On 13/03/07, Shlomi Fish <[EMAIL PROTECTED]> wrote:
On Monday 12 March 2007, Amos Shapira wrote: > On 13/03/07, Oron Peled <[EMAIL PROTECTED]> wrote: > > First, terminology: s/switching/forwarding/ > > > > Either: > > echo 0 > /proc/sys/net/ipv4/ip_forward > > Or the equivalent: > > sysctl -w net.ipv4.ip_forward=0 > > > > Now you can send/receive packets from all your network interfaces > > but there's no routing for packets among them. > > I think what the Shlomi was refering to is that the kernel will > automatically "optimize away" the hardware layer because it knows that the > address it sends to is actually its own. > > I'm not 100% sure he's right (it used to be that you had to have an Who do you mean by "he"?
You, Shlomi. I've since saw that you were satisfied with Oron's reply. I'm still curious, though - will the Linux kernel in your situation go through the hassle to assemble and transmit Ethernet frames or will it notice that it's actually sending to its own address and just say "sod it - I'll just copy the buffer in memory" (which is what I half expect)? --Amos
