On Tue, May 19, 2015 at 04:28:45PM -0400, David Miller wrote:

From: Andy Gospodarek <redacted>
Date: Tue, 19 May 2015 15:47:32 -0400

quoted

Are you actually saying that if users complain loudly enough about
the current behavior (not the change Roopa has proposed) that you
would be open to considering a change the current behavior?

I am saying that we have a contract with users not to break existing
behavior.  Full stop.

I stepped away from the ledge and now definitely agree that the default
behavior should stay the same in the offload or non-offload case.

There is probably room for some level of notification (even if just a
printk) to let one know that offloading is disabled and it would be
great if there was a way to set fib_offload_disabled to false again, but
that's not going to happen until patches are submitted, so I'll get to
that instead of complaining more.  :-)