{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-99343","assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","state":"PUBLISHED","assignerShortName":"Linux","dateReserved":"2026-09-25T10:25:14.409Z","datePublished":"2026-11-19T14:21:51.069Z","dateUpdated":"2026-11-22T11:07:45.748Z"},"containers":{"cna":{"providerMetadata":{"orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux","dateUpdated":"2026-11-22T11:07:45.748Z"},"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nKVM: x86/mmu: Bug the VM if KVM attempts to walk more levels than the MMU has\n\nExtend the \"EFER.LMA && !CR4.PAE\" check, which exists largely to guard\nagainst KVM configuring a paging32 MMU with more than 2 levels of paging,\nwith a very explicit check for exactly that: that KVM isn't trying to walk\nmore levels of paging than the MMU template provides.  I.e. harden KVM\nagainst all bugs that would cause KVM to generates accesses beyond the\nbounds of guest_walker's arrays, regardless of how KVM ended up with the\nmisconfigured MMU.\n\nNote, don't use w->cpu_role.base.level directly as the paging64 template\nonly provides two levels of page tables for PAE paging on 32-bit hosts, and\nhandles the third level by manually emulating the PDPTR access."}],"metrics":[{"cvssV3_1":{"version":"3.1","vectorString":"CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:H/A:H","baseScore":7.1,"baseSeverity":"HIGH"},"scenarios":[{"lang":"en","value":"AV:L - KVM_SET_NESTED_STATE in kvm_vcpu_ioctl calls svm_set_nested_state, which copies the L1 VMCB save area through svm_copy_vmrun_state into vmcb01. nested_svm_init_mmu_context passes that CR4 and EFER to kvm_init_shadow_npt_mmu, so the paging mode is chosen from ioctl payload.\nAC:L - The caller puts EFER.LMA=1, EFER.LME=0 and CR4.PAE=0 in that save area; nested_vmcb_check_save checks PAE only when LME and CR0.PG are both set. KVM_RUN then enters paging32_walk_addr_generic. Nested SVM and NPT both default on (svm.c nested=true, npt_enabled=true), and the caller supplies the VMCB, memslots and guest PTEs.\nPR:L - svm_set_nested_state runs from the KVM_SET_NESTED_STATE vCPU ioctl with no capability check. A process that can open /dev/kvm, create a VM, and set EFER.SVME in guest CPUID can issue it.\nUI:N - That same process writes the nested state and calls KVM_RUN, and paging32_walk_addr_generic runs on the vCPU it created.\nS:U - Only svm_copy_vmrun_state writes the illegal L1 EFER/CR4 into vmcb01. nested_svm_vmrun copies vcpu->arch.efer and vcpu->arch.cr4, which set_efer and kvm_set_cr4 keep consistent, so the bad role is a host-ioctl state. The overflow is in the host stack frame of paging32_walk_addr_generic.\nC:N - gpte_changed and update_accessed_dirty_bits reread out-of-range walker slots that alias prefetch_ptes, ptep_user and pt_access inside the same guest_walker, which hold the caller's PTE bytes and memslot HVA. walker.fault and the GPA returned by paging32_gva_to_gpa are filled from the faulting address and the leaf PTE.\nI:H - paging32_walk_addr_generic writes table_gfn, pte_gpa, ptep_user, ptes and pt_access at index cpu_role.base.level-1 (3 or 4) past PT_MAX_FULL_LEVELS (2). FNAME(fetch) passes the aliased table_gfn to kvm_mmu_get_shadow_page, which allocates a shadow page and write-tracks that attacker-chosen GFN.\nA:H - account_shadowed calls __kvm_write_track_add_gfn with __gfn_to_memslot's result, and update_gfn_write_track dereferences slot->arch.gfn_write_track with no NULL check, so a GFN in no memslot oopses the host. A gpte_changed mismatch instead returns RET_PF_RETRY and spins the vCPU thread."}]}],"affected":[{"product":"Linux","vendor":"Linux","defaultStatus":"unaffected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["arch/x86/kvm/mmu/paging_tmpl.h"],"versions":[{"version":"fe660f7244d7e237ab7726813dc9aec8e94900d6","lessThan":"d6f3877eabf159167ee9972d627396412a04ac74","status":"affected","versionType":"git"},{"version":"fe660f7244d7e237ab7726813dc9aec8e94900d6","lessThan":"54afd548602bb96b3fbb0b4b11ccbbfeb5fb46fc","status":"affected","versionType":"git"},{"version":"fe660f7244d7e237ab7726813dc9aec8e94900d6","lessThan":"73f427e5afd94edd5d0e7cdddc2ecc794f28c492","status":"affected","versionType":"git"},{"version":"fe660f7244d7e237ab7726813dc9aec8e94900d6","lessThan":"115f9405a2ed6d78ee7ad8436bdd638743f6d1d4","status":"affected","versionType":"git"}]},{"product":"Linux","vendor":"Linux","defaultStatus":"affected","repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","programFiles":["arch/x86/kvm/mmu/paging_tmpl.h"],"versions":[{"version":"5.14","status":"affected"},{"version":"0","lessThan":"5.14","status":"unaffected","versionType":"semver"},{"version":"6.12.117","lessThanOrEqual":"6.12.*","status":"unaffected","versionType":"semver"},{"version":"6.18.61","lessThanOrEqual":"6.18.*","status":"unaffected","versionType":"semver"},{"version":"7.3.2","lessThanOrEqual":"7.3.*","status":"unaffected","versionType":"semver"},{"version":"7.4-rc1","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":"5.14","versionEndExcluding":"6.12.117"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"5.14","versionEndExcluding":"6.18.61"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"5.14","versionEndExcluding":"7.3.2"},{"vulnerable":true,"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionStartIncluding":"5.14","versionEndExcluding":"7.4-rc1"}]}]}],"references":[{"url":"https://git.kernel.org/stable/c/d6f3877eabf159167ee9972d627396412a04ac74"},{"url":"https://git.kernel.org/stable/c/54afd548602bb96b3fbb0b4b11ccbbfeb5fb46fc"},{"url":"https://git.kernel.org/stable/c/73f427e5afd94edd5d0e7cdddc2ecc794f28c492"},{"url":"https://git.kernel.org/stable/c/115f9405a2ed6d78ee7ad8436bdd638743f6d1d4"}],"title":"KVM: x86/mmu: Bug the VM if KVM attempts to walk more levels than the MMU has","x_generator":{"engine":"bippy-1.2.0"}}}}