Hi Andreas, On Wed, Sep 20, 2017 at 05:37:52PM +0200, Andreas Schultz wrote: > I think we had this discussion before. The sending IP and port are not part > of the identity of the PDP context. So IMHO the sender is permitted > to change the source IP at random.
Thanks for the reminder: You are correct, at least in the uplink case (MS->GGSN) where there is mobility of the MS. In the downlink case (GGSN->MS), which is the "sending" part for the kernel GTP code used at a GGSN, I'm not sure if that theory holds true in reality. Do you agree that the current behavior of not using automatic source address selection for encapsulated GTP packets but rather using the source address of the socket is intended? Do you further agree that the dst_cache support patch by Tom retains that intended behavior and it should be merged? -- - Harald Welte <lafo...@gnumonks.org> http://laforge.gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)