IP fragment overlap fix in 2.0.32

Philip Gladstone (philip@raptor.com)
Tue, 18 Nov 1997 18:21:55 -0500

I suspect that the fix is not entirely correct as once one of these
little fragments is received and inserted into the fragment queue
then the fragment is unlikely to ever be completed. This is
because the ip_done function checks that all the fragments
connect, and the small (by now negative) fragment 'creates a hole'.

I'm not sure that this situation will ever arise in normal use, but


Philip Gladstone                           +1 617 487 7700
Raptor Systems, Waltham, MA         http://www.raptor.com/
Our new daughter: http://www.mwmc.com/Extweb/Cybernursery/17423662.htm