Skip to contentSkip to conversation

KVM: move KVM_CAP_DEVICE_CTRL to the generic check

PATCH Superseded Wei Wang 2 participants 3 messages Dec 29, 2022 – Mar 8, 2023

Subsystems (9)
KVM PowerPC RISC-V arm64 +5 more subsystems
lore
Share & tools
lore

Wei Wang

author
#0 lore raw

PATCH v2 1/1 KVM: move KVM_CAP_DEVICE_CTRL to the generic check

KVM_CAP_DEVICE_CTRL allows userspace to check if the kvm_device framework (e.g. KVM_CREATE_DEVICE) is supported by KVM. Move KVM_CAP_DEVICE_CTRL to the generic check for the two reasons:
1) it already supports arch agnostic usages (i.e. KVM_DEV_TYPE_VFIO). For example, userspace VFIO implementation may needs to create KVM_DEV_TYPE_VFIO on x86, riscv, or arm etc. It is simpler to have it checked at the generic code than at each arch's code.
2) KVM_CREATE_DEVICE has been added to the generic code.

---
 arch/arm64/kvm/arm.c       | 1 -
 arch/powerpc/kvm/powerpc.c | 1 -
 arch/riscv/kvm/vm.c        | 1 -
 arch/s390/kvm/kvm-s390.c   | 1 -
 virt/kvm/kvm_main.c        | 1 +
 5 files changed, 1 insertion(+), 4 deletions(-)
arch/arm64/kvm/arm.c +0 -1
index 9c5573bc4614..190e9c3b10a7 100644
@@ -212,7 +212,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 		r = vgic_present;
 		break;
 	case KVM_CAP_IOEVENTFD:
-	case KVM_CAP_DEVICE_CTRL:
 	case KVM_CAP_USER_MEMORY:
 	case KVM_CAP_SYNC_MMU:
 	case KVM_CAP_DESTROY_MEMORY_REGION_WORKS:
arch/powerpc/kvm/powerpc.c +0 -1
index 04494a4fb37a..21f9fbe96f6a 100644
@@ -541,7 +541,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 	case KVM_CAP_ENABLE_CAP:
 	case KVM_CAP_ONE_REG:
 	case KVM_CAP_IOEVENTFD:
-	case KVM_CAP_DEVICE_CTRL:
 	case KVM_CAP_IMMEDIATE_EXIT:
 	case KVM_CAP_SET_GUEST_DEBUG:
 		r = 1;
arch/riscv/kvm/vm.c +0 -1
index 65a964d7e70d..6efe93b282e1 100644
@@ -57,7 +57,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 
 	switch (ext) {
 	case KVM_CAP_IOEVENTFD:
-	case KVM_CAP_DEVICE_CTRL:
 	case KVM_CAP_USER_MEMORY:
 	case KVM_CAP_SYNC_MMU:
 	case KVM_CAP_DESTROY_MEMORY_REGION_WORKS:
arch/s390/kvm/kvm-s390.c +0 -1
index e4890e04b210..191d220b6a30 100644
@@ -567,7 +567,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 	case KVM_CAP_ENABLE_CAP:
 	case KVM_CAP_S390_CSS_SUPPORT:
 	case KVM_CAP_IOEVENTFD:
-	case KVM_CAP_DEVICE_CTRL:
 	case KVM_CAP_S390_IRQCHIP:
 	case KVM_CAP_VM_ATTRIBUTES:
 	case KVM_CAP_MP_STATE:
virt/kvm/kvm_main.c +1 -0
index 582757ebdce6..84073464aea0 100644
@@ -4524,6 +4524,7 @@ static long kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
 #endif
 	case KVM_CAP_BINARY_STATS_FD:
 	case KVM_CAP_SYSTEM_EVENT_DATA:
+	case KVM_CAP_DEVICE_CTRL:
 		return 1;
 	default:
 		break;
Signature · 1 line
2.27.0
reviewer Reviewed-by on v2 1/1
#1 lore raw
Quoted 8 lines · Wei Wang

KVM_CAP_DEVICE_CTRL allows userspace to check if the kvm_device framework (e.g. KVM_CREATE_DEVICE) is supported by KVM. Move KVM_CAP_DEVICE_CTRL to the generic check for the two reasons:
1) it already supports arch agnostic usages (i.e. KVM_DEV_TYPE_VFIO). For example, userspace VFIO implementation may needs to create KVM_DEV_TYPE_VFIO on x86, riscv, or arm etc. It is simpler to have it checked at the generic code than at each arch's code.
2) KVM_CREATE_DEVICE has been added to the generic code.

I think it's worth adding a link to the v1 patch, as that thread provides the context for how KVM ended up unconditionally advertising the feature, e.g. instead of trying to be more precise.

Quoted 2 lines

---

#2 lore raw
Quoted 14 lines · Sean Christopherson
Quoted 8 lines · Wei Wang

KVM_CAP_DEVICE_CTRL allows userspace to check if the kvm_device framework (e.g. KVM_CREATE_DEVICE) is supported by KVM. Move KVM_CAP_DEVICE_CTRL to the generic check for the two reasons:
1) it already supports arch agnostic usages (i.e. KVM_DEV_TYPE_VFIO). For example, userspace VFIO implementation may needs to create KVM_DEV_TYPE_VFIO on x86, riscv, or arm etc. It is simpler to have it checked at the generic code than at each arch's code.
2) KVM_CREATE_DEVICE has been added to the generic code.

I think it's worth adding a link to the v1 patch, as that thread provides the context for how KVM ended up unconditionally advertising the feature, e.g. instead of trying to be more precise.

Quoted 7 lines

[email protected]

Quoted 2 lines

---

Thanks.

Is this patch ready to be merged along with the link or need me to re-send to have the link added?