public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Marek Mietus <mmietus97@yahoo.com>
To: netdev@vger.kernel.org, sd@queasysnail.net, kuba@kernel.org,
	pabeni@redhat.com
Cc: Jason@zx2c4.com, Marek Mietus <mmietus97@yahoo.com>
Subject: [PATCH net-next v8 04/11] net: tunnel: return noref dsts in udp_tunnel{,6}_dst_lookup
Date: Thu, 12 Mar 2026 16:56:50 +0100	[thread overview]
Message-ID: <20260312155657.25676-5-mmietus97@yahoo.com> (raw)
In-Reply-To: <20260312155657.25676-1-mmietus97@yahoo.com>

Update udp_tunnel{,6}_dst_lookup to always return noref dsts. The dst
is only valid inside the RCU read-side critical section in which it was
queried.

The dst is fetched from the dst cache (as noref) or returned by a routing
lookup operation, in which case its refcount is either stolen into the
cache, or decremented shortly before returning (in case the cache can't
be used). This is safe, since this code runs in an RCU read-side critical
section, and the dst only lingers until the end of said section.

Update all callers to use the new convention (of no longer calling
dst_release, since all dsts are now noref). This affects the bareudp,
geneve and vxlan tunnels.

Signed-off-by: Marek Mietus <mmietus97@yahoo.com>
---
 drivers/net/bareudp.c          |  6 ------
 drivers/net/geneve.c           | 13 -------------
 drivers/net/vxlan/vxlan_core.c |  8 --------
 net/ipv4/udp_tunnel_core.c     |  6 ++++--
 net/ipv6/ip6_udp_tunnel.c      |  8 ++++++--
 5 files changed, 10 insertions(+), 31 deletions(-)

diff --git a/drivers/net/bareudp.c b/drivers/net/bareudp.c
index 92ee4a36f86f..456bc17c352d 100644
--- a/drivers/net/bareudp.c
+++ b/drivers/net/bareudp.c
@@ -364,11 +364,9 @@ static int bareudp_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 			    !net_eq(bareudp->net, dev_net(bareudp->dev)),
 			    !test_bit(IP_TUNNEL_CSUM_BIT, info->key.tun_flags),
 			    0);
-	ip_rt_put(rt);
 	return 0;
 
 free_dst:
-	dst_release(&rt->dst);
 	return err;
 }
 
@@ -434,11 +432,9 @@ static int bareudp6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 			     !test_bit(IP_TUNNEL_CSUM_BIT,
 				       info->key.tun_flags),
 			     0);
-	dst_release(dst);
 	return 0;
 
 free_dst:
-	dst_release(dst);
 	return err;
 }
 
@@ -524,7 +520,6 @@ static int bareudp_fill_metadata_dst(struct net_device *dev,
 		if (IS_ERR(rt))
 			return PTR_ERR(rt);
 
-		ip_rt_put(rt);
 		info->key.u.ipv4.src = saddr;
 	} else if (ip_tunnel_info_af(info) == AF_INET6) {
 		struct dst_entry *dst;
@@ -538,7 +533,6 @@ static int bareudp_fill_metadata_dst(struct net_device *dev,
 		if (IS_ERR(dst))
 			return PTR_ERR(dst);
 
-		dst_release(dst);
 		info->key.u.ipv6.src = saddr;
 	} else {
 		return -EINVAL;
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 4216a8ffd591..8a918bd009dc 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -1268,7 +1268,6 @@ static int geneve_build_skb(struct dst_entry *dst, struct sk_buff *skb,
 	return 0;
 
 free_dst:
-	dst_release(dst);
 	return err;
 }
 
@@ -1327,7 +1326,6 @@ static int geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 				    geneve_build_gro_hint_opt(geneve, skb),
 				    netif_is_any_bridge_port(dev));
 	if (err < 0) {
-		dst_release(&rt->dst);
 		return err;
 	} else if (err) {
 		struct ip_tunnel_info *info;
@@ -1338,7 +1336,6 @@ static int geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 
 			unclone = skb_tunnel_info_unclone(skb);
 			if (unlikely(!unclone)) {
-				dst_release(&rt->dst);
 				return -ENOMEM;
 			}
 
@@ -1347,13 +1344,11 @@ static int geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 		}
 
 		if (!pskb_may_pull(skb, ETH_HLEN)) {
-			dst_release(&rt->dst);
 			return -EINVAL;
 		}
 
 		skb->protocol = eth_type_trans(skb, geneve->dev);
 		__netif_rx(skb);
-		dst_release(&rt->dst);
 		return -EMSGSIZE;
 	}
 
@@ -1396,7 +1391,6 @@ static int geneve_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 			    !net_eq(geneve->net, dev_net(geneve->dev)),
 			    !test_bit(IP_TUNNEL_CSUM_BIT, info->key.tun_flags),
 			    0);
-	ip_rt_put(rt);
 	return 0;
 }
 
@@ -1439,7 +1433,6 @@ static int geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 				    geneve_build_gro_hint_opt(geneve, skb),
 				    netif_is_any_bridge_port(dev));
 	if (err < 0) {
-		dst_release(dst);
 		return err;
 	} else if (err) {
 		struct ip_tunnel_info *info = skb_tunnel_info(skb);
@@ -1449,7 +1442,6 @@ static int geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 
 			unclone = skb_tunnel_info_unclone(skb);
 			if (unlikely(!unclone)) {
-				dst_release(dst);
 				return -ENOMEM;
 			}
 
@@ -1458,13 +1450,11 @@ static int geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 		}
 
 		if (!pskb_may_pull(skb, ETH_HLEN)) {
-			dst_release(dst);
 			return -EINVAL;
 		}
 
 		skb->protocol = eth_type_trans(skb, geneve->dev);
 		__netif_rx(skb);
-		dst_release(dst);
 		return -EMSGSIZE;
 	}
 
@@ -1488,7 +1478,6 @@ static int geneve6_xmit_skb(struct sk_buff *skb, struct net_device *dev,
 			     !test_bit(IP_TUNNEL_CSUM_BIT,
 				       info->key.tun_flags),
 			     0);
-	dst_release(dst);
 	return 0;
 }
 #endif
@@ -1576,7 +1565,6 @@ static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 		if (IS_ERR(rt))
 			return PTR_ERR(rt);
 
-		ip_rt_put(rt);
 		info->key.u.ipv4.src = saddr;
 #if IS_ENABLED(CONFIG_IPV6)
 	} else if (ip_tunnel_info_af(info) == AF_INET6) {
@@ -1602,7 +1590,6 @@ static int geneve_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 		if (IS_ERR(dst))
 			return PTR_ERR(dst);
 
-		dst_release(dst);
 		info->key.u.ipv6.src = saddr;
 #endif
 	} else {
diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
index 4482a47dbe15..39fb2e6df6c4 100644
--- a/drivers/net/vxlan/vxlan_core.c
+++ b/drivers/net/vxlan/vxlan_core.c
@@ -2320,7 +2320,6 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 	    vxlan->cfg.flags & VXLAN_F_LOCALBYPASS) {
 		struct vxlan_dev *dst_vxlan;
 
-		dst_release(dst);
 		dst_vxlan = vxlan_find_vni(vxlan->net, dst_ifindex, vni,
 					   addr_family, dst_port,
 					   vxlan->cfg.flags);
@@ -2528,7 +2527,6 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				unclone->key.u.ipv4.dst = saddr;
 			}
 			vxlan_encap_bypass(skb, vxlan, vxlan, vni, false);
-			dst_release(ndst);
 			goto out_unlock;
 		}
 
@@ -2545,7 +2543,6 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				    pkey->u.ipv4.dst, tos, ttl, df,
 				    src_port, dst_port, xnet, !udp_sum,
 				    ipcb_flags);
-		ip_rt_put(rt);
 #if IS_ENABLED(CONFIG_IPV6)
 	} else {
 		struct vxlan_sock *sock6;
@@ -2603,7 +2600,6 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 			}
 
 			vxlan_encap_bypass(skb, vxlan, vxlan, vni, false);
-			dst_release(ndst);
 			goto out_unlock;
 		}
 
@@ -2621,7 +2617,6 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				     &saddr, &pkey->u.ipv6.dst, tos, ttl,
 				     pkey->label, src_port, dst_port, !udp_sum,
 				     ip6cb_flags);
-		dst_release(ndst);
 #endif
 	}
 	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);
@@ -2641,7 +2636,6 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		DEV_STATS_INC(dev, collisions);
 	else if (err == -ENETUNREACH)
 		DEV_STATS_INC(dev, tx_carrier_errors);
-	dst_release(ndst);
 	DEV_STATS_INC(dev, tx_errors);
 	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);
 	kfree_skb_reason(skb, reason);
@@ -3248,7 +3242,6 @@ static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 					   &info->dst_cache);
 		if (IS_ERR(rt))
 			return PTR_ERR(rt);
-		ip_rt_put(rt);
 	} else {
 #if IS_ENABLED(CONFIG_IPV6)
 		struct vxlan_sock *sock6 = rcu_dereference(vxlan->vn6_sock);
@@ -3264,7 +3257,6 @@ static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 					      &info->dst_cache);
 		if (IS_ERR(ndst))
 			return PTR_ERR(ndst);
-		dst_release(ndst);
 #else /* !CONFIG_IPV6 */
 		return -EPFNOSUPPORT;
 #endif
diff --git a/net/ipv4/udp_tunnel_core.c b/net/ipv4/udp_tunnel_core.c
index b1f667c52cb2..c9c3fe8f0158 100644
--- a/net/ipv4/udp_tunnel_core.c
+++ b/net/ipv4/udp_tunnel_core.c
@@ -240,7 +240,7 @@ struct rtable *udp_tunnel_dst_lookup(struct sk_buff *skb,
 
 #ifdef CONFIG_DST_CACHE
 	if (dst_cache) {
-		rt = dst_cache_get_ip4(dst_cache, saddr);
+		rt = dst_cache_get_ip4_rcu(dst_cache, saddr);
 		if (rt)
 			return rt;
 	}
@@ -269,8 +269,10 @@ struct rtable *udp_tunnel_dst_lookup(struct sk_buff *skb,
 	}
 #ifdef CONFIG_DST_CACHE
 	if (dst_cache)
-		dst_cache_set_ip4(dst_cache, &rt->dst, fl4.saddr);
+		dst_cache_steal_ip4(dst_cache, &rt->dst, fl4.saddr);
+	else
 #endif
+		ip_rt_put(rt);
 	*saddr = fl4.saddr;
 	return rt;
 }
diff --git a/net/ipv6/ip6_udp_tunnel.c b/net/ipv6/ip6_udp_tunnel.c
index d58815db8182..94901935c9e9 100644
--- a/net/ipv6/ip6_udp_tunnel.c
+++ b/net/ipv6/ip6_udp_tunnel.c
@@ -130,6 +130,8 @@ EXPORT_SYMBOL_GPL(udp_tunnel6_xmit_skb);
  *
  *      It returns a valid dst pointer and stores src address to be used in
  *      tunnel in param saddr on success, else a pointer encoded error code.
+ *      The returned dst pointer is noref and must only be used in the RCU
+ *      read-side critical section in which it was queried.
  */
 
 struct dst_entry *udp_tunnel6_dst_lookup(struct sk_buff *skb,
@@ -147,7 +149,7 @@ struct dst_entry *udp_tunnel6_dst_lookup(struct sk_buff *skb,
 
 #ifdef CONFIG_DST_CACHE
 	if (dst_cache) {
-		dst = dst_cache_get_ip6(dst_cache, saddr);
+		dst = dst_cache_get_ip6_rcu(dst_cache, saddr);
 		if (dst)
 			return dst;
 	}
@@ -175,8 +177,10 @@ struct dst_entry *udp_tunnel6_dst_lookup(struct sk_buff *skb,
 	}
 #ifdef CONFIG_DST_CACHE
 	if (dst_cache)
-		dst_cache_set_ip6(dst_cache, dst, &fl6.saddr);
+		dst_cache_steal_ip6(dst_cache, dst, &fl6.saddr);
+	else
 #endif
+		dst_release(dst);
 	*saddr = fl6.saddr;
 	return dst;
 }
-- 
2.51.0


  parent reply	other threads:[~2026-03-12 15:58 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20260312155657.25676-1-mmietus97.ref@yahoo.com>
2026-03-12 15:56 ` [PATCH net-next v8 00/11] net: tunnel: introduce noref xmit flows for tunnels Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 01/11] net: dst_cache: add noref versions for dst_cache Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 02/11] net: tunnel: convert iptunnel_xmit to noref Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 03/11] net: tunnel: convert udp_tunnel{6,}_xmit_skb " Marek Mietus
2026-03-12 15:56   ` Marek Mietus [this message]
2026-03-12 15:56   ` [PATCH net-next v8 05/11] net: ovpn: convert ovpn_udp{4,6}_output to use a noref dst Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 06/11] wireguard: socket: convert send{4,6} to use a noref dst when possible Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 07/11] net: tunnel: convert ip_md_tunnel_xmit to use noref dsts Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 08/11] net: tunnel: convert ip_tunnel_xmit to use a noref dst when possible Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 09/11] net: sctp: convert sctp_v{4,6}_xmit " Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 10/11] net: sit: convert ipip6_tunnel_xmit to use a noref dst Marek Mietus
2026-03-12 15:56   ` [PATCH net-next v8 11/11] net: tipc: convert tipc_udp_xmit " Marek Mietus
2026-03-17 11:37   ` [PATCH net-next v8 00/11] net: tunnel: introduce noref xmit flows for tunnels Paolo Abeni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260312155657.25676-5-mmietus97@yahoo.com \
    --to=mmietus97@yahoo.com \
    --cc=Jason@zx2c4.com \
    --cc=kuba@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=sd@queasysnail.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox