Re: [PATCH 2/3] KVM: s390: Add UV feature negotiation

[Claudio Imbrenda <imbrenda@linux.ibm.com>]

On Thu, 27 Jul 2023 14:20:52 +0200
Steffen Eiden <seiden@linux.ibm.com> wrote:

> Add a uv_feature list for pv-guests to the kvm cpu-model.
> The feature bits 'AP-interpretation for secure guests' and
> 'AP-interrupt for secure guests' are available.
> 
> Signed-off-by: Steffen Eiden <seiden@linux.ibm.com>
> ---
>  arch/s390/include/asm/kvm_host.h |  2 +
>  arch/s390/include/uapi/asm/kvm.h | 24 +++++++++++
>  arch/s390/kvm/kvm-s390.c         | 70 ++++++++++++++++++++++++++++++++
>  3 files changed, 96 insertions(+)
> 
> diff --git a/arch/s390/include/asm/kvm_host.h b/arch/s390/include/asm/kvm_host.h
> index 2bbc3d54959d..7ae57ac352b2 100644
> --- a/arch/s390/include/asm/kvm_host.h
> +++ b/arch/s390/include/asm/kvm_host.h
> @@ -817,6 +817,8 @@ struct kvm_s390_cpu_model {
>  	__u64 *fac_list;
>  	u64 cpuid;
>  	unsigned short ibc;
> +	/* subset of available uv features for pv-guests enabled by user space */
> +	struct kvm_s390_vm_cpu_uv_feat uv_feat_guest;
>  };
>  
>  typedef int (*crypto_hook)(struct kvm_vcpu *vcpu);
> diff --git a/arch/s390/include/uapi/asm/kvm.h b/arch/s390/include/uapi/asm/kvm.h
> index a73cf01a1606..7ea7e4fbd37e 100644
> --- a/arch/s390/include/uapi/asm/kvm.h
> +++ b/arch/s390/include/uapi/asm/kvm.h
> @@ -159,6 +159,30 @@ struct kvm_s390_vm_cpu_subfunc {
>  	__u8 reserved[1728];
>  };
>  
> +#define KVM_S390_VM_CPU_PROCESSOR_UV_FEAT_GUEST	6
> +#define KVM_S390_VM_CPU_MACHINE_UV_FEAT_GUEST	7
> +
> +#define KVM_S390_VM_CPU_UV_FEAT_NR_BITS	64
> +struct kvm_s390_vm_cpu_uv_feat {
> +	union {
> +		struct {
> +			__u64 res0 : 4;
> +			__u64 ap : 1;		/* bit 4 */
> +			__u64 ap_intr : 1;	/* bit 5 */

I would put padding here for completeness, and maybe a static assert
for the size of struct kvm_s390_vm_cpu_uv_feat

> +		};
> +		__u64 feat;
> +	};
> +};
> +
> +#define KVM_S390_VM_CPU_UV_FEAT_GUEST_MASK \
> +(((struct kvm_s390_vm_cpu_uv_feat) {	\
> +	.ap = 1,		\
> +	.ap_intr = 1,		\
> +}).feat)

please indent the body of the macro, and add tabs on the right so that
the \ are aligned

> +
> +#define KVM_S390_VM_CPU_UV_FEAT_GUEST_DEFAULT \
> +((struct kvm_s390_vm_cpu_uv_feat) { })

please indent the body of the macro

> +
>  /* kvm attributes for crypto */
>  #define KVM_S390_VM_CRYPTO_ENABLE_AES_KW	0
>  #define KVM_S390_VM_CRYPTO_ENABLE_DEA_KW	1
> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> index 339e3190f6dc..50c8b85b89ac 100644
> --- a/arch/s390/kvm/kvm-s390.c
> +++ b/arch/s390/kvm/kvm-s390.c
> @@ -1531,6 +1531,32 @@ static int kvm_s390_set_processor_subfunc(struct kvm *kvm,
>  	return 0;
>  }
>  
> +static int kvm_s390_set_uv_feat(struct kvm *kvm, struct kvm_device_attr *attr)
> +{
> +	struct kvm_s390_vm_cpu_uv_feat data = {};
> +	unsigned long filter =
> +		uv_info.uv_feature_indications &

I think this can go in the previous line?

> +		(unsigned long)KVM_S390_VM_CPU_UV_FEAT_GUEST_MASK;

the cast is not necessary

without the cast the whole thing can fit in one line

> +
> +	if (copy_from_user(&data, (void __user *)attr->addr, sizeof(data)))
> +		return -EFAULT;
> +	if (!bitmap_subset((unsigned long *)&data, &filter,
> +			   KVM_S390_VM_CPU_UV_FEAT_NR_BITS))

this also fits in one line :)

> +		return -EINVAL;
> +
> +	mutex_lock(&kvm->lock);
> +	if (kvm->created_vcpus) {
> +		mutex_unlock(&kvm->lock);
> +		return -EBUSY;
> +	}
> +	kvm->arch.model.uv_feat_guest = data;
> +	mutex_unlock(&kvm->lock);
> +
> +	VM_EVENT(kvm, 3, "SET: guest uv feat: 0x%16.16llx", data.feat);
> +
> +	return 0;
> +}
> +
>  static int kvm_s390_set_cpu_model(struct kvm *kvm, struct kvm_device_attr *attr)
>  {
>  	int ret = -ENXIO;
> @@ -1545,6 +1571,9 @@ static int kvm_s390_set_cpu_model(struct kvm *kvm, struct kvm_device_attr *attr)
>  	case KVM_S390_VM_CPU_PROCESSOR_SUBFUNC:
>  		ret = kvm_s390_set_processor_subfunc(kvm, attr);
>  		break;
> +	case KVM_S390_VM_CPU_PROCESSOR_UV_FEAT_GUEST:
> +		ret = kvm_s390_set_uv_feat(kvm, attr);
> +		break;
>  	}
>  	return ret;
>  }
> @@ -1777,6 +1806,37 @@ static int kvm_s390_get_machine_subfunc(struct kvm *kvm,
>  	return 0;
>  }
>  
> +static int kvm_s390_get_processor_uv_feat(struct kvm *kvm,
> +					  struct kvm_device_attr *attr)
> +{
> +	struct kvm_s390_vm_cpu_uv_feat *src = &kvm->arch.model.uv_feat_guest;
> +
> +	if (copy_to_user((void __user *)attr->addr, src, sizeof(*src)))
> +		return -EFAULT;
> +	VM_EVENT(kvm, 3, "GET: guest  uv feat: 0x%16.16llx",
> +		 kvm->arch.model.uv_feat_guest.feat);

this can actually fit in one line

> +
> +	return 0;
> +}
> +
> +static int kvm_s390_get_machine_uv_feat(struct kvm *kvm,
> +					struct kvm_device_attr *attr)
> +{
> +	struct kvm_s390_vm_cpu_uv_feat data = {
> +		.feat = uv_info.uv_feature_indications &
> +			KVM_S390_VM_CPU_UV_FEAT_GUEST_MASK

this can fit in one line

> +	};
> +
> +	BUILD_BUG_ON(sizeof(struct kvm_s390_vm_cpu_uv_feat) !=
> +		     sizeof(uv_info.uv_feature_indications));
> +	if (copy_to_user((void __user *)attr->addr, &data,
> +			 sizeof(struct kvm_s390_vm_cpu_uv_feat)))

why not sizeof(data) ?

(also, it can fit in one line)

> +		return -EFAULT;
> +	VM_EVENT(kvm, 3, "GET: guest  uv feat: 0x%16.16llx", data.feat);
> +
> +	return 0;
> +}
> +
>  static int kvm_s390_get_cpu_model(struct kvm *kvm, struct kvm_device_attr *attr)
>  {
>  	int ret = -ENXIO;
> @@ -1800,6 +1860,12 @@ static int kvm_s390_get_cpu_model(struct kvm *kvm, struct kvm_device_attr *attr)
>  	case KVM_S390_VM_CPU_MACHINE_SUBFUNC:
>  		ret = kvm_s390_get_machine_subfunc(kvm, attr);
>  		break;
> +	case KVM_S390_VM_CPU_PROCESSOR_UV_FEAT_GUEST:
> +		ret = kvm_s390_get_processor_uv_feat(kvm, attr);
> +		break;
> +	case KVM_S390_VM_CPU_MACHINE_UV_FEAT_GUEST:
> +		ret = kvm_s390_get_machine_uv_feat(kvm, attr);
> +		break;
>  	}
>  	return ret;
>  }
> @@ -1952,6 +2018,8 @@ static int kvm_s390_vm_has_attr(struct kvm *kvm, struct kvm_device_attr *attr)
>  		case KVM_S390_VM_CPU_MACHINE_FEAT:
>  		case KVM_S390_VM_CPU_MACHINE_SUBFUNC:
>  		case KVM_S390_VM_CPU_PROCESSOR_SUBFUNC:
> +		case KVM_S390_VM_CPU_MACHINE_UV_FEAT_GUEST:
> +		case KVM_S390_VM_CPU_PROCESSOR_UV_FEAT_GUEST:
>  			ret = 0;
>  			break;
>  		default:
> @@ -3292,6 +3360,8 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
>  	kvm->arch.model.cpuid = kvm_s390_get_initial_cpuid();
>  	kvm->arch.model.ibc = sclp.ibc & 0x0fff;
>  
> +	kvm->arch.model.uv_feat_guest = KVM_S390_VM_CPU_UV_FEAT_GUEST_DEFAULT;
> +
>  	kvm_s390_crypto_init(kvm);
>  
>  	if (IS_ENABLED(CONFIG_VFIO_PCI_ZDEV_KVM)) {