On Mon, Mar 29, 2021 at 1:10 PM John Fastabend <john.fastabend@xxxxxxxxx> wrote:
>
> Cong Wang wrote:
> > From: Cong Wang <cong.wang@xxxxxxxxxxxxx>
> >
> > Reusing BPF_SK_SKB_STREAM_VERDICT is possible but its name is
> > confusing and more importantly we still want to distinguish them
> > from user-space. So we can just reuse the stream verdict code but
> > introduce a new type of eBPF program, skb_verdict. Users are not
> > allowed to set stream_verdict and skb_verdict at the same time.
> >
> > Cc: John Fastabend <john.fastabend@xxxxxxxxx>
> > Cc: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
> > Cc: Jakub Sitnicki <jakub@xxxxxxxxxxxxxx>
> > Cc: Lorenz Bauer <lmb@xxxxxxxxxxxxxx>
> > Signed-off-by: Cong Wang <cong.wang@xxxxxxxxxxxxx>
> > ---
>
> [...]
>
> > diff --git a/net/core/skmsg.c b/net/core/skmsg.c
> > index 656eceab73bc..a045812d7c78 100644
> > --- a/net/core/skmsg.c
> > +++ b/net/core/skmsg.c
> > @@ -697,7 +697,7 @@ void sk_psock_drop(struct sock *sk, struct sk_psock *psock)
> > rcu_assign_sk_user_data(sk, NULL);
> > if (psock->progs.stream_parser)
> > sk_psock_stop_strp(sk, psock);
> > - else if (psock->progs.stream_verdict)
> > + else if (psock->progs.stream_verdict || psock->progs.skb_verdict)
> > sk_psock_stop_verdict(sk, psock);
> > write_unlock_bh(&sk->sk_callback_lock);
> >
> > @@ -1024,6 +1024,8 @@ static int sk_psock_verdict_recv(read_descriptor_t *desc, struct sk_buff *skb,
> > }
> > skb_set_owner_r(skb, sk);
> > prog = READ_ONCE(psock->progs.stream_verdict);
> > + if (!prog)
> > + prog = READ_ONCE(psock->progs.skb_verdict);
>
> Trying to think through this case. User attachs skb_verdict program
> to map, then updates map with a bunch of TCP sockets. The above
> code will run the skb_verdict program with the TCP socket as far as
> I can tell.
>
> This is OK because there really is no difference, other than by name,
> between a skb_verdict and a stream_verdict program? Do we want something
> to block adding TCP sockets to maps with stream_verdict programs? It
> feels a bit odd in its current state to me.
Yes, it should work too. skb_verdict only extends stream_verdict beyond
TCP, it does not prohibit TCP.
>
> > if (likely(prog)) {
> > skb_dst_drop(skb);
> > skb_bpf_redirect_clear(skb);
> > diff --git a/net/core/sock_map.c b/net/core/sock_map.c
> > index e564fdeaada1..c46709786a49 100644
> > --- a/net/core/sock_map.c
> > +++ b/net/core/sock_map.c
> > @@ -155,6 +155,8 @@ static void sock_map_del_link(struct sock *sk,
> > strp_stop = true;
> > if (psock->saved_data_ready && stab->progs.stream_verdict)
> > verdict_stop = true;
> > + if (psock->saved_data_ready && stab->progs.skb_verdict)
> > + verdict_stop = true;
> > list_del(&link->list);
> > sk_psock_free_link(link);
> > }
> > @@ -227,7 +229,7 @@ static struct sk_psock *sock_map_psock_get_checked(struct sock *sk)
> > static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
> > struct sock *sk)
> > {
> > - struct bpf_prog *msg_parser, *stream_parser, *stream_verdict;
> > + struct bpf_prog *msg_parser, *stream_parser, *stream_verdict, *skb_verdict;
> > struct sk_psock *psock;
> > int ret;
> >
> > @@ -256,6 +258,15 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
> > }
> > }
> >
> > + skb_verdict = READ_ONCE(progs->skb_verdict);
> > + if (skb_verdict) {
> > + skb_verdict = bpf_prog_inc_not_zero(skb_verdict);
> > + if (IS_ERR(skb_verdict)) {
> > + ret = PTR_ERR(skb_verdict);
> > + goto out_put_msg_parser;
> > + }
> > + }
> > +
> > psock = sock_map_psock_get_checked(sk);
> > if (IS_ERR(psock)) {
> > ret = PTR_ERR(psock);
> > @@ -265,6 +276,7 @@ static int sock_map_link(struct bpf_map *map, struct sk_psock_progs *progs,
> > if (psock) {
> > if ((msg_parser && READ_ONCE(psock->progs.msg_parser)) ||
> > (stream_parser && READ_ONCE(psock->progs.stream_parser)) ||
> > + (skb_verdict && READ_ONCE(psock->progs.skb_verdict)) ||
> > (stream_verdict && READ_ONCE(psock->progs.stream_verdict))) {
> > sk_psock_put(sk, psock);
> > ret = -EBUSY;
>
> Do we need another test here,
>
> (skb_verdict && READ_ONCE(psock->progs.stream_verdict)
>
> this way we return EBUSY and avoid having both stream_verdict and
> skb_verdict attached on the same map?
Yes, good catch, we do need a check here. And I will see if I can add a small
test case for this too.
Thanks.
