{ "dataType": "CVE_RECORD", "dataVersion": "5.2", "cveMetadata": { "cveId": "CVE-2026-98927", "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "state": "PUBLISHED", "assignerShortName": "Linux", "dateReserved": "2026-09-25T10:25:14.380Z", "datePublished": "2026-10-23T12:06:01.528Z", "dateUpdated": "2026-10-24T14:09:32.255Z" }, "containers": { "cna": { "providerMetadata": { "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67", "shortName": "Linux", "dateUpdated": "2026-10-24T14:09:32.255Z" }, "descriptions": [ { "lang": "en", "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nKVM: arm64: nv: Fix life cycle of the nested_mmus array\n\nThe nested_mmus array holds the shadow page tables that are used when\na guest is running a nested context. These structures are allocated on\nVCPU_INIT for whole guest, which implies that they may have to be\nrelocated as the array grows.\n\nShould a VCPU_INIT occur whilst a vcpu is actively running an L2 and\nthat the allocation requires relocation, that vcpu will still be\nrunning with a pointer to the previous structure, which will have been\nfreed.\n\nFix this by turning the array of structures to an array of pointers,\nwhich is now allocated at VM creation, sized to the absolute maximum\nthat KVM can handle.\n\nIn turn, each VCPU_INIT contributes S2_MMU_PER_VCPU to the pool. No\nreallocation is ever performed, and the life cycle of each object is\nmuch clearer:\n\n- the nested_mmus array is allocated in kvm_init_nested(), and freed\n in kvm_arch_destroy_vm()\n\n- s2_mmu structures are allocated in kvm_vcpu_init_nested(), and freed\n on kvm_arch_flush_shadow_all()\n\nFinally, the freeing of vcpu->arch.vncr_array is made consistent\nrather than being done on some failure paths, but not others." } ], "metrics": [ { "cvssV3_1": { "version": "3.1", "vectorString": "CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H", "baseScore": 7, "baseSeverity": "HIGH" }, "scenarios": [ { "lang": "en", "value": "AV:L - A local process invokes KVM_ARM_VCPU_INIT on a vCPU fd; kvm_vcpu_ioctl() reaches patched kvm_vcpu_init_nested() through kvm_arch_vcpu_ioctl_vcpu_init() and kvm_setup_vcpu(). No network data path reaches this code.\nAC:H - The attacker can run one L2 vCPU while initializing another, but reachability requires FEAT_NV2 hardware booted in the explicitly non-default experimental kvm-arm.mode=nested mode. A low-privileged attacker cannot establish that host state.\nPR:L - kvm_vcpu_ioctl() requires the VM's owning mm but performs no capable() check. A basic local user granted /dev/kvm access can create the VM and vCPUs and request KVM_ARM_VCPU_HAS_EL2.\nUI:N - The same process can issue KVM_RUN for the L2 vCPU and KVM_ARM_VCPU_INIT for a newly created vCPU from concurrent threads; no other user must act.\nS:U - The required trigger is the host-side KVM_ARM_VCPU_INIT ioctl, which L2 guest code cannot issue. The demonstrated attack is therefore a /dev/kvm user corrupting the same host kernel, not a guest-to-host escape.\nC:H - After kvm_vcpu_init_nested() frees the old array, vcpu->arch.hw_mmu remains stale; __load_stage2() consumes reclaimable pgd_phys and vtcr fields, while kvm_handle_guest_abort() follows the reclaimable pgt pointer, enabling disclosure of host memory.\nI:H - kvm_handle_guest_abort() passes stale hw_mmu->pgt state to stage-2 mapping operations, and kvm_vcpu_put_hw_mmu() modifies the stale refcnt. Reclaiming the freed object can redirect kernel writes and corrupt host code or data.\nA:H - Dereferencing the freed kvm_s2_mmu through __load_stage2(), nested fault handling, or kvm_vcpu_put_hw_mmu() can cause a host kernel oops or panic." } ] } ], "affected": [ { "product": "Linux", "vendor": "Linux", "defaultStatus": "unaffected", "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git", "programFiles": [ "arch/arm64/include/asm/kvm_host.h", "arch/arm64/include/asm/kvm_nested.h", "arch/arm64/kvm/arm.c", "arch/arm64/kvm/nested.c" ], "versions": [ { "version": "4f128f8e1aaac189f83d0f828bcdb2986d8d2e51", "lessThan": "ea242bdc29e8a8ff0e2447a64b3841114779f3af", "status": "affected", "versionType": "git" }, { "version": "4f128f8e1aaac189f83d0f828bcdb2986d8d2e51", "lessThan": "81e3f065a6485626f816d791dfd5e00ddd12eff5", "status": "affected", "versionType": "git" }, { "version": "4f128f8e1aaac189f83d0f828bcdb2986d8d2e51", "lessThan": "512c93f4461b9a3735987f48423d438a9cd49e6a", "status": "affected", "versionType": "git" }, { "version": "4f128f8e1aaac189f83d0f828bcdb2986d8d2e51", "lessThan": "33346f8960c7bb6a3b4e273b5cfe25c5a8be349f", "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/arm64/include/asm/kvm_host.h", "arch/arm64/include/asm/kvm_nested.h", "arch/arm64/kvm/arm.c", "arch/arm64/kvm/nested.c" ], "versions": [ { "version": "6.11", "status": "affected" }, { "version": "0", "lessThan": "6.11", "status": "unaffected", "versionType": "semver" }, { "version": "6.12.113", "lessThanOrEqual": "6.12.*", "status": "unaffected", "versionType": "semver" }, { "version": "6.18.56", "lessThanOrEqual": "6.18.*", "status": "unaffected", "versionType": "semver" }, { "version": "7.2.9", "lessThanOrEqual": "7.2.*", "status": "unaffected", "versionType": "semver" }, { "version": "7.3", "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": "6.11", "versionEndExcluding": "6.12.113" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.11", "versionEndExcluding": "6.18.56" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.11", "versionEndExcluding": "7.2.9" }, { "vulnerable": true, "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.11", "versionEndExcluding": "7.3" } ] } ] } ], "references": [ { "url": "https://git.kernel.org/stable/c/ea242bdc29e8a8ff0e2447a64b3841114779f3af" }, { "url": "https://git.kernel.org/stable/c/81e3f065a6485626f816d791dfd5e00ddd12eff5" }, { "url": "https://git.kernel.org/stable/c/512c93f4461b9a3735987f48423d438a9cd49e6a" }, { "url": "https://git.kernel.org/stable/c/33346f8960c7bb6a3b4e273b5cfe25c5a8be349f" } ], "title": "KVM: arm64: nv: Fix life cycle of the nested_mmus array", "x_generator": { "engine": "bippy-1.2.0" } } } }