[BUG] ali5451 not resumed properly under 2.6.7 ( fine under 2.6.6 )

[BUG] ali5451 not resumed properly under 2.6.7 ( fine under 2.6.6 )

[Fedor Karpelevitch]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I upgraded to 2.6.7 recently and noticed that my ali5451 souncard 
stopped behaving properly after resuming. 

Basically after resume it produces no sound even though everything 
pretends that the card is working properly (mixer "changes" volume 
etc...). 
I found that executing something like 'alsactl -F power A5451 D1' 
makes it produce sound again although the volume is noticably lower 
than it should be.

2.6.6 works just fine.

I looked at the diff for ali5451.c and noticed that in ali_suspend a 
call to snd_pcm_suspend_all(chip->pcm) was added, but in ali_resume 
no call to resume pcm was added. Could that be the cause of the 
problem?

I will try adding that call and see if that fixes the problem.


here is that part of the diff:



+static int ali_suspend(snd_card_t *card, unsigned int state)
 {
+       ali_t *chip = snd_magic_cast(ali_t, card->pm_private_data, 
return -EINVAL);
        ali_image_t *im;
        int i, j;
 
        im = chip->image;
        if (! im)
- -               return;
+               return 0;
+
+       snd_pcm_suspend_all(chip->pcm);
+       snd_ac97_suspend(chip->ac97);
 
        spin_lock_irq(&chip->reg_lock);
        
@@ -1940,16 +1943,18 @@
        outl(0xffffffff, ALI_REG(chip, ALI_STOP));
 
        spin_unlock_irq(&chip->reg_lock);
+       return 0;
 }
 
- -static void ali_resume(ali_t *chip)
+static int ali_resume(snd_card_t *card, unsigned int state)
 {
+       ali_t *chip = snd_magic_cast(ali_t, card->pm_private_data, 
return -EINVAL);
        ali_image_t *im;
        int i, j;
 
        im = chip->image;
        if (! im)
- -               return;
+               return 0;
 
        pci_enable_device(chip->pci);
 
@@ -1967,27 +1972,15 @@
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA2bgYw4m50RG4juoRAqTqAJ4+oHXRvOjz6NZLUS82oxIh+d7SIACgl65l
hCtld2qftBglwS0V0gYNHAw=
=ZLy7
-----END PGP SIGNATURE-----

Re: [BUG] ali5451 not resumed properly under 2.6.7 ( fine under 2.6.6 )

[Takashi Iwai]

At Wed, 23 Jun 2004 10:04:17 -0700,
Fedor Karpelevitch wrote:
> 
> Hi,
> 
> I upgraded to 2.6.7 recently and noticed that my ali5451 souncard 
> stopped behaving properly after resuming. 
> 
> Basically after resume it produces no sound even though everything 
> pretends that the card is working properly (mixer "changes" volume 
> etc...). 
> I found that executing something like 'alsactl -F power A5451 D1' 
> makes it produce sound again although the volume is noticably lower 
> than it should be.
> 
> 2.6.6 works just fine.
> 
> I looked at the diff for ali5451.c and noticed that in ali_suspend a 
> call to snd_pcm_suspend_all(chip->pcm) was added, but in ali_resume 
> no call to resume pcm was added. Could that be the cause of the 
> problem?

No, the resume is done either in pcm_oss.c or by calling the resume
ioctl explicitly from ALSA-native apps later on.

> I will try adding that call and see if that fixes the problem.

There is no function such as snd_pcm_resume_pcm().

Maybe you can remove the calls of snd_pcm_suspend_all(),
snd_ac97_suspend() in suspend callback, and snd_ac97_resume() in
resume callback.


--
Takashi Iwai <tiwai@suse.de>		ALSA Developer - www.alsa-project.org

Re: [BUG] ali5451 not resumed properly under 2.6.7 ( fine under 2.6.6 )

[Fedor Karpelevitch]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Takashi Iwai wrote:
> At Wed, 23 Jun 2004 10:04:17 -0700,
>
> Fedor Karpelevitch wrote:
> > Hi,
> >
> > I upgraded to 2.6.7 recently and noticed that my ali5451 souncard
> > stopped behaving properly after resuming.
> >
> > Basically after resume it produces no sound even though
> > everything pretends that the card is working properly (mixer
> > "changes" volume etc...).
> > I found that executing something like 'alsactl -F power A5451 D1'
> > makes it produce sound again although the volume is noticably
> > lower than it should be.
> >
> > 2.6.6 works just fine.
> >
> > I looked at the diff for ali5451.c and noticed that in
> > ali_suspend a call to snd_pcm_suspend_all(chip->pcm) was added,
> > but in ali_resume no call to resume pcm was added. Could that be
> > the cause of the problem?
>
> No, the resume is done either in pcm_oss.c or by calling the resume
> ioctl explicitly from ALSA-native apps later on.
>
> > I will try adding that call and see if that fixes the problem.
>
> There is no function such as snd_pcm_resume_pcm().

yeh, I figured that out... Should there be one?

>
> Maybe you can remove the calls of snd_pcm_suspend_all(),
> snd_ac97_suspend() in suspend callback, and snd_ac97_resume() in
> resume callback.
>

I tried this but it did not help, same result - no sound after resume, 
tweaking PCM volume does make sound come back, but it is noticably 
quieter than before suspend. I am back to 2.6.6 for now... :-(

Fedor
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFA2zrlw4m50RG4juoRAr5jAKClrmdZDMOKf12WVWQ5qmoyIv+z5ACfZit2
sBRBksHIVbqGF/Yljk1v/XY=
=eFiQ
-----END PGP SIGNATURE-----