Date: Sun, 17 Mar 1996 21:18:03 +0100 (MET) From: guido@gvr.win.tue.nl (Guido van Rooij) To: fenner@freefall.freebsd.org (Bill Fenner) Cc: CVS-committers@freefall.freebsd.org, cvs-all@freefall.freebsd.org, cvs-sys@freefall.freebsd.org Subject: Re: cvs commit: src/sys/netinet in.c Message-ID: <199603172018.VAA21057@gvr.win.tue.nl> In-Reply-To: <199603151708.JAA12346@freefall.freebsd.org> from "Bill Fenner" at Mar 15, 96 09:08:09 am
next in thread | previous in thread | raw e-mail | index | archive | help
Bill Fenner wrote: > > fenner 96/03/15 09:08:08 > > Modified: sys/netinet in.c > Log: > Allow SIOCGIFBRDADDR and SIOCGIFNETMASK to return information about > aliases, if the alias address was passed in the struct ifreq. > Default to first address on the list, for backwards compatibility. > What happens now when you want to do a broadcast for any ip address on a system? The rpc code would generate <n> duplicate broadcasts when there were <n-1> aliases on the same subnet as an interfaces' first ip address. What happens now is that you get one broadcast plus <n-1> directed udp packets to the alias ip itsself. I don't really know a solution. Suppose an interface ed0 has ip address 192.1.1.1 (mask 0xffffff00) and an alias 192.1.1.2. Now the code will return the following broadcast addresses: 192.1.1.0 for the 192.1.1.1 case and 192.1.1.2 for the 192.1.1.2 case I believe that is wrong. But to just make checks like `if the interface has another ip address within the same subnet with a netmask != 0xffffffff then return that other netmask' is a gross hack :-( -Guido
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199603172018.VAA21057>