The vulnerability pertains to the Linux kernel's eventfs subsystem. During a remount operation, eventfs walks through nodes while only holding an rcu_read_lock, which is insufficient for modifications performed elsewhere. This flaw introduces a race condition where list traversal can collide with list deletions, freeing of nodes, and attribute writes, potentially dereferencing poisoned list entries or using freed memory. The result is an unstable kernel that may crash or exhibit memory corruption.

All installations of the Linux kernel that have not applied the commit that corrects the eventfs_remount_lock behavior are affected. No specific versions are listed, so any current kernel not patched by the referenced commit is vulnerable. Sellers or users should verify their kernel versions against the latest stable releases that incorporate commit 340f0c7067a9.

The CVSS score is not provided, but the lack of an EPSS score and absence from the KEV list suggest that exploitation is not widely documented yet. Nevertheless, the vulnerability requires only local access to perform remount operations or write to the kprobe_events interface. An attacker who can exploit the race could trigger an out-of-bounds read or use-after-free, leading to a kernel crash. Because the conditions are simple and the attack vector is local, the risk is significant for systems that expose the /sys/kernel/tracing filesystem to untrusted or low-privileged processes.