Inspiron 8000 running 5.2.1 sound issues

[Scott Long]

Sean Welch wrote:

> I have written about this before asking for help but nothing I have
> tried up to this point has worked.
> 
> I'm running straight 5.2.1-RELEASE code and no matter what I modify
> in the code I get behavior very similar to what I saw in
> 5.2-RELEASE.  At that time someone was circulating a patch that
> fixed it for me.  The symptoms were (and are) periodic freezing of
> the sound for fractions of a second causing a "blurring" of the
> sound.  Something like the audio you hear in the Matrix when Neo has
> been given a "tracer" and is being overcome by the mirror like
> surface right before he wakes up for the first time.  Only on a much
> shorter time scale.
> 
> A similar effect can be heard when audio is playing on a laptop and
> the battery is removed (obviously have the laptop plugged in before
> doing this!).  Everything freezes and the sound is distorted and
> repeated for a bit.
> 
> I tried the patch I had used successfully under 5.2-RELEASE and it
> did nothing to alleviate the issue.  Someone else suggested a much
> more comprehensive patch and I got precisely the same result.
> Finally today I couldn't take it anymore and installed the source
> from 5.2-RELEASE and compiled a new snd_pcm.ko from that with the
> patch that had worked for me with that code -- same result!
> 
> Then I remembered similar issues with a 4.x release a while back
> (4.6 maybe?) that I tracked down to the charging cycle of the
> batteries.  Everything was fine with only one battery but I normally
> run two.  So I removed one battery.  Much better but still some
> anomalies.  I removed the other one and the sound was PERFECT.
> 
> Now this is *really* annoying.  I've found a way around the issue
> but it requires me to lose the batteries.  So much for crisp
> listening away from the wall socket...
> 
> Up until 5.2.1 I've had to replace the dsdt -- I've not been doing
> that with this release but figuring it had something to do with acpi
> I replaced it.  No difference.  I'm now back to running the
> snd_pcm.ko module that shipped with 5.2.1 because it is perfect with
> no batteries and has the same issues as the patched versions with
> batteries.
> 
> Help?  Any clues I might be able to follow up on?
> 
> Sean
> 

This almost sounds like the polling loop in ACPI is taking significantly
too long to do it's work.  With batteries inserted, do you get the same
problem with ACPI disabled?

Scott