On Wed, 2020-11-25 at 10:54 +0100, Paolo Abeni wrote:
Note: I think this later change is actually an independed bug fix, I
fear a peer sending an MP_JOIN with an invalid token will get back an
MPJ_SYNACK, while the subflow_req->msk will be cleared. Later
in subflow_syn_recv_sock(), we will probably get a NULL ptr dereference
@Davide: can we somehow easily check the above with a packet drill?
yes I can try that. I'm now looking once again at the MP_JOIN script
(issue #94) and trying to fix this in a way or another (the problem is
apparently bad values of [s,c]addr[0,1] in the environment. But at least
generation of an inbound MP_JOIN SYN should be under control. I do a
quicjk tyest and let you know, ok?