Re: 5.2-RELEASE TODO

From: Munish Chopra <chopra_at_soulwax.net>
Date: Mon, 15 Dec 2003 15:25:43 -0500
On 2003-12-15 14:09 +0000, Mathew Kanner wrote:
> On Dec 15, Munish Chopra wrote:
> > 
> > I'm not personally seeing any LOR's, but audio playback slows down under
> > load, and every few seconds I hear popping noises during audio playback,
> > whether under load or not.
> > 
> > I've got a simple SB16:
> > pcm0: <Creative CT5880-C> port 0xd800-0xd83f irq 5 at device 15.0 on
> > pci0
> > pcm0: <TriTech TR28602 AC97 Codec>
> 
> 	Hello Munish,
> 	I would realy like to help you and others.  Please describe in
> detail the circumstances that create the problem.  Are you
> experiencing buffer overruns or underruns?  How many devices are
> sharing the IRQ? What are they?  What happens when you cat a raw file
> to the device?  Does this happen with VCHANs?  Have you tried the patch
> I posted to -current under the heading "sound patch for pop &
> crackles"?
> 	Please read the pcm man page for a better description on how
> the answer the above.
> 

There are two distinct problems in my case:

1) Load can be virtually zero, moderate, or heavy, either way every few
   seconds (anywhere between 2 and 10 seconds or so) I hear snaps or
   pops in the output. Some are pretty hard to hear, others are obvious,
   depending on what's actually playing.

2) When untarring large gzip'ed or bzip2'ed tarballs, playback begins to
   "slow down" to something like half the normal speed or less. I first
   noticed this during portupgrade, when the XFree86 tarballs are
   unpacked.

Interestingly enough, even when I hammer the disk with several FTP
transfers and copies, I don't see the "slow down" effect. 

I'm seeing neither buffer overruns nor underruns. With hw.snd.verbose
set to 3, here's the output of `cat /dev/sndstat` towards the end of a
random MP3:

FreeBSD Audio Driver (newpcm)
Installed devices:
pcm0: <Creative CT5880-C> at io 0xd800 irq 5 (1p/1r/0v channels duplex
default)
        [pcm0:record:0]: spd 0, fmt 0x00000000/0x00000008, flags
	0x00000000, 0x00000000
        interrupts 0, overruns 0, hfree 4096, sfree 0
        {hardware} -> feeder_root(0x00000000) -> {userland}
        [pcm0:play:0]: spd 44100/44099, fmt 0x10000010, flags
	0x00005030, 0x00000000, pid 7063
        interrupts 21630, underruns 0, ready 129600
        {userland} -> feeder_root(0x10000010) -> {hardware}

File Versions:
$FreeBSD: src/sys/dev/sound/pci/es137x.c,v 1.48 2003/09/07 16:28:03 cg
Exp $
$FreeBSD: src/sys/dev/sound/isa/sndbuf_dma.c,v 1.2 2003/09/07 16:28:02
cg Exp $
$FreeBSD: src/sys/dev/sound/pcm/vchan.c,v 1.13 2003/09/07 16:28:03 cg
Exp $
$FreeBSD: src/sys/dev/sound/pcm/sound.c,v 1.86 2003/12/08 01:08:03
truckman Exp $
$FreeBSD: src/sys/dev/sound/pcm/sndstat.c,v 1.15 2003/12/08 01:08:03
truckman Exp $
$FreeBSD: src/sys/dev/sound/pcm/mixer.c,v 1.33 2003/11/11 05:38:28
scottl Exp $
$FreeBSD: src/sys/dev/sound/pcm/feeder_rate.c,v 1.10 2003/04/20 17:08:56
orion Exp $
$FreeBSD: src/sys/dev/sound/pcm/feeder_fmt.c,v 1.13 2003/09/07 16:28:03
cg Exp $
$FreeBSD: src/sys/dev/sound/pcm/feeder.c,v 1.32 2003/09/07 16:28:03 cg
Exp $
$FreeBSD: src/sys/dev/sound/pcm/fake.c,v 1.13 2003/09/07 16:28:03 cg Exp
$
$FreeBSD: src/sys/dev/sound/pcm/dsp.c,v 1.67 2003/11/11 05:38:28 scottl
Exp $
$FreeBSD: src/sys/dev/sound/pcm/channel.c,v 1.93 2003/12/05 02:08:13
matk Exp $
$FreeBSD: src/sys/dev/sound/pcm/buffer.c,v 1.21 2003/11/27 19:51:44 matk
Exp $
$FreeBSD: src/sys/dev/sound/pcm/ac97_patch.c,v 1.2 2003/08/21 15:44:55
orion Exp $
$FreeBSD: src/sys/dev/sound/pcm/ac97.c,v 1.49 2003/11/11 22:15:17
kuriyama Exp $

I can't see anything unusual above, but perhaps something's lurking. I
played this with no load on the box, and every few seconds there were
easily discernable popping noises.

There are no IRQ's being shared, though I hadn't noticed this before (it
looks harmless to me?): pcib0: slot 15 INTA is routed to irq 5 

I cat'ed a file to the sound device and received static -- is there
something specific I should be looking for? For future reference, what
clues were you hoping to gain from this?

I'm not using VCHANS, though the last time I did (some months ago) I had
the same problems.

I haven't tried your patch, I fear I was pretty busy when you posted it
and brushed past it assuming it had been committed. I'll hunt it down
now and see if it changes anything.

Thanks for taking interest in this, I appreciate it.

-- 
Munish Chopra
Received on Mon Dec 15 2003 - 11:25:42 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:34 UTC