[14/23] target/arm: Implement SVE Integer Arithmetic - Unary Predicated Group

Message ID	20171218174552.18871-15-richard.henderson@linaro.org
State	Superseded
Headers	show Delivered-To: patch@linaro.org Received-SPF: pass (google.com: domain of qemu-devel-bounces+patch=linaro.org@nongnu.org designates 2001:4830:134:3::11 as permitted sender) client-ip=2001:4830:134:3::11; From: Richard Henderson <richard.henderson@linaro.org> To: qemu-devel@nongnu.org Date: Mon, 18 Dec 2017 09:45:43 -0800 Message-Id: <20171218174552.18871-15-richard.henderson@linaro.org> In-Reply-To: <20171218174552.18871-1-richard.henderson@linaro.org> References: <20171218174552.18871-1-richard.henderson@linaro.org> Subject: [Qemu-devel] [PATCH 14/23] target/arm: Implement SVE Integer Arithmetic - Unary Predicated Group Precedence: list Cc: peter.maydell@linaro.org, qemu-arm@nongnu.org Errors-To: qemu-devel-bounces+patch=linaro.org@nongnu.org Sender: "Qemu-devel" <qemu-devel-bounces+patch=linaro.org@nongnu.org>
Series	target/arm: decode generator and initial sve patches \| expand [RFC,00/23] target/arm: decode generator and initial sve patches [01/23] scripts: Add decodetree.py [02/23] target/arm: Add SVE decode skeleton [03/23] target/arm: Implement SVE Bitwise Logical - Unpredicated Group [04/23] target/arm: Implement PTRUE, PFALSE, SETFFR [05/23] target/arm: Implement SVE predicate logical operations [06/23] target/arm: Implement SVE load vector/predicate [07/23] target/arm: Implement SVE Integer Binary Arithmetic - Predicated Group [08/23] target/arm: Handle SVE registers in write_fp_dreg [09/23] target/arm: Handle SVE registers when using clear_vec_high [10/23] target/arm: Implement SVE Integer Reduction Group [11/23] target/arm: Implement SVE bitwise shift by immediate (predicated) [12/23] target/arm: Implement SVE bitwise shift by vector (predicated) [13/23] target/arm: Implement SVE bitwise shift by wide elements (predicated) [14/23] target/arm: Implement SVE Integer Arithmetic - Unary Predicated Group [15/23] target/arm: Implement SVE Integer Multiply-Add Group [16/23] target/arm: Implement SVE Integer Arithmetic - Unpredicated Group [17/23] target/arm: Implement SVE Index Generation Group [18/23] target/arm: Implement SVE Stack Allocation Group [19/23] target/arm: Implement SVE Bitwise Shift - Unpredicated Group [20/23] target/arm: Implement SVE Compute Vector Address Group [21/23] target/arm: Implement SVE floating-point exponential accelerator [22/23] target/arm: Implement SVE floating-point trig select coefficient [23/23] target/arm: Implement SVE Element Count Group, register destinations

diff --git a/target/arm/helper-sve.h b/target/arm/helper-sve.h index a2db3e2fd9..e9382de300 100644 --- a/target/arm/helper-sve.h +++ b/target/arm/helper-sve.h @@ -279,6 +279,66 @@ DEF_HELPER_FLAGS_4(sve_asrd_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_asrd_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_4(sve_asrd_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cls_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cls_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cls_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cls_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_clz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_clz_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_clz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_clz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_cnt_zpz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cnt_zpz_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cnt_zpz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cnt_zpz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_cnot_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cnot_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cnot_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_cnot_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_fabs_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_fabs_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_fabs_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_fneg_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_fneg_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_fneg_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_not_zpz_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_not_zpz_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_not_zpz_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_not_zpz_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_sxtb_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_sxtb_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_sxtb_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_uxtb_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_uxtb_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_uxtb_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_sxth_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_sxth_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_uxth_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_uxth_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_sxtw_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_uxtw_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_abs_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_abs_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_abs_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_abs_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_4(sve_neg_b, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_neg_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_neg_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_4(sve_neg_d, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, i32) + DEF_HELPER_FLAGS_5(sve_and_pred, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve_bic_pred, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve_eor_pred, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) diff --git a/target/arm/sve_helper.c b/target/arm/sve_helper.c index 3be6d1ae05..481b3bdefe 100644 --- a/target/arm/sve_helper.c +++ b/target/arm/sve_helper.c @@ -553,6 +553,134 @@ DO_ZPZW(sve_lsl_zpzw_s, uint32_t, uint64_t, H1_4, DO_LSL) #undef DO_ZPZW +/* Fully general two-operand expander, controlled by a predicate. + */ +#define DO_ZPZ(NAME, TYPE, H, OP) \ +void HELPER(NAME)(void *vd, void *vn, void *vg, uint32_t desc) \ +{ \ + intptr_t iv = 0, ib = 0, opr_sz = simd_oprsz(desc); \ + for (iv = ib = 0; iv < opr_sz; iv += 16, ib += 2) { \ + uint16_t pg = *(uint16_t *)(vg + H2(ib)); \ + intptr_t i = 0; \ + do { \ + if (pg & 1) { \ + TYPE nn = *(TYPE *)(vn + iv + H(i)); \ + *(TYPE *)(vd + iv + H(i)) = OP(nn); \ + } \ + i += sizeof(TYPE), pg >>= sizeof(TYPE); \ + } while (pg); \ + } \ +} + +/* Similarly, specialized for 64-bit operands. */ +#define DO_ZPZ_D(NAME, TYPE, OP) \ +void HELPER(NAME)(void *vd, void *vn, void *vg, uint32_t desc) \ +{ \ + intptr_t i, opr_sz = simd_oprsz(desc) / 8; \ + TYPE *d = vd, *n = vn; \ + uint8_t *pg = vg; \ + for (i = 0; i < opr_sz; i += 1) { \ + if (pg[H1(i)] & 1) { \ + TYPE nn = n[i]; \ + d[i] = OP(nn); \ + } \ + } \ +} + +#define DO_CLS_B(N) (clrsb32(N) - 24) +#define DO_CLS_H(N) (clrsb32(N) - 16) + +DO_ZPZ(sve_cls_b, int8_t, H1, DO_CLS_B) +DO_ZPZ(sve_cls_h, int16_t, H1_2, DO_CLS_H) +DO_ZPZ(sve_cls_s, int32_t, H1_4, clrsb32) +DO_ZPZ_D(sve_cls_d, int64_t, clrsb64) + +#define DO_CLZ_B(N) (clz32(N) - 24) +#define DO_CLZ_H(N) (clz32(N) - 16) + +DO_ZPZ(sve_clz_b, uint8_t, H1, DO_CLZ_B) +DO_ZPZ(sve_clz_h, uint16_t, H1_2, DO_CLZ_H) +DO_ZPZ(sve_clz_s, uint32_t, H1_4, clz32) +DO_ZPZ_D(sve_clz_d, uint64_t, clz64) + +DO_ZPZ(sve_cnt_zpz_b, uint8_t, H1, ctpop8) +DO_ZPZ(sve_cnt_zpz_h, uint16_t, H1_2, ctpop16) +DO_ZPZ(sve_cnt_zpz_s, uint32_t, H1_4, ctpop32) +DO_ZPZ_D(sve_cnt_zpz_d, uint64_t, ctpop64) + +#define DO_CNOT(N) (N == 0) + +DO_ZPZ(sve_cnot_b, uint8_t, H1, DO_CNOT) +DO_ZPZ(sve_cnot_h, uint16_t, H1_2, DO_CNOT) +DO_ZPZ(sve_cnot_s, uint32_t, H1_4, DO_CNOT) +DO_ZPZ_D(sve_cnot_d, uint64_t, DO_CNOT) + +#define DO_FABS(N) (N & ((__typeof(N))-1 >> 1)) + +DO_ZPZ(sve_fabs_h, uint16_t, H1_2, DO_FABS) +DO_ZPZ(sve_fabs_s, uint32_t, H1_4, DO_FABS) +DO_ZPZ_D(sve_fabs_d, uint64_t, DO_FABS) + +#define DO_FNEG(N) (N ^ ~((__typeof(N))-1 >> 1)) + +DO_ZPZ(sve_fneg_h, uint16_t, H1_2, DO_FNEG) +DO_ZPZ(sve_fneg_s, uint32_t, H1_4, DO_FNEG) +DO_ZPZ_D(sve_fneg_d, uint64_t, DO_FNEG) + +#define DO_NOT(N) (~N) + +DO_ZPZ(sve_not_zpz_b, uint8_t, H1, DO_NOT) +DO_ZPZ(sve_not_zpz_h, uint16_t, H1_2, DO_NOT) +DO_ZPZ(sve_not_zpz_s, uint32_t, H1_4, DO_NOT) +DO_ZPZ_D(sve_not_zpz_d, uint64_t, DO_NOT) + +#define DO_SXTB(N) ((int8_t)N) +#define DO_SXTH(N) ((int16_t)N) +#define DO_SXTS(N) ((int32_t)N) +#define DO_UXTB(N) ((uint8_t)N) +#define DO_UXTH(N) ((uint16_t)N) +#define DO_UXTS(N) ((uint32_t)N) + +DO_ZPZ(sve_sxtb_h, uint16_t, H1_2, DO_SXTB) +DO_ZPZ(sve_sxtb_s, uint32_t, H1_4, DO_SXTB) +DO_ZPZ(sve_sxth_s, uint32_t, H1_4, DO_SXTH) +DO_ZPZ_D(sve_sxtb_d, uint64_t, DO_SXTB) +DO_ZPZ_D(sve_sxth_d, uint64_t, DO_SXTH) +DO_ZPZ_D(sve_sxtw_d, uint64_t, DO_SXTS) + +DO_ZPZ(sve_uxtb_h, uint16_t, H1_2, DO_UXTB) +DO_ZPZ(sve_uxtb_s, uint32_t, H1_4, DO_UXTB) +DO_ZPZ(sve_uxth_s, uint32_t, H1_4, DO_UXTH) +DO_ZPZ_D(sve_uxtb_d, uint64_t, DO_UXTB) +DO_ZPZ_D(sve_uxth_d, uint64_t, DO_UXTH) +DO_ZPZ_D(sve_uxtw_d, uint64_t, DO_UXTS) + +#define DO_ABS(N) (N < 0 ? -N : N) + +DO_ZPZ(sve_abs_b, int8_t, H1, DO_ABS) +DO_ZPZ(sve_abs_h, int16_t, H1_2, DO_ABS) +DO_ZPZ(sve_abs_s, int32_t, H1_4, DO_ABS) +DO_ZPZ_D(sve_abs_d, int64_t, DO_ABS) + +#define DO_NEG(N) (-N) + +DO_ZPZ(sve_neg_b, uint8_t, H1, DO_NEG) +DO_ZPZ(sve_neg_h, uint16_t, H1_2, DO_NEG) +DO_ZPZ(sve_neg_s, uint32_t, H1_4, DO_NEG) +DO_ZPZ_D(sve_neg_d, uint64_t, DO_NEG) + +#undef DO_CLS_B +#undef DO_CLS_H +#undef DO_CLZ_B +#undef DO_CLZ_H +#undef DO_CNOT +#undef DO_FABS +#undef DO_FNEG +#undef DO_ABS +#undef DO_NEG +#undef DO_ZPZ +#undef DO_ZPZ_D + /* Two-operand reduction expander, controlled by a predicate. * The difference between TYPERED and TYPERET has to do with * sign-extension. E.g. for SMAX, TYPERED must be signed, diff --git a/target/arm/translate-sve.c b/target/arm/translate-sve.c index 91f07d57e3..7dbc43fb6e 100644 --- a/target/arm/translate-sve.c +++ b/target/arm/translate-sve.c @@ -250,6 +250,113 @@ DO_ZPZW(LSL, lsl) #undef DO_ZPZW +static void do_zpz_ool(DisasContext *s, arg_rpr_esz *a, gen_helper_gvec_3 *fn) +{ + unsigned vsz = size_for_gvec(vec_full_reg_size(s)); + if (fn == NULL) { + unallocated_encoding(s); + return; + } + tcg_gen_gvec_3_ool(vec_full_reg_offset(s, a->rd), + vec_full_reg_offset(s, a->rn), + pred_full_reg_offset(s, a->pg), + vsz, vsz, 0, fn); +} + +#define DO_ZPZ(NAME, name) \ +void trans_##NAME(DisasContext *s, arg_rpr_esz *a, uint32_t insn) \ +{ \ + static gen_helper_gvec_3 * const fns[4] = { \ + gen_helper_sve_##name##_b, gen_helper_sve_##name##_h, \ + gen_helper_sve_##name##_s, gen_helper_sve_##name##_d, \ + }; \ + do_zpz_ool(s, a, fns[a->esz]); \ +} + +DO_ZPZ(CLS, cls) +DO_ZPZ(CLZ, clz) +DO_ZPZ(CNT_zpz, cnt_zpz) +DO_ZPZ(CNOT, cnot) +DO_ZPZ(NOT_zpz, not_zpz) +DO_ZPZ(ABS, abs) +DO_ZPZ(NEG, neg) + +void trans_FABS(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + static gen_helper_gvec_3 * const fns[4] = { + NULL, + gen_helper_sve_fabs_h, + gen_helper_sve_fabs_s, + gen_helper_sve_fabs_d + }; + do_zpz_ool(s, a, fns[a->esz]); +} + +void trans_FNEG(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + static gen_helper_gvec_3 * const fns[4] = { + NULL, + gen_helper_sve_fneg_h, + gen_helper_sve_fneg_s, + gen_helper_sve_fneg_d + }; + do_zpz_ool(s, a, fns[a->esz]); +} + +void trans_SXTB(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + static gen_helper_gvec_3 * const fns[4] = { + NULL, + gen_helper_sve_sxtb_h, + gen_helper_sve_sxtb_s, + gen_helper_sve_sxtb_d + }; + do_zpz_ool(s, a, fns[a->esz]); +} + +void trans_UXTB(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + static gen_helper_gvec_3 * const fns[4] = { + NULL, + gen_helper_sve_uxtb_h, + gen_helper_sve_uxtb_s, + gen_helper_sve_uxtb_d + }; + do_zpz_ool(s, a, fns[a->esz]); +} + +void trans_SXTH(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + static gen_helper_gvec_3 * const fns[4] = { + NULL, NULL, + gen_helper_sve_sxth_s, + gen_helper_sve_sxth_d + }; + do_zpz_ool(s, a, fns[a->esz]); +} + +void trans_UXTH(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + static gen_helper_gvec_3 * const fns[4] = { + NULL, NULL, + gen_helper_sve_uxth_s, + gen_helper_sve_uxth_d + }; + do_zpz_ool(s, a, fns[a->esz]); +} + +void trans_SXTW(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + do_zpz_ool(s, a, a->esz == 3 ? gen_helper_sve_sxtw_d : NULL); +} + +void trans_UXTW(DisasContext *s, arg_rpr_esz *a, uint32_t insn) +{ + do_zpz_ool(s, a, a->esz == 3 ? gen_helper_sve_uxtw_d : NULL); +} + +#undef DO_ZPZ + typedef void gen_helper_gvec_reduc(TCGv_i64, TCGv_ptr, TCGv_ptr, TCGv_i32); static void do_vpz_ool(DisasContext *s, arg_rpr_esz *a, gen_helper_gvec_reduc *fn) diff --git a/target/arm/sve.def b/target/arm/sve.def index 66be950ca5..955a0275a1 100644 --- a/target/arm/sve.def +++ b/target/arm/sve.def @@ -147,6 +147,27 @@ ASR_zpzw 00000100 .. 011 000 100 ... ..... ..... @rdn_pg_rm_esz LSR_zpzw 00000100 .. 011 001 100 ... ..... ..... @rdn_pg_rm_esz LSL_zpzw 00000100 .. 011 011 100 ... ..... ..... @rdn_pg_rm_esz +### SVE Integer Arithmetic - Unary Predicated Group + +# SVE unary bit operations (predicated) +CLS 00000100 .. 011 000 101 ... ..... ..... @rd_pg_rn_esz +CLZ 00000100 .. 011 001 101 ... ..... ..... @rd_pg_rn_esz +CNT_zpz 00000100 .. 011 010 101 ... ..... ..... @rd_pg_rn_esz +CNOT 00000100 .. 011 011 101 ... ..... ..... @rd_pg_rn_esz +NOT_zpz 00000100 .. 011 110 101 ... ..... ..... @rd_pg_rn_esz +FABS 00000100 .. 011 100 101 ... ..... ..... @rd_pg_rn_esz # Note size != 0 +FNEG 00000100 .. 011 101 101 ... ..... ..... @rd_pg_rn_esz # Note size != 0 + +# SVE integer unary operations (predicated) +ABS 00000100 .. 010 110 101 ... ..... ..... @rd_pg_rn_esz +NEG 00000100 .. 010 111 101 ... ..... ..... @rd_pg_rn_esz +SXTB 00000100 .. 010 000 101 ... ..... ..... @rd_pg_rn_esz # Note size != 0 +UXTB 00000100 .. 010 001 101 ... ..... ..... @rd_pg_rn_esz # Note size != 0 +SXTH 00000100 .. 010 010 101 ... ..... ..... @rd_pg_rn_esz # Note size > 1 +UXTH 00000100 .. 010 011 101 ... ..... ..... @rd_pg_rn_esz # Note size > 1 +SXTW 00000100 .. 010 100 101 ... ..... ..... @rd_pg_rn_esz # Note size == 3 +UXTW 00000100 .. 010 101 101 ... ..... ..... @rd_pg_rn_esz # Note size == 3 + ### SVE Logical - Unpredicated Group # SVE bitwise logical operations (unpredicated)

[14/23] target/arm: Implement SVE Integer Arithmetic - Unary Predicated Group

Commit Message

Patch