The Linux kernel’s cgroup cpuset subsystem can fail to detect a conflict when a remote partition is enabled under an existing local partition. The bug replaces an error check with a warning, allowing the remote partition to claim CPUs already owned by the ancestor local partition. This duplication of exclusive CPU ownership breaks the cpuset isolation guarantee and can lead to scheduling instability or a denial of service if critical processes are affected. The flaw is fixed by returning the appropriate error code, preventing the remote partition from being activated in conflict.

Any system running a Linux kernel that has not incorporated the commit that restores proper conflict testing is affected. This includes release candidates such as 7.3.0-rc3 and all earlier kernels lacking the fix. No official vendor patch information is listed, so the issue applies broadly to all unpatched kernel configurations.

The likely attack vector is a local user with permission to create and configure cgroups; such a user can reproduce the problem using the supplied commands that manipulate the cpuset hierarchy. Because the vulnerability requires local access to the cgroup filesystem and does not allow privilege escalation or remote exploitation, its exploitation probability is low to moderate. The EPSS score is not available, and the vulnerability is not in the CISA KEV catalog, but systems that rely on strict CPU isolation may still suffer service disruption if the conflict is triggered.