On Wed, Jan 3, 2024 at 11:55 AM Yonghong Song <yonghong.song@xxxxxxxxx> wrote: > > > On 1/2/24 6:54 PM, Menglong Dong wrote: > > On Wed, Jan 3, 2024 at 8:52 AM Martin KaFai Lau <martin.lau@xxxxxxxxx> wrote: > >> On 1/2/24 10:11 AM, Stanislav Fomichev wrote: > >>> On 12/29, Menglong Dong wrote: > >>>> For now, we have to call some helpers when we need to update the csum, > >>>> such as bpf_l4_csum_replace, bpf_l3_csum_replace, etc. These helpers are > >>>> not inlined, which causes poor performance. > >>>> > >>>> In fact, we can define our own csum update functions in BPF program > >>>> instead of bpf_l3_csum_replace, which is totally inlined and efficient. > >>>> However, we can't do this for bpf_l4_csum_replace for now, as we can't > >>>> update skb->csum, which can cause skb->csum invalid in the rx path with > >>>> CHECKSUM_COMPLETE mode. > >>>> > >>>> What's more, we can't use the direct data access and have to use > >>>> skb_store_bytes() with the BPF_F_RECOMPUTE_CSUM flag in some case, such > >>>> as modifing the vni in the vxlan header and the underlay udp header has > >>>> no checksum. > >> There is bpf_csum_update(), does it work? > >> A helper call should be acceptable comparing with the csum calculation itself. > > Yeah, this helper works in this case! Now we miss the last > > piece for the tx path: ip_summed. We need to know if it is > > CHECKSUM_PARTIAL to decide if we should update the > > csum in the packet. In the tx path, the csum in the L4 is the > > pseudo header only if skb->ip_summed is CHECKSUM_PARTIAL. > > > > Maybe we can introduce a lightweight kfunc to get its > > value? Such as bpf_skb_csum_mode(). As we need only call > > it once, there shouldn't be overhead on it. > > You don't need kfunc, you can do checking like > struct sk_buff *kskb = bpf_cast_to_kern_ctx(skb); > if (kskb->ip_summed == CHECKSUM_PARTIAL) ... > ... > Great, this is exactly what I need! Thanks~
