[PATCH] ALSA: patch_realtek: handle unset external amp bits

[PATCH] ALSA: patch_realtek: handle unset external amp bits

[Kyle McMartin]

From: Kyle McMartin <kyle@redhat.com>

Commit ef5dbbcc removes the default initialization of init_amp to
ALC_INIT_DEFAULT, with the justification that it should be set by
alc_subsystem_id, which contains a switch case:

        tmp = (ass & 0x38) >> 3;        /* external Amp control */
        switch (tmp) {
        case 1:
	...
	case 5:
	}

Which handles values 1,3,7 for GPIOs, or 5 to set _DEFAULT.

But... what if it's 0? :)

I got a bug report this morning from a user whose sound had broken
between 2.6.34 and 2.6.35, and I bisected it down to this commit.

[ 32.366183] ALSA sound/pci/hda/patch_realtek.c:1374: realtek: Enabling
init ASM_ID=0x8a05 CODEC_ID=10ec0888

(0x8a05 & 0x38) >> 3 = 0;

But otherwise the SKU looks fine (since if I revert that commit, it
seems to work correctly.)

I think the best thing to do would be to either restore the 'redundant'
default initialization, or quirk it.

https://bugzilla.redhat.com/show_bug.cgi?id=657388

Signed-off-by: Kyle McMartin <kyle@redhat.com>
---
    Revert "ALSA: hda - Remove superfluous external amp setup for ALC888"
    
    This reverts commit ef5dbbccbbfa7d2211fa8efcc095a9f4a7912dda.

diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c
index 76c4968..b6ba1d5 100644
--- a/sound/pci/hda/patch_realtek.c
+++ b/sound/pci/hda/patch_realtek.c
@@ -10604,6 +10604,9 @@ static int patch_alc882(struct hda_codec *codec)
 	spec->stream_digital_playback = &alc882_pcm_digital_playback;
 	spec->stream_digital_capture = &alc882_pcm_digital_capture;
 
+	if (codec->vendor_id == 0x10ec0888)
+		spec->init_amp = ALC_INIT_DEFAULT; /* always initialize */
+
 	if (!spec->adc_nids && spec->input_mux) {
 		int i, j;
 		spec->num_adc_nids = 0;

Re: [PATCH] ALSA: patch_realtek: handle unset external amp bits

[Takashi Iwai]

At Thu, 25 Nov 2010 13:57:30 -0500,
Kyle McMartin wrote:
> 
> From: Kyle McMartin <kyle@redhat.com>
> 
> Commit ef5dbbcc removes the default initialization of init_amp to
> ALC_INIT_DEFAULT, with the justification that it should be set by
> alc_subsystem_id, which contains a switch case:
> 
>         tmp = (ass & 0x38) >> 3;        /* external Amp control */
>         switch (tmp) {
>         case 1:
> 	...
> 	case 5:
> 	}
> 
> Which handles values 1,3,7 for GPIOs, or 5 to set _DEFAULT.
> 
> But... what if it's 0? :)
> 
> I got a bug report this morning from a user whose sound had broken
> between 2.6.34 and 2.6.35, and I bisected it down to this commit.
> 
> [ 32.366183] ALSA sound/pci/hda/patch_realtek.c:1374: realtek: Enabling
> init ASM_ID=0x8a05 CODEC_ID=10ec0888
> 
> (0x8a05 & 0x38) >> 3 = 0;
> 
> But otherwise the SKU looks fine (since if I revert that commit, it
> seems to work correctly.)
> 
> I think the best thing to do would be to either restore the 'redundant'
> default initialization, or quirk it.
> 
> https://bugzilla.redhat.com/show_bug.cgi?id=657388
> 
> Signed-off-by: Kyle McMartin <kyle@redhat.com>

How about setting default to ALC_INIT_DEFAULT instead like the patch
below?  This is safer than forcing ALC_INIT_DEFAULT to all, I
suppose.


thanks,

Takashi

---
diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c
index 81a2a49..886d7c7 100644
--- a/sound/pci/hda/patch_realtek.c
+++ b/sound/pci/hda/patch_realtek.c
@@ -1614,6 +1614,7 @@ do_sku:
 		spec->init_amp = ALC_INIT_GPIO3;
 		break;
 	case 5:
+	default:
 		spec->init_amp = ALC_INIT_DEFAULT;
 		break;
 	}

Re: [PATCH] ALSA: patch_realtek: handle unset external amp bits

[Kyle McMartin]

On Fri, Nov 26, 2010 at 06:51:11AM +0100, Takashi Iwai wrote:
> How about setting default to ALC_INIT_DEFAULT instead like the patch
> below?  This is safer than forcing ALC_INIT_DEFAULT to all, I
> suppose.
> 

As I understand the code, that should work fine. I'll build a new kernel
for my tester to confirm, I suspect it will though. :)

Thanks!
--Kyle

Re: [PATCH] ALSA: patch_realtek: handle unset external amp bits

[Takashi Iwai]

At Fri, 26 Nov 2010 08:29:22 -0500,
Kyle McMartin wrote:
> 
> On Fri, Nov 26, 2010 at 06:51:11AM +0100, Takashi Iwai wrote:
> > How about setting default to ALC_INIT_DEFAULT instead like the patch
> > below?  This is safer than forcing ALC_INIT_DEFAULT to all, I
> > suppose.
> > 
> 
> As I understand the code, that should work fine. I'll build a new kernel
> for my tester to confirm, I suspect it will though. :)

Thanks!


Takashi

Re: [PATCH] ALSA: patch_realtek: handle unset external amp bits

[Kyle McMartin]

On Fri, Nov 26, 2010 at 03:04:13PM +0100, Takashi Iwai wrote:
> > As I understand the code, that should work fine. I'll build a new kernel
> > for my tester to confirm, I suspect it will though. :)
> 
> Thanks!
>

Tester confirms audio works after the patch. Thanks!

Re: [PATCH] ALSA: patch_realtek: handle unset external amp bits

[Takashi Iwai]

At Fri, 26 Nov 2010 11:07:14 -0500,
Kyle McMartin wrote:
> 
> On Fri, Nov 26, 2010 at 03:04:13PM +0100, Takashi Iwai wrote:
> > > As I understand the code, that should work fine. I'll build a new kernel
> > > for my tester to confirm, I suspect it will though. :)
> > 
> > Thanks!
> >
> 
> Tester confirms audio works after the patch. Thanks!

Thanks, now applied it.
I'm going to send a pull request to Linus tomorrow.


Takashi