BPF List
 help / color / mirror / Atom feed
From: Martin KaFai Lau <martin.lau@linux.dev>
To: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Cc: Martin KaFai Lau <martin.lau@kernel.org>, bpf <bpf@vger.kernel.org>
Subject: Re: flaky tc_redirect/tc_redirect_dtime
Date: Mon, 15 Jan 2024 07:09:34 -0800	[thread overview]
Message-ID: <8af7e87c-1103-44b7-a0eb-1a28f52666a2@linux.dev> (raw)
In-Reply-To: <CAADnVQJ_Xwk7xp-AybVC7dtSqRnbo1Lkw1Y+vQ+_w6UJTPvhKw@mail.gmail.com>

On 1/13/24 1:09 PM, Alexei Starovoitov wrote:
> Hi Martin,
> 
> I remember you tried to fix tc_redirect/tc_redirect_dtime flakiness,
> but it is still flaky.
> Just running test_progs -t tc_redirect/tc_redirect_dtime
> in a loop it will fail after 30-50 iterations in my VM and always with:
> 
> test_inet_dtime:PASS:setns src 0 nsec
> (network_helpers.c:253: errno: Operation now in progress) 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
> 
> I've added this hack:
> +again:
>          nstoken = open_netns(NS_SRC);
>          if (!ASSERT_OK_PTR(nstoken, "setns src"))
>                  goto done;
> @@ -573,6 +575,11 @@ static void test_inet_dtime(int family, int type,
> const char *addr, __u16 port)
>          if (!ASSERT_GE(client_fd, 0, "connect_to_fd"))
>                  goto done;
> 
> +       if (i++ < 1000 && 0) {
> +               printf("XXXX %d\n", i);
> +               close(client_fd);
> +               goto again;
> +       }
> 
> and realized that only the first connect can succeed.
> The 2nd connect always fails.
> So I suspect bpf prog sees first stray packet, acts on it,
> but the actual connect request comes 2nd and it fails.
> 
> I tried to understand what's going on inside bpf prog,
> but the test is too complicated.
> Please take a look when you have a chance.
> 
> I also added:
> @@ -857,7 +864,7 @@ static void test_tc_redirect_dtime(struct
> netns_setup_result *setup_result)
>                  goto done;
> 
>          test_tcp_clear_dtime(skel);
> -
> +if (0) {
>          test_tcp_dtime(skel, AF_INET, true);
>          test_tcp_dtime(skel, AF_INET6, true);
>          test_udp_dtime(skel, AF_INET, true);
> @@ -878,7 +885,7 @@ static void test_tc_redirect_dtime(struct
> netns_setup_result *setup_result)
>          test_tcp_dtime(skel, AF_INET6, false);
>          test_udp_dtime(skel, AF_INET, false);
>          test_udp_dtime(skel, AF_INET6, false);
> -
> +}
> to speed up a test a bit.

Thanks for the tips to reproduce it. I will look into it.


  reply	other threads:[~2024-01-15 15:09 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-13 21:09 flaky tc_redirect/tc_redirect_dtime Alexei Starovoitov
2024-01-15 15:09 ` Martin KaFai Lau [this message]
2024-01-16 22:38   ` Martin KaFai Lau

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=8af7e87c-1103-44b7-a0eb-1a28f52666a2@linux.dev \
    --to=martin.lau@linux.dev \
    --cc=alexei.starovoitov@gmail.com \
    --cc=bpf@vger.kernel.org \
    --cc=martin.lau@kernel.org \
    /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