netdev - suspicious logic in tcp_make

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-next>] [day] [month] [year] [list]

Message-Id: <20081026.010158.193698298.davem@davemloft.net>
Date:	Sun, 26 Oct 2008 01:01:58 -0700 (PDT)
From:	David Miller <davem@...emloft.net>
To:	agl@...erialviolet.org
CC:	netdev@...r.kernel.org
Subject: suspicious logic in tcp_make_synack()


I've been debugging something over the past few days and noticed
something I can't explain in your change:

commit 33ad798c924b4a1afad3593f2796d465040aadd5
Author: Adam Langley <agl@...erialviolet.org>
Date:   Sat Jul 19 00:04:31 2008 -0700

    tcp: options clean up

specifically this hunk of the change:

@@ -2244,18 +2295,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 		TCP_SKB_CB(skb)->when = cookie_init_timestamp(req);
 	else
 #endif
-	TCP_SKB_CB(skb)->when = tcp_time_stamp;
-	tcp_syn_build_options((__be32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), ireq->tstamp_ok,
-			      ireq->sack_ok, ireq->wscale_ok, ireq->rcv_wscale,
-			      TCP_SKB_CB(skb)->when,
-			      req->ts_recent,
-			      (
-#ifdef CONFIG_TCP_MD5SIG
-			       md5 ? &md5_hash_location :
-#endif
-			       NULL)
-			      );
-
+	tcp_options_write((__be32 *)(th + 1), tp, &opts, &md5_hash_location);

the code right above this stuff reads:

#ifdef CONFIG_SYN_COOKIES
	if (unlikely(req->cookie_ts))
		TCP_SKB_CB(skb)->when = cookie_init_timestamp(req);
	else
#endif

The old code unconditionally built TCP options for the SYNACK, but now
it doesn't build those options if req->cookie_ts, that call falls into
the "else" branch, is that intentional?

That req->cookie_ts test is meant to only control how
TCP_SKB_CB(skb)->when gets calculated.

This looks like a bug to me.  It means also we now won't set ->when if
CONFIG_SYN_COOKIES is enabled and req->cookie_ts is false.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html