The Linux kernel’s zram block device driver does not call bio_endio() when handling partial discard requests, causing the I/O operation to never complete. A process issuing blkdiscard for a sub‑page size blocks indefinitely in submit_bio_wait(), effectively hanging the calling task and stalling system responsiveness. The flaw affects only availability and does not provide any direct path to compromise confidentiality or integrity; a successful exploitation results in a denial of service that can disrupt services relying on the zram device.

All Linux kernel releases that have not yet incorporated the fix associated with commit 2d1f18efccdb8b29552399d024c36b705447e975 are affected. This includes every distribution that has not applied the patch and uses a zram device. Systems that do not enable zram or have disabled the device are not impacted.

The vulnerability requires local execution of blkdiscard on a /dev/zramX device, meaning the attacker must possess sufficient privileges or device permissions. Once the condition is triggered, the affected process remains blocked forever, leading to a system‑level denial of service. The EPSS score is not available and the flaw is not listed in CISA KEV, but the lack of a defensive mechanism and the high impact on availability indicate a significant risk, particularly in environments where zram is used for memory compression.