On Sat, 2012-11-03 at 18:29 +0100, martin f krafft wrote: > also sprach Ben Hutchings <b...@decadent.org.uk> [2012.11.03.1515 +0100]: > > int one = 1; > > setsockopt(sock, IPPROTO_IP, IP_FREEBIND, &one, sizeof(one)); > > bind(sock, ...); > > > > Or we set the net.ipv4.ip_nonlocal_bind sysctl globally, but this is > > probably not a good idea. (Note that 'v4' there is not a typo; this > > would actually affect both v4 and v6.) > > Generally, one should not be able to bind() non-local sockets, don't > you think?
Whyever not? You can get a socket bound to a non-local address even without this option, if the address is removed after you bind. The restriction to current local addresses is only a sanity check which may or may not be useful. Ben. -- Ben Hutchings I'm always amazed by the number of people who take up solipsism because they heard someone else explain it. - E*Borg on alt.fan.pratchett
signature.asc
Description: This is a digitally signed message part