Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

[Takashi Iwai]

At Tue, 23 Dec 2008 10:44:50 -0500,
Kyler Laird wrote:
> 
> My Panasonic CF-74 Toughbooks do not play audio through the headphone
> jack starting with 2.6.28-rc5.  The patch below is only to illustrate
> what changes I made to restore headphone jack functionality.  I do not
> suggest using this patch as part of the mainline kernel because it
> reverts some code that I'm sure is important to other users.
> 
> Output of the ALSA script aadebug follows the patch.
> 
> I'm continuing to investigate but I wanted to get this out in case the
> solution is obvious to someone else.  It would be great to have a fix in
> 2.6.28.  (2.6.28 is the first series to support all of the hardware in
> the Toughbook.)

Please run alsa-info.sh with --no-upload option and attach the
generated files on both the headphone plugged and unplugged.
Also, run preferably on both (old) working and non-working kernels
to compare.


thanks,

Takashi

Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

[Takashi Iwai]

Hi,

sorry for the late reply, as I lately got back from vacation.

At Thu, 1 Jan 2009 09:57:30 -0500,
Kyler Laird wrote:
> 
> On Thu, Jan 01, 2009 at 11:04:06AM +0100, Takashi Iwai wrote:
> 
> > Please run alsa-info.sh with --no-upload option and attach the
> > generated files on both the headphone plugged and unplugged.
> > Also, run preferably on both (old) working and non-working kernels
> > to compare.
> 
> There appears to be no difference in output between headphones plugged
> and unplugged.

Thanks.  So, it implies that the device actually doesn't need
the headphone detection but rather does on hardware by itself.

How about the patch below?
It's against the latest git but should work with 2.6.28, too.


Takashi

---
diff --git a/sound/pci/hda/patch_sigmatel.c b/sound/pci/hda/patch_sigmatel.c
index 35b83dc..b6040b7 100644
--- a/sound/pci/hda/patch_sigmatel.c
+++ b/sound/pci/hda/patch_sigmatel.c
@@ -4408,6 +4408,9 @@ static int patch_stac9200(struct hda_codec *codec)
 		return err;
 	}
 
+	if (spec->board_config == STAC_9200_PANASONIC)
+		spec->hp_detect = 0;
+
 	codec->patch_ops = stac92xx_patch_ops;
 
 	return 0;

Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

[Takashi Iwai]

At Thu, 8 Jan 2009 12:10:27 -0500,
Kyler Laird wrote:
> 
> On Thu, Jan 08, 2009 at 04:52:59PM +0100, Takashi Iwai wrote:
> 
> > > There appears to be no difference in output between headphones plugged
> > > and unplugged.
> > 
> > Thanks.  So, it implies that the device actually doesn't need
> > the headphone detection but rather does on hardware by itself.
> > 
> > How about the patch below?
> > It's against the latest git but should work with 2.6.28, too.
> 
> The patch applied alright but did not seem to help.  I applied my patch
> on top of it and it did work again.  That surprised me.  It seems that
> your patch just disables hp_detect but apparently STAC_HP_EVENT is still
> being received by stac92xx_unsol_event() and that's calling
> stac92xx_hp_detect().
> 
> I added a printk() to stac92xx_unsol_event() and see that it's called
> only once, right as the module is loaded (res=3221225472).  It is not
> called again when I plug or unplug the headphones.
> 
> I agree that there does not seem to be headphone detection but it does
> seem that another audio pathway needs to be enabled in order for the
> headphones to work.

Could you add dump_stack() in stac92xx_unsol_event() and check the
code path?

You are using 2.6.28, not the latest git tree, right?


thanks,

Takashi

Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

[Takashi Iwai]

At Thu, 8 Jan 2009 13:22:41 -0500,
Kyler Laird wrote:
> 
> On Thu, Jan 08, 2009 at 06:54:03PM +0100, Takashi Iwai wrote:
> 
> > Could you add dump_stack() in stac92xx_unsol_event() and check the
> > code path?
> 
> Jan  8 13:19:12 ruff kernel: HDA Intel 0000:00:1b.0: PCI INT A disabled
> Jan  8 13:19:12 ruff kernel: HDA Intel 0000:00:1b.0: power state changed by ACPI to D0
> Jan  8 13:19:12 ruff kernel: HDA Intel 0000:00:1b.0: PCI INT A -> GSI 22 (level, low) -> IRQ 22
> Jan  8 13:19:12 ruff kernel: HDA Intel 0000:00:1b.0: setting latency timer to 64
> Jan  8 13:19:12 ruff kernel: (Kyler) stac92xx_unsol_event: 3221225472
> Jan  8 13:19:12 ruff kernel: Pid: 13602, comm: modprobe Not tainted 2.6.28 #1
> Jan  8 13:19:12 ruff kernel: Call Trace:
> Jan  8 13:19:12 ruff kernel:  [<e0a30af3>] stac92xx_unsol_event+0x33/0x380 [snd_hda_intel]
> Jan  8 13:19:12 ruff kernel:  [<e0a303b0>] stac92xx_init+0x120/0x490 [snd_hda_intel]

Weird.  I thought this path is active only when hp_detect != 0.
Could you check whether it's really correctly patched, e.g. by the
patch like below?

It seems applying my previous patch on 2.6.28 puts the code into a
wrong place.


Takashi

---
diff --git a/sound/pci/hda/patch_sigmatel.c b/sound/pci/hda/patch_sigmatel.c
index b77f330..9982e4b 100644
--- a/sound/pci/hda/patch_sigmatel.c
+++ b/sound/pci/hda/patch_sigmatel.c
@@ -3745,6 +3745,7 @@ static int stac92xx_init(struct hda_codec *codec)
 
 	/* set up pins */
 	if (spec->hp_detect) {
+		printk(KERN_DEBUG "XXX stac92xx_init -- hp_detect is ON!\n");
 		/* Enable unsolicited responses on the HP widget */
 		for (i = 0; i < cfg->hp_outs; i++)
 			enable_pin_detect(codec, cfg->hp_pins[i],
@@ -4129,6 +4130,11 @@ static int patch_stac9200(struct hda_codec *codec)
 		return err;
 	}
 
+	if (spec->board_config == STAC_9200_PANASONIC) {
+		printk(KERN_DEBUG "XXX panasonic quirk enabled\n");
+		spec->hp_detect = 0;
+	}
+
 	codec->patch_ops = stac92xx_patch_ops;
 
 	return 0;

Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

[Takashi Iwai]

At Fri, 9 Jan 2009 10:21:21 -0500,
Kyler Laird wrote:
> 
> On Fri, Jan 09, 2009 at 08:51:58AM +0100, Takashi Iwai wrote:
> 
> > Weird.  I thought this path is active only when hp_detect != 0.
> > Could you check whether it's really correctly patched, e.g. by the
> > patch like below?
> 
> That patch works.  Thanks!
> 
> > It seems applying my previous patch on 2.6.28 puts the code into a
> > wrong place.
> 
> I haven't succeeded in getting either of your recent patches to work
> against 2.6.28-git13.  I'll work on it more later.

The previous patch was for 2.6.28 while the first one was for git.
Maybe just because of a line difference.

Could you try test/cf74-fix branch on sound git tree?
  git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git test/cf74-fix

Takashi

Re: 2.6.28-rc5 broke ALSA headphone output for my Toughbooks.

[Takashi Iwai]

At Sat, 10 Jan 2009 09:34:57 -0500,
Kyler Laird wrote:
> 
> On Fri, Jan 09, 2009 at 05:14:45PM +0100, Takashi Iwai wrote:
> 
> > Could you try test/cf74-fix branch on sound git tree?
> >   git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6.git test/cf74-fix
> 
> That didn't work for me.  I see:
> 	HDA Intel 0000:00:1b.0: power state changed by ACPI to D0
> 	HDA Intel 0000:00:1b.0: PCI INT A -> GSI 22 (level, low) -> IRQ
> 	22
> 	HDA Intel 0000:00:1b.0: setting latency timer to 64
> 	input: HDA Intel at 0xe8640000 irq 22 Line In at Ext Rear Jack
> 	as /class/input/input7
> 	input: HDA Intel at 0xe8640000 irq 22 Mic at Ext Front Jack as
> 	/class/input/input8
> 	input: HDA Intel at 0xe8640000 irq 22 Mic at Ext Rear Jack as
> 	/class/input/input9
> 	input: HDA Intel at 0xe8640000 irq 22 Speaker at Ext Rear Jack
> 	as /class/input/input10
> 	input: HDA Intel at 0xe8640000 irq 22 HP Out at Ext Front Jack
> 	as /class/input/input11

These are all OK.  The new messages regarding input device are for
jack-sensing stuff.

Could you run alsa-info.sh on the latest kernel, too?

Also, to be sure, try to put a printk in the quirk for CF-74
(disabling hp_detect) to confirm it's activated.


thanks,

Takashi