Message ID | 20210112093950.17530-18-frank.chang@sifive.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | support vector extension v1.0 | expand |
On Tue, Jan 12, 2021 at 2:06 AM <frank.chang@sifive.com> wrote: > > From: Frank Chang <frank.chang@sifive.com> > > Signed-off-by: Frank Chang <frank.chang@sifive.com> > Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Alistair Francis <alistair.francis@wdc.com> Alistair > --- > target/riscv/insn_trans/trans_rvv.c.inc | 53 +++++++++---------------- > target/riscv/vector_helper.c | 14 ++++++- > 2 files changed, 31 insertions(+), 36 deletions(-) > > diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc > index ccfa93cf2f8..a3732e76e09 100644 > --- a/target/riscv/insn_trans/trans_rvv.c.inc > +++ b/target/riscv/insn_trans/trans_rvv.c.inc > @@ -132,28 +132,29 @@ static bool require_noover_seg(const int8_t dst, const int8_t nf, > return !is_overlapped(dst, nf, src, 1); > } > > -static bool trans_vsetvl(DisasContext *ctx, arg_vsetvl *a) > +static bool do_vsetvl(DisasContext *ctx, int rd, int rs1, TCGv s2) > { > - TCGv s1, s2, dst; > + TCGv s1, dst; > > if (!require_rvv(ctx) || !has_ext(ctx, RVV)) { > return false; > } > > - s2 = tcg_temp_new(); > dst = tcg_temp_new(); > > - /* Using x0 as the rs1 register specifier, encodes an infinite AVL */ > - if (a->rs1 == 0) { > + if (rd == 0 && rs1 == 0) { > + s1 = tcg_temp_new(); > + tcg_gen_mov_tl(s1, cpu_vl); > + } else if (rs1 == 0) { > /* As the mask is at least one bit, RV_VLEN_MAX is >= VLMAX */ > s1 = tcg_const_tl(RV_VLEN_MAX); > } else { > s1 = tcg_temp_new(); > - gen_get_gpr(s1, a->rs1); > + gen_get_gpr(s1, rs1); > } > - gen_get_gpr(s2, a->rs2); > + > gen_helper_vsetvl(dst, cpu_env, s1, s2); > - gen_set_gpr(a->rd, dst); > + gen_set_gpr(rd, dst); > mark_vs_dirty(ctx); > tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn); > lookup_and_goto_ptr(ctx); > @@ -165,35 +166,17 @@ static bool trans_vsetvl(DisasContext *ctx, arg_vsetvl *a) > return true; > } > > -static bool trans_vsetvli(DisasContext *ctx, arg_vsetvli *a) > +static bool trans_vsetvl(DisasContext *ctx, arg_vsetvl *a) > { > - TCGv s1, s2, dst; > - > - if (!require_rvv(ctx) || !has_ext(ctx, RVV)) { > - return false; > - } > - > - s2 = tcg_const_tl(a->zimm); > - dst = tcg_temp_new(); > - > - /* Using x0 as the rs1 register specifier, encodes an infinite AVL */ > - if (a->rs1 == 0) { > - /* As the mask is at least one bit, RV_VLEN_MAX is >= VLMAX */ > - s1 = tcg_const_tl(RV_VLEN_MAX); > - } else { > - s1 = tcg_temp_new(); > - gen_get_gpr(s1, a->rs1); > - } > - gen_helper_vsetvl(dst, cpu_env, s1, s2); > - gen_set_gpr(a->rd, dst); > - mark_vs_dirty(ctx); > - gen_goto_tb(ctx, 0, ctx->pc_succ_insn); > - ctx->base.is_jmp = DISAS_NORETURN; > + TCGv s2 = tcg_temp_new(); > + gen_get_gpr(s2, a->rs2); > + return do_vsetvl(ctx, a->rd, a->rs1, s2); > +} > > - tcg_temp_free(s1); > - tcg_temp_free(s2); > - tcg_temp_free(dst); > - return true; > +static bool trans_vsetvli(DisasContext *ctx, arg_vsetvli *a) > +{ > + TCGv s2 = tcg_const_tl(a->zimm); > + return do_vsetvl(ctx, a->rd, a->rs1, s2); > } > > /* vector register offset from env */ > diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c > index 89aa7cbf73f..61917d34ffe 100644 > --- a/target/riscv/vector_helper.c > +++ b/target/riscv/vector_helper.c > @@ -31,12 +31,24 @@ target_ulong HELPER(vsetvl)(CPURISCVState *env, target_ulong s1, > { > int vlmax, vl; > RISCVCPU *cpu = env_archcpu(env); > + uint64_t lmul = FIELD_EX64(s2, VTYPE, VLMUL); > uint16_t sew = 8 << FIELD_EX64(s2, VTYPE, VSEW); > uint8_t ediv = FIELD_EX64(s2, VTYPE, VEDIV); > bool vill = FIELD_EX64(s2, VTYPE, VILL); > target_ulong reserved = FIELD_EX64(s2, VTYPE, RESERVED); > > - if ((sew > cpu->cfg.elen) || vill || (ediv != 0) || (reserved != 0)) { > + if (lmul & 4) { > + /* Fractional LMUL. */ > + if (lmul == 4 || > + cpu->cfg.elen >> (8 - lmul) < sew) { > + vill = true; > + } > + } > + > + if ((sew > cpu->cfg.elen) > + || vill > + || (ediv != 0) > + || (reserved != 0)) { > /* only set vill bit. */ > env->vtype = FIELD_DP64(0, VTYPE, VILL, 1); > env->vl = 0; > -- > 2.17.1 > >
diff --git a/target/riscv/insn_trans/trans_rvv.c.inc b/target/riscv/insn_trans/trans_rvv.c.inc index ccfa93cf2f8..a3732e76e09 100644 --- a/target/riscv/insn_trans/trans_rvv.c.inc +++ b/target/riscv/insn_trans/trans_rvv.c.inc @@ -132,28 +132,29 @@ static bool require_noover_seg(const int8_t dst, const int8_t nf, return !is_overlapped(dst, nf, src, 1); } -static bool trans_vsetvl(DisasContext *ctx, arg_vsetvl *a) +static bool do_vsetvl(DisasContext *ctx, int rd, int rs1, TCGv s2) { - TCGv s1, s2, dst; + TCGv s1, dst; if (!require_rvv(ctx) || !has_ext(ctx, RVV)) { return false; } - s2 = tcg_temp_new(); dst = tcg_temp_new(); - /* Using x0 as the rs1 register specifier, encodes an infinite AVL */ - if (a->rs1 == 0) { + if (rd == 0 && rs1 == 0) { + s1 = tcg_temp_new(); + tcg_gen_mov_tl(s1, cpu_vl); + } else if (rs1 == 0) { /* As the mask is at least one bit, RV_VLEN_MAX is >= VLMAX */ s1 = tcg_const_tl(RV_VLEN_MAX); } else { s1 = tcg_temp_new(); - gen_get_gpr(s1, a->rs1); + gen_get_gpr(s1, rs1); } - gen_get_gpr(s2, a->rs2); + gen_helper_vsetvl(dst, cpu_env, s1, s2); - gen_set_gpr(a->rd, dst); + gen_set_gpr(rd, dst); mark_vs_dirty(ctx); tcg_gen_movi_tl(cpu_pc, ctx->pc_succ_insn); lookup_and_goto_ptr(ctx); @@ -165,35 +166,17 @@ static bool trans_vsetvl(DisasContext *ctx, arg_vsetvl *a) return true; } -static bool trans_vsetvli(DisasContext *ctx, arg_vsetvli *a) +static bool trans_vsetvl(DisasContext *ctx, arg_vsetvl *a) { - TCGv s1, s2, dst; - - if (!require_rvv(ctx) || !has_ext(ctx, RVV)) { - return false; - } - - s2 = tcg_const_tl(a->zimm); - dst = tcg_temp_new(); - - /* Using x0 as the rs1 register specifier, encodes an infinite AVL */ - if (a->rs1 == 0) { - /* As the mask is at least one bit, RV_VLEN_MAX is >= VLMAX */ - s1 = tcg_const_tl(RV_VLEN_MAX); - } else { - s1 = tcg_temp_new(); - gen_get_gpr(s1, a->rs1); - } - gen_helper_vsetvl(dst, cpu_env, s1, s2); - gen_set_gpr(a->rd, dst); - mark_vs_dirty(ctx); - gen_goto_tb(ctx, 0, ctx->pc_succ_insn); - ctx->base.is_jmp = DISAS_NORETURN; + TCGv s2 = tcg_temp_new(); + gen_get_gpr(s2, a->rs2); + return do_vsetvl(ctx, a->rd, a->rs1, s2); +} - tcg_temp_free(s1); - tcg_temp_free(s2); - tcg_temp_free(dst); - return true; +static bool trans_vsetvli(DisasContext *ctx, arg_vsetvli *a) +{ + TCGv s2 = tcg_const_tl(a->zimm); + return do_vsetvl(ctx, a->rd, a->rs1, s2); } /* vector register offset from env */ diff --git a/target/riscv/vector_helper.c b/target/riscv/vector_helper.c index 89aa7cbf73f..61917d34ffe 100644 --- a/target/riscv/vector_helper.c +++ b/target/riscv/vector_helper.c @@ -31,12 +31,24 @@ target_ulong HELPER(vsetvl)(CPURISCVState *env, target_ulong s1, { int vlmax, vl; RISCVCPU *cpu = env_archcpu(env); + uint64_t lmul = FIELD_EX64(s2, VTYPE, VLMUL); uint16_t sew = 8 << FIELD_EX64(s2, VTYPE, VSEW); uint8_t ediv = FIELD_EX64(s2, VTYPE, VEDIV); bool vill = FIELD_EX64(s2, VTYPE, VILL); target_ulong reserved = FIELD_EX64(s2, VTYPE, RESERVED); - if ((sew > cpu->cfg.elen) || vill || (ediv != 0) || (reserved != 0)) { + if (lmul & 4) { + /* Fractional LMUL. */ + if (lmul == 4 || + cpu->cfg.elen >> (8 - lmul) < sew) { + vill = true; + } + } + + if ((sew > cpu->cfg.elen) + || vill + || (ediv != 0) + || (reserved != 0)) { /* only set vill bit. */ env->vtype = FIELD_DP64(0, VTYPE, VILL, 1); env->vl = 0;