Date: Wed, 29 Apr 1998 02:39:46 -0400 (EDT)
From: Derrick J Brashear <shadow@DEMENTIA.ORG>
To: sparclinux@vger.rutgers.edu
Subject: for your listening pleasure: audio-2.0-980429

Hi,

I just dropped two tarballs in
ftp://ftp.dementia.org/pub/linux/sparcaudio.
One is amp-0.7.6.tar.gz. It's a quickie port of amp to sparclinux, e.g.
SunOS audio interface, sparclinux include files. Because of a "bug" I
found in mpg123 0.59k for now I recommend amp until I can figure out 
why we're losing, or more importantly, why no one else is since the
midlevel audio code would appear to do the right thing! A binary and the
source are included.

The other is audio-2.0.33-980429.tar.gz. Source and modules are included
but because I have a few small hacks to my kernel it's conceivable that
these modules may not work on a stock 2.0.33 system. I have not tried. 

They *will not* work under 2.1. I can and intend to port the changes to
2.1, but currently I have no way to test any 2.1 kernels. No, I don't know
why audio is (ostensibly) broken in 2.1.95, either. Sorry.

What's here:
midlevel:
-changes to the play side of the midlevel to deal with the DMA that the
4231 does, so it's not reclaiming buffers that the chip is still trying to
play. 
-changes to make play interrupts be serviced immediately instead
of my placing a task on the queue. This makes 44.1 stereo reasonable
though not yet perfect.
-mixer ioctl handling based on patches from Michael Mraka (his were amd
specific; these should be generic though my stereo macros may not
attentuate volume correctly.)

4231:
-changes to deal with the concept of multiple buffers the hardware level
owns per above
-make driver act more like sunos, e.g. don't reset rate, encoding etc on
reopen

What's left to do:
midlevel:
-figure out how to use buffers larger than 4k. this almost certainly would
improve performance at 44.1 stereo. limiting factor may be the 4231 dma
-more linux sound ioctls. this is easy and is coming soon
-better getinfo/setinfo. i have some ideas
-capture changes corresponding to the play changes above

4231:
-make capture work

I'd appreciate it if someone could try this on an AMD-equipped machine and
let me know if it works. I'm having ongoing problems with the disk I have
for my IPC (a shame since otherwise the machine is ok).

Feedback welcome...

Thanks
-D




-- 
To unsubscribe: mail -s unsubscribe sparc-list-request@redhat.com < /dev/null