CVE-2026-99702: mm/vma: write lock source VMA before duplicating it in __split_vma()
In the Linux kernel, the following vulnerability has been resolved:
mm/vma: write lock source VMA before duplicating it in splitvma()
splitvma() duplicates the source VMA before taking its VMA write lock, and until the VMA is write-locked, a concurrent thread might set VMAMAYBEGUARDBIT, resulting in the duplication missing the flag.
If this happens, then fork may fail to duplicate guard markers and, if a file-backed mapping, THP collapse may also discard the markers.
Resolve this by taking the VMA write lock before vmareadup() and, since all callers of vmareadup() now hold the VMA write lock, assert that this is the case there, and add a comment to explain why it's needed.
[ljs@kernel.org: add VMA write assert in vmareadup(), commit message touchup]
Affected Software
Remediation
Recommended actions to resolve this vulnerability, in priority order.
- 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 6.12.107-1Fixed in 6.12.117-1Fixed in 7.3.5-1
Event History
Frequently Asked Questions
What conditions are required for this issue to occur?
A concurrent thread must set VMA_MAYBE_GUARD_BIT after __split_vma() duplicates the source VMA but before it obtains the source VMA write lock. The resulting duplicated VMA can then miss the guard-related flag.
What are the practical consequences if the race occurs?
A subsequent fork may fail to duplicate guard markers. For file-backed mappings, transparent huge page collapse may also discard those markers.
How is the issue fixed?
The fix acquires the source VMA write lock before calling vm_area_dup(), preventing concurrent modification of the relevant flag during duplication. It also adds an assertion that callers hold the VMA write lock.