[PATCH 4.19 018/133] sched: consistently handle layer3 header accesses in the presence of VLANs

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=UTF-8, Size: 12566 bytes --]

From: "Toke Høiland-Jørgensen" <toke@redhat.com>

[ Upstream commit d7bf2ebebc2bd61ab95e2a8e33541ef282f303d4 ]

There are a couple of places in net/sched/ that check skb->protocol and act
on the value there. However, in the presence of VLAN tags, the value stored
in skb->protocol can be inconsistent based on whether VLAN acceleration is
enabled. The commit quoted in the Fixes tag below fixed the users of
skb->protocol to use a helper that will always see the VLAN ethertype.

However, most of the callers don't actually handle the VLAN ethertype, but
expect to find the IP header type in the protocol field. This means that
things like changing the ECN field, or parsing diffserv values, stops
working if there's a VLAN tag, or if there are multiple nested VLAN
tags (QinQ).

To fix this, change the helper to take an argument that indicates whether
the caller wants to skip the VLAN tags or not. When skipping VLAN tags, we
make sure to skip all of them, so behaviour is consistent even in QinQ
mode.

To make the helper usable from the ECN code, move it to if_vlan.h instead
of pkt_sched.h.

v3:
- Remove empty lines
- Move vlan variable definitions inside loop in skb_protocol()
- Also use skb_protocol() helper in IP{,6}_ECN_decapsulate() and
  bpf_skb_ecn_set_ce()

v2:
- Use eth_type_vlan() helper in skb_protocol()
- Also fix code that reads skb->protocol directly
- Change a couple of 'if/else if' statements to switch constructs to avoid
  calling the helper twice

Reported-by: Ilya Ponetayev <i.ponetaev@ndmsystems.com>
Fixes: d8b9605d2697 ("net: sched: fix skb->protocol use in case of accelerated vlan path")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 include/linux/if_vlan.h  |   28 ++++++++++++++++++++++++++++
 include/net/inet_ecn.h   |   23 ++++++++++++++++-------
 include/net/pkt_sched.h  |   11 -----------
 net/core/filter.c        |    8 ++++----
 net/sched/act_connmark.c |    9 ++++++---
 net/sched/act_csum.c     |    2 +-
 net/sched/act_skbedit.c  |    2 +-
 net/sched/cls_api.c      |    2 +-
 net/sched/cls_flow.c     |    8 ++++----
 net/sched/cls_flower.c   |    2 +-
 net/sched/em_ipset.c     |    2 +-
 net/sched/em_meta.c      |    2 +-
 net/sched/sch_cake.c     |    4 ++--
 net/sched/sch_dsmark.c   |    6 +++---
 net/sched/sch_teql.c     |    2 +-
 15 files changed, 70 insertions(+), 41 deletions(-)

--- a/include/linux/if_vlan.h
+++ b/include/linux/if_vlan.h
@@ -313,6 +313,34 @@ static inline bool eth_type_vlan(__be16
 	}
 }
 
+/* A getter for the SKB protocol field which will handle VLAN tags consistently
+ * whether VLAN acceleration is enabled or not.
+ */
+static inline __be16 skb_protocol(const struct sk_buff *skb, bool skip_vlan)
+{
+	unsigned int offset = skb_mac_offset(skb) + sizeof(struct ethhdr);
+	__be16 proto = skb->protocol;
+
+	if (!skip_vlan)
+		/* VLAN acceleration strips the VLAN header from the skb and
+		 * moves it to skb->vlan_proto
+		 */
+		return skb_vlan_tag_present(skb) ? skb->vlan_proto : proto;
+
+	while (eth_type_vlan(proto)) {
+		struct vlan_hdr vhdr, *vh;
+
+		vh = skb_header_pointer(skb, offset, sizeof(vhdr), &vhdr);
+		if (!vh)
+			break;
+
+		proto = vh->h_vlan_encapsulated_proto;
+		offset += sizeof(vhdr);
+	}
+
+	return proto;
+}
+
 static inline bool vlan_hw_offload_capable(netdev_features_t features,
 					   __be16 proto)
 {
--- a/include/net/inet_ecn.h
+++ b/include/net/inet_ecn.h
@@ -4,6 +4,7 @@
 
 #include <linux/ip.h>
 #include <linux/skbuff.h>
+#include <linux/if_vlan.h>
 
 #include <net/inet_sock.h>
 #include <net/dsfield.h>
@@ -142,7 +143,7 @@ static inline void ipv6_copy_dscp(unsign
 
 static inline int INET_ECN_set_ce(struct sk_buff *skb)
 {
-	switch (skb->protocol) {
+	switch (skb_protocol(skb, true)) {
 	case cpu_to_be16(ETH_P_IP):
 		if (skb_network_header(skb) + sizeof(struct iphdr) <=
 		    skb_tail_pointer(skb))
@@ -209,12 +210,16 @@ static inline int IP_ECN_decapsulate(con
 {
 	__u8 inner;
 
-	if (skb->protocol == htons(ETH_P_IP))
+	switch (skb_protocol(skb, true)) {
+	case htons(ETH_P_IP):
 		inner = ip_hdr(skb)->tos;
-	else if (skb->protocol == htons(ETH_P_IPV6))
+		break;
+	case htons(ETH_P_IPV6):
 		inner = ipv6_get_dsfield(ipv6_hdr(skb));
-	else
+		break;
+	default:
 		return 0;
+	}
 
 	return INET_ECN_decapsulate(skb, oiph->tos, inner);
 }
@@ -224,12 +229,16 @@ static inline int IP6_ECN_decapsulate(co
 {
 	__u8 inner;
 
-	if (skb->protocol == htons(ETH_P_IP))
+	switch (skb_protocol(skb, true)) {
+	case htons(ETH_P_IP):
 		inner = ip_hdr(skb)->tos;
-	else if (skb->protocol == htons(ETH_P_IPV6))
+		break;
+	case htons(ETH_P_IPV6):
 		inner = ipv6_get_dsfield(ipv6_hdr(skb));
-	else
+		break;
+	default:
 		return 0;
+	}
 
 	return INET_ECN_decapsulate(skb, ipv6_get_dsfield(oipv6h), inner);
 }
--- a/include/net/pkt_sched.h
+++ b/include/net/pkt_sched.h
@@ -122,17 +122,6 @@ static inline void qdisc_run(struct Qdis
 	}
 }
 
-static inline __be16 tc_skb_protocol(const struct sk_buff *skb)
-{
-	/* We need to take extra care in case the skb came via
-	 * vlan accelerated path. In that case, use skb->vlan_proto
-	 * as the original vlan header was already stripped.
-	 */
-	if (skb_vlan_tag_present(skb))
-		return skb->vlan_proto;
-	return skb->protocol;
-}
-
 /* Calculate maximal size of packet seen by hard_start_xmit
    routine of this device.
  */
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -2695,7 +2695,7 @@ static int bpf_skb_proto_6_to_4(struct s
 
 static int bpf_skb_proto_xlat(struct sk_buff *skb, __be16 to_proto)
 {
-	__be16 from_proto = skb->protocol;
+	__be16 from_proto = skb_protocol(skb, true);
 
 	if (from_proto == htons(ETH_P_IP) &&
 	      to_proto == htons(ETH_P_IPV6))
@@ -2768,7 +2768,7 @@ static const struct bpf_func_proto bpf_s
 
 static u32 bpf_skb_net_base_len(const struct sk_buff *skb)
 {
-	switch (skb->protocol) {
+	switch (skb_protocol(skb, true)) {
 	case htons(ETH_P_IP):
 		return sizeof(struct iphdr);
 	case htons(ETH_P_IPV6):
@@ -2848,7 +2848,7 @@ static int bpf_skb_adjust_net(struct sk_
 	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(skb);
-	__be16 proto = skb->protocol;
+	__be16 proto = skb_protocol(skb, true);
 	bool shrink = len_diff < 0;
 	int ret;
 
@@ -4552,7 +4552,7 @@ static int bpf_push_seg6_encap(struct sk
 
 	switch (type) {
 	case BPF_LWT_ENCAP_SEG6_INLINE:
-		if (skb->protocol != htons(ETH_P_IPV6))
+		if (skb_protocol(skb, true) != htons(ETH_P_IPV6))
 			return -EBADMSG;
 
 		err = seg6_do_srh_inline(skb, srh);
--- a/net/sched/act_connmark.c
+++ b/net/sched/act_connmark.c
@@ -46,17 +46,20 @@ static int tcf_connmark_act(struct sk_bu
 	tcf_lastuse_update(&ca->tcf_tm);
 	bstats_update(&ca->tcf_bstats, skb);
 
-	if (skb->protocol == htons(ETH_P_IP)) {
+	switch (skb_protocol(skb, true)) {
+	case htons(ETH_P_IP):
 		if (skb->len < sizeof(struct iphdr))
 			goto out;
 
 		proto = NFPROTO_IPV4;
-	} else if (skb->protocol == htons(ETH_P_IPV6)) {
+		break;
+	case htons(ETH_P_IPV6):
 		if (skb->len < sizeof(struct ipv6hdr))
 			goto out;
 
 		proto = NFPROTO_IPV6;
-	} else {
+		break;
+	default:
 		goto out;
 	}
 
--- a/net/sched/act_csum.c
+++ b/net/sched/act_csum.c
@@ -577,7 +577,7 @@ static int tcf_csum_act(struct sk_buff *
 		goto drop;
 
 	update_flags = params->update_flags;
-	protocol = tc_skb_protocol(skb);
+	protocol = skb_protocol(skb, false);
 again:
 	switch (protocol) {
 	case cpu_to_be16(ETH_P_IP):
--- a/net/sched/act_skbedit.c
+++ b/net/sched/act_skbedit.c
@@ -51,7 +51,7 @@ static int tcf_skbedit_act(struct sk_buf
 	if (params->flags & SKBEDIT_F_INHERITDSFIELD) {
 		int wlen = skb_network_offset(skb);
 
-		switch (tc_skb_protocol(skb)) {
+		switch (skb_protocol(skb, true)) {
 		case htons(ETH_P_IP):
 			wlen += sizeof(struct iphdr);
 			if (!pskb_may_pull(skb, wlen))
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -969,7 +969,7 @@ int tcf_classify(struct sk_buff *skb, co
 reclassify:
 #endif
 	for (; tp; tp = rcu_dereference_bh(tp->next)) {
-		__be16 protocol = tc_skb_protocol(skb);
+		__be16 protocol = skb_protocol(skb, false);
 		int err;
 
 		if (tp->protocol != protocol &&
--- a/net/sched/cls_flow.c
+++ b/net/sched/cls_flow.c
@@ -84,7 +84,7 @@ static u32 flow_get_dst(const struct sk_
 	if (dst)
 		return ntohl(dst);
 
-	return addr_fold(skb_dst(skb)) ^ (__force u16) tc_skb_protocol(skb);
+	return addr_fold(skb_dst(skb)) ^ (__force u16)skb_protocol(skb, true);
 }
 
 static u32 flow_get_proto(const struct sk_buff *skb,
@@ -108,7 +108,7 @@ static u32 flow_get_proto_dst(const stru
 	if (flow->ports.ports)
 		return ntohs(flow->ports.dst);
 
-	return addr_fold(skb_dst(skb)) ^ (__force u16) tc_skb_protocol(skb);
+	return addr_fold(skb_dst(skb)) ^ (__force u16)skb_protocol(skb, true);
 }
 
 static u32 flow_get_iif(const struct sk_buff *skb)
@@ -155,7 +155,7 @@ static u32 flow_get_nfct(const struct sk
 static u32 flow_get_nfct_src(const struct sk_buff *skb,
 			     const struct flow_keys *flow)
 {
-	switch (tc_skb_protocol(skb)) {
+	switch (skb_protocol(skb, true)) {
 	case htons(ETH_P_IP):
 		return ntohl(CTTUPLE(skb, src.u3.ip));
 	case htons(ETH_P_IPV6):
@@ -168,7 +168,7 @@ fallback:
 static u32 flow_get_nfct_dst(const struct sk_buff *skb,
 			     const struct flow_keys *flow)
 {
-	switch (tc_skb_protocol(skb)) {
+	switch (skb_protocol(skb, true)) {
 	case htons(ETH_P_IP):
 		return ntohl(CTTUPLE(skb, dst.u3.ip));
 	case htons(ETH_P_IPV6):
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -203,7 +203,7 @@ static int fl_classify(struct sk_buff *s
 		/* skb_flow_dissect() does not set n_proto in case an unknown
 		 * protocol, so do it rather here.
 		 */
-		skb_key.basic.n_proto = skb->protocol;
+		skb_key.basic.n_proto = skb_protocol(skb, false);
 		skb_flow_dissect_tunnel_info(skb, &mask->dissector, &skb_key);
 		skb_flow_dissect(skb, &mask->dissector, &skb_key, 0);
 
--- a/net/sched/em_ipset.c
+++ b/net/sched/em_ipset.c
@@ -62,7 +62,7 @@ static int em_ipset_match(struct sk_buff
 	};
 	int ret, network_offset;
 
-	switch (tc_skb_protocol(skb)) {
+	switch (skb_protocol(skb, true)) {
 	case htons(ETH_P_IP):
 		state.pf = NFPROTO_IPV4;
 		if (!pskb_network_may_pull(skb, sizeof(struct iphdr)))
--- a/net/sched/em_meta.c
+++ b/net/sched/em_meta.c
@@ -199,7 +199,7 @@ META_COLLECTOR(int_priority)
 META_COLLECTOR(int_protocol)
 {
 	/* Let userspace take care of the byte ordering */
-	dst->value = tc_skb_protocol(skb);
+	dst->value = skb_protocol(skb, false);
 }
 
 META_COLLECTOR(int_pkttype)
--- a/net/sched/sch_cake.c
+++ b/net/sched/sch_cake.c
@@ -589,7 +589,7 @@ static void cake_update_flowkeys(struct
 	struct nf_conntrack_tuple tuple = {};
 	bool rev = !skb->_nfct;
 
-	if (tc_skb_protocol(skb) != htons(ETH_P_IP))
+	if (skb_protocol(skb, true) != htons(ETH_P_IP))
 		return;
 
 	if (!nf_ct_get_tuple_skb(&tuple, skb))
@@ -1514,7 +1514,7 @@ static u8 cake_handle_diffserv(struct sk
 	u16 *buf, buf_;
 	u8 dscp;
 
-	switch (tc_skb_protocol(skb)) {
+	switch (skb_protocol(skb, true)) {
 	case htons(ETH_P_IP):
 		buf = skb_header_pointer(skb, offset, sizeof(buf_), &buf_);
 		if (unlikely(!buf))
--- a/net/sched/sch_dsmark.c
+++ b/net/sched/sch_dsmark.c
@@ -207,7 +207,7 @@ static int dsmark_enqueue(struct sk_buff
 	if (p->set_tc_index) {
 		int wlen = skb_network_offset(skb);
 
-		switch (tc_skb_protocol(skb)) {
+		switch (skb_protocol(skb, true)) {
 		case htons(ETH_P_IP):
 			wlen += sizeof(struct iphdr);
 			if (!pskb_may_pull(skb, wlen) ||
@@ -300,7 +300,7 @@ static struct sk_buff *dsmark_dequeue(st
 	index = skb->tc_index & (p->indices - 1);
 	pr_debug("index %d->%d\n", skb->tc_index, index);
 
-	switch (tc_skb_protocol(skb)) {
+	switch (skb_protocol(skb, true)) {
 	case htons(ETH_P_IP):
 		ipv4_change_dsfield(ip_hdr(skb), p->mv[index].mask,
 				    p->mv[index].value);
@@ -317,7 +317,7 @@ static struct sk_buff *dsmark_dequeue(st
 		 */
 		if (p->mv[index].mask != 0xff || p->mv[index].value)
 			pr_warn("%s: unsupported protocol %d\n",
-				__func__, ntohs(tc_skb_protocol(skb)));
+				__func__, ntohs(skb_protocol(skb, true)));
 		break;
 	}
 
--- a/net/sched/sch_teql.c
+++ b/net/sched/sch_teql.c
@@ -243,7 +243,7 @@ __teql_resolve(struct sk_buff *skb, stru
 		char haddr[MAX_ADDR_LEN];
 
 		neigh_ha_snapshot(haddr, n, dev);
-		err = dev_hard_header(skb, dev, ntohs(tc_skb_protocol(skb)),
+		err = dev_hard_header(skb, dev, ntohs(skb_protocol(skb, false)),
 				      haddr, NULL, skb->len);
 
 		if (err < 0)