Re: [Openvortex-dev] patch for ac97 fixes/additions

[Jeff Muizelaar]

On Sun, 15 Jun 2003, Ryan Underwood wrote:

>
> Hi,
>
> The attached patches fix some problems and add some features to the
> vortex driver.  They are against an ALSA 0.94 tree that was patched with
> the patch that made its way onto the mailing list earlier.
>
> 1st, the vortex_read was not working because it didn't wait for the PCI
> write to be flushed first.  So it would return a value that was
> prefetched.  This screwed up the ALSA ac97 init functions because they
> write a value to test and read it back, and when they read back the
> wrong value, they would disable some features in the stac codec, such as
> 3d stereo enhancement, etc.  All of those features are now enabled and
> should be tested.
>
> BTW, ReadCodec in the old binary driver never worked correctly either on
> my card.  (It was never used in the driver, but I tried to rely on it
> for debugging the OSS driver and ended up wasting many hours with it)
>
> Also, the hack in au88x0_mixer.c for enabling the output amp and
> changing the attentuation should not be necessary as these are now
> toggleable features as long as the user's particular codec supports
> them.  (I think!)  So they were removed.  If the user's particular card
> needs them, they can set them on.
>
> 2.
> In the ALSA ac97 driver, I added support for the stereo -> 4 speakers
> output.  The rear speakers always run at 100% master volume when this is
> enabled, so the only way their volume will match the front is if the
> front is also at 100%.  Anyone know a way around that?
>
> I also added a button for the phase inversion, which was previously set
> in the alsa_patches.c, but now the user can toggle it (they could in the
> windows driver, so....)
>
> Else:
>
> Does the AU8820 use a STAC9704 ?  (someone with an au8820 tell us)
> I found some more interesting constants in the dissassembled driver
> code.
>
> Let me know if this is ok, and .....

Please send patches in 'diff -u' format

-Jeff