On Tue, Nov 12, 2024 at 8:41 AM Yonghong Song <yonghong.song@xxxxxxxxx> wrote:
>
> +
> +static void priv_stack_check_guard(void __percpu *priv_stack_ptr, int alloc_size,
> + struct bpf_prog *prog)
> +{
> + int cpu, underflow_idx = (alloc_size - PRIV_STACK_GUARD_SZ) >> 3;
> + u64 *stack_ptr;
> +
> + for_each_possible_cpu(cpu) {
> + stack_ptr = per_cpu_ptr(priv_stack_ptr, cpu);
> + if (stack_ptr[0] != PRIV_STACK_GUARD_VAL ||
> + stack_ptr[underflow_idx] != PRIV_STACK_GUARD_VAL) {
> + pr_err("BPF private stack overflow/underflow detected for prog %sx\n",
> + bpf_get_prog_name(prog));
> + break;
> + }
> + }
> +}
I was tempted to change pr_err() to WARN() to make sure this kinda bug
is very obvious, but left it as-is.
I think kasan-ing JITed load/stores and adding poison to guards
will be a bigger win.
The bpf prog/verifier bug will be spotted right away instead of
later during jit_free.
