[9fans] AC'97

[9fans] AC'97

[Devon H. O'Dell]

I just wanted to give a heads up since I know Dave has students working on this
and it seems silly to have duplicated work. Aki has a driver on 9grid.net that
works for my chipset. I just grabbed:

http://9grid.net/aki/src/9/pc/audio.h
http://9grid.net/aki/src/9/pc/audioac97.h
http://9grid.net/aki/src/9/pc/audioac97mix.h
http://9grid.net/aki/src/9/port/devaudio.c

My card probes with:

#A0: ac97 port 0x18c0 mixport 0x1c00 irq 10
#A0: ac97 warm reset
#A0: ac97 codecs ready: sdin0 sdin1
#A0: ac97 codecs resumed:
#A0: ac97 codec ext: vra spdif

9top% madplay '02 - Birdhouse In Your Soul.mp3'
MPEG Audio Decoder version 0.11.4 (beta) - Copyright (C) 2000 Robert Leslie
 Title: Birdhouse In Your Soul          Artist: They Might Be Giants
 Album: Flood                            Genre: Other
  Year: 1990  Track: 2                 Comment:
ID3: version 2.3.0, flags 0x00, size 2171 bytes
 Album/Movie/Show title: Flood
 Title/Songname/Content description: Birdhouse In Your Soul
 Year: 1990
ID3: unhandled MCDI (Music CD Identifier): flags 0x0000, 804 bytes
 Track number/Position in set: 2
 Content type: Alternative
ID3: unhandled PRIV (Private frame): flags 0x0000, 14 bytes
ID3: unhandled PRIV (Private frame): flags 0x0000, 17 bytes
 Lead performer(s)/Soloist(s): They Might Be Giants
 Length: 200333

(I'm a geek, am I not obliged to like TMBG?)

Now, this is fine and dandy, but I have questions.

Aki, why isn't this on sources? Why didn't you post to 9fans saying that
this existed? I suppose writing a driver is good learning material, but I have
the feeling now that Dave's students are doing a good deal of duplicate
work now.

Is there some reason this can't be included into Plan 9? Was this ever
suggested off-list?

--dho

[9fans] Re: AC'97

[Devon H. O'Dell]

> http://9grid.net/aki/src/9/pc/audioac97.h
> http://9grid.net/aki/src/9/pc/audioac97mix.h

Sorry, that should be
http://9grid.net/aki/src/9/pc/audioac97.c
http://9grid.net/aki/src/9/pc/audioac97mix.c

--dho

Re: [9fans] Re: AC'97

[Bankim Bhavsar]

[-- Attachment #1: Type: text/plain, Size: 579 bytes --]

We've implemented the audio playback for AC97 for ICH controllers and the
project description can be found here
http://www.andrew.cmu.edu/user/bbhavsar/ac97.html

We have tested the code thoroughly and in the final stages of code review
with Prof. Dave Eckhardt.

-Bankim.

On 4/16/07, Devon H. O'Dell <devon.odell@gmail.com> wrote:
>
> > http://9grid.net/aki/src/9/pc/audioac97.h
> > http://9grid.net/aki/src/9/pc/audioac97mix.h
>
> Sorry, that should be
> http://9grid.net/aki/src/9/pc/audioac97.c
> http://9grid.net/aki/src/9/pc/audioac97mix.c
>
> --dho
>

[-- Attachment #2: Type: text/html, Size: 1177 bytes --]

Re: [9fans] Re: AC'97

[Steve Simon]

> We've implemented the audio playback for AC97 for ICH controllers and the
> project description can be found here

Excellent work, this will be widely useful, well done.

-Steve

Re: [9fans] Re: AC'97

[erik quanstrom]

i second the motion.

- erik

Re: [9fans] Re: AC'97

[Devon H. O'Dell]

2007/4/16, Steve Simon <steve@quintile.net>:
> > We've implemented the audio playback for AC97 for ICH controllers and the
> > project description can be found here
>
> Excellent work, this will be widely useful, well done.
>
> -Steve

I agree; I just wonder why this work was duplicated.

--dho

Re: [9fans] AC'97

[Uriel]

On 4/16/07, Devon H. O'Dell <devon.odell@gmail.com> wrote:
> I just wanted to give a heads up since I know Dave has students working on this
> and it seems silly to have duplicated work.

It is not silly, it is the Plan 9 way: duplicate as much work as
possible, and if you do something, bury it somewhere where nobody
could ever find it, god forbid someone could find it useful!

If you poke around 9grid.net you will find also references to a
working gcc port, and now I heard that there are different people
working on ape and non-ape python ports.

Honestly, aren't you happy about the amazing transparency and
efficiency of this development model?

uriel

Re: [9fans] AC'97

[Skip Tavakkolian]

> It is not silly, it is the Plan 9 way:

that's bs.

Re: [9fans] AC'97

[erik quanstrom]

let's just suppose there is already an ac97 driver that does everything
available.  the excercize of writing another driver is still useful.  there's
a lot to be learned through the process.  i saw a lot of good information
on the web page describing the process that wasn't present in the link
to source on 9grid.  i may need to write an ac97 driver for a non-plan9
kernel.  i think i could use some of the tips presented there.

so, uriel, could you explain why these folk's education needs to benefit you
or to fit into some grand development model?

- erik

Re: [9fans] AC'97

[Devon H. O'Dell]

2007/4/16, erik quanstrom <quanstro@coraid.com>:
> let's just suppose there is already an ac97 driver that does everything
> available.  the excercize of writing another driver is still useful.  there's
> a lot to be learned through the process.  i saw a lot of good information
> on the web page describing the process that wasn't present in the link
> to source on 9grid.  i may need to write an ac97 driver for a non-plan9
> kernel.  i think i could use some of the tips presented there.

Similar to what I've said before in this thread. In fact, that link
did indeed provide a good bit of useful information. If only all
drivers were written with this model in mind! Porting would be a lot
easier. My original question was certainly not intended to downplay
the efforts of either party.

> so, uriel, could you explain why these folk's education needs to benefit you
> or to fit into some grand development model?

Wait, this is exactly what I wanted to avoid. Let's kill the flames,
please. Or avoid them before they start.

> - erik
>
>

Re: [9fans] AC'97

[Dave Eckhardt]

> Aki has a driver on 9grid.net that works for my chipset.

That looks like very professional code.  We've wanted AC '97
support for a while, so any way we end up with it is good for
the community.  I hope Aki Nyrhinen chooses to contribute this
work.  If so, it looks as if the students here could contribute
rate-scaling code for codecs with bad clocks, which seem to be
common.

As for the wasted-effort issue... I've been wary about
pre-announcing student projects for a couple of reasons:

1. They don't all pan out, so I'm not sure it would make
sense to set up "leases" on unfilled holes in the code
base--for six-month chunks of time.

2. I think student projects benefit from not getting lots
of "Is it ready yet?  Can I help?" mail.

Anyway, it looks like both the educational mission at CMU and
the driver-development mission of the community have been
served.  I'm open to suggestions about how to possibly improve
coordination... but meanwhile it's hard to argue with progress!

Dave Eckhardt