Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Rene Herman]

On 04/12/2007 09:55 PM, Rask Ingemann Lambertsen wrote:

>  - Add code to detect and PnP protocol to manage Media Vision Jazz16 sound
>    cards.
> 
> Tested with a Jazz16 based sound card and Linux 2.6.20.6.

Just got around to giving this a spin again and unfortunately the preffered 
options code isn't working for me:

rene@5bt1:~$ cat /sys/bus/pnp/devices/02\:01/options
Dependent: 01 - Priority preferred
    port 0x220-0x220, align 0x0, size 0x10, 10-bit address decoding
    port 0x330-0x330, align 0x0, size 0x2, 10-bit address decoding
    port 0x388-0x388, align 0x0, size 0x4, 10-bit address decoding
    irq 5 High-Edge
    irq 2/9 High-Edge
    dma 1 8-bit compatible
Dependent: 02 - Priority acceptable
    port 0x210-0x260, align 0xf, size 0x10, 10-bit address decoding
    port 0x300-0x330, align 0xf, size 0x2, 10-bit address decoding
    port 0x388-0x388, align 0x0, size 0x4, 10-bit address decoding
    irq 3,5,7,2/9,10,15 High-Edge
    irq 3,5,7,2/9,10,15 High-Edge
    dma 1,3,5,7 8-bit&16-bit compatible
Dependent: 03 - Priority acceptable
    port 0x210-0x260, align 0xf, size 0x10, 10-bit address decoding
    port 0x300-0x330, align 0xf, size 0x2, 10-bit address decoding
    port 0x388-0x388, align 0x0, size 0x4, 10-bit address decoding
    irq 3,5,7,2/9,10,15 High-Edge
    dma 1,3,5,7 8-bit&16-bit compatible
rene@5bt1:~$ cat /sys/bus/pnp/devices/02\:01/resources
state = disabled
rene@5bt1:~$ su -c "modprobe snd-jazz16"
rene@5bt1:~$ cat /sys/bus/pnp/devices/02\:01/resources
state = active
io 0x210-0x21f
io 0x300-0x301
io 0x388-0x38b
irq 5
irq 7
dma 1

Ports 210 and 300 are first from the acceptables, not the preferreds and the 
irq 7 is also not preffered for MPU. Yes, ports 220, 330 and irq 9 are free.

Looking back through what I sent earlier, I see I sent you a wrong patch 
then so I'm not sure anymore what I tested and I don't have time to look at 
it now. Thought I'd let you know though.

Rene.

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Rask Ingemann Lambertsen]

On Fri, Apr 20, 2007 at 07:27:25PM +0200, Rene Herman wrote:
> 
> Just got around to giving this a spin again and unfortunately the preffered 
> options code isn't working for me:
> 
> rene@5bt1:~$ cat /sys/bus/pnp/devices/02\:01/options
> Dependent: 01 - Priority preferred
>    port 0x220-0x220, align 0x0, size 0x10, 10-bit address decoding
>    port 0x330-0x330, align 0x0, size 0x2, 10-bit address decoding
>    port 0x388-0x388, align 0x0, size 0x4, 10-bit address decoding
>    irq 5 High-Edge
>    irq 2/9 High-Edge
>    dma 1 8-bit compatible

   Do you have a PCI device like this one? /sbin/lspci -v:

00:07.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 01)
        Flags: medium devsel, IRQ 9

   I don't know who to blame, but I know of no way of moving that device off
of IRQ 9. It may also just be a bug in the PCI code, because lspci -v -b
gives:

00:07.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 01)
        Flags: medium devsel

   I.e. no IRQ as seen from the bus.

-- 
Rask Ingemann Lambertsen

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Rene Herman]

[-- Attachment #1: Type: text/plain, Size: 1729 bytes --]

On 04/20/2007 10:00 PM, Rask Ingemann Lambertsen wrote:

> On Fri, Apr 20, 2007 at 07:27:25PM +0200, Rene Herman wrote:

>> Just got around to giving this a spin again and unfortunately the preffered 
>> options code isn't working for me:

>    Do you have a PCI device like this one? /sbin/lspci -v:
> 
> 00:07.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 01)
>         Flags: medium devsel, IRQ 9

Mmm, yes, I do. Was on an Intel 430VX chipset with a PIIX3 last time and on 
a 430TX with the PIIX4 this time.

> I don't know who to blame, but I know of no way of moving that device off
> of IRQ 9.

Indeed; it seems to be a well-known problem.

The Intel 430TX and 440BX (which also uses the PIIX4 if I'm not mistaken) 
were hugely popular chipsets in the later generation of boards that still 
had ISA slots and given that noone will deperately want the MPU-401 IRQ 
assigned; how would you feel about the attached? Just adds the preffered 
options minus the MPU IRQ as the first acceptable. Works for me...

I ran the driver through a complete test again. Playback, capture (from 
line, CD and mic), the OPL3 synth and MPU-401 out- and input (both with and 
without the MPU-401 IRQ) and all work fine. Feel free to add my signed-off 
to the ALSA part.

As to the PnP part; other than that IRQ9 thing everything also working fine 
and if you think the attached is ugly; I don't much care. I still have the 
possible objection with the BIOS maybe needing to know about the card anyway 
so it can route the IRQ to ISA (and the slight discomfort about Jazz16 code 
in drivers/pnp). But next iteration can be final.

If Adam still isn't commenting by then, let's just feed it through Andrew 
Morton and LKML.

Rene.


[-- Attachment #2: jazz16pnp-irq9.diff --]
[-- Type: text/plain, Size: 840 bytes --]

commit ff9c54ec792041f6f7f45d5abceb5af47a71a2f7
Author: Rene Herman <rene.herman@gmail.com>
Date:   Sat Apr 21 09:07:44 2007 +0200

    IRQ 9 is unavailable on PIIX4 boards -- the MPU-401 UART can
    do without an IRQ as well...

diff --git a/drivers/pnp/jazz16pnp.c b/drivers/pnp/jazz16pnp.c
index b71381b..020b6bb 100644
--- a/drivers/pnp/jazz16pnp.c
+++ b/drivers/pnp/jazz16pnp.c
@@ -233,6 +233,11 @@ static int __init jazz16pnp_add_sb(void)
 				   jazz16pnp_pref_dmas, 1))
 		goto out_no_sboption;
 	if (!jazz16pnp_easy_option(sbdev, PNP_RES_PRIORITY_ACCEPTABLE,
+				   jazz16pnp_pref_ports, 3,
+				   jazz16pnp_pref_irqs, 1,
+				   jazz16pnp_pref_dmas, 1))
+		goto out_no_sboption;
+	if (!jazz16pnp_easy_option(sbdev, PNP_RES_PRIORITY_ACCEPTABLE,
 				   jazz16pnp_ports, 3,
 				   jazz16pnp_irqs, 2,
 				   jazz16pnp_dmas, 1))

[-- Attachment #3: Type: text/plain, Size: 160 bytes --]

_______________________________________________
Alsa-devel mailing list
Alsa-devel@alsa-project.org
http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Rask Ingemann Lambertsen]

On Sun, Apr 22, 2007 at 06:27:44PM +0200, Rene Herman wrote:
> 
> The Intel 430TX and 440BX (which also uses the PIIX4 if I'm not mistaken) 
> were hugely popular chipsets in the later generation of boards that still 
> had ISA slots and given that noone will deperately want the MPU-401 IRQ 
> assigned; how would you feel about the attached? Just adds the preffered 
> options minus the MPU IRQ as the first acceptable. Works for me...

   Fine with me. Maybe the MPU IRQ should be omitted alltogether? You tell
me.

   Thanks for the testing you've done.

> As to the PnP part; other than that IRQ9 thing everything also working fine 

   Aren't you having problems loading the module after a resource allocation
failure? I find that I need this additional patch to be able to retry:

--- linux-2.6.20.6-clean/drivers/pnp/driver.c	2007-02-04 19:44:54.000000000 +0100
+++ linux-2.6.20.6-ril/drivers/pnp/driver.c	2007-04-15 10:10:35.000000000 +0200
@@ -96,13 +96,13 @@
 		if (!(pnp_drv->flags & PNP_DRIVER_RES_DO_NOT_CHANGE)) {
 			error = pnp_activate_dev(pnp_dev);
 			if (error < 0)
-				return error;
+				goto fail;
 		}
 	} else if ((pnp_drv->flags & PNP_DRIVER_RES_DISABLE)
 		    == PNP_DRIVER_RES_DISABLE) {
 		error = pnp_disable_dev(pnp_dev);
 		if (error < 0)
-			return error;
+			goto fail;
 	}
 	error = 0;
 	if (pnp_drv->probe) {


   Without it, the device stays attached to the driver. What happens is

	pnp_attach_device (device, driver);
	error = pnp_activate_dev(device);
	if (error < 0)
		return error;
	...
fail:
	pnp_detach_device (device, driver);
	return error;

-- 
Rask Ingemann Lambertsen

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Rene Herman]

On 04/23/2007 08:52 PM, Rask Ingemann Lambertsen wrote:

> On Sun, Apr 22, 2007 at 06:27:44PM +0200, Rene Herman wrote:

>> The Intel 430TX and 440BX (which also uses the PIIX4 if I'm not
>> mistaken) were hugely popular chipsets in the later generation of
>> boards that still had ISA slots and given that noone will deperately
>> want the MPU-401 IRQ assigned; how would you feel about the attached?
>> Just adds the preffered options minus the MPU IRQ as the first
>> acceptable. Works for me...
> 
> Fine with me. Maybe the MPU IRQ should be omitted alltogether? You tell 
> me.

"Using the hardware to its fullest potential" should probably be called the 
preferred option. I mean, what's the fun of playing with this old hardware 
if we're not going to use it fully? :)

So as submitted (or nothing at all but if we suppose the driver will be 
seeing practical use, guarding against that PIIX4 seems best) is nicest I feel.

> Aren't you having problems loading the module after a resource allocation
> failure? I find that I need this additional patch to be able to retry:

I've had to remove the jazz16 again for a bit and I haven't tested an 
allocation failure yet. That looks to be an "obvious fix" though. Now all we 
need is a PnP maintainer...

Rene.

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Takashi Iwai]

At Thu, 26 Apr 2007 14:55:03 +0200,
Rene Herman wrote:
> 
> > Aren't you having problems loading the module after a resource allocation
> > failure? I find that I need this additional patch to be able to retry:
> 
> I've had to remove the jazz16 again for a bit and I haven't tested an 
> allocation failure yet. That looks to be an "obvious fix" though. Now all we 
> need is a PnP maintainer...

I'd suggest to post LKML once, then.  If you get no negative feedback,
let's try to push together with ALSA patch.


Takashi

Re: [PATCH] PnP: Protocol for Media Vision Jazz16 sound cards

[Rene Herman]

On 04/26/2007 03:00 PM, Takashi Iwai wrote:

>> I've had to remove the jazz16 again for a bit and I haven't tested an 
>> allocation failure yet. That looks to be an "obvious fix" though. Now all we 
>> need is a PnP maintainer...
> 
> I'd suggest to post LKML once, then.  If you get no negative feedback,
> let's try to push together with ALSA patch.

I have one other email address for Adam Belay (added to CC). Let's see if 
that one's active...

Rene.