From: Yonghong Song <yonghong.song@linux.dev>
To: Martin KaFai Lau <martin.lau@linux.dev>, bpf@vger.kernel.org
Cc: Alexei Starovoitov <ast@kernel.org>,
Andrii Nakryiko <andrii@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
kernel-team@meta.com
Subject: Re: [PATCH v2 bpf-next 1/2] selftests/bpf: Fix the flaky tc_redirect_dtime test
Date: Mon, 22 Jan 2024 21:02:19 -0800 [thread overview]
Message-ID: <791f4b62-3ea3-4bd6-a9a0-7f84022c9eb4@linux.dev> (raw)
In-Reply-To: <20240120060518.3604920-1-martin.lau@linux.dev>
On 1/19/24 10:05 PM, Martin KaFai Lau wrote:
> From: Martin KaFai Lau <martin.lau@kernel.org>
>
> BPF CI has been reporting the tc_redirect_dtime test failing
> from time to time:
>
> test_inet_dtime:PASS:setns src 0 nsec
> (network_helpers.c:253: errno: No route to host) Failed to connect to server
> close_netns:PASS:setns 0 nsec
> test_inet_dtime:FAIL:connect_to_fd unexpected connect_to_fd: actual -1 < expected 0
> test_tcp_clear_dtime:PASS:tcp ip6 clear dtime ingress_fwdns_p100 0 nsec
>
> The connect_to_fd failure (EHOSTUNREACH) is from the
> test_tcp_clear_dtime() test and it is the very first IPv6 traffic
> after setting up all the links, addresses, and routes.
>
> The symptom is this first connect() is always slow. In my setup, it
> could take ~3s.
>
> After some tracing and tcpdump, the slowness is mostly spent in
> the neighbor solicitation in the "ns_fwd" namespace while
> the "ns_src" and "ns_dst" are fine.
>
> I forced the kernel to drop the neighbor solicitation messages.
> I can then reproduce EHOSTUNREACH. What actually happen could be:
> - the neighbor advertisement came back a little slow.
> - the "ns_fwd" namespace concluded a neighbor discovery failure
> and triggered the ndisc_error_report() => ip6_link_failure() =>
> icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0)
> - the client's connect() reports EHOSTUNREACH after receiving
> the ICMPV6_DEST_UNREACH message.
>
> The neigh table of both "ns_src" and "ns_dst" namespace has already
> been manually populated but not the "ns_fwd" namespace. This patch
I tried one experiment which comments out manual population of neigh mac address
like below:
diff --git a/tools/testing/selftests/bpf/prog_tests/tc_redirect.c b/tools/testing/selftests/bpf/prog_tests/tc_redirect.c
index 518f143c5b0f..feb477366393 100644
--- a/tools/testing/selftests/bpf/prog_tests/tc_redirect.c
+++ b/tools/testing/selftests/bpf/prog_tests/tc_redirect.c
@@ -241,14 +241,14 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result)
SYS(fail, "ip route add " IP4_DST "/32 dev src scope global");
SYS(fail, "ip route add " IP4_NET "/16 dev src scope global");
SYS(fail, "ip route add " IP6_DST "/128 dev src scope global");
-
+#if 0
if (result->dev_mode == MODE_VETH) {
SYS(fail, "ip neigh add " IP4_DST " dev src lladdr %s",
src_fwd_addr);
SYS(fail, "ip neigh add " IP6_DST " dev src lladdr %s",
src_fwd_addr);
}
-
+#endif
close_netns(nstoken);
/** setup in 'fwd' namespace */
@@ -284,12 +284,12 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result)
SYS(fail, "ip route add " IP4_SRC "/32 dev dst scope global");
SYS(fail, "ip route add " IP4_NET "/16 dev dst scope global");
SYS(fail, "ip route add " IP6_SRC "/128 dev dst scope global");
-
+#if 0
if (result->dev_mode == MODE_VETH) {
SYS(fail, "ip neigh add " IP4_SRC " dev dst lladdr " MAC_DST_FWD);
SYS(fail, "ip neigh add " IP6_SRC " dev dst lladdr " MAC_DST_FWD);
}
-
+#endif
close_netns(nstoken);
return 0;
And I can 100% reliably trigger the following error:
$ ./test_progs -t tc_redirect
...
All error logs:
test_tc_redirect:PASS:pthread_create 0 nsec
netns_setup_namespaces:PASS:ip netns add ns_src 0 nsec
netns_setup_namespaces:PASS:ip netns add ns_fwd 0 nsec
netns_setup_namespaces:PASS:ip netns add ns_dst 0 nsec
test_tc_redirect_run_tests:PASS:setup namespaces 0 nsec
...
test_tcp:PASS:setns dst 0 nsec
test_tcp:PASS:listen 0 nsec
close_netns:PASS:setns 0 nsec
open_netns:PASS:malloc token 0 nsec
open_netns:PASS:open /proc/self/ns/net 0 nsec
open_netns:PASS:open netns fd 0 nsec
open_netns:PASS:setns 0 nsec
test_tcp:PASS:setns src 0 nsec
(network_helpers.c:253: errno: Operation now in progress) Failed to connect to server
test_tcp:FAIL:connect_to_fd unexpected connect_to_fd: actual -1 < expected 0
close_netns:PASS:setns 0 nsec
test_ping:FAIL:ip netns exec ns_src ping -i 0.2 -c 3 -w 10 -q 172.16.2.100 > /dev/null unexpected error: 256 (errno 115)
...
open_netns:PASS:setns 0 nsec
test_tcp:PASS:setns src 0 nsec
(network_helpers.c:253: errno: Operation now in progress) Failed to connect to server
test_tcp:FAIL:connect_to_fd unexpected connect_to_fd: actual -1 < expected 0
close_netns:PASS:setns 0 nsec
test_ping:FAIL:ip netns exec ns_src ping6 -i 0.2 -c 3 -w 10 -q 0::2:dead:beef:cafe > /dev/null unexpected error: 256 (errno 115)
close_netns:PASS:setns 0 nsec
The error message:
(network_helpers.c:253: errno: Operation now in progress) Failed to connect to server
not exactly the same but very similar to the failure observed in CI:
(network_helpers.c:253: errno: No route to host) Failed to connect to server
The above error 'Operation now in progress' may indicate some slowness in the kernel
which caused the test failure, with or without my above hack.
> fixes it by manually populating the neigh table also in the "ns_fwd"
> namespace.
>
> Although the namespace configuration part had been existed before
> the tc_redirect_dtime test, still Fixes-tagging the patch when
> the tc_redirect_dtime test was added since it is the only test
> hitting it so far.
>
> Fixes: c803475fd8dd ("bpf: selftests: test skb->tstamp in redirect_neigh")
> Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
> ---
> tools/testing/selftests/bpf/prog_tests/tc_redirect.c | 11 +++++++++++
> 1 file changed, 11 insertions(+)
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/tc_redirect.c b/tools/testing/selftests/bpf/prog_tests/tc_redirect.c
> index 518f143c5b0f..610887157fd8 100644
> --- a/tools/testing/selftests/bpf/prog_tests/tc_redirect.c
> +++ b/tools/testing/selftests/bpf/prog_tests/tc_redirect.c
> @@ -188,6 +188,7 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result)
> {
> struct nstoken *nstoken = NULL;
> char src_fwd_addr[IFADDR_STR_LEN+1] = {};
> + char src_addr[IFADDR_STR_LEN + 1] = {};
> int err;
>
> if (result->dev_mode == MODE_VETH) {
> @@ -208,6 +209,9 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result)
> if (get_ifaddr("src_fwd", src_fwd_addr))
> goto fail;
>
> + if (get_ifaddr("src", src_addr))
> + goto fail;
> +
> result->ifindex_src = if_nametoindex("src");
> if (!ASSERT_GT(result->ifindex_src, 0, "ifindex_src"))
> goto fail;
> @@ -270,6 +274,13 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result)
> SYS(fail, "ip route add " IP4_DST "/32 dev dst_fwd scope global");
> SYS(fail, "ip route add " IP6_DST "/128 dev dst_fwd scope global");
>
> + if (result->dev_mode == MODE_VETH) {
> + SYS(fail, "ip neigh add " IP4_SRC " dev src_fwd lladdr %s", src_addr);
> + SYS(fail, "ip neigh add " IP6_SRC " dev src_fwd lladdr %s", src_addr);
> + SYS(fail, "ip neigh add " IP4_DST " dev dst_fwd lladdr %s", MAC_DST);
> + SYS(fail, "ip neigh add " IP6_DST " dev dst_fwd lladdr %s", MAC_DST);
> + }
> +
> close_netns(nstoken);
>
> /** setup in 'dst' namespace */
next prev parent reply other threads:[~2024-01-23 5:02 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-20 6:05 [PATCH v2 bpf-next 1/2] selftests/bpf: Fix the flaky tc_redirect_dtime test Martin KaFai Lau
2024-01-20 6:05 ` [PATCH v2 bpf-next 2/2] selftests/bpf: Wait for the netstamp_needed_key static key to be turned on Martin KaFai Lau
2024-01-23 5:02 ` Yonghong Song [this message]
2024-01-24 18:20 ` [PATCH v2 bpf-next 1/2] selftests/bpf: Fix the flaky tc_redirect_dtime test patchwork-bot+netdevbpf
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=791f4b62-3ea3-4bd6-a9a0-7f84022c9eb4@linux.dev \
--to=yonghong.song@linux.dev \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=kernel-team@meta.com \
--cc=martin.lau@linux.dev \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.