[Asterisk-bsd] Which drivers version to use for FreeBSD 5.4?

[Alexander Timoshenko]

Thomas Sandford wrote:

> I'm looking for the best version of the Zaptel drivers to use for a 
> FreeBSD 5.4 box.
>
> Specifically the current ports tree drivers (0.10_1) have problems 
> with the FXS board on my TDM411B.
>
> The choices I appear to have are:
> Port (0.10_1)
> SVN Trunk
>    (or nightly snapshot)
> SVN 4.x branch
>   ( which seems to have a snapshot at 
> http://portaone.com/~gonzo/zaptel-freebsd-0.11.tar.gz that I found by 
> mistake when trying to compile the svn head as a port!)
>
> I would have expected the SVN trunk to be the one, except that there 
> appears to be a changelog entry:
>
> "12/10/2005 Alexander Timoshenko <gonzo at pbxpress.com>
>     o Fixed problem with panic on -current and 6.X"
>
> which appears in the Changelog of the 4.X branch and _not_ in the 
> trunk Changelog.   ???

    Let me clarify things a bit. There was an ugly hack called zt_clone 
in order to support
multiple device instances for single device node(linux kernel 
behaviour). It caused kernel
panic on 6.X and -CURRENT so i've cleaned things out and rewrite it 
using FreeBSD cloning
API. Unfortunately it relies on devfs and there is no defvs (at least 
functionaity i need) in 4.X
AFAIK. So to keep FreeBSD 4.X supported this 4.x-branch has been 
created. But some time
ago i've got workaround for all systems from Roman Kurakin, so i've 
commited it to 4.x-branch
it is not in trunk because trunk does not cause kernel panic on 6 or 
-current.     

>
> I'm not familiar enough with SVN to be able to work out what is in 
> each change, unfortunately.
>
> PS - any chance of updating the port to a SVN snapshot that contains 
> the fix for the "ProSLIC 3210 version 2 is too old", since it seems to 
> be a fairly common problem looking at the list archives and Google.
>
    I've send update to Maxim Sobolev and waiting for him to commit the 
new version of port.  0.11
will be snapshot from 4.x-branch. It will be last update with  hack i've 
mentioned above because
this version is known to work (somewhat :)) and trunk code  has not been 
tested fine yet.