ASoC: tegra20: i2s: Filter out unsupported rates

Support new nvidia,fixed-parent-rate device-tree property which instructs
I2S that board wants parent clock rate to stay at a fixed rate. This allows
to play audio over S/PDIF and I2S simultaneously. The root of the problem
is that audio components on Tegra share the same audio PLL, and thus, only
a subset of rates can be supported if we want to play audio simultaneously.
Filter out audio rates that don't match parent clock rate if device-tree
has the nvidia,fixed-parent-rate property.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-14-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Filter out unsupported rates

SPDIF and other SoC components share audio PLL on Tegra, thus only one
component may set the desired base clock rate. This creates problem for
HDMI audio because it uses SPDIF and audio may not work if SPDIF's clock
doesn't exactly match standard audio rate since some receivers may reject
audio in that case. Filter out audio rates which SPDIF output can't
support, assuming that other components won't change rate at runtime.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-13-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Support system suspend

Support system suspend by enforcing runtime PM suspend/resume.
Now there is no doubt that h/w is indeed stopped during suspend
and that h/w state will be properly restored after resume.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-12-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Reset hardware

Reset S/PDIF controller on runtime PM suspend/resume to ensure that we
always have a consistent hardware state.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-11-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Use more resource-managed helpers

Use resource-managed helpers to make code cleaner. Driver's remove callback
isn't needed anymore since driver is completely resource-managed now.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-10-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Improve driver's code

- Clean up whitespaces, defines and variables.

- Remove obsolete code.

- Adhere to upstream coding style.

- Don't override returned error code.

- Replace pr_err with dev_err.

No functional changes are made by this patch. This is a minor code's
refactoring that will ease further maintenance of the driver.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-9-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Support device-tree

Tegra20 S/PDIF driver was added in a pre-DT era and was never used since
that time. Revive driver by adding device-tree support.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-8-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra20: spdif: Set FIFO trigger level

FIFO trigger level must be not less than the size of DMA burst, otherwise
audio will be played x4 faster that it should be because part of the DMA
data will be dropped on FIFO input buffer overflow.

Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-6-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: tegra20-i2s: Document new nvidia,fixed-parent-rate property

Document new nvidia,fixed-parent-rate property which instructs that this
board wants parent clock to stay at a fixed rate. It allows to prevent
conflicts between audio components that share same parent PLL. For
instance, this property allows to have HDMI audio, speaker and headphones
in the system playing audio simultaneously, which is a common pattern for
consumer devices.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Link: https://lore.kernel.org/r/20211204143725.31646-4-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: tegra20-i2s: Convert to schema

Convert NVIDIA Tegra20 I2S binding to schema.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-3-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: Add binding for Tegra20 S/PDIF

Add device-tree binding for Tegra20 S/PDIF controller.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Thierry Reding <treding@nvidia.com>
Link: https://lore.kernel.org/r/20211204143725.31646-2-digetx@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

Merge tag 'dmaengine_topic_slave_id_removal_5.17' of git://git./linux/kernel/git/vkoul/dmaengine into v4_20211204_digetx_support_hdmi_audio_on_nvidia_tegra20

dmaengine_topic_slave_id_removal_5.17

Tag for dmaengine slave_id removal topic branch which should be merged
into v5.17

dmaengine: remove slave_id config field

All references to the slave_id field have been removed, so remove the
field as well to prevent new references from creeping in again.

Originally this allowed slave DMA drivers to configure which device
is accessed with the dmaengine_slave_config() call, but this was
inconsistent, as the same information is also passed while requesting
a channel, and never changes in practice.

In modern kernels, the device is always selected when requesting
the channel, so the .slave_id field is no longer useful.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-12-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dmaengine: xilinx_dpdma: stop using slave_id field

The display driver wants to pass a custom flag to the DMA engine driver,
which it started doing by using the slave_id field that was traditionally
used for a different purpose.

As there is no longer a correct use for the slave_id field, it should
really be removed, and the remaining users changed over to something
different.

The new mechanism for passing nonstandard settings is using the
.peripheral_config field, so use that to pass a newly defined structure
here, making it clear that this will not work in portable drivers.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-10-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dmaengine: qcom-adm: stop abusing slave_id config

The slave_id was previously used to pick one DMA slave instead of another,
but this is now done through the DMA descriptors in device tree.

For the qcom_adm driver, the configuration is documented in the DT
binding to contain a tuple of device identifier and a "crci" field,
but the implementation ends up using only a single cell for identifying
the slave, with the crci getting passed in nonstandard properties of
the device, and passed through the dma driver using the old slave_id
field. Part of the problem apparently is that the nand driver ends up
using only a single DMA request ID, but requires distinct values for
"crci" depending on the type of transfer.

Change both the dmaengine driver and the two slave drivers to allow
the documented binding to work in addition to the ad-hoc passing
of crci values. In order to no longer abuse the slave_id field, pass
the data using the "peripheral_config" mechanism instead.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-9-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dmaengine: sprd: stop referencing config->slave_id

It appears that the code that reads the slave_id from the channel config
was copied incorrectly from other drivers. Nothing ever sets this field
on platforms that use this driver, so remove the reference.

Reviewed-by: Baolin Wang <baolin.wang7@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-8-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dmaengine: pxa/mmp: stop referencing config->slave_id

The last driver referencing the slave_id on Marvell PXA and MMP platforms
was the SPI driver, but this stopped doing so a long time ago, so the
TODO from the earlier patch can no be removed.

Fixes: b729bf34535e ("spi/pxa2xx: Don't use slave_id of dma_slave_config")
Fixes: 13b3006b8ebd ("dma: mmp_pdma: add filter function")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-7-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dmaengine: shdma: remove legacy slave_id parsing

The slave device is picked through either devicetree or a filter
function, and any remaining out-of-tree drivers would have warned
about this usage since 2015.

Stop interpreting the field finally so it can be removed from
the interface.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-6-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

mmc: bcm2835: stop setting chan_config->slave_id

The field is not interpreted by the DMA engine driver, as all the data
is passed from devicetree instead. Remove the assignment so the field
can eventually be deleted.

Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-5-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

spi: pic32: stop setting dma_config->slave_id

Setting slave_id makes no sense with DT based probing, and
should eventually get removed entirely. Address this driver
by no longer setting the field here.

I could not find which DMA driver is used on PIC32, if it's
in the tree at all, but none of the obvious ones even care
about slave_id any more.

Acked-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20211122222203.4103644-4-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

ASoC: dai_dma: remove slave_id field

This field is no longer set from any driver now, so remove the
last references as well.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-3-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

dmaengine: tegra20-apb: stop checking config->slave_id

Nothing sets the slave_id field any more, so stop accessing
it to allow the removal of this field.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-11-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

ASoC: tegra20-spdif: stop setting slave_id

The DMA resource is never set up anywhere, and passing this as slave_id
has not been the proper procedure in a long time.

As a preparation for removing all slave_id references from the ALSA code,
remove this one.

According to Dmitry Osipenko, this driver has never been used and
the mechanism for configuring DMA would not work as it is implemented,
so this part will get rewritten when the driver gets put into use
again in the future.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20211122222203.4103644-2-arnd@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>

 ASoC: Changes to SOF kcontrol data set/get ops

Merge series from Ranjani Sridharan <ranjani.sridharan@linux.intel.com>:

This set of patches deals with modifications to the signature of kcontrol
get/set data functions to make them more intuitive. The last patch deals
with initializing the binary control data size after boot up.

ASoC: AMD: fix depend/select mistake on SND_AMD_ACP_CONFIG

on i386 or x86_64:

when # CONFIG_ACPI is not set,
so SND_SOC_ACPI is not set:

WARNING: unmet direct dependencies detected for SND_AMD_ACP_CONFIG
  Depends on [n]: SOUND [=y] && !UML && SND [=y] && SND_SOC [=y] &&
  SND_SOC_ACPI [=n]
  Selected by [y]:
  - SND_SOC_AMD_ACP_COMMON [=y] && SOUND [=y] && !UML && SND [=y] &&
  SND_SOC [=y] && X86 [=y] && PCI [=y]

This problem is due to the unconditional selection of
SND_AMD_ACP_CONFIG in other options. Using 'depends on' solved an
initial problem but exposed another, let's use select instead.

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Fixes: d9b994cd7641 ('ASoC: AMD: acp-config: fix missing dependency on SND_SOC_ACPI')
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@intel.com>
Reviewed-by: Ajit Kumar Pandey <AjitKumar.Pandey@amd.com>
Link: https://lore.kernel.org/r/20211215163511.151286-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: nvidia,tegra-audio: Convert multiple txt bindings to yaml

Convert Tegra audio complex with the
  * ALC5632
  * MAX98090
  * RT5640
  * RT5677
  * SGTL5000
  * TrimSlice
  * WM8753
  * WM8903
  * WM9712
codec to the YAML format.

Additional changes:
 - added missing HPOUTL to the WM9712 codec.
 - extended rt5677 codec with multiple pins

Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: David Heidelberg <david@ixit.cz>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211211224946.79875-1-david@ixit.cz
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: topology: read back control data from DSP

Read back the control data from the DSP to initialize the control data
size to match that of the data in the DSP. This is particularly useful
for volatile read-only kcontrols in static pipelines.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-9-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Drop ctrl_type parameter for snd_sof_ipc_set_get_comp_data()

The SOF_CTRL_TYPE_VALUE_COMP_* type is not used by the firmware nor in the
kernel side.
It is also not clear what action should be taken for such type.

With this in mind:
The correct ipc_cmd can be selected based on the `ctrl_cmd` and the `set`
parameters:
if the ctrl_cmd is SOF_CTRL_CMD_BINARY then SOF_CTRL_TYPE_DATA_*
otherwise SOF_CTRL_TYPE_VALUE_CHAN_*.

The SET or GET direction can be selected with the use of `set` parameter.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-8-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: control: Do not handle control notification with component type

The component type is not used in firmware nor in the kernel currently and
it is not even clear how it should be handled.

Do not even try to handle it to avoid errors.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-7-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: sof-audio: Drop the `cmd` member from struct snd_sof_control

There is no need to use two variables to store and check the same
information, the scontrol->cmd is the same as scontrol->control_data->cmd.

Drop the former one and when it is needed, access the cmd from the
control_data.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-6-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Drop ctrl_cmd parameter for snd_sof_ipc_set_get_comp_data()

The scontrol->control_data->cmd has been configured during initialization
to the correct sof_ipc_ctrl_cmd.

No need to pass duplicated information, let's use the already available
one via scontrol.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-5-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: topology: Set control_data->cmd alongside scontrol->cmd

Set the scontrol->control_data->cmd early to the same as scontrol->cmd.

This is a preparatory patch to remove the ctrl_cmd parameter for the
snd_sof_ipc_set_get_comp_data() function.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-4-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Drop ipc_cmd parameter for snd_sof_ipc_set_get_comp_data()

The correct ipc_cmd can be selected based on the `ctrl_cmd` and the `set`
parameters:
if the ctrl_cmd is SOF_CTRL_CMD_BINARY then SOF_IPC_COMP_*_DATA
otherwise SOF_IPC_COMP_*_VALUE.

The SET or GET direction can be selected with the use of `set` parameter.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-3-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: ipc: Rename send parameter in snd_sof_ipc_set_get_comp_data()

Rename the send parameter to set in snd_sof_ipc_set_get_comp_data() and
sof_set_get_large_ctrl_data() to be more aligned with the function name.

No functional change.

Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211215180404.53254-2-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: rt5663: Handle device_property_read_u32_array error codes

The return value of device_property_read_u32_array() is not always 0.
To catch the exception in case that devm_kzalloc failed and the
rt5663->imp_table was NULL, which caused the failure of
device_property_read_u32_array.

Fixes: 450f0f6a8fb4 ("ASoC: rt5663: Add the manual offset field to compensate the DC offset")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Link: https://lore.kernel.org/r/20211215031550.70702-1-jiasheng@iscas.ac.cn
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: OF: Avoid reverse module dependency

Similar with commit 8a49cd11e68ed0 ("ASoC: SOF: ACPI: avoid reverse
module dependency") we will be having hardware specific drivers that
link against a common "helper" framework.

sof-of-dev.c becomes a library with the interface defined in the newly
created file sof-of-dev.h.

This is the final step started with Kconfig simplification in
commit 7548a391c53ca ("ASoC: SOF: i.MX: simplify Kconfig")

Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
Reviewed-by: Paul Olaru <paul.olaru@oss.nxp.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211215085703.137414-1-daniel.baluta@oss.nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC : soc-pcm: fix trigger race conditions with shared BE

Merge series from Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>:

We've been adding a 'deep buffer' PCM device to several SOF topologies
in order to reduce power consumption. The typical use-case would be
music playback over a headset: this additional PCM device provides
more buffering and longer latencies, leaving the rest of the system
sleep for longer periods. Notifications and 'regular' low-latency
audio playback would still use the 'normal' PCM device and be mixed
with the 'deep buffer' before rendering on the headphone endpoint. The
tentative direction would be to expose this alternate device to
PulseAudio/PipeWire/CRAS via the UCM SectionModifier definitions.

That seemed a straightforward topology change until our automated
validation stress tests started reporting issues on SoundWire
platforms, when e.g. two START triggers might be send and conversely
the STOP trigger is never sent. The SoundWire stream state management
flagged inconsistent states when the two 'normal' and 'deep buffer'
devices are used concurrently with rapid play/stop/pause monkey
testing.

Looking at the soc-pcm.c code, it seems that the BE state
management needs a lot of love.

a) there is no consistent protection for the BE state. In some parts
of the code, the state updates are protected by a spinlock but in the
trigger they are not. When we open/play/close the two PCM devices in
stress tests, we end-up testing a state that is being modified. That
can't be good.

b) there is a conceptual deadlock: on stop we check the FE states to
see if a shared BE can be stopped, but since we trigger the BE first
the FE states have not been modified yet, so the TRIGGER_STOP is never
sent.

This patchset suggests the removal of the dedicated 'dpcm_lock' and
follows the design suggested by Takashi Iwai.  By default the
protection relies on the 'pcm_mutex', except for the FE and BE
triggers where the mutex cannot be used.  In this case, the FE PCM
lock is used instead. In the cases where a BE is added/removed, the
pcm_mutex and FE PCM lock are both taken.  In addition, the BE PCM
lock is used to serialize access to a shared BE.

With these patches I am able to run our entire validation suite
without any issues with this new 'deep buffer' topology, and no
regressions on existing solutions [1]. The tests were reproduced by
Bard Liao for SoundWire devices.

One might ask 'how come we didn't see this earlier'? The answer is
probably that the .trigger callbacks in most implementations seems to
perform DAPM operations, and sending the triggers multiple times is
not an issue. In the case of SoundWire, we do use the .trigger
callback to reconfigure the bus using the 'bank switch' mechanism. It
could be acceptable to tolerate a trigger multiple times, but the
deadlock on stop cannot be fixed at the SoundWire level alone.

Opens:

1) The issues reported by Nvidia on the RFCv3 may or may not be
present. We'd need test results to make sure the locking update does
not introduce a regression on Tegra.

2) There are other reports of kernel oopses [2] that seem related to
the lack of protection. I'd be good to confirm if this patchset solve
these problems as well.

[1] https://github.com/thesofproject/linux/pull/3146
[2] https://lore.kernel.org/alsa-devel/002f01d7b4f5$c030f4a0$4092dde0$@samsung.com/

changes since RFCv3:
Used two patches from Takashi. We now use the pcm_mutex, the FE stream
lock when adding and deleting a BE, and the BE stream lock to handle
concurrency between streams using the same BE.
Added a patch to use GFP_ATOMIC for the DPCM structure.
Fixed PAUSE_RELEASE transition (GitHub comment from Kai Vehmanen)

changes since RFCv2:
Removal of dpcm_lock to use FE PCM locks (credits to Takashi Iwai for
the suggestion). The FE PCM lock is now used before each use of
for_each_dpcm_be() - with the exception of the trigger where the lock
is already taken. This change is also applied in drivers which make
use of this loop (compress, SH, FSL).
Addition of BE PCM lock to deal with mutual exclusion between triggers
for the same BE.
Alignment of the BE atomicity on the FE on connections, this is
required to avoid sleeping in atomic context.
Additional cleanups (indentation, static functions)

changes since RFC v1:
Removed unused function
Removed exported symbols only used in soc-pcm.c, used static instead
Use a mutex instead of a spinlock
Protect all for_each_dpcm_be() loops
Fix bugs introduced in the refcount

Pierre-Louis Bossart (4):
  ASoC: soc-pcm: use GFP_ATOMIC for dpcm structure
  ASoC: soc-pcm: align BE 'atomicity' with that of the FE
  ASoC: soc-pcm: test refcount before triggering
  ASoC: soc-pcm: fix BE handling of PAUSE_RELEASE

Takashi Iwai (2):
  ASoC: soc-pcm: Fix and cleanup DPCM locking
  ASoC: soc-pcm: serialize BE triggers

 include/sound/soc-dpcm.h |   2 +
 include/sound/soc.h      |   2 -
 sound/soc/soc-core.c     |   1 -
 sound/soc/soc-pcm.c      | 351 +++++++++++++++++++++++++++------------
 4 files changed, 246 insertions(+), 110 deletions(-)

--
2.25.1

ASoC: qdsp6: Fix an IS_ERR() vs NULL bug

The function gpr_alloc_port return ERR_PTR on errors, it doesn't return
null.

Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Link: https://lore.kernel.org/r/20211211065840.1221-1-linmq006@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: qcom: Distinguish headset codec by codec_dai->name

Distinguish which headset codec is on the board by codec_dai->name
instead of card->name.

It fixes the crash of being unable to handle kernel paging requests
at virtual address ADDR by initializing the correct audio codec on
the board.

Call stack of the crash:
```
 Unable to handle kernel paging request at virtual address ...
 ...
 Call trace:
  rt5682_set_component_pll+0xcc/0xb78 [snd_soc_rt5682]
  snd_soc_component_set_pll+0x90/0x154
  snd_soc_dai_set_pll+0xf4/0x1ac
  sc7180_snd_startup+0x268/0x3c0 [snd_soc_sc7180]
  snd_soc_link_startup+0xa4/0x180
  soc_pcm_open+0x35c/0x15c8
  snd_pcm_open_substream+0xa90/0x13b0
  snd_pcm_open+0x1a4/0x55c
  snd_pcm_capture_open+0x7c/0xe8
  snd_open+0x2b8/0x2e4
  chrdev_open+0x364/0x3d4
  do_dentry_open+0x66c/0xc58
  vfs_open+0x7c/0x8c
  path_openat+0x108c/0x2bbc
  do_filp_open+0x15c/0x258
  do_sys_open+0x278/0x62c
  __arm64_compat_sys_openat+0x9c/0xb0
  ...
```

Fixes: 425c5fce8a03 ("ASoC: qcom: Add support for ALC5682I-VS codec")
Signed-off-by: Judy Hsiao <judyhsiao@chromium.org>
Link: https://lore.kernel.org/r/20211214022509.1288245-1-judyhsiao@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: soc-pcm: fix BE handling of PAUSE_RELEASE

A BE connected to more than one FE, e.g. in a mixer case, can go
through the following transitions.

play FE1    -> BE state is START
pause FE1   -> BE state is PAUSED
play FE2    -> BE state is START
stop FE2    -> BE state is STOP (see note [1] below)
release FE1 -> BE state is START
stop FE1    -> BE state is STOP

play FE1    -> BE state is START
pause FE1   -> BE state is PAUSED
play FE2    -> BE state is START
release FE1 -> BE state is START
stop FE2    -> BE state is START
stop FE1    -> BE state is STOP

play FE1    -> BE state is START
play FE2    -> BE state is START (no change)
pause FE1   -> BE state is START (no change)
pause FE2   -> BE state is PAUSED
release FE1 -> BE state is START
release FE2 -> BE state is START (no change)
stop FE1    -> BE state is START (no change)
stop FE2    -> BE state is STOP

The existing code for PAUSE_RELEASE only allows for the case where the
BE is paused, which clearly would not work in the sequences above.

Extend the allowed states to restart the BE when PAUSE_RELEASE is
received, and increase the refcount if the BE is already in START.

[1] the existing logic does not move the BE state back to PAUSED when
the FE2 is stopped. This patch does not change the logic; it would be
painful to keep a history of changes on the FE side, the state machine
is already rather complicated with transitions based on the last BE
state and the trigger type.

Reported-by: Bard Liao <bard.liao@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207173745.15850-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: soc-pcm: test refcount before triggering

On start/pause_release/resume, when more than one FE is connected to
the same BE, it's possible that the trigger is sent more than
once. This is not desirable, we only want to trigger a BE once, which
is straightforward to implement with a refcount.

For stop/pause/suspend, the problem is more complicated: the check
implemented in snd_soc_dpcm_can_be_free_stop() may fail due to a
conceptual deadlock when we trigger the BE before the FE. In this
case, the FE states have not yet changed, so there are corner cases
where the TRIGGER_STOP is never sent - the dual case of start where
multiple triggers might be sent.

This patch suggests an unconditional trigger in all cases, without
checking the FE states, using a refcount protected by the BE PCM
stream lock.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207173745.15850-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: soc-pcm: serialize BE triggers

When more than one FE is connected to a BE, e.g. in a mixing use case,
the BE can be triggered multiple times when the FE are opened/started
concurrently. This race condition is problematic in the case of
SoundWire BE dailinks, and this is not desirable in a general
case.

This patch relies on the existing BE PCM lock, which takes atomicity into
account. The locking model assumes that all interactions start with
the FE, so that there is no deadlock between FE and BE locks.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
[test, checkpatch fix and clarification of commit message by plbossart]
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207173745.15850-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: soc-pcm: Fix and cleanup DPCM locking

The existing locking for DPCM has several issues
a) a confusing mix of card->mutex and card->pcm_mutex.
b) a dpcm_lock spinlock added inconsistently and on paths that could
be recursively taken. The use of irqsave/irqrestore was also overkill.

The suggested model is:

1) The pcm_mutex is the top-most protection of BE links in the FE. The
pcm_mutex is applied always on either the top PCM callbacks or the
external call from DAPM, not taken in the internal functions.

2) the FE stream lock is taken in higher levels before invoking
dpcm_be_dai_trigger()

3) when adding and deleting a BE, both the pcm_mutex and FE stream
lock are taken.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
[clarification of commit message by plbossart]
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207173745.15850-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: soc-pcm: align BE 'atomicity' with that of the FE

Since the flow for DPCM is based on taking a lock for the FE first, we
need to make sure during the connection between a BE and an FE that
they both use the same 'atomicity', otherwise we may sleep in atomic
context.

If the FE is nonatomic, this patch forces the BE to be nonatomic as
well. That should have no negative impact since the BE 'inherits' the
FE properties.

However, if the FE is atomic and the BE is not, then the configuration
is flagged as invalid.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
[ removed FE stream lock by tiwai ]
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207173745.15850-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: soc-pcm: use GFP_ATOMIC for dpcm structure

We allocate a structure in dpcm_be_connect(), which may be called in
atomic context. Using GFP_KERNEL is not quite right, we have to use
GFP_ATOMIC to prevent the allocator from sleeping.

Suggested-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207173745.15850-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: qcom: apq8016_sbc: Allow routing audio through QDSP6

Merge series from Stephan Gerhold <stephan@gerhold.net>:

This series makes it possible to route audio through the combined
audio/modem DSP on MSM8916/APQ8016 devices instead of bypassing it using
the LPASS drivers. This is necessary to support certain functionality such
as voice call audio. See PATCH 4/5 for details.

Also, qcom,apq8016-sbc.txt is converted to DT schema by adding it to the
existing qcom,sm8250.yaml. The bindings are similar enough that it is easier
to share a single schema instead of duplicating everything into multiple ones.

ASoC: mediatek: assign correct type to argument

Fix the following sparse warning: (new ones prefixed by >>)
>> sound/soc/mediatek/mt8192/mt8192-mt6359-rt1015-rt5682.c:370:33:
     sparse: sparse: incorrect type in argument 3 (different base types)
   sound/soc/mediatek/mt8192/mt8192-mt6359-rt1015-rt5682.c:370:33: sparse:
     expected unsigned int to
   sound/soc/mediatek/mt8192/mt8192-mt6359-rt1015-rt5682.c:370:33: sparse:
     got restricted snd_pcm_format_t [usertype]

Correct discription of format, use S32_LE and S24_LE to distinguish the
different 32bit.

Signed-off-by: Jiaxin Yu <jiaxin.yu@mediatek.com>
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Tzung-Bi Shih <tzungbi@google.com>
Link: https://lore.kernel.org/r/20211209073224.21793-1-jiaxin.yu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: sof-probes: Constify sof_probe_compr_ops

The only usage of sof_probe_compr_ops is to assign its address to the
cops field in the snd_soc_dai_driver struct (in
sound/soc/sof/intel/hda-dai.c). Make it const to allow the compiler to
put it in read-only memory.

Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Link: https://lore.kernel.org/r/20211204203220.54712-1-rikard.falkeborn@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Remove pm_runtime_put_autosuspend() for SOF OF device

In SOF OF device, pm_runtime_put_autosuspend() is not
matching any pm_runtime_get_sync().

This is imbalanced for PM runtime.

Also, for consistency we call pm_runtime_mark_last_busy()
before enabling PM runtime.

1. Remove pm_runtime_put_autosuspend() in probe_complete
2. Reorder PM runtime calls int probe_complete

Signed-off-by: Allen-KH Cheng <Allen-KH.Cheng@mediatek.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Reviewed-by: Paul Olaru <paul.olaru@oss.nxp.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
Link: https://lore.kernel.org/r/20211209200830.145005-1-daniel.baluta@oss.nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: qcom: apq8016_sbc: Allow routing audio through QDSP6

The apq8016-sbc-sndcard is designed to be used with the LPASS drivers
(bypassing the combined audio/modem DSP in MSM8916/APQ8016).
Make it possible to use QDSP6 audio instead for the msm8916-qdsp6-sndcard.

This only requires adding some additional hooks that set up the DPCM
backends correctly. Similar code is already used in drivers for newer
SoCs such as apq8096.c, sdm845.c and sm8250.c.

A slightly different initialization sequence is used for the apq8016-sbc
and msm8916-qdsp6 sound card by defining the apq8016_sbc_add_ops()
function as device match data.

Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20211202145505.58852-6-stephan@gerhold.net
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: qcom: Document qcom,msm8916-qdsp6-sndcard compatible

There are two possible audio setups on MSM8916/APQ8016: Normally the audio
is routed through the audio/modem DSP (covered by the qdsp6 driver). During
upstreaming for the DragonBoard 410c it was decided to bypass it and
instead talk directly to the audio controller using the "lpass" driver.

Bypassing the DSP gives more control about the audio configuration but limits
the functionality: For example, routing audio through the audio/modem DSP is
strictly required for voice call audio. Also, without the special changes in
the DB410c firmware other MSM8916 devices can only use the bypass as long as
the modem DSP is not started. Otherwise, the firmware will assume control of
the LPASS hardware block and audio is no longer functional.

Add support for using the DSP audio setup instead using a new
"qcom,msm8916-qdsp6-sndcard" compatible. It is basically a mixture of
the apq8016-sbc-sndcard and the newer sm8250-sndcard, which uses
indirect QDSP6 DAI links instead of the direct LPASS DAI links.

Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211202145505.58852-5-stephan@gerhold.net
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: qcom: apq8016-sbc: Move to qcom,sm8250 DT schema

All the Qualcomm sound card drivers use the same common device tree
parsing code, so the allowed device tree nodes are almost the same
for all of them. Convert the qcom,apq8016-sbc-sndcard documentation
to a DT schema by adding it to the existing qcom,sm8250 schema.

The only speciality of qcom,apq8016-sbc-sndcard is that it has memory
resources for setting up an I/O mux. This can be handled using
a conditional if statement that only requires it for the apq8016-sbc
compatible.

Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211202145505.58852-4-stephan@gerhold.net
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: qcom: sm8250: Document "aux-devs"

The sm8250 audio driver uses the common Qualcomm device tree parser and
therefore already supports the "aux-devs" property that allows adding
additional auxiliary devices to the sound card (e.g. analog speaker
amplifiers that can be connected using "audio-routing").

Document the property in the DT schema for sm8250 as well. The description
is taken from simple-card.yaml which has a very similar property.

Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211202145505.58852-3-stephan@gerhold.net
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: qcom: sm8250: Drop redundant MultiMedia routes

The MultiMedia audio routes can be deduced from other parts of the
device tree (e.g. the definitions of the MultiMedia DAIs) and therefore
specifying them again in "audio-routing" is redundant and prone to
mistakes. This is no longer necessary since commit 6fd8d2d275f7
("ASoC: qcom: qdsp6: Move frontend AIFs to q6asm-dai").

Let's drop them from the example in the DT schema as well
to avoid confusion.

Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211202145505.58852-2-stephan@gerhold.net
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: add comment on JasperLake support

Explain why JasperLake is exposed in cnl.c instead of icl.c
No functionality change.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-8-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: hda-dai: remove unused fields

The existing code does not use the 'host_dma_id', 'link_dma_id',
'host_bps' fields remove them.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: hda: add quirks for HDAudio DMA position information

The code inherited from the Skylake driver does not seem to follow any
known hardware recommendations.

The only two recommended options are
a) use DPIB registers if VC1 traffic is not allowed
b) use DPIB DDR update if VC1 traffic is used

In all of SOF-based updated, VC1 is not supported so we can 'safely'
move to using DPIB registers only.

This patch keeps the legacy code, in case there was an undocumented
issue lost to history, and adds the DPIB DDR update for additional
debug.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: hda-stream: only enable DPIB if needed

The existing code is inconsistent, we should only enable DPIB if the
'use_posbuf' field is true.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: hda-ctrl: apply symmetry for DPIB

we use 'bus->use_posbuf && bus->posbuf.addr' in
hda_dsp_ctrl_init_chip(), use the same for hda_dsp_ctrl_stop_chip()

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: hda-stream: limit PROCEN workaround

The work-around enabled in hda-stream.c is only required on earlier
versions of SOCs/PCH (Skylake, KabyLake, ApolloLake,
GeminiLake). Before setting the format on the host DMA, it is required
to couple the host and link DMA - which as a consequence shall use the
same format.

This patch introduces a quirk field in the platform descriptor and
makes the work-around conditional. Newer platforms have
no limitations on the use of host and link DMA, which can use
different formats.

Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: ICL: move ICL-specific ops to icl.c

Move the ICL specific ops to icl.c. Also introduce a
macro ICL_DSP_HPRO_CORE_ID to define the core that
should be powered up when HPRO is enabled.

Reviewed-by: Bard Liao <bard.liao@intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211207193947.71080-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: test-component: fix null pointer dereference.

Dereferncing of_id pointer will result in exception in current
implementation since of_match_device() will assign it to NULL.
Adding NULL check for protection.

Signed-off-by: Ameer Hamza <amhamza.mgc@gmail.com>
Link: https://lore.kernel.org/r/20211207142309.222820-1-amhamza.mgc@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: tegra: Document interconnects property

Add the interconnects and interconnect-names properties to the bindings
for the sound card on various NVIDIA Tegra based boards. These are used
to describe the device's memory paths to and from memory.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211206154624.229018-1-thierry.reding@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: amd: Convert to new style DAI format definitions

Convert the AMD machine drivers to use the new style defines for clocking
in DAI formats.

Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20210915180957.39996-1-broonie@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: silence uninitialized variable warning

Smatch complains that we might hit the continue path on every iteration
through the loop.

    sound/soc/mediatek/mt8195/mt8195-mt6359-rt1019-rt5682.c:831
      mt8195_mt6359_rt1019_rt5682_card_late_probe()
    error: uninitialized symbol 'sof_comp'.

Initialize "sof_comp" to NULL to silence this warning.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20211208151145.GA29257@kili
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: AMD: acp-config: fix missing dependency on SND_SOC_ACPI

With a custom .config, the following error is thrown:

ERROR: modpost: "snd_soc_acpi_codec_list"
[sound/soc/amd/snd-acp-config.ko] undefined!

Fix by adding a clear dependency on SND_SOC_ACPI

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20211207192309.43883-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: dt-bindings: rt5682s: add AMIC delay time property

Add the AMIC delay time to control how much delay time (ms) to unmute the stereo1 ADC.

Signed-off-by: Shuming Fan <shumingf@realtek.com>
Link: https://lore.kernel.org/r/20211208101654.28925-1-shumingf@realtek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: Intel: sof_rt5682: Move rt1015 speaker amp to common file

Move rt1015 driver code to common file to be consistent with
rt1011 and rt1015p.

No functional change.

Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Yong Zhi <yong.zhi@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211207192458.44007-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: rt5682s: add delay time to fix pop sound issue

There is a pop noise at the beginning of the capture data.
This patch adds the delay time before stereo1 ADC unmute to fix the pop sound issue.

Signed-off-by: Shuming Fan <shumingf@realtek.com>
Link: https://lore.kernel.org/r/20211208101718.28945-1-shumingf@realtek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: Intel: boards: add 'static' qualifiers for max98390 routes

Sparse warnings:

sound/soc/intel/boards/sof_maxim_common.c:140:33: error: symbol
'max_98390_dapm_routes' was not declared. Should it be static?

sound/soc/intel/boards/sof_maxim_common.c:156:33: error: symbol
'max_98390_tt_dapm_routes' was not declared. Should it be static?

Fixes: f316c9d9ba8ea ('ASoC: Intel: boards: add max98390 2/4 speakers support')
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211207211700.115319-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: fsl-asoc-card: Add missing Kconfig option for tlv320aic31xx

Merge series from Ariel D'Alessandro <ariel.dalessandro@collabora.com>:

This is a follow up of patchsets:

  [RFC patch 0/5] Support BCLK input clock in tlv320aic31xx
  [PATCH 0/4] fsl-asoc-card: Add optional dt property for setting mclk-id

Patch "ASoC: fsl-asoc-card: Support fsl,imx-audio-tlv320aic31xx codec"
in "[RFC patch 0/5] Support BCLK input clock in tlv320aic31xx" missed a
Kconfig option. Sending incremental patch fix.

Regards,
Ariel

Ariel D'Alessandro (1):
  ASoC: fsl-asoc-card: Add missing Kconfig option for tlv320aic31xx

 sound/soc/fsl/Kconfig | 1 +
 1 file changed, 1 insertion(+)

--
2.30.2

ASoC: mediatek: support memory-region assignment

Merge series from Trevor Wu <trevor.wu@mediatek.com>:

This series of patches adds support for memory-region assignment, so the
access region of DMA engine could be restricted.
Patches are based on broonie tree "for-next" branch.

Trevor Wu (2):
  ASoC: mediatek: mt8195: support reserved memory assignment
  dt-bindings: mediatek: mt8195: add memory-region property

 .../devicetree/bindings/sound/mt8195-afe-pcm.yaml         | 8 ++++++++
 sound/soc/mediatek/mt8195/mt8195-afe-pcm.c                | 7 +++++++
 2 files changed, 15 insertions(+)

--
2.18.0

ASoC: mediatek: Update MT8195 machine driver

Merge series from Trevor Wu <trevor.wu@mediatek.com>:

This series of patches adds support for RT5682s headset codec in mt8195
machine drivers, and SOF support on card mt8195-mt6359-rt1019-rt5682 is
also included.
Patches are based on broonie tree "for-next" branch.

Changes since v1:
  - remove patch3 and patch4 in v1
  - add SOF support on card mt8195-mt6359-rt1012-rt5682
  - add new propertes to dt-bindings for mt8195-mt6359-rt1019-rt5682

Trevor Wu (4):
  ASoC: mediatek: mt8195: add headset codec rt5682s support
  dt-bindings: mediatek: mt8195: add model property
  ASoC: mediatek: mt8195: add sof support on mt8195-mt6359-rt1019-rt5682
  dt-bindings: mediatek: mt8195: add adsp and dai-link property

 .../sound/mt8195-mt6359-rt1011-rt5682.yaml    |   4 +
 .../sound/mt8195-mt6359-rt1019-rt5682.yaml    |  14 +
 sound/soc/mediatek/Kconfig                    |   2 +
 .../mt8195/mt8195-mt6359-rt1011-rt5682.c      |  29 +-
 .../mt8195/mt8195-mt6359-rt1019-rt5682.c      | 347 +++++++++++++++++-
 5 files changed, 370 insertions(+), 26 deletions(-)

--
2.18.0

ASoC: mediatek: mt8195: add memory-region property

Add a required property "memory-region", which is used to specify memory
for DMA usage.

Signed-off-by: Trevor Wu <trevor.wu@mediatek.com>
Link: https://lore.kernel.org/r/20211130053905.28470-3-trevor.wu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: add adsp and dai-link property

1. adsp phandle can be assigned to the machine driver if adsp is enabled.
2. dai-link supported in the sound card can be specified from DTS.

Signed-off-by: Trevor Wu <trevor.wu@mediatek.com>
Link: https://lore.kernel.org/r/20211129141057.12422-5-trevor.wu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: add sof support on mt8195-mt6359-rt1019-rt5682

In the patch, widgets, routes and dai-link requrird by SOF are included,
and late_probe is introduced for SOF route connection.

Only when adsp phandle could be retrieved from DTS, the SOF related part
of machine driver is executed.

Additionally, supported dai-links could be specified from DTS, so that
we can disable AP side hardware controls when DSP SOF controls the same
audio FE.

Signed-off-by: Trevor Wu <trevor.wu@mediatek.com>
Signed-off-by: YC Hung <yc.hung@mediatek.com>
Link: https://lore.kernel.org/r/20211129141057.12422-4-trevor.wu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: add model property

This patch adds the description of model property used to specify card
name from dts.

Signed-off-by: Trevor Wu <trevor.wu@mediatek.com>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20211129141057.12422-3-trevor.wu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: amd: acp6x-pdm-dma: Constify static snd_soc_dai_ops

The only usage of acp6x_pdm_dai_ops is to assign its address to the ops
field in the snd_soc_dai_driver struct, which is a pointer to const
snd_soc_dai_ops. Make it const to allow the compiler to put it in
read-only memory.

Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Link: https://lore.kernel.org/r/20211204110848.21322-1-rikard.falkeborn@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: fix build issue related to CODEC_PROBE_ENTRIES

Fix following error:
sound/soc/sof/intel/hda-codec.c:132:35: error: use of undeclared identifier 'CODEC_PROBE_RETRIES'

Found with config: i386-randconfig-r033-20211202
(https://download.01.org/0day-ci/archive/20211203/202112031943.Twg19fWT-lkp@intel.com/config)

Fixes: 046aede2f847 ("ASoC: SOF: Intel: Retry codec probing if it fails")
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20211203154721.923496-1-kai.vehmanen@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: ti: davinci-mcasp: Remove unnecessary conditional

Instead of double validating of_node, return value of
the boolean property directly.

We can't remove ifdeffery, because in OF_GPIO=n cases
it might bring unwanted surprises.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://lore.kernel.org/r/20211202205612.76216-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: ti: davinci-mcasp: Get rid of duplicate of_node assignment

GPIO library does copy the of_node from the parent device of
the GPIO chip, there is no need to repeat this in the individual
drivers. Remove assignment here.

For the details one may look into the of_gpio_dev_init() implementation.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://lore.kernel.org/r/20211202205612.76216-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: zl38060: Setup parent device and get rid of unnecessary of_node assignment

Some of the drivers do not set parent device. This may lead to obstacles
during debugging or understanding the device relations from the Linux
point of view. Assign parent device for GPIO chips created by these
drivers.

While at it, let GPIO library to assign of_node from the parent device.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20211202204838.75287-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: test-component: fix null pointer dereference.

Dereferncing of_id pointer will result in exception in current
implementation since of_match_device() will assign it to NULL.
Adding NULL check for protection.

Signed-off-by: Ameer Hamza <amhamza.mgc@gmail.com>
Link: https://lore.kernel.org/r/20211205204200.7852-1-amhamza.mgc@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

sound/soc: remove useless bool conversion to bool variable

This patch remove useless bool conversion to bool variable

Signed-off-by: Bernard Zhao <bernard@vivo.com>
Link: https://lore.kernel.org/r/20211206021100.321170-1-bernard@vivo.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: add headset codec rt5682s support

mt8195 machine driver adds rt5682s support in this patch.
Card name can be specified from dts by model property, and driver makes
use of the name to distinguish which headset codec is on the board.

Signed-off-by: Trevor Wu <trevor.wu@mediatek.com>
Link: https://lore.kernel.org/r/20211129141057.12422-2-trevor.wu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: fsl-asoc-card: Add missing Kconfig option for tlv320aic31xx

The following commit added support for tlv320aic31xx codec to
fsl-asoc-card, but missed the related Kconfig option. Fix this.

  commit 8c9b9cfb7724685ce705f511b882f30597596536
  Author: Ariel D'Alessandro <ariel.dalessandro@collabora.com>
  Date:   Fri Nov 19 12:32:48 2021 -0300

      ASoC: fsl-asoc-card: Support fsl,imx-audio-tlv320aic31xx codec

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>
Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com>
Link: https://lore.kernel.org/r/20211203175018.252641-2-ariel.dalessandro@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: support reserved memory assignment

For security purpose, restrict the memory assess region of AFE memif.
The specified memory region should be assigned from DTS.

Signed-off-by: Trevor Wu <trevor.wu@mediatek.com>
Link: https://lore.kernel.org/r/20211130053905.28470-2-trevor.wu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: cs35l41: Fix undefined reference to core functions

Auto select core driver if i2c or spi bus drivers are
selected

Fixes: a5e0091d62ab ("ASoC: cs35l41: Fix link problem")

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/20211201180004.1402156-2-tanureal@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: amd: Fix dependency for SPI master

Set SPI_MASTER as dependency as is using CS35L41 SPI driver

Fixes: 96792fdd77cd1 ("ASoC: amd: enable vangogh platform machine driver build")

Signed-off-by: Lucas Tanure <tanureal@opensource.cirrus.com>
Reported-by: kernel test robot <lkp@intel.com>
Link: https://lore.kernel.org/r/20211201180004.1402156-1-tanureal@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: rt5640: Fix the wrong state of the JD in the HDA header

The patch fixes the wrong state of the JD with 1M pull up resistor in the
HDA header.

Signed-off-by: Oder Chiou <oder_chiou@realtek.com>
Link: https://lore.kernel.org/r/20211201095629.21818-1-oder_chiou@realtek.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: codecs: wcd934x: remove redundant ret variable

return value form snd_soc_dapm_put_enum_double() directly instead
of taking this in another redundant variable.

Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20211130160507.22180-3-srinivas.kandagatla@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: tegra: Add master volume/mute control support

The MVC module has a per channel control bit, based on which it decides
to apply channel specific volume/mute settings. When per channel control
bit is enabled (which is the default HW configuration), all MVC channel
volume/mute can be independently controlled. If the control is disabled,
channel-0 volume/mute setting is applied by HW to all remaining channels.
Thus add support to leverage this HW feature by exposing master controls
for volume/mute.

With this, now there are per channel and master volume/mute controls.
Users need to just use controls which are suitable for their applications.
The per channel control enable/disable is mananged in driver and hidden
from users, so that they need to just worry about respective volume/mute
controls.

Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Link: https://lore.kernel.org/r/1638278605-28225-1-git-send-email-spujar@nvidia.com
Signed-off-by: Mark Brown <broonie@kernel.org>

Merge branch 'for-5.16' of https://git./linux/kernel/git/broonie/sound into asoc-5.17 so we can apply new Tegra work

ASoC: Intel: boards: add max98390 2/4 speakers support

support 2 hw boards.
1. SSP2 connects max98390, 2 speakers.
2. SSP1 connects max98390, 2/4 speakers.

2 or 4 speakers playback
add echo reference capture
add bt offload support
add DMI_OEM_STRING for board variants
add ALC5682I-VS support

Signed-off-by: Mark Hsieh <mark_hsieh@wistron.corp-partner.google.com>
Signed-off-by: Mac Chiang <mac.chiang@intel.com>
Signed-off-by: Kieth Tzeng <keith.tzeng@quantatw.com>
Signed-off-by: Brent Lu <brent.lu@intel.com>
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211125030453.4382-1-mac.chiang@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: make several arrays static const

Don't populate various arrays on the stack but instead make them
static const. Also makes the object code smaller by a few hundred
bytes.

Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Link: https://lore.kernel.org/r/20211129224236.506883-1-colin.i.king@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

firmware: cs_dsp: Move lockdep asserts to avoid potential null pointer

Move the lockdep asserts until after the ctl pointer has been checked
for NULL, to avoid potentially NULL pointer dereferences.

Fixes: fb2f364fb5b9 ("firmware: cs_dsp: Add lockdep asserts to interface functions")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20211130102842.26410-1-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: intel: boards: bytcht*: Constify static snd_soc_ops

These are only assigned to the ops fields in the snd_soc_dai_link struct
which is a pointer to const struct snd_soc_ops. Make them const to allow
the compiler to put them in read-only memory.

Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20211127091954.12075-1-rikard.falkeborn@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: Intel: hda_dsp_common: don't multiline PCM topology warning

On my T14s Gen2 I saw the following:

    [   16.057258] skl_hda_dsp_generic skl_hda_dsp_generic: hda_dsp_hdmi_build_controls: no PCM in topology for HDMI converter 3

    [   16.057261] skl_hda_dsp_generic skl_hda_dsp_generic: hda_dsp_hdmi_build_controls: no PCM in topology for HDMI converter 4

    [   16.057263] skl_hda_dsp_generic skl_hda_dsp_generic: hda_dsp_hdmi_build_controls: no PCM in topology for HDMI converter 5

    [...and so on.]

It looks like the double newline is a mistake, so remove one.

Signed-off-by: Chris Down <chris@chrisdown.name>
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/YaOS0sBueAfApwOx@chrisdown.name
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: mediatek: mt8195: Constify static snd_soc_ops

These are only assigned to the ops field in the snd_soc_dai_link which
is a pointer to const struct snd_soc_ops. Make them const to allow the
compiler to put them in read-only memory.

Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Link: https://lore.kernel.org/r/20211127093147.17368-1-rikard.falkeborn@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>

ASoC: SOF: Intel: Retry codec probing if it fails

On the latest Lenovo Thinkstation laptops, we often experience the
speaker failure after rebooting, check the dmesg, we could see:
 sof-audio-pci-intel-tgl 0000:00:1f.3: codec #0 probe error, ret: -5

The analogue codec on the machine is ALC287, then we designed a
testcase to reboot and check the codec probing result repeatedly, we
found the analogue codec probing always failed at least once within
several minutes to several hours (roughly 1 reboot per min). This
issue happens on all laptops of this Thinkstation model, but with
legacy HDA driver, we couldn't reproduce this issue on those laptops.
And so far, this issue is not reproduced on machines which don't
belong to this model.

We tried to make the hda_dsp_ctrl_init_chip() same as
hda_intel_init_chip() which is the controller init routine in the
legacy HDA driver, but it didn't help.

We found when issue happens, the resp is -1, and if we let driver
re-run send_cmd() and get_response(), it will get the correct response
10ec0287, then driver continues the rest work, finally boot to the
desktop and all audio function work well.

Here adding codec probing retries to 3 times, it could fix the issue
on this Thinkstation model, and it doesn't bring impact to other
machines.

Reviewed-by: Bard Liao <bard.liao@intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Hui Wang <hui.wang@canonical.com>
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20211130090606.529348-1-kai.vehmanen@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>