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

[PATCH] ASoC: wm_adsp: Fix missing mutex_lock in wm_adsp_write_ctl()
by Richard Fitzgerald 07 Mar '24

07 Mar '24

wm_adsp_write_ctl() must hold the pwr_lock mutex when calling cs_dsp_get_ctl(). This was previously partially fixed by commit 781118bc2fc1 ("ASoC: wm_adsp: Fix missing locking in wm_adsp_[read|write]_ctl()") but this only put locking around the call to cs_dsp_coeff_write_ctrl(), missing the call to cs_dsp_get_ctl(). Signed-off-by: Richard Fitzgerald <rf(a)opensource.cirrus.com> Fixes: 781118bc2fc1 ("ASoC: wm_adsp: Fix missing locking in wm_adsp_[read|write]_ctl()") --- sound/soc/codecs/wm_adsp.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/sound/soc/codecs/wm_adsp.c b/sound/soc/codecs/wm_adsp.c index e451c009f2d9..7d5c096e06cd 100644 --- a/sound/soc/codecs/wm_adsp.c +++ b/sound/soc/codecs/wm_adsp.c @@ -683,11 +683,12 @@ static void wm_adsp_control_remove(struct cs_dsp_coeff_ctl *cs_ctl) int wm_adsp_write_ctl(struct wm_adsp *dsp, const char *name, int type, unsigned int alg, void *buf, size_t len) { - struct cs_dsp_coeff_ctl *cs_ctl = cs_dsp_get_ctl(&dsp->cs_dsp, name, type, alg); + struct cs_dsp_coeff_ctl *cs_ctl; struct wm_coeff_ctl *ctl; int ret; mutex_lock(&dsp->cs_dsp.pwr_lock); + cs_ctl = cs_dsp_get_ctl(&dsp->cs_dsp, name, type, alg); ret = cs_dsp_coeff_write_ctrl(cs_ctl, 0, buf, len); mutex_unlock(&dsp->cs_dsp.pwr_lock); -- 2.30.2

2 1

[PATCH v2 0/2] ASoC: codecs: ES8326: change members of private structure
by Zhang Yi 07 Mar '24

07 Mar '24

We found that using 0x45 as the default value for interrupt-clk would cause a headset detection error.So we took 0x00 as the default value for interrupt-clk and passed the test. We removed mic1-src and mic2-src, which were not used. Zhang Yi (2): ASoC: codecs: ES8326: Changing members of private structure ASoC: codecs: ES8326: change support for ES8326 .../bindings/sound/everest,es8326.yaml | 8 ++++---- sound/soc/codecs/es8326.c | 18 +----------------- 2 files changed, 5 insertions(+), 21 deletions(-) -- 2.17.1

2 3

[PATCH 1/2] ALSA: hda: hda_component: Add missing #include guards
by Richard Fitzgerald 07 Mar '24

07 Mar '24

Add the conventional include guards around the content of the hda_component.h header file. This prevents double-declaration of struct hda_component if the header gets included multiple times. This isn't causing any problems with current code, so no need to backport to older kernels. Signed-off-by: Richard Fitzgerald <rf(a)opensource.cirrus.com> --- sound/pci/hda/hda_component.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/sound/pci/hda/hda_component.h b/sound/pci/hda/hda_component.h index deae9dea01b4..a0fcc723483d 100644 --- a/sound/pci/hda/hda_component.h +++ b/sound/pci/hda/hda_component.h @@ -6,6 +6,9 @@ * Cirrus Logic International Semiconductor Ltd. */ +#ifndef __HDA_COMPONENT_H__ +#define __HDA_COMPONENT_H__ + #include <linux/acpi.h> #include <linux/component.h> @@ -82,3 +85,5 @@ static inline void hda_component_manager_unbind(struct hda_codec *cdc, { component_unbind_all(hda_codec_dev(cdc), comps); } + +#endif /* ifndef __HDA_COMPONENT_H__ */ -- 2.30.2

2 2

[PATCH] firmware: cirrus: cs_dsp: Remove non-existent member from kerneldoc
by Richard Fitzgerald 07 Mar '24

07 Mar '24

The kerneldoc for struct cs_dsp refers to a fw_file_name member but there's no such member. Signed-off-by: Richard Fitzgerald <rf(a)opensource.cirrus.com> --- include/linux/firmware/cirrus/cs_dsp.h | 1 - 1 file changed, 1 deletion(-) diff --git a/include/linux/firmware/cirrus/cs_dsp.h b/include/linux/firmware/cirrus/cs_dsp.h index 29cd11d5a3cf..23384a54d575 100644 --- a/include/linux/firmware/cirrus/cs_dsp.h +++ b/include/linux/firmware/cirrus/cs_dsp.h @@ -123,7 +123,6 @@ struct cs_dsp_client_ops; * @sysclk_mask: Mask of frequency bits within sysclk register (ADSP1 only) * @sysclk_shift: Shift of frequency bits within sysclk register (ADSP1 only) * @alg_regions: List of currently loaded algorithm regions - * @fw_file_name: Filename of the current firmware * @fw_name: Name of the current firmware * @fw_id: ID of the current firmware, obtained from the wmfw * @fw_id_version: Version of the firmware, obtained from the wmfw -- 2.30.2

2 1

[PATCH] firmware: cirrus: cs_dsp: Initialize debugfs_root to invalid
by Richard Fitzgerald 07 Mar '24

07 Mar '24

Initialize debugfs_root to -ENODEV so that if the client never sets a valid debugfs root the debugfs files will not be created. A NULL pointer passed to any of the debugfs_create_*() functions means "create in the root of debugfs". It doesn't mean "ignore". Signed-off-by: Richard Fitzgerald <rf(a)opensource.cirrus.com> --- drivers/firmware/cirrus/cs_dsp.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/firmware/cirrus/cs_dsp.c b/drivers/firmware/cirrus/cs_dsp.c index 79d4254d1f9b..9f3d665cfdcf 100644 --- a/drivers/firmware/cirrus/cs_dsp.c +++ b/drivers/firmware/cirrus/cs_dsp.c @@ -522,7 +522,7 @@ void cs_dsp_cleanup_debugfs(struct cs_dsp *dsp) { cs_dsp_debugfs_clear(dsp); debugfs_remove_recursive(dsp->debugfs_root); - dsp->debugfs_root = NULL; + dsp->debugfs_root = ERR_PTR(-ENODEV); } EXPORT_SYMBOL_NS_GPL(cs_dsp_cleanup_debugfs, FW_CS_DSP); #else @@ -2246,6 +2246,11 @@ static int cs_dsp_common_init(struct cs_dsp *dsp) mutex_init(&dsp->pwr_lock); +#ifdef CONFIG_DEBUG_FS + /* Ensure this is invalid if client never provides a debugfs root */ + dsp->debugfs_root = ERR_PTR(-ENODEV); +#endif + return 0; } -- 2.30.2

2 1

[PATCH 00/22] ASoC: Cleanup MediaTek soundcard machine drivers
by AngeloGioacchino Del Regno 07 Mar '24

07 Mar '24

This series performs a cleanup of most of the MediaTek AFE drivers and soundcard machine drivers, reducing code duplication and setting a base to be consistent with their devicetree bindings, as many of those are using different properties and nodes for no good reason. Summarizing: - Commonizes functions and ops where possible - Adds a common probe mechanism, increasing maintainability of soundcard drivers for older MediaTek SoCs - Migrates all drivers to support the new bindings - Obviously retains compatibility with old device trees - Reduces machine-specific parameters hardcoding in drivers - Can now set machine-specific params in device tree - Uses the `audio-routing` and `dai-link` nodes like some other non-MediaTek SoC sound drivers - Imposes consistency between MediaTek ASoC machine soundcard drivers bindings - Reduces code size and greatly reduces the amount of code that will be required for newer drivers (retaining compatibility with the old bindings was costly in terms of code size, otherwise this series would've removed ~1000 more lines, or something along that line). This series was (manually) tested on MT8173, MT8192, MT8195 and MT8186 Chromebooks. AngeloGioacchino Del Regno (22): ASoC: mediatek: mt8192-afe-pcm: Convert to devm_pm_runtime_enable() ASoC: mediatek: mt8192-afe-pcm: Simplify with dev_err_probe() ASoC: mediatek: Commonize ADDA rate transform functions and enums ASoC: mediatek: Assign dummy when codec not specified for a DAI link ASoC: mediatek: Add common machine soundcard driver probe mechanism ASoC: mediatek: common: Constify struct mtk_sof_priv ASoC: mediatek: mt8188: Migrate to mtk_soundcard_common_probe ASoC: mediatek: mt8195: Migrate to mtk_soundcard_common_probe ASoC: mediatek: mt8192: Migrate to mtk_soundcard_common_probe ASoC: mediatek: mt8186: Migrate to mtk_soundcard_common_probe ASoC: mediatek: Add common snd_soc_ops .startup() callback ASoC: mediatek: mt8195: Migrate to the common mtk_soundcard_startup ASoC: mediatek: mt8192: Migrate to the common mtk_soundcard_startup ASoC: mediatek: mt8186-rt1019: Migrate to the common mtk_soundcard_startup ASoC: mediatek: Add common mtk_afe_component_probe callback ASoC: mediatek: Use common mtk_afe_pcm_platform with common probe cb ASoC: mediatek: mt8186: Unify mt8186-mt6366 machine drivers ASoC: dt-bindings: mt8195: Document audio-routing and dai-link subnode ASoC: dt-bindings: mt8192: Document audio-routing and dai-link subnode ASoC: dt-bindings: mt8186: Document audio-routing and dai-link subnode arm64: dts: mediatek: mt8195-cherry: Specify sound DAI links and routing arm64: dts: mediatek: mt8186-corsola: Specify sound DAI links and routing .../sound/mt8186-mt6366-da7219-max98357.yaml | 112 +- .../sound/mt8186-mt6366-rt1019-rt5682s.yaml | 112 +- .../sound/mt8192-mt6359-rt1015-rt5682.yaml | 129 +- .../bindings/sound/mt8195-mt6359.yaml | 123 ++ .../boot/dts/mediatek/mt8186-corsola.dtsi | 42 +- .../boot/dts/mediatek/mt8195-cherry.dtsi | 45 + sound/soc/mediatek/Kconfig | 24 +- sound/soc/mediatek/common/Makefile | 2 + .../mediatek/common/mtk-afe-platform-driver.c | 18 + .../soc/mediatek/common/mtk-dai-adda-common.c | 70 + .../soc/mediatek/common/mtk-dai-adda-common.h | 45 + .../soc/mediatek/common/mtk-dsp-sof-common.c | 15 +- .../soc/mediatek/common/mtk-dsp-sof-common.h | 1 - sound/soc/mediatek/common/mtk-soc-card.h | 7 +- .../mediatek/common/mtk-soundcard-driver.c | 205 ++- .../mediatek/common/mtk-soundcard-driver.h | 42 + sound/soc/mediatek/mt6797/mt6797-afe-pcm.c | 14 +- sound/soc/mediatek/mt6797/mt6797-dai-adda.c | 85 +- sound/soc/mediatek/mt7986/mt7986-afe-pcm.c | 14 +- sound/soc/mediatek/mt8183/mt8183-afe-pcm.c | 14 +- sound/soc/mediatek/mt8183/mt8183-dai-adda.c | 90 +- sound/soc/mediatek/mt8186/Makefile | 3 +- sound/soc/mediatek/mt8186/mt8186-afe-pcm.c | 19 +- sound/soc/mediatek/mt8186/mt8186-dai-adda.c | 94 +- .../mt8186/mt8186-mt6366-da7219-max98357.c | 1189 ----------------- ...t6366-rt1019-rt5682s.c => mt8186-mt6366.c} | 578 ++++---- sound/soc/mediatek/mt8188/mt8188-afe-pcm.c | 21 +- sound/soc/mediatek/mt8188/mt8188-dai-adda.c | 89 +- sound/soc/mediatek/mt8188/mt8188-mt6359.c | 203 +-- sound/soc/mediatek/mt8192/mt8192-afe-pcm.c | 114 +- sound/soc/mediatek/mt8192/mt8192-dai-adda.c | 90 +- .../mt8192/mt8192-mt6359-rt1015-rt5682.c | 301 ++--- sound/soc/mediatek/mt8195/mt8195-afe-pcm.c | 21 +- sound/soc/mediatek/mt8195/mt8195-dai-adda.c | 90 +- sound/soc/mediatek/mt8195/mt8195-mt6359.c | 487 +++---- 35 files changed, 1739 insertions(+), 2769 deletions(-) create mode 100644 sound/soc/mediatek/common/mtk-dai-adda-common.c create mode 100644 sound/soc/mediatek/common/mtk-dai-adda-common.h delete mode 100644 sound/soc/mediatek/mt8186/mt8186-mt6366-da7219-max98357.c rename sound/soc/mediatek/mt8186/{mt8186-mt6366-rt1019-rt5682s.c => mt8186-mt6366.c} (72%) -- 2.44.0

4 31

Strange trigger_htstamp after snd_pcm_drop
by GitHub issues - opened 07 Mar '24

07 Mar '24

alsa-project/alsa-lib issue #387 was opened from z-s-e: I have conceptually the following situation (actual code can be found [here](https://gitlab.com/z-s-e/alpe/-/blob/main/src/alpe_alsa.cpp?ref_type=…): ``` // sound is currently playing snd_pcm_status(pcm, pcm_status); snd_pcm_status_get_htstamp(pcm_status, &ts1); snd_pcm_drop(pcm); snd_pcm_status(pcm, pcm_status); snd_pcm_status_get_trigger_htstamp(status, &ts2); ``` On my hardware I get a `ts2` timestamp that is earlier than `ts1`. From my understanding that should be impossible, as `ts2` should contain the timestamp of when the drop happened, which should obviously be after `ts1`. Is that a bug or do I misunderstand the meaning of those timestamps? Issue URL : https://github.com/alsa-project/alsa-lib/issues/387 Repository URL: https://github.com/alsa-project/alsa-lib

1 0

RE: [EXTERNAL] Re: [PATCH v10] The tas2783 is a smart audio amplifier with integrated MIPI SoundWire interface (Version 1.2.1 compliant), I2C, and I2S/TDM interfaces designed for portable applications. An on-chip DSP supports Texas Instruments SmartAmp sp
by Ding, Shenghao 07 Mar '24

07 Mar '24

> -----Original Message----- > From: Mark Brown <broonie(a)kernel.org> > Sent: Wednesday, March 6, 2024 3:19 AM > To: Ding, Shenghao <shenghao-ding(a)ti.com> > Cc: andriy.shevchenko(a)linux.intel.com; lgirdwood(a)gmail.com; > perex(a)perex.cz; pierre-louis.bossart(a)linux.intel.com; > 13916275206(a)139.com; alsa-devel(a)alsa-project.org; linux- > kernel(a)vger.kernel.org; liam.r.girdwood(a)intel.com; bard.liao(a)intel.com; > mengdong.lin(a)intel.com; yung-chuan.liao(a)linux.intel.com; Xu, Baojun > <baojun.xu(a)ti.com>; Lu, Kevin <kevin-lu(a)ti.com>; tiwai(a)suse.de; > soyer(a)irl.hu; Baojun.Xu(a)fpt.com; Navada Kanyana, Mukund > <navada(a)ti.com> > Subject: [EXTERNAL] Re: [PATCH v10] The tas2783 is a smart audio amplifier > with integrated MIPI SoundWire interface (Version 1.2.1 compliant), I2C, and > I2S/TDM interfaces designed for portable applications. An on-chip DSP > supports Texas Instruments SmartAmp spe... > > On Tue, Mar 05, 2024 at 04:43:35PM +0800, Shenghao Ding wrote: > > The ASoC component provides the majority of the functionality of the > > device, all the audio functions. > > > +static const struct reg_default tas2783_reg_defaults[] = { > > + /* Default values for ROM mode. Activated. */ > > + { 0x8002, 0x1a }, /* AMP inactive. */ > > + { 0x8097, 0xc8 }, > > + { 0x80b5, 0x74 }, > > + { 0x8099, 0x20 }, > > + { 0xfe8d, 0x0d }, > > + { 0xfebe, 0x4a }, > > + { 0x8230, 0x00 }, > > + { 0x8231, 0x00 }, > > + { 0x8232, 0x00 }, > > + { 0x8233, 0x01 }, > > + { 0x8418, 0x00 }, /* Set volume to 0 dB. */ > > + { 0x8419, 0x00 }, > > + { 0x841a, 0x00 }, > > + { 0x841b, 0x00 }, > > + { 0x8428, 0x40 }, /* Unmute channel */ > > + { 0x8429, 0x00 }, > > + { 0x842a, 0x00 }, > > + { 0x842b, 0x00 }, > > + { 0x8548, 0x00 }, /* Set volume to 0 dB. */ > > + { 0x8549, 0x00 }, > > + { 0x854a, 0x00 }, > > + { 0x854b, 0x00 }, > > + { 0x8558, 0x40 }, /* Unmute channel */ > > + { 0x8559, 0x00 }, > > + { 0x855a, 0x00 }, > > + { 0x855b, 0x00 }, > > + { 0x800a, 0x3a }, /* Enable both channel */ > > These comments sound like this register default table is not actually the > physical default register values that the chip has... > > > +static const struct regmap_config tasdevice_regmap = { > > + .reg_bits = 32, > > + .val_bits = 8, > > + .readable_reg = tas2783_readable_register, > > + .volatile_reg = tas2783_volatile_register, > > + .max_register = 0x44ffffff, > > + .reg_defaults = tas2783_reg_defaults, > > + .num_reg_defaults = ARRAY_SIZE(tas2783_reg_defaults), > > ...but this is set as the register defaults. This will cause problems with things > like cache sync where we don't write values out if they're not the defaults. > We also try to keep default settings from the silicon except in the most > obvious cases, it avoids issues with trying to work out what to set and > accomodate different use cases for different systems. > > If you do need to set non-default values then either just regular writes during > probe or a regmap patch would do it. So, can I remove ".reg_defaults = tas2783_reg_defaults," and tas2783_reg_defaults from the code? > > > + .cache_type = REGCACHE_RBTREE, > > + .use_single_read = true, > > + .use_single_write = true, > > REGCACHE_MAPLE is generally the most sensible choice for modern devices > - it's a newer and fancier data structure underlying it and there's only a few > cases with low end devices, mostly doing block I/O which this doesn't > support, where the RBTREE cache is still better. Accept > > > + u16 dev_info; > > + int ret; > > + > > + if (!tas_dev->sdw_peripheral) { > > + dev_err(tas_dev->dev, "%s: peripheral doesn't exist.\n", > > + __func__); > > + return; > > + } > > + > > + dev_info = tas_dev->sdw_peripheral->bus->link_id | > > + tas_dev->sdw_peripheral->id.unique_id << 16; > > I'm kind of surprised dev_info works as a variable name without something > getting upset that it aliases the function of the same name. Accept

2 1

[PATCH 0/4] ASoC: fsl: Support register and unregister rpmsg sound card through remoteproc
by Chancel Liu 07 Mar '24

07 Mar '24

echo /lib/firmware/fw.elf > /sys/class/remoteproc/remoteproc0/firmware (A) echo start > /sys/class/remoteproc/remoteproc0/state (B) echo stop > /sys/class/remoteproc/remoteproc0/state The rpmsg sound card is registered in (A) and unregistered in (B). After "start", imx-audio-rpmsg registers devices for ASoC platform driver and machine driver. Then sound card is registered. After "stop", imx-audio-rpmsg unregisters devices for ASoC platform driver and machine driver. Then sound card is unregistered. Chancel Liu (4): ASoC: fsl: imx_pcm_rpmsg: Register component with rpmsg channel name ASoC: fsl: imx-audio-rpmsg: Register device with rpmsg channel name ASoC: fsl: Let imx-audio-rpmsg register platform device for card ASoC: fsl: imx-rpmsg: Update to correct DT node sound/soc/fsl/fsl_rpmsg.c | 11 ----------- sound/soc/fsl/imx-audio-rpmsg.c | 22 +++++++++++++++++++--- sound/soc/fsl/imx-pcm-rpmsg.c | 11 ++++++++--- sound/soc/fsl/imx-rpmsg.c | 21 ++++++++++++++++++--- 4 files changed, 45 insertions(+), 20 deletions(-) -- 2.43.0

2 6

[PATCH v1 2/2] ASoC: codecs: ES8326: change support for ES8326
by Zhang Yi 07 Mar '24

07 Mar '24

> > On Wed, Mar 06, 2024 at 09:34:14AM +0800, Zhang Yi wrote: > > > Removed mic1-src and mic2-src. and changed default value > > > of interrupt-clk > > > > We could do with a better changelog here. > > > > > - everest,mic1-src: > > > - $ref: /schemas/types.yaml#/definitions/uint8 > > > - description: > > > - the value of reg 2A when headset plugged. > > > - minimum: 0x00 > > > - maximum: 0x77 > > > - default: 0x22 > > > - > > > - everest,mic2-src: > > > - $ref: /schemas/types.yaml#/definitions/uint8 > > > - description: > > > - the value of reg 2A when headset unplugged. > > > - minimum: 0x00 > > > - maximum: 0x77 > > > - default: 0x44 > > > > This will make any DTs using the properties instantly buggy. I believe > > there's a way of marking properties as deprecated, you should use that > > instead. Or we could just leave the properties there and ignore them at > > runtime. > 'deprecated: true' is the way. OK.I will modify the properties, thank you for your support

1 0