KernelScan.io

HIGH Introduced in 2.6.16

configfs Rmdir UAF

CVE-2026-97949

CVSS 7.0 / 10.0 KernelScan AI

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

01

In the Linux kernel, the following vulnerability has been resolved: configfs: unhash the dentry before dropping the item in rmdir configfs_get_config_item() treats a hashed dentry as proof that sd->s_element is a live config_item. configfs_rmdir() breaks that: simple_rmdir() leaves the dentry hashed, the last reference to the item is dropped right after, and the dentry is only unhashed by d_delete() once ->rmdir() has returned. configfs_symlink() resolves its target holding no lock on it, so get_target() can land in that window: BUG: KASAN: slab-use-after-free in config_item_get+0x26/0x90 get_target fs/configfs/symlink.c:128 [inline] configfs_symlink+0x4ab/0x1030 fs/configfs/symlink.c:185 Unhash in configfs_remove_dir(), while the item is still guaranteed to be there. A reference obtained just before that stays harmless, as create_link() rechecks CONFIGFS_USET_DROPPING, already set by configfs_detach_prep(). Both configfs_unregister_subsystem() paths d_drop() after detaching, so this only makes rmdir match them.

02

Engine v0.6.0

Risk summary

A local user can trigger a use-after-free in the configfs filesystem by racing a directory removal against a symlink creation. The race window allows the symlink path to obtain a pointer to a freed kernel configuration object, leading to memory corruption that can cause kernel crashes or potential privilege escalation. Any system with configfs mounted and accessible to local users is affected.

Affectedfs/configfs/dir.c (configfs filesystem)

Vulnerability analysis

When a directory is removed from configfs, the filesystem briefly leaves the directory entry visible while the underlying configuration object is freed, creating a window where a concurrent symlink operation can obtain a stale pointer to that freed object. This use-after-free corrupts kernel heap memory. The fix removes the directory entry from the cache immediately after the removal succeeds but before dropping the object reference, closing the race window; any reference obtained just before this remains safe because the symlink creation path checks whether teardown is already in progress. This is reachable by a local user performing standard filesystem operations on a mounted configfs filesystem, requiring no special kernel privileges beyond access to the configfs mount point.

03

BranchIntroducedFixed inPatch commit
6.122.6.166.12.112446720051d44
6.182.6.166.18.535d7cbfb65e47
7.22.6.167.2.7925d8564f1b2
mainline2.6.167.3-rc3f06c2d26d199