The kernel’s handling of executable memory for BPF just‑in‑time (JIT) compilation on RISC‑V is incorrect. When CONFIG_STRICT_MODULE_RWX is disabled, the patch_map function assumes vmalloc memory can be written, but recent code makes that memory read‑only. As a result, a page fault occurs during the copy of the BPF program, causing the JIT path to abort and the program to be left un‑compiled. This silent failure manifests as repeated kernel warnings and prevents any dependent subsystem from executing its BPF logic, effectively disabling BPF‑based features and introducing a denial‑of‑service vector.

Linux kernel builds for the RISC‑V architecture, including all kernel versions prior to the application of commit 8718e5a3090b, that are compiled with CONFIG_STRICT_MODULE_RWX set to "no" (the default) and that have BPF JIT support enabled. The regression applies to systems that use the default CONFIG_ARCH_OPTIONAL_KERNEL_RWX configuration or disable strict module memory protection while still building modules, and does not affect non‑MMU builds.

No CVSS score or EPSS data are available, and the vulnerability is not listed in CISA’s KEV catalog, suggesting it has not yet been observed in the wild. However, any process capable of loading a BPF program—typically a privileged user or a compromised kernel module—can trigger the defect, leading to a denied execution path. The risk is local or requires privilege to load BPF, but the impact is broad because many kernel subsystems rely on BPF JIT.