Look up another CVE · open CVE-2026-98164 in the interactive tracker
CVE-2026-98164
KVM: x86/mmu: Check write tracking in all address spaces
Published 2026-09-29 (20 days ago) by the Linux kernel CNA. Subsystem arch/x86. Debian urgency: not yet assigned.
What to do, per Debian release
Debian's own status for CVE-2026-98164 in every release this site tracks, and what each one means in practice. Every status below is the one the Debian Security Team publishes for this CVE. source
| Release | Status | Version | What this means |
|---|---|---|---|
| Debian 12bookworm | fixed | 6.1.187-1 | Update the kernel and reboot. Fixed in 6.1.187-1. A running kernel keeps the old code until the machine restarts. |
| Debian 13trixie | fixed | 6.12.111-1 | Update the kernel and reboot. Fixed in 6.12.111-1. A running kernel keeps the old code until the machine restarts. |
| Debian 14forky | fixed | 7.1.13-1 | Update the kernel and reboot. Fixed in 7.1.13-1. A running kernel keeps the old code until the machine restarts. |
| Debian unstablesid | fixed | 7.1.13-1 | Update the kernel and reboot. Fixed in 7.1.13-1. A running kernel keeps the old code until the machine restarts. |
| Debian 12bookworm · linux-6.12 | fixedDLA-4817-1 · 2026-10-04 · text | 6.12.111-1~deb12u1 | Update the kernel and reboot. DLA-4817-1 shipped the fix on 2026-10-04, in 6.12.111-1~deb12u1. A running kernel keeps the old code until the machine restarts. |
The vulnerability
What the bug is
kvm_gfn_is_write_tracked() checks only the supplied memslot, but page tracking is per-address-space and shadow pages are shared across all address spaces. With SMM, a GFN can therefore be write-tracked in one address space and appear untracked through the other. Check the supplied slot first, then the slot for the other address space. This ensures all callers honor write tracking regardless of the active address space. In particular, it prevents mmu_try_to_unsync_pages() from marking an upper-level shadow page unsync and eventually triggering the BUG in pte_list_remove(). [invert direction of the conditional. - Paolo]
Where it lives
The fix touches these files. If the subsystem is one you do not use, the practical exposure is lower, though the package is still the vulnerable one.
arch/x86/kvm/mmu/page_track.c
- What this is
- KERNEL VIRTUAL MACHINE FOR X86 (KVM/x86)MAINTAINERS section matched on arch/x86/kvm/*/ · kvm@vger.kernel.org · Supported
Show whether Debian builds this, per release
Not established. This site could not find the build rule for the affected file in the kernel trees Debian ships, so it has no answer about whether Debian compiles this code. Read that as unknown, never as absent.
- bookworm, trixie, forky, sid, bookworm · linux-6.12?
- ?
- not established - no Makefile entry for the affected file
Where a release above reads "not established", no build rule for the affected file was found in that release's kernel tree. That is a gap in this lookup, not a finding that the code is absent.
From the configuration Debian publishes for every kernel flavour it builds. It covers only the architectures Debian ships a linux-config package for, and nothing about a kernel you built yourself. Debian's own status above stays authoritative about the source package.MAINTAINERS
Signals
- Published
- 2026-09-29 (20 days ago)source
- CVSS
- 7.1 high
CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:Hsource recompute it - Attack vector
- local (AV:L)source
- EPSS
- 0.15%, higher than 3.7% of all CVEssource
- CISA KEV
- not listedthe KEV catalogue
- Debian urgency
- not yet assignedsource
- Subsystem
- arch/x86
Upstream fixes
The versions the CNA records as introducing and fixing this, and the commit for each. source
| Introduced | Fixed in | Commit |
|---|---|---|
| 4.2 | 6.1.187 | 09aa68552d25 |
| 4.2 | 6.6.156 | 429b6f43b4d8 |
| 4.2 | 6.12.108 | d8636c8f9f95 |
| 4.2 | 6.18.49 | c0a9bd5fca0b |
| 4.2 | 7.1.13 | ec8fcaf354c1 |
| 4.2 | 7.2 | 0f38453cdb2e |
Debian advisories
| Advisory | Date | Shipped to | Debian |
|---|---|---|---|
| DLA-4817-1 | 2026-10-04 | bookworm 6.12.111-1~deb12u1 | tracker |
Why the CNA scored it this way
Quoted from the kernel CNA's own scoring file, word for word. A highlighted metric carries the most severe value CVSS v3.1 defines for it, which is what pushes the score up. source
- AV:LAttack vector, local
- The fix patches kvm_gfn_is_write_tracked(), called from mmu_try_to_unsync_pages() via make_spte() on a vCPU page fault. Reaching it needs a VM with an SMM address space (KVM_SET_USER_MEMORY_REGION with as_id 1) driven through KVM_RUN on /dev/kvm. No remote protocol carries the input.
- AC:LAttack complexity, low
- The attacker builds every precondition: a gPT shadowed and tracked in the as_id 0 slot, then a write to that gfn while in SMM, which checks only the as_id 1 slot and unsyncs an upper-level SP. A later sync deterministically reaches pte_list_remove; no race or outside state is needed.
- PR:LPrivileges required, low
- Needs only an fd on /dev/kvm (commonly 0666 or kvm group) to create the VM, its SMM memslots and its own SMM code. No init-namespace root or capability is checked on the KVM_CREATE_VM/KVM_SET_USER_MEMORY_REGION/KVM_RUN path.
- UI:NUser interaction, none
- The attacker's own process runs the VM and triggers the page fault and the later root sync (kvm_mmu_sync_roots/kvm_sync_page) on its own. No other user does anything.
- S:UScope, unchanged
- I score the local VMM-process attacker, who controls guest SMM code. A guest-only trigger would need firmware to write chosen bytes into a gPT while in SMM, which kernel code cannot demonstrate. So the impact stays within the host kernel's authority.
- C:HConfidentiality impact, high
- Syncing an unsynced upper-level SP runs FNAME(sync_spte) on non-leaf SPTEs. shadowed_translation is uninitialised for those entries (no __GFP_ZERO), so a gfn match makes make_spte() build a huge leaf from spte_to_pfn() of the non-leaf SPTE. That maps KVM's child page-table page, and nearby host memory, readable into the guest.
- I:NIntegrity impact, none
- Non-leaf SPTEs from make_nonleaf_spte() lack shadow_host_writable_mask, so the bogus leaf built in sync_spte has ACC_WRITE stripped. I found no code path where this bug lets the guest write host memory.
- A:HAvailability impact, high
- Otherwise sync_spte calls drop_spte() on the non-leaf SPTE with a mismatched gfn. rmap_remove() then hits KVM_BUG_ON_DATA_CORRUPTION in pte_list_remove(), a BUG_ON with mmu_lock held when CONFIG_BUG_ON_DATA_CORRUPTION=y (the BUG the fix cites), crashing the host.