Re: Dynamic IP hack (PR#294)

Erik Corry (erik@arbat.com)
Tue, 20 Oct 1998 00:22:42 +0200

On Tue, Oct 20, 1998 at 12:19:43AM +0100, Alan Cox wrote:
> Erik Corry wrote:
> > There's one more thing we really ought to fix, and that is
> > telling the application that its socket has died.
>
> It hasn't died. Its source interface has become temporarily unavailable.

The sysadm knows whether failures on a particular interface
are likely to be permanent or temporary, but at the moment
in 2.1 he has no way to tell the system. On most ISPs it
is always permanent. The socket is dead, it just hasn't
stopped moving (and costing money) yet.

Note that the RST-provoking hack has the advantage that if
you are lucky and get your old address back (or if you do
so half the time, or 99% of the time) things 'just work'.

And Andi's solution is per-interface, so it is also usable
for situations where you have DHCP on one port and dynamic
PPP on another.

-- 
Erik Corry erik@arbat.com           Ceterum censeo, Microsoftem esse delendam!


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/