Since OS X 10.7, we must declare whether we want the IPv6 sockets API to
work like RFC 2292 or RFC 3542. As far as I know, we are compatible with
both, so just pick the more recent one, which Apple says will become the
default in the future.
For each rtattr we add to the netlink message, we were adding
RTA_LENGTH(rtattr->rta_len) to the length of the netlink message. But
rtattr->rta_len was already calculated as RTA_LENGTH of something, and
doing RTA_LENGTH twice made the length 4 bytes longer than it should be.
This caused a log in dmesg:
netlink: 4 bytes leftover after parsing attributes.
or
netlink: 8 bytes leftover after parsing attributes.
if there was an IPv6 scope ID (because that causes two rtattrs instead
of one).
The new code is consistent with the rtnetlink(3) man page, which does
rta->rta_len = sizeof(unsigned int);
req.n.nlmsg_len = NLMSG_ALIGN(req.n.nlmsg_len) + RTA_LENGTH(sizeof(unsigned int));
We do the equivalent
rta->rta_len = sizeof(unsigned int);
req.n.nlmsg_len = NLMSG_ALIGN(req.n.nlmsg_len) + rta->rta_len;
This is set nonzero when there is a scope identifier at the end of
an IPv6 address, like fe80::a8bb:ccff:fedd:eeff%eth0 or
fe80::a8bb:ccff:fedd:eeff%1 on Windows. When this happens, we look up
the interface by index and then act as if it was the interface given by
-e. (But -e always has precedence over this.)
This is set nonzero when there is a scope identifier at the end of
an IPv6 address, like fe80::a8bb:ccff:fedd:eeff%eth0. When this
happens, we look up the interface by index and then act as if it was the
interface given by -e. (But -e always has precedence over this.)
This is set nonzero when there is a scope identifier at the end of an
IPv6 address, like fe80::a8bb:ccff:fedd:eeff%eth0. When this happens, we
add an rtattr with type RTA_OIF to request a particular outgoing
interface.
In my tests, this does the right thing when the address is in fact the
assigned address of the interface; the interface becomes lo instead of
the physical interface name.
This was previously gotten by setting the source address to be the same
of the interface address of the matching route. However this can be
wrong; when making a normal socket connection the source address is
chosen differently. We create a SOCK_DGRAM socket, connect it, and read
the local address with getsockname.
These version allow returning an extension header or other
non–upper-layer protocol if it is the final header before the end of the
packet. This is used to parse the broken packets sent as part of
protocol scan.
* Adding path-mtu.nse for Path MTU Discovery
* Nmap now stores the MTU for interfaces (from SIOCGIFMTU or libdnet)
* Scripts can access the MTU for host.interface via host.interface_mtu
* Nmap prints the MTU for interfaces in --iflist
Instead of sending multiple fragments, Nmap would just send the
original whole packet instead. In some circumstances, Nmap would
fail to send on interfaces with low MTUs (such as SLIP lines) with
no way to bump down packet sizes for transport. [Kris]
It looks like this has been broken in trunk since merging libnetutil,
and since r18037 in the dedup branch.
