Re: [PATCH v7 1/5] target/riscv: Ignore reserved bits in PTE for RV64

[Alistair Francis <alistair23@gmail.com>]

On Fri, Jan 28, 2022 at 7:11 PM Weiwei Li <liweiwei@iscas.ac.cn> wrote:
>
> From: Guo Ren <ren_guo@c-sky.com>
>
> Highest bits of PTE has been used for svpbmt, ref: [1], [2], so we
> need to ignore them. They cannot be a part of ppn.
>
> 1: The RISC-V Instruction Set Manual, Volume II: Privileged Architecture
>    4.4 Sv39: Page-Based 39-bit Virtual-Memory System
>    4.5 Sv48: Page-Based 48-bit Virtual-Memory System
>
> 2: https://github.com/riscv/virtual-memory/blob/main/specs/663-Svpbmt-diff.pdf
>
> Signed-off-by: Guo Ren <ren_guo@c-sky.com>
> Reviewed-by: Liu Zhiwei <zhiwei_liu@c-sky.com>
> Cc: Bin Meng <bmeng.cn@gmail.com>
> Cc: Alistair Francis <alistair.francis@wdc.com>
> ---
>  target/riscv/cpu.h        | 15 +++++++++++++++
>  target/riscv/cpu_bits.h   |  3 +++
>  target/riscv/cpu_helper.c | 14 +++++++++++++-
>  3 files changed, 31 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index 55635d68d5..336fe8e3d5 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -341,6 +341,8 @@ struct RISCVCPU {
>          bool ext_counters;
>          bool ext_ifencei;
>          bool ext_icsr;
> +        bool ext_svnapot;
> +        bool ext_svpbmt;
>          bool ext_zfh;
>          bool ext_zfhmin;
>          bool ext_zve32f;

Hello, thanks for the patches.

This looks good, but you might need to rebase it as there are patches
on list that move this into a different struct.

> @@ -495,6 +497,19 @@ static inline int riscv_cpu_xlen(CPURISCVState *env)
>      return 16 << env->xl;
>  }
>
> +#ifdef TARGET_RISCV32
> +#define riscv_cpu_sxl(env)  ((void)(env), MXL_RV32)
> +#else
> +static inline RISCVMXL riscv_cpu_sxl(CPURISCVState *env)
> +{
> +#ifdef CONFIG_USER_ONLY
> +    return env->misa_mxl;
> +#else
> +    return get_field(env->mstatus, MSTATUS64_SXL);
> +#endif
> +}
> +#endif
> +
>  /*
>   * Encode LMUL to lmul as follows:
>   *     LMUL    vlmul    lmul
> diff --git a/target/riscv/cpu_bits.h b/target/riscv/cpu_bits.h
> index 7c87433645..6ea3944423 100644
> --- a/target/riscv/cpu_bits.h
> +++ b/target/riscv/cpu_bits.h
> @@ -493,6 +493,9 @@ typedef enum {
>  /* Page table PPN shift amount */
>  #define PTE_PPN_SHIFT       10
>
> +/* Page table PPN mask */
> +#define PTE_PPN_MASK        0x3FFFFFFFFFFC00ULL
> +
>  /* Leaf page shift amount */
>  #define PGSHIFT             12
>
> diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
> index 327a2c4f1d..5a1c0e239e 100644
> --- a/target/riscv/cpu_helper.c
> +++ b/target/riscv/cpu_helper.c
> @@ -622,7 +622,19 @@ restart:
>              return TRANSLATE_FAIL;
>          }
>
> -        hwaddr ppn = pte >> PTE_PPN_SHIFT;
> +        hwaddr ppn;
> +        RISCVCPU *cpu = env_archcpu(env);

I know there is existing code in this function that does this, but
please don't initiate variables mid function. Can you move this to the
top of the function?

Otherwise:

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Alistair

> +
> +        if (riscv_cpu_sxl(env) == MXL_RV32) {
> +            ppn = pte >> PTE_PPN_SHIFT;
> +        } else if (cpu->cfg.ext_svpbmt || cpu->cfg.ext_svnapot) {
> +            ppn = (pte & (target_ulong)PTE_PPN_MASK) >> PTE_PPN_SHIFT;
> +        } else {
> +            ppn = pte >> PTE_PPN_SHIFT;
> +            if ((pte & ~(target_ulong)PTE_PPN_MASK) >> PTE_PPN_SHIFT) {
> +                return TRANSLATE_FAIL;
> +            }
> +        }
>
>          if (!(pte & PTE_V)) {
>              /* Invalid PTE */
> --
> 2.17.1
>
>