I think that Vendor ID sending is kind of screwed up, particularly our own Vendor ID.
Each case where we emit our Vendor ID is conditional on c->send_vendorid (good). - aggr_inI1_outR1_continue2_tail should send it but does not - aggr_outI1_tail sends VID_LIBRESWANSELF (== libreswan_vendorid) - main_outI1 sends pluto_vendorid - main_inI1_outR1 sends pluto_vendorid Is there any reason for these to send different collections of Vendor IDs? I would guess not. I'm working on function to encapsulate this once rather than have four diverging chunks of code. I don't think you answered my question. Which is correct: libreswan_vendorid or pluto_vendorid. Surely not both! | From: Paul Wouters <[email protected]> | | On Mon, 30 Jul 2018, D. Hugh Redelmeier wrote: | | > Some of our code emits our vendorid payload using libreswan_vendorid | > as our Vendor ID using | > | > out_vid( ..., VID_LIBRESWANSELF); | > | > and some uses pluto_vendorid, via ikev1_out_generic_raw(). | > | > Those are two different things, | > | > Which one is correct? | | both :) | | VID_LIBRESWANSELF is our compiled in version based on various compile | time settings. | | pluto_vendorid is our vendorid which can be set with the myvendorid= | | eg: | | config setup | myvendorid="paulswan" | send-vendorid=yes | | | If send-vendorid=yes and no myvendorid=, the default to use is whatever | is compiled in as VID_LIBRESWANSELF. | | Paul _______________________________________________ Swan-dev mailing list [email protected] https://lists.libreswan.org/mailman/listinfo/swan-dev
