From 048f1efd0cefd37f8249ca746fc84f998f33a5d0 Mon Sep 17 00:00:00 2001 From: Andrew Lukoshko Date: Thu, 6 Aug 2026 13:44:37 +0200 Subject: [PATCH] CVE-2026-64561: KVM: x86: check for invalid/obsolete root after making MMU pages available Taken verbatim from CentOS Stream 10 merge request 3068, which cherry-picks upstream commit 2abd5287f083 ("KVM: x86: Check for invalid/obsolete root *after* making MMU pages available") by Sean Christopherson and Paolo Bonzini. If reclaiming shadow pages zaps an in-use root, KVM would go on to map memory into a root already marked invalid. Child shadow pages inherit their parent's role, so any children created during the map/fetch land on the list of active MMU pages while flagged invalid, breaking the invariant introduced in 5.9 that invalid pages are never on that list. The fix moves the is_page_fault_stale() check to after make_mmu_pages_available(). Applies with --fuzz=0 against arch/x86/kvm/mmu/{mmu.c,paging_tmpl.h}, no fuzz and no offsets. Verified against the centos-stream-10 main branch: the el10_2 z-stream tree this kernel is built from (6.12.0-211.38.1.el10_2) is not published as a branch or tag, but both hunks sit in direct_page_fault() and FNAME(page_fault)(), which are unchanged across 6.12 z-streams. Note that MR 3068 is still a Draft upstream. If Red Hat revises it before merge, this patch needs to be re-synced. JIRA: RHEL-224013 --- config.yaml | 5 + ...-x86-Check-for-invalid-obsolete-root.patch | 97 +++++++++++++++++++ 2 files changed, 102 insertions(+) create mode 100644 files/2013-CVE-2026-64561-KVM-x86-Check-for-invalid-obsolete-root.patch diff --git a/config.yaml b/config.yaml index 7c40ce4..6eddf9e 100644 --- a/config.yaml +++ b/config.yaml @@ -56,6 +56,10 @@ actions: name: "2012-gve-Enable-reading-max-ring-size-in-DQO-QPL-mode.patch" number: 2012 + - type: "patch" + name: "2013-CVE-2026-64561-KVM-x86-Check-for-invalid-obsolete-root.patch" + number: 2013 + - replace: - target: "kernel*.config" find: | @@ -466,6 +470,7 @@ actions: - name: "Andrew Lukoshko" email: "alukoshko@almalinux.org" line: + - "KVM: x86: check for invalid/obsolete root after making MMU pages available {CVE-2026-64561}" - "hpsa: bring back deprecated PCI ids #CFHack #CFHack2024" - "mptsas: bring back deprecated PCI ids #CFHack #CFHack2024" - "megaraid_sas: bring back deprecated PCI ids #CFHack #CFHack2024" diff --git a/files/2013-CVE-2026-64561-KVM-x86-Check-for-invalid-obsolete-root.patch b/files/2013-CVE-2026-64561-KVM-x86-Check-for-invalid-obsolete-root.patch new file mode 100644 index 0000000..0c5c7ed --- /dev/null +++ b/files/2013-CVE-2026-64561-KVM-x86-Check-for-invalid-obsolete-root.patch @@ -0,0 +1,97 @@ +From 884d1cab4bfadf53f05ca36f35f06d3cc1f916a2 Mon Sep 17 00:00:00 2001 +From: CKI Backport Bot +Date: Tue, 4 Aug 2026 10:08:07 +0000 +Subject: [PATCH] KVM: x86: Check for invalid/obsolete root *after* making MMU + pages available + +JIRA: https://redhat.atlassian.net/browse/RHEL-224013 +CVE: CVE-2026-64561 +Backported from tree(s): linux + +KVM: x86: Check for invalid/obsolete root *after* making MMU pages available + +Check for a "stale" page fault, i.e. for an invalid and/or obsolete root, +after making MMU pages available for the shadow MMU. If reclaiming shadow +pages zaps an in-use root, i.e. marks it invalid, then KVM will attempt to +map memory into an invalid root. On its own, populating an invalid root is +"fine", but because child shadow pages inherit their parent's role, any +children created during the map/fetch will be created as invalid pages, +thus violating KVM's invariant that invalid pages are never on the list of +active MMU pages. + +Note, the underlying flaw has existed since KVM first started tracking +invalid roots in 2008 (commit 2e53d63acba7, "KVM: MMU: ignore zapped root +pagetables"), but the true badness only came along in 2020 (Linux 5.9) +with the invariant that invalid shadow pages can't be on the list of +active pages. + +Note #2, inheriting role.invalid when creating child shadow pages is also +far from ideal; that flaw will be addressed separately. + +Reported-by: Hyunwoo Kim +Fixes: f95eec9bed76 ("KVM: x86/mmu: Don't put invalid SPs back on the list of active pages") +Cc: stable@vger.kernel.org +Signed-off-by: Sean Christopherson +Signed-off-by: Paolo Bonzini +(cherry picked from commit 2abd5287f08319fa35764566b15c6e22cb1068db) +Signed-off-by: CKI Backport Bot +--- + arch/x86/kvm/mmu/mmu.c | 9 +++++---- + arch/x86/kvm/mmu/paging_tmpl.h | 10 ++++++---- + 2 files changed, 11 insertions(+), 8 deletions(-) + +diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c +index ab4bbb692064..0c4553e57731 100644 +--- a/arch/x86/kvm/mmu/mmu.c ++++ b/arch/x86/kvm/mmu/mmu.c +@@ -4814,16 +4814,17 @@ static int direct_page_fault(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault + if (r != RET_PF_CONTINUE) + return r; + +- r = RET_PF_RETRY; + write_lock(&vcpu->kvm->mmu_lock); + +- if (is_page_fault_stale(vcpu, fault)) +- goto out_unlock; +- + r = make_mmu_pages_available(vcpu); + if (r) + goto out_unlock; + ++ if (is_page_fault_stale(vcpu, fault)) { ++ r = RET_PF_RETRY; ++ goto out_unlock; ++ } ++ + r = direct_map(vcpu, fault); + + out_unlock: +diff --git a/arch/x86/kvm/mmu/paging_tmpl.h b/arch/x86/kvm/mmu/paging_tmpl.h +index 901cd2bd40b8..6465de820e70 100644 +--- a/arch/x86/kvm/mmu/paging_tmpl.h ++++ b/arch/x86/kvm/mmu/paging_tmpl.h +@@ -827,15 +827,17 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, struct kvm_page_fault *fault + } + #endif + +- r = RET_PF_RETRY; + write_lock(&vcpu->kvm->mmu_lock); + +- if (is_page_fault_stale(vcpu, fault)) +- goto out_unlock; +- + r = make_mmu_pages_available(vcpu); + if (r) + goto out_unlock; ++ ++ if (is_page_fault_stale(vcpu, fault)) { ++ r = RET_PF_RETRY; ++ goto out_unlock; ++ } ++ + r = FNAME(fetch)(vcpu, fault, &walker); + + out_unlock: +-- +GitLab +