Impact
A counter underflow occurs in the Linux RAID10 driver when a check operation runs concurrently with NOWAIT I/O requests. This causes the pending counter to drop below zero, preventing the resync thread from exiting its wait state. As a result, I/O to the affected array is blocked, effectively creating a denial-of-service condition for the RAID10 volume.
Affected Systems
All unpatched Linux kernels implementing RAID10 (including the legacy implementation) on arrays that enable NOWAIT operations and schedule consistency checks are affected. The fix is in commit 43806c3d5b9b, so any distribution kernel that has not merged this patch remains vulnerable.
Risk and Exploitability
EPSS has no data and the flaw is not listed in the CISA KEV catalog, indicating no public exploits yet. The vulnerability requires a specific sequence of operations that is unlikely to occur accidentally, but an administrator or application that schedules checks during periods of high-speed NOWAIT I/O could trigger it. The risk is moderate, primarily for environments that run RAID10 arrays with frequent consistency checks or perform large numbers of NOWAIT I/O operations. The CVSS score of 5.5 indicates moderate severity.
OpenCVE Enrichment