Smatch complains that the call to snd_soc_component_get_drvdata(component)
will dereference "component" so this NULL check is too late. The probe()
function will never be called with a NULL component pointer so just
delete the check.
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Link: https://patch.msgid.link/aScUKqCEhSIZoOmg@stanley.mountain
Signed-off-by: Mark Brown <broonie@kernel.org>
Now that cs_dsp uses regmap_raw_write() instead of regmap_raw_write_async()
it doesn't need to keep multiple DMA-safe buffers of every chunk of data
it wrote.
See commit fe08b7d508 ("firmware: cs_dsp: Remove async regmap writes")
Only one write can be in progress at a time, so one DMA-safe buffer can be
re-used. The single DMA-safe buffer is reallocated if the next write chunk
is larger. Reallocation size is rounded up to reduce the amount of churn.
PAGE_SIZE is used as a convenient size multiple. Typically for .wmfw files
the first chunk is the largest.
A DMA-safe intermediate buffer is used because we can't assume that the
bus underlying regmap can accept non-DMA-safe buffers.
Note that a chunk from the firmware file cannot simply be split into
arbitrarily-sized chunks to avoid buffer reallocation. The data in the
firmware file is preformatted to be written directly to the device as one
block. It already takes account of alignment, write-bursts and write
length requirements of the target hardware, so that the cs_dsp driver can
treat it as an opaque blob.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251126131501.884188-1-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The architectures supported by this driver have a maximum of 32-bits
of address, so we don't need more than 32-bits to store the length of
control data. Change the length in struct cs_dsp_coeff_ctl to an
unsigned int instead of a size_t. Also make a corresponding trivial
change to wm_adsp.c to prevent a compiler warning.
Tested on x86_64 builds this saves at least 4 bytes per control
(another 4 bytes might be saved if the compiler was inserting padding
to align the size_t).
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251124171536.78962-1-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Richard Fitzgerald <rf@opensource.cirrus.com>:
The CAL_SET_STATUS and CAL_DATA_RB controls are volatile and read-only, but
the existing ASoC macros to define controls don't allow setting access
permissions, so those controls were marked as non-volatile read/write.
These four patches fix that.
The first two patches add two new control macros to soc.h. I really don't
like codec drivers open-coding a kcontrol_new content for a control that
will be managed by the ASoC info/get/put handlers. If a new type of ASoC
control definition is needed it's better to have it in soc.h so all the
dependencies between ASoC and the kcontrol_new content are in one place.
Merge series from Charles Keepax <ckeepax@opensource.cirrus.com>:
This series adds an initial SDCA class driver, this consists of a
primary driver attached to the SoundWire device, and auxiliary drivers
representing each of the functions of the SDCA device. These drivers all
use the APIs added over the past series's to provide the class
functionality, as such these final drivers themselves are quite thin.
Also a few fix ups at the start of the series that have gathered up
whilst the last SDCA series was in review.
It is not uncommon for an SDCA Control to have no specified default
value in the DisCo. Non-volatile registers with no defaults will not be
present in the cache until they are accessed. However, if the first
operation user-space performs is a read whilst the device is runtime
suspended this read will fail.
To avoid such problems we should populate values from the hardware into
the cache for all non-volatile readable registers with no defaults.
Update the defaults handling to do this cache population since it is
iterating over the Controls and happens at a time the hardware is
always powered up.
Tested-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Reviewed-by: Maciej Strozek <mstrozek@opensource.cirrus.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Tested-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://patch.msgid.link/20251120153023.2105663-10-ckeepax@opensource.cirrus.com
Reviewed-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Richard Fitzgerald <rf@opensource.cirrus.com>:
These two patches improve the implementation of the debugfs files for
the wmfw and bin file names. First patch removes duplicated code. Second
patch replaces the old clunkiness of storing the filename with an appended
\n. The \n can be appended when the file is read, to keep the stored string
sane.
Use the new SND_SOC_BYTES_E_ACC() macro instead of SND_SOC_BYTES_E()
to define the CAL_DATA_RB control, so that it is marked as read-only
and volatile.
This avoids userland code expecting to be able to write to it and
then getting an unexpected EPERM error, or assuming that its value
can never change.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251120134437.1179191-5-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Use the new SOC_ENUM_EXT_ACC() macro instead of SOC_ENUM_EXT() to
define the CAL_SET_STATUS control, so that it is marked as read-only
and volatile.
This avoids userland code expecting to be able to write to it and
then getting an unexpected EPERM error, or assuming that its value
can never change.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251120134437.1179191-4-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a macro SND_SOC_BYTES_E_ACC() to allow the access permission flags
to be set. This is the same as SND_SOC_BYTES_E() but with an extra
argument for the access flags.
This will be used by the cs35l56.c driver to create a read-only
volatile byte control. It's preferable to avoid custom control macros
in codec drivers. Code maintenance is easier if all control macros are
defined together in soc.h.
This commit only creates this one macro that is actually going to be used.
There's no point cluttering soc.h with unused macros - that just adds a
maintenance burden. People can add equivalents for the other macros if
they need them.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251120134437.1179191-3-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a macro SOC_ENUM_EXT_ACC() to allow the access permission flags
to be set. This is the same as SOC_ENUM_EXT() but with an extra
argument for the access flags.
This will be used by the cs35l56.c driver to create a read-only
volatile enum. It's preferable to avoid custom control macros in codec
drivers. Code maintenance is easier if all control macros are defined
together in soc.h.
This commit only creates this one macro that is actually going to be used.
There's no point cluttering soc.h with unused macros - that just adds a
maintenance burden. People can add equivalents for the other macros if
they need them.
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251120134437.1179191-2-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
cs_dsp_debugfs_wmfw_read() and cs_dsp_debugfs_bin_read() were identical
except for which struct member they printed. Move all this duplicated
code into a common function cs_dsp_debugfs_string_read().
The check for dsp->booted has been removed because this is redundant.
The two strings are set when the DSP is booted and cleared when the
DSP is powered-down.
Access to the string char * must be protected by the pwr_lock mutex. The
string is passed into cs_dsp_debugfs_string_read() as a pointer to the
char * so that the mutex lock can also be factored out into
cs_dsp_debugfs_string_read().
wmfw_file_name and bin_file_name members of struct cs_dsp have been
changed to const char *. It makes for a better API to pass a const
pointer into cs_dsp_debugfs_string_read().
Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
Link: https://patch.msgid.link/20251120130640.1169780-2-rf@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Merge series from Bartosz Golaszewski <brgl@bgdev.pl>:
Problem statement: GPIOs are implemented as a strictly exclusive
resource in the kernel but there are lots of platforms on which single
pin is shared by multiple devices which don't communicate so need some
way of properly sharing access to a GPIO. What we have now is the
GPIOD_FLAGS_BIT_NONEXCLUSIVE flag which was introduced as a hack and
doesn't do any locking or arbitration of access - it literally just hand
the same GPIO descriptor to all interested users.
The proposed solution is composed of three major parts: the high-level,
shared GPIO proxy driver that arbitrates access to the shared pin and
exposes a regular GPIO chip interface to consumers, a low-level shared
GPIOLIB module that scans firmware nodes and creates auxiliary devices
that attach to the proxy driver and finally a set of core GPIOLIB
changes that plug the former into the GPIO lookup path.
The changes are implemented in a way that allows to seamlessly compile
out any code related to sharing GPIOs for systems that don't need it.
The practical use-case for this are the powerdown GPIOs shared by
speakers on Qualcomm db845c platform, however I have also extensively
tested it using gpio-virtuser on arm64 qemu with various DT
configurations.
The current mute_steam() implementation affects all paths instead of
only those in use by the DAI.
For example, playing to 2 DAIs simultaneously with mixing, stopping
one will mute the other.
Rework to use the same logic as hw_params() to mute only the relevant paths.
(also, use "rx->main_clk_users[j] > 0" instead of dsm_reg, which is
equivalent. I also don't think the clock enable should be in this function,
but that's a change for another patch)
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com>
Link: https://patch.msgid.link/20251117051523.16462-9-jonathan@marek.ca
Signed-off-by: Mark Brown <broonie@kernel.org>
"WSA_RX0_CLK",etc. are path clocks, but "WSA RX0",etc. are ports, and there
isn't a correspondence between the ports and paths.
For example "WSA RX0" port could be used by the "RX1 MIX" path.
The problem becomes obvious when RX4,etc. ports are added.
Enabling the path clocks should depend on the path being enabled.
With this fix, the main path clock will be enabled whenever the path is
active: previously using the mix ports only would only activate the mix
path clock and no audio would play.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Tested-by: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com>
Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com>
Link: https://patch.msgid.link/20251117051523.16462-4-jonathan@marek.ca
Signed-off-by: Mark Brown <broonie@kernel.org>
This commit adds a pre-check in ASoC to ensure there are non-zero CPU or
codec DAIs. It also handles early failure of memory allocation for
rtd->dais by going to the cleanup path, preventing potential issues
from invalid configurations or out-of-memory situations.
Signed-off-by: jempty.liang <imntjempty@163.com>
Link: https://patch.msgid.link/20251119101116.78676-1-imntjempty@163.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Linux supports NXP's LMM SCMI protocol so switch to using the appropriate
API. The SIPs were intended to act as placeholders until the support for
said protocol was upstreamed.
The underlying CPU protocol command from IMX_SIP_SRC_M_RESET_ADDR_SET is
replaced by a LMM protocol command with the same effect (i.e. setting the
boot address) since using the CPU protocol would require additional
permissions (which TF-A already had). Apart from this, the SIPs are
replaced by their equivalent Linux LMM commands.
Signed-off-by: Laurentiu Mihalcea <laurentiu.mihalcea@nxp.com>
Reviewed-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Link: https://patch.msgid.link/20251114143503.2139-1-laurentiumihalcea111@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Dan Carpenter
Richard Fitzgerald
Francesco Lavra
Alexander Stein
Mark Brown
HariKrishna Sagala
Charles Keepax
Pierre-Louis Bossart
Baojun Xu
Kuninori Morimoto
Jonathan Marek
Bartosz Golaszewski
jempty.liang
Laurentiu Mihalcea