Emiliano Della Casa ha scritto:
> Hi again,
> 
> I don't know if this could be of any help, but I ran asterisk in debug
> mode and analyzed the core file using gdb.

This is the same bug i found one month later, i dont have time to work
on but seems simple to track and go over.

I think really the problem is this:

Oct  6 12:41:26 [kernel] lapd: visdn0: MDL-ERROR-INDICATION(H)

From the start i found that the visdn driver suffer the "parallel" or
the "not linear" problem.
If everything is perfect the driver works great, if someone in the
middle hangs the phone not as espected all things go bad.

If you are not using my patch, also, you found more more problems, all
problems are related to:

ast_chan that disappear in the middle of a function (why? simple the
call was closed somewhere else)

to_visdn_chan(ast_chan) that returns an invalid pointer (the same reason).

NOTE: i found a lot of problem linking in the same asterisk server 8
ISDN BRI togheter (ISDN1 -> ISDN2 -> ISDN3 -> ISDN4 etc. etc.) and
using this like a single ISDN gateway, this run very fast isdn dial
and very fast isdn hangup AND there is a very HIGH risk to see a
middle interface going much faster that the first or the last, leaving
visdn in a confusion state!

I think also is a BAD idea to track every pointer as the real problem
is "why" the channel is closed and some function try to work on!

I am waiting some response from the author of visdn... probably is
working on the problem and this kind of work require a lot of time...

Bye!

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Visdn-hackers mailing list
[email protected]
https://mailman.uli.it/mailman/listinfo/visdn-hackers

Reply via email to