[PATCH net 1/6] net: Introduce skb ttl field to track packet loops

[Jamal Hadi Salim <jhs@mojatatu.com>]

In order to keep track of loops across the stack, in particular when going
across from egress->ingress and back,we need to _remember the global loop
state in the skb_.
We introduce a per-skb ttl field to keep track of this state.

This patch liberates two bits:
1) The bit "skb->from_ingress" is reclaimed for ttl. Since it is currently
only used for ifb, it is safe to move this to local-per-layer skb/tc state
on the qdisc_skb_cb struct.
2) A second bit that was available on the skb.

Use cases:
1) Mirred increments the ttl whenever it sees an skb. If the skb shows
up multiple times we catch it when it exceeds MIRRED_NEST_LIMIT iterations
of the loop.

2) netem increments when using the "duplicate" feature and catches it when
it sees the packet the second time.

Fixes: fe946a751d9b ("net/sched: act_mirred: add loop detection")
Fixes: 0afb51e72855 ("[PKT_SCHED]: netem: reinsert for duplication")
Tested-by: Victor Nogueira <victor@mojatatu.com>
Signed-off-by: Jamal Hadi Salim <jhs@mojatatu.com>
---
 drivers/net/ifb.c              |  2 +-
 include/linux/skbuff.h         | 24 ++----------------------
 include/net/sch_generic.h      | 22 ++++++++++++++++++++++
 net/netfilter/nft_fwd_netdev.c |  1 +
 4 files changed, 26 insertions(+), 23 deletions(-)

diff --git a/drivers/net/ifb.c b/drivers/net/ifb.c
index d3dc0914450a..137a20e4bf8c 100644
--- a/drivers/net/ifb.c
+++ b/drivers/net/ifb.c
@@ -124,7 +124,7 @@ static void ifb_ri_tasklet(struct tasklet_struct *t)
 		rcu_read_unlock();
 		skb->skb_iif = txp->dev->ifindex;
 
-		if (!skb->from_ingress) {
+		if (!qdisc_skb_cb(skb)->from_ingress) {
 			dev_queue_xmit(skb);
 		} else {
 			skb_pull_rcsum(skb, skb->mac_len);
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 86737076101d..7f18b0c28728 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -840,6 +840,7 @@ enum skb_tstamp_type {
  *	@no_fcs:  Request NIC to treat last 4 bytes as Ethernet FCS
  *	@encapsulation: indicates the inner headers in the skbuff are valid
  *	@encap_hdr_csum: software checksum is needed
+ *	@ttl: time to live count when a packet loops.
  *	@csum_valid: checksum is already valid
  *	@csum_not_inet: use CRC32c to resolve CHECKSUM_PARTIAL
  *	@csum_complete_sw: checksum was completed by software
@@ -1000,6 +1001,7 @@ struct sk_buff {
 	/* Indicates the inner headers are valid in the skbuff. */
 	__u8			encapsulation:1;
 	__u8			encap_hdr_csum:1;
+	__u8			ttl:2;
 	__u8			csum_valid:1;
 #ifdef CONFIG_IPV6_NDISC_NODETYPE
 	__u8			ndisc_nodetype:2;
@@ -1016,9 +1018,6 @@ struct sk_buff {
 	__u8			offload_l3_fwd_mark:1;
 #endif
 	__u8			redirected:1;
-#ifdef CONFIG_NET_REDIRECT
-	__u8			from_ingress:1;
-#endif
 #ifdef CONFIG_NETFILTER_SKIP_EGRESS
 	__u8			nf_skip_egress:1;
 #endif
@@ -5352,30 +5351,11 @@ static inline bool skb_is_redirected(const struct sk_buff *skb)
 	return skb->redirected;
 }
 
-static inline void skb_set_redirected(struct sk_buff *skb, bool from_ingress)
-{
-	skb->redirected = 1;
-#ifdef CONFIG_NET_REDIRECT
-	skb->from_ingress = from_ingress;
-	if (skb->from_ingress)
-		skb_clear_tstamp(skb);
-#endif
-}
-
 static inline void skb_reset_redirect(struct sk_buff *skb)
 {
 	skb->redirected = 0;
 }
 
-static inline void skb_set_redirected_noclear(struct sk_buff *skb,
-					      bool from_ingress)
-{
-	skb->redirected = 1;
-#ifdef CONFIG_NET_REDIRECT
-	skb->from_ingress = from_ingress;
-#endif
-}
-
 static inline bool skb_csum_is_sctp(struct sk_buff *skb)
 {
 #if IS_ENABLED(CONFIG_IP_SCTP)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index c3a7268b567e..42d8a1a9db4c 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -459,6 +459,9 @@ struct qdisc_skb_cb {
 	u8			post_ct:1;
 	u8			post_ct_snat:1;
 	u8			post_ct_dnat:1;
+#ifdef CONFIG_NET_REDIRECT
+	u8			from_ingress:1;
+#endif
 };
 
 typedef void tcf_chain_head_change_t(struct tcf_proto *tp_head, void *priv);
@@ -1140,6 +1143,25 @@ static inline void qdisc_dequeue_drop(struct Qdisc *q, struct sk_buff *skb,
 	q->to_free = skb;
 }
 
+static inline void skb_set_redirected(struct sk_buff *skb, bool from_ingress)
+{
+	skb->redirected = 1;
+#ifdef CONFIG_NET_REDIRECT
+	qdisc_skb_cb(skb)->from_ingress = from_ingress;
+	if (qdisc_skb_cb(skb)->from_ingress)
+		skb_clear_tstamp(skb);
+#endif
+}
+
+static inline void skb_set_redirected_noclear(struct sk_buff *skb,
+					      bool from_ingress)
+{
+	skb->redirected = 1;
+#ifdef CONFIG_NET_REDIRECT
+	qdisc_skb_cb(skb)->from_ingress = from_ingress;
+#endif
+}
+
 /* Instead of calling kfree_skb() while root qdisc lock is held,
  * queue the skb for future freeing at end of __dev_xmit_skb()
  */
diff --git a/net/netfilter/nft_fwd_netdev.c b/net/netfilter/nft_fwd_netdev.c
index 152a9fb4d23a..d62c856ef96a 100644
--- a/net/netfilter/nft_fwd_netdev.c
+++ b/net/netfilter/nft_fwd_netdev.c
@@ -16,6 +16,7 @@
 #include <net/netfilter/nf_dup_netdev.h>
 #include <net/neighbour.h>
 #include <net/ip.h>
+#include <net/sch_generic.h>
 
 struct nft_fwd_netdev {
 	u8	sreg_dev;
-- 
2.34.1