From: Eric Dumazet <edumazet@kernel.org>
To: "David S . Miller" <davem@davemloft.net>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
Neal Cardwell <ncardwell@google.com>
Cc: Kuniyuki Iwashima <kuniyu@google.com>,
Simon Horman <horms@kernel.org>,
Flavius-Romeo Mazilu <flavius.mazilu@stud.acs.upb.ro>,
Michael Tuexen <tuexen@fh-muenster.de>,
netdev@vger.kernel.org, Eric Dumazet <edumazet@kernel.org>
Subject: [PATCH net-next 2/2] selftests/net: packetdrill: add tests for ACK of ISS
Date: Wed, 18 Nov 2026 18:26:54 +0100 [thread overview]
Message-ID: <20261118172654.1267071-3-edumazet@kernel.org> (raw)
In-Reply-To: <20261118172654.1267071-1-edumazet@kernel.org>
Once a connection is established, a segment with SEG.ACK == ISS
acknowledges bytes we never sent. It must be dropped and a challenge
ACK sent back (RFC 5961 Section 5.2), regardless of which side
opened the connection.
Add four tests covering the ways a socket can be established:
- tcp_rfc5961_ack-iss-server.pkt: passive open
- tcp_rfc5961_ack-iss-client.pkt: active open
- tcp_rfc5961_ack-iss-fastopen-client.pkt: active open, data in SYN
- tcp_rfc5961_ack-iss-simultaneous-open.pkt: crossed SYNs
Each test also checks that SEG.ACK == ISS + 1 is still silently
accepted as a genuine old ACK.
Only the passive open test passes without the previous patch.
Link: https://mailarchive.ietf.org/arch/msg/tcpm/ym6lb-UkrW2298LRnTLxCysoo0g/
Signed-off-by: Eric Dumazet <edumazet@kernel.org>
---
.../tcp_rfc5961_ack-iss-client.pkt | 46 +++++++++++++++++++
.../tcp_rfc5961_ack-iss-fastopen-client.pkt | 40 ++++++++++++++++
.../tcp_rfc5961_ack-iss-server.pkt | 45 ++++++++++++++++++
.../tcp_rfc5961_ack-iss-simultaneous-open.pkt | 42 +++++++++++++++++
4 files changed, 173 insertions(+)
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-client.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-fastopen-client.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-server.pkt
create mode 100644 tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-simultaneous-open.pkt
diff --git a/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-client.pkt b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-client.pkt
new file mode 100644
index 0000000000000000000000000000000000000000..50c0638a5219d8389ab433765f9f65fa0f0a8ffc
--- /dev/null
+++ b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-client.pkt
@@ -0,0 +1,46 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Active open: once the connection is established, SND.UNA is at least
+// ISS + 1. A segment with SEG.ACK = ISS acknowledges nothing we ever
+// sent past the SYN, and the SYN itself can not be acked "again" as a
+// stale duplicate: no ACK carrying ISS was ever valid on this
+// connection. It must be treated like any other ACK for bytes we never
+// sent (RFC 5961 Section 5.2): dropped, and a challenge ACK sent back.
+//
+// The SYN consumes one sequence number but is not a payload byte, so it
+// must not widen the range of acceptable old ACKs.
+
+`./defaults.sh
+sysctl -q net.ipv4.tcp_invalid_ratelimit=0
+`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 fcntl(3, F_GETFL) = 0x2 (flags O_RDWR)
+ +0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+ +0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress)
+ +0 > S 0:0(0) <...>
++.01 < S. 0:0(0) ack 1 win 10000 <mss 1000,nop,nop,sackOK,nop,wscale 0>
+ +0 > . 1:1(0) ack 1
+
+// SND.UNA = SND.NXT = ISS + 1, no payload acked yet.
+// SEG.ACK = ISS must trigger a challenge ACK.
++.01 < . 1:1(0) ack 0 win 10000
+ +0 > . 1:1(0) ack 1
+
+// Get 1000 bytes of payload acked.
+ +0 write(3, ..., 1000) = 1000
+ +0 > P. 1:1001(1000) ack 1
++.01 < . 1:1(0) ack 1001 win 10000
+
+// SND.UNA = ISS + 1001. Old ACKs in [ISS + 1, ISS + 1001] are valid
+// duplicates, but SEG.ACK = ISS is still below anything the peer could
+// have legitimately sent.
++.01 < . 1:1(0) ack 0 win 10000
+ +0 > . 1001:1001(0) ack 1
+
+// SEG.ACK = ISS + 1 is a genuine old ACK: silently accepted, no
+// challenge ACK. The next packet on the wire must be our new data.
++.01 < . 1:1(0) ack 1 win 10000
++.01 write(3, ..., 1000) = 1000
+ +0 > P. 1001:2001(1000) ack 1
++.01 < . 1:1(0) ack 2001 win 10000
diff --git a/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-fastopen-client.pkt b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-fastopen-client.pkt
new file mode 100644
index 0000000000000000000000000000000000000000..5e8c7f8ee7ddb812761dd5dbc72e2f01ff46c644
--- /dev/null
+++ b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-fastopen-client.pkt
@@ -0,0 +1,40 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Fast Open variant of tcp_rfc5961_ack-iss-client.pkt.
+//
+// The SYN carries data, and the SYN-ACK acks both the SYN and its
+// payload. SEG.ACK = ISS must still trigger a challenge ACK, while
+// SEG.ACK = ISS + 1 (SYN acked, payload not yet) is a valid old ACK.
+
+`./defaults.sh
+ ./set_sysctls.py /proc/sys/net/ipv4/tcp_timestamps=0 \
+ /proc/sys/net/ipv4/tcp_invalid_ratelimit=0`
+
+// Cache warmup: send a Fast Open cookie request
+ 0 socket(..., SOCK_STREAM|SOCK_NONBLOCK, IPPROTO_TCP) = 3
+ +0 sendto(3, ..., 0, MSG_FASTOPEN, ..., ...) = -1 EINPROGRESS (Operation is now in progress)
+ +0 > S 0:0(0) <mss 1460,nop,nop,sackOK,nop,wscale 8,FO,nop,nop>
+ +0 < S. 123:123(0) ack 1 win 14600 <mss 1460,nop,nop,sackOK,nop,wscale 6,FO abcd1234,nop,nop>
+ +0 > . 1:1(0) ack 1
+ +0 close(3) = 0
+ +0 > F. 1:1(0) ack 1
+ +0 < F. 1:1(0) ack 2 win 92
+ +0 > . 2:2(0) ack 2
+
+// SYN + 500 bytes of payload, all acked by the SYN-ACK.
+ +0 socket(..., SOCK_STREAM|SOCK_NONBLOCK, IPPROTO_TCP) = 4
+ +0 sendto(4, ..., 500, MSG_FASTOPEN, ..., ...) = 500
+ +0 > S 0:500(500) <mss 1460,nop,nop,sackOK,nop,wscale 8,FO abcd1234,nop,nop>
+ +.01 < S. 0:0(0) ack 501 win 14600 <mss 1460,nop,nop,sackOK,nop,wscale 6>
+ +0 > . 501:501(0) ack 1
+
+// SND.UNA = ISS + 501. SEG.ACK = ISS must trigger a challenge ACK.
+ +.01 < . 1:1(0) ack 0 win 1000
+ +0 > . 501:501(0) ack 1
+
+// SEG.ACK = ISS + 1 is a genuine old ACK: silently accepted, no
+// challenge ACK. The next packet on the wire must be our new data.
+ +.01 < . 1:1(0) ack 1 win 1000
+ +.01 write(4, ..., 100) = 100
+ +0 > P. 501:601(100) ack 1
+ +.01 < . 1:1(0) ack 601 win 1000
diff --git a/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-server.pkt b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-server.pkt
new file mode 100644
index 0000000000000000000000000000000000000000..508380ff1e72d2a31b13471989f0cb0f5377011a
--- /dev/null
+++ b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-server.pkt
@@ -0,0 +1,45 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Passive open counterpart of tcp_rfc5961_ack-iss-client.pkt.
+//
+// Once the connection is established, a segment with SEG.ACK = ISS
+// acknowledges bytes we never sent. It must be dropped and a challenge
+// ACK sent back (RFC 5961 Section 5.2), whatever amount of payload was
+// acked so far.
+
+`./defaults.sh
+sysctl -q net.ipv4.tcp_invalid_ratelimit=0
+`
+
+ 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+ +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+ +0 bind(3, ..., ...) = 0
+ +0 listen(3, 1) = 0
+
+ +0 < S 0:0(0) win 10000 <mss 1000,nop,nop,sackOK,nop,wscale 0>
+ +0 > S. 0:0(0) ack 1 <...>
++.01 < . 1:1(0) ack 1 win 10000
+ +0 accept(3, ..., ...) = 4
+
+// SND.UNA = SND.NXT = ISS + 1, no payload acked yet.
+// SEG.ACK = ISS must trigger a challenge ACK.
++.01 < . 1:1(0) ack 0 win 10000
+ +0 > . 1:1(0) ack 1
+
+// Get 1000 bytes of payload acked.
+ +0 write(4, ..., 1000) = 1000
+ +0 > P. 1:1001(1000) ack 1
++.01 < . 1:1(0) ack 1001 win 10000
+
+// SND.UNA = ISS + 1001. Old ACKs in [ISS + 1, ISS + 1001] are valid
+// duplicates, but SEG.ACK = ISS is still below anything the peer could
+// have legitimately sent.
++.01 < . 1:1(0) ack 0 win 10000
+ +0 > . 1001:1001(0) ack 1
+
+// SEG.ACK = ISS + 1 is a genuine old ACK: silently accepted, no
+// challenge ACK. The next packet on the wire must be our new data.
++.01 < . 1:1(0) ack 1 win 10000
++.01 write(4, ..., 1000) = 1000
+ +0 > P. 1001:2001(1000) ack 1
++.01 < . 1:1(0) ack 2001 win 10000
diff --git a/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-simultaneous-open.pkt b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-simultaneous-open.pkt
new file mode 100644
index 0000000000000000000000000000000000000000..17a545712a929760332a0cbceabb00ef1a38f234
--- /dev/null
+++ b/tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-iss-simultaneous-open.pkt
@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: GPL-2.0
+//
+// Simultaneous open variant of tcp_rfc5961_ack-iss-client.pkt.
+//
+// Our SYN is acked by the SYN-ACK of the peer while we are in SYN-RECV
+// state. Once established, SEG.ACK = ISS must trigger a challenge ACK.
+
+`./defaults.sh
+sysctl -q net.ipv4.tcp_invalid_ratelimit=0
+`
+
+ 0 socket(..., SOCK_STREAM|SOCK_NONBLOCK, IPPROTO_TCP) = 3
+ +0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress)
+ +0 > S 0:0(0) <...>
+
+// Crossed SYN.
++.01 < S 0:0(0) win 10000 <mss 1000,nop,nop,sackOK,nop,wscale 0>
+ +0 > S. 0:0(0) ack 1 <...>
++.01 < S. 0:0(0) ack 1 win 10000 <mss 1000,nop,nop,sackOK,nop,wscale 0>
+// The SYN part of this SYN-ACK is a duplicate, and is DSACKed.
+ +0 > . 1:1(0) ack 1 <nop,nop,sack 0:1>
+ +0 %{ assert tcpi_state == TCP_ESTABLISHED, tcpi_state }%
+
+// SND.UNA = SND.NXT = ISS + 1, no payload acked yet.
+// SEG.ACK = ISS must trigger a challenge ACK.
++.01 < . 1:1(0) ack 0 win 10000
+ +0 > . 1:1(0) ack 1
+
+// Get 1000 bytes of payload acked.
+ +0 write(3, ..., 1000) = 1000
+ +0 > P. 1:1001(1000) ack 1
++.01 < . 1:1(0) ack 1001 win 10000
+
++.01 < . 1:1(0) ack 0 win 10000
+ +0 > . 1001:1001(0) ack 1
+
+// SEG.ACK = ISS + 1 is a genuine old ACK: silently accepted, no
+// challenge ACK. The next packet on the wire must be our new data.
++.01 < . 1:1(0) ack 1 win 10000
++.01 write(3, ..., 1000) = 1000
+ +0 > P. 1001:2001(1000) ack 1
++.01 < . 1:1(0) ack 2001 win 10000
--
2.53.0
prev parent reply other threads:[~2026-11-18 17:27 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-18 17:26 [PATCH net-next 0/2] tcp: do not accept ACK of our ISS on active opens Eric Dumazet
2026-11-18 17:26 ` [PATCH net-next 1/2] " Eric Dumazet
2026-11-18 17:26 ` Eric Dumazet [this message]
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=20261118172654.1267071-3-edumazet@kernel.org \
--to=edumazet@kernel.org \
--cc=davem@davemloft.net \
--cc=flavius.mazilu@stud.acs.upb.ro \
--cc=horms@kernel.org \
--cc=kuba@kernel.org \
--cc=kuniyu@google.com \
--cc=ncardwell@google.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=tuexen@fh-muenster.de \
/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