UBUNTU-CVE-2026-99120: Medium severity vulnerability
In the Linux kernel, the following vulnerability has been resolved: KVM: x86/mmu: Cope with the SS bit being set in #NPF exit codes When running Hyper-V from Windows 11, KVM is seeing the SS bit set in nested page fault error codes produced for shadowed NPT. The AMD manual is unclear about when SS is set, including whether it can be set when SupervisorShadowStackEn (bit 6) is not set in the MISCCTL field of the VMCB. Nevertheless, the unexpected SS bit causes out of bounds accesses in permissionfault(), seen as either WARNs or UBSAN reports: kernel: UBSAN: array-index-out-of-bounds in arch/x86/kvm/mmu.h:225:27 kernel: index 34 is out of range for type 'u16 [16]' (in this case, the page fault error code is 342=0x44, i.e. SS|U; the report shows SS|U|W happening as well). Fix the handling of unknown error codes by dropping SS and (with a WARN) any bits not among those that permissionfault() expects to receive. At the same time, given the uncertainty about when hardware sets SS, follow the processor's behavior when constructing the fault's error code, and preserve the SS bit as passed to FNAME(walkaddrgeneric). While the behavior is surprising, there's a possibility that Hyper-V expects it (Hyper-V does not even enable CET unless it finds MBEC/GMET!), so retain the information when reflecting the fault to L1 instead of unconditionally discarding it. Bisection shows that this started happening with commit 687ee95c1b6d ("KVM: nSVM: enable GMET for guests"), even though SHSTK was available to guests since Linux 6.18. However, this is not because bit 1 of MISCCTL (GMET) somehow enables reporting the SS bit; rather, it's simply that Hyper-V never uses CET unless GMET is also present. The WARN is reproducible even with gmet=0, with some hacking of kvm-unit-tests's cet.c testcase to ensure that the first access to the shadow stack triggers a page fault.
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Compensating control
Update KVM x86 MMU nested page-fault handling to drop the SS bit and, with a WARN, any error-code bits not expected by permission_fault() before indexing its permission table; preserve the SS bit when reflecting the fault to L1 and when passing it to FNAME(walk_addr_generic).
Event History
Frequently Asked Questions
Which systems are most likely to encounter this issue?
The described trigger occurs when KVM runs Hyper-V from Windows 11 and receives nested page-fault exit codes with the SS bit set for shadowed NPT. The issue is in the x86 KVM MMU path.
What is the observed impact before the fix?
Unexpected SS bits in nested page-fault error codes can cause an out-of-bounds access in permission_fault(). This has been observed as kernel WARN messages or UBSAN array-index-out-of-bounds reports.
How does the fix mitigate the problem?
The fix drops the SS bit and warns on other unexpected bits before passing an error code to permission_fault(), limiting that function to the error-code bits it expects. It also preserves SS when constructing the fault error code passed to the address-walk path.