CVE-2025-22120

What it actually does

Missing unlock of mapping->invalidate_lock in ext4_setattr when ext4_inode_attach_jinode fails, leaving a read lock held and causing subsequent tasks to block indefinitely on the invalidate semaphore.

Why it matters to the business

Linux systems using ext4 can experience prolonged task hangs and system unresponsiveness under memory pressure during file attribute changes, increasing downtime risk for workloads on affected kernels.

How an attack unfolds

An attacker with local access triggers ext4_setattr under low-memory conditions so that ext4_inode_attach_jinode returns an error; the missing unlock leaves invalidate_lock held, blocking later readers and producing a hung task detectable via kernel logs.

How to fix it

Upgrade to a fixed kernel version: mainline 6.6.89 or later, or the specific Red Hat packages listed (0:5.14.0-284.96.1.rt14.381.el9_2 and equivalents for all listed kernel-rt variants). Reboot after package installation.

How to confirm the fix worked

Confirm the running kernel version matches or exceeds the fixed release; reproduce the original error path under memory pressure and verify no hung-task messages appear and that invalidate_lock is released.

Typical attack path

1. Obtain local code execution on target system
2. Induce memory pressure while performing setattr operations on ext4 files
3. Trigger OOM path inside ext4_inode_attach_jinode
4. Observe indefinite blocking on invalidate_lock during subsequent file reads

How to detect it

Search kernel logs for EXT4-fs error messages mentioning ext4_setattr together with hung-task backtraces showing rwsem_down_read_slowpath and filemap_get_pages; correlate with processes blocked for >120 seconds.

What your security team should do

• Prioritize kernel updates on all ext4-using systems before the CISA remediation window if applicable
• Monitor for hung-task warnings referencing ext4_setattr in kernel logs
• Test the updated kernel in staging environments that exercise ext4 setattr and read workloads