ice1712 and the delta 1010

ice1712 and the delta 1010

[David Santamauro]

Hi all,

this is my first post as I was guided here from the linux audio user
list to follow up on a problem I've been having with my delta 1010.

Background:
Since upgrading my audio PC to 64-bit, the delta 1010 analog inputs
emit a constant 'hiss' -36 -30dB. Analog outputs and digital ins and
outs are fine. The delta 1010 from a hardware perspective is fine as I
tested it on the same pc hardware under windows7 64-bit.

So far, I have tested it in a 32-bit fedora, 32-bit windows vista and
64-bit windows7 all successfully. This problem arises only in 64-bit
fedora.

My kernel:
Linux 2.6.31.12-1.rt21.1.fc12.ccrma.x86_64.rt

My alsa packages and versions:
alsa-lib.x86_64                          1.0.23-1.fc12
alsa-lib-devel.x86_64                    1.0.23-1.fc12
alsa-tools.x86_64                        1.0.22-1.1.fc12.ccrma
alsa-utils.x86_64                        1.0.23-3.fc12

My modinfo:

$ modinfo snd_ice1712
filename:       /lib/modules/2.6.31.12-1.rt21.1.fc12.ccrma.x86_64.rt/kernel/sound/pci/ice1712/snd-ice1712.ko
license:        GPL
description:    ICEnsemble ICE1712 (Envy24)
author:         Jaroslav Kysela <perex@perex.cz>
srcversion:     0B183929429B4C5102B7D09
alias:          pci:v00001412d00001712sv*sd*bc*sc*i*
depends:
snd-pcm,snd,snd-i2c,snd-mpu401-uart,snd-ice17xx-ak4xxx,snd-cs8427,snd-ak4xxx-adda,snd-ac97-codec
vermagic:       2.6.31.12-1.rt21.1.fc12.ccrma.x86_64.rt SMP preempt
mod_unload parm:           index:Index value for ICE1712 soundcard.
(array of int) parm:           id:ID string for ICE1712 soundcard.
(array of charp) parm:           enable:Enable ICE1712 soundcard.
(array of bool) parm:           omni:Enable Midiman M-Audio Delta Omni
I/O support. (array of bool) parm:           cs8427_timeout:Define
reset timeout for cs8427 chip in msec resolution. (array of int)
parm:           model:Use the given board model. (array of charp)
parm:           dxr_enable:Enable DXR support for Terratec DMX6FIRE.
(array of int)


So, my question is simply: has this been reported before, in
particular regarding the 64-bit ice1712 module?

I'm pretty convinced it is a software/driver 32/64-bit issue based on my
observations above and confirmation of the same problem under the same
conditions by another user, but am willing to be corrected.

I'm a pretty competent C/C++ coder but a novice coder in the linux audio
world so I'd be willing to learn and help from a driver perspective
given the right direction -- at the least help debug the problem (if it
is even at the driver level).

Thanks for any input.

David

Re: ice1712 and the delta 1010

[Clemens Ladisch]

David Santamauro wrote:
> Since upgrading my audio PC to 64-bit,

Hardware or software upgrade?

> the delta 1010 analog inputs emit a constant 'hiss' -36 -30dB. Analog
> outputs and digital ins and outs are fine. The delta 1010 from
> a hardware perspective is fine as I tested it on the same pc hardware
> under windows7 64-bit.

In theory, this sounds like a driver bug.

> So far, I have tested it in a 32-bit fedora, 32-bit windows vista and
> 64-bit windows7 all successfully. This problem arises only in 64-bit
> fedora.

Does the 32-bit Linux use the same driver version?
If yes, then it's possible that the 32-bit and 64-bit kernels configure
the PC differently; for example, there might be noise on the power line
due to a different kind of power management.


Regards,
Clemens

Re: ice1712 and the delta 1010

[David Santamauro]

Hi,

On Thu, 26 Aug 2010 15:02:04 +0200
Clemens Ladisch <clemens@ladisch.de> wrote:

> David Santamauro wrote:
> > Since upgrading my audio PC to 64-bit,
> 
> Hardware or software upgrade?

Both ... I should say the 32-bit linux test was done on different
hardware. The events were:

a) pc1, fedora 11 32-bit both maudio delta 1010 and 2496 coexist happily
b) pc2 new build to be a 64-bit machine
c) pulled delta 1010 out of pc1, inserted into pc2 -- noise
d) installed win7 64-bit in dual boot configuration on pc2, no noise
with delta 1010

> 
> > the delta 1010 analog inputs emit a constant 'hiss' -36 -30dB.
> > Analog outputs and digital ins and outs are fine. The delta 1010
> > from a hardware perspective is fine as I tested it on the same pc
> > hardware under windows7 64-bit.
> 
> In theory, this sounds like a driver bug.
> 
> > So far, I have tested it in a 32-bit fedora, 32-bit windows vista
> > and 64-bit windows7 all successfully. This problem arises only in
> > 64-bit fedora.
> 
> Does the 32-bit Linux use the same driver version?

To be honest, I'm not sure exactly how to check. They are the latest
versions of fedora 11 (32-bit) and fedora 12 (64-bit).
srcversion from a modinfo yields different results.

32-bit fedora 11
Linux beethoven 2.6.29.6-1.rt23.4.fc11.ccrma.i586.rt
srcversion:     F7E1C076BE18C5DFF723A6E


64-bit fedora 12
Linux debussy 2.6.31.12-1.rt21.1.fc12.ccrma.x86_64.rt
srcversion:     0B183929429B4C5102B7D09

> If yes, then it's possible that the 32-bit and 64-bit kernels
> configure the PC differently; for example, there might be noise on
> the power line due to a different kind of power management.

another interesting angle... I thought it might be power or some kind
of internal noise. How could I pursue that further?

David

Re: ice1712 and the delta 1010

[Clemens Ladisch]

David Santamauro wrote:
> Clemens Ladisch <clemens@ladisch.de> wrote:
> > Does the 32-bit Linux use the same driver version?
> 
> To be honest, I'm not sure exactly how to check. They are the latest
> versions of fedora 11 (32-bit) and fedora 12 (64-bit).

> Linux beethoven 2.6.29.6-1.rt23.4.fc11.ccrma.i586.rt
> Linux debussy 2.6.31.12-1.rt21.1.fc12.ccrma.x86_64.rt

That's different.

> > If yes, then it's possible that the 32-bit and 64-bit kernels
> > configure the PC differently; for example, there might be noise on
> > the power line due to a different kind of power management.
> 
> another interesting angle... I thought it might be power or some kind
> of internal noise. How could I pursue that further?

Try running something on all CPU cores so that it doesn't switch into
power saving mode.


Regards,
Clemens

Re: ice1712 and the delta 1010

[David Santamauro]

On Thu, 26 Aug 2010 16:55:32 +0200
Clemens Ladisch <clemens@ladisch.de> wrote:

> > > If yes, then it's possible that the 32-bit and 64-bit kernels
> > > configure the PC differently; for example, there might be noise on
> > > the power line due to a different kind of power management.  
> > 
> > another interesting angle... I thought it might be power or some
> > kind
> > of internal noise. How could I pursue that further?  
> > 
> > Try running something on all CPU cores so that it doesn't switch
> > into
> > power saving mode.

I wrote a small cpu-hog to exercise all 8 cores. I didn't notice any
difference with the metering on jkmeter. My CPU fans increased speed
though.

Now that I'm thinking about it, wouldn't any type of interference
register under windows as well?

Have there been any other reports of this nature regarding 32 vs 64-bit
anomalies?

Re: ice1712 and the delta 1010

[Clemens Ladisch]

David Santamauro wrote:
> Now that I'm thinking about it, wouldn't any type of interference
> register under windows as well?

If Windows uses a different interrupt frequency that isn't in the
audible range, the power fluctuations wouldn't be noticeable.  Or it
could use some different power management mechanism.

> Have there been any other reports of this nature regarding 32 vs 64-bit
> anomalies?

You are using a different version of the distribution, so it might be
as well any other configuration change.


Regards,
Clemens