linux-kernel - [RFC PATCH 09/30] KVM: x86/mmu: Handle page fault for private MMIO

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20250529053513.1592088-10-yilun.xu@linux.intel.com>
Date: Thu, 29 May 2025 13:34:52 +0800
From: Xu Yilun <yilun.xu@...ux.intel.com>
To: kvm@...r.kernel.org,
	sumit.semwal@...aro.org,
	christian.koenig@....com,
	pbonzini@...hat.com,
	seanjc@...gle.com,
	alex.williamson@...hat.com,
	jgg@...dia.com,
	dan.j.williams@...el.com,
	aik@....com,
	linux-coco@...ts.linux.dev
Cc: dri-devel@...ts.freedesktop.org,
	linux-media@...r.kernel.org,
	linaro-mm-sig@...ts.linaro.org,
	vivek.kasireddy@...el.com,
	yilun.xu@...el.com,
	yilun.xu@...ux.intel.com,
	linux-kernel@...r.kernel.org,
	lukas@...ner.de,
	yan.y.zhao@...el.com,
	daniel.vetter@...ll.ch,
	leon@...nel.org,
	baolu.lu@...ux.intel.com,
	zhenzhong.duan@...el.com,
	tao1.su@...el.com,
	linux-pci@...r.kernel.org,
	zhiw@...dia.com,
	simona.vetter@...ll.ch,
	shameerali.kolothum.thodi@...wei.com,
	aneesh.kumar@...nel.org,
	iommu@...ts.linux.dev,
	kevin.tian@...el.com
Subject: [RFC PATCH 09/30] KVM: x86/mmu: Handle page fault for private MMIO

Add support for resolving page faults on private MMIO. This is part of
the effort to enable private assigned devices (known as TDI in TDISP
spec).

Private MMIOs are set to KVM as vfio_dmabuf typed memory slot, which is
another type of can-be-private memory slot just like the gmem slot.
Like gmem slot, KVM needs to map its GFN as shared or private based on
the current state of the GFN's memory attribute. When page fault
happens for private MMIO but private <-> shared conversion is needed,
KVM still exits to userspace with exit reason KVM_EXIT_MEMORY_FAULT and
toggles KVM_MEMORY_EXIT_FLAG_PRIVATE. Unlike gmem slot, vfio_dmabuf
slot has only one backend MMIO resource, the switching of GFN's
attribute won't change the way of getting PFN, the vfio_dmabuf specific
way, kvm_vfio_dmabuf_get_pfn().

Signed-off-by: Xu Yilun <yilun.xu@...ux.intel.com>
---
 arch/x86/kvm/mmu/mmu.c   | 9 +++++++--
 include/linux/kvm_host.h | 2 +-
 2 files changed, 8 insertions(+), 3 deletions(-)

diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 40d33bd6b532..547fb645692b 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -4501,8 +4501,13 @@ static int kvm_mmu_faultin_pfn_private(struct kvm_vcpu *vcpu,
 		return -EFAULT;
 	}
 
-	r = kvm_gmem_get_pfn(vcpu->kvm, fault->slot, fault->gfn, &fault->pfn,
-			     &fault->refcounted_page, &max_order);
+	if (kvm_slot_is_vfio_dmabuf(fault->slot))
+		r = kvm_vfio_dmabuf_get_pfn(vcpu->kvm, fault->slot, fault->gfn,
+					    &fault->pfn, &max_order);
+	else
+		r = kvm_gmem_get_pfn(vcpu->kvm, fault->slot, fault->gfn,
+				     &fault->pfn, &fault->refcounted_page,
+				     &max_order);
 	if (r) {
 		kvm_mmu_prepare_memory_fault_exit(vcpu, fault);
 		return r;
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index b850d3cff83c..dd9c876374b8 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -620,7 +620,7 @@ struct kvm_memory_slot {
 
 static inline bool kvm_slot_can_be_private(const struct kvm_memory_slot *slot)
 {
-	return slot && (slot->flags & KVM_MEM_GUEST_MEMFD);
+	return slot && (slot->flags & (KVM_MEM_GUEST_MEMFD | KVM_MEM_VFIO_DMABUF));
 }
 
 static inline bool kvm_slot_is_vfio_dmabuf(const struct kvm_memory_slot *slot)
-- 
2.25.1