{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-99312","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-09-25T10:25:14.407Z","datePublished":"2026-11-18T08:42:06.136Z","dateUpdated":"2026-11-19T13:52:45.586Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-11-19T13:52:45.586Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nKVM: x86: Unconditionally reset MMU context when emulating INIT\n\nForce an MMU reset when emulating an INIT even if paging was already\ndisabled, as the MMU page role includes guest_mode (L2), i.e. an INIT while\nL2 is running with CR0.PG=0 will result in KVM running L1 with a stale MMU\ncontext.\n\nAlternatively, the reset could be scoped strictly to paging being enabled\nor L2 being active, but SMM is also encoded in the page role.  Even though\nit _should_ be impossible for the vCPU to be INIT'd while SMM is active,\nthere's zero value in not handling SMM.  At that point, the only scenario\nwhere KVM _doesn't_ need to reset the MMU context is on an INIT with the\nvCPU running in Real Mode or Unpaged Protect Mode, and while INIT in those\nmodes is somewhat common, it's far from a hot path and re-initializing the\nMMU is not expensive, especially when the role isn't changing (KVM elides\nmost of the re-configuration in that case).\n\nNote, like the Fixes commit, this bug only affects nested SVM as INIT is an\nunconditional intercept on VMX, i.e. it's impossible to emulate INIT while\nL2 is active.\n\nTemporarily keep the guest TLB flush when CR0.PG is being cleared even\nthough it is fully redundant with the flush that is emitted on all INIT\nevents (KVM itself only does RESET when a vCPU is being created, i.e. when\nCR0.PG must be '0')."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H","baseScore":8.8,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - L1 reaches kvm_vcpu_reset from kvm_apic_accept_events after nested_svm_vmrun: L2 is entered with CR0.PG clear, and a sibling vCPU's INIT is taken in-guest because svm_check_nested_events returns 0 when INTERCEPT_INIT is clear. No network message carries the trigger.\nAC:L - That same user clears L2 CR0.PG and INTERCEPT_INIT, then KVM_SET_USER_MEMORY_REGION calls kvm_arch_flush_shadow_memslot. Default KVM_X86_QUIRK_SLOT_ZAP_ALL makes kvm_tdp_mmu_invalidate_roots drop the root still named by vmcb01.control.nested_cr3. kvm_amd nested=1 is the default.\nPR:L - KVM_CREATE_VM, KVM_SET_USER_MEMORY_REGION, and KVM_RUN need only a non-root kvm user. VMRUN and the INIT IPI run inside that VM; kvm_apic_accept_events calls kvm_vcpu_reset with no host capability check.\nUI:N - The attacker issues the memslot ioctl and the INIT IPI. kvm_vcpu_reset does not wait for another user to mount media or acknowledge a prompt.\nS:C - svm_leave_nested switches to vmcb01 without rewriting nested_cr3, and the skipped kvm_mmu_reset_context leaves kvm_mmu_reload seeing a valid root.hpa, so VM-entry walks a freed host page as the NPT root.\nC:H - vmcb01.control.nested_cr3 still holds the HPA of the TDP root freed by kvm_tdp_mmu_put_root after kvm_tdp_mmu_invalidate_roots. The CPU walks that freed page as the top-level NPT, so a reclaimed page-table page discloses host PFNs.\nI:H - The same dangling nested_cr3 lets the attacker aim the walked NPT at guest-controlled data and write arbitrary host physical pages. It is a page-table use-after-free, not only a guest crash.\nA:H - Until the freed root is reused, npf_interception livelocks: kvm_mmu_page_fault installs SPTEs into the guest_mode=1 root while hardware still walks vmcb01's stale nested_cr3, so the vCPU thread never completes the instruction."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["arch/x86/kvm/x86.c"],"versions":[{"version":"ed129ec9057f89d615ba0c81a4984a90345a1684","lessThan":"f55a3e4e08dc004b4bab7c1e9323d7e7e96e266a","status":"affected","versionType":"git"},{"version":"ed129ec9057f89d615ba0c81a4984a90345a1684","lessThan":"24f644ea26c2ebb7b14c3baa2aa92d37fde3095f","status":"affected","versionType":"git"},{"version":"ed129ec9057f89d615ba0c81a4984a90345a1684","lessThan":"d47bae9eb33b3ca0843a0954b7d0cafb354c99b0","status":"affected","versionType":"git"},{"version":"ed129ec9057f89d615ba0c81a4984a90345a1684","lessThan":"c7c288d178cb0b78d3464a616b16b1aa0b316049","status":"affected","versionType":"git"},{"version":"6425c590d0cc6914658a630a40b7f8226aa028c3","status":"affected","versionType":"git"},{"version":"d40ef0a511676bd65ca9acb295430c07af59ab85","status":"affected","versionType":"git"},{"version":"5.15.81","lessThan":"5.16","status":"affected","versionType":"semver"},{"version":"6.0.11","lessThan":"6.1","status":"affected","versionType":"semver"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["arch/x86/kvm/x86.c"],"versions":[{"version":"6.1","status":"affected"},{"version":"0","lessThan":"6.1","status":"unaffected","versionType":"semver"},{"version":"6.12.118","lessThanOrEqual":"6.12.*","status":"unaffected","versionType":"semver"},{"version":"6.18.64","lessThanOrEqual":"6.18.*","status":"unaffected","versionType":"semver"},{"version":"7.3.5","lessThanOrEqual":"7.3.*","status":"unaffected","versionType":"semver"},{"version":"7.4-rc2","lessThanOrEqual":"*","status":"unaffected","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"operator":"OR","negate":false,"cpeMatch":[{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.1","versionEndExcluding":"6.12.118"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.1","versionEndExcluding":"6.18.64"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.1","versionEndExcluding":"7.3.5"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.1","versionEndExcluding":"7.4-rc2"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"5.15.81"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"6.0.11"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/f55a3e4e08dc004b4bab7c1e9323d7e7e96e266a"},{"url":"https://git.kernel.org/stable/c/24f644ea26c2ebb7b14c3baa2aa92d37fde3095f"},{"url":"https://git.kernel.org/stable/c/d47bae9eb33b3ca0843a0954b7d0cafb354c99b0"},{"url":"https://git.kernel.org/stable/c/c7c288d178cb0b78d3464a616b16b1aa0b316049"}],"title":"KVM: x86: Unconditionally reset MMU context when emulating INIT","x_generator":{"engine":"bippy-1.2.0"}}}}