Re: [PATCH bpf-next v2 2/4] netfilter: add bpf_xdp_flow_offload_lookup kfunc

[Lorenzo Bianconi <lorenzo@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1714 bytes --]

> On Sat, May 18, 2024 at 3:13 AM Lorenzo Bianconi <lorenzo@kernel.org> wrote:

[...]
> 
> I think it needs to be KF_RET_NULL.
> And most likely KF_TRUSTED_ARGS as well.

ack, I will fix it in v2.

> 
> Also the "offload" doesn't fit in the name.
> The existing code calls it "offload", because it's actually
> pushing the rules to HW (if I understand the code),
> but here it's just a lookup from xdp.
> So call it
> bpf_xdp_flow_lookup() ?

ack fine, I do not have a strong opinion on it. I will fix it in v2.

> 
> Though "flow" is a bit too generic here.
> nf_flow maybe?

ack, I will fix it in v2.

Regards,
Lorenzo

> 
> > +BTF_KFUNCS_END(nf_ft_kfunc_set)
> > +
> > +static const struct btf_kfunc_id_set nf_flow_offload_kfunc_set = {
> > +       .owner = THIS_MODULE,
> > +       .set   = &nf_ft_kfunc_set,
> > +};
> > +
> > +int nf_flow_offload_register_bpf(void)
> > +{
> > +       return register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP,
> > +                                        &nf_flow_offload_kfunc_set);
> > +}
> > +EXPORT_SYMBOL_GPL(nf_flow_offload_register_bpf);
> > diff --git a/net/netfilter/nf_flow_table_inet.c b/net/netfilter/nf_flow_table_inet.c
> > index 6eef15648b7b0..6175f7556919d 100644
> > --- a/net/netfilter/nf_flow_table_inet.c
> > +++ b/net/netfilter/nf_flow_table_inet.c
> > @@ -98,7 +98,7 @@ static int __init nf_flow_inet_module_init(void)
> >         nft_register_flowtable_type(&flowtable_ipv6);
> >         nft_register_flowtable_type(&flowtable_inet);
> >
> > -       return 0;
> > +       return nf_flow_offload_register_bpf();
> >  }
> >
> >  static void __exit nf_flow_inet_module_exit(void)
> > --
> > 2.45.1
> >

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]