The flaw lies in the Linux kernel’s md RAID1 and RAID10 drivers: error detection relies on r1_bio or r10_bio->read_slot, which can become NULL or invalid after a bio is split and resubmitted. When the kernel fails to recognize the error path, it uses the memory allocation flag GFP_NOIO instead of the intended GFP_NOIO | __GFP_HIGH. Under memory pressure this incorrect flag can cause a deadlock, halting system operation. This impacts the integrity and availability of any system that runs the affected kernel with these RAID modules.

All Linux systems that load the md RAID1 or RAID10 kernel modules and have not incorporated the commits referenced by the advisory URLs. The advisory provides a git reference, but specific kernel version ranges are not provided, so any installation that has not applied the patch may be affected.

The CVSS score of 7.5 indicates high severity. The EPSS score of <1% shows a very low probability of exploitation, and the vulnerability is not listed in the CISA KEV catalog. No publicly available exploit exists. However, an attacker who can trigger repeated I/O failures on an MD array in a memory‑constrained environment—such as a local user or process with write access to the device—could activate the bug and force a kernel deadlock. The likely attack vector is local privilege escalation or denial of service through induced I/O errors.