Re: [PATCH bpf-next 06/18] bpf: Prepare architecture JIT support for stack arguments

[Yonghong Song <yonghong.song@linux.dev>]

On 4/24/26 10:48 AM, bot+bpf-ci@kernel.org wrote:
>> diff --git a/include/linux/filter.h b/include/linux/filter.h
>> index b77d0b06db6e..911205dd670e 100644
>> --- a/include/linux/filter.h
>> +++ b/include/linux/filter.h
> [ ... ]
>
>> diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
>> index cfb35a2decf6..fa4b971c5289 100644
>> --- a/kernel/bpf/btf.c
>> +++ b/kernel/bpf/btf.c
> [ ... ]
>
>> diff --git a/kernel/bpf/core.c b/kernel/bpf/core.c
>> index ec8523e6e4eb..7522b3d7b267 100644
>> --- a/kernel/bpf/core.c
>> +++ b/kernel/bpf/core.c
> [ ... ]
>
>> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
>> index e041c182c614..0ba140dabe93 100644
>> --- a/kernel/bpf/verifier.c
>> +++ b/kernel/bpf/verifier.c
>> @@ -12254,6 +12254,11 @@ static int check_kfunc_args(struct bpf_verifier_env *env, struct bpf_kfunc_call_
>>   			MAX_BPF_FUNC_REG_ARGS);
>>   		return -EINVAL;
>>   	}
>> +	if (nargs > MAX_BPF_FUNC_REG_ARGS && !bpf_jit_supports_stack_args()) {
>> +		verbose(env, "JIT does not support kfunc %s() with %d args\n",
>> +			func_name, nargs);
>> +		return -ENOTSUPP;
>> +	}
> Can this new check ever execute? The existing check immediately above already
> rejects all kfuncs with nargs > MAX_BPF_FUNC_REG_ARGS by returning -EINVAL.
>
> Since control only reaches the new if statement when the previous check
> passes, nargs must be <= MAX_BPF_FUNC_REG_ARGS. The condition
> 'nargs > MAX_BPF_FUNC_REG_ARGS' in the new check is therefore always false
> at this point.
>
> Looking at the commit message's statement "Reject BPF functions and kfuncs
> with more than 5 parameters at verification time if the architecture does not
> support stack arguments," it appears the intention was to conditionally reject
> based on JIT support. However, for kfuncs, the unconditional rejection
> happens first.
>
> The btf_prepare_func_args() change does work as intended because its earlier
> guards allow nargs up to MAX_BPF_FUNC_ARGS for certain subprogs before
> reaching the stack-args check.
>
> A later commit in this series (73a990fcc429) raises the first check in
> check_kfunc_args() from MAX_BPF_FUNC_REG_ARGS to MAX_BPF_FUNC_ARGS, which
> would make the new branch reachable. Should this hunk be combined with that
> change to avoid introducing unreachable code?

You are right. The above check does not really work as it is a noop.
And it really works until "bpf: Support stack arguments for kfunc calls".
Therefore, as you mentioned, it should be combined with
"bpf: Support stack arguments for kfunc calls".

Will do with next revision.

>
>
> ---
> AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
> See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
>
> CI run summary: https://github.com/kernel-patches/bpf/actions/runs/24902767240