[alsa-devel] Mute LED on HP Pavilion DV7-1160eg

[Takashi Iwai]

At Mon, 10 May 2010 22:10:23 +0200,
Thoralf Freitag wrote:
> 
> Takashi Iwai schrieb:
> > At Sat, 01 May 2010 22:19:14 +0200,
> > Thoralf Freitag wrote:
> >   
> >> I tried to find a way for switching/control Mute LED on an HP DV 7
> >> notebook. I found a solution which is very ugly. According the solution
> >> described here
> >> (http://mailman.alsa-project.org/pipermail/alsa-devel/2009-April/016126.html)
> >> I changed the value for LED control in this manner:
> >>
> >> diff -u alsa-driver-1.0.23/sound/pci/hda/patch_sigmatel.c
> >> alsa-driver-1.0.23_b/sound/pci/hda/patch_sigmatel.c
> >> --- alsa-driver-1.0.23/sound/pci/hda/patch_sigmatel.c    2010-04-16
> >> 13:10:10.000000000 +0200
> >> +++ alsa-driver-1.0.23_b/sound/pci/hda/patch_sigmatel.c    2010-05-01
> >> 21:37:36.719063655 +0200
> >> @@ -4952,9 +4952,9 @@
> >>          }
> >>      }
> >>      if (muted)
> >> -        spec->gpio_data &= ~spec->gpio_led; /* orange */
> >> +        spec->gpio_data &= ~0x01; /* orange */
> >>      else
> >> -        spec->gpio_data |= spec->gpio_led; /* white */
> >> +        spec->gpio_data |= 0x01; /* white */
> >>  
> >>      if (!spec->gpio_led_polarity) {
> >>          /* LED state is inverted on these systems */
> >>
> >> I hate this solution. Maybe someone of the programming experts could
> >> give me a way to make it better. On which position should I make which
> >> changes ?
> >>     
> >
> > The mute LED is set up in find_mute_led_gpio().
> > If your BIOS doesn't give the proper information, you can add the device
> > statically in hp_blike_system().
> >
> >
> > Takashi
> > _______________________________________________
> > Alsa-devel mailing list
> > Alsa-devel at alsa-project.org
> > http://mailman.alsa-project.org/mailman/listinfo/alsa-devel
> >   
> Takashi, thanks for your advice. Maybe I found the problem. Adding my
> card to hp_blike_system() doesn't change anything. The driver identifies
> the card (laptop) successfully as a HP DV7 (because the bass speaker is
> usable and can switched on or off with tools like Alsamixergui), with or
> without an entry in hp_blike_system() like this: case 0x103c30fc: .
> 
> As I identified, the problem is located in set_hp_led_gpio. This
> function assumes, if it counts more than 3 io's, the GPIO 3 is
> responsible for controlling of mute LED. Maybe it is right in some
> cases, but not in all. For the HP7-1160eg (and  probably not only on
> this notebook) GPIO 0 controlls the mute LED, but the counter is
> 
> I could fix it by changing the value from 3 to 8 in static void
> set_hp_led_gpio(struct hda_codec *codec). I don't know, there the value
> of 3 comes from. But I afraid, it could causing a lot of side effects.
> Who knows, why the threshold value is set greater then 3 ?
> 
> kaktus at elefant-lin:~$ diff
> alsa-driver-1.0.23_a/alsa-kernel/pci/hda/patch_sigmatel.c
> alsa-driver-1.0.23_b/alsa-kernel/pci/hda/patch_sigmatel.c
> 
> 4769c4769
> <     if (gpio > 3)
> ---
> >     if (gpio > 8)
> 
> 
> This is the function:
> 
> static void set_hp_led_gpio(struct hda_codec *codec)
> {
>         struct sigmatel_spec *spec = codec->spec;
>         unsigned int gpio;
> 
>         gpio = snd_hda_param_read(codec, codec->afg, AC_PAR_GPIO_CAP);
>         gpio &= AC_GPIO_IO_COUNT;
>         if (gpio > 8)
>                 spec->gpio_led = 0x08; /* GPIO 3 */
>         else
>                 spec->gpio_led = 0x01; /* GPIO 0 */
> }

This isn't correct -- gpio here returned gives the number of GPIO
pins, and your codec chip has indeed 8 pins, thus its bitmask is
up to 0x80.  GPIO bit 0x08 means 3 pins.  That's how the number 3
was deduced in the original code.

> Here are some information about my sound-card. hda_analyzer shows:
> 
> Codec: 0x111d76b2
> Address: 0
> Function Id: 0x101
> Vendor Id: 0x111d76b2
> Subsystem Id: 0x103c30fd
> Revision Id: 0x100302
> No Modem Function Group found
> Default PCM:
>     rates [0x7e0]: 44100 48000 88200 96000 176400 192000
>     bits [0xe]: 16 20 24
>     formats [0x1]: PCM
> Default Amp-In caps: ofs=0x00, nsteps=0x00, stepsize=0x00, mute=1
> Default Amp-Out caps: ofs=0x7f, nsteps=0x7f, stepsize=0x02, mute=1
> GPIO: io=8, o=0, i=0, unsolicited=1, wake=1
>   IO[0]: enable=1, dir=1, wake=0, sticky=0, data=1, unsol=0
>   IO[1]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0
>   IO[2]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0
>   IO[3]: enable=1, dir=1, wake=0, sticky=0, data=0, unsol=0
>   IO[4]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0
>   IO[5]: enable=1, dir=1, wake=0, sticky=0, data=1, unsol=0
>   IO[6]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0
>   IO[7]: enable=0, dir=0, wake=0, sticky=0, data=0, unsol=0
> 
> 
> What should I do for for making a reliable patch. Any ideas ?

Based on the information from HP, I assumed that the large package
always contains 8 GPIO pins and GPIO 3 is used for mute LED.  But,
it seems some don't follow that rule, apparently.

The right fix is simply to add an exception check.
Could you try the patch below?


Takashi

---
diff --git a/sound/pci/hda/patch_sigmatel.c b/sound/pci/hda/patch_sigmatel.c
index 12825aa..fa8c25b 100644
--- a/sound/pci/hda/patch_sigmatel.c
+++ b/sound/pci/hda/patch_sigmatel.c
@@ -4766,6 +4766,12 @@ static void set_hp_led_gpio(struct hda_codec *codec)
 	struct sigmatel_spec *spec = codec->spec;
 	unsigned int gpio;
 
+	if (codec->vendor_id == 0x111d76b2) {
+		/* use GPIO 3 exceptionally */
+		spec->gpio_led = 0x08; /* GPIO 3 */
+		return;
+	}
+
 	gpio = snd_hda_param_read(codec, codec->afg, AC_PAR_GPIO_CAP);
 	gpio &= AC_GPIO_IO_COUNT;
 	if (gpio > 3)