netdev - [PATCH net-next 4/4] sch_sfb: use skb_flow

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: <1322493902.2292.72.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC>
Date:	Mon, 28 Nov 2011 16:25:02 +0100
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	David Miller <davem@...emloft.net>
Cc:	john.r.fastabend@...el.com, jpettit@...ira.com,
	Florian Westphal <fw@...len.de>,
	Stephen Hemminger <shemminger@...tta.com>, jesse@...ira.com,
	jhs@...atatu.com, dev@...nvswitch.org,
	netdev <netdev@...r.kernel.org>, Dan Siemon <dan@...erfire.com>,
	herbert@...dor.hengli.com.au, chrisw@...hat.com
Subject: [PATCH net-next 4/4] sch_sfb: use skb_flow_dissect()

Current SFB double hashing is not fulfilling SFB theory, if two flows
share same rxhash value.

Using skb_flow_dissect() permits to really have better hash dispersion,
and get tunnelling support as well.

Double hashing point was mentioned by Florian Westphal

Signed-off-by: Eric Dumazet <eric.dumazet@...il.com>
---
 net/sched/sch_sfb.c |   17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/net/sched/sch_sfb.c b/net/sched/sch_sfb.c
index e83c272..96e42ca 100644
--- a/net/sched/sch_sfb.c
+++ b/net/sched/sch_sfb.c
@@ -26,6 +26,7 @@
 #include <net/ip.h>
 #include <net/pkt_sched.h>
 #include <net/inet_ecn.h>
+#include <net/flow_keys.h>
 
 /*
  * SFB uses two B[l][n] : L x N arrays of bins (L levels, N bins per level)
@@ -286,6 +287,7 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	u32 minqlen = ~0;
 	u32 r, slot, salt, sfbhash;
 	int ret = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS;
+	struct flow_keys keys;
 
 	if (unlikely(sch->q.qlen >= q->limit)) {
 		sch->qstats.overlimits++;
@@ -309,13 +311,19 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 		/* If using external classifiers, get result and record it. */
 		if (!sfb_classify(skb, q, &ret, &salt))
 			goto other_drop;
+		keys.src = salt;
+		keys.dst = 0;
+		keys.ports = 0;
 	} else {
-		salt = skb_get_rxhash(skb);
+		skb_flow_dissect(skb, &keys);
 	}
 
 	slot = q->slot;
 
-	sfbhash = jhash_1word(salt, q->bins[slot].perturbation);
+	sfbhash = jhash_3words((__force u32)keys.dst,
+			       (__force u32)keys.src,
+			       (__force u32)keys.ports,
+			       q->bins[slot].perturbation);
 	if (!sfbhash)
 		sfbhash = 1;
 	sfb_skb_cb(skb)->hashes[slot] = sfbhash;
@@ -347,7 +355,10 @@ static int sfb_enqueue(struct sk_buff *skb, struct Qdisc *sch)
 	if (unlikely(p_min >= SFB_MAX_PROB)) {
 		/* Inelastic flow */
 		if (q->double_buffering) {
-			sfbhash = jhash_1word(salt, q->bins[slot].perturbation);
+			sfbhash = jhash_3words((__force u32)keys.dst,
+					       (__force u32)keys.src,
+					       (__force u32)keys.ports,
+					       q->bins[slot].perturbation);
 			if (!sfbhash)
 				sfbhash = 1;
 			sfb_skb_cb(skb)->hashes[slot] = sfbhash;


--
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