From d8e81e3c18231e4dbab562f53e3e6cf7ad77735f Mon Sep 17 00:00:00 2001 From: Weiwei Li Date: Sat, 23 Apr 2022 10:34:58 +0800 Subject: target/riscv: rvk: add support for zbkb extension - reuse partial instructions of zbb extension, update extension check for them - add brev8, pack, packh, packw, unzip, zip instructions Signed-off-by: Weiwei Li Signed-off-by: Junqiang Wang Acked-by: Alistair Francis Reviewed-by: Richard Henderson Message-Id: <20220423023510.30794-3-liweiwei@iscas.ac.cn> Signed-off-by: Alistair Francis --- target/riscv/bitmanip_helper.c | 53 +++++++++++++++++++ target/riscv/helper.h | 3 ++ target/riscv/insn32.decode | 45 ++++++++++------ target/riscv/insn_trans/trans_rvb.c.inc | 94 ++++++++++++++++++++++++++++----- target/riscv/translate.c | 7 +++ 5 files changed, 174 insertions(+), 28 deletions(-) (limited to 'target/riscv') diff --git a/target/riscv/bitmanip_helper.c b/target/riscv/bitmanip_helper.c index f1b5e55..e003e8b 100644 --- a/target/riscv/bitmanip_helper.c +++ b/target/riscv/bitmanip_helper.c @@ -49,3 +49,56 @@ target_ulong HELPER(clmulr)(target_ulong rs1, target_ulong rs2) return result; } + +static inline target_ulong do_swap(target_ulong x, uint64_t mask, int shift) +{ + return ((x & mask) << shift) | ((x & ~mask) >> shift); +} + +target_ulong HELPER(brev8)(target_ulong rs1) +{ + target_ulong x = rs1; + + x = do_swap(x, 0x5555555555555555ull, 1); + x = do_swap(x, 0x3333333333333333ull, 2); + x = do_swap(x, 0x0f0f0f0f0f0f0f0full, 4); + return x; +} + +static const uint64_t shuf_masks[] = { + dup_const(MO_8, 0x44), + dup_const(MO_8, 0x30), + dup_const(MO_16, 0x0f00), + dup_const(MO_32, 0xff0000) +}; + +static inline target_ulong do_shuf_stage(target_ulong src, uint64_t maskL, + uint64_t maskR, int shift) +{ + target_ulong x = src & ~(maskL | maskR); + + x |= ((src << shift) & maskL) | ((src >> shift) & maskR); + return x; +} + +target_ulong HELPER(unzip)(target_ulong rs1) +{ + target_ulong x = rs1; + + x = do_shuf_stage(x, shuf_masks[0], shuf_masks[0] >> 1, 1); + x = do_shuf_stage(x, shuf_masks[1], shuf_masks[1] >> 2, 2); + x = do_shuf_stage(x, shuf_masks[2], shuf_masks[2] >> 4, 4); + x = do_shuf_stage(x, shuf_masks[3], shuf_masks[3] >> 8, 8); + return x; +} + +target_ulong HELPER(zip)(target_ulong rs1) +{ + target_ulong x = rs1; + + x = do_shuf_stage(x, shuf_masks[3], shuf_masks[3] >> 8, 8); + x = do_shuf_stage(x, shuf_masks[2], shuf_masks[2] >> 4, 4); + x = do_shuf_stage(x, shuf_masks[1], shuf_masks[1] >> 2, 2); + x = do_shuf_stage(x, shuf_masks[0], shuf_masks[0] >> 1, 1); + return x; +} diff --git a/target/riscv/helper.h b/target/riscv/helper.h index a669d01..8a3a761 100644 --- a/target/riscv/helper.h +++ b/target/riscv/helper.h @@ -66,6 +66,9 @@ DEF_HELPER_FLAGS_1(fclass_d, TCG_CALL_NO_RWG_SE, tl, i64) /* Bitmanip */ DEF_HELPER_FLAGS_2(clmul, TCG_CALL_NO_RWG_SE, tl, tl, tl) DEF_HELPER_FLAGS_2(clmulr, TCG_CALL_NO_RWG_SE, tl, tl, tl) +DEF_HELPER_FLAGS_1(brev8, TCG_CALL_NO_RWG_SE, tl, tl) +DEF_HELPER_FLAGS_1(unzip, TCG_CALL_NO_RWG_SE, tl, tl) +DEF_HELPER_FLAGS_1(zip, TCG_CALL_NO_RWG_SE, tl, tl) /* Floating Point - Half Precision */ DEF_HELPER_FLAGS_3(fadd_h, TCG_CALL_NO_RWG, i64, env, i64, i64) diff --git a/target/riscv/insn32.decode b/target/riscv/insn32.decode index 1d3ff1e..fdceaf6 100644 --- a/target/riscv/insn32.decode +++ b/target/riscv/insn32.decode @@ -717,8 +717,22 @@ sh2add_uw 0010000 .......... 100 ..... 0111011 @r sh3add_uw 0010000 .......... 110 ..... 0111011 @r slli_uw 00001 ............ 001 ..... 0011011 @sh -# *** RV32 Zbb Standard Extension *** +# *** RV32 Zbb/Zbkb Standard Extension *** andn 0100000 .......... 111 ..... 0110011 @r +rol 0110000 .......... 001 ..... 0110011 @r +ror 0110000 .......... 101 ..... 0110011 @r +rori 01100 ............ 101 ..... 0010011 @sh +# The encoding for rev8 differs between RV32 and RV64. +# rev8_32 denotes the RV32 variant. +rev8_32 011010 011000 ..... 101 ..... 0010011 @r2 +# The encoding for zext.h differs between RV32 and RV64. +# zext_h_32 denotes the RV32 variant. +{ + zext_h_32 0000100 00000 ..... 100 ..... 0110011 @r2 + pack 0000100 ..... ..... 100 ..... 0110011 @r +} +xnor 0100000 .......... 100 ..... 0110011 @r +# *** RV32 extra Zbb Standard Extension *** clz 011000 000000 ..... 001 ..... 0010011 @r2 cpop 011000 000010 ..... 001 ..... 0010011 @r2 ctz 011000 000001 ..... 001 ..... 0010011 @r2 @@ -728,23 +742,15 @@ min 0000101 .......... 100 ..... 0110011 @r minu 0000101 .......... 101 ..... 0110011 @r orc_b 001010 000111 ..... 101 ..... 0010011 @r2 orn 0100000 .......... 110 ..... 0110011 @r -# The encoding for rev8 differs between RV32 and RV64. -# rev8_32 denotes the RV32 variant. -rev8_32 011010 011000 ..... 101 ..... 0010011 @r2 -rol 0110000 .......... 001 ..... 0110011 @r -ror 0110000 .......... 101 ..... 0110011 @r -rori 01100 ............ 101 ..... 0010011 @sh sext_b 011000 000100 ..... 001 ..... 0010011 @r2 sext_h 011000 000101 ..... 001 ..... 0010011 @r2 -xnor 0100000 .......... 100 ..... 0110011 @r -# The encoding for zext.h differs between RV32 and RV64. -# zext_h_32 denotes the RV32 variant. -zext_h_32 0000100 00000 ..... 100 ..... 0110011 @r2 +# *** RV32 extra Zbkb Standard Extension *** +brev8 0110100 00111 ..... 101 ..... 0010011 @r2 #grevi +packh 0000100 .......... 111 ..... 0110011 @r +unzip 0000100 01111 ..... 101 ..... 0010011 @r2 #unshfl +zip 0000100 01111 ..... 001 ..... 0010011 @r2 #shfl -# *** RV64 Zbb Standard Extension (in addition to RV32 Zbb) *** -clzw 0110000 00000 ..... 001 ..... 0011011 @r2 -ctzw 0110000 00001 ..... 001 ..... 0011011 @r2 -cpopw 0110000 00010 ..... 001 ..... 0011011 @r2 +# *** RV64 Zbb/Zbkb Standard Extension (in addition to RV32 Zbb/Zbkb) *** # The encoding for rev8 differs between RV32 and RV64. # When executing on RV64, the encoding used in RV32 is an illegal # instruction, so we use different handler functions to differentiate. @@ -755,7 +761,14 @@ rorw 0110000 .......... 101 ..... 0111011 @r # The encoding for zext.h differs between RV32 and RV64. # When executing on RV64, the encoding used in RV32 is an illegal # instruction, so we use different handler functions to differentiate. -zext_h_64 0000100 00000 ..... 100 ..... 0111011 @r2 +{ + zext_h_64 0000100 00000 ..... 100 ..... 0111011 @r2 + packw 0000100 ..... ..... 100 ..... 0111011 @r +} +# *** RV64 extra Zbb Standard Extension (in addition to RV32 Zbb) *** +clzw 0110000 00000 ..... 001 ..... 0011011 @r2 +ctzw 0110000 00001 ..... 001 ..... 0011011 @r2 +cpopw 0110000 00010 ..... 001 ..... 0011011 @r2 # *** RV32 Zbc Standard Extension *** clmul 0000101 .......... 001 ..... 0110011 @r diff --git a/target/riscv/insn_trans/trans_rvb.c.inc b/target/riscv/insn_trans/trans_rvb.c.inc index e8519a6..f365775 100644 --- a/target/riscv/insn_trans/trans_rvb.c.inc +++ b/target/riscv/insn_trans/trans_rvb.c.inc @@ -1,5 +1,5 @@ /* - * RISC-V translation routines for the Zb[abcs] Standard Extension. + * RISC-V translation routines for the Zb[abcs] and Zbk[bcx] Standard Extension. * * Copyright (c) 2020 Kito Cheng, kito.cheng@sifive.com * Copyright (c) 2020 Frank Chang, frank.chang@sifive.com @@ -42,6 +42,12 @@ } \ } while (0) +#define REQUIRE_ZBKB(ctx) do { \ + if (!ctx->cfg_ptr->ext_zbkb) { \ + return false; \ + } \ +} while (0) + static void gen_clz(TCGv ret, TCGv arg1) { tcg_gen_clzi_tl(ret, arg1, TARGET_LONG_BITS); @@ -85,19 +91,19 @@ static bool trans_cpop(DisasContext *ctx, arg_cpop *a) static bool trans_andn(DisasContext *ctx, arg_andn *a) { - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_logic(ctx, a, tcg_gen_andc_tl); } static bool trans_orn(DisasContext *ctx, arg_orn *a) { - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_logic(ctx, a, tcg_gen_orc_tl); } static bool trans_xnor(DisasContext *ctx, arg_xnor *a) { - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_logic(ctx, a, tcg_gen_eqv_tl); } @@ -247,7 +253,7 @@ static void gen_rorw(TCGv ret, TCGv arg1, TCGv arg2) static bool trans_ror(DisasContext *ctx, arg_ror *a) { - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_shift_per_ol(ctx, a, EXT_NONE, tcg_gen_rotr_tl, gen_rorw, NULL); } @@ -264,7 +270,7 @@ static void gen_roriw(TCGv ret, TCGv arg1, target_long shamt) static bool trans_rori(DisasContext *ctx, arg_rori *a) { - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_shift_imm_fn_per_ol(ctx, a, EXT_NONE, tcg_gen_rotri_tl, gen_roriw, NULL); } @@ -289,7 +295,7 @@ static void gen_rolw(TCGv ret, TCGv arg1, TCGv arg2) static bool trans_rol(DisasContext *ctx, arg_rol *a) { - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_shift_per_ol(ctx, a, EXT_NONE, tcg_gen_rotl_tl, gen_rolw, NULL); } @@ -301,14 +307,14 @@ static void gen_rev8_32(TCGv ret, TCGv src1) static bool trans_rev8_32(DisasContext *ctx, arg_rev8_32 *a) { REQUIRE_32BIT(ctx); - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_unary(ctx, a, EXT_NONE, gen_rev8_32); } static bool trans_rev8_64(DisasContext *ctx, arg_rev8_64 *a) { REQUIRE_64BIT(ctx); - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); return gen_unary(ctx, a, EXT_NONE, tcg_gen_bswap_tl); } @@ -403,7 +409,7 @@ static bool trans_cpopw(DisasContext *ctx, arg_cpopw *a) static bool trans_rorw(DisasContext *ctx, arg_rorw *a) { REQUIRE_64BIT(ctx); - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); ctx->ol = MXL_RV32; return gen_shift(ctx, a, EXT_NONE, gen_rorw, NULL); } @@ -411,7 +417,7 @@ static bool trans_rorw(DisasContext *ctx, arg_rorw *a) static bool trans_roriw(DisasContext *ctx, arg_roriw *a) { REQUIRE_64BIT(ctx); - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); ctx->ol = MXL_RV32; return gen_shift_imm_fn(ctx, a, EXT_NONE, gen_roriw, NULL); } @@ -419,7 +425,7 @@ static bool trans_roriw(DisasContext *ctx, arg_roriw *a) static bool trans_rolw(DisasContext *ctx, arg_rolw *a) { REQUIRE_64BIT(ctx); - REQUIRE_ZBB(ctx); + REQUIRE_EITHER_EXT(ctx, zbb, zbkb); ctx->ol = MXL_RV32; return gen_shift(ctx, a, EXT_NONE, gen_rolw, NULL); } @@ -504,3 +510,67 @@ static bool trans_clmulr(DisasContext *ctx, arg_clmulh *a) REQUIRE_ZBC(ctx); return gen_arith(ctx, a, EXT_NONE, gen_helper_clmulr, NULL); } + +static void gen_pack(TCGv ret, TCGv src1, TCGv src2) +{ + tcg_gen_deposit_tl(ret, src1, src2, + TARGET_LONG_BITS / 2, + TARGET_LONG_BITS / 2); +} + +static void gen_packh(TCGv ret, TCGv src1, TCGv src2) +{ + TCGv t = tcg_temp_new(); + + tcg_gen_ext8u_tl(t, src2); + tcg_gen_deposit_tl(ret, src1, t, 8, TARGET_LONG_BITS - 8); + tcg_temp_free(t); +} + +static void gen_packw(TCGv ret, TCGv src1, TCGv src2) +{ + TCGv t = tcg_temp_new(); + + tcg_gen_ext16s_tl(t, src2); + tcg_gen_deposit_tl(ret, src1, t, 16, TARGET_LONG_BITS - 16); + tcg_temp_free(t); +} + +static bool trans_brev8(DisasContext *ctx, arg_brev8 *a) +{ + REQUIRE_ZBKB(ctx); + return gen_unary(ctx, a, EXT_NONE, gen_helper_brev8); +} + +static bool trans_pack(DisasContext *ctx, arg_pack *a) +{ + REQUIRE_ZBKB(ctx); + return gen_arith(ctx, a, EXT_NONE, gen_pack, NULL); +} + +static bool trans_packh(DisasContext *ctx, arg_packh *a) +{ + REQUIRE_ZBKB(ctx); + return gen_arith(ctx, a, EXT_NONE, gen_packh, NULL); +} + +static bool trans_packw(DisasContext *ctx, arg_packw *a) +{ + REQUIRE_64BIT(ctx); + REQUIRE_ZBKB(ctx); + return gen_arith(ctx, a, EXT_NONE, gen_packw, NULL); +} + +static bool trans_unzip(DisasContext *ctx, arg_unzip *a) +{ + REQUIRE_32BIT(ctx); + REQUIRE_ZBKB(ctx); + return gen_unary(ctx, a, EXT_NONE, gen_helper_unzip); +} + +static bool trans_zip(DisasContext *ctx, arg_zip *a) +{ + REQUIRE_32BIT(ctx); + REQUIRE_ZBKB(ctx); + return gen_unary(ctx, a, EXT_NONE, gen_helper_zip); +} diff --git a/target/riscv/translate.c b/target/riscv/translate.c index 6495726..121ead7 100644 --- a/target/riscv/translate.c +++ b/target/riscv/translate.c @@ -688,6 +688,13 @@ EX_SH(12) } \ } while (0) +#define REQUIRE_EITHER_EXT(ctx, A, B) do { \ + if (!ctx->cfg_ptr->ext_##A && \ + !ctx->cfg_ptr->ext_##B) { \ + return false; \ + } \ +} while (0) + static int ex_rvc_register(DisasContext *ctx, int reg) { return 8 + reg; -- cgit v1.1