[Bug 667850] rtpsession: creation should be signaled before validation

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=667850
  GStreamer | gst-plugins-good | unspecified

--- Comment #11 from Wim Taymans <wim.taymans at gmail.com> 2012-05-10 09:10:31 UTC ---
(In reply to comment #10)
> Well that seams count intuitive to me .. I would say that any time a new ssrc
> is seen then on_new_ssrc should be called, and only once that ssrc has been
> validated as good then on_validated_ssrc should be called. But to be fair you
> did write this right :).

> 
> Also although the code in process_rtp calls on_validated first I think in
> actual fact the first RTP packet of a new SSRC will cause the source to be
> placed on probation and only after that probation has completed will it be
> validated. That means the first time on_new_ssrc will be called and the second
> packet will cause on_validated_to be called.

It is true, I was hoping that on_new_ssrc would only be called for validated
new sources but I can see in the code that it isn't..

Because this is the case, I agree that on_new_ssrc goes first and then
validated.  I'm wondering however if the on_new_ssrc is done correctly.

AFAIU the spec, you should do minimal processing on SSRCs that you have seen
only once because an attacker might simply send random packets and cause
massive amounts of on_new_ssrc signals and overflow the session table. Those
new unvalidated SSRCs should probably be put into a separate (limited in size)
table and the docs for on_new_ssrc should state that this new SSRC might not
validate or something.

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.