CVE-2026-72477 PUBLISHED

fs/ntfs3: call _ntfs_bad_inode() when failing to rename

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

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

fs/ntfs3: call _ntfs_bad_inode() when failing to rename

It is safe to call _ntfs_bad_inode on live inodes since: commit 519b078998ce ("fs/ntfs3: Exclude call make_bad_inode for live nodes.")

The WARN_ON was added when it wasn't safe by: commit d99208b91933 ("fs/ntfs3: cancle set bad inode after removing name fails")

Replace the WARN_ON with a call to _ntfs_bad_inode() to prevent further operations on the inconsistent inode.

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 - ntfs3 registers export_operations and rename on an NFS- or SMB-exported NTFS volume reaches ni_rename() through nfsd/ksmbd vfs_rename without any local syscall; hybrid NAS and Windows/Linux file-server deployments are a reasonable high-impact reach path for this VFS rename flaw. AC:L - An attacker controls the on-disk NTFS layout and rename arguments so ni_add_name succeeds and both ni_remove_name steps fail deterministically; syzkaller reproduces this with a crafted loop image and no races or victim-specific timing outside attacker influence. PR:N - Exploitation needs only permission to rename on an exported NTFS share (including guest/anonymous NFS or SMB write access) or to supply removable media later mounted by the host; no init-namespace root, CAP_SYS_ADMIN, or other elevated host credentials are required. UI:N - Once an administrator has mounted and exported the NTFS volume (or automounted dual-boot/shared storage), the attacker triggers the flawed rename directly; no further victim clicks, file opens, or cooperative actions are needed beyond that standing configuration. S:U - Impact is kernel heap/metadata corruption and denial of service within the same kernel security authority; this is standard local or network filesystem exploitation, not a VM escape, IOMMU bypass, or cross-scope sandbox breakout. C:H - Leaving the inode operable after a failed rename rollback preserves inconsistent MFT/directory index state; subsequent lookups and index walks parse attacker-influenced corrupt metadata, matching ntfs3 guidance to score metadata corruption that can yield kernel memory disclosure as High. I:H - Without ni_bad set, further writes, unlinks, and renames continue against mismatched $FILE_NAME and $I30 state, propagating attacker-steerable directory/MFT corruption and heap metadata damage that can be leveraged for arbitrary kernel writes or control-flow hijack. A:H - The vulnerable path emits WARN_ON in ni_rename (kernel panic on panic_on_warn) and syzbot also hit a kernel paging fault in folio_wait_bit_common after follow-on operations on the inconsistent inode; any kernel oops, hang, or panic from continued use scores Availability High.

Product Status

Vendor Linux
Product Linux
Versions Default: unaffected
  • affected from 519b078998ce6e729f98dccf35505b4756985d11 to ff825bf0521f6da2f30878cbad18ab7b341bc31b (excl.)
  • affected from 519b078998ce6e729f98dccf35505b4756985d11 to e8ed78f40eecd0176fda71d673f6957c98e7ffbe (excl.)
Vendor Linux
Product Linux
Versions Default: affected
  • Version 6.17 is affected
  • unaffected from 0 to 6.17 (excl.)
  • unaffected from 7.1.5 to 7.1.* (incl.)
  • unaffected from 7.2 to * (incl.)

References