> On 18 Aug 97 at 10:07, Andrew E. Mileski wrote:
>
> > > I have tried the 2.1.43 kernel with the AWE32 support.
> > >
> > > My soundcard is a SB32 with 512kB RAM (SIMM) installed in the
> > > sockets. My card does not have the "Advanced Signal Processor, AWE",
> > > but otherwise has the EMU8000 stuff.
> >
> > Mine has the ASP.
The "Advanced Signal Processor" (ASP) option has nothing to do with "AWE"
or EMU8000. It just an useless chip introduced when Creative released the
original SB16. This chip was supposed to be used for speech
recognition/synthesis and other heavy DSP tasks but in practice it has no
use (there is no software that supports it).
Having the ASP chip there doesn't affect the AWE driver in any ways. It's
just a dead piece of silicon in Linux.
> As far as I have learned there is some mysterious reason that
> prevents the PnP BIOS to activate the "WaveTable" device it seems.
> The driver then does not detect the AWE32. Even though I have read
> the PnP spec, I did not find a clue that a device can request itself
> not being initialized. The idea to have to use isapnp and a sound
> module seems not the very best solution.
The synth device has some requirements about reletive offsets of the I/O
ports that can't be expressed using PnP resource (EEPROM) information. For
this reason the card lists only the first EMU8000 I/O port in it's
resource data. The other two register blocks are hidden and only an
initialization routine specially written for AWE32 can handle them. Of
cause it's possible to implement the BIOS so that it recognizes AWE32 and
can initialize it but...
Best regards,
Hannu
-----
Hannu Savolainen (hannu@4front-tech.com)
http://www.4Front-Tech.com/oss.html (Open Sound System (OSS))
http://personal.eunet.fi/pp/voxware (OSS Free/TASD/VoxWare)