- 01 Oct, 2021 21 commits
-
-
Mark Brown authored
Merge series "Add support for on demand pipeline setup/destroy" from Peter Ujfalusi <peter.ujfalusi@linux.intel.com>: Hi, The previous, v2 of this series was sent by Daniel Baluta: https://lore.kernel.org/alsa-devel/20210917143659.401102-1-daniel.baluta@oss.nxp.com/ We have agreed that it might be better that someone from Intel is going to take it from here as we already have the infrastructure up to test and verify the dynamic pipelines support. Changes since v2 (sent by Daniel Baluta): - patch 10: Fix NULL point dereference in hda_dai_update_config() - I have kept Daniel's SoB for the series. Changes since v1: - Signed-off-by tag added by Daniel This series implements initial support for dynamic pipelines to setup/teardown pipeline as needed when a PCM is open/closed. Initially dynamic pipelines are only supported with single core setup which will be expanded with a follow-up series. Review with SOF community at https://github.com/thesofproject/linux/pull/2794 The feature has been merged on 1st of April to sof-dev, all issues found since has been fixed and squashed to this upstream series. Regards, Peter --- Ranjani Sridharan (12): ASoC: topology: change the complete op in snd_soc_tplg_ops to return int ASoC: SOF: control: Add access field in struct snd_sof_control ASoC: SOF: topology: Add new token for dynamic pipeline ASoC: SOF: sof-audio: add helpers for widgets, kcontrols and dai config set up AsoC: dapm: export a couple of functions ASoC: SOF: Add new fields to snd_sof_route ASoC: SOF: restore kcontrols for widget during set up ASoC: SOF: Don't set up widgets during topology parsing ASoC: SOF: Introduce widget use_count ASoC: SOF: Intel: hda: make sure DAI widget is set up before IPC ASoC: SOF: Add support for dynamic pipelines ASoC: SOF: topology: Add kernel parameter for topology verification include/sound/soc-dpcm.h | 1 + include/sound/soc-topology.h | 2 +- include/uapi/sound/sof/tokens.h | 1 + sound/soc/intel/skylake/skl-topology.c | 6 +- sound/soc/soc-dapm.c | 2 + sound/soc/soc-pcm.c | 4 +- sound/soc/soc-topology.c | 10 +- sound/soc/sof/intel/hda-dai.c | 174 +++--- sound/soc/sof/intel/hda.c | 177 ++++-- sound/soc/sof/intel/hda.h | 5 + sound/soc/sof/ipc.c | 22 + sound/soc/sof/pcm.c | 58 +- sound/soc/sof/pm.c | 4 +- sound/soc/sof/sof-audio.c | 709 +++++++++++++++++++------ sound/soc/sof/sof-audio.h | 32 +- sound/soc/sof/sof-priv.h | 1 + sound/soc/sof/topology.c | 362 +++++-------- 17 files changed, 1032 insertions(+), 538 deletions(-) -- 2.33.0
-
Viorel Suman authored
Implement SPDIF bypass mode. It implies internal SoC routing of SPDIF input signal to SPDIF output signal. The test bed requires two boards: B1 configured in bypass mode, and B2 to feed B1 SPDIF RX port and read B1 SPDIF TX port: B2 TX -> B1 RX, B2 RX <- B1 TX. The test procedure: a) Boot both boards b) B2: start "arecord <spdifcard> -r 48kHz | aplay <local DAC>" c) B2: start "aplay <spdifcard> -r 48kHz <2ch 48kHz audio file>" d) B1: enable bypass mode: amixer -cimxspdif cset numid=8,iface=PCM,name='Bypass Mode' on e) B2: check DAC audio, make sure the same sample rate is used at steps b) and c), in example above the rate is 48kHz. f) B1: try to run "aplay" or "arecord" on imxspdif card while in bypass mode - both must fail until bypass mode is disabled g) B1: disable bypass mode: amixer -cimxspdif cset numid=8,iface=PCM,name='Bypass Mode' off h) B1: check the usual playback and capture on imxspdif card. During this test try to set bypass mode - must not be allowed while playback or capture is running. Signed-off-by:
Viorel Suman <viorel.suman@nxp.com> Signed-off-by:
Shengjiu Wang <shengjiu.wang@nxp.com> Link: https://lore.kernel.org/r/1632649760-1651-1-git-send-email-shengjiu.wang@nxp.comSigned-off-by:
Mark Brown <broonie@kernel.org>
-
Ranjani Sridharan authored
Add a kernel debug flag to enable a one-shot topology verification for all pipelines including the dynamic ones. If the debug flag is set, all the topology component loading will be verified during the complete callback. Signed-off-by:
Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Reviewed-by:
Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com> Reviewed-by:
Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Reviewed-by:
Kai Vehmanen <kai.vehmanen@linux.intel.com> Signed-off-by:
Daniel Baluta <daniel.baluta@nxp.com> Signed-off-by:
Peter Ujfalusi <peter.ujfalusi@linux.intel.com> Link: https://lore.kernel.org/r/20210927120517.20505-13-peter.ujfalusi@linux.intel.comSigned-off-by:
-
Ranjani Sridharan authored
Add support for dynamic pipelines by modifying the PCM hw_params ioctl implementation to determine the widgets required for a PCM stream by querying the list of connected DAPM widgets. This list is saved as part of snd_sof_pcm_stream struct and will be used to setup the widgets. The sof_widget_list_setup/free routines setup and free connected DAPM widgets when a PCM is opened/closed. These routines accept a list of connected DAPM widgets as input and determine the SOF widgets, their corresponding pipeline widgets and connections between them that need to be setup before the PCM is triggered. Please note that the dynamic pipeline feature will only be enabled for those pipelines whose dynamic_pipeline_widget flag is set in topologies. Add a new token called SOF_TKN_SCHED_DYNAMIC_PIPELINE that when set in topology will be applied to the dynamic_pipeline_widget flag of the pipeline widget. Signed-off-by:
-
Ranjani Sridharan authored
With the implementation of the dynamic pipeline feature, widgets will only be setup when a PCM is opened during the hw_params ioctl. The BE hw_params callback is responsible for sending the DAI_CONFIG for the DAI widgets in the DSP. With dynamic pipelines, the DAI widgets will need to set up first before sending the DAI_CONFIG IPC in the BE hw_params. Update the BE hw_params/hw_free callbacks for all ALH, HDA and SSP DAIs to set up/free the DAI widget before/after DAI_CONFIG IPC. Signed-off-by:
-
Ranjani Sridharan authored
Add a new field, use_count to struct snd_sof_widget to keep track of the usage count for each widget. Since widgets can belong to multiple pipelines, this field will ensure that the widget is setup only when the first pipeline that needs it is started and freed when the last pipeline that needs it is stopped. There is no need to protect the widget use_count access as the core already handles mutual exclusion at the PCM level. Add a new helper sof_widget_free() to handle freeing the SOF widgets and export the sof_widget_setup/free() functions. Signed-off-by:
-
Ranjani Sridharan authored
In preparation for supporting dynamic pipelines, move the widget setup, DAI config IPCs to the complete callback during topology loading. For current topology where all the pipelines are static, all the pipelines will be set up during complete. For topologies with dynamic and static pipelines, this will enable setting up only the static ones during topology loading. Reuse the sof_restore_pipelines() function for this purpose and rename it to sof_set_up_pipelines(). Signed-off-by:
-
Ranjani Sridharan authored
Restore kcontrols for each widget after it has been set up successfully. Signed-off-by:
Seppo Ingalsuo <seppo.ingalsuo@intel.com> Signed-off-by:
-
Ranjani Sridharan authored
Add two new fields to save the source widget and sink widget pointers in struct snd_sof_route to make it easier to look up routes by source/sink widget. Also, add a flag to indicate if the route has been set up in the DSP. These will be used when the dynamic pipeline feature is implemented and routes will have to be set up at run time. Also, add a new sof_tear_down_pipelines() callback, that will used to reset the set up status for all routes during suspend. Signed-off-by:
-
Ranjani Sridharan authored
Export a couple of DAPM functions that can be used by ASoC drivers to determine connected widgets when a PCM is started. Signed-off-by:
-
Ranjani Sridharan authored
Refactor the existing code to use helper functions to set up/free widgets, send dai config and set up kcontrols for widgets. These will be reused later on for setting up widgets in the connected DAPM widgets list for a particular PCM when the dynamic pipeline feature is implemented. Signed-off-by:
-
Ranjani Sridharan authored
Today, we set up all widgets required for all PCM streams at the time of topology parsing even if they are not used. An optimization would be to only set up the widgets required for currently active PCM streams. This would give the FW the opportunity to power gate unused memory blocks, thereby saving power. For dynamic pipelines, the widgets in the connected DAPM path for each PCM will need to be set up at runtime. This patch introduces a new token, DYNAMIC_PIPELINE, for scheduler type widgets that indicate whether a pipeline should be set up statically during topology load or at runtime when the PCM is opened. Introduce a new field called dynamic_pipeline_widget in struct snd_sof_widget to save the value of the parsed token. The token is set only for the pipeline (scheduler type) widget and must be propagated to all widgets in the same pipeline during topology load. Introduce another field called pipe_widget in struct snd_sof_widget that saves the pointer to the scheduler widget with the same pipeline ID as that of the widget. This field is populated when the pipeline completion callback is invoked during topology loading. Signed-off-by:
-
Ranjani Sridharan authored
Add a new field to save the access setting for all controls in struct snd_sof_control. This will be used to ensure that only widgets belonging to static pipelines have volatile controls. Signed-off-by:
-
Ranjani Sridharan authored
In the SOF driver, the operations performed in the complete callback can fail and therefore topology loading should return an error in such cases. So, change the signature of the complete op in struct snd_soc_tplg_ops to return an int to return the error. Also, amend the complete callback functions in the SOF driver and the SKL driver to conform with the new signature. Signed-off-by:
-
Rikard Falkeborn authored
The struct iapq8096_ops is only assigned to the ops field in the snd_soc_dai_link struct which is a pointer to const struct snd_soc_ops. Make it const to allow the compiler to put it in read-only memory. Signed-off-by:
Rikard Falkeborn <rikard.falkeborn@gmail.com> Link: https://lore.kernel.org/r/20211001115030.10402-1-rikard.falkeborn@gmail.comSigned-off-by:
-
Derek Fang authored
When the system resumes from S3, if the system plays a beep, there is continuous "Zizi.." noise from HP that could be heard. It is caused by the SAR mode switch during the combo jack re-detection which be executed parallelly in a workqueue after the system resumes. This patch changes the behavior of SAR mode switch to avoid this issue. Signed-off-by:
Derek Fang <derek.fang@realtek.com> Link: https://lore.kernel.org/r/20211001074113.2223-2-derek.fang@realtek.comSigned-off-by:
-
Derek Fang authored
When codec's ASRC was enabled, the DA and AD filters use the reference clk which tracks the I2S clks. And if the I2S clks' timing of open and close are not expected, this patch allows the filters to switch to use sysclk if ASRC's ref clks disappeared and could fix the below possible issues: 1. Avoid DA filter to keep surplus samples. 2. Avoid that AD filter works failed during dapm's power on. For example, if I2S clks were closed before dacdat during playback off due to jack unplug, it causes ref clks disappeared and DA filter remained some samples which will produce pop noise on the next HP playback. ASRC auto-disable could clear the samples during the playback off. Signed-off-by:
-
Trevor Wu authored
platforms->of_node and codes->of_node are assigned in probe function, and of_node_put is called at the end of probe function, because of_node seems to be not used after probe functon In this patch, of_node_put is moved to platform remove function in case of_node is used at any occasions after probe function in the future. Signed-off-by:
Trevor Wu <trevor.wu@mediatek.com> Link: https://lore.kernel.org/r/20211001031601.3953-1-trevor.wu@mediatek.comSigned-off-by:
-
Malik_Hsu authored
Add a board config adl_mx98360a_rt5682 to support alc5682 headset codec and max98360a speaker amplifier. Follow Intel BT offload design by connecting alc5682 to SSP0 and max98360a to SSP1. Signed-off-by:
Malik_Hsu <malik_hsu@wistron.corp-partner.google.com> Signed-off-by:
Brent Lu <brent.lu@intel.com> Acked-by:
-
Rikard Falkeborn authored
The struct tegra_machine_snd_ops is only assigned to the ops field in the snd_soc_dai_link struct which is a pointer to const struct snd_soc_ops. Make it const to allow the compiler to put it in read-only memory. Signed-off-by:
-
Simon Trimmer authored
This simplification allows the use of the standard kernel pattern of static inline dummy functions for debugfs code. Most systems will only have a small number of snd_soc_components so the memory impact is minimal. Signed-off-by:
Simon Trimmer <simont@opensource.cirrus.com> Suggested-by:
-
- 30 Sep, 2021 8 commits
-
-
Jack Yu authored
The DAI clock is only used in I2S mode, to make it clear and to fix clock resource release issue, we move CCF clock related code to rt5682_i2c_probe to fix clock register/unregister issue. Signed-off-by:
Jack Yu <jack.yu@realtek.com> Link: https://lore.kernel.org/r/20210929054344.12112-1-jack.yu@realtek.comSigned-off-by:
-
Shengjiu Wang authored
On i.MX8ULP the audio interface and codec are controlled by Cortex-M domain, Cortex-M core provides audio service over rpmsg. The rpmsg audio function is almost same as i.MX7ULP platform, so share same configuration. Signed-off-by:
-
Guo Zhengkui authored
Remove a repeated "#include <linux/firmware.h>" in line 32. Signed-off-by:
Guo Zhengkui <guozhengkui@vivo.com> Acked-by:
-
Peter Ujfalusi authored
We have growing number of options under SND_SOC_SOF_TOPLEVEL as SOF adaptation is growing (Intel, NXP, AMD and Mediatek) and new features are added. It will make the menuconfig user experience much cleaner if we move the SOF options under a separate page. Signed-off-by:
Bard Liao <bard.liao@intel.com> Reviewed-by:
Daniel Baluta <daniel.baluta@gmail.com> Reviewed-by:
-
Derek Fang authored
Fix the device-tree schema errors that be reported by using 'make dt_binding_check'. Signed-off-by:
-
Derek Fang authored
Revise the macro RT5682S_PLLB_SRC_MASK to 0x1 because the mux is only two-source. Signed-off-by:
-
Derek Fang authored
It could observe the debug messages more clearly by using dev_dbg. Signed-off-by:
-
Derek Fang authored
This reg is for SW reset. It shouldn't have default value, so remove. Signed-off-by:
-
- 29 Sep, 2021 2 commits
-
-
Rikard Falkeborn authored
The struct mop500_ab8500_ops is only assigned to the ops field in the snd_soc_dai_link struct which is a pointer to const struct snd_soc_ops. Make it const to allow the compiler to put it in read-only memory. Signed-off-by:
-
Rikard Falkeborn authored
These are only assigned to the ops field in the snd_soc_dai_link struct which is a pointer to const struct snd_soc_ops. Make them const to allow the compiler to put them in read-only memory. Signed-off-by:
-
- 28 Sep, 2021 8 commits
-
-
Brent Lu authored
To follow 20-character length limitation of platform device name, we have only 7 character space for amplifier. Therefore, the last character of mx98357a and mx98360a is removed to save space. Signed-off-by:
-
Peter Ujfalusi authored
If we get an error on reply (msg->reply_error) then we should print the error value out. At the same time extend the print to include the message size as well and do the same in case of a timeout. Signed-off-by:
Paul Olaru <paul.olaru@oss.nxp.com> Reviewed-by:
-
Marc Herbert authored
These helped troubleshoot some DMA issue in SOF. Signed-off-by:
Marc Herbert <marc.herbert@intel.com> Reviewed-by:
-
Guennadi Liakhovetski authored
If an invalid stream is passed to snd_sof_ipc_msg_data() it won't fill the provided object with data. The caller has to be able to recognise such cases to avoid handling invalid data. Make the function return an error when failing. Signed-off-by:
Peter Ujfalusi <peter.ujfalusi@intel.com> Reviewed-by:
-
Trevor Wu authored
dp node and hdmi node are retrieved from of_parse_phandle(), so using of_node_put() on them before return. Signed-off-by:
-
Pierre-Louis Bossart authored
Sparse (make C=2) complains about undeclared variables. Fix by adding a real prototype instead of 'extern' in sof-of-dev.c Signed-off-by:
-
Ranjani Sridharan authored
There is no restore_stream flag anymmore. Signed-off-by:
-
Peter Ujfalusi authored
In order to provide more information in case of timeout observed while reading STREAM_SD_OFFSET, print out the stream name or in case there is no audio stream associated (like dma-trace), print "--" Signed-off-by:
-
- 27 Sep, 2021 1 commit
-
-
Mark Brown authored
Merge series "add driver to support firmware loading on Cirrus Logic DSPs" from Simon Trimmer <simont@opensource.cirrus.com>: This series of patches gradually separates the ASoC specific wm_adsp code from that required to manage firmware in Cirrus Logic DSPs. The series starts with renaming, progresses to splitting the functionality before finally moving the independent functionality into drivers/firmware so that it can be used by both the existing ASoC wm_adsp and new non-audio parts. Charles Keepax (3): ASoC: wm_adsp: Move check for control existence ASoC: wm_adsp: Switch to using wm_coeff_read_ctrl for compressed buffers ASoC: wm_adsp: Move sys_config_size to wm_adsp Simon Trimmer (13): ASoC: wm_adsp: Remove use of snd_ctl_elem_type_t ASoC: wm_adsp: Cancel ongoing work when removing controls ASoC: wm_adsp: Rename generic DSP support ASoC: wm_adsp: Introduce cs_dsp logging macros ASoC: wm_adsp: Separate some ASoC and generic functions ASoC: wm_adsp: Split DSP power operations into helper functions ASoC: wm_adsp: Separate generic cs_dsp_coeff_ctl handling ASoC: wm_adsp: Move check of dsp->running to better place ASoC: wm_adsp: Pass firmware names as parameters when starting DSP core ASoC: wm_adsp: move firmware loading to client ASoC: wm_adsp: Split out struct cs_dsp from struct wm_adsp ASoC: wm_adsp: Separate wm_adsp specifics in cs_dsp_client_ops firmware: cs_dsp: add driver to support firmware loading on Cirrus Logic DSPs MAINTAINERS | 11 + drivers/firmware/Kconfig | 1 + drivers/firmware/Makefile | 1 + drivers/firmware/cirrus/Kconfig | 5 + drivers/firmware/cirrus/Makefile | 3 + drivers/firmware/cirrus/cs_dsp.c | 3109 ++++++++++++++++ include/linux/firmware/cirrus/cs_dsp.h | 242 ++ .../linux/firmware/cirrus}/wmfw.h | 8 +- sound/soc/codecs/Kconfig | 1 + sound/soc/codecs/cs47l15.c | 22 +- sound/soc/codecs/cs47l24.c | 20 +- sound/soc/codecs/cs47l35.c | 24 +- sound/soc/codecs/cs47l85.c | 32 +- sound/soc/codecs/cs47l90.c | 34 +- sound/soc/codecs/cs47l92.c | 20 +- sound/soc/codecs/madera.c | 18 +- sound/soc/codecs/wm2200.c | 30 +- sound/soc/codecs/wm5102.c | 16 +- sound/soc/codecs/wm5110.c | 24 +- sound/soc/codecs/wm_adsp.c | 3188 ++--------------- sound/soc/codecs/wm_adsp.h | 105 +- 21 files changed, 3794 insertions(+), 3120 deletions(-) create mode 100644 drivers/firmware/cirrus/Kconfig create mode 100644 drivers/firmware/cirrus/Makefile create mode 100644 drivers/firmware/cirrus/cs_dsp.c create mode 100644 include/linux/firmware/cirrus/cs_dsp.h rename {sound/soc/codecs => include/linux/firmware/cirrus}/wmfw.h (91%) -- 2.33.0
-
