Netdev List
 help / color / mirror / Atom feed
From: Eric Dumazet <eric.dumazet@gmail.com>
To: Changli Gao <xiaosuo@gmail.com>, David Miller <davem@davemloft.net>
Cc: therbert@google.com, netdev@vger.kernel.org
Subject: [PATCH net-next-2.6] rps: consistent rxhash
Date: Tue, 20 Apr 2010 09:56:38 +0200	[thread overview]
Message-ID: <1271750198.3845.216.camel@edumazet-laptop> (raw)
In-Reply-To: <1271743164.3845.128.camel@edumazet-laptop>

In case we compute a software skb->rxhash, we can generate a consistent
hash : Its value will be the same in both flow directions.

This helps some workloads, like conntracking, since the same state needs
to be accessed in both directions.

tbench + RFS + this patch gives better results than tbench with default
kernel configuration (no RPS, no RFS)

Also fixed some sparse warnings.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
 net/core/dev.c |   25 ++++++++++++++++++-------
 1 files changed, 18 insertions(+), 7 deletions(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index 05a2b29..cb150ec 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1974,7 +1974,7 @@ u16 skb_tx_hash(const struct net_device *dev, const struct sk_buff *skb)
 	if (skb->sk && skb->sk->sk_hash)
 		hash = skb->sk->sk_hash;
 	else
-		hash = skb->protocol;
+		hash = (__force u16) skb->protocol;
 
 	hash = jhash_1word(hash, hashrnd);
 
@@ -2253,8 +2253,8 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
 
 		ip = (struct iphdr *) skb->data;
 		ip_proto = ip->protocol;
-		addr1 = ip->saddr;
-		addr2 = ip->daddr;
+		addr1 = (__force u32) ip->saddr;
+		addr2 = (__force u32) ip->daddr;
 		ihl = ip->ihl;
 		break;
 	case __constant_htons(ETH_P_IPV6):
@@ -2263,8 +2263,8 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
 
 		ip6 = (struct ipv6hdr *) skb->data;
 		ip_proto = ip6->nexthdr;
-		addr1 = ip6->saddr.s6_addr32[3];
-		addr2 = ip6->daddr.s6_addr32[3];
+		addr1 = (__force u32) ip6->saddr.s6_addr32[3];
+		addr2 = (__force u32) ip6->daddr.s6_addr32[3];
 		ihl = (40 >> 2);
 		break;
 	default:
@@ -2279,14 +2279,25 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
 	case IPPROTO_AH:
 	case IPPROTO_SCTP:
 	case IPPROTO_UDPLITE:
-		if (pskb_may_pull(skb, (ihl * 4) + 4))
-			ports = *((u32 *) (skb->data + (ihl * 4)));
+		if (pskb_may_pull(skb, (ihl * 4) + 4)) {
+			__be16 *hports = (__be16 *) (skb->data + (ihl * 4));
+			u32 sport, dport;
+
+			sport = (__force u16) hports[0];
+			dport = (__force u16) hports[1];
+			if (dport < sport)
+				swap(sport, dport);
+			ports = (sport << 16) + dport;
+		}
 		break;
 
 	default:
 		break;
 	}
 
+	/* get a consistent hash (same value on both flow directions) */
+	if (addr2 < addr1)
+		swap(addr1, addr2);
 	skb->rxhash = jhash_3words(addr1, addr2, ports, hashrnd);
 	if (!skb->rxhash)
 		skb->rxhash = 1;



  reply	other threads:[~2010-04-20  7:56 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-16  5:47 [PATCH v5] rfs: Receive Flow Steering Tom Herbert
2010-04-16  6:33 ` David Miller
2010-04-16  6:56   ` Eric Dumazet
2010-04-16  7:18     ` Eric Dumazet
2010-04-16  7:26       ` David Miller
2010-04-16  7:48         ` Eric Dumazet
2010-04-17  7:52           ` [PATCH net-next-2.6] rps: rps_sock_flow_table is mostly read Eric Dumazet
2010-04-17  7:57             ` David Miller
2010-04-16 15:35     ` [PATCH v5] rfs: Receive Flow Steering Tom Herbert
2010-04-16 18:15       ` Eric Dumazet
2010-04-16 18:35     ` Tom Herbert
2010-04-16 18:53       ` Eric Dumazet
2010-04-16 20:42         ` Tom Herbert
2010-04-16 21:12           ` Eric Dumazet
2010-04-16 21:25             ` Eric Dumazet
2010-04-17 16:10             ` Eric Dumazet
2010-04-17 17:38               ` Tom Herbert
2010-04-18  0:06                 ` Changli Gao
2010-04-18 11:06                   ` Franco Fichtner
2010-04-19 20:09               ` David Miller
2010-04-19 20:23                 ` David Miller
2010-04-19 20:32                   ` Eric Dumazet
2010-04-19 21:19                     ` David Miller
2010-04-26  8:41                       ` Eric Dumazet
2010-04-27 21:59                         ` David Miller
2010-04-27 22:08                           ` Eric Dumazet
2010-04-27 22:10                             ` David Miller
2010-04-19 23:38                     ` Changli Gao
2010-04-20  5:59                       ` Eric Dumazet
2010-04-20  7:56                         ` Eric Dumazet [this message]
2010-04-20  8:18                           ` [PATCH net-next-2.6] rps: consistent rxhash David Miller
2010-04-20 12:48                           ` Franco Fichtner
2010-04-20 13:16                             ` Eric Dumazet
2010-04-20 14:03                               ` Franco Fichtner
2010-04-20 14:57                                 ` Eric Dumazet
2010-04-20 21:41                                   ` David Miller
2010-04-20 23:35                                     ` Changli Gao
2010-04-20 23:38                                       ` David Miller
2010-04-21 19:12                                     ` Tom Herbert
2010-04-23 20:44                                       ` David Miller
2010-05-06  8:06                                       ` David Miller
2010-05-06 14:45                                         ` Tom Herbert
2010-04-20 15:09                             ` Tom Herbert
2010-04-21  9:29                               ` Franco Fichtner
2010-04-21  9:39                                 ` Eric Dumazet
2010-04-21 11:06                                   ` Franco Fichtner
2010-04-21 11:16                                     ` Eric Dumazet
2010-04-20 15:04                         ` [PATCH v5] rfs: Receive Flow Steering Tom Herbert
2010-04-20 15:39                           ` Eric Dumazet
2010-04-16 19:37   ` Eric Dumazet
2010-04-16 22:49     ` David Miller
2010-04-16 22:53       ` David Miller
2010-04-16 22:57         ` David Miller
2010-04-17  0:22           ` Tom Herbert
2010-04-17  0:58             ` David Miller
2010-04-16 11:57 ` Andi Kleen
2010-04-16 13:32   ` jamal
2010-04-16 13:42     ` Andi Kleen
2010-04-16 14:05       ` jamal
2010-04-16 15:28         ` Andi Kleen

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=1271750198.3845.216.camel@edumazet-laptop \
    --to=eric.dumazet@gmail.com \
    --cc=davem@davemloft.net \
    --cc=netdev@vger.kernel.org \
    --cc=therbert@google.com \
    --cc=xiaosuo@gmail.com \
    /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