The CS4236B is a vastly improved superset of the AD1848.  It has many
more registers and can do full duplex with different modes for record
and play.  The existing wss code seems to be tuned for AD1848 and is
quite limited in what it can do, viz it is not duplex.  I began
organizing the pieces of the wss driver into what would/could become
a CS4236/37B driver.  I can dust this off and see about committing it
then we can procede to enhance it.  I was/am worried whether handling
the CS42356 extensions in the wss driver might break the simple
functionality and I have no ad1848 hardware to test it on.

Note 1: The hardware claims to be SB compatible -- isn't everything.
Playback worked fine, but recording didn't.  The problem with 
recording is that if you access the wss registers you can get an extra
20db boost which you need.

Note 2: I did my experiments on a CS4237B in a TP 560X.

Note 3: I did not understand why the wss driver was split into a wss.c
top part and a ad1848.c bottom part.  I just fused these two to
together to get a new cs423x.c driver.  Was this a "bad" thing to do?


Havard.Eidnes@runit.sintef.no writes:

> Hi,
> 
> I'm thinking of getting a newer i386-based machine, and I'd like
> for NetBSD's audio to work well on the machine.  Quite a few of
> the machines I've been looking at have the Cirrus Logic CS4236B
> "audio system on a chip" audio interface.  Which, if any, of the
> drivers in NetBSD-current will support that hardware?  What about
> NetBSD-1.3.2?  Can it operate in full duplex?
> 
> The front of the spec booklet says it is compatible with the WSS
> standard and will run with software written for SoundBlaster and
> SoundBlaster Pro, but it's not entirely clear to me how much of
> that is colored and/or covered by Microsoft-related software,
> which isn't of much use to me, of course.
> 
> Just FYI, the machines I'm primarily looking at are the newish
> Dell PCs.
> 
> - H=E5vard