Re: [PATCH v2 3/8] RISC-V: KVM: Allow some SBI extensions to be disabled by default

[Andrew Jones <ajones@ventanamicro.com>]

On Thu, Oct 12, 2023 at 10:45:04AM +0530, Anup Patel wrote:
> Currently, all SBI extensions are enabled by default which is
> problematic for SBI extensions (such as DBCN) which are forwarded
> to the KVM user-space because we might have an older KVM user-space
> which is not aware/ready to handle newer SBI extensions. Ideally,
> the SBI extensions forwarded to the KVM user-space must be
> disabled by default.
> 
> To address above, we allow certain SBI extensions to be disabled
> by default so that KVM user-space must explicitly enable such
> SBI extensions to receive forwarded calls from Guest VCPU.
> 
> Signed-off-by: Anup Patel <apatel@ventanamicro.com>
> ---
>  arch/riscv/include/asm/kvm_vcpu_sbi.h |  4 +++
>  arch/riscv/kvm/vcpu.c                 |  6 ++++
>  arch/riscv/kvm/vcpu_sbi.c             | 45 ++++++++++++++++-----------
>  3 files changed, 36 insertions(+), 19 deletions(-)
> 
> diff --git a/arch/riscv/include/asm/kvm_vcpu_sbi.h b/arch/riscv/include/asm/kvm_vcpu_sbi.h
> index 8d6d4dce8a5e..c02bda5559d7 100644
> --- a/arch/riscv/include/asm/kvm_vcpu_sbi.h
> +++ b/arch/riscv/include/asm/kvm_vcpu_sbi.h
> @@ -35,6 +35,9 @@ struct kvm_vcpu_sbi_return {
>  struct kvm_vcpu_sbi_extension {
>  	unsigned long extid_start;
>  	unsigned long extid_end;
> +
> +	bool default_unavail;
> +
>  	/**
>  	 * SBI extension handler. It can be defined for a given extension or group of
>  	 * extension. But it should always return linux error codes rather than SBI
> @@ -59,6 +62,7 @@ int kvm_riscv_vcpu_get_reg_sbi_ext(struct kvm_vcpu *vcpu,
>  const struct kvm_vcpu_sbi_extension *kvm_vcpu_sbi_find_ext(
>  				struct kvm_vcpu *vcpu, unsigned long extid);
>  int kvm_riscv_vcpu_sbi_ecall(struct kvm_vcpu *vcpu, struct kvm_run *run);
> +void kvm_riscv_vcpu_sbi_init(struct kvm_vcpu *vcpu);
>  
>  #ifdef CONFIG_RISCV_SBI_V01
>  extern const struct kvm_vcpu_sbi_extension vcpu_sbi_ext_v01;
> diff --git a/arch/riscv/kvm/vcpu.c b/arch/riscv/kvm/vcpu.c
> index c061a1c5fe98..e087c809073c 100644
> --- a/arch/riscv/kvm/vcpu.c
> +++ b/arch/riscv/kvm/vcpu.c
> @@ -141,6 +141,12 @@ int kvm_arch_vcpu_create(struct kvm_vcpu *vcpu)
>  	if (rc)
>  		return rc;
>  
> +	/*
> +	 * Setup SBI extensions
> +	 * NOTE: This must be the last thing to be initialized.
> +	 */
> +	kvm_riscv_vcpu_sbi_init(vcpu);

With this, we no longer defer probing to the first access (whether that's
by the guest or KVM userspace). With our current small set of SBI
extensions where only a single one has a probe function, then this
simpler approach is good enough. We can always go back to the lazy
approach later if needed.

> +
>  	/* Reset VCPU */
>  	kvm_riscv_reset_vcpu(vcpu);
>  
> diff --git a/arch/riscv/kvm/vcpu_sbi.c b/arch/riscv/kvm/vcpu_sbi.c
> index 9cd97091c723..1b1cee86efda 100644
> --- a/arch/riscv/kvm/vcpu_sbi.c
> +++ b/arch/riscv/kvm/vcpu_sbi.c
> @@ -155,14 +155,8 @@ static int riscv_vcpu_set_sbi_ext_single(struct kvm_vcpu *vcpu,
>  	if (!sext)
>  		return -ENOENT;
>  
> -	/*
> -	 * We can't set the extension status to available here, since it may
> -	 * have a probe() function which needs to confirm availability first,
> -	 * but it may be too early to call that here. We can set the status to
> -	 * unavailable, though.
> -	 */
> -	if (!reg_val)
> -		scontext->ext_status[sext->ext_idx] =
> +	scontext->ext_status[sext->ext_idx] = (reg_val) ?
> +			KVM_RISCV_SBI_EXT_AVAILABLE :
>  			KVM_RISCV_SBI_EXT_UNAVAILABLE;

We're missing the change to riscv_vcpu_get_sbi_ext_single() which should
also drop the comment block explaining the limits to status knowledge
without initial probing (which we now do) and then just check for
available, i.e.

diff --git a/arch/riscv/kvm/vcpu_sbi.c b/arch/riscv/kvm/vcpu_sbi.c
index bb76c3cf633f..92c42d9aba1c 100644
--- a/arch/riscv/kvm/vcpu_sbi.c
+++ b/arch/riscv/kvm/vcpu_sbi.c
@@ -186,15 +186,8 @@ static int riscv_vcpu_get_sbi_ext_single(struct kvm_vcpu *vcpu,
 	if (!sext)
 		return -ENOENT;
 
-	/*
-	 * If the extension status is still uninitialized, then we should probe
-	 * to determine if it's available, but it may be too early to do that
-	 * here. The best we can do is report that the extension has not been
-	 * disabled, i.e. we return 1 when the extension is available and also
-	 * when it only may be available.
-	 */
-	*reg_val = scontext->ext_status[sext->ext_idx] !=
-				KVM_RISCV_SBI_EXT_UNAVAILABLE;
+	*reg_val = scontext->ext_status[sext->ext_idx] ==
+				KVM_RISCV_SBI_EXT_AVAILABLE;
 
 	return 0;
 }

>  
>  	return 0;
> @@ -337,18 +331,8 @@ const struct kvm_vcpu_sbi_extension *kvm_vcpu_sbi_find_ext(
>  			    scontext->ext_status[entry->ext_idx] ==
>  						KVM_RISCV_SBI_EXT_AVAILABLE)
>  				return ext;
> -			if (scontext->ext_status[entry->ext_idx] ==
> -						KVM_RISCV_SBI_EXT_UNAVAILABLE)
> -				return NULL;
> -			if (ext->probe && !ext->probe(vcpu)) {
> -				scontext->ext_status[entry->ext_idx] =
> -					KVM_RISCV_SBI_EXT_UNAVAILABLE;
> -				return NULL;
> -			}
>  
> -			scontext->ext_status[entry->ext_idx] =
> -				KVM_RISCV_SBI_EXT_AVAILABLE;
> -			return ext;
> +			return NULL;
>  		}
>  	}
>  
> @@ -419,3 +403,26 @@ int kvm_riscv_vcpu_sbi_ecall(struct kvm_vcpu *vcpu, struct kvm_run *run)
>  
>  	return ret;
>  }
> +
> +void kvm_riscv_vcpu_sbi_init(struct kvm_vcpu *vcpu)
> +{
> +	struct kvm_vcpu_sbi_context *scontext = &vcpu->arch.sbi_context;
> +	const struct kvm_riscv_sbi_extension_entry *entry;
> +	const struct kvm_vcpu_sbi_extension *ext;
> +	int i;
> +
> +	for (i = 0; i < ARRAY_SIZE(sbi_ext); i++) {
> +		entry = &sbi_ext[i];
> +		ext = entry->ext_ptr;
> +
> +		if (ext->probe && !ext->probe(vcpu)) {
> +			scontext->ext_status[entry->ext_idx] =
> +				KVM_RISCV_SBI_EXT_UNAVAILABLE;
> +			continue;
> +		}
> +
> +		scontext->ext_status[entry->ext_idx] = ext->default_unavail ?
> +					KVM_RISCV_SBI_EXT_UNAVAILABLE :
> +					KVM_RISCV_SBI_EXT_AVAILABLE;
> +	}
> +}
> -- 
> 2.34.1
>

Thanks,
drew