On 6/11/21 4:42 PM, Frank Chang wrote:
LIU Zhiwei <zhiwei_...@c-sky.com <mailto:zhiwei_...@c-sky.com>> 於 2021年6月11日 週五 下午4:30寫道:


    On 6/11/21 4:15 PM, Frank Chang wrote:
    LIU Zhiwei <zhiwei_...@c-sky.com <mailto:zhiwei_...@c-sky.com>> 於
    2021年4月9日 週五 下午3:52寫道:

        The CSR can be used by software to service the next
        horizontal interrupt
        when it has greater level than the saved interrupt context
        (held in xcause`.pil`) and greater level than the interrupt
        threshold of
        the corresponding privilege mode,

        Signed-off-by: LIU Zhiwei <zhiwei_...@c-sky.com
        <mailto:zhiwei_...@c-sky.com>>
        ---
         target/riscv/cpu_bits.h |  16 ++++++
         target/riscv/csr.c      | 114
        ++++++++++++++++++++++++++++++++++++++++
         2 files changed, 130 insertions(+)

        diff --git a/target/riscv/cpu_bits.h b/target/riscv/cpu_bits.h
        index 7922097776..494e41edc9 100644
        --- a/target/riscv/cpu_bits.h
        +++ b/target/riscv/cpu_bits.h
        @@ -166,6 +166,7 @@
         #define CSR_MCAUSE          0x342
         #define CSR_MTVAL           0x343
         #define CSR_MIP             0x344
        +#define CSR_MNXTI           0x345 /* clic-spec-draft */
         #define CSR_MINTSTATUS      0x346 /* clic-spec-draft */
         #define CSR_MINTTHRESH      0x347 /* clic-spec-draft */

        @@ -187,6 +188,7 @@
         #define CSR_SCAUSE          0x142
         #define CSR_STVAL           0x143
         #define CSR_SIP             0x144
        +#define CSR_SNXTI           0x145 /* clic-spec-draft */
         #define CSR_SINTSTATUS      0x146 /* clic-spec-draft */
         #define CSR_SINTTHRESH      0x147 /* clic-spec-draft */

        @@ -596,10 +598,24 @@
         #define MINTSTATUS_SIL  0x0000ff00 /* sil[7:0] */
         #define MINTSTATUS_UIL  0x000000ff /* uil[7:0] */

        +/* mcause */
        +#define MCAUSE_MINHV  0x40000000 /* minhv */
        +#define MCAUSE_MPP  0x30000000 /* mpp[1:0] */
        +#define MCAUSE_MPIE 0x08000000 /* mpie */
        +#define MCAUSE_MPIL 0x00ff0000 /* mpil[7:0] */
        +#define MCAUSE_EXCCODE  0x00000fff /* exccode[11:0] */
        +
         /* sintstatus */
         #define SINTSTATUS_SIL  0x0000ff00 /* sil[7:0] */
         #define SINTSTATUS_UIL  0x000000ff /* uil[7:0] */

        +/* scause */
        +#define SCAUSE_SINHV  0x40000000 /* sinhv */
        +#define SCAUSE_SPP  0x10000000 /* spp */
        +#define SCAUSE_SPIE 0x08000000 /* spie */
        +#define SCAUSE_SPIL 0x00ff0000 /* spil[7:0] */
        +#define SCAUSE_EXCCODE  0x00000fff /* exccode[11:0] */
        +
         /* MIE masks */
         #define MIE_SEIE                           (1 << IRQ_S_EXT)
         #define MIE_UEIE                           (1 << IRQ_U_EXT)
        diff --git a/target/riscv/csr.c b/target/riscv/csr.c
        index e12222b77f..72cba080bf 100644
        --- a/target/riscv/csr.c
        +++ b/target/riscv/csr.c
        @@ -774,6 +774,80 @@ static int rmw_mip(CPURISCVState *env,
        int csrno, target_ulong *ret_value,
             return 0;
         }

        +static bool get_xnxti_status(CPURISCVState *env)
        +{
        +    CPUState *cs = env_cpu(env);
        +    int clic_irq, clic_priv, clic_il, pil;
        +
        +    if (!env->exccode) { /* No interrupt */
        +        return false;
        +    }
        +    /* The system is not in a CLIC mode */
        +    if (!riscv_clic_is_clic_mode(env)) {
        +        return false;
        +    } else {
        + riscv_clic_decode_exccode(env->exccode, &clic_priv, &clic_il,
        +                                  &clic_irq);
        +
        +        if (env->priv == PRV_M) {
        +            pil = MAX(get_field(env->mcause, MCAUSE_MPIL),
        env->mintthresh);
        +        } else if (env->priv == PRV_S) {
        +            pil = MAX(get_field(env->scause, SCAUSE_SPIL),
        env->sintthresh);
        +        } else {
        +            qemu_log_mask(LOG_GUEST_ERROR,
        +                          "CSR: rmw xnxti with unsupported
        mode\n");
        +            exit(1);
        +        }
        +
        +        if ((clic_priv != env->priv) || /* No horizontal
        interrupt */
        +            (clic_il <= pil) || /* No higher level interrupt */
        + (riscv_clic_shv_interrupt(env->clic, clic_priv, cs->cpu_index,
        +                                      clic_irq))) { /* CLIC
        vector mode */
        +            return false;
        +        } else {
        +            return true;
        +        }
        +    }
        +}
        +
        +static int rmw_mnxti(CPURISCVState *env, int csrno,
        target_ulong *ret_value,
        +                     target_ulong new_value, target_ulong
        write_mask)
        +{
        +    int clic_priv, clic_il, clic_irq;
        +    bool ready;
        +    CPUState *cs = env_cpu(env);
        +    if (write_mask) {
        +        env->mstatus |= new_value & (write_mask & 0b11111);
        +    }
        +
        +    qemu_mutex_lock_iothread();


    Hi Zhiwei,

    May I ask what's the purpose to request the BQL here with
    /qemu_mutex_lock_iothread()/?
    Is there any critical data we need to protect in /rmw_mnxti()/?
    As far I see, /rmw_mnxti()/ won't call /cpu_interrupt()/ which
    need BQL to be held before calling.
    Am I missing anything?
    In my opinion, if you read or write any  MMIO register, you need
    to hold the BQL. As you can quickly see,
    it calls riscv_clic_clean_pending. That's why it should hold the BQL.

    Zhiwei


Oh, I see.
The MMIO register reads and writes should also be protected by BQL.
Thanks for the explanation.

I am glad to know that you are reviewing this patch set. As Sifive implements the initial v0.7 CLIC, I think you may need this patch set for your SOC. If you like to, I am happy to see that you connect this patch set to your SOC, and resend it again. I can also provide the qtest of this patch set if you need.

As you may see, the v6.1 soft freeze will come in July. I am afraid I can't upstream a new SOC in so short time.

Zhiwei


Regards,
Frank Chang


    Regard,
    Frank Chang

        +    ready = get_xnxti_status(env);
        +    if (ready) {
        + riscv_clic_decode_exccode(env->exccode, &clic_priv, &clic_il,
        +                                  &clic_irq);
        +        if (write_mask) {
        +            bool edge = riscv_clic_edge_triggered(env->clic,
        clic_priv,
        + cs->cpu_index, clic_irq);
        +            if (edge) {
        + riscv_clic_clean_pending(env->clic, clic_priv,
        +  cs->cpu_index, clic_irq);
        +            }
        +            env->mintstatus = set_field(env->mintstatus,
        + MINTSTATUS_MIL, clic_il);
        +            env->mcause = set_field(env->mcause,
        MCAUSE_EXCCODE, clic_irq);
        +        }
        +        if (ret_value) {
        +            *ret_value = (env->mtvt & ~0x3f) +
        sizeof(target_ulong) * clic_irq;
        +        }
        +    } else {
        +        if (ret_value) {
        +            *ret_value = 0;
        +        }
        +    }
        +    qemu_mutex_unlock_iothread();
        +    return 0;
        +}
        +
         static int read_mintstatus(CPURISCVState *env, int csrno,
        target_ulong *val)
         {
             *val = env->mintstatus;
        @@ -982,6 +1056,44 @@ static int rmw_sip(CPURISCVState *env,
        int csrno, target_ulong *ret_value,
             return ret;
         }

        +static int rmw_snxti(CPURISCVState *env, int csrno,
        target_ulong *ret_value,
        +                     target_ulong new_value, target_ulong
        write_mask)
        +{
        +    int clic_priv, clic_il, clic_irq;
        +    bool ready;
        +    CPUState *cs = env_cpu(env);
        +    if (write_mask) {
        +        env->mstatus |= new_value & (write_mask & 0b11111);
        +    }
        +
        +    qemu_mutex_lock_iothread();
        +    ready = get_xnxti_status(env);
        +    if (ready) {
        + riscv_clic_decode_exccode(env->exccode, &clic_priv, &clic_il,
        +                                  &clic_irq);
        +        if (write_mask) {
        +            bool edge = riscv_clic_edge_triggered(env->clic,
        clic_priv,
        + cs->cpu_index, clic_irq);
        +            if (edge) {
        + riscv_clic_clean_pending(env->clic, clic_priv,
        +  cs->cpu_index, clic_irq);
        +            }
        +            env->mintstatus = set_field(env->mintstatus,
        + MINTSTATUS_SIL, clic_il);
        +            env->scause = set_field(env->scause,
        SCAUSE_EXCCODE, clic_irq);
        +        }
        +        if (ret_value) {
        +            *ret_value = (env->stvt & ~0x3f) +
        sizeof(target_ulong) * clic_irq;
        +        }
        +    } else {
        +        if (ret_value) {
        +            *ret_value = 0;
        +        }
        +    }
        +    qemu_mutex_unlock_iothread();
        +    return 0;
        +}
        +
         static int read_sintstatus(CPURISCVState *env, int csrno,
        target_ulong *val)
         {
             target_ulong mask = SINTSTATUS_SIL | SINTSTATUS_UIL;
        @@ -1755,6 +1867,7 @@ riscv_csr_operations
        csr_ops[CSR_TABLE_SIZE] = {

             /* Machine Mode Core Level Interrupt Controller */
             [CSR_MTVT] = { "mtvt", clic,  read_mtvt, write_mtvt      },
        +    [CSR_MNXTI] = { "mnxti", clic,  NULL,  NULL, rmw_mnxti   },
             [CSR_MINTSTATUS] = { "mintstatus", clic, read_mintstatus },
             [CSR_MINTTHRESH] = { "mintthresh", clic, read_mintthresh,
                                  write_mintthresh },
        @@ -1766,6 +1879,7 @@ riscv_csr_operations
        csr_ops[CSR_TABLE_SIZE] = {

             /* Supervisor Mode Core Level Interrupt Controller */
             [CSR_STVT] = { "stvt", clic,  read_stvt, write_stvt       },
        +    [CSR_SNXTI] = { "snxti", clic,  NULL,  NULL, rmw_snxti   },

         #endif /* !CONFIG_USER_ONLY */
         };
-- 2.25.1


Reply via email to