Messages in this thread | | | Date | Mon, 19 Oct 1998 17:15:41 +1300 | From | Chris Wedgwood <> | Subject | Re: Dynamic IP hack (PR#294) |
| |
On Sun, Oct 18, 1998 at 12:47:46PM -0700, Eric Schenk wrote:
> [...] Frankly the real problem is that the internet was never > designed to allow for dynamic IP addresses, and anything we do to > hack them in results in one problem or another. Anyway, I don't > like this solution much, but I can't offer the right solution > either.
Why not do what some other OSs do - that is, attempt to dial on demand via some other mechanism only when the applications needs to, and defer socket creation undit the interface is up?
This might be when an AF_INET socket is first created, or perhaps we defer this until someone either uses connect or bind on the socket. Deferring this means we can be clever, so bind or connect to localhost still works...
Anyone using O_NONBLOCK will have to expect EAGAIN or atleast sanely handle it.
I know this solution isn't optimal - but on another OS, it works for millions on people. (With Win32, you call WSAStartup, which can then do the dial-on-demand thing).
Basically, what I'm saying is address translation is a hack, can't we make the applications smarters, perhaps giving them some help in places by using netlink or whatever to trigger DOD?
-Chris
P.S. Your back!
- 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/
| |