When using the S/PDIF DAI, there is no requirement to call
snd_soc_dai_set_fmt() as there is no DAI format definition that defines
S/PDIF.  In any case, S/PDIF does not have separate clocks, this is
embedded into the data stream.

Consequently, when attempting to use TDA998x in S/PDIF mode, the attempt
to configure TDA998x via the hw_params callback fails as the
hdmi_codec_daifmt is left initialised to zero.

Since the S/PDIF DAI will only be used by S/PDIF, prepare the
hdmi_codec_daifmt structure for this format.
Signed-off-by: NRussell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: NJyri Sarha <jsarha@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

add direct loopback path from rx to tx
Signed-off-by: NKaiChieh Chuang <kaichieh.chuang@mediatek.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Dummy write in capture master mode is used to gate
bus clocks. This write is useless in slave mode
as the clocks are not managed by slave.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

When snd_pcm_stop_xrun() is called in interrupt routine,
substream context may have already been released.
Add protection on substream context.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Clocks do not need to be released on driver removal,
as this is already managed before.
Remove useless remove callback.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

DMA configuration is not balanced on start/stop.
Move DMA configuration to trigger callback.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Move counter handling to trigger start section
to manage multiple start/stop events.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

I2S supports 16 bits data in 32 channel length.
However the expected driver behavior, is to
set channel length to 16 bits when data format is 16 bits.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Because of regmap cache, interrupts may not be cleared
as expected.
Declare IFCR register as write only and make writings
to IFCR register unconditional.
Signed-off-by: NOlivier Moysan <olivier.moysan@st.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Merge branch 'for-5.0' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into asoc-5.1

SND_SOC_CROS_EC_CODEC depends on MFD_CROS_EC.
Add that dependency to SND_SOC_SDM845 to fix unmet direct dependencies
warning.

Fixes: 74c6ecf4 (ASoC: qcom: Kconfig: select dmic for sdm845)
Signed-off-by: NCheng-Yi Chiang <cychiang@chromium.org>
Reported-by: NRandy Dunlap <rdunlap@infradead.org>
Tested-by: NEnric Balletbo i Serra <enric.balletbo@collabora.com>
Tested-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NMark Brown <broonie@kernel.org>

This patch enables the reuse of kbl_da7219_max98927 machine driver to
support max98373. The same machine driver is modified for cases where one
amplifier is swapped out with another. Most of the changes are about
renaming the codec and codec_dai names, with minor differences due to
support for 24 bits in one case and 16 in the other.
Signed-off-by: NJenny TC <jenny.tc@intel.com>
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently each SSI unit 's busif mode/adinr/dalign address is
registered by: (in busif4 case)
RSND_GEN_M_REG(SSI_BUSIF4_MODE, 0x500, 0x80)
RSND_GEN_M_REG(SSI_BUSIF4_ADINR,0x504, 0x80)
RSND_GEN_M_REG(SSI_BUSIF4_DALIGN, 0x508, 0x80)

But according to user manual 41.1.4 Register Configuration
ssi9 4/5/6/7 busif mode/adinr/dalign register address
( SSI9-[4/5/6/7]_BUSIF_[MODE/ADINR/DALIGN] )
are out of this rule.

This patch registers ssi9 4/5/6/7 mode/adinr/dalign register
as single register, and access these registers in case of
SSI9 BUSIF 4/5/6/7.

Fixes: commit 8c9d7503 ("ASoC: rsnd: ssiu: Support BUSIF other than BUSIF0")
Signed-off-by: NJiada Wang <jiada_wang@mentor.com>
Signed-off-by: NTimo Wischer <twischer@de.adit-jv.com>
Acked-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

As the compressed stream implementation has acquired support for
multiple DAI links and compressed streams it has become harder to
interpret messages in the kernel log. Add additional macros to include
the compressed DAI name in the log messages, allowing different streams
to be easily disambiguated.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently, only a single compressed stream is supported per firmware.
Add support for multiple compressed streams on a single firmware, this
allows additional features like completely independent trigger words or
separate debug capture streams to be implemented.
Signed-off-by: NStuart Henderson <stuarth@opensource.cirrus.com>
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Make the code slightly clearer and prepare things for the addition of
multiple compressed streams on a single DSP core.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

In preparation for more refactoring add a helper function to strip the
padding from ADSP data.
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The irq_get_irq_data() function doesn't return error pointers, it
returns NULL.

Fixes: 6ba9dd6c ("ASoC: cs35l36: Add support for Cirrus CS35L36 Amplifier")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

This addresses an issue pointed out by compiler warning:

sound/soc/samsung/odroid.c: In function ‘odroid_audio_probe’:
sound/soc/samsung/odroid.c:298:22: warning: ‘cpu_dai’ may be used
uninitialized in this function [-Wmaybe-uninitialized]
  priv->clk_i2s_bus = of_clk_get_by_name(cpu_dai, "iis");
                      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Although qcom_snd_parse_of() tries to manage the of-node refcount,
there are still a few places that lead to the unblanced refcount in
the error code path.  Namely,

- for_each_child_of_node() needs to unreference the iterator node if
  aborting the loop in the middle,
- cpu, codec and platform node objects have to be unreferenced at each
  iteration,
- platform and codec node objects have to be referred before jumping
  to the error handling code that unreference them unconditionally.

This patch tries to address these by moving the assignment of platform
and codec node objects to the beginning of the loop and adding the
of_node_put() calls adequately.

Fixes: c25e295c ("ASoC: qcom: Add support to parse common audio device nodes")
Cc: Patrick Lai <plai@codeaurora.org>
Cc: Banajit Goswami <bgoswami@codeaurora.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

The apq8016 driver leaves the of-node refcount at aborting from the
loop of for_each_child_of_node() in the error path.  Not only the
iterator node of for_each_child_of_node(), the children nodes referred
from it for codec and cpu have to be properly unreferenced.

Fixes: bdb052e8 ("ASoC: qcom: add apq8016 sound card support")
Cc: Patrick Lai <plai@codeaurora.org>
Cc: Banajit Goswami <bgoswami@codeaurora.org>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NMark Brown <broonie@kernel.org>

In odroid_audio_probe() some OF nodes are left without reference count
decrease after use. Fix it by ensuring required of_node_calls() are done
before exiting probe.
Reported-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The function simple_for_each_link() has a few missing places that
forgot unrefereing of-nodes after the use.  The main do-while loop
may abort when loop=0, and this leaves the node object still
referenced.  A similar leak is found in the error handling of NULL
codec that aborts the loop as well.  Last but not least, the inner
for_each_child_of_node() loop may abort in the middle, and this leaks
the refcount of the iterator node.

This patch addresses these missing refcount issues.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Acked-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

We forgot to unreference the platform node object obtained from
of_get_child_by_name().  This leads to the unbalance of node
refcount.

Fixes: e0ae225b ("ASoC: simple-card: support platform in dts parse")
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Acked-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The node obtained from of_find_node_by_path() has to be unreferenced
after the use, but we forgot it for the root node.

Fixes: f0fba2ad ("ASoC: multi-component - ASoC Multi-Component Support")
Cc: Timur Tabi <timur@kernel.org>
Cc: Nicolin Chen <nicoleotsuka@gmail.com>
Cc: Xiubo Li <Xiubo.Lee@gmail.com>
Cc: Fabio Estevam <festevam@gmail.com>
Signed-off-by: NTakashi Iwai <tiwai@suse.de>
Acked-by: NNicolin Chen <nicoleotsuka@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Currently, compressed buffers can only be specified in the XM memory
region. There is no reason to have such a restriction with the newer
meta-data based way of specifying the buffers, so remove it.
Signed-off-by: NAndrew Ford <aford@opensource.cirrus.com>
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

If a compressed stream is restarted after getting an error, the cached
error value will still be used on the next pointer request, preventing
the stream from starting.  Resolve this by ensuring the error status is
updated on trigger start.
Signed-off-by: NStuart Henderson <stuarth@opensource.cirrus.com>
Signed-off-by: NCharles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The ADCs are sleeping when the SLEEP bit is set and running when it's
cleared, so the bit should be inverted.
Tested on pcm1863.
Signed-off-by: NCodrin Ciubotariu <codrin.ciubotariu@microchip.com>
Acked-by: NAndrew F. Davis <afd@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

According to DS, the gain is between -12 dB and 40 dB, with a 0.5 dB step.
Tested on pcm1863.
Signed-off-by: NCodrin Ciubotariu <codrin.ciubotariu@microchip.com>
Acked-by: NAndrew F. Davis <afd@ti.com>
Signed-off-by: NMark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org

Merge branch 'for-5.0' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into asoc-5.1

On some SoCs (e.g. Exynos5433) there are multiple "IIS multi audio
interfaces" and the driver will try to register there multiple times
same platform device for the secondary FIFO, which of course fails
miserably.  To fix this we derive the secondary platform device name
from the primary device name. The secondary device name will now
be <primary_dev_name>-sec instead of fixed "samsung-i2s-sec".

The fixed platform_device_id table entry is removed as the secondary
device name is now dynamic and device/driver matching is done through
driver_override.
Reported-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Suggested-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Acked-by: NKrzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: NMark Brown <broonie@kernel.org>

This fixes unregistration of the secondary platform device so all
resources are properly released.  Additionally the removal sequence
is corrected so it is in reverse order comparing to probe sequence.
The test against NULL priv->pdev_sec is removed as it is not necessary.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Acked-by: NKrzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: NMark Brown <broonie@kernel.org>

Add FE DAI link to support parallel playback on 2 ports
simultaneously.
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NNaveen Manohar <naveen.m@intel.com>
Signed-off-by: NYong Zhi <yong.zhi@intel.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

This patch enables support for GeminiLake with the DA7219 codec and
MAX98357A amplifier. To avoid duplicating code, the existing machine
driver for ApolloLake is reused with only changes in hardware
connectivity (SSP2 for DA7219 and SSP1 for MAX98357A).

The dailinks are directly modified in this patch. Using a helper would
be nicer, but it'll be done in a follow-up step with validation done
across multiple machine drivers.
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NYong Zhi <yong.zhi@intel.com>
Signed-off-by: NNaveen Manohar <naveen.m@intel.com>
Signed-off-by: NHarsha Priya <harshapriya.n@intel.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Map the 4 headset buttons to KEY_PAUSE, KEY_VOLUMEUP, KEY_VOLUMEDOWN and
KEY_VOICECOMMAND.
Acked-by: NPierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: NNaveen Manohar <naveen.m@intel.com>
Signed-off-by: NYong Zhi <yong.zhi@intel.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

The A64 datasheet lists the supply rail for the headphone amp's charge
pump as "CPVDD". cpvdd-supply is the name of the property for this power
rail specified in the device tree bindings. "HPVCC" was the name used in
the A33 datasheet for the same function.

Rename the supply so it matches the datasheet, bindings, and the subject
from the original commit.

Fixes: ca0412a0 ("ASoC: sunxi: sun50i-codec-analog: Add support for cpvdd regulator supply")
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NMark Brown <broonie@kernel.org>

If platform_data is NULL add reading of optional adi,micbias
property from DT. If adi,micbias is not set keep the default
value for micbias.
Signed-off-by: NBogdan Togorean <bogdan.togorean@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

Add MICBIAS property to the optional devicetree bindings.
Signed-off-by: NBogdan Togorean <bogdan.togorean@gmail.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

We should free "w" on the error path.

Fixes: 199ed3e8 ("ASoC: dapm: fix use-after-free issue with dailink sname")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NMark Brown <broonie@kernel.org>