Rewrite the handling of ASP1 TX mixer mux initialization to prevent a
deadlock during component_remove().
The firmware can overwrite the ASP1 TX mixer registers with
system-specific settings. This is mainly for hardware that uses the
ASP as a chip-to-chip link controlled by the firmware. Because of this
the driver cannot know the starting state of the ASP1 mixer muxes until
the firmware has been downloaded and rebooted.
The original workaround for this was to queue a work function from the
dsp_work() job. This work then read the register values (populating the
regmap cache the first time around) and then called
snd_soc_dapm_mux_update_power(). The problem with this is that it was
ultimately triggered by cs35l56_component_probe() queueing dsp_work,
which meant that it would be running in parallel with the rest of the
ASoC component and card initialization. To prevent accessing DAPM before
it was fully initialized the work function took the card mutex. But this
would deadlock if cs35l56_component_remove() was called before the work job
had completed, because ASoC calls component_remove() with the card mutex
held.
This new version removes the work function. Instead the regmap cache and
DAPM mux widgets are initialized the first time any of the associated ALSA
controls is read or written.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Fixes: 07f7d6e7a1 ("ASoC: cs35l56: Fix for initializing ASP1 mixer registers")
Link: https://lore.kernel.org/r/20240208123742.1278104-1-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
If the "spk-id-gpios" property is present it points to GPIOs whose
value must be used to select the correct bin file to match the
speakers.
Some manufacturers use multiple sources of speakers, which need
different tunings for best performance. On these models the type of
speaker fitted is indicated by the values of one or more GPIOs. The
number formed by the GPIOs identifies the tuning required.
The speaker ID must be used in combination with the subsystem ID
(either from PCI SSID or cirrus,firmware-uid property), because the
GPIOs can only indicate variants of a specific model.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Fixes: 1a1c3d794e ("ASoC: cs35l56: Use PCI SSID as the firmware UID")
Link: https://msgid.link/r/20240129162737.497-14-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Defer initializing the state of the ASP1 mixer registers until
the firmware has been downloaded and rebooted.
On a SoundWire system the ASP is free for use as a chip-to-chip
interconnect. This can be either for the firmware on multiple
CS35L56 to share reference audio; or as a bridge to another
device. If it is a firmware interconnect it is owned by the
firmware and the Linux driver should avoid writing the registers.
However, if it is a bridge then Linux may take over and handle
it as a normal codec-to-codec link. Even if the ASP is used
as a firmware-firmware interconnect it is useful to have
ALSA controls for the ASP mixer. They are at least useful for
debugging.
CS35L56 is designed for SDCA and a generic SDCA driver would
know nothing about these chip-specific registers. So if the
ASP is being used on a SoundWire system the firmware sets up the
ASP mixer registers. This means that we can't assume the default
state of these registers. But we don't know the initial state
that the firmware set them to until after the firmware has been
downloaded and booted, which can take several seconds when
downloading multiple amps.
DAPM normally reads the initial state of mux registers during
probe() but this would mean blocking probe() for several seconds
until the firmware has initialized them. To avoid this, the
mixer muxes are set SND_SOC_NOPM to prevent DAPM trying to read
the register state. Custom get/set callbacks are implemented for
ALSA control access, and these can safely block waiting for the
firmware download.
After the firmware download has completed, the state of the
mux registers is known so a work job is queued to call
snd_soc_dapm_mux_update_power() on each of the mux widgets.
Backport note:
This won't apply cleanly to kernels older than v6.6.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Fixes: e496112529 ("ASoC: cs35l56: Add driver for Cirrus Logic CS35L56")
Link: https://msgid.link/r/20240129162737.497-11-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The majority of runtime_suspend and runtime_resume handling
doesn't have anything specific to the ASoC driver, so can be
shared by the HDA driver. Move this code into the shared
library.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230721132120.5523-6-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The ASoC and HDA drivers have structures that contain some of the same
information - instead of maintaining two locations for this data the
drivers should share a common data structure as this will enable common
utility functions to be created.
The first step is to move the location of these members in the ASoC
driver.
Signed-off-by: Simon Trimmer <simont@opensource.cirrus.com>
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Acked-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20230721132120.5523-2-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
cs35l56_remove() always returns 0. Two of the functions that call
it are void and the other one should only return 0. So there's no
point returning anything from cs35l56_remove().
Signed-off-by: Simon Trimmer <simont@opensource.cirrus.com>
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230414133753.653139-6-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
dsp_ready_completion is redundant and can be replaced by a call
flush_work() to wait for cs35l56_dsp_work() to complete.
As the dsp_work is queued by component_probe() it must run before other
ASoC component callbacks and therefore there is no risk of calling
flush_work() before the dsp_work() has been queued.
Signed-off-by: Simon Trimmer <simont@opensource.cirrus.com>
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230414133753.653139-5-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The irq member was being set before calling the init function and then
cs35l56_irq_request() was called only when the init was successful.
However cs35l56_release() calls devm_free_irq() when the irq member is
set and therefore if init() fails then this will cause an attempted free
of an unallocated IRQ.
Instead pass the desired IRQ number to the cs35l56_irq_request()
function and set cs35l56->irq only when it has been successfully
allocated.
Signed-off-by: Simon Trimmer <simont@opensource.cirrus.com>
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/168147949598.26.711670799488943454@mailman-core.alsa-project.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Delete the 'removing' flag and don't kick init_completion to make a
quick cancel of dsp_work(). Just let it timeout on the wait for the
completion.
Simplify the code to standard cancelling or flushing of the work.
This avoids introducing corner cases from a layer of custom signalling.
It also avoids potential race conditions when system-suspend handling
is added.
Unless the hardware is broken, the dsp_work() will already have started
and passed the completion before the driver would want to cancel it.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/168122674746.26.16881587647873355224@mailman-core.alsa-project.org
Signed-off-by: Mark Brown <broonie@kernel.org>
This adds the main handling for system suspend but does not handle
re-patching the firmware after system resume.
This is a multi-stage suspend and resume because if there is a
RESET line it is almost certain that it will be shared by all the
amps. So every amp must have done its suspend before we can
assert RESET. Likewise we must de-assert RESET before the amps
can resume.
It's preferable to assert RESET before we turning off regulators, and
while they power up.
The actual suspend and resume is done by using the pair
pm_runtime_force_suspend() and pm_runtime_force_resume() to
re-use our runtime suspend/resume sequences.
pm_runtime_force_suspend() will disable our pm_runtime. If we were
runtime-resumed it calls our runtime_suspend().
pm_runtime_force_resume() re-enables pm_runtime and if we were
originally runtime-resumed before the pm_runtime_force_suspend()
it calls our runtime_resume(). Otherwise it leaves us
runtime-suspended.
The general process is therefore:
suspend() -> finish dsp_work and then run our runtime_suspend
suspend_late() -> assert RESET and turn off supplies
resume_early() -> enable supplies and de-assert RESET
resume() -> pm_runtime_force_resume()
In addition, to prevent the IRQ handler running in the period
between pm_runtime_force_suspend() and pm_runtime_force_resume()
the parent IRQ is temporarily disabled:
- from suspend until suspend_noirq
- from resume_noirq until resume
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230411152528.329803-6-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The CS35L56 combines a high-performance mono audio amplifier, Class-H
tracking inductive boost converter, Halo Core(TM) DSP and a DC-DC boost
converter supporting Class-H tracking.
Supported control interfaces are I2C, SPI or SoundWire.
Supported audio interfaces are I2S/TDM or SoundWire.
Most chip functionality is controlled by on-board ROM firmware that is
always running. The driver must apply patch/tune to the firmware
before using the CS35L56.
Signed-off-by: Simon Trimmer <simont@opensource.cirrus.com>
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230320112245.115720-9-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
