Third Try: [PATCH] cs4231 alsa driver: Fix IRQ and initialisation

Third Try: [PATCH] cs4231 alsa driver: Fix IRQ and initialisation

[Chini, Georg]

[-- Attachment #1: Type: text/plain, Size: 244 bytes --]

Hi,

I don't know what my mail program does to the patch. When I
send it there is always a + sign too much in one line. So I attach
it as a file now. I hope this works, and once again sorry for the
confusion.

Regards
              Georg Chini

[-- Attachment #2: cs4231_061229.diff --]
[-- Type: text/plain, Size: 2412 bytes --]

From: Georg Chini georg.chini@triaton-webhosting.com

SBUS: Change IRQ-handler return value from 0 to IRQ_HANDLED and
fix some initialisation problems.
Change period_bytes_min from 4096 to 256 to allow driver to work with
low latency (VOIP) applications. Hope this does not break EBUS.

Signed-off-by: Georg Chini <georg.chini@triaton-webhosting.com>

diff -uprN linux.orig/sound/sparc/cs4231.c linux/sound/sparc/cs4231.c
--- linux.orig/sound/sparc/cs4231.c	2006-12-29 22:08:10.000000000 +0100
+++ linux/sound/sparc/cs4231.c	2006-12-29 21:45:42.000000000 +0100
@@ -1268,7 +1268,7 @@ static struct snd_pcm_hardware snd_cs423
 	.channels_min		= 1,
 	.channels_max		= 2,
 	.buffer_bytes_max	= (32*1024),
-	.period_bytes_min	= 4096,
+	.period_bytes_min	= 256,
 	.period_bytes_max	= (32*1024),
 	.periods_min		= 1,
 	.periods_max		= 1024,
@@ -1288,7 +1288,7 @@ static struct snd_pcm_hardware snd_cs423
 	.channels_min		= 1,
 	.channels_max		= 2,
 	.buffer_bytes_max	= (32*1024),
-	.period_bytes_min	= 4096,
+	.period_bytes_min	= 256,
 	.period_bytes_max	= (32*1024),
 	.periods_min		= 1,
 	.periods_max		= 1024,
@@ -1796,7 +1796,7 @@ static irqreturn_t snd_cs4231_sbus_inter
 	snd_cs4231_outm(chip, CS4231_IRQ_STATUS, ~CS4231_ALL_IRQS | ~status, 0);
 	spin_unlock_irqrestore(&chip->lock, flags);
 
-	return 0;
+	return IRQ_HANDLED;
 }
 
 /*
@@ -1821,7 +1821,6 @@ static int sbus_dma_request(struct cs423
 	if (!(csr & test))
 		goto out;
 	err = -EBUSY;
-	csr = sbus_readl(base->regs + APCCSR);
 	test = APC_XINT_CNVA;
 	if ( base->dir == APC_PLAY )
 		test = APC_XINT_PNVA;
@@ -1862,17 +1861,16 @@ static void sbus_dma_enable(struct cs423
 
 	spin_lock_irqsave(&base->lock, flags);
 	if (!on) {
-		if (base->dir == APC_PLAY) { 
-			sbus_writel(0, base->regs + base->dir + APCNVA); 
-			sbus_writel(1, base->regs + base->dir + APCC); 
-		}
-		else
-		{
-			sbus_writel(0, base->regs + base->dir + APCNC); 
-			sbus_writel(0, base->regs + base->dir + APCVA); 
-		} 
+		sbus_writel(0, base->regs + base->dir + APCNC); 
+		sbus_writel(0, base->regs + base->dir + APCNVA); 
+		sbus_writel(0, base->regs + base->dir + APCC); 
+		sbus_writel(0, base->regs + base->dir + APCVA); 
+		
+		/* ACK any APC interrupts. */
+		csr = sbus_readl(base->regs + APCCSR);
+		sbus_writel(csr, base->regs + APCCSR);
 	} 
-	udelay(600); 
+	udelay(1000); 
 	csr = sbus_readl(base->regs + APCCSR);
 	shift = 0;
 	if ( base->dir == APC_PLAY )

Re: Third Try: [PATCH] cs4231 alsa driver: Fix IRQ and

[David Miller]

From: "Chini, Georg" <georg.chini@hp.com>
Date: Fri, 29 Dec 2006 23:16:02 +0100

> I don't know what my mail program does to the patch. When I
> send it there is always a + sign too much in one line. So I attach
> it as a file now. I hope this works, and once again sorry for the
> confusion.

Thanks for your fix, I've applied your patch.

BTW, the period_bytes_min settings I simply copied over from the
existing ISA cs4231 driver at the time.

But it seems that nowadays the ISA CS4231 driver is using a value of
"64" for period_bytes_min.  Perhaps the Sparc driver should do that
too :-)

Re: Third Try: [PATCH] cs4231 alsa driver: Fix IRQ and initialisation

[Georg Chini]

David Miller wrote:

>But it seems that nowadays the ISA CS4231 driver is using a value of
>"64" for period_bytes_min.  Perhaps the Sparc driver should do that
>too :-)
>
>
>  
>
My machine seems too slow to handle this. With values less
than 256 sound will not work properly. I think it starts missing
interrupts.

Regards
              Georg

Re: Third Try: [PATCH] cs4231 alsa driver: Fix IRQ and initialisation

[Chini, Georg]

Georg Chini wrote:

>
> David Miller wrote:
>
>> But it seems that nowadays the ISA CS4231 driver is using a value of
>> "64" for period_bytes_min.  Perhaps the Sparc driver should do that
>> too :-)
>>
>>
>>  
>>
> My machine seems too slow to handle this. With values less
> than 256 sound will not work properly. I think it starts missing
> interrupts.
>
>
Wrong again, I had some other problem when I tried 64. It works
fine, so please change the period_bytes_min to 64.

Regards
              Georg

Re: Third Try: [PATCH] cs4231 alsa driver: Fix IRQ and

[David Miller]

From: "Chini, Georg" <georg.chini@hp.com>
Date: Wed, 03 Jan 2007 19:32:57 +0100

> Georg Chini wrote:
> 
> >
> > David Miller wrote:
> >
> >> But it seems that nowadays the ISA CS4231 driver is using a value of
> >> "64" for period_bytes_min.  Perhaps the Sparc driver should do that
> >> too :-)
> >>
> >>
> >>  
> >>
> > My machine seems too slow to handle this. With values less
> > than 256 sound will not work properly. I think it starts missing
> > interrupts.
> >
> >
> Wrong again, I had some other problem when I tried 64. It works
> fine, so please change the period_bytes_min to 64.

Great, thanks for testing.  I'll make the change.