Linux Kernel Netfs Cache Write Completion Race Condition Vulnerability

A race condition vulnerability has been identified in the Linux kernel's netfs component, specifically related to cache write completion and the signaling of all subrequests being queued. This issue arises when netfslib processes subrequests asynchronously. The problem occurs if the application thread indicates that all subrequests have been queued after the last subrequest has completed, particularly in scenarios involving the copy2cache functionality used by Ceph. In such cases, the request can hang indefinitely because the final collection and cleanup are not properly executed. The vulnerability has been addressed by adjusting the timing of the collection process and adding tracepoints for better monitoring.

Impact

Exploitation of this vulnerability can lead to requests hanging indefinitely, causing delays and potential failures in data processing and caching operations.