CVE-2026-89560: landlock: Require LANDLOCK_ACCESS_FS_MAKE_REG for whiteout creation
In the Linux kernel, the following vulnerability has been resolved:
landlock: Require LANDLOCKACCESSFSMAKEREG for whiteout creation
Whiteout objects are used in the upper layer of an OverlayFS to indicate that the file with this name does not exist in the unified view, even if it is present in one of the lower layer file systems.
For the userspace implementations of OverlayFS (fuse-overlayfs), whiteout objects can be created from userspace as well:
mknod(2) with SIFCHR and makedev(0, 0) renameat2(2) with RENAMEWHITEOUT, creating the whiteout in the old place of the moved file.
This commit guards whiteout creation in both of these cases with LANDLOCKACCESSFSMAKEREG. Whiteout objects are not considered character devices and are not bound to a driver.
LANDLOCKACCESSFSMAKEREG describes the same permission class as a whiteout object: creating one is the only SIFCHR creation that the VFS exempts from CAPMKNOD, so it is as unprivileged as creating a regular file, while LANDLOCKACCESSFSMAKECHAR and LANDLOCKACCESSFSMAKEBLOCK keep meaning the creation of devices that expose a kernel interface [1].
For the mknod(2) case, introduce a Landlock erratum. The creation of whiteout objects through mknod(2) was previously guarded using LANDLOCKACCESSFSMAKECHAR, and it is now guarded using LANDLOCKACCESSFSMAKEREG.
For the renameat2(2) case, fix a bug: Before this commit, renameat2(2) with RENAMEWHITEOUT would create a directory entry even when all LANDLOCKACCESSFSMAKE rights were denied.
This does not affect normal renames within layered OverlayFS mounts: When doing a regular rename() on a mounted fuse-overlayfs, it is the fuse-overlayfs daemon that exercises renameat2() with RENAMEWHITEOUT, and only the Landlock domain of that daemon is checked there.
Depends-on: 49c9e09d9610 ("landlock: Fix handling of disconnected directories") Depends-on: fe72ce6710cb ("landlock: Add errata documentation section") [mic: Record why LANDLOCKACCESSFSMAKEREG is the matching right, and add link(2) to the user doc]
Affected Software
Event History
Frequently Asked Questions
Which userspace operations are relevant when assessing exposure?
Userspace OverlayFS implementations such as fuse-overlayfs can create whiteouts using mknod(2) with S_IFCHR and makedev(0, 0), or renameat2(2) with RENAME_WHITEOUT. The latter creates the whiteout at the moved file's old location.
What Landlock right should policies require for these operations?
Whiteout creation is guarded by LANDLOCK_ACCESS_FS_MAKE_REG in both the mknod(2) and renameat2(2) cases. Policies governing locations where whiteouts may be created should account for this permission.
Does this involve access to a character-device driver or other kernel device interface?
No. Whiteouts are not considered character devices and are not bound to a driver. Unlike ordinary character or block device creation, whiteout creation is exempt from CAP_MKNOD and is treated as comparable to regular-file creation.