CVE-2026-99735: mm/mglru: don't use stack pointers when walking PUD and PMD tables
In the Linux kernel, the following vulnerability has been resolved:
mm/mglru: don't use stack pointers when walking PUD and PMD tables
MGLRU passes stack copies of page-table entries in walkpudrange() and walkpmdrange() when walking to the next level. These copies can become stale when khugepaged or MADVDONTNEED (CONFIGPTRECLAIM) removes a PTE table concurrently.
This pattern dates back to commit bd74fdaea146 ("mm: multi-gen LRU: support page table walks"). Likely, earlier locking masked the latent bug. Later changes exposed it: commit 1d65b771bc08 ("mm/khugepaged: retractpagetables() without mmap or vma lock") allowed concurrent retraction, and commit cf4a65539c13 ("memcg-v1: no need for memcg locking for MGLRU") dropped the walker's outer RCU protection.
The fix is simple, so just fix the original commit so we don't have to speculate. To fix it, pass the actual PMD/PUD pointers. This restores the RCU-protected lookup and validation against the live PMD. Note that the PUD handling is required as well to handle folded page tables correctly in all cases.
Fix generated by an LLM and sent by Josef Bacik in private.
[ david: handle walkpudrange() as well, clarify+simplify description, adjust Fixes: ]
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 7.3.5-1
Event History
Frequently Asked Questions
What conditions are required for this race to occur?
The issue requires MGLRU to be walking PUD or PMD page tables while another activity concurrently removes a PTE table. The described concurrent removals can occur through khugepaged or MADV_DONTNEED when CONFIG_PT_RECLAIM is enabled.
Are systems with folded page tables affected by the same issue?
Yes. The fix specifically includes PUD handling because it is needed to handle folded page tables correctly in all cases.
What should be checked when assessing whether a kernel contains the fix?
Verify that the MGLRU page-table walker passes live PMD and PUD pointers rather than stack copies of page-table entries. The corrected behavior restores RCU-protected lookup and validation against the live PMD.