[alsa-devel] module snd-atiixp issue with suspend/hibernate/resume

[Takashi Iwai]

At Mon, 4 May 2009 20:21:24 -0400,
Ryan Dunn wrote:
> 
> Thanks Takashi, I appreciate your help so far.  I would normally agree with
> you that it is a BIOS issue, however, the machine is a dualboot with WinXP,
> where the mute LED works properly after a resume.  Do you have any other
> suggestions on where the issue may be?  I hate to keep bugging the list for
> such a small issue, but I'm not familiar with the code at all.  When I get
> more free time, I'll try to trace through the resume path to see if I can find
> anything.  How does the driver actually toggle the LED?  Is the register a
> physical register on the card that the hardware reads?  Or does the code do
> something else to toggle the LED?

To be sure, you can toggle the bit of register 0x26 manually by writing
the proc file.  If it doesn't work, it's a BIOS issue indeed.

However, there might be some fix / workaround in the suspend part.
So, after confirming that the register change doesn't work, report it
on either LKML or kernel bugzilla.


Takashi

> 
> Ryan
> 
> On Mon, May 4, 2009 at 3:02 AM, Takashi Iwai <tiwai at suse.de> wrote:
> 
>     At Thu, 30 Apr 2009 22:24:35 -0400,
>     Ryan Dunn wrote:
>     >
>     > I apologize for the list spam.  I didn't realize that this bit is only
>     set
>     > when the led is on.  So with that here are diffs of the regs with and
>     without
>     > mute enabled, before and after suspend.
>     > Before suspend:
>     > diff regs_w-quirk.txt regs_w-quirk+mute.txt
>     > 2c2
>     > < 0:02 = 1e1e
>     > ---
>     > > 0:02 = 9e1e
>     > 20c20
>     > < 0:26 = 000f
>     > ---
>     > > 0:26 = 800f
>     >
>     > After resume:
>     > diff regs_w-quirk_after_resume.txt regs_w-quirk_after_resume+mute.txt
>     > 2c2
>     > < 0:02 = 1e1e
>     > ---
>     > > 0:02 = 9e1e
>     > 20c20
>     > < 0:26 = 000f
>     > ---
>     > > 0:26 = 800f
>     >
>     > So it looks as if the registers are all ok after a resume, but the LED
>     doesn't
>     > turn back on.
>    
>     Then I'd say it's rather a BIOS problem.
> 
>     Takashi
> 
>