[PATCH net v3 0/2] seg6: fix dst_cache sharing in seg6 lwtunnel

[PATCH net v3 0/2] seg6: fix dst_cache sharing in seg6 lwtunnel

[Andrea Mayer]

The seg6 lwtunnel encap uses a single per-route dst_cache shared
between seg6_input_core() and seg6_output_core(). These two paths
can perform the post-encap SID lookup in different routing contexts
(e.g., ip rules matching on the ingress interface, or VRF table
separation). Whichever path runs first populates the cache, and the
other reuses it blindly, bypassing its own lookup.

Patch 1 fixes this by splitting the cache into cache_input and
cache_output. Patch 2 adds a selftest that validates the isolation.

Changes v2 -> v3:
 - Patch 2: add root privilege check (Justin)
 - Patch 2: extend prerequisite tool checks and use ping instead
   of ping6 for consistency with other srv6 selftests

Changes v1 -> v2:
 - Patch 2: fix SKIP message wording (Nicolas)

Andrea Mayer (2):
  seg6: separate dst_cache for input and output paths in seg6 lwtunnel
  selftests: seg6: add test for dst_cache isolation in seg6 lwtunnel

 net/ipv6/seg6_iptunnel.c                      |  34 ++-
 tools/testing/selftests/net/Makefile          |   1 +
 .../selftests/net/srv6_iptunnel_cache.sh      | 197 ++++++++++++++++++
 3 files changed, 221 insertions(+), 11 deletions(-)
 create mode 100755 tools/testing/selftests/net/srv6_iptunnel_cache.sh

-- 
2.43.0

[PATCH net v3 1/2] seg6: separate dst_cache for input and output paths in seg6 lwtunnel

[Andrea Mayer]

The seg6 lwtunnel uses a single dst_cache per encap route, shared
between seg6_input_core() and seg6_output_core(). These two paths
can perform the post-encap SID lookup in different routing contexts
(e.g., ip rules matching on the ingress interface, or VRF table
separation). Whichever path runs first populates the cache, and the
other reuses it blindly, bypassing its own lookup.

Fix this by splitting the cache into cache_input and cache_output,
so each path maintains its own cached dst independently.

Fixes: 6c8702c60b88 ("ipv6: sr: add support for SRH encapsulation and injection with lwtunnels")
Cc: stable@vger.kernel.org
Signed-off-by: Andrea Mayer <andrea.mayer@uniroma2.it>
Reviewed-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reviewed-by: Justin Iurman <justin.iurman@gmail.com>
---
 net/ipv6/seg6_iptunnel.c | 34 +++++++++++++++++++++++-----------
 1 file changed, 23 insertions(+), 11 deletions(-)

diff --git a/net/ipv6/seg6_iptunnel.c b/net/ipv6/seg6_iptunnel.c
index 3e1b9991131a..d6a0f7df9080 100644
--- a/net/ipv6/seg6_iptunnel.c
+++ b/net/ipv6/seg6_iptunnel.c
@@ -48,7 +48,8 @@ static size_t seg6_lwt_headroom(struct seg6_iptunnel_encap *tuninfo)
 }
 
 struct seg6_lwt {
-	struct dst_cache cache;
+	struct dst_cache cache_input;
+	struct dst_cache cache_output;
 	struct seg6_iptunnel_encap tuninfo[];
 };
 
@@ -488,7 +489,7 @@ static int seg6_input_core(struct net *net, struct sock *sk,
 	slwt = seg6_lwt_lwtunnel(lwtst);
 
 	local_bh_disable();
-	dst = dst_cache_get(&slwt->cache);
+	dst = dst_cache_get(&slwt->cache_input);
 	local_bh_enable();
 
 	err = seg6_do_srh(skb, dst);
@@ -504,7 +505,7 @@ static int seg6_input_core(struct net *net, struct sock *sk,
 		/* cache only if we don't create a dst reference loop */
 		if (!dst->error && lwtst != dst->lwtstate) {
 			local_bh_disable();
-			dst_cache_set_ip6(&slwt->cache, dst,
+			dst_cache_set_ip6(&slwt->cache_input, dst,
 					  &ipv6_hdr(skb)->saddr);
 			local_bh_enable();
 		}
@@ -564,7 +565,7 @@ static int seg6_output_core(struct net *net, struct sock *sk,
 	slwt = seg6_lwt_lwtunnel(orig_dst->lwtstate);
 
 	local_bh_disable();
-	dst = dst_cache_get(&slwt->cache);
+	dst = dst_cache_get(&slwt->cache_output);
 	local_bh_enable();
 
 	err = seg6_do_srh(skb, dst);
@@ -591,7 +592,7 @@ static int seg6_output_core(struct net *net, struct sock *sk,
 		/* cache only if we don't create a dst reference loop */
 		if (orig_dst->lwtstate != dst->lwtstate) {
 			local_bh_disable();
-			dst_cache_set_ip6(&slwt->cache, dst, &fl6.saddr);
+			dst_cache_set_ip6(&slwt->cache_output, dst, &fl6.saddr);
 			local_bh_enable();
 		}
 
@@ -701,11 +702,13 @@ static int seg6_build_state(struct net *net, struct nlattr *nla,
 
 	slwt = seg6_lwt_lwtunnel(newts);
 
-	err = dst_cache_init(&slwt->cache, GFP_ATOMIC);
-	if (err) {
-		kfree(newts);
-		return err;
-	}
+	err = dst_cache_init(&slwt->cache_input, GFP_ATOMIC);
+	if (err)
+		goto err_free_newts;
+
+	err = dst_cache_init(&slwt->cache_output, GFP_ATOMIC);
+	if (err)
+		goto err_destroy_input;
 
 	memcpy(&slwt->tuninfo, tuninfo, tuninfo_len);
 
@@ -720,11 +723,20 @@ static int seg6_build_state(struct net *net, struct nlattr *nla,
 	*ts = newts;
 
 	return 0;
+
+err_destroy_input:
+	dst_cache_destroy(&slwt->cache_input);
+err_free_newts:
+	kfree(newts);
+	return err;
 }
 
 static void seg6_destroy_state(struct lwtunnel_state *lwt)
 {
-	dst_cache_destroy(&seg6_lwt_lwtunnel(lwt)->cache);
+	struct seg6_lwt *slwt = seg6_lwt_lwtunnel(lwt);
+
+	dst_cache_destroy(&slwt->cache_input);
+	dst_cache_destroy(&slwt->cache_output);
 }
 
 static int seg6_fill_encap_info(struct sk_buff *skb,
-- 
2.43.0

[PATCH net v3 2/2] selftests: seg6: add test for dst_cache isolation in seg6 lwtunnel

[Andrea Mayer]

Add a selftest that verifies the dst_cache in seg6 lwtunnel is not
shared between the input (forwarding) and output (locally generated)
paths.

The test creates three namespaces (ns_src, ns_router, ns_dst)
connected in a line. An SRv6 encap route on ns_router encapsulates
traffic destined to cafe::1 with SID fc00::100. The SID is
reachable only for forwarded traffic (from ns_src) via an ip rule
matching the ingress interface (iif veth-r0 lookup 100), and
blackholed in the main table.

The test verifies that:

  1. A packet generated locally on ns_router does not reach
     ns_dst with an empty cache, since the SID is blackholed;
  2. A forwarded packet from ns_src populates the input cache
     from table 100 and reaches ns_dst;
  3. A packet generated locally on ns_router still does not
     reach ns_dst after the input cache is populated,
     confirming the output path does not reuse the input
     cache entry.

Both the forwarded and local packets are pinned to the same CPU
with taskset, since dst_cache is per-cpu.

Cc: Shuah Khan <shuah@kernel.org>
Cc: linux-kselftest@vger.kernel.org
Signed-off-by: Andrea Mayer <andrea.mayer@uniroma2.it>
Reviewed-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reviewed-by: Justin Iurman <justin.iurman@gmail.com>
---
Changes v2 -> v3:
 - add root privilege check (Justin)
 - extend prerequisite tool checks and use ping instead of ping6
   for consistency with other srv6 selftests

Changes v1 -> v2:
 - fix SKIP message wording (Nicolas)
---
 tools/testing/selftests/net/Makefile          |   1 +
 .../selftests/net/srv6_iptunnel_cache.sh      | 197 ++++++++++++++++++
 2 files changed, 198 insertions(+)
 create mode 100755 tools/testing/selftests/net/srv6_iptunnel_cache.sh

diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
index 605c54c0e8a3..c709523c99c6 100644
--- a/tools/testing/selftests/net/Makefile
+++ b/tools/testing/selftests/net/Makefile
@@ -89,6 +89,7 @@ TEST_PROGS := \
 	srv6_end_x_next_csid_l3vpn_test.sh \
 	srv6_hencap_red_l3vpn_test.sh \
 	srv6_hl2encap_red_l2vpn_test.sh \
+	srv6_iptunnel_cache.sh \
 	stress_reuseport_listen.sh \
 	tcp_fastopen_backup_key.sh \
 	test_bpf.sh \
diff --git a/tools/testing/selftests/net/srv6_iptunnel_cache.sh b/tools/testing/selftests/net/srv6_iptunnel_cache.sh
new file mode 100755
index 000000000000..62638ab679d9
--- /dev/null
+++ b/tools/testing/selftests/net/srv6_iptunnel_cache.sh
@@ -0,0 +1,197 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+#
+# author: Andrea Mayer <andrea.mayer@uniroma2.it>
+
+# This test verifies that the seg6 lwtunnel does not share the dst_cache
+# between the input (forwarding) and output (locally generated) paths.
+#
+# A shared dst_cache allows a forwarded packet to populate the cache and a
+# subsequent locally generated packet to silently reuse that entry, bypassing
+# its own route lookup. To expose this, the SID is made reachable only for
+# forwarded traffic (via an ip rule matching iif) and blackholed for everything
+# else. A local ping on ns_router must always hit the blackhole;
+# if it succeeds after a forwarded packet has populated the
+# cache, the bug is confirmed.
+#
+# Both forwarded and local packets are pinned to the same CPU with taskset,
+# since dst_cache is per-cpu.
+#
+#
+# +--------------------+                        +--------------------+
+# |      ns_src        |                        |      ns_dst        |
+# |                    |                        |                    |
+# |  veth-s0           |                        |           veth-d0  |
+# |  fd00::1/64        |                        |        fd01::2/64  |
+# +-------+------------+                        +----------+---------+
+#         |                                                |
+#         |            +--------------------+              |
+#         |            |     ns_router      |              |
+#         |            |                    |              |
+#         +------------+ veth-r0    veth-r1 +--------------+
+#                      | fd00::2    fd01::1 |
+#                      +--------------------+
+#
+#
+# ns_router: encap (main table)
+# +---------+---------------------------------------+
+# | dst     | action                                |
+# +---------+---------------------------------------+
+# | cafe::1 | encap seg6 mode encap segs fc00::100  |
+# +---------+---------------------------------------+
+#
+# ns_router: post-encap SID resolution
+# +-------+------------+----------------------------+
+# | table | dst        | action                     |
+# +-------+------------+----------------------------+
+# | 100   | fc00::100  | via fd01::2 dev veth-r1    |
+# +-------+------------+----------------------------+
+# | main  | fc00::100  | blackhole                  |
+# +-------+------------+----------------------------+
+#
+# ns_router: ip rule
+# +------------------+------------------------------+
+# | match            | action                       |
+# +------------------+------------------------------+
+# | iif veth-r0      | lookup 100                   |
+# +------------------+------------------------------+
+#
+# ns_dst: SRv6 decap (main table)
+# +--------------+----------------------------------+
+# | SID          | action                           |
+# +--------------+----------------------------------+
+# | fc00::100    | End.DT6 table 255 (local)        |
+# +--------------+----------------------------------+
+
+source lib.sh
+
+readonly SID="fc00::100"
+readonly DEST="cafe::1"
+
+readonly SRC_MAC="02:00:00:00:00:01"
+readonly RTR_R0_MAC="02:00:00:00:00:02"
+readonly RTR_R1_MAC="02:00:00:00:00:03"
+readonly DST_MAC="02:00:00:00:00:04"
+
+cleanup()
+{
+	cleanup_ns "${NS_SRC}" "${NS_RTR}" "${NS_DST}"
+}
+
+check_prerequisites()
+{
+	if ! command -v ip &>/dev/null; then
+		echo "SKIP: ip tool not found"
+		exit "${ksft_skip}"
+	fi
+
+	if ! command -v ping &>/dev/null; then
+		echo "SKIP: ping not found"
+		exit "${ksft_skip}"
+	fi
+
+	if ! command -v sysctl &>/dev/null; then
+		echo "SKIP: sysctl not found"
+		exit "${ksft_skip}"
+	fi
+
+	if ! command -v taskset &>/dev/null; then
+		echo "SKIP: taskset not found"
+		exit "${ksft_skip}"
+	fi
+}
+
+setup()
+{
+	setup_ns NS_SRC NS_RTR NS_DST
+
+	ip link add veth-s0 netns "${NS_SRC}" type veth \
+		peer name veth-r0 netns "${NS_RTR}"
+	ip link add veth-r1 netns "${NS_RTR}" type veth \
+		peer name veth-d0 netns "${NS_DST}"
+
+	ip -n "${NS_SRC}" link set veth-s0 address "${SRC_MAC}"
+	ip -n "${NS_RTR}" link set veth-r0 address "${RTR_R0_MAC}"
+	ip -n "${NS_RTR}" link set veth-r1 address "${RTR_R1_MAC}"
+	ip -n "${NS_DST}" link set veth-d0 address "${DST_MAC}"
+
+	# ns_src
+	ip -n "${NS_SRC}" link set veth-s0 up
+	ip -n "${NS_SRC}" addr add fd00::1/64 dev veth-s0 nodad
+	ip -n "${NS_SRC}" -6 route add "${DEST}"/128 via fd00::2
+
+	# ns_router
+	ip -n "${NS_RTR}" link set veth-r0 up
+	ip -n "${NS_RTR}" addr add fd00::2/64 dev veth-r0 nodad
+	ip -n "${NS_RTR}" link set veth-r1 up
+	ip -n "${NS_RTR}" addr add fd01::1/64 dev veth-r1 nodad
+	ip netns exec "${NS_RTR}" sysctl -qw net.ipv6.conf.all.forwarding=1
+
+	ip -n "${NS_RTR}" -6 route add "${DEST}"/128 \
+		encap seg6 mode encap segs "${SID}" dev veth-r0
+	ip -n "${NS_RTR}" -6 route add "${SID}"/128 table 100 \
+		via fd01::2 dev veth-r1
+	ip -n "${NS_RTR}" -6 route add blackhole "${SID}"/128
+	ip -n "${NS_RTR}" -6 rule add iif veth-r0 lookup 100
+
+	# ns_dst
+	ip -n "${NS_DST}" link set veth-d0 up
+	ip -n "${NS_DST}" addr add fd01::2/64 dev veth-d0 nodad
+	ip -n "${NS_DST}" addr add "${DEST}"/128 dev lo nodad
+	ip -n "${NS_DST}" -6 route add "${SID}"/128 \
+		encap seg6local action End.DT6 table 255 dev veth-d0
+	ip -n "${NS_DST}" -6 route add fd00::/64 via fd01::1
+
+	# static neighbors
+	ip -n "${NS_SRC}" -6 neigh add fd00::2 dev veth-s0 \
+		lladdr "${RTR_R0_MAC}" nud permanent
+	ip -n "${NS_RTR}" -6 neigh add fd00::1 dev veth-r0 \
+		lladdr "${SRC_MAC}" nud permanent
+	ip -n "${NS_RTR}" -6 neigh add fd01::2 dev veth-r1 \
+		lladdr "${DST_MAC}" nud permanent
+	ip -n "${NS_DST}" -6 neigh add fd01::1 dev veth-d0 \
+		lladdr "${RTR_R1_MAC}" nud permanent
+}
+
+test_cache_isolation()
+{
+	RET=0
+
+	# local ping with empty cache: must fail (SID is blackholed)
+	if ip netns exec "${NS_RTR}" taskset -c 0 \
+			ping -c 1 -W 2 "${DEST}" &>/dev/null; then
+		echo "SKIP: local ping succeeded, topology broken"
+		exit "${ksft_skip}"
+	fi
+
+	# forward from ns_src to populate the input cache
+	if ! ip netns exec "${NS_SRC}" taskset -c 0 \
+			ping -c 1 -W 2 "${DEST}" &>/dev/null; then
+		echo "SKIP: forwarded ping failed, topology broken"
+		exit "${ksft_skip}"
+	fi
+
+	# local ping again: must still fail; if the output path reuses
+	# the input cache, it bypasses the blackhole and the ping succeeds
+	if ip netns exec "${NS_RTR}" taskset -c 0 \
+			ping -c 1 -W 2 "${DEST}" &>/dev/null; then
+		echo "FAIL: output path used dst cached by input path"
+		RET="${ksft_fail}"
+	else
+		echo "PASS: output path dst_cache is independent"
+	fi
+
+	return "${RET}"
+}
+
+if [ "$(id -u)" -ne 0 ]; then
+	echo "SKIP: Need root privileges"
+	exit "${ksft_skip}"
+fi
+
+trap cleanup EXIT
+
+check_prerequisites
+setup
+test_cache_isolation
+exit "${RET}"
-- 
2.43.0

Re: [PATCH net v3 1/2] seg6: separate dst_cache for input and output paths in seg6 lwtunnel

[Jakub Kicinski]

This is an AI-generated review of your patch. The human sending this
email says "Seems orthogonal but also legit"
---
seg6: separate dst_cache for input and output paths in seg6 lwtunnel

This commit splits the shared seg6 lwtunnel dst_cache into separate input
and output caches to prevent blind reuse across different routing contexts.

> diff --git a/net/ipv6/seg6_iptunnel.c b/net/ipv6/seg6_iptunnel.c
> index 3e1b9991131a2..d6a0f7df90807 100644
> --- a/net/ipv6/seg6_iptunnel.c
> +++ b/net/ipv6/seg6_iptunnel.c

[ ... ]

> @@ -504,7 +505,7 @@ static int seg6_input_core(struct net *net, struct sock *sk,
>  		/* cache only if we don't create a dst reference loop */
>  		if (!dst->error && lwtst != dst->lwtstate) {
>  			local_bh_disable();
> -			dst_cache_set_ip6(&slwt->cache, dst,
> +			dst_cache_set_ip6(&slwt->cache_input, dst,
>  					  &ipv6_hdr(skb)->saddr);
>  			local_bh_enable();
>  		}

This code appears to cache the dst entry resolved just prior by calling
ip6_route_input(skb).

Does ip6_route_input() assign a NOREF dst to the skb?

If so, this dst is protected only by RCU. If the underlying route is
concurrently deleted, its refcount can drop to zero. When
dst_cache_set_ip6() is called, it attempts to use dst_hold(dst).

If the refcount is already zero, dst_hold() will warn and skip incrementing
the refcount, but dst_cache_set_ip6() will still store the pointer.

Could this cause the cache to hold a dangling pointer, leading to a
use-after-free once the RCU grace period expires?

Should this path use skb_dst_force(skb) or dst_hold_safe(dst) to safely
upgrade the dst to a refcounted reference before it is cached?
-- 
pw-bot: cr

Re: [PATCH net v3 0/2] seg6: fix dst_cache sharing in seg6 lwtunnel

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Sat,  4 Apr 2026 02:44:03 +0200 you wrote:
> The seg6 lwtunnel encap uses a single per-route dst_cache shared
> between seg6_input_core() and seg6_output_core(). These two paths
> can perform the post-encap SID lookup in different routing contexts
> (e.g., ip rules matching on the ingress interface, or VRF table
> separation). Whichever path runs first populates the cache, and the
> other reuses it blindly, bypassing its own lookup.
> 
> [...]

Here is the summary with links:
  - [net,v3,1/2] seg6: separate dst_cache for input and output paths in seg6 lwtunnel
    https://git.kernel.org/netdev/net/c/c3812651b522
  - [net,v3,2/2] selftests: seg6: add test for dst_cache isolation in seg6 lwtunnel
    https://git.kernel.org/netdev/net/c/32dfd742f06a

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html