Skip to contentSkip to conversation

net: fix lwtunnel reentry loops

PATCH Applied Justin Iurman 5 participants 11 messages Mar 14, 2025 – Apr 16, 2025

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

v2 cover net: fix lwtunnel reentry loops

v2:
- removed some patches from the -v1 series
- added a patch that was initially sent separately
- code style for the selftest (thanks Paolo)
v1:
- https://lore.kernel.org/all/[email protected]/

When the destination is the same after the transformation, we enter a lwtunnel loop. This is true for most of lwt users: ioam6, rpl, seg6, seg6_local, ila_lwt, and lwt_bpf. It can happen in their input() and output() handlers respectively, where either dst_input() or dst_output() is called at the end. It can also happen in xmit() handlers.

Here is an example for rpl_input():

dump_stack_lvl+0x60/0x80
rpl_input+0x9d/0x320
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
[...]
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
lwtunnel_input+0x64/0xa0
ip6_sublist_rcv_finish+0x85/0x90
ip6_sublist_rcv+0x236/0x2f0

... until rpl_do_srh() fails, which means skb_cow_head() failed.

This series provides a fix at the core level of lwtunnel to catch such loops when they're not caught by the respective lwtunnel users, and handle the loop case in ioam6 which is one of the users. This series also comes with a new selftest to detect some dst cache reference loops in lwtunnel users.

Justin Iurman (3):
net: lwtunnel: fix recursion loops
net: ipv6: ioam6: fix lwtunnel_output() loop
selftests: net: test for lwtunnel dst ref loops

 net/core/lwtunnel.c                           |  65 ++++-
 net/ipv6/ioam6_iptunnel.c                     |   8 +-
 tools/testing/selftests/net/Makefile          |   1 +
 tools/testing/selftests/net/config            |   2 +
 .../selftests/net/lwt_dst_cache_ref_loop.sh   | 246 ++++++++++++++++++
 5 files changed, 306 insertions(+), 16 deletions(-)
 create mode 100755 tools/testing/selftests/net/lwt_dst_cache_ref_loop.sh
Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 1/3 net: lwtunnel: fix recursion loops

This patch acts as a parachute, catch all solution, by detecting recursion loops in lwtunnel users and taking care of them (e.g., a loop between routes, a loop within the same route, etc). In general, such loops are the consequence of pathological configurations. Each lwtunnel user is still free to catch such loops early and do whatever they want with them. It will be the case in a separate patch for, e.g., seg6 and seg6_local, in order to provide drop reasons and update statistics. Another example of a lwtunnel user taking care of loops is ioam6, which has valid use cases that include loops (e.g., inline mode), and which is addressed by the next patch in this series. Overall, this patch acts as a last resort to catch loops and drop packets, since we don't want to leak something unintentionally because of a pathological configuration in lwtunnels.

The solution in this patch reuses dev_xmit_recursion(), dev_xmit_recursion_inc(), and dev_xmit_recursion_dec(), which seems fine considering the context.

Fixes ffce41962ef6 ("lwtunnel: support dst output redirect function")
Fixes 2536862311d2 ("lwt: Add support to redirect dst.input")
Fixes 14972cbd34ff ("net: lwtunnel: Handle fragmentation")

---
Cc: Roopa Prabhu <[email protected]>
Cc: Andrea Mayer <[email protected]>
Cc: Stefano Salsano <[email protected]>
Cc: Ahmed Abdelsalam <[email protected]>
Cc: Ido Schimmel <[email protected]>

---
 net/core/lwtunnel.c | 65 ++++++++++++++++++++++++++++++++++++---------
 1 file changed, 53 insertions(+), 12 deletions(-)
net/core/lwtunnel.c +53 -12
index 711cd3b4347a..4417a18b3e95 100644
@@ -23,6 +23,8 @@
 #include <net/ip6_fib.h>
 #include <net/rtnh.h>
 
+#include "dev.h"
+
 DEFINE_STATIC_KEY_FALSE(nf_hooks_lwtunnel_enabled);
 EXPORT_SYMBOL_GPL(nf_hooks_lwtunnel_enabled);
 
@@ -325,13 +327,23 @@ EXPORT_SYMBOL_GPL(lwtunnel_cmp_encap);
 
 int lwtunnel_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 {
-	struct dst_entry *dst = skb_dst(skb);
 	const struct lwtunnel_encap_ops *ops;
 	struct lwtunnel_state *lwtstate;
-	int ret = -EINVAL;
+	struct dst_entry *dst;
+	int ret;
+
+	if (dev_xmit_recursion()) {
+		net_crit_ratelimited("%s(): recursion limit reached on datapath\n",
+				     __func__);
+		ret = -ENETDOWN;
+		goto drop;
+	}
 
-	if (!dst)
+	dst = skb_dst(skb);
+	if (!dst) {
+		ret = -EINVAL;
 		goto drop;
+	}
 	lwtstate = dst->lwtstate;
 
 	if (lwtstate->type == LWTUNNEL_ENCAP_NONE ||
@@ -341,8 +353,11 @@ int lwtunnel_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 	ret = -EOPNOTSUPP;
 	rcu_read_lock();
 	ops = rcu_dereference(lwtun_encaps[lwtstate->type]);
-	if (likely(ops && ops->output))
+	if (likely(ops && ops->output)) {
+		dev_xmit_recursion_inc();
 		ret = ops->output(net, sk, skb);
+		dev_xmit_recursion_dec();
+	}
 	rcu_read_unlock();
 
 	if (ret == -EOPNOTSUPP)
@@ -359,13 +374,23 @@ EXPORT_SYMBOL_GPL(lwtunnel_output);
 
 int lwtunnel_xmit(struct sk_buff *skb)
 {
-	struct dst_entry *dst = skb_dst(skb);
 	const struct lwtunnel_encap_ops *ops;
 	struct lwtunnel_state *lwtstate;
-	int ret = -EINVAL;
+	struct dst_entry *dst;
+	int ret;
+
+	if (dev_xmit_recursion()) {
+		net_crit_ratelimited("%s(): recursion limit reached on datapath\n",
+				     __func__);
+		ret = -ENETDOWN;
+		goto drop;
+	}
 
-	if (!dst)
+	dst = skb_dst(skb);
+	if (!dst) {
+		ret = -EINVAL;
 		goto drop;
+	}
 
 	lwtstate = dst->lwtstate;
 
@@ -376,8 +401,11 @@ int lwtunnel_xmit(struct sk_buff *skb)
 	ret = -EOPNOTSUPP;
 	rcu_read_lock();
 	ops = rcu_dereference(lwtun_encaps[lwtstate->type]);
-	if (likely(ops && ops->xmit))
+	if (likely(ops && ops->xmit)) {
+		dev_xmit_recursion_inc();
 		ret = ops->xmit(skb);
+		dev_xmit_recursion_dec();
+	}
 	rcu_read_unlock();
 
 	if (ret == -EOPNOTSUPP)
@@ -394,13 +422,23 @@ EXPORT_SYMBOL_GPL(lwtunnel_xmit);
 
 int lwtunnel_input(struct sk_buff *skb)
 {
-	struct dst_entry *dst = skb_dst(skb);
 	const struct lwtunnel_encap_ops *ops;
 	struct lwtunnel_state *lwtstate;
-	int ret = -EINVAL;
+	struct dst_entry *dst;
+	int ret;
 
-	if (!dst)
+	if (dev_xmit_recursion()) {
+		net_crit_ratelimited("%s(): recursion limit reached on datapath\n",
+				     __func__);
+		ret = -ENETDOWN;
 		goto drop;
+	}
+
+	dst = skb_dst(skb);
+	if (!dst) {
+		ret = -EINVAL;
+		goto drop;
+	}
 	lwtstate = dst->lwtstate;
 
 	if (lwtstate->type == LWTUNNEL_ENCAP_NONE ||
@@ -410,8 +448,11 @@ int lwtunnel_input(struct sk_buff *skb)
 	ret = -EOPNOTSUPP;
 	rcu_read_lock();
 	ops = rcu_dereference(lwtun_encaps[lwtstate->type]);
-	if (likely(ops && ops->input))
+	if (likely(ops && ops->input)) {
+		dev_xmit_recursion_inc();
 		ret = ops->input(skb);
+		dev_xmit_recursion_dec();
+	}
 	rcu_read_unlock();
 
 	if (ret == -EOPNOTSUPP)
Signature · 1 line
2.34.1
#2 lore raw

PATCH v2 2/3 net: ipv6: ioam6: fix lwtunnel_output() loop

In reply to Justin Iurman

Fix the lwtunnel_output() reentry loop in ioam6_iptunnel when the destination is the same after transformation. Note that a check on the destination address was already performed, but it was not enough. This is the example of a lwtunnel user taking care of loops without relying only on the last resort detection offered by lwtunnel.

Fixes 8cb3bf8bff3c ("ipv6: ioam: Add support for the ip6ip6 encapsulation")

---
Cc: Ido Schimmel <[email protected]>

---
 net/ipv6/ioam6_iptunnel.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
net/ipv6/ioam6_iptunnel.c +4 -4
index 2c383c12a431..09065187378e 100644
@@ -337,7 +337,6 @@ static int ioam6_do_encap(struct net *net, struct sk_buff *skb,
 static int ioam6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 {
 	struct dst_entry *dst = skb_dst(skb), *cache_dst = NULL;
-	struct in6_addr orig_daddr;
 	struct ioam6_lwt *ilwt;
 	int err = -EINVAL;
 	u32 pkt_cnt;
@@ -352,8 +351,6 @@ static int ioam6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 	if (pkt_cnt % ilwt->freq.n >= ilwt->freq.k)
 		goto out;
 
-	orig_daddr = ipv6_hdr(skb)->daddr;
-
 	local_bh_disable();
 	cache_dst = dst_cache_get(&ilwt->cache);
 	local_bh_enable();
@@ -422,7 +419,10 @@ static int ioam6_output(struct net *net, struct sock *sk, struct sk_buff *skb)
 			goto drop;
 	}
 
-	if (!ipv6_addr_equal(&orig_daddr, &ipv6_hdr(skb)->daddr)) {
+	/* avoid lwtunnel_output() reentry loop when destination is the same
+	 * after transformation (e.g., with the inline mode)
+	 */
+	if (dst->lwtstate != cache_dst->lwtstate) {
 		skb_dst_drop(skb);
 		skb_dst_set(skb, cache_dst);
 		return dst_output(net, sk, skb);
Signature · 1 line
2.34.1
#3 lore raw

PATCH v2 3/3 selftests: net: test for lwtunnel dst ref loops

In reply to Justin Iurman

As recently specified by commit 0ea09cbf8350 ("docs: netdev: add a note on selftest posting") in net-next, the selftest is therefore shipped in this series. However, this selftest does not really test this series. It needs this series to avoid crashing the kernel. What it really tests, thanks to kmemleak, is what was fixed by the following commits:
- commit c71a192976de ("net: ipv6: fix dst refleaks in rpl, seg6 and ioam6 lwtunnels")
- commit 92191dd10730 ("net: ipv6: fix dst ref loops in rpl, seg6 and ioam6 lwtunnels")
- commit c64a0727f9b1 ("net: ipv6: fix dst ref loop on input in seg6 lwt")
- commit 13e55fbaec17 ("net: ipv6: fix dst ref loop on input in rpl lwt")
- commit 0e7633d7b95b ("net: ipv6: fix dst ref loop in ila lwtunnel")
- commit 5da15a9c11c1 ("net: ipv6: fix missing dst ref drop in ila lwtunnel")

---
Cc: Shuah Khan <[email protected]>
Cc: [email protected]

---
 tools/testing/selftests/net/Makefile          |   1 +
 tools/testing/selftests/net/config            |   2 +
 .../selftests/net/lwt_dst_cache_ref_loop.sh   | 246 ++++++++++++++++++
 3 files changed, 249 insertions(+)
 create mode 100755 tools/testing/selftests/net/lwt_dst_cache_ref_loop.sh
tools/testing/selftests/net/Makefile +1 -0
index 5916f3b81c39..843ab747645d 100644
@@ -101,6 +101,7 @@ TEST_PROGS += vlan_bridge_binding.sh
 TEST_PROGS += bpf_offload.py
 TEST_PROGS += ipv6_route_update_soft_lockup.sh
 TEST_PROGS += busy_poll_test.sh
+TEST_PROGS += lwt_dst_cache_ref_loop.sh
 
 # YNL files, must be before "include ..lib.mk"
 YNL_GEN_FILES := busy_poller netlink-dumps
tools/testing/selftests/net/config +2 -0
index 5b9baf708950..61e5116987f3 100644
@@ -107,3 +107,5 @@ CONFIG_XFRM_INTERFACE=m
 CONFIG_XFRM_USER=m
 CONFIG_IP_NF_MATCH_RPFILTER=m
 CONFIG_IP6_NF_MATCH_RPFILTER=m
+CONFIG_IPV6_ILA=m
+CONFIG_IPV6_RPL_LWTUNNEL=y
tools/testing/selftests/net/lwt_dst_cache_ref_loop.sh new file +246 -0

246 lines not on this page · raw message

Signature · 1 line
2.34.1
maintainer
#4 lore raw

Re: [PATCH net v2 3/3] selftests: net: test for lwtunnel dst ref loops

Quoted 5 lines · Justin Iurman

As recently specified by commit 0ea09cbf8350 ("docs: netdev: add a note on selftest posting") in net-next, the selftest is therefore shipped in this series. However, this selftest does not really test this series. It needs this series to avoid crashing the kernel. What it really tests, thanks to kmemleak,

As a net-next follow-up you could force a kmemleak scan and check the result after each test case to really output a pass/fail message.

Also, still for net-next, please investigate if dropping or reducing the many sleep below could be possible (it's not clear to me why they are needed).

I'll take is as-is to avoid blocking the fixes for trivial matters.

Thanks,

Paolo

#5 lore raw

Re: [PATCH net v2 3/3] selftests: net: test for lwtunnel dst ref loops

Quoted 9 lines · Paolo Abeni
Quoted 5 lines · Justin Iurman

As recently specified by commit 0ea09cbf8350 ("docs: netdev: add a note on selftest posting") in net-next, the selftest is therefore shipped in this series. However, this selftest does not really test this series. It needs this series to avoid crashing the kernel. What it really tests, thanks to kmemleak,

As a net-next follow-up you could force a kmemleak scan and check the result after each test case to really output a pass/fail message.

Sure. Could you point me to an example doing that? Not sure how I'd force it within the selftest.

Quoted 3 lines

Also, still for net-next, please investigate if dropping or reducing the many sleep below could be possible (it's not clear to me why they are needed).

Of course, will do.

Quoted 1 line

I'll take is as-is to avoid blocking the fixes for trivial matters.

Thanks,
Justin

Quoted 4 lines

Thanks,

Paolo

#6 lore raw

Re: [PATCH net v2 3/3] selftests: net: test for lwtunnel dst ref loops: manual merge

In reply to Justin Iurman

Hi Justin,

Quoted 45 lines · Justin Iurman

As recently specified by commit 0ea09cbf8350 ("docs: netdev: add a note on selftest posting") in net-next, the selftest is therefore shipped in this series. However, this selftest does not really test this series. It needs this series to avoid crashing the kernel. What it really tests, thanks to kmemleak, is what was fixed by the following commits:
- commit c71a192976de ("net: ipv6: fix dst refleaks in rpl, seg6 and ioam6 lwtunnels")
- commit 92191dd10730 ("net: ipv6: fix dst ref loops in rpl, seg6 and ioam6 lwtunnels")
- commit c64a0727f9b1 ("net: ipv6: fix dst ref loop on input in seg6 lwt")
- commit 13e55fbaec17 ("net: ipv6: fix dst ref loop on input in rpl lwt")
- commit 0e7633d7b95b ("net: ipv6: fix dst ref loop in ila lwtunnel")
- commit 5da15a9c11c1 ("net: ipv6: fix missing dst ref drop in ila lwtunnel")

---
Cc: Shuah Khan <[email protected]>
Cc: [email protected]

---
 tools/testing/selftests/net/Makefile          |   1 +
 tools/testing/selftests/net/config            |   2 +
 .../selftests/net/lwt_dst_cache_ref_loop.sh   | 246 ++++++++++++++++++
 3 files changed, 249 insertions(+)
 create mode 100755 tools/testing/selftests/net/lwt_dst_cache_ref_loop.sh
tools/testing/selftests/net/Makefile +1 -0
index 5916f3b81c39..843ab747645d 100644
@@ -101,6 +101,7 @@ TEST_PROGS += vlan_bridge_binding.sh
 TEST_PROGS += bpf_offload.py
 TEST_PROGS += ipv6_route_update_soft_lockup.sh
 TEST_PROGS += busy_poll_test.sh
+TEST_PROGS += lwt_dst_cache_ref_loop.sh
 
 # YNL files, must be before "include ..lib.mk"
 YNL_GEN_FILES := busy_poller netlink-dumps
tools/testing/selftests/net/config +0 -0
index 5b9baf708950..61e5116987f3 100644

tools/testing/selftests/net/config CONFIG_XFRM_INTERFACE=m · line 111

Quoted 5 lines
tools/testing/selftests/net/config +2 -0
@@ -107,3 +107,5 @@ CONFIG_XFRM_INTERFACE=m
 CONFIG_XFRM_USER=m
 CONFIG_IP_NF_MATCH_RPFILTER=m
 CONFIG_IP6_NF_MATCH_RPFILTER=m
+CONFIG_IPV6_ILA=m
+CONFIG_IPV6_RPL_LWTUNNEL=y

FYI, we got a small conflict when merging 'net' in 'net-next' in the MPTCP tree due to this patch applied in 'net':

3ed61b8938c6 ("selftests: net: test for lwtunnel dst ref loops")

and these ones from 'net-next':

03544faad761 ("selftest: net: add proc_net_pktgen") 85cb3711acb8 ("selftests: net: Add test cases for link and peer netns")

----- Generic Message -----
The best is to avoid conflicts between 'net' and 'net-next' trees but if they cannot be avoided when preparing patches, a note about how to fix them is much appreciated.

The conflict has been resolved on our side[1] and the resolution we suggest is attached to this email. Please report any issues linked to this conflict resolution as it might be used by others. If you worked on the mentioned patches, don't hesitate to ACK this conflict resolution.
---------------------------

Regarding these conflicts, that's because the new test and configs have been added at the end. It is usually recommended to avoid that for long list, e.g. inserting new entries at the right place if the list is sorted by alphabetical order (not the case here), grouping by theme, or keeping the same order as in the .config, etc. For the resolutions here, I simply kept the new tests and configs from both sides at the same place, adding the new ones from net at the end.

Rerere cache is available in [2].

Cheers,
Matt

[1] https://github.com/multipath-tcp/mptcp_net-next/commit/89220b9164c7
[2] https://github.com/multipath-tcp/mptcp-upstream-rr-cache/commit/d240

Cheers,
Matt

Signature · 1 line
Sponsored by the NGI0 Core fund.
tools/testing/selftests/net/Makefile +0 -0
diff --cc tools/testing/selftests/net/Makefile
index f03a0399e7a3,8f32b4f01aee..6d718b478ed8
@@@ -31,15 -31,11 +31,14 @@@ TEST_PROGS += veth.s
  TEST_PROGS += ioam6.sh
  TEST_PROGS += gro.sh
  TEST_PROGS += gre_gso.sh
- TEST_PROGS += gre_ipv6_lladdr.sh
  TEST_PROGS += cmsg_so_mark.sh
  TEST_PROGS += cmsg_so_priority.sh
 -TEST_PROGS += cmsg_time.sh cmsg_ipv6.sh
 +TEST_PROGS += test_so_rcv.sh
 +TEST_PROGS += cmsg_time.sh cmsg_ip.sh
  TEST_PROGS += netns-name.sh
 +TEST_PROGS += link_netns.py
  TEST_PROGS += nl_netdev.py
 +TEST_PROGS += rtnetlink.py
  TEST_PROGS += srv6_end_dt46_l3vpn_test.sh
  TEST_PROGS += srv6_end_dt4_l3vpn_test.sh
  TEST_PROGS += srv6_end_dt6_l3vpn_test.sh
@@@ -105,7 -100,7 +104,8 @@@ TEST_PROGS += vlan_bridge_binding.s
  TEST_PROGS += bpf_offload.py
  TEST_PROGS += ipv6_route_update_soft_lockup.sh
  TEST_PROGS += busy_poll_test.sh
 +TEST_GEN_PROGS += proc_net_pktgen
+ TEST_PROGS += lwt_dst_cache_ref_loop.sh
  
  # YNL files, must be before "include ..lib.mk"
  YNL_GEN_FILES := busy_poller netlink-dumps
tools/testing/selftests/net/config +0 -0
diff --cc tools/testing/selftests/net/config
index b0d0eda829d0,61e5116987f3..130d532b7e67
@@@ -109,9 -107,5 +109,11 @@@ CONFIG_XFRM_INTERFACE=
  CONFIG_XFRM_USER=m
  CONFIG_IP_NF_MATCH_RPFILTER=m
  CONFIG_IP6_NF_MATCH_RPFILTER=m
 +CONFIG_IPVLAN=m
 +CONFIG_CAN=m
 +CONFIG_CAN_DEV=m
 +CONFIG_CAN_VXCAN=m
 +CONFIG_NETKIT=y
 +CONFIG_NET_PKTGEN=m
+ CONFIG_IPV6_ILA=m
+ CONFIG_IPV6_RPL_LWTUNNEL=y
#7 lore raw
In reply to Justin Iurman
This series was applied to netdev/net.git (main) by Paolo Abeni <[email protected]>:

Hello:

This series was applied to netdev/net.git (main)
by Paolo Abeni <[email protected]>:

Quoted 8 lines · On Fri, 14 Mar 2025 13:00:45 +0100 you wrote:

v2:
- removed some patches from the -v1 series
- added a patch that was initially sent separately
- code style for the selftest (thanks Paolo)
v1:
- https://lore.kernel.org/all/[email protected]/

[...]

Here is the summary with links:
- [net,v2,1/3] net: lwtunnel: fix recursion loops

    https://git.kernel.org/netdev/net/c/986ffb3a57c5
  - [net,v2,2/3] net: ipv6: ioam6: fix lwtunnel_output() loop
    https://git.kernel.org/netdev/net/c/3e7a60b368ea
  - [net,v2,3/3] selftests: net: test for lwtunnel dst ref loops
    https://git.kernel.org/netdev/net/c/3ed61b8938c6

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
#8 lore raw
In reply to Justin Iurman
Quoted 6 lines · Justin Iurman

v2:
- removed some patches from the -v1 series
- added a patch that was initially sent separately
- code style for the selftest (thanks Paolo)
v1:
- https://lore.kernel.org/all/[email protected]/

Hi Justin,

I've noticed a BUG splat likely introduced by this patch. The splat is reported when executing some BPF selftests, e.g. lwt_ip_encap_ipv4/egress
(defined in tools/testing/selftests/bpf/prog_tests/lwt_ip_encap.c and
tools/testing/selftests/bpf/progs/test_lwt_ip_encap.c).

Decoded splat is at the end of the email.
Line numbers correspond to commit
a27a97f71394 ("Merge branch 'bpf-support-atomic-update-for-htab-of-maps'") from the kernel/git/bpf/bpf-next.git tree.

Thanks,
Eduard

---

[  193.993893] BUG: using __this_cpu_add() in preemptible [00000000] code: test_progs/206
[  193.994292] caller is lwtunnel_xmit (net/core/dev.h:340 net/core/lwtunnel.c:408) 
[  193.994601] Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
[  193.994603] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-4.el9 04/01/2014
[  193.994605] Call Trace:
[  193.994608]  <TASK>
[  193.994611] dump_stack_lvl (lib/dump_stack.c:122) 
[  193.994622] check_preemption_disabled (lib/smp_processor_id.c:0) 
[  193.994630] ? lwtunnel_xmit (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 net/core/lwtunnel.c:403) 
[  193.994637] lwtunnel_xmit (net/core/dev.h:340 net/core/lwtunnel.c:408) 
[  193.994648] ip_finish_output2 (net/ipv4/ip_output.c:222) 
[  193.994655] ? ip_skb_dst_mtu (./include/net/ip.h:517) 
[  193.994659] ? ip_skb_dst_mtu (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 ./include/net/ip.h:471 ./include/net/ip.h:512) 
[  193.994669] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182) 
[  193.994672] ? __ip_finish_output (net/ipv4/ip_output.c:306) 
[  193.994683] ? __ip_queue_xmit (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 net/ipv4/ip_output.c:470) 
[  193.994688] __ip_queue_xmit (net/ipv4/ip_output.c:527) 
[  193.994693] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182) 
[  193.994711] ? __ip_queue_xmit (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 net/ipv4/ip_output.c:470) 
[  193.994726] __tcp_transmit_skb (net/ipv4/tcp_output.c:1479) 
[  193.994800] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182) 
[  193.994804] ? __asan_memset (mm/kasan/shadow.c:84) 
[  193.994810] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182) 
[  193.994824] tcp_connect (net/ipv4/tcp_output.c:0 net/ipv4/tcp_output.c:4155) 
[  193.994890] tcp_v4_connect (net/ipv4/tcp_ipv4.c:343) 
[  193.994926] __inet_stream_connect (net/ipv4/af_inet.c:678) 
[  193.994944] ? __local_bh_enable_ip (./arch/x86/include/asm/irqflags.h:42 ./arch/x86/include/asm/irqflags.h:119 kernel/softirq.c:412) 
[  193.994950] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182) 
[  193.994953] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4473) 
[  193.994967] inet_stream_connect (net/ipv4/af_inet.c:748) 
[  193.994976] ? __pfx_inet_stream_connect (net/ipv4/af_inet.c:744) 
[  193.994981] __sys_connect (./include/linux/file.h:62 ./include/linux/file.h:83 net/socket.c:2058) 
[  193.995013] __x64_sys_connect (net/socket.c:2063 net/socket.c:2060 net/socket.c:2060) 
[  193.995022] do_syscall_64 (arch/x86/entry/syscall_64.c:0) 
[  193.995026] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182) 
[  193.995030] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4473) 
[  193.995038] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130) 
[  193.995042] RIP: 0033:0x7faec2d0f9cb
[ 193.995047] Code: 83 ec 18 89 54 24 0c 48 89 34 24 89 7c 24 08 e8 4b 70 f7 ff 8b 54 24 0c 48 8b 34 24 41 89 c0 8b 7c 24 08 b8 2a 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 35 44 89 c7 89 44 24 08 e8 a1 70 f7 ff 8b 44
All code
========
   0:	83 ec 18             	sub    $0x18,%esp
   3:	89 54 24 0c          	mov    %edx,0xc(%rsp)
   7:	48 89 34 24          	mov    %rsi,(%rsp)
   b:	89 7c 24 08          	mov    %edi,0x8(%rsp)
   f:	e8 4b 70 f7 ff       	call   0xfffffffffff7705f
  14:	8b 54 24 0c          	mov    0xc(%rsp),%edx
  18:	48 8b 34 24          	mov    (%rsp),%rsi
  1c:	41 89 c0             	mov    %eax,%r8d
  1f:	8b 7c 24 08          	mov    0x8(%rsp),%edi
  23:	b8 2a 00 00 00       	mov    $0x2a,%eax
  28:	0f 05                	syscall
  2a:*	48 3d 00 f0 ff ff    	cmp    $0xfffffffffffff000,%rax		<-- trapping instruction
  30:	77 35                	ja     0x67
  32:	44 89 c7             	mov    %r8d,%edi
  35:	89 44 24 08          	mov    %eax,0x8(%rsp)
  39:	e8 a1 70 f7 ff       	call   0xfffffffffff770df
  3e:	8b                   	.byte 0x8b
  3f:	44                   	rex.R

Code starting with the faulting instruction
===========================================

   0:	48 3d 00 f0 ff ff    	cmp    $0xfffffffffffff000,%rax
   6:	77 35                	ja     0x3d
   8:	44 89 c7             	mov    %r8d,%edi
   b:	89 44 24 08          	mov    %eax,0x8(%rsp)
   f:	e8 a1 70 f7 ff       	call   0xfffffffffff770b5
  14:	8b                   	.byte 0x8b
  15:	44                   	rex.R
[  193.995050] RSP: 002b:00007fff992d3a20 EFLAGS: 00000293 ORIG_RAX: 000000000000002a
[  193.995054] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007faec2d0f9cb
[  193.995057] RDX: 0000000000000010 RSI: 00007fff992d3ad8 RDI: 0000000000000035
[  193.995059] RBP: 00007fff992d3ac0 R08: 0000000000000000 R09: 0000000000000004
[  193.995062] R10: 00007fff992d39b0 R11: 0000000000000293 R12: 00007fff992d7b78
[  193.995064] R13: 000000000095f760 R14: 0000000002e38b90 R15: 00007faec373d000
[  193.995091]  </TASK>
#9 lore raw
Quoted 17 lines · Eduard Zingerman
Quoted 6 lines · Justin Iurman

v2:
- removed some patches from the -v1 series
- added a patch that was initially sent separately
- code style for the selftest (thanks Paolo)
v1:
- https://lore.kernel.org/all/[email protected]/

Hi Justin,

I've noticed a BUG splat likely introduced by this patch. The splat is reported when executing some BPF selftests, e.g. lwt_ip_encap_ipv4/egress
(defined in tools/testing/selftests/bpf/prog_tests/lwt_ip_encap.c and
tools/testing/selftests/bpf/progs/test_lwt_ip_encap.c).

Quoted 4 lines

Decoded splat is at the end of the email.
Line numbers correspond to commit
a27a97f71394 ("Merge branch 'bpf-support-atomic-update-for-htab-of-maps'") from the kernel/git/bpf/bpf-next.git tree.

Hi Eduard,

This is currently discussed in [1]. We have a solution, but no consensus on the fix yet.

[1]
https://lore.kernel.org/netdev/[email protected]/T/#t

Thanks for the report,
Justin

Quoted 82 lines

Thanks,
Eduard

---

[  193.993893] BUG: using __this_cpu_add() in preemptible [00000000] code: test_progs/206
[  193.994292] caller is lwtunnel_xmit (net/core/dev.h:340 net/core/lwtunnel.c:408)
[  193.994601] Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
[  193.994603] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-4.el9 04/01/2014
[  193.994605] Call Trace:
[  193.994608]  <TASK>
[  193.994611] dump_stack_lvl (lib/dump_stack.c:122)
[  193.994622] check_preemption_disabled (lib/smp_processor_id.c:0)
[  193.994630] ? lwtunnel_xmit (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 net/core/lwtunnel.c:403)
[  193.994637] lwtunnel_xmit (net/core/dev.h:340 net/core/lwtunnel.c:408)
[  193.994648] ip_finish_output2 (net/ipv4/ip_output.c:222)
[  193.994655] ? ip_skb_dst_mtu (./include/net/ip.h:517)
[  193.994659] ? ip_skb_dst_mtu (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 ./include/net/ip.h:471 ./include/net/ip.h:512)
[  193.994669] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182)
[  193.994672] ? __ip_finish_output (net/ipv4/ip_output.c:306)
[  193.994683] ? __ip_queue_xmit (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 net/ipv4/ip_output.c:470)
[  193.994688] __ip_queue_xmit (net/ipv4/ip_output.c:527)
[  193.994693] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182)
[  193.994711] ? __ip_queue_xmit (./include/linux/rcupdate.h:331 ./include/linux/rcupdate.h:841 net/ipv4/ip_output.c:470)
[  193.994726] __tcp_transmit_skb (net/ipv4/tcp_output.c:1479)
[  193.994800] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182)
[  193.994804] ? __asan_memset (mm/kasan/shadow.c:84)
[  193.994810] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182)
[  193.994824] tcp_connect (net/ipv4/tcp_output.c:0 net/ipv4/tcp_output.c:4155)
[  193.994890] tcp_v4_connect (net/ipv4/tcp_ipv4.c:343)
[  193.994926] __inet_stream_connect (net/ipv4/af_inet.c:678)
[  193.994944] ? __local_bh_enable_ip (./arch/x86/include/asm/irqflags.h:42 ./arch/x86/include/asm/irqflags.h:119 kernel/softirq.c:412)
[  193.994950] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182)
[  193.994953] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4473)
[  193.994967] inet_stream_connect (net/ipv4/af_inet.c:748)
[  193.994976] ? __pfx_inet_stream_connect (net/ipv4/af_inet.c:744)
[  193.994981] __sys_connect (./include/linux/file.h:62 ./include/linux/file.h:83 net/socket.c:2058)
[  193.995013] __x64_sys_connect (net/socket.c:2063 net/socket.c:2060 net/socket.c:2060)
[  193.995022] do_syscall_64 (arch/x86/entry/syscall_64.c:0)
[  193.995026] ? srso_alias_return_thunk (arch/x86/lib/retpoline.S:182)
[  193.995030] ? lockdep_hardirqs_on (kernel/locking/lockdep.c:4473)
[  193.995038] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
[  193.995042] RIP: 0033:0x7faec2d0f9cb
[ 193.995047] Code: 83 ec 18 89 54 24 0c 48 89 34 24 89 7c 24 08 e8 4b 70 f7 ff 8b 54 24 0c 48 8b 34 24 41 89 c0 8b 7c 24 08 b8 2a 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 35 44 89 c7 89 44 24 08 e8 a1 70 f7 ff 8b 44
All code
========
    0:	83 ec 18             	sub    $0x18,%esp
    3:	89 54 24 0c          	mov    %edx,0xc(%rsp)
    7:	48 89 34 24          	mov    %rsi,(%rsp)
    b:	89 7c 24 08          	mov    %edi,0x8(%rsp)
    f:	e8 4b 70 f7 ff       	call   0xfffffffffff7705f
   14:	8b 54 24 0c          	mov    0xc(%rsp),%edx
   18:	48 8b 34 24          	mov    (%rsp),%rsi
   1c:	41 89 c0             	mov    %eax,%r8d
   1f:	8b 7c 24 08          	mov    0x8(%rsp),%edi
   23:	b8 2a 00 00 00       	mov    $0x2a,%eax
   28:	0f 05                	syscall
   2a:*	48 3d 00 f0 ff ff    	cmp    $0xfffffffffffff000,%rax		<-- trapping instruction
   30:	77 35                	ja     0x67
   32:	44 89 c7             	mov    %r8d,%edi
   35:	89 44 24 08          	mov    %eax,0x8(%rsp)
   39:	e8 a1 70 f7 ff       	call   0xfffffffffff770df
   3e:	8b                   	.byte 0x8b
   3f:	44                   	rex.R

Code starting with the faulting instruction
===========================================

    0:	48 3d 00 f0 ff ff    	cmp    $0xfffffffffffff000,%rax
    6:	77 35                	ja     0x3d
    8:	44 89 c7             	mov    %r8d,%edi
    b:	89 44 24 08          	mov    %eax,0x8(%rsp)
    f:	e8 a1 70 f7 ff       	call   0xfffffffffff770b5
   14:	8b                   	.byte 0x8b
   15:	44                   	rex.R
[  193.995050] RSP: 002b:00007fff992d3a20 EFLAGS: 00000293 ORIG_RAX: 000000000000002a
[  193.995054] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007faec2d0f9cb
[  193.995057] RDX: 0000000000000010 RSI: 00007fff992d3ad8 RDI: 0000000000000035
[  193.995059] RBP: 00007fff992d3ac0 R08: 0000000000000000 R09: 0000000000000004
[  193.995062] R10: 00007fff992d39b0 R11: 0000000000000293 R12: 00007fff992d7b78
[  193.995064] R13: 000000000095f760 R14: 0000000002e38b90 R15: 00007faec373d000
[  193.995091]  </TASK>