On Fri, Apr 2, 2021 at 8:49 PM Alistair Francis
<alistair.fran...@wdc.com> wrote:
>
> The RISC-V spec says:
> if PMP entry i is locked and pmpicfg.A is set to TOR, writes to
> pmpaddri-1 are ignored.
>
> The current QEMU code ignores accesses to pmpaddri-1 and pmpcfgi-1 which
> is incorrect.
>
> Update the pmp_is_locked() function to not check the suppording fields
typo: supported?
> and instaed enforce the lock functionality in the pmpaddr write operation.
typo: instead
>
> Signed-off-by: Alistair Francis <alistair.fran...@wdc.com>
> ---
> target/riscv/pmp.c | 25 +++++++++++++++----------
> 1 file changed, 15 insertions(+), 10 deletions(-)
>
> diff --git a/target/riscv/pmp.c b/target/riscv/pmp.c
> index cff020122a..6141d0f8f9 100644
> --- a/target/riscv/pmp.c
> +++ b/target/riscv/pmp.c
> @@ -59,16 +59,6 @@ static inline int pmp_is_locked(CPURISCVState *env,
> uint32_t pmp_index)
> return 0;
> }
>
> - /* In TOR mode, need to check the lock bit of the next pmp
> - * (if there is a next)
> - */
> - const uint8_t a_field =
> - pmp_get_a_field(env->pmp_state.pmp[pmp_index + 1].cfg_reg);
> - if ((env->pmp_state.pmp[pmp_index + 1u].cfg_reg & PMP_LOCK) &&
> - (PMP_AMATCH_TOR == a_field)) {
> - return 1;
> - }
> -
> return 0;
> }
>
> @@ -380,7 +370,22 @@ void pmpaddr_csr_write(CPURISCVState *env, uint32_t
> addr_index,
> target_ulong val)
> {
> trace_pmpaddr_csr_write(env->mhartid, addr_index, val);
> +
> if (addr_index < MAX_RISCV_PMPS) {
> + /* In TOR mode, need to check the lock bit of the next pmp
nits: should use correct multi-line comment block format
> + * (if there is a next).
> + */
> + if (addr_index + 1 < MAX_RISCV_PMPS) {
> + uint8_t pmp_cfg = env->pmp_state.pmp[addr_index + 1].cfg_reg;
> +
> + if (pmp_cfg & PMP_LOCK &&
> + PMP_AMATCH_TOR == pmp_get_a_field(pmp_cfg)) {
> + qemu_log_mask(LOG_GUEST_ERROR,
> + "ignoring pmpaddr write - pmpcfg + 1
> locked\n");
> + return;
> + }
> + }
> +
> if (!pmp_is_locked(env, addr_index)) {
> env->pmp_state.pmp[addr_index].addr_reg = val;
> pmp_update_rule(env, addr_index);
> --
Otherwise,
Reviewed-by: Bin Meng <bmeng...@gmail.com>
