isapnp problems with opl3sa2

isapnp problems with opl3sa2

[Gerald Teschl]

Hi,

first of all, I don't know C and I am no expert, so please apologize if 
some of
my questions/comments are stupid.

I have a yamaha based sound card which works with the opl3sa2 driver. 
However,
the isapnp activation of this card fails (this has been broken ever 
since the 2.4 series).
So I did some investigations and eventually found out that the reason is 
the following
check from isapnp_check_dma(..) in isapnp.c:

-------------
    /* Some machines allow DMA 0, but others don't. In fact on some
       boxes DMA 0 is the memory refresh. Play safe */

    if (dma < 1 || dma == 4 || dma > 7)
        return 1;
--------------

However, the opl3sa2 card insists on dma=0,dma2=1. In particular, if I 
change
"(dma < 1 || dma == 4 || dma > 7)" to "(dma < 0 || dma == 4 || dma > 7)" 
everything
works fine.

Now I am wondering what should be done:

1) Make the above change to isapnp.c. According to the above comment 
this will break
some other boxes, so probably this should be configurable via /proc/isapnp.
2) Change opl3sa2.c and force isapnp to use dma=0 (a patch can be found 
at the end of this
message).

What do the experts think is the right way to fix this problem?

A few further remarks concerning my patch:
*) The patch also adds a line "opl3sa2_state[card].activated = 1" which 
is an
obvious omission in the original driver.
*) I have to force the driver to use both dma=0 and dma2=1. From what I
understand dma=0 should be sufficient, but this will not work. Looks like
a bug in isapnp.c to me. However, this should do no harm since according
to /proc/isapnp, the card only offers these values anyway.

Thanks for your comments,
Gerald


-----------------------------------------------
--- drivers/sound/opl3sa2.c.orig        Thu May  2 23:36:45 2002
+++ drivers/sound/opl3sa2.c     Sun Jun  2 15:19:44 2002
@@ -57,6 +57,7 @@
  *                         (Jan 7, 2001)
  * Zwane Mwaikambo        Added PM support. (Dec 4 2001)
  * Zwane Mwaikambo        Code, data structure cleanups. (Feb 15 2002)
+ * Gerald Teschl          Fixed ISA PnP activate. (Jun 02 2002)
  *
  */
 
@@ -868,8 +869,15 @@
                opl3sa2_state[card].activated = 1;
        }
        else {
+               /* isapnp.c disallows dma=0 but the opl3sa2 card only 
accepts this value */
+               if (!dev->ro) {
+                       isapnp_resource_change(&dev->dma_resource[0], 0, 1);
+                       isapnp_resource_change(&dev->dma_resource[1], 1, 1);
+               }
+               opl3sa2_state[card].activated = 1;
+
                if(dev->activate(dev) < 0) {
-                       printk(KERN_WARNING PFX "ISA PnP activate 
failed\n");
+                       printk(KERN_WARNING PFX "ISA PnP activate 
failed.\n");
                        opl3sa2_state[card].activated = 0;
                        return -ENODEV;
                }
------------------------------------------------

Re: isapnp problems with opl3sa2

[Zwane Mwaikambo]

On Mon, 3 Jun 2002, Gerald Teschl wrote:

> the isapnp activation of this card fails (this has been broken ever 
> since the 2.4 series).So I did some investigations and eventually found 
> out that the reason is the following check from isapnp_check_dma(..) in 
> isapnp.c:

You should be able to load it without any parameters with ISAPNP, i have a 
box at home with that card which only requires a modprobe opl3sa2. Could 
you show me how you're loading it as well as dmesg and /proc/isapnp output 
after loading.

> A few further remarks concerning my patch:
> *) The patch also adds a line "opl3sa2_state[card].activated = 1" which 
> is an obvious omission in the original driver.

Thanks, put it right at the end before the return and i'll be happy.

> *) I have to force the driver to use both dma=0 and dma2=1. From what I
> understand dma=0 should be sufficient, but this will not work. Looks like
> a bug in isapnp.c to me. However, this should do no harm since according
> to /proc/isapnp, the card only offers these values anyway.

isapnp.c is fine

> -----------------------------------------------
> --- drivers/sound/opl3sa2.c.orig        Thu May  2 23:36:45 2002
> +++ drivers/sound/opl3sa2.c     Sun Jun  2 15:19:44 2002

When you rediff, can you diff so that the patch can be applied with a 
strip of '1' ie diff -u linux/drivers...

Cheers,
	Zwane Mwaikambo

-- 
http://function.linuxpower.ca
		

Re: isapnp problems with opl3sa2

[Gerald Teschl]

Zwane Mwaikambo wrote:

>You should be able to load it without any parameters with ISAPNP, i have a 
>box at home with that card which only requires a modprobe opl3sa2. Could 
>you show me how you're loading it as well as dmesg and /proc/isapnp output 
>after loading.
>
I use redhat-7.3 (kernel-2.4.18-4) and this is an ASUS L7300 laptop.
In /etc/modules.conf I have

alias sound-slot-0 opl3sa2
options sound dmabuf=1
alias synth0 opl3
options opl3 io=0x388
options opl3sa2 ymode=2 isapnp=1

and /proc/isapnp looks like
-------------------------------------------------
Card 1 'YMH0800:YAMAHA OPL3-SAx Audio System' PnP version 1.0
  Logical device 0 'YMH0021:Unknown'
    Device is not active
    Active DMA 0,0
    Resources 0
      Priority preferred
      Port 0x220-0x220, align 0xf, size 0x10, 16-bit address decoding
      Port 0x530-0x530, align 0x7, size 0x8, 16-bit address decoding
      Port 0x388-0x388, align 0x7, size 0x8, 16-bit address decoding
      Port 0x330-0x330, align 0x1, size 0x2, 16-bit address decoding
      Port 0x370-0x370, align 0x1, size 0x2, 16-bit address decoding
      IRQ 5 High-Edge
      DMA 0 8-bit byte-count type-A
      DMA 1 8-bit byte-count type-A
      Alternate resources 0:1
      ...........
---------------------------------------------
before inserting the module. If I insert the (original) module using 
"modporbe opl3sa2" dmesg read:

ad1848/cs4248 codec driver Copyright (C) by Hannu Savolainen 1993-1996
ad1848: No ISAPnP cards found, trying standard ones...
opl3sa2: ISA PnP activate failed
opl3sa2: No PnP cards found
opl3sa2: 0 PnP card(s) found.

and /proc/isapnp looks as before. My investigations showed that the 
error occures in
isapnp_valid_dma() invoked from isapnp_config_activate() in isapnp.c.

If I insert my patched module I get

opl3sa2: Activated ISA PnP card 0 (active=1)
opl3sa2: chipset version = 0x4
opl3sa2: Found OPL3-SA3 (YMF715E or YMF719E)
ad1848: Interrupt test failed (IRQ5)
opl3sa2: 1 PnP card(s) found.

and /proc/isapnp looks like
-------------------------------------------
Card 1 'YMH0800:YAMAHA OPL3-SAx Audio System' PnP version 1.0
  Logical device 0 'YMH0021:Unknown'
    Device is active
    Active port 0x220,0x530,0x388,0x330,0x370
    Active IRQ 5 [0x2]
    Active DMA 0,1
    Resources 0
    .................
-------------------------------------------

>
>
>>A few further remarks concerning my patch:
>>*) The patch also adds a line "opl3sa2_state[card].activated = 1" which 
>>is an obvious omission in the original driver.
>>
>
>Thanks, put it right at the end before the return and i'll be happy.
>
Yes, you are right, probably more readable.

>
>
>>*) I have to force the driver to use both dma=0 and dma2=1. From what I
>>understand dma=0 should be sufficient, but this will not work. Looks like
>>a bug in isapnp.c to me. However, this should do no harm since according
>>to /proc/isapnp, the card only offers these values anyway.
>>
>
>isapnp.c is fine
>
I still think this might affect other hardware as well. Does your card 
use the same
irq/prot/dma settings? Does your /proc/isapnp look the same as mine? I don't
quite understand why it works for you. 

>When you rediff, can you diff so that the patch can be applied with a 
>strip of '1' ie diff -u linux/drivers...
>
No problem.

Thanks for your help!

Gerald