -781,6 +781,8 @@ CMEQ_s 0111 1110 111 ..... 10001 1 ..... ..... @rrr_d SQDMULH_s 0101 1110 ..1 ..... 10110 1 ..... ..... @rrr_e SQRDMULH_s 0111 1110 ..1 ..... 10110 1 ..... ..... @rrr_e +SQRDMLAH_s 0111 1110 ..0 ..... 10000 1 ..... ..... @rrr_e +SQRDMLSH_s 0111 1110 ..0 ..... 10001 1 ..... ..... @rrr_e ### Advanced SIMD scalar pairwise @@ -941,6 +943,8 @@ MLS_v 0.10 1110 ..1 ..... 10010 1 ..... ..... @qrrr_e SQDMULH_v 0.00 1110 ..1 ..... 10110 1 ..... ..... @qrrr_e SQRDMULH_v 0.10 1110 ..1 ..... 10110 1 ..... ..... @qrrr_e +SQRDMLAH_v 0.10 1110 ..0 ..... 10000 1 ..... ..... @qrrr_e +SQRDMLSH_v 0.10 1110 ..0 ..... 10001 1 ..... ..... @qrrr_e ### Advanced SIMD scalar x indexed element @@ -966,6 +970,12 @@ SQDMULH_si 0101 1111 10 .. .... 1100 . 0 ..... ..... @rrx_s SQRDMULH_si 0101 1111 01 .. .... 1101 . 0 ..... ..... @rrx_h SQRDMULH_si 0101 1111 10 . ..... 1101 . 0 ..... ..... @rrx_s +SQRDMLAH_si 0111 1111 01 .. .... 1101 . 0 ..... ..... @rrx_h +SQRDMLAH_si 0111 1111 10 .. .... 1101 . 0 ..... ..... @rrx_s + +SQRDMLSH_si 0111 1111 01 .. .... 1111 . 0 ..... ..... @rrx_h +SQRDMLSH_si 0111 1111 10 .. .... 1111 . 0 ..... ..... @rrx_s + ### Advanced SIMD vector x indexed element FMUL_vi 0.00 1111 00 .. .... 1001 . 0 ..... ..... @qrrx_h @@ -1004,6 +1014,12 @@ SQDMULH_vi 0.00 1111 10 . ..... 1100 . 0 ..... ..... @qrrx_s SQRDMULH_vi 0.00 1111 01 .. .... 1101 . 0 ..... ..... @qrrx_h SQRDMULH_vi 0.00 1111 10 . ..... 1101 . 0 ..... ..... @qrrx_s +SQRDMLAH_vi 0.10 1111 01 .. .... 1101 . 0 ..... ..... @qrrx_h +SQRDMLAH_vi 0.10 1111 10 .. .... 1101 . 0 ..... ..... @qrrx_s + +SQRDMLSH_vi 0.10 1111 01 .. .... 1111 . 0 ..... ..... @qrrx_h +SQRDMLSH_vi 0.10 1111 10 .. .... 1111 . 0 ..... ..... @qrrx_s + # Floating-point conditional select FCSEL 0001 1110 .. 1 rm:5 cond:4 11 rn:5 rd:5 esz=%esz_hsd diff --git a/target/arm/tcg/translate-a64.c b/target/arm/tcg/translate-a64.c index 93543da39c..32c24c7422 100644 --- a/target/arm/tcg/translate-a64.c +++ b/target/arm/tcg/translate-a64.c @@ -5235,6 +5235,43 @@ static const ENVScalar2 f_scalar_sqrdmulh = { }; TRANS(SQRDMULH_s, do_env_scalar2_hs, a, &f_scalar_sqrdmulh) +typedef struct ENVScalar3 { + NeonGenThreeOpEnvFn *gen_hs[2]; +} ENVScalar3; + +static bool do_env_scalar3_hs(DisasContext *s, arg_rrr_e *a, + const ENVScalar3 *f) +{ + TCGv_i32 t0, t1, t2; + + if (a->esz != MO_16 && a->esz != MO_32) { + return false; + } + if (!fp_access_check(s)) { + return true; + } + + t0 = tcg_temp_new_i32(); + t1 = tcg_temp_new_i32(); + t2 = tcg_temp_new_i32(); + read_vec_element_i32(s, t0, a->rn, 0, a->esz); + read_vec_element_i32(s, t1, a->rm, 0, a->esz); + read_vec_element_i32(s, t2, a->rd, 0, a->esz); + f->gen_hs[a->esz - 1](t0, tcg_env, t0, t1, t2); + write_fp_sreg(s, a->rd, t0); + return true; +} + +static const ENVScalar3 f_scalar_sqrdmlah = { + { gen_helper_neon_qrdmlah_s16, gen_helper_neon_qrdmlah_s32 } +}; +TRANS_FEAT(SQRDMLAH_s, aa64_rdm, do_env_scalar3_hs, a, &f_scalar_sqrdmlah) + +static const ENVScalar3 f_scalar_sqrdmlsh = { + { gen_helper_neon_qrdmlsh_s16, gen_helper_neon_qrdmlsh_s32 } +}; +TRANS_FEAT(SQRDMLSH_s, aa64_rdm, do_env_scalar3_hs, a, &f_scalar_sqrdmlsh) + static bool do_cmop_d(DisasContext *s, arg_rrr_e *a, TCGCond cond) { if (fp_access_check(s)) { @@ -5552,6 +5589,8 @@ TRANS(CMTST_v, do_gvec_fn3, a, gen_gvec_cmtst) TRANS(SQDMULH_v, do_gvec_fn3_no8_no64, a, gen_gvec_sqdmulh_qc) TRANS(SQRDMULH_v, do_gvec_fn3_no8_no64, a, gen_gvec_sqrdmulh_qc) +TRANS_FEAT(SQRDMLAH_v, aa64_rdm, do_gvec_fn3_no8_no64, a, gen_gvec_sqrdmlah_qc) +TRANS_FEAT(SQRDMLSH_v, aa64_rdm, do_gvec_fn3_no8_no64, a, gen_gvec_sqrdmlsh_qc) /* * Advanced SIMD scalar/vector x indexed element @@ -5681,6 +5720,29 @@ static bool do_env_scalar2_idx_hs(DisasContext *s, arg_rrx_e *a, TRANS(SQDMULH_si, do_env_scalar2_idx_hs, a, &f_scalar_sqdmulh) TRANS(SQRDMULH_si, do_env_scalar2_idx_hs, a, &f_scalar_sqrdmulh) +static bool do_env_scalar3_idx_hs(DisasContext *s, arg_rrx_e *a, + const ENVScalar3 *f) +{ + if (a->esz < MO_16 || a->esz > MO_32) { + return false; + } + if (fp_access_check(s)) { + TCGv_i32 t0 = tcg_temp_new_i32(); + TCGv_i32 t1 = tcg_temp_new_i32(); + TCGv_i32 t2 = tcg_temp_new_i32(); + + read_vec_element_i32(s, t0, a->rn, 0, a->esz); + read_vec_element_i32(s, t1, a->rm, a->idx, a->esz); + read_vec_element_i32(s, t2, a->rd, 0, a->esz); + f->gen_hs[a->esz - 1](t0, tcg_env, t0, t1, t2); + write_fp_sreg(s, a->rd, t0); + } + return true; +} + +TRANS_FEAT(SQRDMLAH_si, aa64_rdm, do_env_scalar3_idx_hs, a, &f_scalar_sqrdmlah) +TRANS_FEAT(SQRDMLSH_si, aa64_rdm, do_env_scalar3_idx_hs, a, &f_scalar_sqrdmlsh) + static bool do_fp3_vector_idx(DisasContext *s, arg_qrrx_e *a, gen_helper_gvec_3_ptr * const fns[3]) { @@ -5838,6 +5900,20 @@ static gen_helper_gvec_4 * const f_vector_idx_sqrdmulh[2] = { }; TRANS(SQRDMULH_vi, do_int3_qc_vector_idx, a, f_vector_idx_sqrdmulh) +static gen_helper_gvec_4 * const f_vector_idx_sqrdmlah[2] = { + gen_helper_neon_sqrdmlah_idx_h, + gen_helper_neon_sqrdmlah_idx_s, +}; +TRANS_FEAT(SQRDMLAH_vi, aa64_rdm, do_int3_qc_vector_idx, a, + f_vector_idx_sqrdmlah) + +static gen_helper_gvec_4 * const f_vector_idx_sqrdmlsh[2] = { + gen_helper_neon_sqrdmlsh_idx_h, + gen_helper_neon_sqrdmlsh_idx_s, +}; +TRANS_FEAT(SQRDMLSH_vi, aa64_rdm, do_int3_qc_vector_idx, a, + f_vector_idx_sqrdmlsh) + /* * Advanced SIMD scalar pairwise */ @@ -9536,84 +9612,6 @@ static void disas_simd_scalar_three_reg_diff(DisasContext *s, uint32_t insn) } } -/* AdvSIMD scalar three same extra - * 31 30 29 28 24 23 22 21 20 16 15 14 11 10 9 5 4 0 - * +-----+---+-----------+------+---+------+---+--------+---+----+----+ - * | 0 1 | U | 1 1 1 1 0 | size | 0 | Rm | 1 | opcode | 1 | Rn | Rd | - * +-----+---+-----------+------+---+------+---+--------+---+----+----+ - */ -static void disas_simd_scalar_three_reg_same_extra(DisasContext *s, - uint32_t insn) -{ - int rd = extract32(insn, 0, 5); - int rn = extract32(insn, 5, 5); - int opcode = extract32(insn, 11, 4); - int rm = extract32(insn, 16, 5); - int size = extract32(insn, 22, 2); - bool u = extract32(insn, 29, 1); - TCGv_i32 ele1, ele2, ele3; - TCGv_i64 res; - bool feature; - - switch (u * 16 + opcode) { - case 0x10: /* SQRDMLAH (vector) */ - case 0x11: /* SQRDMLSH (vector) */ - if (size != 1 && size != 2) { - unallocated_encoding(s); - return; - } - feature = dc_isar_feature(aa64_rdm, s); - break; - default: - unallocated_encoding(s); - return; - } - if (!feature) { - unallocated_encoding(s); - return; - } - if (!fp_access_check(s)) { - return; - } - - /* Do a single operation on the lowest element in the vector. - * We use the standard Neon helpers and rely on 0 OP 0 == 0 - * with no side effects for all these operations. - * OPTME: special-purpose helpers would avoid doing some - * unnecessary work in the helper for the 16 bit cases. - */ - ele1 = tcg_temp_new_i32(); - ele2 = tcg_temp_new_i32(); - ele3 = tcg_temp_new_i32(); - - read_vec_element_i32(s, ele1, rn, 0, size); - read_vec_element_i32(s, ele2, rm, 0, size); - read_vec_element_i32(s, ele3, rd, 0, size); - - switch (opcode) { - case 0x0: /* SQRDMLAH */ - if (size == 1) { - gen_helper_neon_qrdmlah_s16(ele3, tcg_env, ele1, ele2, ele3); - } else { - gen_helper_neon_qrdmlah_s32(ele3, tcg_env, ele1, ele2, ele3); - } - break; - case 0x1: /* SQRDMLSH */ - if (size == 1) { - gen_helper_neon_qrdmlsh_s16(ele3, tcg_env, ele1, ele2, ele3); - } else { - gen_helper_neon_qrdmlsh_s32(ele3, tcg_env, ele1, ele2, ele3); - } - break; - default: - g_assert_not_reached(); - } - - res = tcg_temp_new_i64(); - tcg_gen_extu_i32_i64(res, ele3); - write_fp_dreg(s, rd, res); -} - static void handle_2misc_64(DisasContext *s, int opcode, bool u, TCGv_i64 tcg_rd, TCGv_i64 tcg_rn, TCGv_i32 tcg_rmode, TCGv_ptr tcg_fpstatus) @@ -10892,14 +10890,6 @@ static void disas_simd_three_reg_same_extra(DisasContext *s, uint32_t insn) int rot; switch (u * 16 + opcode) { - case 0x10: /* SQRDMLAH (vector) */ - case 0x11: /* SQRDMLSH (vector) */ - if (size != 1 && size != 2) { - unallocated_encoding(s); - return; - } - feature = dc_isar_feature(aa64_rdm, s); - break; case 0x02: /* SDOT (vector) */ case 0x12: /* UDOT (vector) */ if (size != MO_32) { @@ -10957,6 +10947,8 @@ static void disas_simd_three_reg_same_extra(DisasContext *s, uint32_t insn) } break; default: + case 0x10: /* SQRDMLAH (vector) */ + case 0x11: /* SQRDMLSH (vector) */ unallocated_encoding(s); return; } @@ -10969,14 +10961,6 @@ static void disas_simd_three_reg_same_extra(DisasContext *s, uint32_t insn) } switch (opcode) { - case 0x0: /* SQRDMLAH (vector) */ - gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sqrdmlah_qc, size); - return; - - case 0x1: /* SQRDMLSH (vector) */ - gen_gvec_fn3(s, is_q, rd, rn, rm, gen_gvec_sqrdmlsh_qc, size); - return; - case 0x2: /* SDOT / UDOT */ gen_gvec_op4_ool(s, is_q, rd, rn, rm, rd, 0, u ? gen_helper_gvec_udot_b : gen_helper_gvec_sdot_b); @@ -12059,13 +12043,6 @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) case 0x0b: /* SQDMULL, SQDMULL2 */ is_long = true; break; - case 0x1d: /* SQRDMLAH */ - case 0x1f: /* SQRDMLSH */ - if (!dc_isar_feature(aa64_rdm, s)) { - unallocated_encoding(s); - return; - } - break; case 0x0e: /* SDOT */ case 0x1e: /* UDOT */ if (is_scalar || size != MO_32 || !dc_isar_feature(aa64_dp, s)) { @@ -12127,6 +12104,8 @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) case 0x18: /* FMLAL2 */ case 0x19: /* FMULX */ case 0x1c: /* FMLSL2 */ + case 0x1d: /* SQRDMLAH */ + case 0x1f: /* SQRDMLSH */ unallocated_encoding(s); return; } @@ -12320,33 +12299,13 @@ static void disas_simd_indexed(DisasContext *s, uint32_t insn) tcg_op, tcg_idx); } break; - case 0x1d: /* SQRDMLAH */ - read_vec_element_i32(s, tcg_res, rd, pass, - is_scalar ? size : MO_32); - if (size == 1) { - gen_helper_neon_qrdmlah_s16(tcg_res, tcg_env, - tcg_op, tcg_idx, tcg_res); - } else { - gen_helper_neon_qrdmlah_s32(tcg_res, tcg_env, - tcg_op, tcg_idx, tcg_res); - } - break; - case 0x1f: /* SQRDMLSH */ - read_vec_element_i32(s, tcg_res, rd, pass, - is_scalar ? size : MO_32); - if (size == 1) { - gen_helper_neon_qrdmlsh_s16(tcg_res, tcg_env, - tcg_op, tcg_idx, tcg_res); - } else { - gen_helper_neon_qrdmlsh_s32(tcg_res, tcg_env, - tcg_op, tcg_idx, tcg_res); - } - break; default: case 0x01: /* FMLA */ case 0x05: /* FMLS */ case 0x09: /* FMUL */ case 0x19: /* FMULX */ + case 0x1d: /* SQRDMLAH */ + case 0x1f: /* SQRDMLSH */ g_assert_not_reached(); } @@ -12538,7 +12497,6 @@ static const AArch64DecodeTable data_proc_simd[] = { { 0x0e000000, 0xbf208c00, disas_simd_tb }, { 0x0e000800, 0xbf208c00, disas_simd_zip_trn }, { 0x2e000000, 0xbf208400, disas_simd_ext }, - { 0x5e008400, 0xdf208400, disas_simd_scalar_three_reg_same_extra }, { 0x5e200000, 0xdf200c00, disas_simd_scalar_three_reg_diff }, { 0x5e200800, 0xdf3e0c00, disas_simd_scalar_two_reg_misc }, { 0x5f000000, 0xdf000400, disas_simd_indexed }, /* scalar indexed */ diff --git a/target/arm/tcg/vec_helper.c b/target/arm/tcg/vec_helper.c index d477479bb1..98604d170f 100644 --- a/target/arm/tcg/vec_helper.c +++ b/target/arm/tcg/vec_helper.c @@ -347,6 +347,42 @@ void HELPER(neon_sqrdmulh_idx_h)(void *vd, void *vn, void *vm, clear_tail(d, opr_sz, simd_maxsz(desc)); } +void HELPER(neon_sqrdmlah_idx_h)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, j, opr_sz = simd_oprsz(desc); + int idx = simd_data(desc); + int16_t *d = vd, *n = vn, *m = (int16_t *)vm + H2(idx); + intptr_t elements = opr_sz / 2; + intptr_t eltspersegment = MIN(16 / 2, elements); + + for (i = 0; i < elements; i += 16 / 2) { + int16_t mm = m[i]; + for (j = 0; j < eltspersegment; ++j) { + d[i + j] = do_sqrdmlah_h(n[i + j], mm, d[i + j], false, true, vq); + } + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + +void HELPER(neon_sqrdmlsh_idx_h)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, j, opr_sz = simd_oprsz(desc); + int idx = simd_data(desc); + int16_t *d = vd, *n = vn, *m = (int16_t *)vm + H2(idx); + intptr_t elements = opr_sz / 2; + intptr_t eltspersegment = MIN(16 / 2, elements); + + for (i = 0; i < elements; i += 16 / 2) { + int16_t mm = m[i]; + for (j = 0; j < eltspersegment; ++j) { + d[i + j] = do_sqrdmlah_h(n[i + j], mm, d[i + j], true, true, vq); + } + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + void HELPER(sve2_sqrdmlah_h)(void *vd, void *vn, void *vm, void *va, uint32_t desc) { @@ -546,6 +582,42 @@ void HELPER(neon_sqrdmulh_idx_s)(void *vd, void *vn, void *vm, clear_tail(d, opr_sz, simd_maxsz(desc)); } +void HELPER(neon_sqrdmlah_idx_s)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, j, opr_sz = simd_oprsz(desc); + int idx = simd_data(desc); + int32_t *d = vd, *n = vn, *m = (int32_t *)vm + H4(idx); + intptr_t elements = opr_sz / 4; + intptr_t eltspersegment = MIN(16 / 4, elements); + + for (i = 0; i < elements; i += 16 / 4) { + int32_t mm = m[i]; + for (j = 0; j < eltspersegment; ++j) { + d[i + j] = do_sqrdmlah_s(n[i + j], mm, d[i + j], false, true, vq); + } + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + +void HELPER(neon_sqrdmlsh_idx_s)(void *vd, void *vn, void *vm, + void *vq, uint32_t desc) +{ + intptr_t i, j, opr_sz = simd_oprsz(desc); + int idx = simd_data(desc); + int32_t *d = vd, *n = vn, *m = (int32_t *)vm + H4(idx); + intptr_t elements = opr_sz / 4; + intptr_t eltspersegment = MIN(16 / 4, elements); + + for (i = 0; i < elements; i += 16 / 4) { + int32_t mm = m[i]; + for (j = 0; j < eltspersegment; ++j) { + d[i + j] = do_sqrdmlah_s(n[i + j], mm, d[i + j], true, true, vq); + } + } + clear_tail(d, opr_sz, simd_maxsz(desc)); +} + void HELPER(sve2_sqrdmlah_s)(void *vd, void *vn, void *vm, void *va, uint32_t desc) {

[04/13] target/arm: Convert SQRDMLAH, SQRDMLSH to decodetree

Commit Message

Comments

Patch