Re: [PATCH qemu] target/riscv: rvv: Fix early exit condition for whole register load/store

[Alistair Francis]

On Fri, May 6, 2022 at 7:17 AM ~eopxd <eopxd@git.sr.ht> wrote:
>
> From: eopXD <eop.chen@sifive.com>
>
> Vector whole register load instructions have EEW encoded in the opcode,
> so we shouldn't take SEW here. Vector whole register store instructions
> are always EEW=8.
>
> Signed-off-by: eop Chen <eop.chen@sifive.com>
> Reviewed-by: Frank Chang <frank.chang@sifive.com>

Thanks!

Applied to riscv-to-apply.next

Alistair

> ---
>  target/riscv/insn_trans/trans_rvv.c.inc | 58 +++++++++++++------------
>  1 file changed, 31 insertions(+), 27 deletions(-)
>
> diff --git a/target/riscv/insn_trans/trans_rvv.c.inc 
> b/target/riscv/insn_trans/trans_rvv.c.inc
> index 90327509f7..391c61fe93 100644
> --- a/target/riscv/insn_trans/trans_rvv.c.inc
> +++ b/target/riscv/insn_trans/trans_rvv.c.inc
> @@ -1118,10 +1118,10 @@ GEN_VEXT_TRANS(vle64ff_v, MO_64, r2nfvm, ldff_op, 
> ld_us_check)
>  typedef void gen_helper_ldst_whole(TCGv_ptr, TCGv, TCGv_env, TCGv_i32);
>
>  static bool ldst_whole_trans(uint32_t vd, uint32_t rs1, uint32_t nf,
> -                             gen_helper_ldst_whole *fn, DisasContext *s,
> -                             bool is_store)
> +                             uint32_t width, gen_helper_ldst_whole *fn,
> +                             DisasContext *s, bool is_store)
>  {
> -    uint32_t evl = (s->cfg_ptr->vlen / 8) * nf / (1 << s->sew);
> +    uint32_t evl = (s->cfg_ptr->vlen / 8) * nf / width;
>      TCGLabel *over = gen_new_label();
>      tcg_gen_brcondi_tl(TCG_COND_GEU, cpu_vstart, evl, over);
>
> @@ -1153,38 +1153,42 @@ static bool ldst_whole_trans(uint32_t vd, uint32_t 
> rs1, uint32_t nf,
>   * load and store whole register instructions ignore vtype and vl setting.
>   * Thus, we don't need to check vill bit. (Section 7.9)
>   */
> -#define GEN_LDST_WHOLE_TRANS(NAME, ARG_NF, IS_STORE)                      \
> +#define GEN_LDST_WHOLE_TRANS(NAME, ARG_NF, WIDTH, IS_STORE)               \
>  static bool trans_##NAME(DisasContext *s, arg_##NAME * a)                 \
>  {                                                                         \
>      if (require_rvv(s) &&                                                 \
>          QEMU_IS_ALIGNED(a->rd, ARG_NF)) {                                 \
> -        return ldst_whole_trans(a->rd, a->rs1, ARG_NF, gen_helper_##NAME, \
> -                                s, IS_STORE);                             \
> +        return ldst_whole_trans(a->rd, a->rs1, ARG_NF, WIDTH,             \
> +                                gen_helper_##NAME, s, IS_STORE);          \
>      }                                                                     \
>      return false;                                                         \
>  }
>
> -GEN_LDST_WHOLE_TRANS(vl1re8_v,  1, false)
> -GEN_LDST_WHOLE_TRANS(vl1re16_v, 1, false)
> -GEN_LDST_WHOLE_TRANS(vl1re32_v, 1, false)
> -GEN_LDST_WHOLE_TRANS(vl1re64_v, 1, false)
> -GEN_LDST_WHOLE_TRANS(vl2re8_v,  2, false)
> -GEN_LDST_WHOLE_TRANS(vl2re16_v, 2, false)
> -GEN_LDST_WHOLE_TRANS(vl2re32_v, 2, false)
> -GEN_LDST_WHOLE_TRANS(vl2re64_v, 2, false)
> -GEN_LDST_WHOLE_TRANS(vl4re8_v,  4, false)
> -GEN_LDST_WHOLE_TRANS(vl4re16_v, 4, false)
> -GEN_LDST_WHOLE_TRANS(vl4re32_v, 4, false)
> -GEN_LDST_WHOLE_TRANS(vl4re64_v, 4, false)
> -GEN_LDST_WHOLE_TRANS(vl8re8_v,  8, false)
> -GEN_LDST_WHOLE_TRANS(vl8re16_v, 8, false)
> -GEN_LDST_WHOLE_TRANS(vl8re32_v, 8, false)
> -GEN_LDST_WHOLE_TRANS(vl8re64_v, 8, false)
> -
> -GEN_LDST_WHOLE_TRANS(vs1r_v, 1, true)
> -GEN_LDST_WHOLE_TRANS(vs2r_v, 2, true)
> -GEN_LDST_WHOLE_TRANS(vs4r_v, 4, true)
> -GEN_LDST_WHOLE_TRANS(vs8r_v, 8, true)
> +GEN_LDST_WHOLE_TRANS(vl1re8_v,  1, 1, false)
> +GEN_LDST_WHOLE_TRANS(vl1re16_v, 1, 2, false)
> +GEN_LDST_WHOLE_TRANS(vl1re32_v, 1, 4, false)
> +GEN_LDST_WHOLE_TRANS(vl1re64_v, 1, 8, false)
> +GEN_LDST_WHOLE_TRANS(vl2re8_v,  2, 1, false)
> +GEN_LDST_WHOLE_TRANS(vl2re16_v, 2, 2, false)
> +GEN_LDST_WHOLE_TRANS(vl2re32_v, 2, 4, false)
> +GEN_LDST_WHOLE_TRANS(vl2re64_v, 2, 8, false)
> +GEN_LDST_WHOLE_TRANS(vl4re8_v,  4, 1, false)
> +GEN_LDST_WHOLE_TRANS(vl4re16_v, 4, 2, false)
> +GEN_LDST_WHOLE_TRANS(vl4re32_v, 4, 4, false)
> +GEN_LDST_WHOLE_TRANS(vl4re64_v, 4, 8, false)
> +GEN_LDST_WHOLE_TRANS(vl8re8_v,  8, 1, false)
> +GEN_LDST_WHOLE_TRANS(vl8re16_v, 8, 2, false)
> +GEN_LDST_WHOLE_TRANS(vl8re32_v, 8, 4, false)
> +GEN_LDST_WHOLE_TRANS(vl8re64_v, 8, 8, false)
> +
> +/*
> + * The vector whole register store instructions are encoded similar to
> + * unmasked unit-stride store of elements with EEW=8.
> + */
> +GEN_LDST_WHOLE_TRANS(vs1r_v, 1, 1, true)
> +GEN_LDST_WHOLE_TRANS(vs2r_v, 2, 1, true)
> +GEN_LDST_WHOLE_TRANS(vs4r_v, 4, 1, true)
> +GEN_LDST_WHOLE_TRANS(vs8r_v, 8, 1, true)
>
>  /*
>   *** Vector Integer Arithmetic Instructions
> --
> 2.34.2
>