The Linux sh_tmu driver normally powers the device on during initialization and then turns it off when it is not required. In builds where the driver is used as a clockevent provider and PREEMPT_RT is enabled, the driver interacts with the kernel’s clockevent framework while holding a raw spinlock. In that context the driver cannot safely invoke runtime power‑management or clock APIs, because those APIs may sleep and rely on regular spinlocks. The resulting conflict triggers a lockdep warning, "Invalid wait context", indicating that the driver is attempting a power or clock operation while a raw spinlock is held. The description suggests that this situation can cause incorrect power or clock management, which may lead to kernel instability on PREEMPT_RT builds. This conclusion is inferred because the advisory states the issue may be problematic and recommends caution, but it does not explicitly claim that a system crash will occur.

The problem affects Linux kernel builds that include the sh_tmu driver compiled as a clockevent provider and that enable PREEMPT_RT. The advisory references kernel version 6.18 and indicates that the driver’s behavior has remained unchanged for a long time. Thus any distribution or deployment that ships a kernel 6.18 or newer with the sh_tmu driver enabled for clock events and builds it with PREEMPT_RT is potentially impacted.

The CVSS score of 5.5 indicates a moderate severity, and the EPSS score is not available. The vulnerability is not catalogued in the CISA KEV list. Because the flaw involves internal kernel synchronization and requires that an attacker either have kernel‑level privileges or be able to modify the kernel source, exploitation in the wild is unlikely. The most probable risk is the generation of lockdep warnings and potential instability when the system is under heavy real‑time workload. The attack vector is inferred to involve interaction with the driver during its initialization phase under a raw spinlock context.