CVE-2026-72210 PUBLISHED

ntfs: fix off-by-one in mapping pairs decoding bounds checks

Assigner: Linux
Reserved: 09.08.2026 Published: 15.08.2026 Updated: 17.08.2026

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

ntfs: fix off-by-one in mapping pairs decoding bounds checks

In ntfs_mapping_pairs_decompress(), attr_end points one byte past the end of the attribute record:

<pre>attr_end = (u8 *)attr + le32_to_cpu(attr->length); </pre>

The two bounds checks validating that mapping pair data bytes fit within the attribute use strict greater-than (>), which allows a one-byte out-of-bounds read when the data extends exactly to attr_end:

b = *buf & 0xf; if (b) { if (unlikely(buf + b > attr_end)) // off-by-one goto io_error; for (deltaxcn = (s8)buf[b--]; b; b--) deltaxcn = (deltaxcn << 8) + buf[b]; }

When buf + b == attr_end, the check evaluates to false and buf[b] reads one byte past the valid attribute boundary. The same pattern appears in the LCN delta bytes check.

Fix both checks to use >= so that buf[b] at exactly attr_end is correctly rejected as out of bounds.

Metrics

CVSS Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
CVSS Score: 9.8

AV:N - The legacy NTFS driver is exported via nfsd and commonly shared over ksmbd; remote NFS/SMB reads on an NTFS-backed export reach ntfs_mapping_pairs_decompress() through VFS read/iomap and ntfs_map_runlist() without further local access. AC:L - An attacker fully controls a crafted NTFS image offline and can place mapping pairs to hit the exact off-by-one boundary, reliably triggering the flaw on mount or on the next access that decompresses that attribute extent. PR:N - Exploitation needs only the ability to trigger server-side parsing of the mounted volume (e.g., guest/anonymous SMB or permissive NFS export); no init-namespace root or CAP_SYS_ADMIN on the victim is required at trigger time. UI:N - After an administrator mounts the poisoned volume, a remote client can trigger the bug with protocol read/open operations alone; no additional end-user clicks, plug-in events, or mount actions are needed during exploitation. S:U - Impact stays within kernel/host context (metadata parsing, runlist corruption, and resulting I/O); it does not cross a VM, container, or IOMMU security boundary to another authority. C:H - The flawed bounds check permits a one-byte out-of-bounds kernel read past the attribute end, often into MFT record slack that may contain stale cached data, matching OOB-read disclosure impact. I:H - The leaked byte is folded into runlist length/LCN deltas, yielding attacker-influenced cluster mappings that can misdirect subsequent kernel block I/O and corrupt arbitrary on-volume metadata or file data. A:H - Corrupt runlists and failed decompression paths surface as ntfs_error/EIO during mount or I/O and can panic or wedge the node under repeated remote access, giving a practical denial-of-service.

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 to bfe835e535fe0aa5767fdd8116f62e835ba50b55 (excl.)
  • affected from 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 to 18760a74ef7c28df93726445b5595162e62ed341 (excl.)
  • affected from 0 to 7.1.5 (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • unaffected from 7.1.5 to 7.1.* (incl.)
  • unaffected from 7.2 to * (incl.)

References