[alsa-devel] [PATCH] ALSA: hda - Fix clicking noise on Thinkpad T61/R61i
The bug reporter reports that setting the speaker pin to D3 before turning off its pinctl fixes the clicking noise on powersave for Thinkpad T61.
Thanks to c4pp4 for doing most of the work with this bug, including reporting and testing, and writing preliminary patches.
BugLink: https://bugs.launchpad.net/bugs/886975 Signed-off-by: David Henningsson david.henningsson@canonical.com --- sound/pci/hda/patch_conexant.c | 36 ++++++++++++++++++++++++++++++++---- 1 file changed, 32 insertions(+), 4 deletions(-)
Note: c4pp4 pointed out that in his case, only the speaker needed to go to D3 before pinctl, but noticed no regressions from the below implementation, so I made the fixup more generic (easier to reuse later).
diff --git a/sound/pci/hda/patch_conexant.c b/sound/pci/hda/patch_conexant.c index a3a2263..27e6ca7 100644 --- a/sound/pci/hda/patch_conexant.c +++ b/sound/pci/hda/patch_conexant.c @@ -144,6 +144,7 @@ struct conexant_spec { unsigned int asus:1; unsigned int pin_eapd_ctrls:1; unsigned int fixup_stereo_dmic:1; + unsigned int d3_before_shutup:1; /* Needs speaker to go to D3 before pinctl */
unsigned int adc_switching:1;
@@ -558,13 +559,22 @@ static int conexant_build_controls(struct hda_codec *codec) return 0; }
-#ifdef CONFIG_PM static int conexant_suspend(struct hda_codec *codec) { + struct conexant_spec *spec = codec->spec; + + if (spec->d3_before_shutup) + snd_hda_codec_set_power_to_all(codec, codec->afg, AC_PWRST_D3, + false); snd_hda_shutup_pins(codec); return 0; } -#endif + +static void conexant_reboot_notify(struct hda_codec *codec) +{ + conexant_suspend(codec); +} +
static const struct hda_codec_ops conexant_patch_ops = { .build_controls = conexant_build_controls, @@ -575,7 +585,7 @@ static const struct hda_codec_ops conexant_patch_ops = { #ifdef CONFIG_PM .suspend = conexant_suspend, #endif - .reboot_notify = snd_hda_shutup_pins, + .reboot_notify = conexant_reboot_notify, };
#ifdef CONFIG_SND_HDA_INPUT_BEEP @@ -4408,7 +4418,7 @@ static const struct hda_codec_ops cx_auto_patch_ops = { #ifdef CONFIG_PM .suspend = conexant_suspend, #endif - .reboot_notify = snd_hda_shutup_pins, + .reboot_notify = conexant_reboot_notify, };
/* @@ -4421,6 +4431,7 @@ enum { CXT_PINCFG_LEMOTE_A1205, CXT_FIXUP_STEREO_DMIC, CXT_FIXUP_INC_MIC_BOOST, + CXT_FIXUP_D3_BEFORE_SHUTUP, };
static void cxt_fixup_stereo_dmic(struct hda_codec *codec, @@ -4430,6 +4441,13 @@ static void cxt_fixup_stereo_dmic(struct hda_codec *codec, spec->fixup_stereo_dmic = 1; }
+static void cxt_fixup_d3_before_shutup(struct hda_codec *codec, + const struct hda_fixup *fix, int action) +{ + struct conexant_spec *spec = codec->spec; + spec->d3_before_shutup = 1; +} + static void cxt5066_increase_mic_boost(struct hda_codec *codec, const struct hda_fixup *fix, int action) { @@ -4499,6 +4517,15 @@ static const struct hda_fixup cxt_fixups[] = { .type = HDA_FIXUP_FUNC, .v.func = cxt5066_increase_mic_boost, }, + [CXT_FIXUP_D3_BEFORE_SHUTUP] = { + .type = HDA_FIXUP_FUNC, + .v.func = cxt_fixup_d3_before_shutup, + }, +}; + +static const struct snd_pci_quirk cxt5045_fixups[] = { + SND_PCI_QUIRK(0x17aa, 0x20ac, "Lenovo T61/R61i", CXT_FIXUP_D3_BEFORE_SHUTUP), + {} };
static const struct snd_pci_quirk cxt5051_fixups[] = { @@ -4553,6 +4580,7 @@ static int patch_conexant_auto(struct hda_codec *codec) switch (codec->vendor_id) { case 0x14f15045: codec->single_adc_amp = 1; + snd_hda_pick_fixup(codec, NULL, cxt5045_fixups, cxt_fixups); break; case 0x14f15051: add_cx5051_fake_mutes(codec);
At Wed, 12 Dec 2012 18:02:04 +0100, David Henningsson wrote:
The bug reporter reports that setting the speaker pin to D3 before turning off its pinctl fixes the clicking noise on powersave for Thinkpad T61.
Thanks to c4pp4 for doing most of the work with this bug, including reporting and testing, and writing preliminary patches.
BugLink: https://bugs.launchpad.net/bugs/886975 Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_conexant.c | 36 ++++++++++++++++++++++++++++++++---- 1 file changed, 32 insertions(+), 4 deletions(-)
Note: c4pp4 pointed out that in his case, only the speaker needed to go to D3 before pinctl, but noticed no regressions from the below implementation, so I made the fixup more generic (easier to reuse later).
Well, if going to D3 is already achieved there, there is no point to call snd_hda_shutup_pins(). snd_hda_shutup_pins() itself is for reducing the click noise while the codec is going down to D3.
Thus, for this particular machine, simply disable snd_hda_shutup_pins() call from suspend, and make D3 call in reboot_notify.
Takashi
diff --git a/sound/pci/hda/patch_conexant.c b/sound/pci/hda/patch_conexant.c index a3a2263..27e6ca7 100644 --- a/sound/pci/hda/patch_conexant.c +++ b/sound/pci/hda/patch_conexant.c @@ -144,6 +144,7 @@ struct conexant_spec { unsigned int asus:1; unsigned int pin_eapd_ctrls:1; unsigned int fixup_stereo_dmic:1;
unsigned int d3_before_shutup:1; /* Needs speaker to go to D3 before pinctl */
unsigned int adc_switching:1;
@@ -558,13 +559,22 @@ static int conexant_build_controls(struct hda_codec *codec) return 0; }
-#ifdef CONFIG_PM static int conexant_suspend(struct hda_codec *codec) {
- struct conexant_spec *spec = codec->spec;
- if (spec->d3_before_shutup)
snd_hda_codec_set_power_to_all(codec, codec->afg, AC_PWRST_D3,false);} -#endif
+static void conexant_reboot_notify(struct hda_codec *codec) +{
- conexant_suspend(codec);
+}
static const struct hda_codec_ops conexant_patch_ops = { .build_controls = conexant_build_controls, @@ -575,7 +585,7 @@ static const struct hda_codec_ops conexant_patch_ops = { #ifdef CONFIG_PM .suspend = conexant_suspend, #endif
- .reboot_notify = snd_hda_shutup_pins,
- .reboot_notify = conexant_reboot_notify,
};
#ifdef CONFIG_SND_HDA_INPUT_BEEP @@ -4408,7 +4418,7 @@ static const struct hda_codec_ops cx_auto_patch_ops = { #ifdef CONFIG_PM .suspend = conexant_suspend, #endif
- .reboot_notify = snd_hda_shutup_pins,
- .reboot_notify = conexant_reboot_notify,
};
/* @@ -4421,6 +4431,7 @@ enum { CXT_PINCFG_LEMOTE_A1205, CXT_FIXUP_STEREO_DMIC, CXT_FIXUP_INC_MIC_BOOST,
- CXT_FIXUP_D3_BEFORE_SHUTUP,
};
static void cxt_fixup_stereo_dmic(struct hda_codec *codec, @@ -4430,6 +4441,13 @@ static void cxt_fixup_stereo_dmic(struct hda_codec *codec, spec->fixup_stereo_dmic = 1; }
+static void cxt_fixup_d3_before_shutup(struct hda_codec *codec,
const struct hda_fixup *fix, int action)+{
- struct conexant_spec *spec = codec->spec;
- spec->d3_before_shutup = 1;
+}
static void cxt5066_increase_mic_boost(struct hda_codec *codec, const struct hda_fixup *fix, int action) { @@ -4499,6 +4517,15 @@ static const struct hda_fixup cxt_fixups[] = { .type = HDA_FIXUP_FUNC, .v.func = cxt5066_increase_mic_boost, },
- [CXT_FIXUP_D3_BEFORE_SHUTUP] = {
.type = HDA_FIXUP_FUNC,.v.func = cxt_fixup_d3_before_shutup,- },
+};
+static const struct snd_pci_quirk cxt5045_fixups[] = {
- SND_PCI_QUIRK(0x17aa, 0x20ac, "Lenovo T61/R61i", CXT_FIXUP_D3_BEFORE_SHUTUP),
- {}
};
static const struct snd_pci_quirk cxt5051_fixups[] = { @@ -4553,6 +4580,7 @@ static int patch_conexant_auto(struct hda_codec *codec) switch (codec->vendor_id) { case 0x14f15045: codec->single_adc_amp = 1;
snd_hda_pick_fixup(codec, NULL, cxt5045_fixups, cxt_fixups);-- 1.7.9.5
On 12/12/2012 06:15 PM, Takashi Iwai wrote:
At Wed, 12 Dec 2012 18:02:04 +0100, David Henningsson wrote:
The bug reporter reports that setting the speaker pin to D3 before turning off its pinctl fixes the clicking noise on powersave for Thinkpad T61.
Thanks to c4pp4 for doing most of the work with this bug, including reporting and testing, and writing preliminary patches.
BugLink: https://bugs.launchpad.net/bugs/886975 Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_conexant.c | 36 ++++++++++++++++++++++++++++++++---- 1 file changed, 32 insertions(+), 4 deletions(-)
Note: c4pp4 pointed out that in his case, only the speaker needed to go to D3 before pinctl, but noticed no regressions from the below implementation, so I made the fixup more generic (easier to reuse later).
Well, if going to D3 is already achieved there, there is no point to call snd_hda_shutup_pins(). snd_hda_shutup_pins() itself is for reducing the click noise while the codec is going down to D3.
Thus, for this particular machine, simply disable snd_hda_shutup_pins() call from suspend, and make D3 call in reboot_notify.
Looking at the bug report, c4pp4 has just reported back that reverting the old commit 697c373e34613609cb5 improves the situation - there are less clicks compared to the current situation.
Quoting c4pp4 below:
"The patch [697c373e34613609cb5] in fact caused click noise hell because the initial click noise is still alive and in addition there were born new click noises, louder click noise and double click noise. They start to appear when using power-save mode, when suspending computer and when rebooting computer."
So maybe revert that patch?
At Mon, 17 Dec 2012 16:33:10 +0100, David Henningsson wrote:
On 12/12/2012 06:15 PM, Takashi Iwai wrote:
At Wed, 12 Dec 2012 18:02:04 +0100, David Henningsson wrote:
The bug reporter reports that setting the speaker pin to D3 before turning off its pinctl fixes the clicking noise on powersave for Thinkpad T61.
Thanks to c4pp4 for doing most of the work with this bug, including reporting and testing, and writing preliminary patches.
BugLink: https://bugs.launchpad.net/bugs/886975 Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_conexant.c | 36 ++++++++++++++++++++++++++++++++---- 1 file changed, 32 insertions(+), 4 deletions(-)
Note: c4pp4 pointed out that in his case, only the speaker needed to go to D3 before pinctl, but noticed no regressions from the below implementation, so I made the fixup more generic (easier to reuse later).
Well, if going to D3 is already achieved there, there is no point to call snd_hda_shutup_pins(). snd_hda_shutup_pins() itself is for reducing the click noise while the codec is going down to D3.
Thus, for this particular machine, simply disable snd_hda_shutup_pins() call from suspend, and make D3 call in reboot_notify.
Looking at the bug report, c4pp4 has just reported back that reverting the old commit 697c373e34613609cb5 improves the situation - there are less clicks compared to the current situation.
Quoting c4pp4 below:
"The patch [697c373e34613609cb5] in fact caused click noise hell because the initial click noise is still alive and in addition there were born new click noises, louder click noise and double click noise. They start to appear when using power-save mode, when suspending computer and when rebooting computer."
So maybe revert that patch?
Yes, feel free to submit the patch.
Takashi
participants (2)
-
David Henningsson -
Takashi Iwai