Jason Thorpe <thorpej@nas.nasa.gov> writes:

> 
> On Tue, 7 May 1996 23:55:35 +0200 (MET DST) 
>  Paul Kranenburg <pk@cs.few.eur.nl> wrote:
> 
>  > Note that some releases of "vat" for SunOS try to be clever and detect
>  > what system they're running on (SunOS vs. BSD) by trying to manipulate
>  > `/dev/audioctl'. Since the process, when run under NetBSD, is still in
>  > SunOS emulation mode, this defeats the sunos_compat ioctl conversion
>  > code.
>  > 
>  > Since I understand there's now a source release of vat, this point should
>  > now be moot pretty soon.
> 
> Unfortunately, I've had bad luck getting a native NetBSD/sparc vat to 
> work.  Charles once mentioned that this might be a problem in the NetBSD 
> audio layer, and if that's the case, I'd like to try and get it fixed.
> 
> Charles: could you provide some details of what the problem might be?

The problem is that when John Brezak whacked on the audio layer, he
put the numbers for input and output ports in the same space, which
forces them to be different from the numbers used under SunOS.  Thus
the AUDIO_SETINFO ioctl doesn't set the ports as you might expect.  I
still can't see a good reason for this, and I think it should be
fixed, as well as reintroducing the generic AUDIO_* names for
applicable ports, so that software can have some hope of being
device-independent without being required to use the mixer interface.