[OpenSIPS-Devel] [ opensips-Patches-2887877 ] nathelper : allow millisecond timeout for Rtpproxy reply

[SourceForge.net]

Patches item #2887877, was opened at 2009-10-28 14:29
Message generated for change (Comment added) made by bogdan_iancu
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=1086412&aid=2887877&group_id=232389

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: modules
Group: trunk
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Christophe Sollet (csollet)
Assigned to: Bogdan-Andrei Iancu (bogdan_iancu)
Summary: nathelper : allow millisecond timeout for Rtpproxy reply

Initial Comment:
Hello,

Current parameter 'rtpproxy_tout' can be used to set a timeout with second resolution.

Sadly, one second is generally too long to wait generating a lot of SIP retransmission

This patch add a 'rtpproxy_tout_ms' parameter to set a timeout with millisecond resolution. The used timeout become (rtpproxy_tout * 1000) + rtpproxy_tout_ms)

Setting 'rtpproxy_tout' to 0, this new parameter can be used to set a sub second timeout.

----------------------------------------------------------------------

>Comment By: Bogdan-Andrei Iancu (bogdan_iancu)
Date: 2009-10-30 13:22

Message:
I was thinking also about BC, but mixing two params may be even more
confusing.

if we change it to accept floating vals, we need to change the type of the
param from INT to STR (modparam does not natively support floats)....So, if
we have to change something and to have BC, I suggest:
- adding a new param 'rtpproxy_timeout' - accepts an INT with full timeout
in milisecs
- 'rtpproxy_tout' will be keept, but warning will be fired if used (as
will become obsolete)
- if both are used from the script, some error should be fired...

What do you say?

Regards,
Bogdan

----------------------------------------------------------------------

Comment By: Christophe Sollet (csollet)
Date: 2009-10-30 13:13

Message:
Hi Bogdan,

In fact, our first implementation  was just to change to millisecond. This
new attribute was added only to keep BC.
But you're right, the best solution is to accept floating numbers. New
patch will come next week.

Regards,
Christophe.


----------------------------------------------------------------------

Comment By: Bogdan-Andrei Iancu (bogdan_iancu)
Date: 2009-10-30 13:04

Message:
Hi Christophe,

Thanks for the patch - shouldn't be simpler to use only one parameter
('rtpproxy_tout') and to change the value from secs to milisecs ? or to
keep secs, but to accept floating numbers ("0.01") ?

Regards,
Bogdan

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=1086412&aid=2887877&group_id=232389