CVE-2023-54123

What it actually does

The vulnerability is a memory leak in the Linux kernel RAID10 driver where conf->bio_split is not freed on error paths in raid10_run(), despite multiple sites freeing the conf structure.

Why it matters to the business

Memory leaks in kernel RAID10 configurations can cause gradual resource exhaustion on storage systems, leading to instability, crashes, or unplanned downtime; low EPSS and absence from KEV indicate limited immediate urgency but still require patching on affected Linux kernels.

How an attack unfolds

An attacker would trigger repeated error conditions during RAID10 array initialization to induce leaks, but the flaw is a non-exploitable memory management error with no known threat actors or active campaigns.

How to fix it

Apply the upstream fix present in commit 133008af833b4f2e021d2c294c29c70364a3f0ba; upgrade affected Linux kernel trees to a version containing this commit.

How to confirm the fix worked

Verify the presence of commit 133008af833b4f2e021d2c294c29c70364a3f0ba in the kernel source; monitor kernel memory statistics for RAID10 allocations after initialization errors.

Typical attack path

1. Trigger error path in raid10_run() during array setup
2. Cause repeated allocation without corresponding free of bio_split
3. Accumulate unreleased kernel memory over time

How to detect it

Examine kernel logs and memory profiling tools for repeated RAID10 initialization failures or growing memory usage tied to bio_split structures; no public exploit artifacts exist.

What your security team should do

• Prioritize upgrade to kernel containing commit 133008af833b4f2e021d2c294c29c70364a3f0ba on systems using software RAID10
• Review kernel changelogs for memory management fixes in md/raid drivers
• Include this kernel update in routine patching cycles given stable low EPSS