On Friday 08 August 2014 08:11:18 Takashi Iwai wrote:

At Thu, 07 Aug 2014 18:35:39 +0200, Gabriele Mazzotta wrote:

...

On init, mic-in is always set as input source, indipendently on what is plugged in. Since setting/unsetting mic-in as input source causes a pop noise, make sure the internal microphone is selected as input source on boot.

On shutdown, make sure the codec is not suspended as that would cause a pop noise.

Signed-off-by: Gabriele Mazzotta gabriele.mzt@gmail.com

sound/pci/hda/hda_codec.c | 2 ++ sound/pci/hda/hda_codec.h | 1 + sound/pci/hda/patch_realtek.c | 12 ++++++++++++ 3 files changed, 15 insertions(+)

diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 4c20277..92d8292 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -5348,6 +5348,8 @@ void snd_hda_bus_reboot_notify(struct hda_bus *bus) if (!bus) return; list_for_each_entry(codec, &bus->codec_list, list) {

• if (codec->resume_at_reboot)
  

• 	hda_call_codec_resume(codec);
  

Do you really need the resume at reboot? I thought the codec needs power down at reboot, i.e. rather it needs suspend.

Yes, the pop noise happens if it's suspended. Now that I payed more attention, the pop noise happens only when rebooting the laptop and not when powering it down. Also, the noise is heard quite late, I think the kernel is no longer running when I hear it. I will try to better understand what is actually happening and why resuming prevents the noise.

Since you apply actually two individual changes, it's better to split patches: one for adding the power-down at reboot, another for initializing cur_mux properly.

Also, it's recommended to put a bugzilla link in the changelog, so that other people can follow the information and discussion.

Last but not least, please put maintainers to Cc at the next time, so that it won't be missed.

I will do it.

Thanks,

Gabriele

thanks,

Takashi

...

if (hda_codec_is_power_on(codec) &&
    codec->patch_ops.reboot_notify)
	codec->patch_ops.reboot_notify(codec);

diff --git a/sound/pci/hda/hda_codec.h b/sound/pci/hda/hda_codec.h index 5825aa1..5c3c66e 100644 --- a/sound/pci/hda/hda_codec.h +++ b/sound/pci/hda/hda_codec.h @@ -366,6 +366,7 @@ struct hda_codec { unsigned int cached_write:1; /* write only to caches */ unsigned int dp_mst:1; /* support DP1.2 Multi-stream transport */ unsigned int dump_coef:1; /* dump processing coefs in codec proc file */

• unsigned int resume_at_reboot:1; /* resume codec at reboot */

#ifdef CONFIG_PM unsigned int power_on :1; /* current (global) power-state */ unsigned int d3_stop_clk:1; /* support D3 operation without BCLK */ diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index b60824e..defaa2a 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -4018,8 +4018,20 @@ static void alc_fixup_dell_xps13(struct hda_codec *codec, { if (action == HDA_FIXUP_ACT_PROBE) { struct alc_spec *spec = codec->spec;

• struct hda_input_mux *imux = &spec->gen.input_mux;
  

• int i;
  

• spec->shutup = alc_no_shutup; codec->power_filter = alc_power_filter_xps13;

• codec->resume_at_reboot = 1;
  

• /* Make the internal mic the default input source. */
  

• for (i = 0; i < imux->num_items; i++) {
  

• 	if (spec->gen.imux_pins[i] == 0x12) {
  

• 		spec->gen.cur_mux[0] = i;
  

• 		break;
  

• 	}
  

• }
  

  }

}

On Friday 08 August 2014 12:13:38 you wrote:

At Fri, 08 Aug 2014 10:49:25 +0200, Gabriele Mazzotta wrote:

...

On Friday 08 August 2014 08:11:18 Takashi Iwai wrote:

...

At Thu, 07 Aug 2014 18:35:39 +0200, Gabriele Mazzotta wrote:

...

On init, mic-in is always set as input source, indipendently on what is plugged in. Since setting/unsetting mic-in as input source causes a pop noise, make sure the internal microphone is selected as input source on boot.

On shutdown, make sure the codec is not suspended as that would cause a pop noise.

Signed-off-by: Gabriele Mazzotta gabriele.mzt@gmail.com

sound/pci/hda/hda_codec.c | 2 ++ sound/pci/hda/hda_codec.h | 1 + sound/pci/hda/patch_realtek.c | 12 ++++++++++++ 3 files changed, 15 insertions(+)

diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 4c20277..92d8292 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -5348,6 +5348,8 @@ void snd_hda_bus_reboot_notify(struct hda_bus *bus) if (!bus) return; list_for_each_entry(codec, &bus->codec_list, list) {

• if (codec->resume_at_reboot)
  

• 	hda_call_codec_resume(codec);
  

Do you really need the resume at reboot? I thought the codec needs power down at reboot, i.e. rather it needs suspend.

Yes, the pop noise happens if it's suspended.

Then the patch is wrong in anyway. You cannot call it unconditionally. The codec might be still active, and calling resume in that state does re-initialize the whole things doubly.

Now I see that it's wrong. In addition, I guess it also breaks the build when CONFIG_PM is not set.

...

Now that I payed more attention, the pop noise happens only when rebooting the laptop and not when powering it down. Also, the noise is heard quite late, I think the kernel is no longer running when I hear it. I will try to better understand what is actually happening and why resuming prevents the noise.

If so, it might be BIOS causes the noise by re-initializing the codec.

It seems that if 0x15 (Headphone playback switch) is in D0, the pop noise cannot be heard. I tried to see using powertop whether it's worth making the current code more complex and force nid 0x15 in D0 only at reboot or not and I couldn't notice any major difference. Between 0x02, 0x15 and AFG only the last is causing a noticeable difference in the power consumption.

That said, I would simply add 0x15 to alc_power_filter_xps13(). I will send the patches soon.

Gabriele

Takashi

...

...

Since you apply actually two individual changes, it's better to split patches: one for adding the power-down at reboot, another for initializing cur_mux properly.

Also, it's recommended to put a bugzilla link in the changelog, so that other people can follow the information and discussion.

Last but not least, please put maintainers to Cc at the next time, so that it won't be missed.

I will do it.

Thanks,

Gabriele

...

thanks,

Takashi

...

if (hda_codec_is_power_on(codec) &&
    codec->patch_ops.reboot_notify)
	codec->patch_ops.reboot_notify(codec);

diff --git a/sound/pci/hda/hda_codec.h b/sound/pci/hda/hda_codec.h index 5825aa1..5c3c66e 100644 --- a/sound/pci/hda/hda_codec.h +++ b/sound/pci/hda/hda_codec.h @@ -366,6 +366,7 @@ struct hda_codec { unsigned int cached_write:1; /* write only to caches */ unsigned int dp_mst:1; /* support DP1.2 Multi-stream transport */ unsigned int dump_coef:1; /* dump processing coefs in codec proc file */

• unsigned int resume_at_reboot:1; /* resume codec at reboot */

#ifdef CONFIG_PM unsigned int power_on :1; /* current (global) power-state */ unsigned int d3_stop_clk:1; /* support D3 operation without BCLK */ diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index b60824e..defaa2a 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -4018,8 +4018,20 @@ static void alc_fixup_dell_xps13(struct hda_codec *codec, { if (action == HDA_FIXUP_ACT_PROBE) { struct alc_spec *spec = codec->spec;

• struct hda_input_mux *imux = &spec->gen.input_mux;
  

• int i;
  

• spec->shutup = alc_no_shutup; codec->power_filter = alc_power_filter_xps13;

• codec->resume_at_reboot = 1;
  

• /* Make the internal mic the default input source. */
  

• for (i = 0; i < imux->num_items; i++) {
  

• 	if (spec->gen.imux_pins[i] == 0x12) {
  

• 		spec->gen.cur_mux[0] = i;
  

• 		break;
  

• 	}
  

• }
  

  }

}