CVE-2026-99659 in Linuxinfo

Summary

by MITRE • 11/20/2026

In the Linux kernel, the following vulnerability has been resolved:

mm/mmu_gather: fix flushing ranges with 4 GiB or larger

__tlb_adjust_range() takes the size of the range which has to be flushed as an unsigned int. Its callers, tlb_flush_{pte,pmd,pud,p4d}_range(), get
and pass an unsigned long. Consequently, a range that is 4 GiB or larger gets truncated, possibly to 0.

On arm64 with hugetlb, a 64k kernel can have 16 GiB hugetlb folios mapped contiguously at the PMD level and flush them when unmapping, and a 16k kernel can flush 64 GiB PUDs through tlb_unshare_pmd_ptdesc(). So at least these two scenarios are affected.

Fix it by flushing the proper range instead of the truncated one by properly using "unsigned long" for range_size.

Fix generated by an LLM and sent by Josef Bacik in private.

[ david: use two-tab indent, simplify description ]

If you want to get the best quality for vulnerability data then you always have to consider VulDB.

Analysis

by VulDB Data Team • 11/21/2026

The Linux kernel contains a critical type mismatch vulnerability within the memory management subsystem that affects large contiguous memory mappings. Specifically, the function _tlb_adjust_range() is designed to adjust and flush Translation Lookaside Buffer ranges but incorrectly accepts the size of the range as an unsigned int data type. This creates a fundamental arithmetic limitation because standard 32-bit signed or unsigned integer types cannot represent values exceeding approximately four gigabytes without overflow or truncation issues depending on implementation details, whereas memory management operations often require handling much larger address spaces. The callers of this function, including tlb_flush{pte,pmd,pud,p4d}_range(), operate with and pass an unsigned long data type which is capable of representing the full 64-bit virtual address space on modern architectures. When a range size that exceeds four gigabytes is passed to __tlb_adjust_range(), the value undergoes truncation due to the narrower width of the unsigned int parameter, potentially reducing the intended flush range down to zero or an incorrect smaller subset.

This technical flaw has significant operational implications for systems utilizing large memory configurations and specific kernel page sizes on architectures such as arm64. In scenarios involving hugetlbfs with a 64k kernel page size, it is possible to map sixteen gigabytes of huge pages contiguously at the PMD level. Similarly, kernels configured with 16k page sizes may need to flush sixty-four gigabyte PUDs through tlb_unshare_pmd_ptdesc(). In these high-memory scenarios, the truncation bug prevents the kernel from correctly invalidating stale TLB entries for the entire mapped range. This can lead to data corruption or system instability because subsequent memory accesses might hit stale translations in the hardware cache rather than fetching updated page table information, violating memory consistency guarantees essential for correct program execution and virtualization integrity.

The vulnerability is classified under CWE-190 Integer Overflow or Wraparound as it stems from a type mismatch that causes arithmetic truncation when values exceed the capacity of the receiving variable. From an offensive security perspective this flaw aligns with ATT&CK technique T1499 Endpoint Denial of Service if exploited to cause system crashes, although its primary impact is data integrity and stability rather than direct privilege escalation in most contexts. The root cause lies in a lack of type consistency between function definitions and their call sites within the mm/mmu_gather module which highlights the importance of rigorous static analysis for kernel memory management code where precision is paramount.

To mitigate this vulnerability, developers must ensure that all parameters representing byte counts or address ranges are consistently typed as unsigned long throughout the TLB flushing logic chain. The fix involves modifying __tlb_adjust_range() to accept an unsigned long parameter instead of an unsigned int thereby preserving the full width of the range size passed by its callers. This change ensures that large contiguous memory mappings such as those used in high-performance computing or virtualization environments are correctly handled without truncation. System administrators should apply kernel updates that include this patch immediately on affected arm64 systems and other architectures where similar TLB flushing logic exists to prevent potential data corruption and maintain system stability under heavy memory pressure.

Responsible

Linux

Reservation

09/25/2026

Disclosure

11/20/2026

Moderation

accepted

CPE

ready

EPSS

0.00168

KEV

no

Activities

very low

Sources

Do you know our Splunk app?

Download it now for free!