[nf-next,v6,2/8] netfilter:nf_flow_table_core: Separate inet operation to single function
diff mbox series

Message ID 1564053176-28605-3-git-send-email-wenxu@ucloud.cn
State Under Review
Delegated to: Pablo Neira
Headers show
Series
  • netfilter: nf_flow_offload: support bridge family offload for both ipv4 and ipv6
Related show

Commit Message

wenxu July 25, 2019, 11:12 a.m. UTC
From: wenxu <wenxu@ucloud.cn>

This patch separate the inet family operation in nf_flow_table_core to
single function. Prepare for support the bridge family.

Signed-off-by: wenxu <wenxu@ucloud.cn>
---
v6: no change

 net/netfilter/nf_flow_table_core.c | 52 ++++++++++++++++++++++++++++----------
 1 file changed, 38 insertions(+), 14 deletions(-)

Patch
diff mbox series

diff --git a/net/netfilter/nf_flow_table_core.c b/net/netfilter/nf_flow_table_core.c
index 7e0b5bd..2bec409 100644
--- a/net/netfilter/nf_flow_table_core.c
+++ b/net/netfilter/nf_flow_table_core.c
@@ -22,6 +22,20 @@  struct flow_offload_entry {
 static DEFINE_MUTEX(flowtable_lock);
 static LIST_HEAD(flowtables);
 
+static struct dst_entry *
+flow_offload_fill_inet_dst(struct flow_offload_tuple *ft,
+			   struct nf_flow_route *route,
+			   enum flow_offload_tuple_dir dir)
+{
+	struct dst_entry *other_dst = route->tuple[!dir].dst;
+	struct dst_entry *dst = route->tuple[dir].dst;
+
+	ft->iifidx = other_dst->dev->ifindex;
+	ft->dst.dst_cache = dst;
+
+	return dst;
+}
+
 static void
 flow_offload_fill_dir(struct flow_offload *flow, struct nf_conn *ct,
 		      struct nf_flow_dst *flow_dst,
@@ -29,9 +43,9 @@  struct flow_offload_entry {
 {
 	struct flow_offload_tuple *ft = &flow->tuplehash[dir].tuple;
 	struct nf_conntrack_tuple *ctt = &ct->tuplehash[dir].tuple;
-	struct dst_entry *other_dst = flow_dst->route.tuple[!dir].dst;
-	struct dst_entry *dst = flow_dst->route.tuple[dir].dst;
+	struct dst_entry *dst;
 
+	dst = flow_offload_fill_inet_dst(ft, &flow_dst->route, dir);
 	ft->dir = dir;
 
 	switch (ctt->src.l3num) {
@@ -51,9 +65,19 @@  struct flow_offload_entry {
 	ft->l4proto = ctt->dst.protonum;
 	ft->src_port = ctt->src.u.tcp.port;
 	ft->dst_port = ctt->dst.u.tcp.port;
+}
 
-	ft->iifidx = other_dst->dev->ifindex;
-	ft->dst_cache = dst;
+static int flow_offload_dst_hold(struct nf_flow_dst *flow_dst)
+{
+	if (!dst_hold_safe(flow_dst->route.tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst))
+		return -1;
+
+	if (!dst_hold_safe(flow_dst->route.tuple[FLOW_OFFLOAD_DIR_REPLY].dst)) {
+		dst_release(flow_dst->route.tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst);
+		return -1;
+	}
+
+	return 0;
 }
 
 struct flow_offload *
@@ -72,11 +96,8 @@  struct flow_offload *
 
 	flow = &entry->flow;
 
-	if (!dst_hold_safe(flow_dst->route.tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst))
-		goto err_dst_cache_original;
-
-	if (!dst_hold_safe(flow_dst->route.tuple[FLOW_OFFLOAD_DIR_REPLY].dst))
-		goto err_dst_cache_reply;
+	if (flow_offload_dst_hold(flow_dst))
+		goto err_dst_cache;
 
 	entry->ct = ct;
 
@@ -90,9 +111,7 @@  struct flow_offload *
 
 	return flow;
 
-err_dst_cache_reply:
-	dst_release(flow_dst->route.tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst);
-err_dst_cache_original:
+err_dst_cache:
 	kfree(entry);
 err_ct_refcnt:
 	nf_ct_put(ct);
@@ -135,12 +154,17 @@  static void flow_offload_fixup_ct_state(struct nf_conn *ct)
 	ct->timeout = nfct_time_stamp + timeout;
 }
 
+static void flow_offload_dst_release(struct flow_offload *flow)
+{
+	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst.dst_cache);
+	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst.dst_cache);
+}
+
 void flow_offload_free(struct flow_offload *flow)
 {
 	struct flow_offload_entry *e;
 
-	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst.dst_cache);
-	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst.dst_cache);
+	flow_offload_dst_release(flow);
 	e = container_of(flow, struct flow_offload_entry, flow);
 	if (flow->flags & FLOW_OFFLOAD_DYING)
 		nf_ct_delete(e->ct, 0, 0);