Re: [PATCH bpf-next v3 10/15] bpf: Allow all struct_ops to use bpf_dynptr_from_skb()

["Emil Tsalapatis" <emil@etsalapatis.com>]

On Mon Jul 6, 2026 at 1:19 PM EDT, Amery Hung wrote:
> bpf_dynptr_from_skb() was only made available to bpf_qdisc, so far the
> only struct_ops type that needs to read an skb. The upcoming bpf_tcp_ops
> header-option hooks (parse_hdr/write_hdr_opt) also want to access the TCP
> options of an skb through a dynptr.
>
> All struct_ops programs share BPF_PROG_TYPE_STRUCT_OPS, so register
> bpf_kfunc_set_skb (which holds bpf_dynptr_from_skb) for that program type
> once, instead of per struct_ops. This makes bpf_dynptr_from_skb()
> available to bpf_tcp_ops and any future struct_ops.
>
> With the kfunc now provided to all of struct_ops, the bpf_qdisc-specific
> registration becomes redundant and is dropped: bpf_qdisc_kfunc_filter()
> only constrains kfuncs listed in qdisc_kfunc_ids, so removing
> bpf_dynptr_from_skb from that set (and from qdisc_common_kfunc_set) lets
> it fall through the filter unchanged, and bpf_qdisc keeps access via the
> generic struct_ops registration.
>
> Widening the registration is safe: a struct_ops that does not receive an
> skb in its context has nothing to pass to the helper.
>
> Signed-off-by: Amery Hung <ameryhung@gmail.com>

Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>

> ---
>  net/core/filter.c     | 1 +
>  net/sched/bpf_qdisc.c | 2 --
>  2 files changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 4f5cbcac3e78..05f492c215e0 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -12635,6 +12635,7 @@ static int __init bpf_kfunc_init(void)
>  	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_LWT_SEG6LOCAL, &bpf_kfunc_set_skb);
>  	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_NETFILTER, &bpf_kfunc_set_skb);
>  	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_TRACING, &bpf_kfunc_set_skb);
> +	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_STRUCT_OPS, &bpf_kfunc_set_skb);
>  	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_SCHED_CLS, &bpf_kfunc_set_skb_meta);
>  	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_SCHED_ACT, &bpf_kfunc_set_skb_meta);
>  	ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP, &bpf_kfunc_set_xdp);
> diff --git a/net/sched/bpf_qdisc.c b/net/sched/bpf_qdisc.c
> index 098ca02aed89..5691c13781a8 100644
> --- a/net/sched/bpf_qdisc.c
> +++ b/net/sched/bpf_qdisc.c
> @@ -280,7 +280,6 @@ BTF_KFUNCS_START(qdisc_kfunc_ids)
>  BTF_ID_FLAGS(func, bpf_skb_get_hash)
>  BTF_ID_FLAGS(func, bpf_kfree_skb, KF_RELEASE)
>  BTF_ID_FLAGS(func, bpf_qdisc_skb_drop, KF_RELEASE)
> -BTF_ID_FLAGS(func, bpf_dynptr_from_skb)
>  BTF_ID_FLAGS(func, bpf_qdisc_watchdog_schedule)
>  BTF_ID_FLAGS(func, bpf_qdisc_init_prologue)
>  BTF_ID_FLAGS(func, bpf_qdisc_reset_destroy_epilogue)
> @@ -290,7 +289,6 @@ BTF_KFUNCS_END(qdisc_kfunc_ids)
>  BTF_SET_START(qdisc_common_kfunc_set)
>  BTF_ID(func, bpf_skb_get_hash)
>  BTF_ID(func, bpf_kfree_skb)
> -BTF_ID(func, bpf_dynptr_from_skb)
>  BTF_SET_END(qdisc_common_kfunc_set)
>  
>  BTF_SET_START(qdisc_enqueue_kfunc_set)