Hello,
The mail with the patch was not delivered to netdev ML for some reasons.
Resending it..
The function sit_add_v4_addrs() calls ipv6_add_addr() with plen 64 if local
endpoint of tunnel is not configured, so I do not see a reason to set plen
to 128 if local endpoint is configured.
We are running patched kernel in our test-lab for about week and it is
working fine.
On Fri, Sep 26, 2003 at 08:19:53PM -0700, David S. Miller wrote:
On Fri, 26 Sep 2003 19:54:05 +0200
Jan Oravec [off-list ref] wrote:
quoted
please apply the following patch.
It sets the prefix length 64 instead of 128 for link-local addresses on sit interfaces.
I've seen this patch before, and I remember it being rejected
by other ipv6 developers.
Do the USAGI or other folks remember what is going on here?
--- linux/net/ipv6/addrconf.c.orig 2003-09-08 21:50:29.000000000 +0200
+++ linux/net/ipv6/addrconf.c 2003-09-26 19:37:27.000000000 +0200
@@ -1681,7 +1681,7 @@
}
if (addr.s6_addr32[3]) {
- ifp = ipv6_add_addr(idev, &addr, 128, scope, IFA_F_PERMANENT);
+ ifp = ipv6_add_addr(idev, &addr, 64, scope, IFA_F_PERMANENT);
if (!IS_ERR(ifp)) {
spin_lock_bh(&ifp->lock);
ifp->flags &= ~IFA_F_TENTATIVE;
Best Regards,
--
Jan Oravec
6COM s.r.o.
http://www.6com.sk