MME respons to S1AP on wrong PPID about open5gs HOT 4 CLOSED

Interesting. Seems like the value 18 is not stored in net_sock->ppid, but in net_sock->sock_id. Can that be a coincidence?

Thread 39 "nextepc-mmed" hit Breakpoint 1, net_write (net_sock=0x555555a02508 <net_pool+4200>, buffer=0x555555a70c30 <cluster_2048_pool+6800> " \021", size=27, dest_addr=0x555555a02518 <net_pool+4216>, addrsize=16) at net_lib.c:514
514     {
(gdb) p *net_sock
$2 = {type = 0, proto = 132, ppid = 0, sock_id = 18, remote = {sin_family = 2, sin_port = 15246, sin_addr = {s_addr = 3060050112}, sin_zero = "\000\000\000\000\000\000\000"}, opt = 0, app_index = 1, sndrcv_errno = 11}

from open5gs.

Please see #6 for a proposed fix. I'm actually surprised this was working at all or has passed any tests in the current state. Or maybe I'm misunderstanding the code. In any case, this fixes the PPID in my SCTP DATA chunks from MME -> ENB

from open5gs.

Correct! I've missed it. I saw the same result in './testepc'.
Thanks a lot!

from open5gs.

proposed fix #6 has been merged, closing issue.

from open5gs.