All of lore.kernel.org
 help / color / mirror / Atom feed
From: Pablo Neira <pablo@eurodev.net>
To: Harald Welte <laforge@netfilter.org>,
	netfilter-devel@lists.netfilter.org
Subject: init_conntrack optimization
Date: Wed, 25 Feb 2004 13:40:05 +0100	[thread overview]
Message-ID: <403C97A5.7070109@eurodev.net> (raw)

[-- Attachment #1: Type: text/plain, Size: 778 bytes --]

Hi Harald and list!

I patched the init_conntrack function to optimize the expectation handling.

With this patch, if an expectation is not found for the conntrack, it 
will look for an helper and go out, this way we save three "if's". By 
other hand, if an expectation is found we save one "if".

I consider that the general case is that an expectation is not found, 
isn't it? because most connection tracked by the system don't need a 
helper, so they not need to handle expectations.

BTW, can we consider that if there's no helper for a given tuple, it 
won't have expectations? because while writing the patch i had more 
ideas, maybe this could be optimized a bit more. Well, I'm not sure...

hope that i'm not missing anything! if so, please let me know.

cheers,
Pablo

[-- Attachment #2: init_conntrack-optimize.patch --]
[-- Type: text/plain, Size: 3644 bytes --]

diff -Nru --exclude .depend --exclude '*.o' --exclude '*.ko' --exclude '*.ver' --exclude '.*.flags' --exclude '*.orig' --exclude '*.rej' --exclude '*.cmd' --exclude '*.mod.c' --exclude '*~' linux-2.6.3-old/net/ipv4/netfilter/ip_conntrack_core.c old-compiled/linux-2.6.3/net/ipv4/netfilter/ip_conntrack_core.c
--- linux-2.6.3-old/net/ipv4/netfilter/ip_conntrack_core.c	2004-02-22 13:29:30.000000000 +0100
+++ old-compiled/linux-2.6.3/net/ipv4/netfilter/ip_conntrack_core.c	2004-02-25 13:02:35.000000000 +0100
@@ -691,42 +691,50 @@
 			     struct ip_conntrack_expect *, tuple);
 	READ_UNLOCK(&ip_conntrack_expect_tuple_lock);
 
-	/* If master is not in hash table yet (ie. packet hasn't left
-	   this machine yet), how can other end know about expected?
-	   Hence these are not the droids you are looking for (if
-	   master ct never got confirmed, we'd hold a reference to it
-	   and weird things would happen to future packets). */
-	if (expected && !is_confirmed(expected->expectant))
-		expected = NULL;
-
-	/* Look up the conntrack helper for master connections only */
-	if (!expected)
-		conntrack->helper = ip_ct_find_helper(&repl_tuple);
-
-	/* If the expectation is dying, then this is a loser. */
-	if (expected
-	    && expected->expectant->helper->timeout
-	    && ! del_timer(&expected->timeout))
-		expected = NULL;
-
 	if (expected) {
-		DEBUGP("conntrack: expectation arrives ct=%p exp=%p\n",
-			conntrack, expected);
-		/* Welcome, Mr. Bond.  We've been expecting you... */
-		IP_NF_ASSERT(master_ct(conntrack));
-		__set_bit(IPS_EXPECTED_BIT, &conntrack->status);
-		conntrack->master = expected;
-		expected->sibling = conntrack;
-		LIST_DELETE(&ip_conntrack_expect_list, expected);
-		expected->expectant->expecting--;
-		nf_conntrack_get(&master_ct(conntrack)->infos[0]);
-	}
-	atomic_inc(&ip_conntrack_count);
+		/* If master is not in hash table yet (ie. packet hasn't left
+		   this machine yet), how can other end know about expected?
+		   Hence these are not the droids you are looking for (if
+		   master ct never got confirmed, we'd hold a reference to it
+		   and weird things would happen to future packets). */
+		if (!is_confirmed(expected->expectant)) {
+			
+			conntrack->helper = ip_ct_find_helper(&repl_tuple);
+			goto end;
+		}
+
+		/* Expectation is dying... */
+		if (expected->expectant->helper->timeout
+		    && ! del_timer(&expected->timeout)) {
+			goto end;	
+		}
+
+                DEBUGP("conntrack: expectation arrives ct=%p exp=%p\n",
+                       conntrack, expected);
+                /* Welcome, Mr. Bond.  We've been expecting you... */
+                IP_NF_ASSERT(master_ct(conntrack));
+                __set_bit(IPS_EXPECTED_BIT, &conntrack->status);
+                conntrack->master = expected;
+                expected->sibling = conntrack;
+                LIST_DELETE(&ip_conntrack_expect_list, expected);
+                expected->expectant->expecting--;
+                nf_conntrack_get(&master_ct(conntrack)->infos[0]);
+
+		/* this is a braindead... --pablo */
+	        atomic_inc(&ip_conntrack_count);
+	        WRITE_UNLOCK(&ip_conntrack_lock);
+
+	        if (expected->expectfn)
+			expected->expectfn(conntrack);
+
+		goto ret;
+        } else 
+                conntrack->helper = ip_ct_find_helper(&repl_tuple);
+
+end:	atomic_inc(&ip_conntrack_count);
 	WRITE_UNLOCK(&ip_conntrack_lock);
 
-	if (expected && expected->expectfn)
-		expected->expectfn(conntrack);
-	return &conntrack->tuplehash[IP_CT_DIR_ORIGINAL];
+ret:	return &conntrack->tuplehash[IP_CT_DIR_ORIGINAL];
 }
 
 /* On success, returns conntrack ptr, sets skb->nfct and ctinfo */

             reply	other threads:[~2004-02-25 12:40 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-02-25 12:40 Pablo Neira [this message]
2004-02-25 14:40 ` init_conntrack optimization Harald Welte
2004-02-25 15:58   ` Pablo Neira

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=403C97A5.7070109@eurodev.net \
    --to=pablo@eurodev.net \
    --cc=laforge@netfilter.org \
    --cc=netfilter-devel@lists.netfilter.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.