CVE-2026-99647: mm/huge_memory: split huge PUDs when adjusting VMA boundaries
In the Linux kernel, the following vulnerability has been resolved:
mm/hugememory: split huge PUDs when adjusting VMA boundaries
PFN mappings such as VFIO PCI BAR mappings can use PUD leaves while allowing VMA splits at page granularity. However, vmaadjusttranshuge() only handles PMDs.
After faulting a writable PUD-sized VFIO BAR mapping, changing its first page to read-only with mprotect() makes the PMD walk treat the PUD leaf as a PMD table. Applying MADVDONTFORK to the first PMD-sized chunk leaves the PUD spanning two VMAs: the PMD-aligned boundary skips splitting.
Walk the page tables once at each adjusted boundary and split a huge PUD or PMD leaf as needed. Handle PUD leaves even when the boundary is PMD-aligned.
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.12.118 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 6.18.64 - Upgrade
Upgrade
linux/Kernelto a version that resolves this vulnerability.Fixed in 7.3.5 - Upgrade
Upgrade
debian/linuxto a version that resolves this vulnerability.Fixed in 6.1.176-1Fixed in 6.1.192-2Fixed in 7.3.5-1
Event History
Frequently Asked Questions
What conditions are required to trigger this issue?
A process must have a PFN mapping that uses a PUD leaf, such as a VFIO PCI BAR mapping, and the mapping must permit VMA splits at page granularity. The described sequence involves faulting a writable PUD-sized mapping, changing its first page to read-only with mprotect(), and applying MADV_DONTFORK to the first PMD-sized chunk.
Are ordinary memory mappings implicated?
The provided information specifically identifies PFN mappings using PUD leaves, with VFIO PCI BAR mappings given as an example. It does not establish that ordinary anonymous or file-backed mappings are affected.
How can administrators determine whether their environment is exposed?
Review whether workloads use VFIO PCI BAR PFN mappings that can be backed by PUD-sized leaf entries and subsequently alter protections or fork behavior on subranges using mprotect() or MADV_DONTFORK. The data does not provide a runtime detection method or affected kernel-version list.