[PATCH net-next 0/4] vxlan: generalize and refine drop reasons

[PATCH net-next 0/4] vxlan: generalize and refine drop reasons

[Anton Danilov]

This series makes two vxlan drop reasons generic, so that other tunnels
can report the same conditions with them, and makes the reasons vxlan
reports for two other failures more precise.

Patch 1 renames VXLAN_INVALID_HDR and VXLAN_VNI_NOT_FOUND to
TUNNEL_INVALID_HDR and TUNNEL_NOT_FOUND. Their values stay the same;
the names vxlan drops are reported with change.

Patch 2 makes __iptunnel_pull_header() and iptunnel_pull_header() return
a drop reason. Until now they returned -ENOMEM for any failure, so a
packet too short to pull looked like an allocation failure. The three
callers in ip_gre and ip6_gre that tested the result with "< 0" test for
a non-zero value instead. Patch 3 has vxlan report the reason these
functions now return: PKT_TOO_SMALL instead of NOMEM for an inner frame
shorter than an Ethernet header.

Patch 4 has vxlan report a route that goes out of the vxlan device
itself as RECURSION_LIMIT instead of IP_OUTNOROUTES, and adds this case
to the description of RECURSION_LIMIT.

These were patches 1-3 and 14 of v5 of "tunnels: add core and gre drop
reasons", split out as Ido suggested. The conversion of ip_tunnel,
ip6_tunnel and GRE, the rest of that series, will follow once these
are in.

Changes since v5 of "tunnels: add core and gre drop reasons":
- split out of that series (Ido)
- the messages of patches 1, 2 and 4 no longer refer to the rest of it
- the message of patch 2 says that vxlan keeps reporting NOMEM until
  patch 3
- patch 4 adds the tunnel case to the description of RECURSION_LIMIT,
  which v5 did in its patch 9

- v5: https://lore.kernel.org/netdev/20260930183910.3151873-1-littlesmilingcloud@gmail.com/

Anton Danilov (4):
  vxlan: rename the drop reasons for use by other tunnels
  ip_tunnel: make __iptunnel_pull_header() return a drop reason
  vxlan: report the drop reason of __iptunnel_pull_header()
  vxlan: report a circular route as SKB_DROP_REASON_RECURSION_LIMIT

 drivers/net/vxlan/vxlan_core.c | 26 ++++++++++++++++----------
 include/net/dropreason-core.h  | 25 ++++++++++++++++---------
 include/net/ip_tunnels.h       | 10 ++++++----
 net/ipv4/ip_gre.c              |  4 ++--
 net/ipv4/ip_tunnel_core.c      | 22 +++++++++++++++-------
 net/ipv6/ip6_gre.c             |  2 +-
 6 files changed, 56 insertions(+), 33 deletions(-)

-- 
2.47.3

[PATCH net-next 1/4] vxlan: rename the drop reasons for use by other tunnels

[Anton Danilov]

VXLAN_INVALID_HDR and VXLAN_VNI_NOT_FOUND describe conditions that are
not specific to vxlan: a malformed tunnel header, and a packet for
which no tunnel device is found. The GRE receive paths drop packets
for the same two conditions, and can report them with the same reasons
once they are converted to drop reasons.

Rename them to TUNNEL_INVALID_HDR and TUNNEL_NOT_FOUND and reword their
descriptions, the way VXLAN_NO_REMOTE became NO_TX_TARGET in commit
46e0ccfb88f0 ("net: vxlan: rename SKB_DROP_REASON_VXLAN_NO_REMOTE").
The numeric values stay the same; the names that the skb:kfree_skb
tracepoint and drop_monitor report change.

Suggested-by: Ido Schimmel <idosch@nvidia.com>
Assisted-by: LLM
Signed-off-by: Anton Danilov <littlesmilingcloud@gmail.com>
---
 drivers/net/vxlan/vxlan_core.c |  8 ++++----
 include/net/dropreason-core.h  | 19 +++++++++++--------
 2 files changed, 15 insertions(+), 12 deletions(-)

diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
index 27b0b6567d52..35a47cc27200 100644
--- a/drivers/net/vxlan/vxlan_core.c
+++ b/drivers/net/vxlan/vxlan_core.c
@@ -1700,7 +1700,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 	if (!(vh->vx_flags & VXLAN_HF_VNI)) {
 		netdev_dbg(skb->dev, "invalid vxlan flags=%#x vni=%#x\n",
 			   ntohl(vh->vx_flags), ntohl(vh->vx_vni));
-		reason = SKB_DROP_REASON_VXLAN_INVALID_HDR;
+		reason = SKB_DROP_REASON_TUNNEL_INVALID_HDR;
 		/* Return non vxlan pkt */
 		goto drop;
 	}
@@ -1713,7 +1713,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 
 	vxlan = vxlan_vs_find_vni(vs, skb->dev->ifindex, vni, &vninode);
 	if (!vxlan) {
-		reason = SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND;
+		reason = SKB_DROP_REASON_TUNNEL_NOT_FOUND;
 		goto drop;
 	}
 
@@ -1729,7 +1729,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		 * previous stack code, and also is more robust and provides a
 		 * little more security in adding extensions to VXLAN.
 		 */
-		reason = SKB_DROP_REASON_VXLAN_INVALID_HDR;
+		reason = SKB_DROP_REASON_TUNNEL_INVALID_HDR;
 		DEV_STATS_INC(vxlan->dev, rx_frame_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
 		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
@@ -2389,7 +2389,7 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 			DEV_STATS_INC(dev, tx_errors);
 			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_TX_ERRORS, 0);
-			kfree_skb_reason(skb, SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND);
+			kfree_skb_reason(skb, SKB_DROP_REASON_TUNNEL_NOT_FOUND);
 
 			return -ENOENT;
 		}
diff --git a/include/net/dropreason-core.h b/include/net/dropreason-core.h
index 12f909651591..40a27d8887af 100644
--- a/include/net/dropreason-core.h
+++ b/include/net/dropreason-core.h
@@ -111,8 +111,8 @@
 	FN(PACKET_SOCK_ERROR)		\
 	FN(TC_CHAIN_NOTFOUND)		\
 	FN(TC_RECLASSIFY_LOOP)		\
-	FN(VXLAN_INVALID_HDR)		\
-	FN(VXLAN_VNI_NOT_FOUND)		\
+	FN(TUNNEL_INVALID_HDR)		\
+	FN(TUNNEL_NOT_FOUND)		\
 	FN(MAC_INVALID_SOURCE)		\
 	FN(VXLAN_ENTRY_EXISTS)		\
 	FN(NO_TX_TARGET)		\
@@ -539,13 +539,16 @@ enum skb_drop_reason {
 	 */
 	SKB_DROP_REASON_TC_RECLASSIFY_LOOP,
 	/**
-	 * @SKB_DROP_REASON_VXLAN_INVALID_HDR: VXLAN header is invalid. E.g.:
-	 * 1) reserved fields are not zero
-	 * 2) "I" flag is not set
+	 * @SKB_DROP_REASON_TUNNEL_INVALID_HDR: tunnel header is invalid. E.g.:
+	 * 1) VXLAN reserved fields are not zero
+	 * 2) VXLAN "I" flag is not set
 	 */
-	SKB_DROP_REASON_VXLAN_INVALID_HDR,
-	/** @SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND: no VXLAN device found for VNI */
-	SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND,
+	SKB_DROP_REASON_TUNNEL_INVALID_HDR,
+	/**
+	 * @SKB_DROP_REASON_TUNNEL_NOT_FOUND: no tunnel device found for the
+	 * packet, e.g. no VXLAN device for its VNI
+	 */
+	SKB_DROP_REASON_TUNNEL_NOT_FOUND,
 	/** @SKB_DROP_REASON_MAC_INVALID_SOURCE: source mac is invalid */
 	SKB_DROP_REASON_MAC_INVALID_SOURCE,
 	/**
-- 
2.47.3

[PATCH net-next 2/4] ip_tunnel: make __iptunnel_pull_header() return a drop reason

[Anton Danilov]

__iptunnel_pull_header() returns -ENOMEM whenever it fails. It can
fail in two pskb_may_pull() calls, one for the tunnel header and one
for the inner Ethernet header of ETH_P_TEB, and in the skb_unclone()
done for GSO packets. pskb_may_pull() fails when the packet is shorter
than the requested length as well as when pulling from the frags cannot
allocate, so a truncated packet and an allocation failure look the same
to the callers. The ones that report a drop reason can only pick
SKB_DROP_REASON_NOMEM, as vxlan_rcv() does, and so would the GRE
receive paths once they report drop reasons. In ip6_gre, gre_rcv()
pulls the header before the tunnel lookup, so a packet from any sender
whose ETH_P_TEB inner Ethernet header or WCCPv2 extra word is cut short
would be reported as an out of memory condition.

Make __iptunnel_pull_header() and iptunnel_pull_header() return the
reason pskb_may_pull_reason() already computes, SKB_DROP_REASON_NOMEM
when skb_unclone() fails, and SKB_NOT_DROPPED_YET on success. A failure
is still non-zero, so the callers that only test the result keep
working. Three callers, in ip_gre and ip6_gre, test it with "< 0"
instead; the enum has no negative values, so that test would always be
false, and the compiler does not warn about it. Make them test for a
non-zero value. vxlan_rcv() keeps reporting NOMEM for any failure until
the next patch has it report the returned reason.

Suggested-by: Ido Schimmel <idosch@nvidia.com>
Assisted-by: LLM
Signed-off-by: Anton Danilov <littlesmilingcloud@gmail.com>
---
 include/net/ip_tunnels.h  | 10 ++++++----
 net/ipv4/ip_gre.c         |  4 ++--
 net/ipv4/ip_tunnel_core.c | 22 +++++++++++++++-------
 net/ipv6/ip6_gre.c        |  2 +-
 4 files changed, 24 insertions(+), 14 deletions(-)

diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h
index 7102aa11fae2..5cccf4c0e691 100644
--- a/include/net/ip_tunnels.h
+++ b/include/net/ip_tunnels.h
@@ -614,11 +614,13 @@ static inline u8 ip_tunnel_ecn_encap(u8 tos, const struct iphdr *iph,
 	return INET_ECN_encapsulate(tos, inner);
 }
 
-int __iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
-			   __be16 inner_proto, bool raw_proto, bool xnet);
+enum skb_drop_reason
+__iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
+		       __be16 inner_proto, bool raw_proto, bool xnet);
 
-static inline int iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
-				       __be16 inner_proto, bool xnet)
+static inline enum skb_drop_reason
+iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
+		     __be16 inner_proto, bool xnet)
 {
 	return __iptunnel_pull_header(skb, hdr_len, inner_proto, false, xnet);
 }
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 51fcd603939c..3ee437fa3eae 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -312,7 +312,7 @@ static int erspan_rcv(struct sk_buff *skb, struct tnl_ptk_info *tpi,
 		if (__iptunnel_pull_header(skb,
 					   len,
 					   htons(ETH_P_TEB),
-					   false, false) < 0)
+					   false, false))
 			goto drop;
 
 		if (tunnel->collect_md) {
@@ -378,7 +378,7 @@ static int __ipgre_rcv(struct sk_buff *skb, const struct tnl_ptk_info *tpi,
 		const struct iphdr *tnl_params;
 
 		if (__iptunnel_pull_header(skb, hdr_len, tpi->proto,
-					   raw_proto, false) < 0)
+					   raw_proto, false))
 			goto drop;
 
 		/* Special case for ipgre_header_parse(), which expects the
diff --git a/net/ipv4/ip_tunnel_core.c b/net/ipv4/ip_tunnel_core.c
index bab42b9e277f..6c2855adff28 100644
--- a/net/ipv4/ip_tunnel_core.c
+++ b/net/ipv4/ip_tunnel_core.c
@@ -106,19 +106,24 @@ void iptunnel_xmit(struct sock *sk, struct rtable *rt, struct sk_buff *skb,
 }
 EXPORT_SYMBOL_GPL(iptunnel_xmit);
 
-int __iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
-			   __be16 inner_proto, bool raw_proto, bool xnet)
+enum skb_drop_reason
+__iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
+		       __be16 inner_proto, bool raw_proto, bool xnet)
 {
-	if (unlikely(!pskb_may_pull(skb, hdr_len)))
-		return -ENOMEM;
+	enum skb_drop_reason reason;
+
+	reason = pskb_may_pull_reason(skb, hdr_len);
+	if (unlikely(reason))
+		return reason;
 
 	skb_pull_rcsum(skb, hdr_len);
 
 	if (!raw_proto && inner_proto == htons(ETH_P_TEB)) {
 		struct ethhdr *eh;
 
-		if (unlikely(!pskb_may_pull(skb, ETH_HLEN)))
-			return -ENOMEM;
+		reason = pskb_may_pull_reason(skb, ETH_HLEN);
+		if (unlikely(reason))
+			return reason;
 
 		eh = (struct ethhdr *)skb->data;
 		if (likely(eth_proto_is_802_3(eh->h_proto)))
@@ -135,7 +140,10 @@ int __iptunnel_pull_header(struct sk_buff *skb, int hdr_len,
 	skb_set_queue_mapping(skb, 0);
 	skb_scrub_packet(skb, xnet);
 
-	return iptunnel_pull_offloads(skb);
+	if (unlikely(iptunnel_pull_offloads(skb)))
+		return SKB_DROP_REASON_NOMEM;
+
+	return SKB_NOT_DROPPED_YET;
 }
 EXPORT_SYMBOL_GPL(__iptunnel_pull_header);
 
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index 258239a7c53b..0392f6ba862b 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -512,7 +512,7 @@ static int ip6erspan_rcv(struct sk_buff *skb,
 
 		if (__iptunnel_pull_header(skb, len,
 					   htons(ETH_P_TEB),
-					   false, false) < 0)
+					   false, false))
 			return PACKET_REJECT;
 
 		if (tunnel->parms.collect_md) {
-- 
2.47.3

[PATCH net-next 3/4] vxlan: report the drop reason of __iptunnel_pull_header()

[Anton Danilov]

vxlan_rcv() reports SKB_DROP_REASON_NOMEM whenever
__iptunnel_pull_header() fails, as the helper used to return -ENOMEM
for any failure. The VXLAN header is already in the linear area at this
point, but the inner Ethernet header of a packet without GPE may not
be, so a packet whose inner frame is shorter than an Ethernet header is
reported as an out of memory condition.

Report the reason __iptunnel_pull_header() now returns instead:
PKT_TOO_SMALL for such a packet, NOMEM when an allocation fails.

Suggested-by: Ido Schimmel <idosch@nvidia.com>
Assisted-by: LLM
Signed-off-by: Anton Danilov <littlesmilingcloud@gmail.com>
---
 drivers/net/vxlan/vxlan_core.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
index 35a47cc27200..0d98e29e2d72 100644
--- a/drivers/net/vxlan/vxlan_core.c
+++ b/drivers/net/vxlan/vxlan_core.c
@@ -1743,11 +1743,11 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		raw_proto = true;
 	}
 
-	if (__iptunnel_pull_header(skb, VXLAN_HLEN, protocol, raw_proto,
-				   !net_eq(vxlan->net, dev_net(vxlan->dev)))) {
-		reason = SKB_DROP_REASON_NOMEM;
+	reason = __iptunnel_pull_header(skb, VXLAN_HLEN, protocol, raw_proto,
+					!net_eq(vxlan->net,
+						dev_net(vxlan->dev)));
+	if (reason)
 		goto drop;
-	}
 
 	if (cfg->flags & VXLAN_F_REMCSUM_RX) {
 		reason = vxlan_remcsum(skb, cfg->flags);
-- 
2.47.3

[PATCH net-next 4/4] vxlan: report a circular route as SKB_DROP_REASON_RECURSION_LIMIT

[Anton Danilov]

udp_tunnel_dst_lookup() and udp_tunnel6_dst_lookup() return -ELOOP
when the route to the remote end goes out of the vxlan device itself,
and vxlan_xmit_one() counts that in collisions, apart from the
-ENETUNREACH of a failed lookup. The drop reason does not tell them
apart: both are reported as SKB_DROP_REASON_IP_OUTNOROUTES, which
describes a failed route lookup.

Such a route would send the encapsulated packet back into the device
it leaves, which is the dead loop on a virtual device that
SKB_DROP_REASON_RECURSION_LIMIT describes. Report the -ELOOP case with
that reason, and add the tunnel case to the description of the reason
as an example. ip_tunnel and ip6_tunnel make the same check,
rt->dst.dev == dev in ip_tunnel_xmit() and ip_md_tunnel_xmit() and
tdev == dev in ip6_tnl_xmit(), and can report it with the same reason.

Assisted-by: LLM
Signed-off-by: Anton Danilov <littlesmilingcloud@gmail.com>
---
 drivers/net/vxlan/vxlan_core.c | 10 ++++++++--
 include/net/dropreason-core.h  |  6 +++++-
 2 files changed, 13 insertions(+), 3 deletions(-)

diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
index 0d98e29e2d72..7c94a630d184 100644
--- a/drivers/net/vxlan/vxlan_core.c
+++ b/drivers/net/vxlan/vxlan_core.c
@@ -2547,7 +2547,10 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 					   tos, use_cache ? dst_cache : NULL);
 		if (IS_ERR(rt)) {
 			err = PTR_ERR(rt);
-			reason = SKB_DROP_REASON_IP_OUTNOROUTES;
+			if (err == -ELOOP)
+				reason = SKB_DROP_REASON_RECURSION_LIMIT;
+			else
+				reason = SKB_DROP_REASON_IP_OUTNOROUTES;
 			goto tx_error;
 		}
 
@@ -2633,7 +2636,10 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (IS_ERR(ndst)) {
 			err = PTR_ERR(ndst);
 			ndst = NULL;
-			reason = SKB_DROP_REASON_IP_OUTNOROUTES;
+			if (err == -ELOOP)
+				reason = SKB_DROP_REASON_RECURSION_LIMIT;
+			else
+				reason = SKB_DROP_REASON_IP_OUTNOROUTES;
 			goto tx_error;
 		}
 
diff --git a/include/net/dropreason-core.h b/include/net/dropreason-core.h
index 40a27d8887af..664249924219 100644
--- a/include/net/dropreason-core.h
+++ b/include/net/dropreason-core.h
@@ -613,7 +613,11 @@ enum skb_drop_reason {
 	SKB_DROP_REASON_PSP_INPUT,
 	/** @SKB_DROP_REASON_PSP_OUTPUT: PSP output checks failed */
 	SKB_DROP_REASON_PSP_OUTPUT,
-	/** @SKB_DROP_REASON_RECURSION_LIMIT: Dead loop on virtual device. */
+	/**
+	 * @SKB_DROP_REASON_RECURSION_LIMIT: Dead loop on virtual device, e.g. a
+	 * tunnel whose route to its remote end goes out of the tunnel device
+	 * itself.
+	 */
 	SKB_DROP_REASON_RECURSION_LIMIT,
 	/**
 	 * @SKB_DROP_REASON_MAX: the maximum of core drop reasons, which
-- 
2.47.3