net: netfilter: Make ct zone opts configurable for bpf ct helpers

Add ct zone id and direction to bpf_ct_opts so that arbitrary ct zones
can be used for xdp/tc bpf ct helper functions bpf_{xdp,skb}_ct_alloc
and bpf_{xdp,skb}_ct_lookup.

Signed-off-by: Brad Cowie <brad@faucet.nz>
Link: https://lore.kernel.org/r/20240522050712.732558-1-brad@faucet.nz
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>

net/netfilter/nf_conntrack_bpf.c

@@ -32,7 +32,9 @@
  *		   -EINVAL - Passed NULL for bpf_tuple pointer
  *		   -EINVAL - opts->reserved is not 0
  *		   -EINVAL - netns_id is less than -1
- *		   -EINVAL - opts__sz isn't NF_BPF_CT_OPTS_SZ (12)
+ *		   -EINVAL - opts__sz isn't NF_BPF_CT_OPTS_SZ (16) or 12
+ *		   -EINVAL - opts->ct_zone_id set when
+			     opts__sz isn't NF_BPF_CT_OPTS_SZ (16)
  *		   -EPROTO - l4proto isn't one of IPPROTO_TCP or IPPROTO_UDP
  *		   -ENONET - No network namespace found for netns_id
  *		   -ENOENT - Conntrack lookup could not find entry for tuple
@@ -42,6 +44,8 @@
  *		 Values:
  *		   IPPROTO_TCP, IPPROTO_UDP
  * @dir:       - connection tracking tuple direction.
+ * @ct_zone_id - connection tracking zone id.
+ * @ct_zone_dir - connection tracking zone direction.
  * @reserved   - Reserved member, will be reused for more options in future
  *		 Values:
  *		   0
@@ -51,11 +55,13 @@ struct bpf_ct_opts {
 	s32 error;
 	u8 l4proto;
 	u8 dir;
-	u8 reserved[2];
+	u16 ct_zone_id;
+	u8 ct_zone_dir;
+	u8 reserved[3];
 };
 
 enum {
-	NF_BPF_CT_OPTS_SZ = 12,
+	NF_BPF_CT_OPTS_SZ = 16,
 };
 
 static int bpf_nf_ct_tuple_parse(struct bpf_sock_tuple *bpf_tuple,
@@ -104,12 +110,21 @@ __bpf_nf_ct_alloc_entry(struct net *net, struct bpf_sock_tuple *bpf_tuple,
 			u32 timeout)
 {
 	struct nf_conntrack_tuple otuple, rtuple;
+	struct nf_conntrack_zone ct_zone;
 	struct nf_conn *ct;
 	int err;
 
-	if (!opts || !bpf_tuple || opts->reserved[0] || opts->reserved[1] ||
+	if (!opts || !bpf_tuple)
-	    opts_len != NF_BPF_CT_OPTS_SZ)
 		return ERR_PTR(-EINVAL);
+	if (!(opts_len == NF_BPF_CT_OPTS_SZ || opts_len == 12))
+		return ERR_PTR(-EINVAL);
+	if (opts_len == NF_BPF_CT_OPTS_SZ) {
+		if (opts->reserved[0] || opts->reserved[1] || opts->reserved[2])
+			return ERR_PTR(-EINVAL);
+	} else {
+		if (opts->ct_zone_id)
+			return ERR_PTR(-EINVAL);
+	}
 
 	if (unlikely(opts->netns_id < BPF_F_CURRENT_NETNS))
 		return ERR_PTR(-EINVAL);
@@ -130,7 +145,16 @@ __bpf_nf_ct_alloc_entry(struct net *net, struct bpf_sock_tuple *bpf_tuple,
 			return ERR_PTR(-ENONET);
 	}
 
-	ct = nf_conntrack_alloc(net, &nf_ct_zone_dflt, &otuple, &rtuple,
+	if (opts_len == NF_BPF_CT_OPTS_SZ) {
+		if (opts->ct_zone_dir == 0)
+			opts->ct_zone_dir = NF_CT_DEFAULT_ZONE_DIR;
+		nf_ct_zone_init(&ct_zone,
+				opts->ct_zone_id, opts->ct_zone_dir, 0);
+	} else {
+		ct_zone = nf_ct_zone_dflt;
+	}
+
+	ct = nf_conntrack_alloc(net, &ct_zone, &otuple, &rtuple,
 				GFP_ATOMIC);
 	if (IS_ERR(ct))
 		goto out;
@@ -152,12 +176,21 @@ static struct nf_conn *__bpf_nf_ct_lookup(struct net *net,
 {
 	struct nf_conntrack_tuple_hash *hash;
 	struct nf_conntrack_tuple tuple;
+	struct nf_conntrack_zone ct_zone;
 	struct nf_conn *ct;
 	int err;
 
-	if (!opts || !bpf_tuple || opts->reserved[0] || opts->reserved[1] ||
+	if (!opts || !bpf_tuple)
-	    opts_len != NF_BPF_CT_OPTS_SZ)
 		return ERR_PTR(-EINVAL);
+	if (!(opts_len == NF_BPF_CT_OPTS_SZ || opts_len == 12))
+		return ERR_PTR(-EINVAL);
+	if (opts_len == NF_BPF_CT_OPTS_SZ) {
+		if (opts->reserved[0] || opts->reserved[1] || opts->reserved[2])
+			return ERR_PTR(-EINVAL);
+	} else {
+		if (opts->ct_zone_id)
+			return ERR_PTR(-EINVAL);
+	}
 	if (unlikely(opts->l4proto != IPPROTO_TCP && opts->l4proto != IPPROTO_UDP))
 		return ERR_PTR(-EPROTO);
 	if (unlikely(opts->netns_id < BPF_F_CURRENT_NETNS))
@@ -174,7 +207,16 @@ static struct nf_conn *__bpf_nf_ct_lookup(struct net *net,
 			return ERR_PTR(-ENONET);
 	}
 
-	hash = nf_conntrack_find_get(net, &nf_ct_zone_dflt, &tuple);
+	if (opts_len == NF_BPF_CT_OPTS_SZ) {
+		if (opts->ct_zone_dir == 0)
+			opts->ct_zone_dir = NF_CT_DEFAULT_ZONE_DIR;
+		nf_ct_zone_init(&ct_zone,
+				opts->ct_zone_id, opts->ct_zone_dir, 0);
+	} else {
+		ct_zone = nf_ct_zone_dflt;
+	}
+
+	hash = nf_conntrack_find_get(net, &ct_zone, &tuple);
 	if (opts->netns_id >= 0)
 		put_net(net);
 	if (!hash)
@@ -245,7 +287,7 @@ __bpf_kfunc_start_defs();
  * @opts	- Additional options for allocation (documented above)
  *		    Cannot be NULL
  * @opts__sz	- Length of the bpf_ct_opts structure
- *		    Must be NF_BPF_CT_OPTS_SZ (12)
+ *		    Must be NF_BPF_CT_OPTS_SZ (16) or 12
  */
 __bpf_kfunc struct nf_conn___init *
 bpf_xdp_ct_alloc(struct xdp_md *xdp_ctx, struct bpf_sock_tuple *bpf_tuple,
@@ -279,7 +321,7 @@ bpf_xdp_ct_alloc(struct xdp_md *xdp_ctx, struct bpf_sock_tuple *bpf_tuple,
  * @opts	- Additional options for lookup (documented above)
  *		    Cannot be NULL
  * @opts__sz	- Length of the bpf_ct_opts structure
- *		    Must be NF_BPF_CT_OPTS_SZ (12)
+ *		    Must be NF_BPF_CT_OPTS_SZ (16) or 12
  */
 __bpf_kfunc struct nf_conn *
 bpf_xdp_ct_lookup(struct xdp_md *xdp_ctx, struct bpf_sock_tuple *bpf_tuple,
@@ -312,7 +354,7 @@ bpf_xdp_ct_lookup(struct xdp_md *xdp_ctx, struct bpf_sock_tuple *bpf_tuple,
  * @opts	- Additional options for allocation (documented above)
  *		    Cannot be NULL
  * @opts__sz	- Length of the bpf_ct_opts structure
- *		    Must be NF_BPF_CT_OPTS_SZ (12)
+ *		    Must be NF_BPF_CT_OPTS_SZ (16) or 12
  */
 __bpf_kfunc struct nf_conn___init *
 bpf_skb_ct_alloc(struct __sk_buff *skb_ctx, struct bpf_sock_tuple *bpf_tuple,
@@ -347,7 +389,7 @@ bpf_skb_ct_alloc(struct __sk_buff *skb_ctx, struct bpf_sock_tuple *bpf_tuple,
  * @opts	- Additional options for lookup (documented above)
  *		    Cannot be NULL
  * @opts__sz	- Length of the bpf_ct_opts structure
- *		    Must be NF_BPF_CT_OPTS_SZ (12)
+ *		    Must be NF_BPF_CT_OPTS_SZ (16) or 12
  */
 __bpf_kfunc struct nf_conn *
 bpf_skb_ct_lookup(struct __sk_buff *skb_ctx, struct bpf_sock_tuple *bpf_tuple,