Bad access to IP address in memory

[Russ Allbery]

Julien ÉLIE <julien at trigofacile.com> writes:

> I believe this bug comes from network_sockaddr_sprint because it does
> not handle a 0 family value:

I don't think it should.  0 isn't a valid family, and that's intended to
be generic code.  An all-zero address, including a zero family, doesn't
mean localhost; it's an invalid value.  It just happens to be what innd
uses to indicate that the given channel isn't a network channel at all.
(So the real bug is in ever passing such a thing to any network_*
function, since it isn't a network address or connection.)

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.