This patch adds a matrix arithmetic library start. Currently it supports
matrix initialization, clear, set values, get values, and a multiply
function for 16 bit fractional format. Multiplication functions exist
for normal matrix multiply and element-wise multiplication.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
This fix replaces the previous copier_copy() implementation there
some required input/output combinations were not supported.
TODO: multichannel ALH case when single stream is produced/consumed
by multiple gateways is NOT IMPLEMENTED by this fix. However,
it was not (properly) implemented in the code being replaced.
So the fix does not bring any loss here.
Signed-off-by: Serhiy Katsyuba <serhiy.katsyuba@intel.com>
No runtime semantic change. Use C library when RTOS uses
C library otherwise use own C library calls.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
No functional runtime change, but changes to rtos partitioning and the
layout of headers .
This patch creates RTOS specifc header paths and updates spinlock.h
and kernel.h to show the new usage. Other headers will incrementally follow.
It reuses the current zephyr topleve directory and creates a new
toplevel xtos directory for xtos specific files.
Due to the mixing of RTOS, driver and library headers at the top level include
directory it was necessary to create rtos specific header directories i.e.
src/include/rtos-xtos
src/include/rtos-zephyr
These RTOS include directories will eventually contain RTOS specific headers
whilst common logic and structures will be placed in non RTOS directories.
This will also mean
"#include <sof/spinlock.h>"
will become
"#include <rtos/spinlock.h>"
and will allow easier visualisation of where and why RTOS headers are being used.
This will help to eliminate cross usage of headers between RTOSes.
Subsequqnt patches will move more headers and rtos specific wrppaer
source files into rtos specific locations.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
With the release of RTNR version 20220728, period for capturing pipeline requires at least 5ms.
Periods setting for 16KHz capturing pipeline is also removed since RTNR is currently running in 48KHz capturing pipeline only.
Signed-off-by: Ming Jen Tai <mingjen_tai@realtek.com>
Note this change does NOT affect Zephyr builds in any shape or form as
Zephyr builds simply don't use the CMake files changed by this commit.
Downloading missing submodules at build time was never a good idea;
always a hack. Downloading and building should always be kept separate
from each other for version control and bill of materials reasons and to
support "off-line" builds; build inputs should always be available.
Now that Zephyr builds have just moved away from git submodules
(replaced by west), stop sneakily downloading missing submodules at
build time while the user cannot notice, overwhelmed by the volume of
build logs. Someone building XTOS first and Zephyr second could
unknowningly end up in a "hybrid" and undesired situation where git
submodules and west would be BOTH pointing at the same rimage clone.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Tool to stream data from Linux SOF driver "mtrace" debugfs
interface to standard output. Plain "cat" is not sufficient
as each read() syscall returns log data with a 32bit binary
header, containing the payload length.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Enable the Zephyr based mtrace backend for all Intel IPC4 platforms.
Also enable routing SOF logs to Zephyr and bump the default
logging level to INFO.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Handle IPC4_ENABLE_LOGS basefw message. If firmware is built with
Zephyr and CONFIG_LOG_BACKEND_ADSP_MTRACE is defined, initialize
the mtrace logging backend.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Extend the EDF scheduler adaptation to Zephyr to allow repeating
tasks. This allows SOF code to use the per core threads used by
EDF for more usages.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Recommendation from Zephyr infrastructure is to use a specific
version of the docker-image. This time we need this to upgrade
to SDK 0.15, which is a requirement to use latest main of
Zephyr. The current "latest" tag does not have 0.15 SDK.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Update zephyr to commit dcda3eab8df7 ("tests: net: socket: tcp: move to
new ztest API").
Needed to bring in support for Intel mtrace logging backend.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
comp_dbg() statements may be compiled out, so having a 'dev'
variable just for debugs is not useful and just ends up with
compiler warnings on unused variables.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
At the moment CONFIG_INCOHERENT is automatically enabled when
CONFIG_TIGERLAKE is set and cannot be switched from menuconfig.
Enable that to make debugging with limited cache use possible.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Cppcheck found this issue:
src/include/sof/math/fft.h:133:51: error: Shifting by a negative
value is undefined behaviour [shiftNegative]
output->real = sat_int16((((int32_t)input->real >> n_rnd) + 1) >> 1);
^
The case n == 0 need to be handled with the upper if branch non-rounded
shift version.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
Cppcheck finds this issue
src/math/fft/fft_16.c:23:27: warning: Either the condition '!plan'
is redundant or there is possible null pointer dereference:
plan. [nullPointerRedundantCheck]
struct icomplex16 *inb = plan->inb16;
There's same issue with inb and outb in both fft_16.c and fft_32.c
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
Fix compiler error related to endpoint field:
src/audio/copier/copier.c:662:13: error: the comparison will always
evaluate as 'true' for the address of 'endpoint' will never be
NULL [-Werror=address]
662 | if (cd->endpoint) {
| ^~
src/include/ipc4/copier.h:325:26: note: 'endpoint' declared here
325 | struct comp_dev *endpoint[IPC4_COPIER_MODULE_OUTPUT_PINS_COUNT];
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
PCM capture followed by PCM playback would not work if
done immediately (less than 2s, sof suspend timeout)
Signed-off-by: Laurentiu Mihalcea <laurentiu.mihalcea@nxp.com>
In a topology with two connected pipelines, e.g. two playback
pipelines, connected on a mixer, if one PCM is already streaming and
then the other one is started too, two bugs trigger:
1. when adding pipelines to the scheduler, all pipelines are added,
including any common pipelines behind the mixer
2. the mixer pipeline then gets to run first, while it has no data
available to it yet.
To fix these bugs
1. don't re-add any active pipelines
2. when a pipeline is added for streaming, let it begin copying
immediately instead of skipping a period.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
This patch adds cmocka test cases for window functions library.
The calculated windows are compared with RMS error criteria
to Matlab calculated windows in header files. The header files
are directly generated by script ref_window.m.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
This patch adds a library for window functions. The window functions
are used to mitigate artefacts (from nature of FFT) to spectra when
computing FFT for a finite long chunk of audio. The currently provided
window functions are Blackman, Hamming, Povey, and rectangular.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
This is another tplg_parser and testbench update on the road to provide
further flexibility around supporting new modules and IPC versions.
There will more to follow.
Changes are mostly mechanical code movements i.e. moving testbench related code
to the testbench, splitting the files into per module/component files and
making some functions static and public.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
On host based builds we sometimes see that panic.h redefines assert().
Fix this. This will go away in the future as Zephyr will fully take care
of assert() on host.
Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
An IPC4 widget may support multiple I/O queue. Queue ID is
required during widget binding in route setup.
This patch adds sink_pin_binging and src_pin_binding classes
for multiple I/O queue support.
For widget requires special pin(queue) binding, sink_pin_binding
or src_pin_binding should be defined in the widget object. So
that the kernel knows which pin to use during widget binding
by parsing topology.
Note that pin binding should be defined for all sink/source pins
or for none. Even if there is only one sink pin requires special
pin binding, pin binding should be defined for all sink pins, and
for pins that are not used, give the value "NotConnected". This rule
applies to source pins, too.
Example: smart_amp widget has two sink pins and one source pin,
and its feedback sink pin requires special pin binding, so we have to
define sink_pin_binding objects within smart_amp object like this:
Object.Base.sink_pin_binding.1 {
# the main sink pin is connected to host for playback data
sink_pin_binding_wname "copier.host.1.1"
}
Object.Base.src_pin_binding.2 {
# the feedback sink pin is connected to capture dai for codec feedback.
sink_pin_binding_wname "copier.SSP.2.1"
}
In this example, smart_amp is connected to host and DAI directly,
it is also possible to be connected to host and DAI indirectly (with
gain or other widget in the middle).
Signed-off-by: Chao Song <chao.song@linux.intel.com>
These two tokens will be used by the kernel to decide the
supported number of sink/source pins, thus to allocate
correct widget queue id in route setup.
These two tokens are mandatory for IPC4 widgets, because
the default value in Linux kernel is zero, which is invalid
for queue id allocation.
Signed-off-by: Chao Song <chao.song@linux.intel.com>
Restores ability to compile on Windows with MSYS.
Fixes commit dcf0577a77 ("logger: allow starting before the driver is
loaded")
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Usually when processing COMP_TRIGGER_PAUSE or COMP_TRIGGER_STOP if a
component returns PPL_STATUS_PATH_STOP it means, that the trigger
propagation should stop and the pipeline should be kept alive.
However, in some cases propagation should be stopped, but the
pipeline should be terminated. This is the case with the selector
component, when used in a keyword detection pipeline. Add a new
PPL_STATUS_PATH_TERMINATE return value for such cases.
BugLink: https://github.com/thesofproject/sof/issues/5398
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Moved hardcoded cycles values for DMIC IO from dmic.h header
to per-platform configuration file (defconfig files for
non-Zephyr builds and app/boards for Zephyr builds).
Added new entry to intel drivers Kconfig - CONFIG_DMIC_HW_IOCLK.
Modification of this clock value may be used for testing purposes
like building firmware for FPGA or simulator.
Signed-off-by: Andrey Borisovich <andrey.borisovich@intel.com>
It was a typo while adding libutil.a to CMakeLists.txt.
This commit unifies the naming with other platforms.
Signed-off-by: Ming Jen Tai <mingjen_tai@realtek.com>
This reverts commit a0b832686c.
There was a PR merge "conflict" as the log defines we're in different
place in the file. This same log define was already added in commit:
2c4e403d4d
So remove the duplicate.
Signed-off-by: Jaska Uimonen <jaska.uimonen@intel.com>
This patch adds tests for 16-bit FFT library version. The test SNR
is improved by using sine frequency that matches with length of FFT.
It allows to get highest SNR in transform without need for window
functions in test and attenuation in the signal.
The 32 bit FFT tests are updated to use the same signal generator.
With this criteria the minimum SNRs for 32 bit to pass could be
increased to 95 - 132 dB.
Also tests were added for 512 size FFT in addition to 256 and 1024.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
This patch adds lower RAM and lower MCPS consuming FFT
for 16 bit data. The twiddle factors save from .bss
4096 bytes. For e.g. 512 bin FFT the saving for runtime is
additional 4096 bytes.
The previous fft.c has been split to fft_common.c and fft_32.c for
32 bits version. The new 16 bit version is in fft_16.c. The name
of previous fft_execute() is changed to fft_execute_32(). Common
function fft_plan_new() takes FFT word length 16 or 32 as new
last argument.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
This patch adds script tools/tune/fft/export_twiddle.m. Since
the data in original twiddle.h differs by one LSB in few
items the file is replaced with one generated with the script.
The identifier 32 is added to header name and variables names
as preparation to add 16 bit FFT version.
Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
In mixout_prepare() mixin->pipeline can be NULL, check it before
dereferencing.
BugLink: https://github.com/thesofproject/sof/issues/6118
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Remove mt8186 unused topologies.
We use sof-mt8186 on our demo board, sof-mt8186-mt6366-rt1019-rt5682s
and sof-mt8186-mt6366-da7219-max98357 will no be used anymore.
Signed-off-by: Chunxu Li <chunxu.li@mediatek.com>
Remove duplicate configs with base IPC4 overlay and add PM related
configs to be able to compile upstream Zephyr SSP changes.
Also enable ALH native driver.
Signed-off-by: Jaska Uimonen <jaska.uimonen@intel.com>
Upstream zephyr has removed "Label" from xtensa device tree definitions
thus device_get_binding stopped to work for sof. So change the code to
use device tree macros.
Signed-off-by: Jaska Uimonen <jaska.uimonen@intel.com>
Seppo Ingalsuo