Re: [PATCH bpf-next v1 1/2] resolve_btfids: Emit arena attributes from kfunc parameter suffixes

> diff --git a/tools/bpf/resolve_btfids/main.c b/tools/bpf/resolve_btfids/main.c
> index d2e4176339da..3c88ea192f7b 100644
> --- a/tools/bpf/resolve_btfids/main.c
> +++ b/tools/bpf/resolve_btfids/main.c

[ ... ]

> @@ -1067,6 +1070,22 @@ static int collect_decl_tags(struct btf2btf_context *ctx)
>  	return 0;
>  }
>
> +static bool param_name_has_suffix(const char *name, const char *suffix)
> +{
> +	size_t name_len = strlen(name);
> +	size_t suffix_len = strlen(suffix);
> +
> +	return name_len >= suffix_len && !strcmp(name + name_len - suffix_len, suffix);
> +}

Should this be `name_len > suffix_len` instead of `name_len >= suffix_len`?

The kernel's btf_param_match_suffix() in kernel/bpf/btf.c rejects a parameter
whose name is exactly equal to the suffix:

kernel/bpf/btf.c:btf_param_match_suffix() {
    ...
    len = strlen(param_name);
    if (len <= suffix_len)
        return false;
    param_name += len - suffix_len;
    return !strncmp(param_name, suffix, suffix_len);
}

That `len <= suffix_len` early return is what both is_kfunc_arg_arena() in
kernel/bpf/verifier.c and btf_func_model_flags() in kernel/bpf/btf.c rely on.
The same helper is also used by kernel/bpf/bpf_struct_ops.c for arena suffix
matching.

With the current implementation, a parameter named exactly "__arena" would be
treated as an arena pointer by resolve_btfids (causing address_space(1) to be
emitted in vmlinux.h), but would be treated as an ordinary kernel pointer by
the verifier's is_kfunc_arg_arena().

No in-tree kfunc currently uses a parameter named exactly "__arena" or
"__arena__nullable" (checked kernel/bpf/arena.c, test_kmods/bpf_testmod.c,
and btf_data.c), and the selftests added by commit 52b915d08490 only use
names like "a__arena", so this would be a latent inconsistency rather than a
currently-firing failure.

> +
> +static bool is_arena_param(const struct btf *btf, const struct btf_param *param)
> +{
> +	const char *name = btf__name_by_offset(btf, param->name_off);
> +
> +	return param_name_has_suffix(name, PARAM_SUFFIX_ARENA) ||
> +	       param_name_has_suffix(name, PARAM_SUFFIX_ARENA_NULLABLE);
> +}

[ ... ]


---
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/31305217088