CVE-2022-50730: 
Linux Kernel vulnerability analysis and mitigation

CVE-2022-50730 is a Linux kernel vulnerability in the ext4 filesystem subsystem related to a spurious warning triggered during inode eviction when the dioread_nolock mount option is in use. When evicting an inode, a race condition between the eviction path and an unwritten extent conversion kworker can cause the inode to be marked dirty again after I_FREEING is set, triggering a false-positive WARN_ON_ONCE in ext4_evict_inode() at fs/ext4/inode.c:227. The fix removes the overly broad warning check rather than addressing a true security flaw. It was published on December 24, 2025, and affects the Linux kernel across multiple stable branches. The CVSS category is estimated as Medium, with an EPSS score of approximately 0.018% (Feedly).

The root cause is a race condition (CWE-362) in the ext4 filesystem's inode eviction path (ext4_evict_inode()) when the dioread_nolock mount option is active. After writeback of newly allocated dirty blocks, a kworker converting unwritten extents to written may merge extents and free extent blocks, causing the inode to be re-dirtied even after I_FREEING has been set. The existing inode->i_io_list check and WARN_ON_ONCE in ext4_evict_inode() did not account for this legitimate scenario. Crucially, ext4_evict_inode() already waits for all extent conversions to complete before the check, so no actual inode use-after-free occurs — the issue is purely a false-positive kernel warning. The fix removes the problematic WARN_ON_ONCE check (Feedly).

The practical impact of CVE-2022-50730 is limited to a spurious kernel warning (WARN_ON_ONCE) being emitted in system logs under specific workload conditions involving ext4 with dioread_nolock. There is no inode use-after-free, no data corruption, no privilege escalation, and no confidentiality or integrity impact. The warning could cause concern for system administrators or trigger automated alerting systems, but does not represent an exploitable security condition (Feedly).

• Logs: Kernel log entries containing WARNING: CPU: <N> PID: <N> at fs/ext4/inode.c:227 ext4_evict_inode+0x875/0xc60 with a call trace through evict, iput, do_unlinkat, and __x64_sys_unlinkat may indicate the affected code path was triggered on an unpatched kernel (Feedly).

The fix is to apply the upstream Linux kernel patch that removes the false-positive WARN_ON_ONCE check in ext4_evict_inode(). Patches have been committed to multiple stable kernel branches (commit hashes: 0d041b72, 3b893cc9, b085fb43, bc12ac98, bdc698ce). Distributions including SUSE have issued kernel security updates incorporating this fix (e.g., SUSE advisories SUSE-2026-0263, SUSE-2026-0317, SUSE-2026-0350, SUSE-2026-0411). Users should update to a patched kernel version provided by their Linux distribution. As a temporary workaround, avoiding the dioread_nolock mount option for ext4 filesystems will prevent the race condition from triggering the warning (Feedly, Linux Security).