Joshua C. Colp

> @asteriskteam Would someone mind giving this review in particular a look soon, if they have time? Don't want to nag, but because of the monolithic architecture of chan_dahdi, it's...

How frequent is frequent? Did this just start happening? A console log leading up to the crash would also be nice to have, to see the state of things.

It feels like a race condition/reference counting issue/off-nominal where the RTP instance is already destroyed, but there is still a pointer on the session_media.

Or another thread jumping in and destroying it, what that thread would be... remains unknown.

Is it possible to get a debug level log to see if it shows prior state or operations on the underlying channel, session, and RTP instance?

This isn't the identified area. This is where it crashes, yes, but that doesn't mean it is the direct cause of it. It could be elsewhere within PJSIP land for...

Honestly, the more the better to determine the root cause. I don't have any specifics because there's not enough information to begin with to scope it.

It would seem to indicate a reference counting issue such that things are being prematurely destroyed, probably due to the "Unable to cancel schedule ID".

To be specific - we won't drop what we're doing and get right on this. I've moved it up in the queue, but there's still no time frame and other...

You're going to need to provide dialplan, configuration, and step by step instructions that reproduce this issue.