So far, forwarding the hw_params of the input to output relied on the
.hw_params() callback of the cpu side of the codec2codec link to be called
first. This is a bit weak.
Instead, override the stream params of the codec2codec to link to set it up
correctly.
ASoC: Intel: Skylake: Remove static table index when parsing topology
Currently when we remove and reload driver we use previous ref_count
value to start iterating over skl->modules which leads to out of table
access. To fix this just inline the function and calculate indexes
everytime we parse UUID token.
Colin Ian King [Fri, 26 Jul 2019 12:33:27 +0000 (13:33 +0100)]
ASoC: codec2codec: fix missing return of error return code
Currently in function snd_soc_dai_link_event_pre_pmu the error return
code in variable err is being set but this is not actually being returned,
the function just returns zero even when there are failures. Fix this by
returning the error return code.
Addresses-Coverity: ("Unused value") Fixes: 3dcfb397dad2 ("ASoC: codec2codec: deal with params when necessary") Signed-off-by: Colin Ian King <colin.king@canonical.com> Link: https://lore.kernel.org/r/20190726123327.10467-1-colin.king@canonical.com Signed-off-by: Mark Brown <broonie@kernel.org>
Peter Ujfalusi [Fri, 26 Jul 2019 06:42:44 +0000 (09:42 +0300)]
ASoC: ti: davinci-mcasp: Support for correct symmetric sample bits
Implement custom snd_pcm_hw_rule to filter the available formats for the
second stream to make it symmetric and allow only formats which require
the same amount of bits on the bus as the running stream.
A simple constraint is not working correctly because for example:
the first stream is started with S24_LE
If we place 24 as constraint for the SAMPLE_BITS then the second stream
can not use S24_LE as it is physically 32bits.
If we would place 32 as constraint (physical width) then S32_LE would have
been allowed, but S24_3LE is not.
The slot_width is a property for the bus while the constraint for
SNDRV_PCM_HW_PARAM_SAMPLE_BITS is for the in memory format.
Applying slot_width constraint to sample_bits works most of the time, but
it will blacklist valid formats in some cases.
With slot_width 24 we can support S24_3LE and S24_LE formats as they both
look the same on the bus, but a a 24 constraint on sample_bits would not
allow S24_LE as it is stored in 32bits in memory.
Implement a simple hw_rule function to allow all formats which require less
or equal number of bits on the bus as slot_width (if configured).
explains the issue of the patch.
While device is configured as 1-ch, hardware is still
generating a 2-ch stream.
When user space reads the data and assumes it is a 1-ch stream,
the rate will be slower by 2x.
Revert the change so 1-ch is not supported.
User space can selectively take one channel data out of two channel
if 1-ch is preferred.
Currently, both channels record identical data.
When running McASP as master capture alone will not record any audio unless
a parallel playback stream is running. As soon as the playback stops the
captured data is going to be silent again.
In McASP master mode we need to set the PDIR for the clock pins and fix
the mcasp_set_axr_pdir() to skip the bits in the PDIR registers above
AMUTE.
This went unnoticed as most of the boards uses McASP as slave and neither
of these issues are visible (audible) in those setups.
ASoC: codec2codec: deal with params when necessary
When there is an event on codec to codec dai_link, we only need to deal
with params if the event is SND_SOC_DAPM_PRE_PMU, when .hw_params() is
called. For the other events, it is useless.
Also, dealing with the codec to codec params just before calling
.hw_params() callbacks give change to either party on the link to alter
params content in .startup(), which might be useful in some cases
ASoC: codec2codec: name link using stream direction
At the moment, codec to codec dai link widgets are named after the
cpu dai and the 1st codec valid on the link. This might be confusing
if there is multiple valid codecs on the link for one stream
direction.
Instead, use the dai link name and the stream direction to name the
the dai link widget
When handling dai_link events on codec to codec links, run all .startup()
callbacks on sinks and sources before running any .hw_params(). Same goes
for hw_free() and shutdown(). This is closer to the behavior of regular
dai links
ASoC: cs47l15: Add codec driver for Cirrus Logic CS47L15
Adds the codec driver for the CS47L15 SmartCodec. This is a
multi-functional codec based on the Cirrus Logic Madera platform.
Signed-off-by: Richard Fitzgerald <rf@opensource.wolfsonmicro.com> Signed-off-by: Jaswinder Jassal <jjassal@opensource.wolfsonmicro.com> Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com> Link: https://lore.kernel.org/r/20190725163931.24964-2-ckeepax@opensource.cirrus.com Signed-off-by: Mark Brown <broonie@kernel.org>
Charles Keepax [Thu, 25 Jul 2019 16:39:29 +0000 (17:39 +0100)]
ASoC: wm_adsp: Allow bus error handler to be called directly
There is no need for end drivers to add helper functions to allow the
bus error handler to be called, simply update the prototype so it can be
called directly.
Peter Ujfalusi [Thu, 25 Jul 2019 08:34:32 +0000 (11:34 +0300)]
ASoC: ti: davinci-mcasp: Improve serializer handling in multi AXR setups
When multiple serializers are used we need to track the number of
serializers used by the other stream direction to avoid killing data lines
when the first stream used more serializers than the second would need.
We are still protected against the case when the second stream uses more
serializers which had affected the running stream as well.
To take advantage of the improved serializer logic we need to modify the
channel constraints rule as well to allow the use of multiple serializers
for the second stream as additional ones will not affect the FS/BCLK on
the bus.
With removal of MCPS, CPS and CPC ambiguity, ibs and obs params for
struct skl_module_cfg have been left unused. Update struct declaration
by removing these two.
ASoC: Intel: Skylake: Make MCPS and CPS params obsolete
As per FW Interface Modules Configuration, init instance IPC request
requires base initial module configuration. This configuration structure
is made of:
- cpc (chunks per cycle)
- ibs (input buffer size)
- obs (output buffer size)
- is_pages (memory pages required)
- audio_fmt (self explanatory)
Skylake topology accepts following tokens: MCPS, CPS and CPC. All of
these are directly connected. Moreover, assigning one of these allows
to calculate the remaining two. In simplest scenario and assuming 1ms
scheduling, following is true:
CPS = CPC times 1000
MCPS = CPS times 1000 000
Note: these calculations vary depending on scenario and scheduling
requirements.
Given the current implementation, userspace is allowed to provide
different values for all three causing informational chaos. On top of
that, struct skl_base_cfg which represents base module configuration,
incorrectly takes CPS param instead of CPC.
This ambiguity may lead to user unintentionally providing improper
values to DSP firmware and thus impacting module scheduling in
unexpected fashion. Fix by making MCPS and CPS topology params obsolete
and relying solely on CPC value.
ASoC: Intel: Skylake: Do not disable FW notifications
As per FW team recommendation we should not disable notifications.
By default, all notifications are enabled in DSP firmware. These
notifications provide a vital information whenever an error occurs.
Currently, driver disables them during boot sequences. By doing so,
Skylake may silently ignore severe stream errors.
Correct that by removing permissive code.
ASoC: Intel: Skylake: Remove memory available check
Current memory availability check is a stub, while actual memory
management takes place in firmware. Leave this task to firmware entirely
and remove redundant code.
The entire logic for MCPS calculation and DSP scheduling is found
within DSP firmware. Currently driver implements simplistic, inaccurate
logic itself which may prevent pipeline creation despite firmware being
completely fine its parameters.
Remove that logic and leave the MCPS calculation to DSP alone.
ASoC: Intel: Skylake: Combine snd_soc_skl_ipc and snd_soc_skl
As both modules are core part of Skylake driver and none can live
without the other, combine snd_soc_skl_ipc and snd_soc_skl.
It's highly probable IPC module was to be treated as an interface for
platform specific code implementations e.g.: possibility of existence of
BXT specific code without SKL one. However, most funtionalities are
being inherited from one DSP firmware to another, and thus this
assumption fails.
skl-sst, bxt-sst and cnl-sst are not individuals pointing respectively
to SKL (cAVS 1.5), BXT (cAVS 1.5+) & CNL (cAVS 1.8) standalone
implementations. Code found within these is shared among all platforms
whenever necessary to avoid code duplication and reduce development
burden.
Merge also helps in cleaning up internal code in future changes.
ASoC: Intel: Skylake: Merge skl_sst and skl into skl_dev struct
Skylake driver is divided into two modules:
- snd_soc_skl
- snd_soc_skl_ipc
and nothing would be wrong if not for the fact that both cannot exist
without one another. IPC module is not some kind of extension, as it is
the case for snd_hda_ext_core which is separated from snd_hda_core -
legacy hda interface. It's as much core Skylake module as snd_soc_skl
is.
Statement backed up by existence of circular dependency between this
two. To eliminate said problem, struct skl_sst has been created. From
that very momment, Skylake has been plagued by header errors (incomplete
structs, unknown references etc.) whenever something new is to be added
or code is cleaned up.
As this design is being corrected, struct skl_sst is no longer needed,
so combine it with struct skl. To avoid ambiguity when searching for skl
stuff (struct skl *skl) it has also been renamed to skl_dev.
ASoC: Fail card instantiation if DAI format setup fails
If the DAI format setup fails, there is no valid communication format
between CPU and CODEC, so fail card instantiation, rather than continue
with a card that will most likely not function properly.
ASoC: soc-dai: move snd_soc_dai_stream_valid() to soc-dai.c
snd_soc_dai_stream_valid() is function to check stream validity.
But, some code is using it, some code are checking stream->channels_min
directly. Doing samethings by different method is confusable.
This patch uses same funcntion for same purpose.
Current ALSA SoC is directly using dai->driver->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_compress_new() and use it.
Current ALSA SoC is directly using dai->driver->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_remvoe() and use it.
Current ALSA SoC is directly using dai->driver->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_probe() and use it.
Current ALSA SoC is directly using dai->driver->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_resume() and use it.
Current ALSA SoC is directly using dai->driver->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_suspend() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_delay() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_bespoke_trigger() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_trigger() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_prepare() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_shutdown() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_startup() and use it.
Current ALSA SoC is directly using dai->driver->ops->xxx,
thus, it has deep nested bracket, and it makes code unreadable.
This patch adds new snd_soc_dai_hw_free() and use it.
Sometimes ALSA SoC naming is very random.
Current soc_dai_hw_params() should use snd_soc_dai_xxx() style.
And then, 1st parameter should be dai. Otherwise it is confusable.
- soc_dai_hw_params(..., dai);
+ snd_soc_dai_hw_params(dai, ...);
Current ALSA SoC has many snd_soc_dai_xxx() function which is
using dai->driver->ops->xxx.
But, some of them are implemented as snd_soc_dai_xxx(),
but others are directly using dai->driver->ops->xxx.
Because of it, the code is not easy to read.
This patch creats new soc-dai.c and moves snd_soc_dai_xxx()
functions into it.
One exception is snd_soc_dai_is_dummy() which is based on
soc-utils local variable. We need to keep it as-is there.
Others which is directly using dai->driver->ops->xxx will be
implemented at soc-dai.c by incremental patches.
Timo Wischer [Mon, 22 Jul 2019 07:24:01 +0000 (16:24 +0900)]
ASoC: rsnd: Support hw_free() callback at DAI level
This patch provides the needed infrastructure to support calling hw_free()
at the DAI level. This is for example required to free resources allocated
in hw_params() callback.
The modification of __rsnd_mod_add_hw_params does not have any side
effects because rsnd_mod_ops::hw_params callback is not used by anyone
until now.
Rander Wang [Mon, 22 Jul 2019 14:13:59 +0000 (09:13 -0500)]
ASoC: SOF: Intel: hda: fix link DMA config
For this bug, there are two capture pcm streams active, with one
stream and its related stream tag released before suspend. Later
when system suspend is done, the stream tag for the remaining
active stream is released by SOF driver. After system resume, hda
codec driver restores the stream tag for the active pcm stream,
but SOF goes to assign a new one, which now doesn't match with the
stream tag used by codec driver, and this causes DMA to fail
receiving data, leading to unrecoverable XRUN condition in FW.
For stream tag is stored in both hda codec and SOF driver, it
shouldn't be released only in SOF driver. This patch just keeps the
stream information in dma data and checks whether there is a stored
DMA data for stream resuming from S3 and restores it. And it also
removes DMA data when the stream is released.
Shengjiu Wang [Thu, 11 Jul 2019 10:49:46 +0000 (18:49 +0800)]
ASoC: fsl_esai: recover the channel swap after xrun
There is chip errata ERR008000, the reference doc is
(https://www.nxp.com/docs/en/errata/IMX6DQCE.pdf),
The issue is "While using ESAI transmit or receive and
an underrun/overrun happens, channel swap may occur.
The only recovery mechanism is to reset the ESAI."
This issue exist in imx3/imx5/imx6(partial) series.
In this commit add a tasklet to handle reset of ESAI
after xrun happens to recover the channel swap.
Janusz Jankowski [Mon, 22 Jul 2019 14:14:02 +0000 (09:14 -0500)]
ASoC: SOF: Intel: ssp: BCLK delay parameter
Some codecs require BCLK to be on for some time, before sending
any data. SOF can enable BCLK and then wait for guaranteed time,
before starting DMA on SSP start.
Kai Vehmanen [Mon, 22 Jul 2019 14:13:58 +0000 (09:13 -0500)]
ASoC: SOF: Intel: hda: reset link DMA state in prepare
When application goes through SUSPEND/STOP->PREPARE->START
cycle, we should always reprogram the DAI link DMA to ensure
it is in sync with the host PCM DMA.
Use same state tracking logic to handle both restart and
system resume flows. Use link_prepared field of
'struct hdac_ext_stream' to store the state, instead of
adding redundant fields to SOF specific structs.
ASoC: SOF: Intel: hda: add a parameter to disable MSI
Enabling MSI on HDA can fail, in which case the legacy PCI IRQ mode
will be used. To make testing this mode easier add an "enable_msi"
module parameter, which is only enabled if debugging is enabled too.
ASoC: SOF: Intel: hda: use SOF defined init chip in resume
Unify resume code by using SOF common function hda_dsp_ctrl_init_chip()
which can handle both HDA and non-HDA cases. Move code to reset
stream-to-link mapping into hda_dsp_ctrl_init_chip().
ASoC: SOF: Intel: hda: set position buffer in init chip
Set the HDA stream position buffer during init chip. The position buffer
needs to be set in both HDA codec and nocodec cases. Using SOF defined
function and move it to common code.
Rander Wang [Mon, 22 Jul 2019 14:13:53 +0000 (09:13 -0500)]
ASoC: SOF: Intel: hda: Enable jack detection
In commit 7d4f606c50ff ("ALSA: hda - WAKEEN feature enabling for
runtime pm"), legacy HD-A driver sets hda controller in reset mode after
entering runtime-suspend. And when resuming from suspend mode, it checks
hda controller & codec status to detect headphone hotplug event. Now
this patch does the same job in SOF runtime pm functions.
And we need to check all the non-hdmi codecs for some cases like playback
with HDMI or capture with DMIC connected to dsp. In these cases, only
controller is active and codecs are suspended, so codecs can't send
unsolicited event to controller. The jack polling operation will activate
codecs and unsolicited event can work even codecs become suspended later.
Tested on whiskylake with hda codecs.
Signed-off-by: Rander Wang <rander.wang@linux.intel.com> Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20190722141402.7194-13-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown <broonie@kernel.org>
FW encapsulates information about section types (e.g DRAM, IRAM)
inside module block header. This information can be used in order
to correctly load the section to the appropriate place in memory.
SOF Linux driver needs to know for each platform how to map the
section type with the corresponding memory BAR. So, this patch
introduces get_bar_index, a new operation inside snd_sof_dsp_ops.
Intel platforms, usually load all the section in a contiguous memory
area (usually denoted by sdev->mmio_bar) so things are relatively
simple there. Anyhow, on i.MX8 IRAM and DRAM for example are mapped
to distinct BARs.
By default, if no get_bar function is provided the core implementation
will always return sdev->mmio_bar so that there will be no need for
a change to existing Intel code.
Kai Vehmanen [Mon, 22 Jul 2019 14:13:46 +0000 (09:13 -0500)]
ASoC: SOF: core: increase default IPC timeouts
Increase the default timeout values for boot (100ms to 2sec) and
IPC message sending (5ms to 500ms). The values should be overridden
with values from platform data.
There is no functional need to have such short timeouts as both boot
and IPC send errors are considered fatal errors. More relaxed timeouts
are convenient when running the driver on top of emulation such as QEMU.
Kai Vehmanen [Mon, 22 Jul 2019 14:13:45 +0000 (09:13 -0500)]
ASoC: SOF: ipc: use timeout configured at probe
Do not hardcode IPC timeout value in ipc.c, but rather use the timeout
value configured during device probe. For platforms that do not override
the IPC timeout, default value TIMEOUT_DEFAULT_IPC_MS has already been
defined in core.c.
Kai Vehmanen [Mon, 22 Jul 2019 14:13:43 +0000 (09:13 -0500)]
ASoC: SOF: reset DMA state in prepare
When application goes through SUSPEND/STOP->PREPARE->START
cycle, we should always reprogram the SOF device to start
DMA from a known state so that hw_ptr/appl_ptrs remain valid.
This is expected by ALSA core as it resets the buffer
state as part of prepare (see snd_pcm_do_prepare()).
Fix the issue by forcing reconfiguration of the FW with
STREAM_PCM_PARAMS in prepare(). Use combined logic to handle
prepare and the existing flow to reprogram hw-params after
system suspend.
Without the fix, first call to pcm pointer() will return
an invalid hw_ptr and application may immediately observe XRUN
status, unless "start_threshold" SW parameter is set to maximum
value by the application.
Pan Xiuli [Mon, 22 Jul 2019 14:13:42 +0000 (09:13 -0500)]
ASoC: SOF: pci: mark last_busy value at runtime PM init
If last_busy value is not set at runtime PM enable, the device will be
suspend immediately after usage counter is 0. Set the last_busy value to
make sure delay is working at first boot up.
Downgrade "nothing to do in IRQ thread" message from error to a debug
message in the IPC interrupt handler thread.
The spurious wake-up can happen if a HDA stream interrupt is
raised while the IPC interrupt thread is running. IPC functionality
is not impacted by this condition, so debug is a more appropriate
trace level.
Stephan Gerhold [Mon, 22 Jul 2019 13:03:52 +0000 (15:03 +0200)]
ASoC: qcom: apq8016_sbc: Fix oops with multiple DAI links
apq8016_sbc_parse_of() sets up multiple DAI links, depending on the
number of nodes in the device tree. However, at the moment
CPU and platform components are only allocated for the first link.
This causes an oops when more than one link is defined:
Move the allocation inside the loop to ensure that each link is
properly initialized.
Fixes: 98b232ca9e0e ("ASoC: qcom: apq8016_sbc: use modern dai_link style") Signed-off-by: Stephan Gerhold <stephan@gerhold.net> Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com> Link: https://lore.kernel.org/r/20190722130352.95874-1-stephan@gerhold.net Signed-off-by: Mark Brown <broonie@kernel.org>
Wenwen Wang [Mon, 22 Jul 2019 13:57:44 +0000 (08:57 -0500)]
ASoC: dapm: fix a memory leak bug
In snd_soc_dapm_new_control_unlocked(), a kernel buffer is allocated in
dapm_cnew_widget() to hold the new dapm widget. Then, different actions are
taken according to the id of the widget, i.e., 'w->id'. If any failure
occurs during this process, snd_soc_dapm_new_control_unlocked() should be
terminated by going to the 'request_failed' label. However, the allocated
kernel buffer is not freed on this code path, leading to a memory leak bug.
To fix the above issue, free the buffer before returning from
snd_soc_dapm_new_control_unlocked() through the 'request_failed' label.
ASoC: sgtl5000: Improve VAG power and mute control
VAG power control is improved to fit the manual [1]. This patch fixes as
minimum one bug: if customer muxes Headphone to Line-In right after boot,
the VAG power remains off that leads to poor sound quality from line-in.
I.e. after boot:
- Connect sound source to Line-In jack;
- Connect headphone to HP jack;
- Run following commands:
$ amixer set 'Headphone' 80%
$ amixer set 'Headphone Mux' LINE_IN
Change VAG power on/off control according to the following algorithm:
- turn VAG power ON on the 1st incoming event.
- keep it ON if there is any active VAG consumer (ADC/DAC/HP/Line-In).
- turn VAG power OFF when there is the latest consumer's pre-down event
come.
- always delay after VAG power OFF to avoid pop.
- delay after VAG power ON if the initiative consumer is Line-In, this
prevents pop during line-in muxing.
According to the data sheet [1], to avoid any pops/clicks,
the outputs should be muted during input/output
routing changes.
ASoC: rockchip-max98090: Remove MICBIAS as supply of input pin IN34
Commit ec0d23b295b9 ("ASoC: rockchip-max98090: Fix the Headset Mic
route.") moved the MICBIAS widget to supply Headset Mic but forget to
remove the MICBIAS widget to supply IN34 which is not really needed, so
remove that path so we have:
Shuming Fan [Fri, 19 Jul 2019 06:32:49 +0000 (14:32 +0800)]
ASoC: rt1308: add silence detection and manual PDB control
We enable the silence detection function in initial settings.
PDB control changes to manual mode, hence the driver could
fully control the AMP output on/off.
Lucas Stach [Fri, 19 Jul 2019 14:36:37 +0000 (16:36 +0200)]
ASoC: tlv320aic31xx: suppress error message for EPROBE_DEFER
Both the supplies and reset GPIO might need a probe deferral for the
resource to be available. Don't print a error message in that case, as
it is a normal operating condition.