[PATCH] ipv6: order addresses by scope

[PATCH] ipv6: order addresses by scope

[Brian Haley]

[-- Attachment #1: Type: text/plain, Size: 250 bytes --]

If IPv6 addresses are ordered by scope, then ipv6_dev_get_saddr() can 
break-out of the device addr_list for() loop when the candidate source 
address scope is less than the destination address scope.

Signed-off-by: Brian Haley <brian.haley@hp.com>

[-- Attachment #2: ip6order.patch --]
[-- Type: text/x-patch, Size: 1287 bytes --]

diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 445006e..e1d6a6f 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -509,6 +509,25 @@ void inet6_ifa_finish_destroy(struct ine
 	kfree(ifp);
 }
 
+static void
+ipv6_link_dev_addr(struct inet6_dev *idev, struct inet6_ifaddr *ifp)
+{
+	struct inet6_ifaddr *ifa, **ifap;
+
+	/*
+	 * Each device address list is sorted in order of scope -
+	 * global before linklocal.
+	 */
+	for (ifap = &idev->addr_list; (ifa = *ifap) != NULL;
+	     ifap = &ifa->if_next) {
+		if (ifp->scope > ifa->scope)
+			break;
+	}
+
+	ifp->if_next = *ifap;
+	*ifap = ifp;
+}
+
 /* On success it returns ifp with increased reference count */
 
 static struct inet6_ifaddr *
@@ -574,8 +593,7 @@ ipv6_add_addr(struct inet6_dev *idev, co
 
 	write_lock(&idev->lock);
 	/* Add to inet6_dev unicast addr list. */
-	ifa->if_next = idev->addr_list;
-	idev->addr_list = ifa;
+	ipv6_link_dev_addr(idev, ifa);
 
 #ifdef CONFIG_IPV6_PRIVACY
 	if (ifa->flags&IFA_F_TEMPORARY) {
@@ -982,7 +1000,7 @@ int ipv6_dev_get_saddr(struct net_device
 					continue;
 			} else if (score.scope < hiscore.scope) {
 				if (score.scope < daddr_scope)
-					continue;
+					break; /* addresses sorted by scope */
 				else {
 					score.rule = 2;
 					goto record_it;

Re: [PATCH] ipv6: order addresses by scope

[David Miller]

From: Brian Haley <brian.haley@hp.com>
Date: Fri, 09 Jun 2006 14:14:50 -0400

> If IPv6 addresses are ordered by scope, then ipv6_dev_get_saddr() can 
> break-out of the device addr_list for() loop when the candidate source 
> address scope is less than the destination address scope.
> 
> Signed-off-by: Brian Haley <brian.haley@hp.com>

This patch looks fine to me, it's an optimization and not a
correctness issue.

But I'd like an ACK from an ipv6 developer, such as Yoshifuji, before
adding this patch.

Thanks.

Re: [PATCH] ipv6: order addresses by scope

[YOSHIFUJI Hideaki / 吉藤英明]

In article <20060710.150045.128593827.davem@davemloft.net> (at Mon, 10 Jul 2006 15:00:45 -0700 (PDT)), David Miller <davem@davemloft.net> says:

> > If IPv6 addresses are ordered by scope, then ipv6_dev_get_saddr() can 
> > break-out of the device addr_list for() loop when the candidate source 
> > address scope is less than the destination address scope.
> > 
> > Signed-off-by: Brian Haley <brian.haley@hp.com>
> 
> This patch looks fine to me, it's an optimization and not a
> correctness issue.
> 
> But I'd like an ACK from an ipv6 developer, such as Yoshifuji, before
> adding this patch.

I agree, too.
Note: People should not depend on the order of addresses.

Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>

--yoshfuji

Re: [PATCH] ipv6: order addresses by scope

[David Miller]

From: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Date: Mon, 10 Jul 2006 18:18:16 -0400 (EDT)

> In article <20060710.150045.128593827.davem@davemloft.net> (at Mon, 10 Jul 2006 15:00:45 -0700 (PDT)), David Miller <davem@davemloft.net> says:
> 
> > > If IPv6 addresses are ordered by scope, then ipv6_dev_get_saddr() can 
> > > break-out of the device addr_list for() loop when the candidate source 
> > > address scope is less than the destination address scope.
> > > 
> > > Signed-off-by: Brian Haley <brian.haley@hp.com>
> > 
> > This patch looks fine to me, it's an optimization and not a
> > correctness issue.
> > 
> > But I'd like an ACK from an ipv6 developer, such as Yoshifuji, before
> > adding this patch.
> 
> I agree, too.
> Note: People should not depend on the order of addresses.
> 
> Acked-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>

Thank you for reviewing, patch applied.