- 17 Sep, 2021 2 commits
-
-
Peter Ujfalusi authored
From the name sof_arch_ops one can not decipher that these ops are DSP architecture ops. Rename it to dsp_arch_ops and change also the macro to retrieve the DSP architecture specific ops as well. Signed-off-by:
Peter Ujfalusi <peter.ujfalusi@linux.intel.com> Reviewed-by:
Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com> Reviewed-by:
Kai Vehmanen <kai.vehmanen@linux.intel.com> Reviewed-by:
Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> Link: https://lore.kernel.org/r/20210916130308.7969-1-peter.ujfalusi@linux.intel.comSigned-off-by:
Mark Brown <broonie@kernel.org>
-
Peter Ujfalusi authored
If the snd_sof_dsp_send_msg() failed then we have already returned from sof_ipc_tx_message_unlocked() with the error message. There is no need to check if ret is really 0 after this and we can return directly the return value from tx_wait_done() At the same time make the remaining checks for error (ret) to match. Signed-off-by:
-
- 16 Sep, 2021 16 commits
-
-
Mark Brown authored
Merge series "ASoC: cs42l42: Implement Manual Type detection as fallback" from Vitaly Rodionov <vitalyr@opensource.cirrus.com>: For some headsets CS42L42 autodetect mode is not working correctly. They will be detected as unknown types or as headphones. According to the CS42L42 datasheet, if the headset autodetect failed, then the driver should switch to manual mode and perform a manual steps sequence. These steps were missing in the current driver code. This patch will add manual mode fallback steps in case autodetect failed. The default behavior is not affected, manual mode runs only when autodetect failed. Tested for regression with autodetect with all known headsets - no regression. Tested with all headsets customers reported as false detected: Gumdrop DropTech B1 - detected as headset OK HUAWEI AM115 - detected as headset OK UGREEN EP103 - detected as headset OK HONOR AM116 - detected as headset OK Stefan Binding (1): ASoC: cs42l42: Implement Manual Type detection as fallback sound/soc/codecs/cs42l42.c | 104 ++++++++++++++++++++++++++++++++----- sound/soc/codecs/cs42l42.h | 54 +++++++++++++++++++ 2 files changed, 146 insertions(+), 12 deletions(-) -- 2.25.1
-
Mark Brown authored
Merge series "ASoC: SOF: Clean up the probe support" from Peter Ujfalusi <peter.ujfalusi@linux.intel.com>: Hi, The probe debug feature of SOF can be used to extract streams of data from a given point of a pipeline for analysis. The support is implemented by using the ALSA/ASoC compress support for the capture stream, but the code can not be used by/for a normal compressed data stream. It is a debug feature. Merge the probe implementation in the core (compress.c/h and probe.c/h) into one file: sof-probes.c/h Rename the Intel HDA specific probe implementation from hda-compressc.c to hda-probes.c We also need to add IPC logging support for the probes messages and drop the unused references to SOF compress to have reasonably clean code. Regards, Peter --- Peter Ujfalusi (5): ASoC: SOF: ipc: Add probe message logging to ipc_log_header() ASoC: SOF: pcm: Remove non existent CONFIG_SND_SOC_SOF_COMPRESS reference ASoC: SOF: probe: Merge and clean up the probe and compress files ASoC: SOF: Intel: Rename hda-compress.c to hda-probes.c ASoC: SOF: sof-probes: Correct the function names used for snd_soc_cdai_ops Ranjani Sridharan (1): ASoC: SOF: compress: move and export sof_probe_compr_ops sound/soc/sof/Makefile | 3 +- sound/soc/sof/compress.c | 147 --------- sound/soc/sof/compress.h | 32 -- sound/soc/sof/core.c | 2 +- sound/soc/sof/debug.c | 2 +- sound/soc/sof/intel/Makefile | 2 +- sound/soc/sof/intel/hda-dai.c | 16 +- .../intel/{hda-compress.c => hda-probes.c} | 0 sound/soc/sof/ipc.c | 23 ++ sound/soc/sof/pcm.c | 6 +- sound/soc/sof/probe.h | 85 ------ sound/soc/sof/sof-priv.h | 5 - sound/soc/sof/{probe.c => sof-probes.c} | 280 +++++++++++------- sound/soc/sof/sof-probes.h | 38 +++ 14 files changed, 248 insertions(+), 393 deletions(-) delete mode 100644 sound/soc/sof/compress.c delete mode 100644 sound/soc/sof/compress.h rename sound/soc/sof/intel/{hda-compress.c => hda-probes.c} (100%) delete mode 100644 sound/soc/sof/probe.h rename sound/soc/sof/{probe.c => sof-probes.c} (52%) create mode 100644 sound/soc/sof/sof-probes.h -- 2.33.0 -
Mark Brown authored
As part of retiring the old macros defining the DAI clocking mode in the DAI format update the au1x drivers to use the new style macros. Signed-off-by: -
Mark Brown authored
Convert the Atmel drivers to use the new style defines for clocking in DAI formats. Signed-off-by:
Codrin Ciubotariu <codrin.ciubotariu@microchip.com> Acked-by:
Peter Rosin <peda@axentia.se>
-
David Rhodes authored
Fix warnings and errors in DT bindings Add newline at end of file Replace 'unevaluatedProperties' with 'additionalProperties' Add spi context to DT example Add #sound-dai-cells to DT example Rename to 'cirrus,cs35l41.yaml' Signed-off-by:
David Rhodes <drhodes@opensource.cirrus.com> Link: https://lore.kernel.org/r/20210915191422.2371623-1-drhodes@opensource.cirrus.comSigned-off-by:
-
Charles Keepax authored
Remove pdn variable that was made redundant in an earlier patch. Fixes: c2f14cc2 ("ASoC: cs35l41: Fix use of an uninitialised variable") Signed-off-by:
Charles Keepax <ckeepax@opensource.cirrus.com> Link: https://lore.kernel.org/r/20210916082346.12001-1-ckeepax@opensource.cirrus.comSigned-off-by:
-
Viorel Suman authored
Add SAI1 instance to imx8m_dai array. Signed-off-by:
Viorel Suman <viorel.suman@nxp.com> Reviewed-by:
Paul Olaru <paul.olaru@oss.nxp.com> Reviewed-by:
-
Vitaly Rodionov authored
Signed-off-by:
Vitaly Rodionov <vitalyr@opensource.cirrus.com> Link: https://lore.kernel.org/r/20210916110932.10293-1-vitalyr@opensource.cirrus.comSigned-off-by:
-
Peter Ujfalusi authored
Relocate the INIT_WORK() at the same place where we schedule the work to make the code simpler and easier to follow. Signed-off-by:
-
Stefan Binding authored
Some headsets are not detected correctly by Automatic Type Detection on cs42l42. Instead, Manual Type Detection can be used to give a more accurate value. Signed-off-by:
Stefan Binding <sbinding@opensource.cirrus.com> Signed-off-by:
-
Peter Ujfalusi authored
The snd_soc_cdai_ops have startup and shutdown callbacks defined unlike the component callbacks where open and free is used. Signed-off-by:
Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Link: https://lore.kernel.org/r/20210916103211.1573-7-peter.ujfalusi@linux.intel.comSigned-off-by:
-
Peter Ujfalusi authored
The hda-compress.c is implementing the SOF probe support for intel HDA platforms using compress API. To avoid the confusion, rename it to reflect this. Signed-off-by:
-
Peter Ujfalusi authored
The probe debug functionality is implemented via compress support and it was spread across two set of files: probe.c/h compress.c/h Merge the two files into sof-probes.s/h and clean them up by removing unused struct definitions, functions. We can also move most of the functions static as they are only used internally. Signed-off-by:
-
Ranjani Sridharan authored
sof_probe_compr_ops are not platform-specific. So move it to common compress code and export the symbol. The compilation of the common compress code is already dependent on the selection of CONFIG_SND_SOC_SOF_DEBUG_PROBES, so no need to check the Kconfig section for defining sof_probe_compr_ops again. Signed-off-by:
-
Peter Ujfalusi authored
The SND_SOC_SOF_COMPRESS is not valid Kconfig option, remove it. At the same time remove the also the declaration of the non existent sof_compressed_ops. Signed-off-by:
-
Peter Ujfalusi authored
Probe related messages are missing from the logging, for example the PROBE_INIT would show up as: ipc tx: 0xc0010000: unknown GLB command ipc tx succeeded: 0xc0010000: unknown GLB command Add code to handle the probe messages to have human readable output Signed-off-by:
-
- 15 Sep, 2021 22 commits
-
-
Peter Ujfalusi authored
The snd_sof_debugfs_io_item() only used within debug.c, no need to export it. Signed-off-by:
Daniel Baluta <daniel.baluta@gmail.com> Reviewed-by:
-
Peter Ujfalusi authored
Do not access the sdev->bar[] directly to make the code generic, use the new generic ops for handing the regions for debugfs. Signed-off-by:
-
Peter Ujfalusi authored
Set the generic iomem callback for debugfs_add_region_item to avoid regression when the core switches to use the generic interface for the regions. Signed-off-by:
-
Peter Ujfalusi authored
Set the generic iomem callback for debugfs_add_region_item to avoid regression when the core switches to use the generic interface for the regions. Signed-off-by:
-
Peter Ujfalusi authored
Add new debugfs_add_region_item along with a generic wrapper snd_sof_debugfs_add_region_item() to abstract away the DSP regions related debugfs support. At the same commit add iomem based generic implementation for the new ops Signed-off-by:
-
Peter Ujfalusi authored
The use of bar in the core poses limits on the portability of the code to other, non iomapped platforms. To make the API more generic, remove the use of 'bar' as parameter for the block copy API. Signed-off-by:
-
Peter Ujfalusi authored
snd_sof_fw_parse_ext_data() is used only internally within loader.c and there is no need to export it. Signed-off-by:
-
Peter Ujfalusi authored
The read64 operation is not used by IMX along with other IO functions. No need to set it for the ops. Signed-off-by:
Bard Liao <bard.liao@intel.com> Reviewed-by:
-
Peter Ujfalusi authored
The snd_sof_dsp_mailbox_init() is called only from sof_get_windows() to set the sdev->dsp_box.offset/size and sdev->host_box.offset/size Instead of using a function, set the offsets and sizes like we do for the other boxes in sof_get_windows(). Signed-off-by:
-
Peter Ujfalusi authored
To align with other platforms, set only the sdev->dsp_box.offset in bdw_probe(). The mailbox offset must be set in order to be able to receive the firmware ready message. The offsets and sizes will be re-configured after the FW ready message based on the window information. Signed-off-by:
-
Mark Brown authored
Merge series "ASoC: SOF: Remove unused members from struct sof_dev_desc" from Peter Ujfalusi <peter.ujfalusi@linux.intel.com>: Hi, dma_engine, dma_size and resindex_dma_base is unused from sof_dev_desc, drop them. resindex_dma_base is initialized to -1 for Intel platforms, but it is not used. Regards, Peter --- Peter Ujfalusi (2): ASoC: SOF: intel: Do no initialize resindex_dma_base ASoC: SOF: Drop resindex_dma_base, dma_engine, dma_size from sof_dev_desc include/sound/sof.h | 5 ----- sound/soc/sof/intel/pci-apl.c | 2 -- sound/soc/sof/intel/pci-cnl.c | 3 --- sound/soc/sof/intel/pci-icl.c | 2 -- sound/soc/sof/intel/pci-tgl.c | 5 ----- sound/soc/sof/intel/pci-tng.c | 1 - 6 files changed, 18 deletions(-) -- 2.33.0 -
Mark Brown authored
Merge series "ASoC: SOF: Intel: hda: Cleanups for local function uses" from Peter Ujfalusi <peter.ujfalusi@linux.intel.com>: Hi, Small cleanups regarding to HDA function locations and dropping of definitions of not implemented functions. Regards, Peter --- Peter Ujfalusi (3): ASoC: SOF: Intel: hda-dsp: Declare locally used functions as static ASoC: SOF: Intel: hda: Remove boot_firmware skl and iccmax_icl declarations ASoC: SOF: Intel: hda: Relocate inline definitions from hda.h to hda.c for sdw sound/soc/sof/intel/hda-dsp.c | 61 +++++++++++++++++------------------ sound/soc/sof/intel/hda.c | 33 ++++++++++++++++++- sound/soc/sof/intel/hda.h | 41 ----------------------- 3 files changed, 62 insertions(+), 73 deletions(-) -- 2.33.0 -
Mark Brown authored
Support the ALC5682I-VS codec in Intel's rt5682 machine driver with three board configurations. Brent Lu (4): ASoC: Intel: sof_rt5682: support ALC5682I-VS codec ASoC: intel: sof_rt5682: support jsl_rt5682s_rt1015p board ASoC: intel: sof_rt5682: support jsl_rt5682s_rt1015 board ASoC: intel: sof_rt5682: support jsl_rt5682s_mx98360a board sound/soc/intel/boards/Kconfig | 1 + sound/soc/intel/boards/sof_rt5682.c | 105 +++++++++++++++--- .../intel/common/soc-acpi-intel-jsl-match.c | 24 ++++ 3 files changed, 112 insertions(+), 18 deletions(-) -- 2.25.1
-
Charles Keepax authored
Signed-off-by:
-
Charles Keepax authored
It is not idiomatic for ASoC to print the function name in the error messages, however it is expected to show the return code. Update the error messages to follow these conventions. Signed-off-by:
-
Charles Keepax authored
In multiple places the driver overwrites the error code returned with a static error code, this is not helpful for debugging. Update to pass the error codes straight through. Signed-off-by:
-
Charles Keepax authored
cs35l41 is often connected over I2C which is a very slow bus, as such timings can be greatly improved combining writes where acceptable. Update several points where the driver does multiple register writes when a single one would suffice. Signed-off-by:
-
Charles Keepax authored
Just clean up the code a little by using the helper rather than open coding waiting for OTP_BOOT_DONE. Signed-off-by:
-
Charles Keepax authored
The loop checking PDN_DONE doesn't check the return value from regmap_read, nor does it initialise val. This means if regmap_read fails val will be checked for the PDN_DONE bit whilst being uninitialised. Fix this up by switching to regmap_read_poll_timeout which tidies up the code and avoids the uninitialised variable. Signed-off-by:
-
Trevor Wu authored
mt6359-sound is a MFD driver. Because its regmap is retrieved from its parent, it shouldn't be freed in mt6359-sound driver. snd_soc_component_exit_regmap() will do regmap_exit(), this results in unexpected results if sound card unregister flow is invoked when users try to bind/unbind audio codec. Remove the usage of snd_soc_component_exit_regmap(). Instead, set component->regmap = NULL in the component remove function. Signed-off-by:
Trevor Wu <trevor.wu@mediatek.com> Link: https://lore.kernel.org/r/20210915034659.25044-1-trevor.wu@mediatek.comSigned-off-by:
-
Peter Ujfalusi authored
Since the load_firmware callback in snd_sof_dsp_ops is mandatory and it is tested during probe. Move the snd_sof_load_firmware() wrapper to ops.h as inline and drop the check of sof_ops(sdev)->load_firmware Signed-off-by:
-
Colin Ian King authored
Don't populate the array adda_dai_list on the stack but instead it static const. Also makes the object code smaller by 33 bytes: Before: text data bss dec hex filename 28271 11640 0 39911 9be7 mt8195/mt8195-dai-adda.o After: text data bss dec hex filename 28142 11736 0 39878 9bc6 mt8195/mt8195-dai-adda.o (gcc version 11.2.0) Signed-off-by:
Colin Ian King <colin.king@canonical.com> Link: https://lore.kernel.org/r/20210915105027.10805-1-colin.king@canonical.comSigned-off-by:
-
