Colin Ian King [Thu, 2 Jul 2020 11:48:35 +0000 (12:48 +0100)]
ASoC: Intel: bxt-da7219-max98357a: return -EINVAL on unrecognized speaker amplifier
Currently if the ctx->spkamp is not recognized an error message is
reported but the code continues to set up the device with uninitialized
variables such as the number of widgets. Fix this by returning -EINVAL
for unrecognized speaker amplifier types.
Fixes: e1435a1feb18 ("ASoC: Intel: bxt-da7219-max98357a: support MAX98390 speaker amp") Signed-off-by: Colin Ian King <colin.king@canonical.com> Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Addresses-Coverity: ("Uninitialized scalar variable") Link: https://lore.kernel.org/r/20200702114835.37889-1-colin.king@canonical.com Signed-off-by: Mark Brown <broonie@kernel.org>
cs4270.c:508: warning: Function parameter or member 'component' not
described in 'cs4270_probe'
cs4270.c:508: warning: Excess function parameter 'pdev' description in
'cs4270_probe'
cs4270.c:548: warning: Function parameter or member 'component' not
described in 'cs4270_remove'
cs4270.c:548: warning: Excess function parameter 'pdev' description in
'cs4270_remove'
ASoC: Intel: atom: fix 'defined but not used' warning
Fix W=1 warning. The VOIP controls were not used in the mainline but
in special versions of Android. Keep and use __maybe_used to make
warning go away.
Brent Lu [Wed, 1 Jul 2020 09:36:51 +0000 (17:36 +0800)]
ASoC: Intel: bxt-da7219-max98357a: support MAX98390 speaker amp
Support MAX98390 speaker amplifier on cometlake platform. Driver now
detects amplifier type in the probe function and installs corresponding
controls and DAPM widgets/routes in the late_probe function.
Mark Brown [Wed, 1 Jul 2020 16:21:09 +0000 (17:21 +0100)]
Merge series "ASoC: ti: Add support for audio on J721e EVM" from Peter Ujfalusi <peter.ujfalusi@ti.com>:
Hi,
Changes since v3:
- Fix the single clock source handling and typo
Changes since v2:
- DT binding:
- use proper (?) patch subject for the binding docuemtn patch
- drop pll4 and pll15 from DT - driver should check the rate via
clk_get_parent. If it is not available (as it is not currently) then use the
match_data provided rates.
- add simple explanation for the clocking setup
- Use descriptive names for clocks: cpb/ivi-mcasp-auxclk and cpb/ivi-codec-scki
- dt_binding_check shows no errors/warnings
- ASoC machine driver:
- Try to read the PLL4/15 rate with clk API (parent of the two clock divider)
if it is not available then use the match_data provided numbers.
- Support for single PLL setup
Changes since v1:
- Fixed DT binding documentation errors
- Rebased on ASoC head and updated the driver to compile and work
This series adds support for the analog audio setup on the j721e EVM.
The audio setup of the EVM is:
Common Processor Board (CPB): McASP10 <-> pcm3168a
Infotainment Expansion Board (IVI): McASP0 <-> 2x pcm3168a
Both CPB and IVI wired in parallel serializer setup.
The first patch adds the stream_name for McASP driver as it is needed in
multicodec (and would be needed in DPCM) setup for proper DAPM handling.
The second patch adds two DT schema, one for the cpb and one for the cpb+ivi
card.
Regards,
Peter
---
Peter Ujfalusi (3):
ASoC: ti: davinci-mcasp: Specify stream_name for playback/capture
ASoC: dt-bindings: Add documentation for TI j721e EVM (CPB and IVI)
ASoC: ti: Add custom machine driver for j721e EVM (CPB and IVI)
Audio Graph Card is using "audio-graph-card" prefix instead of
"simple-graph-card", and moreover "widgets / routing" doesn't need it.
This patch removes unsupported "simple-graph-card" prefix from
motorola-mapphone-common.dtsi and vendor-prefixes.yaml.
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.com> Acked-by: Tony Lindgren <tony@atomide.com> Link: https://lore.kernel.org/r/87r1ub39hq.wl-kuninori.morimoto.gx@renesas.com Signed-off-by: Mark Brown <broonie@kernel.org>
Shengjiu Wang [Tue, 30 Jun 2020 13:56:07 +0000 (21:56 +0800)]
ASoC: fsl_asrc: Add an option to select internal ratio mode
The ASRC not only supports ideal ratio mode, but also supports
internal ratio mode.
For internal rato mode, the rate of clock source should be divided
with no remainder by sample rate, otherwise there is sound
distortion.
Add function fsl_asrc_select_clk() to find proper clock source for
internal ratio mode, if the clock source is available then internal
ratio mode will be selected.
With change, the ideal ratio mode is not the only option for user.
Akshu Agrawal [Tue, 30 Jun 2020 18:37:46 +0000 (00:07 +0530)]
ASoC: amd: Enable interrupt in dma_open
Fixes interrupt enable condition check with which now
interrupt gets enabled in dma_open.
Prior to this patch it was getting enabled in runtime_resume only.
The following build warnings are seen with 'make dt_binding_check':
Documentation/devicetree/bindings/sound/simple-card.example.dts:209.46-211.15: Warning (unit_address_vs_reg): /example-4/sound/simple-audio-card,cpu@0: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:213.37-215.15: Warning (unit_address_vs_reg): /example-4/sound/simple-audio-card,cpu@1: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:250.42-261.15: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@0: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:263.42-288.15: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@1: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:270.32-272.19: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@1/cpu@0: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:273.23-275.19: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@1/cpu@1: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:276.23-278.19: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@1/cpu@2: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:279.23-281.19: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@1/cpu@3: node has a unit name, but no reg or ranges property
Documentation/devicetree/bindings/sound/simple-card.example.dts:290.42-303.15: Warning (unit_address_vs_reg): /example-5/sound/simple-audio-card,dai-link@2: node has a unit name, but no reg or ranges property
In file included from sound/soc/intel/keembay/kmb_platform.c:14:
sound/soc/intel/keembay/kmb_platform.h:9:9: warning: 'KMB_PLATFORM_H_'
is used as a header guard here, followed by #define of a different
macro [-Wheader-guard]
#ifndef KMB_PLATFORM_H_
^~~~~~~~~~~~~~~
sound/soc/intel/keembay/kmb_platform.h:10:9: note: 'KMB_PLATFORMP_H_'
is defined here; did you mean 'KMB_PLATFORM_H_'?
#define KMB_PLATFORMP_H_
^~~~~~~~~~~~~~~~
KMB_PLATFORM_H_
1 warning generated.
Fix the typo so that the header guard works as intended.
Fixes: c5477e966728 ("ASoC: Intel: Add KeemBay platform driver") Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> Reviewed-by: Nick Desaulniers <ndesaulniers@google.com> Cc: Sia, Jee Heng <jee.heng.sia@intel.com>; alsa-devel@alsa-project.org; linux-kernel@vger.kernel.org; clang-built-linux@googlegroups.com; Nathan Chancellor <natechancellor@gmail.com> Link: https://github.com/ClangBuiltLinux/linux/issues/1053
To: Rojewski, Cezary <cezary.rojewski@intel.com>; Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>; Liam Girdwood <liam.r.girdwood@linux.intel.com>; Jie Yang <yang.jie@linux.intel.com>; Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/20200617010232.23222-1-natechancellor@gmail.com Signed-off-by: Mark Brown <broonie@kernel.org>
Shengjiu Wang [Mon, 29 Jun 2020 06:42:33 +0000 (14:42 +0800)]
ASoC: fsl_sai: Refine regcache usage with pm runtime
When there is dedicated power domain bound with device, after probing
the power will be disabled, then registers are not accessible in
fsl_sai_dai_probe(), so regcache only need to be enabled in end of
probe() and regcache_mark_dirty should be moved to pm runtime resume
callback function.
Peter Ujfalusi [Tue, 30 Jun 2020 12:58:43 +0000 (15:58 +0300)]
ASoC: ti: Add custom machine driver for j721e EVM (CPB and IVI)
The audio support on the board is using pcm3168a codec connected to McASP10
serializers in parallel setup.
The pcm3168a SCKI clock is coming via the j721e AUDIO_REFCLK2 pin.
In order to support 48KHz and 44.1KHz family of sampling rates the parent clock
for AUDIO_REFCLK2 needs to be changed between PLL4 (for 48KHz) and PLL15 (for
44.1KHz). The same PLLs are used for McASP10's AUXCLK clock via different
HSDIVIDER.
Generic card can not be used for the board as we need to switch between
clock paths for different sampling rate families and also need to change
the slot_width between 16 and 24 bit audio.
The audio support on the Infotainment Expansion Board consists of McASP0
connected to two pcm3168a codecs with dedicated set of serializers to each.
The SCKI for pcm3168a is sourced from j721e AUDIO_REFCLK0 pin.
It is extending the audio support on the CPB.
Due to the fact that the same PLL4/15 is used by both domains (CPB/IVI)
there are cross restriction on sampling rates.
The IVI side is represented as multicodec setup.
PCMs available on a plain CPB (no IVI addon):
hw:0,0 - cpb playback (8 channels)
hw:0,1 - cpb capture (6 channels)
When the IVI addon is present, additional two PCMs will be present:
hw:0,2 - ivi multicodec playback (16 channels)
hw:0,3 - ivi multicodec capture (12 channels)
Peter Ujfalusi [Tue, 30 Jun 2020 12:58:42 +0000 (15:58 +0300)]
ASoC: dt-bindings: Add documentation for TI j721e EVM (CPB and IVI)
The audio support on the Common Processor Board board is using
pcm3168a codec connected to McASP10 serializers in parallel setup.
The Infotainment board plugs into the Common Processor Board, the support
of the extension board is extending the CPB audio support by adding
the two codecs on the expansion board.
The audio support on the Infotainment Expansion Board consists of McASP0
connected to two pcm3168a codecs with dedicated set of serializers to each.
The SCKI for pcm3168a is sourced from j721e AUDIO_REFCLK0 pin.
Hans de Goede [Sun, 28 Jun 2020 15:52:29 +0000 (17:52 +0200)]
ASoC: rt5670: Fix dac- and adc- vol-tlv values being off by a factor of 10
The adc_vol_tlv volume-control has a range from -17.625 dB to +30 dB,
not -176.25 dB to + 300 dB. This wrong scale is esp. a problem in userspace
apps which translate the dB scale to a linear scale. With the logarithmic
dB scale being of by a factor of 10 we loose all precision in the lower
area of the range when apps translate things to a linear scale.
E.g. the 0 dB default, which corresponds with a value of 47 of the
0 - 127 range for the control, would be shown as 0/100 in alsa-mixer.
Since the centi-dB values used in the TLV struct cannot represent the
0.375 dB step size used by these controls, change the TLV definition
for them to specify a min and max value instead of min + stepsize.
Hans de Goede [Sun, 28 Jun 2020 15:52:28 +0000 (17:52 +0200)]
ASoC: rt5670: Add new gpio1_is_ext_spk_en quirk and enable it on the Lenovo Miix 2 10
The Lenovo Miix 2 10 has a keyboard dock with extra speakers in the dock.
Rather then the ACL5672's GPIO1 pin being used as IRQ to the CPU, it is
actually used to enable the amplifier for these speakers
(the IRQ to the CPU comes directly from the jack-detect switch).
Add a quirk for having an ext speaker-amplifier enable pin on GPIO1
and replace the Lenovo Miix 2 10's dmi_system_id table entry's wrong
GPIO_DEV quirk (which needs to be renamed to GPIO1_IS_IRQ) with the
new RT5670_GPIO1_IS_EXT_SPK_EN quirk, so that we enable the external
speaker-amplifier as necessary.
Also update the ident field for the dmi_system_id table entry, the
Miix models are not Thinkpads.
Hans de Goede [Sun, 28 Jun 2020 15:52:27 +0000 (17:52 +0200)]
ASoC: rt5670: Correct RT5670_LDO_SEL_MASK
The RT5670_PWR_ANLG1 register has 3 bits to select the LDO voltage,
so the correct mask is 0x7 not 0x3.
Because of this wrong mask we were programming the ldo bits
to a setting of binary 001 (0x05 & 0x03) instead of binary 101
when moving to SND_SOC_BIAS_PREPARE.
According to the datasheet 001 is a reserved value, so no idea
what it did, since the driver was working fine before I guess we
got lucky and it does something which is ok.
Hans de Goede [Sun, 28 Jun 2020 15:52:26 +0000 (17:52 +0200)]
ASoC: Intel: cht_bsw_rt5672: Change bus format to I2S 2 channel
The default mode for SSP configuration is TDM 4 slot and so far we were
using this for the bus format on cht-bsw-rt56732 boards.
One board, the Lenovo Miix 2 10 uses not 1 but 2 codecs connected to SSP2.
The second piggy-backed, output-only codec is inside the keyboard-dock
(which has extra speakers). Unlike the main rt5672 codec, we cannot
configure this codec, it is hard coded to use 2 channel 24 bit I2S.
Using 4 channel TDM leads to the dock speakers codec (which listens in on
the data send from the SSP to the rt5672 codec) emiting horribly distorted
sound.
Since we only support 2 channels anyways, there is no need for TDM on any
cht-bsw-rt5672 designs. So we can simply use I2S 2ch everywhere.
This commit fixes the Lenovo Miix 2 10 dock speakers issue by changing
the bus format set in cht_codec_fixup() to I2S 2 channel.
This change has been tested on the following devices with a rt5672 codec:
Dan Murphy [Fri, 26 Jun 2020 15:41:41 +0000 (10:41 -0500)]
ASoC: tas2562: Add voltage sense slot property
Add a property to configure the slot for the voltage sense monitoring of
the device. Vsense data will be sent to the processor via the slot
defined by the property
Dan Murphy [Fri, 26 Jun 2020 15:41:42 +0000 (10:41 -0500)]
ASoC: tas2562: Add voltage sense slot configuration
Add Vsense slot configuration based on the device tree. Adding this
property enables the slot programming to be moved to the tdm_set_slot
callback. This in affect sets the slots for the Isense and Vsense and
enabling this these modes are now based on whether these features were
powered on or not.
Randy Dunlap [Mon, 29 Jun 2020 03:23:33 +0000 (20:23 -0700)]
ASoC: Documentation: fix reference to renamed source file
sound/soc/soc-io.c was merged into sound/soc/soc-component.c, so fixup
the Documentation to use the updated file name.
Error: Cannot open file ../sound/soc/soc-io.c
WARNING: kernel-doc '../scripts/kernel-doc -rst -enable-lineno ../sound/soc/soc-io.c' failed with return code 1
Fixes: 460b42d162e3 ("ASoC: soc-component: merge soc-io.c into soc-component.c") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Cc: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Cc: Mark Brown <broonie@kernel.org> Link: https://lore.kernel.org/r/a9f59f30-8cf2-ea82-567c-1706fd64fe62@infradead.org Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: qcom: Drop HAS_DMA dependency to fix link failure
When building on allyesconfig kernel for a NO_DMA=y platform (e.g.
Sun-3), CONFIG_SND_SOC_QCOM_COMMON=y, but CONFIG_SND_SOC_QDSP6_AFE=n,
leading to a link failure:
sound/soc/qcom/common.o: In function `qcom_snd_parse_of':
common.c:(.text+0x2e2): undefined reference to `q6afe_is_rx_port'
While SND_SOC_QDSP6 depends on HAS_DMA, SND_SOC_MSM8996 and SND_SOC_SDM845
don't, so the following warning is seen:
Until recently, this warning was harmless (from a compile-testing
point-of-view), but the new user of q6afe_is_rx_port() turned this into
a hard failure.
As the QDSP6 driver itself builds fine if NO_DMA=y, and it depends on
QCOM_APR (which in turns depends on ARCH_QCOM || COMPILE_TEST), it is
safe to increase compile testing coverage. Hence fix the link failure
by dropping the HAS_DMA dependency of SND_SOC_QDSP6.
Fixes: a2120089251f1fe2 ("ASoC: qcom: common: set correct directions for dailinks") Fixes: 6b1687bf76ef84cb ("ASoC: qcom: add sdm845 sound card support") Fixes: a6f933f63f2ffdb2 ("ASoC: qcom: apq8096: Add db820c machine driver") Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> Link: https://lore.kernel.org/r/20200629122443.21736-1-geert@linux-m68k.org Signed-off-by: Mark Brown <broonie@kernel.org>
Naveen Manohar [Thu, 25 Jun 2020 19:26:20 +0000 (14:26 -0500)]
ASoC: Intel: sof_sdw: Add MAX98373 support
Add max98373-sdw helper function, which configures 2x MAX98373 codecs to
Link1. This patch shares code between the I2S and SoundWire modes of
MAX98373 and adds the trigger already added for I2S.
Signed-off-by: Rander Wang <rander.wang@linux.intel.com> Signed-off-by: Naveen Manohar <naveen.m@intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com> Link: https://lore.kernel.org/r/20200625192620.4312-1-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Fred Oh [Thu, 25 Jun 2020 19:13:02 +0000 (14:13 -0500)]
ASoC: Intel: Boards: cml_rt1011_rt5682: use statically define codec config
When the cml_rt1011_rt5682_dailink[].codecs pointer is overridden by
a quirk with a devm allocated structure and the probe is deferred,
in the next probe we will see an use-after-free condition
(verified with KASAN). This can be avoided by using statically allocated
configurations - which simplifies the code quite a bit as well.
Dharageswari R [Thu, 25 Jun 2020 19:12:59 +0000 (14:12 -0500)]
ASoC: Intel: Boards: tgl_max98373: Fix the comment for max_98373_components
MAX_98373_DEV0_NAME is the Right speaker and MAX_98373_DEV1_NAME is the
Left speaker, hence updating the comments for max98373 dailink components
accordingly.
Signed-off-by: Dharageswari R <dharageswari.r@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: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/20200625191308.3322-9-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Dharageswari R [Thu, 25 Jun 2020 19:12:58 +0000 (14:12 -0500)]
ASoC: Intel: Boards: tgl_max98373: add dai_trigger function
Speaker amplifier feedback is not modeled as being dependent on any
active output. Even when there is no playback happening, parts of the
graph, specifically the IV sense->speaker protection->output remains
active and this prevents the DSP from entering low-power states.
This patch suggests a machine driver level approach where the speaker
pins are enabled/disabled dynamically depending on stream start/stop
events. DPAM graph representations show the feedback loop is indeed
disabled and low-power states can be reached.
Signed-off-by: Dharageswari R <dharageswari.r@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: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/20200625191308.3322-8-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
ASoC: Intel: sof_sdw: add quirk override with kernel parameter
During the bring-up of new platforms, or to take care of specific
hardware reworks, it's useful to add a kernel parameter to override
the default DMI-based quirks.
For example, adding the following line in a .conf file in
/etc/modprobe.d/ will change the default quirk and log the changes if
dynamic debug is enabled.
Tzung-Bi Shih [Thu, 25 Jun 2020 15:35:42 +0000 (23:35 +0800)]
ASoC: dapm: declare missing structure prototypes
To fix compilation warnings:
- struct 'snd_soc_pcm_runtime' declared inside parameter list will not
be visible outside of this definition or declaration
- struct 'soc_enum' declared inside parameter list will not be visible
outside of this definition or declaration
Jaroslav Kysela [Thu, 25 Jun 2020 11:58:29 +0000 (13:58 +0200)]
AsoC: amd: add missing snd- module prefix to the acp3x-rn driver kernel module
Signed-off-by: Jaroslav Kysela <perex@perex.cz> Acked-by: Alex Deucher <alexander.deucher@amd.com> Cc: Mark Brown <broonie@kernel.org> Cc: vijendar.mukunda@amd.com Cc: Alexander.Deucher@amd.com Link: https://lore.kernel.org/r/20200625115829.791750-1-perex@perex.cz Signed-off-by: Mark Brown <broonie@kernel.org>
Mark Brown [Tue, 23 Jun 2020 11:54:12 +0000 (12:54 +0100)]
Merge series "ASoC: add dailink .exit() callback" from Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>:
While looking at reboot issues and module load/unload tests, I found
out some resources allocated in the dailink .init() callback are not
properly released - there is no existing mechanism in the soc-core to
do so.
The addition of a dailink .exit() callback seems to be the simplest
solution overall. It can be argued that the existing machine platform
device .remove() callback can also perform the necessary cleanups,
however as shown in the last two examples this might require a loop to
identify components whereas the dailink .exit() already has all the
necessary information to revert the actions done in the .init() step.
Changes since RFC:
Better commit messages and explanations
rt5682 cases with snd_soc_component_set_jack() called in the .exit()
Fred Oh (2):
ASoC: intel: sof_rt5682: move disabling jack to dai link's exit()
ASoC: intel: cml_rt1011_rt5682: disable jack in dailink .exit()
Fred Oh [Mon, 22 Jun 2020 15:13:48 +0000 (10:13 -0500)]
ASoC: rt1011: fix KASAN out-of-bounds bug in find_next_bit()
KASAN throws the following warning in rt1011.c:
[ 170.777603] BUG: KASAN: stack-out-of-bounds in _find_next_bit.constprop.0+0x3e/0xf0
find_next_bit() relies on unsigned long pointer arguments, but this driver
uses a type cast that generates the KASAN warning. Replace find_next_bit()
and find_last_bit() with __ffs() and __fls() to pass the value and avoid
casting pointers to make the warning go away.
Fred Oh [Mon, 22 Jun 2020 15:42:41 +0000 (10:42 -0500)]
ASoC: intel: cml_rt1011_rt5682: disable jack in dailink .exit()
When removing the machine driver, the rt5682 jack handler will oops if jack
detection is not disabled. The jack can be disabled in the dai link's exit().
This is symmetrical change as jack is enabled in init().
The mainline code currently prevents modules from being removed.
The BE dailink .init() function calls devm_gpiod_get() using the codec
component device as argument. When the machine driver is removed, the
references to the gpiod are not released, and it's not possible to
remove the codec driver module - which is the only entity which could
free the gpiod.
This conceptual deadlock can be avoided by invoking gpiod_get() in the
.init() callback, and calling gpiod_put() in the exit() callback.
Tested on SAMUS Chromebook with SOF driver.
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com> Reviewed-by: Curtis Malainey <curtis@malainey.com> Link: https://lore.kernel.org/r/20200622154241.29053-3-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Some machine drivers allocate or request resources with
snd_soc_link_init() phase of the card probe. These resources need to
be properly released when removing a card, and this patch suggests a
dual exit() callback.
The exit() is invoked in soc_remove_pcm_runtime(), which is not
completely symmetric with the init() invoked in soc_init_pcm_runtime().
Alternate solutions were considered, e.g. adding a .remove() callback
for the platform driver, but that's not symmetrical at all and would
be difficult to handle if there are more than one dailink implementing
an .init(). We looked also into using .remove_dai_link() callback, but
that would also be imbalanced.
Note that because of the error handling in snd_soc_bind_card(), which
jumps to probe_end, there is no way to guarantee the exit() is invoked
with resources allocated in the init(). Prior to releasing those
resources, implementations of the exit() callback shall check the
resources are valid.
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com> Reviewed-by: Curtis Malainey <curtis@malainey.com> Link: https://lore.kernel.org/r/20200622154241.29053-2-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
Shengjiu Wang [Tue, 23 Jun 2020 06:01:11 +0000 (14:01 +0800)]
ASoC: fsl_mqs: Don't check clock is NULL before calling clk API
Because clk_prepare_enable and clk_disable_unprepare should
check input clock parameter is NULL or not internally, then
we don't need to check them before calling the function.
Mark Brown [Mon, 22 Jun 2020 14:36:06 +0000 (15:36 +0100)]
Merge series "ASoC: merge snd_soc_component_read() and snd_soc_component_read32()" from Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>:
Hi Mark
Because we can use "read" function which is using
"regmap" or "driver", current ALSA SoC drivers are using both
snd_soc_component_read() // for regmap
snd_soc_component_read32() // for driver callback
These are similar but needs different parameter.
This patch aggressively merge snd_soc_component_read() and _read32(),
and makes snd_soc_component_read/write() as generally style.
New read doesn't return error if it failed,
thus, it can't keep compatibility,
but assuming it is not a big problem.
Because 1) it will indicate error message,
2) it can do nothing anyway if it fails "read".
[02/16] patch is not directly connected to read function merging,
but need to be apply after [01/16].
Kuninori Morimoto (16):
ASoC: soc-component: merge snd_soc_component_read() and snd_soc_component_read32()
ASoC: soc-component: use io_mutex correctly
ASoC: pxa: rename to snd_soc_component_read()
ASoC: atmel: rename to snd_soc_component_read()
ASoC: codecs: rename to snd_soc_component_read()
ASoC: codecs: wcd*: rename to snd_soc_component_read()
ASoC: codecs: tlv*: rename to snd_soc_component_read()
ASoC: codecs: max*: rename to snd_soc_component_read()
ASoC: codecs: msm*: rename to snd_soc_component_read()
ASoC: codecs: alc*: rename to snd_soc_component_read()
ASoC: codecs: wm*: rename to snd_soc_component_read()
ASoC: codecs: rt*: rename to snd_soc_component_read()
ASoC: codecs: da*: rename to snd_soc_component_read()
ASoC: codecs: cs*: rename to snd_soc_component_read()
ASoC: codecs: ak*: rename to snd_soc_component_read()
ASoC: remove snd_soc_component_read32()
Signed-off-by: Vlad Karpovich <Vlad.Karpovich@cirrus.com> Signed-off-by: David Rhodes <david.rhodes@cirrus.com> Acked-by: Charles Keepax <ckeepax@opensource.cirrus.com> Link: https://lore.kernel.org/r/20200619212651.2739-1-david.rhodes@cirrus.com Signed-off-by: Mark Brown <broonie@kernel.org>
component has io_mutex, but it had been used at
snd_soc_component_update_bits_legacy() only which does read and write.
static int snd_soc_component_update_bits_legacy(...)
{
...
=> mutex_lock(&component->io_mutex);
...
old = snd_soc_component_read(...);
...
ret = snd_soc_component_write(...);
...
=> mutex_unlock(&component->io_mutex);
...
}
It is pointless if it is not used with both read and write functions.
This patch uses io_mutex correctly with read/write.
Here, xxx_no_lock() is local functions.
static int snd_soc_component_read(...)
{
...
=> mutex_lock(&component->io_mutex);
val = soc_component_read_no_lock(...);
=> mutex_unlock(&component->io_mutex);
...
}
static int snd_soc_component_write(...)
{
...
=> mutex_lock(&component->io_mutex);
ret = soc_component_write_no_lock(...);
=> mutex_unlock(&component->io_mutex);
...
}
static int snd_soc_component_update_bits_legacy(...)
{
...
=> mutex_lock(&component->io_mutex);
...
old = soc_component_read_no_lock(...);
...
ret = soc_component_write_no_lock(...);
...
=> mutex_unlock(&component->io_mutex);
...
}
ASoC: soc-component: merge snd_soc_component_read() and snd_soc_component_read32()
We had read/write function for Codec, Platform, etc,
but these has been merged into snd_soc_component_read/write().
Internally, it is using regmap or driver function.
In read case, each styles are like below
regmap
ret = regmap_read(..., reg, &val);
driver function
val = xxx->read(..., reg);
Because of this kind of different style, to keep same read style,
when we merged each read function into snd_soc_component_read(),
we created snd_soc_component_read32(), like below.
commit 738b49efe6c6 ("ASoC: add snd_soc_component_read32")
(1) val = snd_soc_component_read32(component, reg);
(2) ret = snd_soc_component_read(component, reg, &val);
Many drivers are using snd_soc_component_read32(), and
some drivers are using snd_soc_component_read() today.
In generally, we don't check read function successes,
because, we will have many other issues at initial timing
if read function didn't work.
Now we can use soc_component_err() when error case.
This means, it is easy to notice if error occurred.
This patch aggressively merge snd_soc_component_read() and _read32(),
and makes snd_soc_component_read/write() as generally style.
This patch do
1) merge snd_soc_component_read() and snd_soc_component_read32()
2) it uses soc_component_err() when error case (easy to notice)
3) keeps read32 for now by #define
4) update snd_soc_component_read() for all drivers
Because _read() user drivers are not too many, this patch changes
all user drivers.
Qiushi Wu [Sat, 13 Jun 2020 20:51:58 +0000 (15:51 -0500)]
ASoC: rockchip: Fix a reference count leak.
Calling pm_runtime_get_sync increments the counter even in case of
failure, causing incorrect ref count if pm_runtime_put is not called in
error handling paths. Call pm_runtime_put if pm_runtime_get_sync fails.
Fixes: fc05a5b22253 ("ASoC: rockchip: add support for pdm controller") Signed-off-by: Qiushi Wu <wu000273@umn.edu> Reviewed-by: Heiko Stuebner <heiko@sntech.de> Link: https://lore.kernel.org/r/20200613205158.27296-1-wu000273@umn.edu Signed-off-by: Mark Brown <broonie@kernel.org>
In trigger we already get the selected instance details
from runtime->private_data.So, removing the local
initialization which may corrupt the instance selected
details and this leads to corrupt data.
ASoC: Intel: broadwell: simplify card names for SOF uses
Blindly adding an sof- prefix to the card name is not user friendly
and causes UCM issues with a driver name truncated to 16 characters.
Simplify to use "sof-bdw <codec_name>" pattern for all Broadwell
machine drivers. The sof- prefix is added by the core. A generic "SOF"
driver name is used, and UCMv2 will detect the configuration for this
driver by testing the card name.
projects / mirror_ubuntu-hirsute-kernel.git / log