Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux
1
fork

Configure Feed

Select the types of activity you want to include in your feed.

Merge branch 'tcp-symmetric-challenge-ack-for-seg-ack-snd-nxt'

Jiayuan Chen says:

====================
tcp: symmetric challenge ACK for SEG.ACK > SND.NXT

Commit 354e4aa391ed ("tcp: RFC 5961 5.2 Blind Data Injection Attack
Mitigation") quotes RFC 5961 Section 5.2 in full, which requires
that any incoming segment whose ACK value falls outside
[SND.UNA - MAX.SND.WND, SND.NXT] MUST be discarded and an ACK sent
back. Linux currently sends that challenge ACK only on the lower
edge (SEG.ACK < SND.UNA - MAX.SND.WND); on the symmetric upper edge
(SEG.ACK > SND.NXT) the segment is silently dropped with
SKB_DROP_REASON_TCP_ACK_UNSENT_DATA.

Patch 1 completes the mitigation by emitting a rate-limited challenge
ACK on that branch, reusing tcp_send_challenge_ack() and honouring
FLAG_NO_CHALLENGE_ACK for consistency with the lower-edge case. It
also updates the existing tcp_ts_recent_invalid_ack.pkt selftest,
which drives this exact path, to consume the new challenge ACK so
bisect stays clean.

Patch 2 adds a new packetdrill selftest that exercises RFC 5961
Section 5.2 on both edges of the acceptable window, filling a gap in
the selftests tree (neither edge had dedicated coverage before).
====================

Link: https://patch.msgid.link/20260422123605.320000-1-jiayuan.chen@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

+58 -4
+7 -3
net/ipv4/tcp_input.c
··· 4286 4286 goto old_ack; 4287 4287 } 4288 4288 4289 - /* If the ack includes data we haven't sent yet, discard 4290 - * this segment (RFC793 Section 3.9). 4289 + /* If the ack includes data we haven't sent yet, drop the 4290 + * segment. RFC 793 Section 3.9 and RFC 5961 Section 5.2 4291 + * require us to send an ACK back in that case. 4291 4292 */ 4292 - if (after(ack, tp->snd_nxt)) 4293 + if (after(ack, tp->snd_nxt)) { 4294 + if (!(flag & FLAG_NO_CHALLENGE_ACK)) 4295 + tcp_send_challenge_ack(sk, false); 4293 4296 return -SKB_DROP_REASON_TCP_ACK_UNSENT_DATA; 4297 + } 4294 4298 4295 4299 if (after(ack, prior_snd_una)) { 4296 4300 flag |= FLAG_SND_UNA_ADVANCED;
+48
tools/testing/selftests/net/packetdrill/tcp_rfc5961_ack-out-of-window.pkt
··· 1 + // SPDX-License-Identifier: GPL-2.0 2 + // 3 + // RFC 5961 Section 5.2 / RFC 793 Section 3.9: an incoming segment's 4 + // ACK value must lie in [SND.UNA - MAX.SND.WND, SND.NXT]; otherwise 5 + // the receiver MUST discard the segment and send a challenge ACK 6 + // back. Exercise both edges of that window in a single connection. 7 + 8 + `./defaults.sh 9 + sysctl -q net.ipv4.tcp_invalid_ratelimit=0 10 + ` 11 + 12 + 0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3 13 + +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0 14 + +0 bind(3, ..., ...) = 0 15 + +0 listen(3, 1) = 0 16 + 17 + // Three-way handshake. Peer advertises rwnd = 1000 (no wscale), so 18 + // MAX.SND.WND is tracked as 1000. 19 + +0 < S 0:0(0) win 1000 <mss 1000,sackOK,nop,nop,nop,wscale 0> 20 + +0 > S. 0:0(0) ack 1 <...> 21 + +.1 < . 1:1(0) ack 1 win 1000 22 + +0 accept(3, ..., ...) = 4 23 + 24 + // ---- Upper edge: SEG.ACK > SND.NXT -------------------------------- 25 + // Server has sent nothing yet, so SND.UNA = SND.NXT = 1. 26 + // Peer sends a pure ACK with SEG.ACK = 2, beyond SND.NXT. 27 + +0 < . 1:1(0) ack 2 win 1000 28 + // Expect a challenge ACK: <SEQ = SND.NXT = 1, ACK = RCV.NXT = 1>. 29 + +0 > . 1:1(0) ack 1 30 + 31 + // Advance SND.UNA past MAX.SND.WND so that the lower edge becomes 32 + // reachable. Issue two 1-MSS writes so each skb is exactly one MSS 33 + // and PSH is set by tcp_push() at the end of each sendmsg, keeping 34 + // the setup independent of the TSO / tcp_fragment split path. 35 + +0 write(4, ..., 1000) = 1000 36 + +0 > P. 1:1001(1000) ack 1 37 + +.01 < . 1:1(0) ack 1001 win 1000 38 + +0 write(4, ..., 1000) = 1000 39 + +0 > P. 1001:2001(1000) ack 1 40 + +.01 < . 1:1(0) ack 2001 win 1000 41 + // Now SND.UNA = SND.NXT = 2001, MAX.SND.WND = 1000, bytes_acked = 2000. 42 + 43 + // ---- Lower edge: SEG.ACK < SND.UNA - MAX.SND.WND ------------------ 44 + // SND.UNA - MAX.SND.WND = 2001 - 1000 = 1001, so SEG.ACK = 1000 falls 45 + // below the acceptable range. 46 + +0 < . 1:1(0) ack 1000 win 1000 47 + // Expect a challenge ACK: <SEQ = SND.NXT = 2001, ACK = RCV.NXT = 1>. 48 + +0 > . 2001:2001(0) ack 1
+3 -1
tools/testing/selftests/net/packetdrill/tcp_ts_recent_invalid_ack.pkt
··· 19 19 20 20 // bad packet with high tsval (its ACK sequence is above our sndnxt) 21 21 +0 < F. 1:1(0) ack 9999 win 20000 <nop,nop,TS val 200000 ecr 100> 22 - 22 + // Challenge ACK for SEG.ACK > SND.NXT (RFC 5961 5.2 / RFC 793 3.9). 23 + // ecr=200 (not 200000) proves ts_recent was not updated from the bad packet. 24 + +0 > . 1:1(0) ack 1 <nop,nop,TS val 200 ecr 200> 23 25 24 26 +0 < . 1:1001(1000) ack 1 win 20000 <nop,nop,TS val 201 ecr 100> 25 27 +0 > . 1:1(0) ack 1001 <nop,nop,TS val 200 ecr 201>