Re: [PATCH v6 1/2] Set KVM initial privilege mode and mp_state

[Andrew Jones <ajones@ventanamicro.com>]

On Wed, Sep 10, 2025 at 05:35:27PM +0800, Xie Bo wrote:
> For KVM mode, the privilege mode should not include M-mode, and the 
> initial value should be set to S-mode. Additionally, a following patch 
> adds the implementation of putting the vCPU privilege mode to KVM. 
> When the vCPU runs for the first time, QEMU will first put the privilege 
> state to KVM. If the initial value is set to M-mode, KVM will encounter
> an error.
> 
> In addition, this patch introduces the 'mp_state' field to RISC-V 
> vCPUs, following the convention used by KVM on x86. The 'mp_state' 
> reflects the multiprocessor state of a vCPU, and is used to control 
> whether the vCPU is runnable by KVM. Randomly select one CPU as the 
> boot CPU.
> 
> Signed-off-by: Xie Bo <xb@ultrarisc.com>
> ---
>  target/riscv/cpu.c | 17 ++++++++++++++++-
>  target/riscv/cpu.h |  2 ++
>  2 files changed, 18 insertions(+), 1 deletion(-)
> 
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 09ded6829a..57b8c421bd 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -38,6 +38,7 @@
>  #include "kvm/kvm_riscv.h"
>  #include "tcg/tcg-cpu.h"
>  #include "tcg/tcg.h"
> +#include "hw/boards.h"
>  
>  /* RISC-V CPU definitions */
>  static const char riscv_single_letter_exts[] = "IEMAFDQCBPVH";
> @@ -1031,18 +1032,32 @@ static void riscv_cpu_reset_hold(Object *obj, ResetType type)
>  #ifndef CONFIG_USER_ONLY
>      uint8_t iprio;
>      int i, irq, rdzero;
> +    static int boot_cpu_index = -1;
>  #endif
>      CPUState *cs = CPU(obj);
>      RISCVCPU *cpu = RISCV_CPU(cs);
>      RISCVCPUClass *mcc = RISCV_CPU_GET_CLASS(obj);
>      CPURISCVState *env = &cpu->env;
> +    MachineState *ms = MACHINE(qdev_get_machine());
>  
>      if (mcc->parent_phases.hold) {
>          mcc->parent_phases.hold(obj, type);
>      }
>  #ifndef CONFIG_USER_ONLY
>      env->misa_mxl = mcc->misa_mxl_max;
> -    env->priv = PRV_M;
> +    if (kvm_enabled()) {
> +        env->priv = PRV_S;
> +    } else {
> +        env->priv = PRV_M;
> +    }
> +    if (boot_cpu_index < 0) {
> +        boot_cpu_index = g_random_int() % ms->smp.cpus;

I don't think we're obliged to use the same boot cpu on each reset
after initially randomly selecting it. We can randomly select each
time.

Also, we can use g_random_int_range() instead of the mod.

Thanks,
drew

> +    }
> +    if (cs->cpu_index == boot_cpu_index) {
> +        env->mp_state = KVM_MP_STATE_RUNNABLE;
> +    } else {
> +        env->mp_state = KVM_MP_STATE_STOPPED;
> +    }
>      env->mstatus &= ~(MSTATUS_MIE | MSTATUS_MPRV);
>      if (env->misa_mxl > MXL_RV32) {
>          /*
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index 51e49e03de..4b1c5bf0e4 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -256,6 +256,8 @@ struct CPUArchState {
>  #endif
>  
>      target_ulong priv;
> +    /* Current multiprocessor state of this vCPU. */
> +    uint32_t mp_state;
>      /* CSRs for execution environment configuration */
>      uint64_t menvcfg;
>      target_ulong senvcfg;
> -- 
> 2.43.0
>