On Fri, Jul 12, 2024 at 3:14 AM Geliang Tang <geliang@xxxxxxxxxx> wrote:
>
> From: Geliang Tang <tanggeliang@xxxxxxxxxx>
>
> Run bloom_filter_map selftests (./test_progs -t bloom_filter_map) on a
> Loongarch platform, an error message "JIT doesn't support bpf-to-bpf calls"
> is got in user space, together with an unexpected errno EINVAL (22), not
> ENOTSUPP (524):
>
> libbpf: prog 'inner_map': BPF program load failed: Invalid argument
> libbpf: prog 'inner_map': -- BEGIN PROG LOAD LOG --
> JIT doesn't support bpf-to-bpf calls
> callbacks are not allowed in non-JITed programs
> processed 37 insns (limit 1000000) max_states_per_insn 1 total_states
> -- END PROG LOAD LOG --
> libbpf: prog 'inner_map': failed to load: -22
> libbpf: failed to load object 'bloom_filter_map'
> libbpf: failed to load BPF skeleton 'bloom_filter_map': -22
> setup_progs:FAIL:bloom_filter_map__open_and_load unexpected error: -22
> #16 bloom_filter_map:FAIL
>
> Although in jit_subprogs(), the error number does be set as "ENOTSUPP":
>
> verbose(env, "JIT doesn't support bpf-to-bpf calls\n");
> err = -ENOTSUPP;
> goto out_free;
>
> But afterwards in fixup_call_args(), such error number is ignored, and
> overwritten as "-EINVAL":
>
> verbose(env, "callbacks are not allowed in non-JITed programs\n");
> return -EINVAL;
>
> This patch fixes this by changing return values of fixup_call_args() from
> "-EINVAL" to "err ?: -EINVAL". With this change, errno 524 is got in user
> space now:
>
> libbpf: prog 'inner_map': BPF program load failed: unknown error (-524)
> libbpf: prog 'inner_map': -- BEGIN PROG LOAD LOG --
> JIT doesn't support bpf-to-bpf calls
> processed 37 insns (limit 1000000) max_states_per_insn 1 total_states
> -- END PROG LOAD LOG --
> libbpf: prog 'inner_map': failed to load: -524
> libbpf: failed to load object 'bloom_filter_map'
> libbpf: failed to load BPF skeleton 'bloom_filter_map': -524
> setup_progs:FAIL:bloom_filter_map__open_and_load unexpected error: -524
>
> Signed-off-by: Geliang Tang <tanggeliang@xxxxxxxxxx>
> ---
> kernel/bpf/verifier.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index c0263fb5ca4b..aa589fedd036 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -19717,14 +19717,14 @@ static int fixup_call_args(struct bpf_verifier_env *env)
> #ifndef CONFIG_BPF_JIT_ALWAYS_ON
> if (has_kfunc_call) {
> verbose(env, "calling kernel functions are not allowed in non-JITed programs\n");
> - return -EINVAL;
> + return err ?: -EINVAL;
> }
> if (env->subprog_cnt > 1 && env->prog->aux->tail_call_reachable) {
> /* When JIT fails the progs with bpf2bpf calls and tail_calls
> * have to be rejected, since interpreter doesn't support them yet.
> */
> verbose(env, "tail_calls are not allowed in non-JITed programs with bpf-to-bpf calls\n");
> - return -EINVAL;
> + return err ?: -EINVAL;
> }
> for (i = 0; i < prog->len; i++, insn++) {
> if (bpf_pseudo_func(insn)) {
> @@ -19732,7 +19732,7 @@ static int fixup_call_args(struct bpf_verifier_env *env)
> * have to be rejected, since interpreter doesn't support them yet.
> */
> verbose(env, "callbacks are not allowed in non-JITed programs\n");
> - return -EINVAL;
> + return err ?: -EINVAL;
Nack.
We're not changing this.
pw-bot: cr
