January 2019 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH v3 0/9] ASoC: sun4i-i2s: Updates to the driver
by codekipper＠gmail.com 09 Jan '19

09 Jan '19

From: Marcus Cooper <codekipper(a)gmail.com> Hi All, here is a patch series to add some improvements to the sun4i-i2s driver found whilst getting slave clocking and hdmi audio working on the newer SoCs. Since the last push there has been some activity getting surround sound working and this is included. The functionality included with the new patch set has been extended to cover more sample resolutions, multi-lane data output for HDMI audio and some bug fixes that have been discovered along the way. I can see more usage of the tdm property since I last attempted to push these patches and the examples currently in mainline sort of the opposite to what I'm trying to achieve. When we first started looking at the i2s driver, the codecs that we were using allowed for the frame width to be determined based on the sampling resolution but in most use cases it seems that a fixed width is required(my highest priority should be to get HDMI audio support in). We're using the tdm property to override the old way to calculate the frame width. What I've seen in what has already been mainlined is that the i2s driver has a frame width that is fixed to 32 bits and this can be overridden using the tdm property. Anyway, I've moved the more controversial patches to the top of the stack as I'm expecting feedback. Have a great Xmas and New Year, BR, CK --- v3 changes compared to v2 are: - added back slave mode changes - added back the use of tdm properties - changes to regmap and caching - removed loopback functionality - fixes to the channel offset mask v2 changes compared to v1 are: - removed slave mode changes which didn't set mclk and bclk div. - removed use of tdm and now use a dedicated property. - fix commit message to better explain reason for sign extending - add divider calculations for newer SoCs. - add support for multi-lane i2s data output. - add support for 20, 24 and 32 bit samples. - add loopback property so blocks can be tested without a codec. --- Marcus Cooper (9): ASoC: sun4i-i2s: Adjust regmap settings ASoC: sun4i-i2s: Add regmap field to sign extend sample ASoc: sun4i-i2s: Add 20, 24 and 32 bit support ASoC: sun4i-i2s: Fix offset mask ASoC: sun4i-i2s: Correct divider calculations ASoC: sun4i-i2s: Add multi-lane functionality ASoC: sun4i-i2s: Do not divide clocks when slave ASoC: sun4i-i2s: Add set_tdm_slot functionality ASoC: sun4i-i2s: Add multichannel functionality sound/soc/sunxi/sun4i-i2s.c | 399 ++++++++++++++++++++++++------------ 1 file changed, 267 insertions(+), 132 deletions(-) -- 2.20.1

4 19

Re: [alsa-devel] [PATCH] ASoC: tlv320aic32x4: SND_SOC_DAPM_MICBIAS is deprecated
by Mark Brown 09 Jan '19

09 Jan '19

On Wed, Jan 09, 2019 at 10:41:21PM +0530, b-ak wrote: > SND_SOC_DAPM_MICBIAS is deprecated, replace it with SND_SOC_DAPM_SUPPLY. > > MICBIAS voltage wasn't supplied to the microphone with the older > SND_SOC_DAPM_MICBIAS widget, hence the microphone wouldn't work. > > This patch fixes the problem. Note that the reason we've not done that already for everything is that the DAPM graph needs updating for the new routing in machine drivers as the two widgets are wired differently. However this is fine here as there aren't any in tree users I can see so I'll apply, thanks!

1 0

[alsa-devel] [PATCH 0/2] ASoC: da7219: Improvements to DAI clocks handling in driver
by Adam Thomson 09 Jan '19

09 Jan '19

This small patchset provides improvements to CCF based clock handling relating to the DAI clocks. In the first patch MCLK is made parent of DAI clocks, if MCLK has been provided to the driver, which means that MCLK will automatically be enabled as a prerequisite to DAI clocks. The second patch adds the ability to read the WCLK rate provided by the codec when DAI clocks are featured. This is more useful than the returned rate simply being the rate of the parent clock, which in this case would be MCLK. Adam Thomson (2): ASoC: da7219: MCLK should be enabled before DAI clocks ASoC: da7219: Add recalc_rate function to return DAI clock rate sound/soc/codecs/da7219.c | 77 ++++++++++++++++++++++++++++++++++++++++------- sound/soc/codecs/da7219.h | 1 + 2 files changed, 67 insertions(+), 11 deletions(-) -- 1.9.1

2 5

Re: [alsa-devel] hp_pin was NULL value
by Takashi Iwai 09 Jan '19

09 Jan '19

On Wed, 09 Jan 2019 12:19:14 +0100, Kailang wrote: > > Hi Takashi, > > But spec->init_hook will call at resume back. > alc294_hp_init() function just need to call at one time. As already mentioned, even if it has to be called only once and does something for the hardware, it'll be missing if the machine goes hibernate (S4) and resume. In S4, the machine boots from scratch and switches immediately to the saved image that calls resume callback. What's worse for us is that this switch happens usually in initrd, i.e. before the sound driver gets loaded. So, the machine goes to resume without processing alc294_hp_init(). IOW, even if it's needed only once, you'll have to call it at resume unless it has a significant drawback. (Strictly speaking, we may use thaw PM callback, but the HD-audio codec driver doesn't use it, so there is only resume callback for now.) > And it need to call it before call spec->init_hook. But ALC294 (and ALC700) have no init_hook, so far? > ALC225 also need to call similar procedure. I will post it later. OK, thanks. Takashi > > BR, > Kailang > ________________________________________ > 從: Takashi Iwai [tiwai(a)suse.de] > 寄件日期: 2019年1月9日下午 05:56 > 至: Kailang > 副本: (alsa-devel(a)alsa-project.org) > 主旨: Re: hp_pin was NULL value > > On Wed, 09 Jan 2019 10:45:25 +0100, > Kailang wrote: > > > > >>But this made me wonder whether we don't need to call this function at resume as well? Currently it's called only at probing. > > Yes, it only call at probing. Just need to call it at boot time. > > > > -----Original Message----- > > From: Takashi Iwai <tiwai(a)suse.de> > > Sent: Wednesday, January 9, 2019 5:43 PM > > To: Kailang <kailang(a)realtek.com> > > Cc: (alsa-devel(a)alsa-project.org) <alsa-devel(a)alsa-project.org> > > Subject: Re: hp_pin was NULL value > > > > On Wed, 09 Jan 2019 10:31:33 +0100, > > Kailang wrote: > > > > > > Hi Takashi, > > > > > > Could I move the alc294_hp_init(codec) to below line. > > > Because hp_pin = spec->gen.autocfg.hp_pins[0] was null value when alc294_hp_init(codec) at original line. > > > Or move alc269_parse_auto_config() upward. > > > > It looks OK to me. But this made me wonder whether we don't need to call this function at resume as well? Currently it's called only at probing. > > Hrm, but it modifies many COEFs, and are these all preserved with > suspend? For example, there is also hibernation (S4), which is a > switch from the boot without the audio driver initialization. Then > we'll resume from the uninitialized state. I guess other COEF changes > in that area should be moved as the additional init hook as well. > > Actually, if we do call this at resume, the change would be easier, a > patch like below. > > > thanks, > > Takashi > > --- a/sound/pci/hda/patch_realtek.c > +++ b/sound/pci/hda/patch_realtek.c > @@ -7404,6 +7404,20 @@ static void alc294_hp_init(struct hda_codec *codec) > msleep(50); > } > > +static void alc294_init(struct hda_codec *codec) > +{ > + /* UAJ MIC Vref control by verb */ > + alc_update_coef_idx(codec, 0x6b, 0x0018, (1<<4) | (1<<3)); > + alc294_hp_init(codec); > +} > + > +static void alc700_init(struct hda_codec *codec) > +{ > + /* Combo jack auto trigger control */ > + alc_update_coef_idx(codec, 0x4a, 1 << 15, 0); > + alc294_hp_init(codec); > +} > + > /* > */ > static int patch_alc269(struct hda_codec *codec) > @@ -7528,8 +7542,7 @@ static int patch_alc269(struct hda_codec *codec) > case 0x10ec0294: > spec->codec_variant = ALC269_TYPE_ALC294; > spec->gen.mixer_nid = 0; /* ALC2x4 does not have any loopback mixer path */ > - alc_update_coef_idx(codec, 0x6b, 0x0018, (1<<4) | (1<<3)); /* UAJ MIC Vref control by verb */ > - alc294_hp_init(codec); > + spec->init_hook = alc294_init; > break; > case 0x10ec0300: > spec->codec_variant = ALC269_TYPE_ALC300; > @@ -7540,8 +7553,7 @@ static int patch_alc269(struct hda_codec *codec) > case 0x10ec0703: > spec->codec_variant = ALC269_TYPE_ALC700; > spec->gen.mixer_nid = 0; /* ALC700 does not have any loopback mixer path */ > - alc_update_coef_idx(codec, 0x4a, 1 << 15, 0); /* Combo jack auto trigger control */ > - alc294_hp_init(codec); > + spec->init_hook = alc700_init; > break; > > } > > > ------Please consider the environment before printing this e-mail.

1 0

[alsa-devel] pcm_hw bug when calling snd_pcm_sw_params twice
by Jamey Sharp 09 Jan '19

09 Jan '19

I would have reported this to your bug tracker, but I can't find it; the links on alsa-project.org are dead. I've attached a small test program that demonstrates that calling snd_pcm_sw_params twice changes the value of the period_event flag in the sw_params struct, at least on pcm_hw devices. In src/pcm/pcm_hw.c, in snd_pcm_hw_sw_params, there's an early `sw_set_period_event(params, 0)` call. Its effect is not undone if the function returns early. This is easiest to trigger by calling it with unchanged parameters, although there are other early-exit paths with the same problem. I don't actually care about this bug, but since I noticed it while reading the alsa-lib source code trying to figure out what a period event is for, I thought I'd go ahead and report it. Jamey

2 1

[alsa-devel] hp_pin was NULL value
by Kailang 09 Jan '19

09 Jan '19

Hi Takashi, Could I move the alc294_hp_init(codec) to below line. Because hp_pin = spec->gen.autocfg.hp_pins[0] was null value when alc294_hp_init(codec) at original line. Or move alc269_parse_auto_config() upward. BR, Kailang --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -7514,7 +7514,6 @@ static int patch_alc269(struct hda_codec *codec) spec->codec_variant = ALC269_TYPE_ALC294; spec->gen.mixer_nid = 0; /* ALC2x4 does not have any loopback mixer path */ alc_update_coef_idx(codec, 0x6b, 0x0018, (1<<4) | (1<<3)); /* UAJ MIC Vref control by verb */ - alc294_hp_init(codec); break; case 0x10ec0300: spec->codec_variant = ALC269_TYPE_ALC300; @@ -7526,7 +7525,6 @@ static int patch_alc269(struct hda_codec *codec) spec->codec_variant = ALC269_TYPE_ALC700; spec->gen.mixer_nid = 0; /* ALC700 does not have any loopback mixer path */ alc_update_coef_idx(codec, 0x4a, 1 << 15, 0); /* Combo jack auto trigger control */ - alc294_hp_init(codec); break; } @@ -7553,6 +7551,16 @@ static int patch_alc269(struct hda_codec *codec) if (err < 0) goto error; + switch (codec->core.vendor_id) { + case 0x10ec0234: + case 0x10ec0274: + case 0x10ec0294: + case 0x10ec0700: + case 0x10ec0701: + case 0x10ec0703: + alc294_hp_init(codec); + break; + } if (!spec->gen.no_analog && spec->gen.beep_nid && spec->gen.mixer_nid) { err = set_beep_amp(spec, spec->gen.mixer_nid, 0x04, HDA_INPUT); if (err < 0)

2 3

[alsa-devel] [PATCH 1/2] ALSA: hda: Fix mismatches for register mask and value in hdac controller
by Keyon Jie 09 Jan '19

09 Jan '19

E.g. for azx_int_enable(), we should set both mask and value to be "AZX_INT_CTRL_EN | AZX_INT_GLOBAL_EN"(the mask was 0) to enable controller CIE and GIE. We have similar issues on setting AZX_GCTL_RESET and AZX_GCTL_UNSOL, here try to correct all of them. Signed-off-by: Keyon Jie <yang.jie(a)linux.intel.com> --- sound/hda/hdac_controller.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/sound/hda/hdac_controller.c b/sound/hda/hdac_controller.c index 74244d8e2909..b2e9454f5816 100644 --- a/sound/hda/hdac_controller.c +++ b/sound/hda/hdac_controller.c @@ -376,7 +376,7 @@ void snd_hdac_bus_exit_link_reset(struct hdac_bus *bus) { unsigned long timeout; - snd_hdac_chip_updateb(bus, GCTL, 0, AZX_GCTL_RESET); + snd_hdac_chip_updateb(bus, GCTL, AZX_GCTL_RESET, AZX_GCTL_RESET); timeout = jiffies + msecs_to_jiffies(100); while (!snd_hdac_chip_readb(bus, GCTL) && time_before(jiffies, timeout)) @@ -415,7 +415,7 @@ int snd_hdac_bus_reset_link(struct hdac_bus *bus, bool full_reset) } /* Accept unsolicited responses */ - snd_hdac_chip_updatel(bus, GCTL, 0, AZX_GCTL_UNSOL); + snd_hdac_chip_updatel(bus, GCTL, AZX_GCTL_UNSOL, AZX_GCTL_UNSOL); /* detect codecs */ if (!bus->codec_mask) { @@ -431,7 +431,9 @@ EXPORT_SYMBOL_GPL(snd_hdac_bus_reset_link); static void azx_int_enable(struct hdac_bus *bus) { /* enable controller CIE and GIE */ - snd_hdac_chip_updatel(bus, INTCTL, 0, AZX_INT_CTRL_EN | AZX_INT_GLOBAL_EN); + snd_hdac_chip_updatel(bus, INTCTL, + AZX_INT_CTRL_EN | AZX_INT_GLOBAL_EN, + AZX_INT_CTRL_EN | AZX_INT_GLOBAL_EN); } /* disable interrupts */ -- 2.17.1

2 3

Re: [alsa-devel] This will solve jack plug and unplug noise
by Kailang 09 Jan '19

09 Jan '19

Hi Takashi, Update ALC225 depop for jack plug and unplug noise. Add this two patch. It will solve jack plug and unplug noise for ALC225 of Dell headset mode. BR, Kailang

2 1

[alsa-devel] [PATCH -next] ALSA: usb-audio: Remove set but not used variable 'first_ch_bits'
by YueHaibing 09 Jan '19

09 Jan '19

Fixes gcc '-Wunused-but-set-variable' warning: sound/usb/mixer.c: In function 'parse_audio_feature_unit': sound/usb/mixer.c:1838:28: warning: variable 'first_ch_bits' set but not used [-Wunused-but-set-variable] It never used since 2.6 Signed-off-by: YueHaibing <yuehaibing(a)huawei.com> --- sound/usb/mixer.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c index e7d441d..8ad1a24 100644 --- a/sound/usb/mixer.c +++ b/sound/usb/mixer.c @@ -1835,7 +1835,7 @@ static int parse_audio_feature_unit(struct mixer_build *state, int unitid, { int channels, i, j; struct usb_audio_term iterm; - unsigned int master_bits, first_ch_bits; + unsigned int master_bits; int err, csize; struct uac_feature_unit_descriptor *hdr = _ftr; __u8 *bmaControls; @@ -1926,10 +1926,6 @@ static int parse_audio_feature_unit(struct mixer_build *state, int unitid, break; } - if (channels > 0) - first_ch_bits = snd_usb_combine_bytes(bmaControls + csize, csize); - else - first_ch_bits = 0; if (state->mixer->protocol == UAC_VERSION_1) { /* check all control types */

2 1

[alsa-devel] [PATCH 0/2] ASoC: soc-core: add snd_soc_of_lookup_component()
by Rohit kumar 09 Jan '19

09 Jan '19

Platform DAI component probe is not called if it is not present in component list during sound card registration. Add snd_soc_of_lookup_component() to get component info from device_node. This can be used by machine driver to make sure that card components are registered before calling snd_soc_register_card(). Ajit Pandey (2): ASoC: soc-core: add snd_soc_of_lookup_component() ASoC: qcom: defer probe if platform dai is not registered include/sound/soc.h | 1 + sound/soc/qcom/common.c | 9 +++++++++ sound/soc/soc-core.c | 28 ++++++++++++++++++++++++++++ 3 files changed, 38 insertions(+) -- Qualcomm India Private Limited, on behalf of Qualcomm Innovation Center, Inc., is a member of Code Aurora Forum, a Linux Foundation Collaborative Project.

2 4