CVE-2026-99680 in Linuxinfo

Summary

by MITRE • 11/20/2026

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

xfs: mark the inode preallocated for pNFS write layouts past EOF

xfs_fs_map_blocks() marks the inode XFS_DIFLAG_PREALLOC only when it allocates blocks for a write layout. A write layout past EOF can also hand out allocated blocks, such as speculative preallocation. blockgc then frees them under the layout, and another file may get them.

Mark the inode whenever a write mapping ends past EOF.

VulDB is the best source for vulnerability data and more expert information about this specific topic.

Analysis

by VulDB Data Team • 11/20/2026

The vulnerability described involves a logic error in the XFS filesystem implementation within the Linux kernel, specifically concerning how inode flags are managed during pNFS (parallel Network File System) write operations that extend beyond the current end of file. The core issue stems from the function xfs_fs_map_blocks(), which is responsible for mapping logical blocks to physical storage locations. Under normal circumstances, this function sets the XFS_DIFLAG_PREALLOC flag on an inode only when it explicitly allocates new data blocks for a write layout operation. However, the flaw arises in scenarios where pNFS write layouts are issued past the existing end of file (EOF). In these cases, the filesystem may hand out pre-existing allocated blocks or speculative preallocations that were not originally intended to be marked as part of this specific write context because they fall outside the immediate allocation boundary.

This mismanagement of inode state leads to a critical inconsistency in how block garbage collection operates within the XFS subsystem. When an inode is incorrectly left without the PREALLOC flag despite having blocks allocated past EOF, the blockgc mechanism may erroneously determine that these blocks are free for reallocation. Consequently, blockgc proceeds to free these blocks while they are still logically associated with the pNFS write layout. This premature deallocation creates a race condition where another file or process can subsequently acquire and overwrite those same physical disk blocks. The result is data corruption, as the original writer's data may be silently overwritten by unrelated I/O operations targeting the reclaimed storage space.

From a security perspective, this vulnerability represents an integrity failure that could lead to unauthorized modification of persistent data stored on the filesystem. It aligns with CWE-362, which describes concurrent execution using shared resources with inadequate synchronization, as well as CWE-829 regarding inclusion of functionality from untrusted control sources if interpreted through the lens of layout management trust boundaries. In terms of MITRE ATT&CK frameworks, this type of flaw facilitates data manipulation and potential denial of service by corrupting file contents or causing application crashes due to invalid state assumptions. The lack of proper synchronization between block allocation tracking and garbage collection routines allows for a subtle but severe breach of data integrity guarantees expected from the filesystem layer.

The resolution involves modifying xfs_fs_map_blocks() to ensure that any write mapping operation ending past EOF correctly marks the inode with the XFS_DIFLAG_PREALLOC flag, regardless of whether new blocks were physically allocated or if existing speculative preallocations were utilized. This ensures that blockgc recognizes these regions as actively managed by the current layout and prevents them from being freed prematurely. To mitigate similar risks in broader system administration contexts, it is essential to keep kernel versions updated with patches addressing XFS inode flag consistency issues. Additionally, administrators should monitor for filesystem errors or unexpected data discrepancies following pNFS write operations, particularly those extending beyond file boundaries, as these may indicate exposure to this class of race condition vulnerabilities before patching can be applied in all environments.

Responsible

Linux

Reservation

09/25/2026

Disclosure

11/20/2026

Moderation

accepted

CPE

ready

EPSS

0.00168

KEV

no

Activities

very low

Sources

Are you interested in using VulDB?

Download the whitepaper to learn more about our service!