On Fri, Jan 6, 2023 at 2:55 PM <sdf@xxxxxxxxxx> wrote:
>
> On 01/06, Ziyang Xuan wrote:
> > Add ipip6 and ip6ip decap support for bpf_skb_adjust_room().
> > Main use case is for using cls_bpf on ingress hook to decapsulate
> > IPv4 over IPv6 and IPv6 over IPv4 tunnel packets.
>
> CC'd Willem since he has done bpf_skb_adjust_room changes in the past.
> There might be a lot of GRO/GSO context I'm missing.
>
> > Signed-off-by: Ziyang Xuan <william.xuanziyang@xxxxxxxxxx>
> > ---
> > net/core/filter.c | 34 ++++++++++++++++++++++++++++++++--
> > 1 file changed, 32 insertions(+), 2 deletions(-)
>
> > diff --git a/net/core/filter.c b/net/core/filter.c
> > index 929358677183..73982fb4fe2e 100644
> > --- a/net/core/filter.c
> > +++ b/net/core/filter.c
> > @@ -3495,6 +3495,12 @@ static int bpf_skb_net_grow(struct sk_buff *skb,
> > u32 off, u32 len_diff,
> > static int bpf_skb_net_shrink(struct sk_buff *skb, u32 off, u32 len_diff,
> > u64 flags)
> > {
> > + union {
> > + struct iphdr *v4;
> > + struct ipv6hdr *v6;
> > + unsigned char *hdr;
> > + } ip;
> > + __be16 proto;
> > int ret;
>
> > if (unlikely(flags & ~(BPF_F_ADJ_ROOM_FIXED_GSO |
> > @@ -3512,10 +3518,19 @@ static int bpf_skb_net_shrink(struct sk_buff
> > *skb, u32 off, u32 len_diff,
> > if (unlikely(ret < 0))
> > return ret;
>
> > + ip.hdr = skb_inner_network_header(skb);
> > + if (ip.v4->version == 4)
> > + proto = htons(ETH_P_IP);
> > + else
> > + proto = htons(ETH_P_IPV6);
> > +
> > ret = bpf_skb_net_hdr_pop(skb, off, len_diff);
> > if (unlikely(ret < 0))
> > return ret;
>
> > + /* Match skb->protocol to new outer l3 protocol */
> > + skb->protocol = proto;
> > +
> > if (skb_is_gso(skb)) {
> > struct skb_shared_info *shinfo = skb_shinfo(skb);
>
> > @@ -3578,10 +3593,14 @@ BPF_CALL_4(bpf_skb_adjust_room, struct sk_buff *,
> > skb, s32, len_diff,
> > u32, mode, u64, flags)
> > {
> > u32 len_cur, len_diff_abs = abs(len_diff);
> > - u32 len_min = bpf_skb_net_base_len(skb);
> > - u32 len_max = BPF_SKB_MAX_LEN;
> > + u32 len_min, len_max = BPF_SKB_MAX_LEN;
> > __be16 proto = skb->protocol;
> > bool shrink = len_diff < 0;
> > + union {
> > + struct iphdr *v4;
> > + struct ipv6hdr *v6;
> > + unsigned char *hdr;
> > + } ip;
> > u32 off;
> > int ret;
>
> > @@ -3594,6 +3613,9 @@ BPF_CALL_4(bpf_skb_adjust_room, struct sk_buff *,
> > skb, s32, len_diff,
> > proto != htons(ETH_P_IPV6)))
> > return -ENOTSUPP;
>
> > + if (unlikely(shrink && !skb->encapsulation))
> > + return -ENOTSUPP;
> > +
This new restriction might break existing users.
There is no pre-existing requirement that shrink is used solely with
packets encapsulated by the protocol stack.
Indeed, skb->encapsulation is likely not set on packets arriving from
the wire, even if encapsulated. Referring to your comment "Main use
case is for using cls_bpf on ingress hook to decapsulate"
Can a combination of the existing bpf_skb_adjust_room and
bpf_skb_change_proto address your problem?
> > off = skb_mac_header_len(skb);
> > switch (mode) {
> > case BPF_ADJ_ROOM_NET:
> > @@ -3605,6 +3627,14 @@ BPF_CALL_4(bpf_skb_adjust_room, struct sk_buff *,
> > skb, s32, len_diff,
> > return -ENOTSUPP;
> > }
>
> > + if (shrink) {
> > + ip.hdr = skb_inner_network_header(skb);
> > + if (ip.v4->version == 4)
> > + len_min = sizeof(struct iphdr);
> > + else
> > + len_min = sizeof(struct ipv6hdr);
> > + }
> > +
> > len_cur = skb->len - skb_network_offset(skb);
> > if ((shrink && (len_diff_abs >= len_cur ||
> > len_cur - len_diff_abs < len_min)) ||
> > --
> > 2.25.1
>
