Improper TLB invalidation in Linux kernel KVM/arm64 risks cross-VM memory exposure
AI analysis
CVE-2026-89775 is a flaw in the Linux kernel's KVM implementation for arm64, where the helper pgshift_level_to_ttl() mishandles negative walk levels when computing the size covered by a stage-1 TLB invalidation on the VNCR (nested virtualization) path. When a guest's stage-1 MMU is off, the level is encoded as the special value -127 (S1_MMU_DISABLED), but this is cast to a u8 and truncated to its bottom two bits, producing an invalidation size of 0 so stale TLB entries survive an operation that should have removed them. On a hypervisor, stale translations can allow memory mappings to outlive the pages they refer to, potentially letting a local attacker or guest read or corrupt memory across the trust boundary, consistent with the critical 9.3 CVSS (local attack, scope change, high confidentiality/integrity/availability impact). Only arm64 Linux systems running KVM with the nested-virtualization (VNCR/FEAT_NV2) code paths are affected; the advisory does not specify concrete affected or fixed version numbers. No public proof-of-concept, CISA KEV listing, or known exploitation has been reported.
What to do: Install a kernel update containing the upstream fix "KVM: arm64: Handle negative S1 walk levels in VNCR TLB size evaluation" from your vendor's stable tree once available. As an interim mitigation, do not enable nested virtualization (FEAT_NV2/NV2) for KVM guests on arm64 hosts. If you operate KVM guests on arm64, confirm whether your running kernel includes the fix by checking the package changelog or git history for this commit.
Affected
| Linux kernel - KVM arm64 nested virtualization (VNCR) TLB invalidation path | — |
Estimated exposure
nichelikely well under ~10,000 ARM64 KVM hosts with nested virtualization enabled — The vulnerable code exists only in recent arm64 kernels carrying KVM nested-virtualization (FEAT_NV2/VNCR) support, a configuration rarely enabled in production on ARM64 servers, and no public exposure-scan data covers this condition.
Order-of-magnitude estimate by the model from install counts, market share and public scan data it knows; verify before quoting.
Description
In the Linux kernel, the following vulnerability has been resolved: KVM: arm64: Handle negative S1 walk levels in VNCR TLB size evaluation Computing the effects of a TLB invalidation involves looking at the size of the mapping cached by the TLB. For S1 mappings such as VNCR, this is deducted from the combination of the base granule size and the mapping level. However, this implies that the S1 MMU is *on*. When the MMU is off, we indicate this with the level being set to a "creative" value of -127 (S1_MMU_DISABLED). This ends-up being misinterpreted by pgshift_level_to_ttl() as it doesn't handle negative levels at all (the level is immediately cast to a u8 and only the bottom two bits considered), leading to an invalidation size of 0. Not helpful. Tidy-up pgshift_level_to_ttl() to handle these negative levels, and ttl_to_size() to always return SZ_1G when no valid TTL is present. This allows the removal of open-coded checks for similar situations. Note that the check for a negative value not explicitely checking for S1_MMU_DISABLED is deliberate, so that actual negative levels introduced with LVA2 and D128 can take the same path if we ever support them.