From: Benny Sjostrand <gorm@cucumelo.org>
To: "alsa-devel@alsa-project.org" <alsa-devel@alsa-project.org>
Subject: Some cs4630 progress
Date: Sun, 21 Jul 2002 20:36:53 +0200 [thread overview]
Message-ID: <3D3AFF45.8060806@cucumelo.org> (raw)
In-Reply-To: Pine.LNX.4.33.0207211030310.755-100000@pnote.perex-int.cz
Hi!
In this moment i've been hacking on the cs46xx alsa driver and hopefully
i've got something that can make any fustraded cs46xx user/developer a
little bit happier.
Public documentation about the cs4630 dont seems to exist and as we
already know, and Cirrus have not been very helpful providing us such
necesary technical information. Finally Tom. W at Cirrus have provided
me some DSP assembler sources, and some other related C++ sources. With
that references i've able to find out some internals of the cs4630 and
get a little bit further.
The CS4630 runs a mini Real-Time OS inside, called SPOS (Stream
Processing OS I guess ...), parts of the SPOS is located in ROM memory
on the CS4630 and CS4612, on the CS4610 got no ROM, and the hole SPOS
needs be loaded in RAM.
Then it's up to user/driver to setup which tasks should be running
inside the SPOS, basically here we one basic entry to care of, the SCB
(stream controll block), then we have 4 task tree, Hyper Foreground,
Foreground, and Background. A SCB entry got a few basic things to care
of, a address pointer to the next SCB list, a sublist address pointer,
a function entry pointer. Then it has a lot of other relevant variables
which depends on which task uses the SCB, but no documentation about
this, so it's only to test && guess.
In the DSP image with the current ALSA driver all tasks are statically
setup when load the image.
When the drivers pokes and peeks on the BA1 address space what it
actually does is to modify and read parameters from a SCB task running
in the DSP.
So I've been working on using the .osp files provided by the Windows
driver. To .osp contains a EXPORT segment where from the driver obtain
the task entry when setting up a SCB, (or a task tree).
The .osp provided by Windows driver dont contain any parameter data,
which means that all tasks and SCB needs to be setup by the driver. When
a the CODE in a .osp file is loaded into the DSP on other addres then
0x0000 all "call()", and "goto xxxx" instructions needs to be fixed up.
In this moment I've wroten a program named "ospparser" then generates
.h from a .osp which contains a table with all exported symbol, code,
sample and parameter segments.
Indirectly i can successfully load cwc4630.osp, cwcasync.osp,
cwcsnoop.osp code into DSP, do all necesary fixups, and have everything
working just a before, (as with current ALSA driver) All SCB's and
tasktrees are setup by the driver.
Today i manage to get "signals" out from the SPDIFO interface, however
the sound is very distorcionated, totally broken, and my minidisc dont
wont record it, course some copyright
bit in the stream that i've got no idea how i been able to configure it
in such way.
To make multiple PCM channel, secondary CODEC to work is just to setup
the necesary SCB's entries, which is hard to know how to do without any
documenation.
Well, if you want take a look at my changes to the cs46xx driver,
however the difference is not "hearable" yet, there a lot of work done,
download my code at this locations:
http://www.cucumelo.org/~gorm/alsa-driver-0.9.0rc2-bs20020721.tar.gz
http://www.cucumelo.org/~gorm/ospparser.tar.gz
/Benny
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
prev parent reply other threads:[~2002-07-21 19:28 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-07-20 9:55 snd_pcm_poll_descriptors_revents() FAU
2002-07-21 8:34 ` snd_pcm_poll_descriptors_revents() Jaroslav Kysela
2002-07-21 18:36 ` Benny Sjostrand [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3D3AFF45.8060806@cucumelo.org \
--to=gorm@cucumelo.org \
--cc=alsa-devel@alsa-project.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.