An ext4 kernel bug was identified in which the function ext4_write_inline_data_end is called without verifying that the inode still allows inline data. When a previous attempt to convert inline data to an extent fails, a flag is cleared while a related flag remains set. Subsequent writes trigger a BUG_ON in ext4_write_inline_data, causing the kernel to panic. This flaw does not directly expose secrets; its primary impact is system unavailability if an attacker can force writes to a filesystem with the data=journal option enabled.

All Linux kernel builds supplying the ext4 filesystem prior to the commit that introduced the guard on ext4_journalled_write_end. The issue applies to any kernel that supported the data=journal mount option and inline data proposals.

The CVSS score is 5.5, and the EPSS score is < 1%, but the vulnerability is not featured in CISA KEV. Because the trigger requires local filesystem writes with the data=journal option, the attack vector is local. An attacker who can write to a mounted ext4 volume could cause a denial‑of‑service through a kernel panic, but the lack of public exploit and restricted prerequisites keep the risk moderate. Updating the kernel to a version with the patch eliminates the vulnerability.