|Subject:||Re: ipv6 and ipv4 interaction weirdness|
|From:||Simon Kelley <simon@xxxxxxxxxxxxxxxxx>|
|Date:||Fri, 11 Mar 2005 14:21:02 +0000|
|Cc:||linux-kernel@xxxxxxxxxxxxxxx, netdev@xxxxxxxxxxx, pekkas@xxxxxxxxxx, yoshfuji@xxxxxxxxxxxxxx|
|User-agent:||Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.3) Gecko/20041007 Debian/1.7.3-5|
I just had some issues with ssh and trying to get it to bind to all ipv6 and ipv4 addresses to it via :: and 0.0.0.0. The problem was that it'd only let one succeed. If 0.0.0.0:22 was successful then :: port 22 could not happen and neither could my ipv6 addy port 22 as it would get the 'address already in use' error from bind(). The reverse was also true. If it bound to :: port 22 then 0.0.0.0:22 would fail.
A solution is to set the IPV6_V6ONLY sockopt on the IPv6 socket (or just use IPv6 sockets and their ability to accept IPv4 connections in a corner of the IPv6 address space).
It seems unlikely that a released ssh would have that problem, but I haven't checked.
|<Prev in Thread]||Current Thread||[Next in Thread>|
|Previous by Date:||Re: [stable] [BK PATCHES] 2.6.x net driver oops fixes, Daniele Venzano|
|Next by Date:||Re: More IPSEC trouble, Steve Hill|
|Previous by Thread:||Re: ipv6 and ipv4 interaction weirdness, CaT|
|Next by Thread:||2.6.10 - "netdev=" kernel boot commands and the Intel e1000 nic, Alex Upton|
|Indexes:||[Date] [Thread] [Top] [All Lists]|