2.6.16-rt11 and low-latency audio xruns (interrupt latency problem?)

2.6.16-rt11 and low-latency audio xruns (interrupt latency problem?)

[Valin, Jean-Marc (ICT Centre, Marsfield)]

Hi,

I'm getting xruns while trying to do low-latency capture/playback of
audio. After enabling the alsa debug options, I'm seeing errors like
these in my logs:

[66205.003000] Unexpected hw_pointer value [1] (stream = 0, delta: -3,
max jitter = 16): wrong interrupt acknowledge?
[66205.003000]  [<f0ae2ec1>] snd_pcm_period_elapsed+0x136/0x27a
[snd_pcm] (8)
[66205.003000]  [<f0af7780>] snd_intel8x0_interrupt+0x18d/0x1e5
[snd_intel8x0] (20)
[66205.003000]  [<b013495b>] handle_IRQ_event+0x4e/0xb6 (44)
[66205.003000]  [<b0135443>] thread_edge_irq+0x6f/0xc4 (48)
[66205.003000]  [<b0135519>] do_irqd+0x0/0x8a (16)
[66205.003000]  [<b01354d8>] do_hardirq+0x40/0x81 (4)
[66205.003000]  [<b0135575>] do_irqd+0x5c/0x8a (16)
[66205.003000]  [<b01261f6>] kthread+0x79/0xa3 (16)
[66205.003000]  [<b012617d>] kthread+0x0/0xa3 (24)
[66205.003000]  [<b0101319>] kernel_thread_helper+0x5/0xb (16)
[66205.004000] XRUN: pcmC0D0p
[66205.004000]  [<f0ae2f72>] snd_pcm_period_elapsed+0x1e7/0x27a
[snd_pcm] (8)
[66205.004000]  [<f0af7780>] snd_intel8x0_interrupt+0x18d/0x1e5
[snd_intel8x0] (24)
[66205.004000]  [<b013495b>] handle_IRQ_event+0x4e/0xb6 (44)
[66205.004000]  [<b0135443>] thread_edge_irq+0x6f/0xc4 (48)
[66205.004000]  [<b0135519>] do_irqd+0x0/0x8a (16)
[66205.004000]  [<b01354d8>] do_hardirq+0x40/0x81 (4)
[66205.004000]  [<b0135575>] do_irqd+0x5c/0x8a (16)
[66205.004000]  [<b01261f6>] kthread+0x79/0xa3 (16)
[66205.004000]  [<b012617d>] kthread+0x0/0xa3 (24)
[66205.004000]  [<b0101319>] kernel_thread_helper+0x5/0xb (16)

The only info I could get after searching for a while is that the
"Unexpected hw_pointer value" *could* be related to the audio interrupt
being serviced too late
(http://www.ussg.iu.edu/hypermail/linux/kernel/0407.1/1613.html).
However, I'm running the audio IRQ (IRQ 16) at priority 99 and my app at
98 (I also tried the opposite and get the same result). In this case,
I'm running with two periods of 16 samples each (48 kHz), but I've seen
that error occur with periods up to 64 samples. 

Any idea how to fix/further investigate that? My machine is a Dell D600
laptop with a Pentium-M 2.13 GHz, which I'm running at full-speed
(performance governor). Kernel is stock 2.6.16 + rt11 patch. Distro is
Ubuntu 5.10, kernel is not tainted.

Thanks,

        Jean-Marc

P.S. Please include me in CC since I'm not subscribed to lkml.

Re: 2.6.16-rt11 and low-latency audio xruns (interrupt latency problem?)

[Lee Revell]

On Thu, 2006-03-30 at 16:13 +1100, Valin, Jean-Marc (ICT Centre,
Marsfield) wrote:
> In this case,
> I'm running with two periods of 16 samples each (48 kHz), but I've
> seen
> that error occur with periods up to 64 samples.  

It's possible that your hardware simply isn't capable of this and you'll
have to get a multichannel soundcard (more channels means you can use a
smaller buffer before hitting the lower limit of a PCI transfer) or use
a period of 128 samples or more.

Lee

Re: 2.6.16-rt11 and low-latency audio xruns (interrupt latency problem?)

[Valin, Jean-Marc (ICT Centre, Marsfield)]

> It's possible that your hardware simply isn't capable of this and you'll
> have to get a multichannel soundcard (more channels means you can use a
> smaller buffer before hitting the lower limit of a PCI transfer) or use
> a period of 128 samples or more.

>From the ALSA verbose mode, I can see that the hardware does accept as
low as 8 samples per period -- and it actually works outside of the few
xruns per minute. 

	Jean-Marc

Re: 2.6.16-rt11 and low-latency audio xruns (interrupt latency problem?)

[Valin, Jean-Marc (ICT Centre, Marsfield)]

On Thu, 2006-03-30 at 17:19 +1100, Valin, Jean-Marc (ICT Centre,
Marsfield) wrote:
> > It's possible that your hardware simply isn't capable of this and you'll
> > have to get a multichannel soundcard (more channels means you can use a
> > smaller buffer before hitting the lower limit of a PCI transfer) or use
> > a period of 128 samples or more.

Could you give me more info on that PCI transfer limit? I couldn't find
any reference to that on the web. Also, any way to see if that's the
real cause of the problem. I've been doing some experiments and noticed
that the percentage or xruns I get is almost independent of the period
size, which is a little odd. Any idea?

	Jean-Marc