Skip to contentSkip to conversation

selftests/bpf: Use ping_command() for IPv6 pings in lwt_ip_encap

PATCH Applied Andrii Nakryiko 4 participants 4 messages Aug 13 – Aug 14

Subsystems (3)
lore
Share & tools
lore

Andrii Nakryiko

author maintainer
#0 lore raw

PATCH 1/1 selftests/bpf: Use ping_command() for IPv6 pings in lwt_ip_encap

lwt_ip_encap hardcodes the ping6 binary for its IPv6 pings. iputils merged ping6 into ping long ago and distros have started dropping the compat symlink -- Arch's iputils 20250605 ships only arping, clockdiff, ping and tracepath. There, every lwt_ip_encap subtest fails:

  check_ping_ok:FAIL:ip netns exec ns-lwt-ip-encap-1-0101330 ping6 -c 1 \
    -W1 -I veth1 fb04::1 > /dev/null unexpected error: 256 (errno 2)
  #217/1   lwt_ip_encap_ipv4/egress:FAIL

The IPv4 subtests fail too, because check_ping_ok() pings both families. SYS() runs the command through system(), so a missing binary is indistinguishable from an unreachable peer.

network_helpers.c has had ping_command() for exactly this since commit 372642ea83ff ("selftests/bpf: Move netcnt test under test_progs"): it falls back to "ping -6" when ping6 is not present. lwt_ip_encap.c is the last hardcoded ping6 user. Fix that.

Fixes f5e288943e2c ("selftests/bpf: Move test_lwt_ip_encap to test_progs")
---
 tools/testing/selftests/bpf/prog_tests/lwt_ip_encap.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)
tools/testing/selftests/bpf/prog_tests/lwt_ip_encap.c +7 -4
index 6606f0ed9a9a..39e8a3b8b6af 100644
@@ -410,7 +410,8 @@ static int test_gso_fix(const char *ns1, const char *ns3, int family)
 static int check_ping_ok(const char *ns1)
 {
 	SYS(fail, "ip netns exec %s ping -c 1 -W1 -I veth1 %s > /dev/null", ns1, IP4_ADDR_DST);
-	SYS(fail, "ip netns exec %s ping6 -c 1 -W1 -I veth1 %s > /dev/null", ns1, IP6_ADDR_DST);
+	SYS(fail, "ip netns exec %s %s -c 1 -W1 -I veth1 %s > /dev/null", ns1,
+	    ping_command(AF_INET6), IP6_ADDR_DST);
 	return 0;
 fail:
 	return -1;
@@ -424,7 +425,8 @@ static int check_ping_fails(const char *ns1)
 	if (!ret)
 		return -1;
 
-	ret = SYS_NOFAIL("ip netns exec %s ping6 -c 1 -W1 -I veth1 %s", ns1, IP6_ADDR_DST);
+	ret = SYS_NOFAIL("ip netns exec %s %s -c 1 -W1 -I veth1 %s", ns1,
+			 ping_command(AF_INET6), IP6_ADDR_DST);
 	if (!ret)
 		return -1;
 
@@ -657,9 +659,10 @@ static void lwt_ip_encap_vxlan(bool ipv4_encap)
 	skel->bss->fexit_triggered = false;
 
 	if (ipv4_encap)
-		SYS(out, "ip netns exec %s ping  -c 1 -W1 %s", ns1, IP4_ADDR_DST);
+		SYS(out, "ip netns exec %s ping -c 1 -W1 %s", ns1, IP4_ADDR_DST);
 	else
-		SYS(out, "ip netns exec %s ping6 -c 1 -W1 %s", ns1, IP6_ADDR_DST);
+		SYS(out, "ip netns exec %s %s -c 1 -W1 %s", ns1,
+		    ping_command(AF_INET6), IP6_ADDR_DST);
 
 	if (!ASSERT_TRUE(skel->bss->fexit_triggered, "fexit_triggered"))
 		goto out;
Signature · 1 line
2.53.0-Meta
Acked-by on v1 1/1
#1 lore raw
Quoted 14 lines · Andrii Nakryiko' via Kernel Team

lwt_ip_encap hardcodes the ping6 binary for its IPv6 pings. iputils merged ping6 into ping long ago and distros have started dropping the compat symlink -- Arch's iputils 20250605 ships only arping, clockdiff, ping and tracepath. There, every lwt_ip_encap subtest fails:

  check_ping_ok:FAIL:ip netns exec ns-lwt-ip-encap-1-0101330 ping6 -c 1 \
    -W1 -I veth1 fb04::1 > /dev/null unexpected error: 256 (errno 2)
  #217/1   lwt_ip_encap_ipv4/egress:FAIL

The IPv4 subtests fail too, because check_ping_ok() pings both families. SYS() runs the command through system(), so a missing binary is indistinguishable from an unreachable peer.

Quoted 7 lines

network_helpers.c has had ping_command() for exactly this since commit 372642ea83ff ("selftests/bpf: Move netcnt test under test_progs"): it falls back to "ping -6" when ping6 is not present. lwt_ip_encap.c is the last hardcoded ping6 user. Fix that.

Fixes f5e288943e2c ("selftests/bpf: Move test_lwt_ip_encap to test_progs")
#2 lore raw
In reply to Andrii Nakryiko
The sentence "every lwt_ip_encap subtest fails" isn't quite accurate. Looking at test_lwt_ip_encap_vxlan_ipv4() in the patched code, it calls lwt_ip_encap_vxlan(IPV4_ENCAP), which only ever runs the…
Quoted 6 lines

commit 0717c3442da506b8360051a6f87406ecc25dde4c

Author: Andrii Nakryiko <[email protected]>
Date:   Thu Aug 13 14:35:58 2026 -0700

selftests/bpf: Use ping_command() for IPv6 pings in lwt_ip_encap

Quoted 8 lines

lwt_ip_encap hardcodes the ping6 binary for its IPv6 pings. iputils merged ping6 into ping long ago and distros have started dropping the compat symlink -- Arch's iputils 20250605 ships only arping, clockdiff, ping and tracepath. There, every lwt_ip_encap subtest fails:

check_ping_ok:FAIL:ip netns exec ns-lwt-ip-encap-1-0101330 ping6 -c 1 \
-W1 -I veth1 fb04::1 > /dev/null unexpected error: 256 (errno 2)
#217/1 lwt_ip_encap_ipv4/egress:FAIL

The sentence "every lwt_ip_encap subtest fails" isn't quite accurate. Looking at test_lwt_ip_encap_vxlan_ipv4() in the patched code, it calls lwt_ip_encap_vxlan(IPV4_ENCAP), which only ever runs the IPv4 arm:

    if (ipv4_encap)
            SYS(out, "ip netns exec %s ping -c 1 -W1 %s", ns1, IP4_ADDR_DST);
    else
            SYS(out, "ip netns exec %s %s -c 1 -W1 %s", ns1,
                ping_command(AF_INET6), IP6_ADDR_DST);

This path never touches ping6 or ping_command(), so the test_lwt_ip_encap_vxlan_ipv4 subtest passes on a ping6-less host both before and after this patch.

The other 9 subtests do fail: lwt_ip_encap_ipv4/{egress,ingress}, lwt_ip_encap_ipv6/{egress,ingress}, lwt_ip_encap_vrf_ipv4/{egress,ingress}, lwt_ip_encap_vrf_ipv6/{egress,ingress}, and lwt_ip_encap_vxlan_ipv6.

Would it be more accurate to say "every subtest that pings over IPv6 fails" so the reproducer description matches the code behavior?

---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/31747915907

#3 lore raw
In reply to Andrii Nakryiko
This patch was applied to bpf/bpf-next.git (master) by Daniel Borkmann <[email protected]>:

Hello:

This patch was applied to bpf/bpf-next.git (master) by Daniel Borkmann <[email protected]>:

Quoted 10 lines · On Thu, 13 Aug 2026 14:35:58 -0700 you wrote:

lwt_ip_encap hardcodes the ping6 binary for its IPv6 pings. iputils merged ping6 into ping long ago and distros have started dropping the compat symlink -- Arch's iputils 20250605 ships only arping, clockdiff, ping and tracepath. There, every lwt_ip_encap subtest fails:

  check_ping_ok:FAIL:ip netns exec ns-lwt-ip-encap-1-0101330 ping6 -c 1 \
    -W1 -I veth1 fb04::1 > /dev/null unexpected error: 256 (errno 2)
  #217/1   lwt_ip_encap_ipv4/egress:FAIL

[...]

Here is the summary with links:
- [bpf-next] selftests/bpf: Use ping_command() for IPv6 pings in lwt_ip_encap

    https://git.kernel.org/bpf/bpf-next/c/073574da7a8e

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html