netdev - Re: [PATCH net-next] tcp: accept RST if SEQ matches right edge of right-most SACK block

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CADVnQyn_AciWUFpOQdXPA97rih9JYkO8OZQUbC8+ADe0WiRoaA@mail.gmail.com>
Date:	Tue, 7 Jun 2016 12:05:47 -0400
From:	Neal Cardwell <ncardwell@...gle.com>
To:	Eric Dumazet <eric.dumazet@...il.com>
Cc:	Pau Espin Pedrol <pau.espin@...sares.net>,
	Netdev <netdev@...r.kernel.org>,
	Yuchung Cheng <ycheng@...gle.com>
Subject: Re: [PATCH net-next] tcp: accept RST if SEQ matches right edge of
 right-most SACK block

On Tue, Jun 7, 2016 at 11:21 AM, Eric Dumazet <eric.dumazet@...il.com> wrote:
> On Tue, 2016-06-07 at 16:30 +0200, Pau Espin Pedrol wrote:
>> RFC 5961 advises to only accept RST packets containing a seq number
>> matching the next expected seq number instead of the whole receive
>> window in order to avoid spoofing attacks.
>
>> Signed-off-by: Pau Espin Pedrol <pau.espin@...sares.net>

Acked-by: Neal Cardwell <ncardwell@...gle.com>
Tested-by: Neal Cardwell <ncardwell@...gle.com>

I verified that it passes the following packetdrill test, which is
based on the script posted by Pau in a recent thread on this topic.

Thanks!

neal

----

// Test that we accept an incoming RST if it matches the rightmost
// SACK block end.

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 500 <mss 1460,sackOK,nop,nop,nop,wscale 7>
+0    > S. 0:0(0) ack 1 win 29200 <mss 1460,nop,nop,sackOK,nop,wscale 7>
+0    < . 1:1(0) ack 1 win 500
+0    accept(3, ..., ...) = 4

// First roundtrip of data client->server (upload)
+0    < P. 1:1001(1000) ack 1 win 500
+0    > . 1:1(0) ack 1001

// A packet arrives without prev packet being seen.
+0    < P. 5001:6001(1000) ack 1 win 500
+0    > . 1:1(0) ack 1001 <nop,nop,sack 5001:6001>

// Another out-of-order packet arrives.
+0    < P. 3001:4001(1000) ack 1 win 500
+0    > . 1:1(0) ack 1001 <nop,nop,sack 3001:4001 5001:6001>

// We send a challenge ACK for a RST matching the right edge of the
inner SACK block:
+0    < R. 4001:4001(0) ack 1 win 500
+0    > . 1:1(0) ack 1001 <nop,nop,sack 3001:4001 5001:6001>

// RST arrives from client, for instance caused by setsockopt(fd,
SOL_SOCKET, SO_LINGER, ...)
// Matches right edge of right-most SACK block, so we reset the connection.
+0    < R. 6001:6001(0) ack 1 win 500

// Verify that the RST caused the kernel to reset the connection:
+.010 write(4, ..., 10000) = -1 ECONNRESET (Connection reset by peer)

// Double-check  that the socket is gone:
+0    < P. 2001:3001(1000) ack 1 win 500
+0    > R 1:1(0) ack 0