Change in osmo-mgw[master]: generate shorter 'I:' conn IDs

This is merely a historical archive of years 2008-2021, before the migration to mailman3.

A maintained and still updated list archive can be found at https://lists.osmocom.org/hyperkitty/list/gerrit-log@lists.osmocom.org/.

Neels Hofmeyr gerrit-no-reply at lists.osmocom.org
Thu Aug 30 14:19:22 UTC 2018


Neels Hofmeyr has posted comments on this change. ( https://gerrit.osmocom.org/10676 )

Change subject: generate shorter 'I:' conn IDs
......................................................................


Patch Set 1:

Note: *not* CallID, just *Connection ID*.
CallID is the one that the MGCP clients send to the MGW to indicate which call reference is served, like an MNCC call reference.

The Connection ID in contrast only is the name of one connection on an endpoint, sort of like a fancy array index. Osmo-MGW invents these freely, and the clients only ever echo them back later to reference the particular "half call leg" on an endpoint.

For example, osmo-bsc does a CRCX on 1 at mgw.
osmo-mgw replies with an OK, including "I: 1234abcd".
osmo-bsc remembers ths 1234abcd as the MGCP referene to pass for MDCX or DLCX later.

The proper way would be to refactor osmo-mgw to have a maximum connection id length and make it configurable from vty what length is actually dealt out. But in fact, the length of the Conn ID is entirely and solely up to osmo-mgw to determine. If we set one specific conn ID to be generated, they will *always* be that length, because no other entity besides osmo-mgw is allowed to create these IDs.

So we cannot modify libosmo-mgcp-client to create shorter IDs, because osmo-mgw creates these IDs. libosmo-mgcp-client should still be able to receive and handle any length of Connection ID it receives from osmo-mgw, we're just making osmo-mgw hand out shorter ones.

But true, if a client sends a Connection ID that is too long, we should make sure to mismatch and not truncate that when parsing the incoming message. I think that's the case, but let me verify.

Does this patch make more sense now?


-- 
To view, visit https://gerrit.osmocom.org/10676
To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings

Gerrit-Project: osmo-mgw
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Ia290c22a91fca0e5aa44515fca6df00064aff100
Gerrit-Change-Number: 10676
Gerrit-PatchSet: 1
Gerrit-Owner: Neels Hofmeyr <nhofmeyr at sysmocom.de>
Gerrit-Reviewer: Harald Welte <laforge at gnumonks.org>
Gerrit-Reviewer: Jenkins Builder (1000002)
Gerrit-Reviewer: Neels Hofmeyr <nhofmeyr at sysmocom.de>
Gerrit-Comment-Date: Thu, 30 Aug 2018 14:19:22 +0000
Gerrit-HasComments: No
Gerrit-HasLabels: No
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/gerrit-log/attachments/20180830/dff3eba1/attachment.htm>


More information about the gerrit-log mailing list