On 07/24/2012 05:10 PM, Mark Andrews wrote:
No. It was a kernel bug. The kernel wouldn't let you un-bind the socket. When you sent to 127.0.0.1 the local address was set to 127.0.0.1 then when you sent to some other address 127.0.0.1 was used as the source address which doesn't work. Modern resolvers check for this and open a new socket.
Thanks, Mark. I should have read a bit more carefully: "in some Berkeley-derived TCP/IP implementations" was the key phrase.
I did some more careful searching, and it looks like this topic came up on the list back in 2007:
https://lists.isc.org/pipermail/bind-users/2007-February/065751.html Good to know that the resolver handles this nowadays. John _______________________________________________ Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
