{"id":"CVE-2026-98150","published":"2026-09-25T11:17:46.580","lastModified":"2026-09-25T15:18:06.997","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nbpf: Fix BPF_F_CPU validation for sparse CPU IDs\n\nBPF_F_CPU stores the target CPU ID in the upper 32 bits of the map\noperation flags. bpf_map_check_op_flags() currently compares that ID\nwith num_possible_cpus(), which is the number of possible CPUs rather\nthan a bound on CPU IDs.\n\nOn an arm64 QEMU guest with a CPU device-tree hole, the possible CPU\nmask was 0,2-3. A userspace program using raw bpf() syscalls creates\na BPF_MAP_TYPE_PERCPU_ARRAY and performs update and lookup operations\nfor each CPU by setting BPF_F_CPU and the CPU ID in the flags.\n\nWith the old check, CPU 1 is incorrectly accepted while valid CPU 3 is\nrejected with -ERANGE. The CPU 1 update then reaches the per-CPU map\naccess path and triggers:\n\n  Unable to handle kernel paging request at virtual address ...\n  pc : __pi_memcpy_generic+0x5c/0x22c\n  lr : bpf_percpu_array_update+0x2dc/0x2e8\n  Call trace:\n    __pi_memcpy_generic\n    bpf_map_update_value\n    map_update_elem\n    __sys_bpf\n\nCheck the CPU ID against nr_cpu_ids and cpu_possible() instead. This\nrejects CPU IDs outside the valid range and CPUs absent from the\npossible mask, while allowing valid sparse CPU IDs.","cvssScore":7,"cvssSeverity":"HIGH","cvssVector":"CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:H/I:H/A:H","cwes":[],"vendors":[],"products":[],"references":[{"url":"https://git.kernel.org/stable/c/bdc5941f6eeef90b76a07fd8ca38ac1933ba2195","tags":[]},{"url":"https://git.kernel.org/stable/c/ed54bf564ac52699cf4def3d0c2125d493e756f9","tags":[]}],"exploitRefs":[],"hasPoc":false,"ai":null}