On Fri, Feb 02, 2024 at 05:18:48PM +0800, Geliang Tang wrote:
> From: Geliang Tang <tanggeliang@kylinos.cn>
>
> Similar to the handling in the functions __register_btf_kfunc_id_set() and
> register_btf_id_dtor_kfuncs(), this patch adds CONFIG_DEBUG_INFO_BTF and
> CONFIG_DEBUG_INFO_BTF_MODULES checks for __register_bpf_struct_ops() on
> error path too when btf_get_module_btf() returns NULL.
>
> Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
> ---
> kernel/bpf/btf.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c
> index ef380e546952..381676add335 100644
> --- a/kernel/bpf/btf.c
> +++ b/kernel/bpf/btf.c
> @@ -8880,8 +8880,15 @@ int __register_bpf_struct_ops(struct bpf_struct_ops *st_ops)
> int err = 0;
>
> btf = btf_get_module_btf(st_ops->owner);
> - if (!btf)
> - return -EINVAL;
> + if (!btf) {
> + if (!st_ops->owner && IS_ENABLED(CONFIG_DEBUG_INFO_BTF)) {
> + pr_err("missing vmlinux BTF, cannot register structs\n");
> + return -EINVAL;
> + }
> + if (st_ops->owner && IS_ENABLED(CONFIG_DEBUG_INFO_BTF_MODULES))
> + pr_warn("missing module BTF, cannot register structs\n");
> + return 0;
given that we have the same code in 2 other functions
would it make sense to add function for that?
jirka
> + }
>
> log = kzalloc(sizeof(*log), GFP_KERNEL | __GFP_NOWARN);
> if (!log) {
> --
> 2.40.1
>