Hi Takashi, Massimo, anyone

Takashi, in response to this last e-mail I patially solved the problem. Just to be in 
the safe side, I:
1. cleaned the rc2 installation.
2. downloaded the cvs version again (10-08-2002)
3. did a �./cvscompile�
4. did a �make install�

It all went fine. I then proceeded to insert �snd_printk� statements in the code of 
the drive to see what was happening. The process is slow and somewhat stressfull 8-[ 
(Cyrix 200+ with 64Mega). I would  like to know:
1. How may I compile/make and install just the modules that I am interested in 
(--with-card ?)
2. Do  have to ./cvscompile all the time? I did try just using �make� after the 1st 
config, but this did  not seem to work.
3. Do you have any pointers on debugging ALSA diver module? I am sorry but my 
experioence in this platform is somewhat lacking (make. autoconfig, etc.).
 
Massimo, 

I hope you are reading the list. I did as you indicated:
1. I compiled and installed the drivers, libs, tools and utils (from cvs)
2. I  did not create any /etc/isapnp.conf file, nor did I execute any �pnpdump� or 
�isapnp� commands
3. I then used a working /etc/modules.conf that used the OSS driver (mad 16-pro)
4. Finaly I executed (repeatedly) a modprobe.

Small note: the O.S�s messages (kernel 2.4.8) indicate that the isa card was detected.

The error I get is always a �chip still busy�, which is written by the 
snd_opti93x_busy_wait(..). With the aid of the �snd_printk� I see that all parameters 
are passed correctly and that all the steps are executed until the call to 
snd_opti93x_probe() in the snd_opti93x_create() routine.  On furher investigation I 
found that snd_opti93x_busy_wait(..)'s �inb� call always seems to return 0xFF from the 
(WssBase + 4) address. So the card is considered not initialized. Naturally the test 
for �&& 0x80" always fails.

Ok, I have some questions you most probably can answer:

1. Because I am compiling via cvs, the isa-pnp is set to true, and because the 
parameters are set and isa_pnp option is set to 1 by default, then no pnp scan is made 
for available resources. Is this the way you set up and tested the cards? I would like 
to replicate you test-bed.

2. Once probing has been sucessfully done, I take it that only then you configure the 
card, or am I wrong? Do we need to configure the card so that it sucessfully 
terminates initialization?

3. If I read correctly,  on dma channel is 16 and the other an 8 bit channel. In this 
case, is dma1 the 8 bit channel and dma2 the 16 bit channel? I would like to set this 
as per OSS.

4. On one of your answers to the list you stated that you had one card working and 
another not. I would like to confirm that the card that is not working is the MED931 
ver 2.0? 

5. How should the /proc/isapnp show up? Does it only have to have the correct settings 
after the driver is load?

6. When I used the OSS drivers, dma1=dma2=0. When I try this out with the driver, this 
is not allowed. Furthermore, you do not allow dma2=2 or 3 also. Is this correct?

Ok, this is a long e-mail, but pleae bear with me. After a trying weekend, I finally 
decided to look at the mad16.c source code. Low and behold, the detection code (which 
is executed when probing the card), makes provision for:
1.  testing for the 0x80 status (as per alsa driver)
2.  if this is not sucessfull, tries a software reset. [line 238]
3. if this does not work, powers the card off and on again. [line 245]
4. finally, if nothing else works, the pnp mode is turned off. [line 257]. (Potential 
PnP Bios problem is indicated in this case.)

Ok, it would be interresting to know what this driver actually does on my box. To be 
honest, I don�t really want to check this because it would mean configuring the kernel 
and activating the OSS drivers again. For anyone out there that thinks I am being 
lazy... it takes nearly a day to compile and install a new kernel. 8-[

Massimo, what do you think of the above strategy? Worth a try?

To anyone/everyone else,

I would like to collect some statistics. I would like to know:
1. Who has a Opti based ISA card working with Alsa (MED 931 from ExpertColor)
2. I would like to know what chip is used (92c931)
3. I would like to know what revision  of the PCB/card is being used (ex. version 2.0)
4. I would like to know what set-up or configuration is being done (isapnptools or 
Alsa�s ISA support)
5. What OS is being used (kernel version 2.4.x)
6. Is Pnp is is supported by BIOS
7. Is Pnp is supported by the OS

Hmm, I think this is it. If you have read until this point, thank you for attention. I 
hope our efforts will solve this small dilemma  because I know I am not the only one.

Regards,.
Hugo Ferreira.

> 
> De: Takashi Iwai <[EMAIL PROTECTED]>
> Data: 09/08/2002 11:29:52
> Para: [EMAIL PROTECTED]
> CC: <[EMAIL PROTECTED]>
> Assunto: Re: [Alsa-devel] Opti 931: Med931 ver.2.0
> 
> At Fri, 9 Aug 2002 10:03:09 +0100,
> [EMAIL PROTECTED] wrote:
> > 
> > Takashi,
> > 
> > >the patch was already in cvs.  but i'm not sure whether it was >already
> > >fixed on rc2.  please try the cvs version.
> > >you'll need to update alsa-lib, too, if you use the cvs drivers.
> > 
> > I downloaded the cvs version and attempted the ?vscompile?in the
> > driver, libs, tools and utils. Unfortunately I could not compile the 
> > lib because of an error. Seems to be occurring in the control/ 
> > control_shm.c file. I have attached the output for you to see.
>  
> hmm, not found.  did you forget to attach it actually?
> 
> anyway, to track down the bug, we need the precise information about
> your environment - which distribution, glibc, and kernel version,
> etc.
> 
> > I am sorry I took so long to answer but my disk was full and I had
> > to move some stuff around to be able to compile ok.
>  
> perhaps that was the reason -- disk full.
> 
> > I would like to confirm the following though:
> > 1. All I need to do is a ?vscompile? the modules will be 
> > automatically installed (no need to depmod ? for example?).
> 
> no.  cvscompile does only compile.
> after that, you need to install the drivers by running "make install"
> as root.
> 
> > 2. I need not remove the rc2 compilation output (no ?ake clean?
> > necessary?)
> 
> no.  the very latest makefile on alsa-driver will remove all old
> modules before installation.
> 
> 
> ciao,
> 
> Takashi
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Alsa-devel mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/alsa-devel
> 



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Alsa-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/alsa-devel

Reply via email to