Commit Graph

7640 Commits

Author SHA1 Message Date
Lionel Koenig c48a5281a0 topology1: google-aec topology
Add topology to include google-rtc-audio-processing to a main topology.

Signed-off-by: Lionel Koenig <lionelk@google.com>
2022-01-26 13:48:00 +00:00
Xiaoyun Wu(Iris) 12675d51ff configs: Configure converter in tglIPC4 config
When build tgl firmware with IPC4
configuration, disable HIFI3 conv-
erter and enable format converters.

Signed-off-by: Xiaoyun Wu(Iris) <xiaoyun.wu@intel.com>
2022-01-26 13:47:38 +00:00
Seppo Ingalsuo 9ca2f7655e Audio: Buffer: Remove too verbose buffer debug traces for CONFIG_LIBRARY
This patch leaves out low value debug traces with CONFIG_LIBRARY for
comp_update_buffer_produce() and comp_update_buffer_consume(). They
are too far too verbose and slow down testbench runs, and cause
massive size data files if traces are captured to files.

The change is done by defining bug_dbg() macro as empty. The variable
addr needed to be eliminated because it is used only for trace.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-26 13:37:49 +00:00
Rander Wang 8ed3bce0ea ipc4: add general S24_3LE support
S24_3LE is not a native supported format for hda dma. Driver
treats it as 16bit stream or 8bit stream, .e.g. 24bit 3LE
2ch stream will be set as 16bit 3ch stream. In FW side,
24bit stream will be converted to S24LE.

Currently HIFI3 optimization is not supported now and also
S24_LE unit test is not implemented. These will be supported
later.

Tested on windows platform.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-26 13:28:31 +00:00
Rander Wang 46d718e6c8 ipc4: add pcm conversion support for link dma & alh
For HDA gateway, 24bits sample will be in high 24bits with
32 bits container case. Alh requires 32 bits container and
all sample bits starting from bit 31 to bit 0.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-26 13:28:31 +00:00
Rander Wang 10a5bdb829 ipc4: change frame_fmt and valid_sample_fmt for gateway
First we reused ipc3 pcm conversion function by by setting
frame_fmt to S24_4LE, but later found that we need different
conversion for copier with gateways support.

Now frame_fmt will be set by default config.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-26 13:28:31 +00:00
Rander Wang 458bda34b4 ipc4: adjust pcm conversion function for gateway
Compared to non-copier module, copier module uses
different conversion algorithoms for different
gateways, .e.g HDA dma requires sample bit starting
from MSB bits and also alh engine. These hw features
make pcm conversion algorithom different for different
modules.

This patch add gateway type to select conversion function.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-26 13:28:31 +00:00
Rander Wang c7b9df0106 ipc4: fix compiling issue for ipc4
Move endif to include general ipc4 conversion.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-26 13:28:31 +00:00
Rander Wang ffffa4b0bd ipc4: remove redundant calculation for copy info
The src & sink info can be gotten from struct comp_copy_limits,
no need to calculate it again.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-26 13:28:31 +00:00
Tinghan Shen 59a32d5bbe platform: Add mt8186 to Kconfig
Add mt8186 to Kconfig

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Tinghan Shen 7c8e272a2a platform: Add build support for mtk mt8186 platform
Add CMakefile for mtk mt8186 platform to build.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Tinghan Shen eff005f8b0 config: Add default config for mt8186
Add default config for mt8186 platform.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Tinghan Shen c46186be23 platform: mtk: Add support for mt8186 platform
mt8186 platform integrates a single-core HIFI5 DSP.

The DSP operates at different highest clock frequencies with corresponding
voltage, 800MHz(0.8v)/400MHz(0.7v)/300MHz(0.65v)/26MHz(0.6v).

The DSP core and DMA in DSP subsys both can access DRAM directly.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Tinghan Shen d4df1f647b drivers: mtk: Add drivers for mt8186
Add interrupt, timer, and ipc drivers for mt8186

mt8186 DSP has 32 interrupts.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Tinghan Shen 24e1b7f6ca platform: mtk: add mt8186 memory layout and register definitions
Add memory layout and register address for mtk mt8186

Cache
  I-Cache: 64KB, 4-way Associativity
  D-Cache: 128KB, 4-way Associativity

External Memory
  DRAM: DSP can access DRAM shared with CPU
  L2TCM: 512KB DSP SRAM POOL

Currently, use
phy addr:0x60000000, size:0x1000000
dma phy addr: 0x61000000, size: 0x0100000

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Tinghan Shen c6ea9b9c20 platform: mtk: add xtensa headers for mt8186
Add xtensa headers for mtk mt8186 platform.

Signed-off-by: Tinghan Shen <tinghan.shen@mediatek.com>
Signed-off-by: Allen-KH Cheng <allen-kh.cheng@mediatek.com>
2022-01-26 12:24:24 +00:00
Seppo Ingalsuo 7d05389b63 Tools: Tune: EQ: Add legend to frequency response plot
This patch adds legend to plot since it was difficult to see to
which play/capture channel which curve corresponded to. In script
selftest case a label called Reference is added to the legend.

Also an indentation issue is fixed in the script.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-25 15:53:09 +00:00
Seppo Ingalsuo 6e2084ee8e Tools: Tune: EQ: In mls_freq_resp wait longer for capture to end
The 3s wait was not enough long time to wait for recording to complete
before copy for analysis so the wait is increased to 5s. The total
recording lengths were variable and not same 7s for a completed
recording.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-25 15:53:09 +00:00
Seppo Ingalsuo a9fb42836f Tools: Tune: EQ: Improve robustness of mls_freq_resp script
This patch increases end marker seek length from 3s to 5s to
allow more drift for arecord and aplay timing for test signal.
The captured waveform is now always plotted and when start
and end markers are found the test signal is plotted with green
color.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-25 15:53:09 +00:00
Seppo Ingalsuo 387d74ea6d Tools: Tune: EQ: Fix script compatibility with Matlab
The script mls_freq_resp.m has worked previously with Octave
while failed with Matlab because it does not have source
command. This patch ads function get_config() that reads
the configuration file, evaluates it, and returns the
needed variable.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-25 15:53:09 +00:00
Seppo Ingalsuo f976333a81 Tools: Tune: TDFB: Updates to design scripts for sound direction estimate
This patch contains improvements to TDFB setup scripts and examples
those are related to sound direction estimation. The changes are add
of 15 degree angle spacing to enum control, add of rotated noise sound
create for sound direction testing in simulated acoustics.

The example line, circular, and xyz array scripts are updated for single
dual mono beam that is controllable in 15 degree or 30 degree steps. All
the two beam design examples are moved into example_two_beams.m.

The patch also contains noise field covariance matrix diagonal load
increase to -40 dB. I noticed uneven frequency response in some
designs due to numerical instability. The increased diagonal load makes
the automatic design procedure more robust at minimal hit in noise
suppression performance.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-25 15:50:54 +00:00
Guennadi Liakhovetski 6ab8c11136 ipc4: fix trace format mismatches
Two tr_err() calls in src/ipc/ipc4/handler.c have mismatching formats
and arguments. When built with Zephyr, tr_err() translates to
printk() which then checks arguments. This allows an easy detection
of such errors.

Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
2022-01-25 11:44:16 +00:00
Guennadi Liakhovetski 2097ba2d27 ipc4: fix uninitialised variable warnings
The ret variable in ipc4_trigger_chain_dma() can be used
uninitialised. Fix it.

Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
2022-01-25 11:44:16 +00:00
Liam Girdwood 8a24a326ac ipc4: eliminate a superfluous initialization
The err variable in ipc_cmd() doesn't have to be initialized, it is
always assigned a value.

Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
2022-01-25 11:44:16 +00:00
Guennadi Liakhovetski 6f50f85633 mixer: prevent NULL dereference
User IPC can provide an invalid component ID, this shouldn't lead to
a NULL dereference.

Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
2022-01-25 11:44:16 +00:00
Liam Girdwood cc5d6efe53 ipc4: build: fix GCC11 build warning in mixer.
In file included from /home/lrg/work/sof/sof/src/include/sof/audio/buffer.h:18,
                 from /home/lrg/work/sof/sof/src/audio/mixer.c:8:
/home/lrg/work/sof/sof/src/audio/mixer.c:45:31: error: 'mixin_uuid_ldc' defined but not used [-Werror=unused-const-variable=]
   45 | DECLARE_SOF_RT_UUID("mix_in", mixin_uuid, 0x39656eb2, 0x3b71, 0x4049,
      |                               ^~~~~~~~~~
/home/lrg/work/sof/sof/src/include/sof/lib/uuid.h:94:44: note: in definition of macro 'DECLARE_SOF_UUID'
   94 |         static const struct sof_uuid_entry uuid_name ## _ldc = {        \
      |                                            ^~~~~~~~~
/home/lrg/work/sof/sof/src/audio/mixer.c:45:1: note: in expansion of macro 'DECLARE_SOF_RT_UUID'
   45 | DECLARE_SOF_RT_UUID("mix_in", mixin_uuid, 0x39656eb2, 0x3b71, 0x4049,
      | ^~~~~~~~~~~~~~~~~~~
/home/lrg/work/sof/sof/src/audio/mixer.c: In function 'mixin_bind':
/home/lrg/work/sof/sof/src/audio/mixer.c:743:17: error: 'sink_buf' may be used uninitialized in this function [-Werror=maybe-uninitialized]
  743 |                 pipeline_connect(dev, sink_buf, PPL_CONN_DIR_BUFFER_TO_COMP);
      |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
make[2]: *** [CMakeFiles/sof.dir/build.make:786: CMakeFiles/sof.dir/src/audio/mixer.c.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:1970: CMakeFiles/sof.dir/all] Error 2
make: *** [Makefile:136: all] Error 2
"make all" terminated with exit cod

Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
[guennadi.liakhovetski@linux.intel.com: mixin_tr only for IPC4]
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
2022-01-25 11:44:16 +00:00
Liam Girdwood 30aeff5493 ipc4: build: fix GCC11 build warning in copier
/home/lrg/work/sof/sof/src/audio/copier.c: In function 'copier_reset':
/home/lrg/work/sof/sof/src/audio/copier.c:519:29: error: argument to 'sizeof' in 'memset' call is the same expression as the destination; did you mean to dereference it? [-Werror=sizeof-pointer-memaccess]
  519 |         memset(cd, 0, sizeof(cd));
      |                             ^
cc1: all warnings being treated as errors
make[2]: *** [CMakeFiles/sof.dir/build.make:996: CMakeFiles/sof.dir/src/audio/copier.c.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:1970: CMakeFiles/sof.dir/all] Error 2
make: *** [Makefile:136: all] Error 2
"make all" terminated with exit code 2. Build might be incomplete.

Signed-off-by: Liam Girdwood <liam.r.girdwood@linux.intel.com>
2022-01-25 11:44:16 +00:00
Ranjani Sridharan cb80023b22 topology2: fix build issues with no command line definitions
If there are no command line definitions, the build fails because the
4th argument to the add_alsatplg2_command macro is NULL.

Fix this by using named arguments for the macro and checking for the
optional argument for command line definitions.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-24 12:51:11 +00:00
Seppo Ingalsuo 95d8e941cf Testbench: Enable verbose traces
Some useful algorithms internal traces are output with debug traces.
This patch enables the verbose trace (CONFIG_TRACEV) for library build.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-24 11:23:21 +00:00
Seppo Ingalsuo e0ec195bcd Test: Audio: Fix stdnotch_get() function issue in Matlab
The path append of current directory (.) is not correct for
the signal processing package function iirnotch. The issue does
not happen with Octave since it has pei_tseng_notch().

The standard notch function is used in all THD+N tests so this
issue has caused fail of nearly all tests with Matlab.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-24 11:23:21 +00:00
Seppo Ingalsuo 30012d1454 Tools: Testbench: Support for different in/out channels count and other
This patch adds to command line switch -n for output channels count.
Existing -c is for in channels, new -n is for out channels count. Out
channels count is same as input if -n is not present. Switch -q is
added to quiet the trace output if it is not needed.

The Matlab language test scripts for components are updated to use
the -t config.sh interface of comp_run.sh. It allows more flexible
control of input and output streams.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>
2022-01-24 11:23:21 +00:00
Marc Herbert 0031664403 xtensa-build-all.sh: distinguish default/working platform and WIP
The -a option was recently broken by the addition of platform(s) that
don't have a toolchain available in the Docker image. Every platform can
be built by someone but no one can built -a(ll) platforms right now.

Add a new platform array to make the distinction between default
platforms built by the -a option versus work in progress.

Signed-off-by: Marc Herbert <marc.herbert@intel.com>
2022-01-21 17:17:09 +00:00
Marc Herbert da1d1b041e xtensa-build-all.sh: rn: remove odd ARCH=xtensa definition
No one else has it.

Gets rid of the only shellcheck warning.

Fixes commit 7453e3d571 ("scripts: add renoir support")

Signed-off-by: Marc Herbert <marc.herbert@intel.com>
2022-01-21 17:17:09 +00:00
Marc Herbert 7f45cb7bde .github: add imx8ulp gcc build to pull request tests
Fixes commit b40a2f7d36 ("docker: Add support for i.MX8ULP platform")

Signed-off-by: Marc Herbert <marc.herbert@intel.com>
2022-01-21 17:17:09 +00:00
Marc Herbert 62d57d699b .github: reformat platform list for qemu-boot-test
More future-proof. No functional change

Signed-off-by: Marc Herbert <marc.herbert@intel.com>
2022-01-21 17:17:09 +00:00
Bard Liao 7f1572f4d9 topology2: cavs-nocodec: add audio format tokens
Add audio format tokens for I2S + DMIC.

Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
2022-01-21 17:15:38 +00:00
Guennadi Liakhovetski 92f6db8ba9 zephyr: add support for IPC4
Select correct IPC implementation depending on the selected version.

Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
2022-01-21 17:05:58 +00:00
Rander Wang db4d081040 ipc4: fix a regression issue on windows
The issue was found when testing 44.1khz stream on windows.
We should use input format of copier to set source buffer fmt,
not output fmt which may be dai output fmt and different from
source fmt.

According to spec, copier only support one input source and
four output sinks, so we don't need to check source list.

Tested on windows.

Signed-off-by: Rander Wang <rander.wang@intel.com>
2022-01-21 16:57:38 +00:00
Adrian Bonislawski 34be14f644 ipc4: add Aria module support for ipc4
Automatic Regressive Input Amplifier Module
ARIA applies variable gain into incoming signal.
Applied gain is in range <1, 2 power attenuation>
Currently ARIA introduces gain transition and
algorithmic latency equal to 1 ms.
At initial release only hifi3 version is supported

Signed-off-by: Adrian Bonislawski <adrian.bonislawski@intel.com>
2022-01-21 16:56:58 +00:00
Ranjani Sridharan ddcdadb68f codec_adapter: cadence: Do not free codec data during reset
During reset, it is only important to free the memory associated with
the runtime params, so that they can be reallocated during prepare. So
move the allocation of the cadence_codec_data to use rballoc, so that
module_free_all_memory() doesn't end up freeing it. It will be freed
during cadence_free() when the device is freed.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan bfd5e61768 topology1: remove setup config info from all codec_adapter topologies
Remove the setup config info from all initial config data in the
codec_adapter kcontrol's. This is no longer needed.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 692299f2a6 codec_adapter: Remove setup config from module data
Remove the setup cfg from struct module_datai and have one just common
config. The config sent during module init should be saved a set up config
within the module's private data if the module needs it. Modify the
cadence_codec_data and the waves_codec_data to add the setup config to
save the initial config that can be applied during prepare when it is
invoked during the reset.

Remove enum module_cfg_type and struct ca_config as they are no longer
needed.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 5c3262a6a9 codec_adapter: cadece: remove unnecessary goto's and fix memory leaks
Just return the appropriate value right away when possible and fix
memory leaks.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan eaaea60b65 codec_adapter: remove macro CODEC_INTERFACE_ID()
Interface ID has been replaced with UUID. So no need for the unused
macro.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 8effa4ff3d codec_adapter: fix panic during reset
If module_prepare() fails, the pipeline will be reset resulting in
module_reset() getting invoked. Just return if the state is <
MODULE_IDLE which indicates that the module was never prepared.
Also, in this case, local_buff is never allocated. So, check if it is
NULL before zeroing it out to avoid DSP panic.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan c75d794913 codec_adapter: remove printing the error with -ENOSPC
The error message should be skipped for -ENOSPC in module_process().

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 2874b17a8b codec_adapter: waves: remove included header
waves.h has been deleted

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 7475929bc7 codec_adapter: dts: remove included header
dts.h has been deleted.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 993037ceb1 codec_adapter: Remove redundant setting of avail to true
module_load_config() already sets the config avail flag to true.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00
Ranjani Sridharan 2dba175d61 module_interface: remove module_id
Remove the module ID fields from struct processing_module and struct
ca_config. The comp ID is unique across all components in topology and
can be reused to identify the modules as well.

Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
2022-01-21 12:55:38 +00:00