May 2018 - Sound-open-firmware - mailman.alsa-project.org

[Sound-open-firmware] [PATCH] topology: add fake dapm route supported in topology file
by Rander Wang 18 May '18

18 May '18

To reuse some machine drivers in Linux kernel without hacking dapm route info, some unsupported routes and widgets could be added in toplogy file. Signed-off-by: Rander Wang <rander.wang(a)linux.intel.com> --- test on cnl apl and byt pass SOF: master da9baf3f3d kernel: v4.14 390a3ece3 SOF-tools: master 86fe688a2b --- topology/m4/utils.m4 | 38 ++++++++++++++++++++++++++++++++++++++ topology/sof-cnl-rt274.m4 | 11 +++++++++++ 2 files changed, 49 insertions(+) diff --git a/topology/m4/utils.m4 b/topology/m4/utils.m4 index ba87c8d..8821576 100644 --- a/topology/m4/utils.m4 +++ b/topology/m4/utils.m4 @@ -70,5 +70,43 @@ define(`W_DATA', ` tuples STR($2)' `}') +dnl FAKE_DAPM_ROUTE_OUT(name, dai type, dai index, direction, index) +define(`FAKE_DAPM_ROUTE_OUT', +`SectionWidget.STR($1) {' +` index STR($5)' +` type "aif_out"' +` no_pm "true"' +`}' +`SectionGraph.STR($2) {' +` index STR($5)' +`' +` lines [' +` dapm($1,$2$3.$4)' +` ]' +`}') + +dnl FAKE_DAPM_ROUTE_IN(name, dai type, dai index, direction, index) +define(`FAKE_DAPM_ROUTE_IN', +`SectionWidget.STR($1) {' +` index STR($5)' +` type "aif_in"' +` no_pm "true"' +`}' +`SectionGraph.STR($2) {' +` index STR($5)' +`' +` lines [' +` dapm($2$3.$4, $1)' +` ]' +`}') + +dnl FAKE_WIDGET(name, index) +define(`FAKE_WIDGET', +`SectionWidget.STR($1) {' +` index STR($2)' +` type "mixer"' +` no_pm "true"' +`}') + divert(0) dnl diff --git a/topology/sof-cnl-rt274.m4 b/topology/sof-cnl-rt274.m4 index 89d5a25..f592b44 100644 --- a/topology/sof-cnl-rt274.m4 +++ b/topology/sof-cnl-rt274.m4 @@ -69,3 +69,14 @@ DAI_CONFIG(SSP, 0, SSP0-Codec, DSP_B, 24, DAI_CLOCK(bclk, 4800000, codec_slave), DAI_CLOCK(fsync, 48000, codec_slave), DAI_TDM(4, 25, 3, 3)) + +FAKE_DAPM_ROUTE_OUT(codec0_out, SSP, 0, OUT, 0) +FAKE_DAPM_ROUTE_OUT(codec1_out, SSP, 0, OUT, 1) +FAKE_DAPM_ROUTE_OUT(ssp0 Tx, SSP, 0, OUT, 2) +FAKE_DAPM_ROUTE_OUT(Capture, SSP, 0, OUT, 3) +FAKE_DAPM_ROUTE_OUT(SoC DMIC, SSP, 0, OUT, 4) +FAKE_DAPM_ROUTE_IN(codec0_in, SSP, 0, IN, 5) +FAKE_WIDGET(DMIC01 Rx, 6) +FAKE_WIDGET(DMic, 7) +FAKE_WIDGET(dmic01_hifi, 8) +FAKE_WIDGET(ssp0 Rx, 9) -- 2.14.1

3 2

[Sound-open-firmware] [PATCH] topology: add virtual dapm route supported in topology file
by Rander Wang 18 May '18

18 May '18

To reuse some machine drivers in Linux kernel without hacking dapm route info, some unsupported routes and widgets could be added in toplogy file. Signed-off-by: Rander Wang <rander.wang(a)linux.intel.com> --- test on cnl apl and byt pass SOF: master da9baf3f3d kernel: v4.14 390a3ece3 SOF-tools: master 86fe688a2b --- topology/m4/utils.m4 | 38 ++++++++++++++++++++++++++++++++++++++ topology/sof-cnl-rt274.m4 | 11 +++++++++++ 2 files changed, 49 insertions(+) diff --git a/topology/m4/utils.m4 b/topology/m4/utils.m4 index ba87c8d..8b397b6 100644 --- a/topology/m4/utils.m4 +++ b/topology/m4/utils.m4 @@ -70,5 +70,43 @@ define(`W_DATA', ` tuples STR($2)' `}') +dnl VIRTUAL_DAPM_ROUTE_OUT(name, dai type, dai index, direction, index) +define(`VIRTUAL_DAPM_ROUTE_OUT', +`SectionWidget.STR($1) {' +` index STR($5)' +` type "aif_out"' +` no_pm "true"' +`}' +`SectionGraph.STR($2) {' +` index STR($5)' +`' +` lines [' +` dapm($1,$2$3.$4)' +` ]' +`}') + +dnl VIRTUAL_DAPM_ROUTE_IN(name, dai type, dai index, direction, index) +define(`VIRTUAL_DAPM_ROUTE_IN', +`SectionWidget.STR($1) {' +` index STR($5)' +` type "aif_in"' +` no_pm "true"' +`}' +`SectionGraph.STR($2) {' +` index STR($5)' +`' +` lines [' +` dapm($2$3.$4, $1)' +` ]' +`}') + +dnl VIRTUAL_WIDGET(name, index) +define(`VIRTUAL_WIDGET', +`SectionWidget.STR($1) {' +` index STR($2)' +` type "mixer"' +` no_pm "true"' +`}') + divert(0) dnl diff --git a/topology/sof-cnl-rt274.m4 b/topology/sof-cnl-rt274.m4 index 89d5a25..36cd91f 100644 --- a/topology/sof-cnl-rt274.m4 +++ b/topology/sof-cnl-rt274.m4 @@ -69,3 +69,14 @@ DAI_CONFIG(SSP, 0, SSP0-Codec, DSP_B, 24, DAI_CLOCK(bclk, 4800000, codec_slave), DAI_CLOCK(fsync, 48000, codec_slave), DAI_TDM(4, 25, 3, 3)) + +VIRTUAL_DAPM_ROUTE_OUT(codec0_out, SSP, 0, OUT, 0) +VIRTUAL_DAPM_ROUTE_OUT(codec1_out, SSP, 0, OUT, 1) +VIRTUAL_DAPM_ROUTE_OUT(ssp0 Tx, SSP, 0, OUT, 2) +VIRTUAL_DAPM_ROUTE_OUT(Capture, SSP, 0, OUT, 3) +VIRTUAL_DAPM_ROUTE_OUT(SoC DMIC, SSP, 0, OUT, 4) +VIRTUAL_DAPM_ROUTE_IN(codec0_in, SSP, 0, IN, 5) +VIRTUAL_WIDGET(DMIC01 Rx, 6) +VIRTUAL_WIDGET(DMic, 7) +VIRTUAL_WIDGET(dmic01_hifi, 8) +VIRTUAL_WIDGET(ssp0 Rx, 9) -- 2.14.1

2 1

[Sound-open-firmware] [PATCH] trace: Modfiy trace event to accept uint64_t data
by Xiuli Pan 18 May '18

18 May '18

From: Pan Xiuli <xiuli.pan(a)linux.intel.com> In DMA trace we store the data as uint64_t, modify the function prototype that the trace_value can get uint64_t as the src. The added 32 bit can be used to store some channel info to help debug with multi stream. Signed-off-by: Pan Xiuli <xiuli.pan(a)linux.intel.com> --- Test with: Mininow max rt5651 and UP2 nocodec and CNL nocodec SOF master: da9baf3f3dfef7378ee922620e3a31a04d15504c SOF-Tool master: 86fe688a2b4f68a1ce87e0951686be12a00f1a3c https://github.com/plbossart/sound/tree/topic/sof-v4.14: 33807ca8baa20bab2f98e9cc8a23c51a0822ad9f --- src/include/sof/trace.h | 2 +- src/lib/trace.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/include/sof/trace.h b/src/include/sof/trace.h index 2fd77fe..15e620b 100644 --- a/src/include/sof/trace.h +++ b/src/include/sof/trace.h @@ -103,7 +103,7 @@ #define TRACEE 1 #define TRACEM 0 /* send all trace messages to mbox and local trace buffer */ -void _trace_event(uint32_t event); +void _trace_event(uint64_t event); void _trace_event_mbox(uint32_t event); void _trace_event_atomic(uint32_t event); void _trace_event_mbox_atomic(uint32_t event); diff --git a/src/lib/trace.c b/src/lib/trace.c index 0c37858..38def6d 100644 --- a/src/lib/trace.c +++ b/src/lib/trace.c @@ -46,7 +46,7 @@ struct trace { static struct trace trace; /* send trace events only to the local trace buffer */ -void _trace_event(uint32_t event) +void _trace_event(uint64_t event) { uint64_t dt[2]; -- 2.7.4

2 1

[Sound-open-firmware] [PATCH] host: remove first_copy workaround for HDA DMA
by Keyon Jie 18 May '18

18 May '18

Here remove the first_copy workaround(it was needed as HDA DMA is not available at the trigger start stage). This require the fix from host driver side: ASoC: SOF: hw-apl: start HDA DMA at stream_prepare() stage and remove stream_trigger() Signed-off-by: Keyon Jie <yang.jie(a)linux.intel.com> --- Test with: Mininow max rt5651 and APL GPMRB and CNL nocodec SOF master: 7c5dd8cfdf4f8ca1fa5cfdd7352b81cbfc8edb88 SOF-Tool master: 86fe688a2b4f68a1ce87e0951686be12a00f1a3c https://github.com/plbossart/sound/tree/topic/sof-v4.14: 22c8b2b68186e6af83f21c302d8086559b5d2e23 note: this need be merged with driver patch together: 'ASoC: SOF: start HDA DMA at hw_params() stage and remove stream_trigger()' --- src/audio/host.c | 41 ++++++++++++----------------------------- 1 file changed, 12 insertions(+), 29 deletions(-) diff --git a/src/audio/host.c b/src/audio/host.c index 998c5ca..a0956b3 100644 --- a/src/audio/host.c +++ b/src/audio/host.c @@ -84,9 +84,6 @@ struct host_data { uint32_t period_bytes; uint32_t period_count; -#if defined CONFIG_DMA_GW - uint32_t first_copy; -#endif /* stream info */ struct sof_ipc_stream_posn posn; /* TODO: update this */ }; @@ -477,12 +474,12 @@ static int host_trigger(struct comp_dev *dev, int cmd) goto out; } - /* preload first playback period for preloader task */ + /* + * host dma will copy the first period once it is started, + * automatically. + * Here update the pointers to reflect the real case. + */ if (dev->params.direction == SOF_IPC_STREAM_PLAYBACK) { - /* - * host dma will not start copy at this point yet, - * just produce empty period bytes for it. - */ comp_update_buffer_produce(hd->dma_buffer, hd->period_bytes); } @@ -736,10 +733,6 @@ static int host_prepare(struct comp_dev *dev) hd->split_remaining = 0; dev->position = 0; -#if defined CONFIG_DMA_GW - hd->first_copy = 1; -#endif - return 0; } @@ -863,19 +856,6 @@ static int host_copy(struct comp_dev *dev) if (dev->state != COMP_STATE_ACTIVE) return 0; -#if defined CONFIG_DMA_GW - if (dev->params.direction == SOF_IPC_STREAM_PLAYBACK && - hd->first_copy) { - /* - * host dma will not start copy at this point yet, just produce - * empty period bytes for it. - */ - comp_update_buffer_produce(hd->dma_buffer, - hd->period_bytes); - hd->first_copy = 0; - return 0; - } -#endif local_elem = list_first_item(&hd->config.elem_list, struct dma_sg_elem, list); @@ -896,14 +876,17 @@ static int host_copy(struct comp_dev *dev) } #if defined CONFIG_DMA_GW - - /* update host pointers from last period */ - host_gw_dma_update(dev); - /* tell gateway to copy another period */ ret = dma_copy(hd->dma, hd->chan, hd->period_bytes); if (ret < 0) goto out; + + /* + * update host pointers for the new copied period. + * fixme: do we need wait and check to make sure + * the new copy is finished here? + */ + host_gw_dma_update(dev); #else /* do DMA transfer */ ret = dma_set_config(hd->dma, hd->chan, &hd->config); -- 2.14.1

3 2

[Sound-open-firmware] [PATCH v2] [RFC]SOF: FIX: adjust the SSC3's TFT/RFT field in SSP
by Wu Zhigang 18 May '18

18 May '18

Actually I did not have full confidence about this patch. if someone can help do more test, that would be better. before this, do not accept this patch. Incorrect TFT/RFT will cause DMA IRQ happened too frequently. the interval will be less than 1ms, which will cause XRUN easily. The old code will has this issue in TDM4-s32le test. Signed-off-by: Wu Zhigang <zhigang.wu(a)linux.intel.com> Acked-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> --- v2: the sample-width should be taken into account. --- Test with: APL-GPMRB with TDF8532 codec Linux sof-v4.14: 1d4a25d72d1e87c182f15a68d32c9599b734ea7b SOF master: 7c5dd8cfdf4f8ca1fa5cfdd7352b81cbfc8edb88 Tools master: 86fe688a2b4f68a1ce87e0951686be12a00f1a3c Pass the Sanity test: BYT(Minnowboard Turbo) APL(UP^2) CNL(CNL-RVP) --- src/drivers/apl-ssp.c | 25 ++++++++++++++----------- 1 file changed, 14 insertions(+), 11 deletions(-) diff --git a/src/drivers/apl-ssp.c b/src/drivers/apl-ssp.c index a51f91c..53bb8be 100644 --- a/src/drivers/apl-ssp.c +++ b/src/drivers/apl-ssp.c @@ -36,6 +36,7 @@ #include <sof/ssp.h> #include <sof/alloc.h> #include <sof/interrupt.h> +#include <sof/math/numbers.h> #include <config.h> /* tracing */ @@ -114,6 +115,7 @@ static inline int ssp_set_config(struct dai *dai, uint32_t rft; uint32_t active_tx_slots = 2; uint32_t active_rx_slots = 2; + uint32_t sample_width = 2; bool inverted_frame = false; int ret = 0; @@ -499,22 +501,23 @@ static inline int ssp_set_config(struct dai *dai, /* setting TFT and RFT */ switch (config->ssp.sample_valid_bits) { case 16: - /* use 2 bytes for each slot */ - tft = active_tx_slots * 2; - rft = active_rx_slots * 2; - break; + /* use 2 bytes for each slot */ + sample_width = 2; + break; case 24: case 32: - /* use 4 bytes for each slot */ - tft = active_tx_slots * 4; - rft = active_rx_slots * 4; - break; + /* use 4 bytes for each slot */ + sample_width = 4; + break; default: - trace_ssp_error("ecd"); - ret = -EINVAL; - goto out; + trace_ssp_error("ecd"); + ret = -EINVAL; + goto out; } + tft = MIN(8, sample_width * active_tx_slots); + rft = MIN(8, sample_width * active_rx_slots); + sscr3 |= SSCR3_TX(tft) | SSCR3_RX(rft); trace_ssp("coe"); -- 2.14.1

2 1

[Sound-open-firmware] [PATCH] [RFC]SOF: FIX: adjust the SSC3's TFT/RFT field in SSP
by Wu Zhigang 18 May '18

18 May '18

Actually I did not have full confidence about this patch. if someone can help do more test, that would be better. before this, do not accept this patch. Incorrect TFT/RFT will cause DMA IRQ happened too frequently. the interval will be less than 1ms, which will cause XRUN easily. The old code will has this issue in TDM4-s32le test. Signed-off-by: Wu Zhigang <zhigang.wu(a)linux.intel.com> Acked-by: Pierre-Louis Bossart <pierre-louis.bossart(a)linux.intel.com> --- Test with: APL-GPMRB with TDF8532 codec Linux sof-v4.14: 1d4a25d72d1e87c182f15a68d32c9599b734ea7b SOF master: 7c5dd8cfdf4f8ca1fa5cfdd7352b81cbfc8edb88 Tools master: 86fe688a2b4f68a1ce87e0951686be12a00f1a3c Pass the Sanity test: BYT(Minnowboard Turbo) APL(UP^2) CNL(CNL-RVP) --- src/drivers/apl-ssp.c | 21 +++------------------ 1 file changed, 3 insertions(+), 18 deletions(-) diff --git a/src/drivers/apl-ssp.c b/src/drivers/apl-ssp.c index a51f91c..5fb9b27 100644 --- a/src/drivers/apl-ssp.c +++ b/src/drivers/apl-ssp.c @@ -36,6 +36,7 @@ #include <sof/ssp.h> #include <sof/alloc.h> #include <sof/interrupt.h> +#include <sof/math/numbers.h> #include <config.h> /* tracing */ @@ -496,24 +497,8 @@ static inline int ssp_set_config(struct dai *dai, /* bypass divider for MCLK */ mdivr = 0x00000fff; - /* setting TFT and RFT */ - switch (config->ssp.sample_valid_bits) { - case 16: - /* use 2 bytes for each slot */ - tft = active_tx_slots * 2; - rft = active_rx_slots * 2; - break; - case 24: - case 32: - /* use 4 bytes for each slot */ - tft = active_tx_slots * 4; - rft = active_rx_slots * 4; - break; - default: - trace_ssp_error("ecd"); - ret = -EINVAL; - goto out; - } + tft = MIN(8, 2 * active_tx_slots); + rft = MIN(8, 2 * active_rx_slots); sscr3 |= SSCR3_TX(tft) | SSCR3_RX(rft); -- 2.14.1

1 0

[Sound-open-firmware] how to find a suitable value for the TFT/RFT parameter on multi-platform
by zhigangw 17 May '18

17 May '18

Hello Liam: I think it is better to mention this status again in the mail list. I am working on playback with the TDM8(4slot used)-48KHz-32bit on APL-GPMRB platform. Currently the status is: The music can be heard from the all four channels. But the it has the noise on the background. the noise exists all the time during playing. No xrun happened during the audio playing. I have not tried the pause/resume feature. I have not tried keyon's patch for the noise issue. Issue: During the development, the TFT (in SSP's SSC3 register) parameter will affect the playing operation. If we keep code: tft = active_tx_slots * 4; the tft will be 16. during the playing, i found the DMA IRQ will happened very frequently, less than 1ms, which will cause XRUN during playing. But it will be normal if the tft value is below 8 (tft < 8). I did some tests on other platform. but I have not found suitable way to fit all of them. the RFT also have the similar concern. Do you have any suggestions about this? thank you br ~zhigang

2 4

[Sound-open-firmware] [RFC PATCH v2 00/29] Trace level control based on modules.
by yan.wang＠linux.intel.com 17 May '18

17 May '18

From: Yan Wang <yan.wang(a)linux.intel.com> This patch set is only for review. This patch set is tested on CNL no-codec only currently. User could check current trace level status like the following: cat /sys/kernel/debug/sof/trace_level irq >> 1 ipc >> 1 dma >> 1 ssp >> 1 wait >> 1 lock >> 1 mem >> 1 value >> 1 PIPELINE.2.SSP2.IN >> 1 SSP2.IN >> 1 BUF2.0 >> 1 PCM0C >> 0 PIPELINE.1.SSP2.OUT >> 1 SSP2.OUT >> 1 BUF1.0 >> 1 PCM0P >> 1 It includes trace levels of global trace types and loaded widgets. When level = 0, only error trace will be enabled. When level = 1, error and normal trace will be enabled. When level = 2, all trace including verbose trace will be enabled. Set trace level like the following: echo PCM0C 0 > /sys/kernel/debug/sof/trace_level will disable the trace of capture host component of fw pipeline. In the future, more module information will bind with trace. Yan Wang (29): Define IPC message type and data structure for trace level. Add utility functions and variable for DMA trace level control. Add trace level variable and set function for SOF modules. Add trace level IPC message processing logic. Change trace macro definition and API for trace level checking. Change macro definition and trace calling for dma-trace.c Change macro definition and trace calling for dma-copy.c Change macro definition and trace calling for alloc.c Change macro definition and trace calling for agent.c Change macro definition and trace calling for schdule module. Change macro definition and trace calling for wait.h Change macro definition and trace calling for SSP. Change macro definition and trace calling for dw-dma.c Change macro definition and trace calling for hda-dma.c Change macro definition and trace calling for interrupt.h Change macro definition and trace calling for comp_dev. Change macro definition and trace calling for comp_buffer. Change macro definition and trace calling for pipeline. Change macro definition and trace calling for dai. Change macro definition and trace calling for volume. Change macro definition and trace calling for switch. Change macro definition and trace calling for tone. Change macro definition and trace calling for mixer. Change macro definition and trace calling for host. Change macro definition and trace calling for mux. Change macro definition and trace calling for eq_fir. Change macro definition and trace calling for eq_iir. Change macro definition and trace calling for src. Change macro definition and trace calling for IPC. src/audio/buffer.c | 13 +-- src/audio/component.c | 31 +++--- src/audio/dai.c | 80 +++++++++------- src/audio/eq_fir.c | 63 ++++++------ src/audio/eq_iir.c | 63 ++++++------ src/audio/host.c | 71 ++++++++------ src/audio/mixer.c | 39 +++++--- src/audio/mux.c | 17 +++- src/audio/pipeline.c | 191 +++++++++++++++++++----------------- src/audio/pipeline_static.c | 2 +- src/audio/src.c | 97 ++++++++++--------- src/audio/src.h | 5 +- src/audio/switch.c | 17 +++- src/audio/tone.c | 79 ++++++++------- src/audio/volume.c | 121 ++++++++++++----------- src/drivers/apl-ssp.c | 29 +++--- src/drivers/byt-ssp.c | 7 +- src/drivers/dw-dma.c | 25 +++-- src/drivers/hda-dma.c | 44 +++++---- src/include/sof/audio/buffer.h | 41 +++++--- src/include/sof/audio/component.h | 34 ++++--- src/include/sof/audio/pipeline.h | 18 +++- src/include/sof/dma-trace.h | 7 ++ src/include/sof/interrupt.h | 12 ++- src/include/sof/ipc.h | 18 +++- src/include/sof/schedule.h | 10 ++ src/include/sof/ssp.h | 14 ++- src/include/sof/trace.h | 197 +++++++++++++++++++++++++++++--------- src/include/sof/wait.h | 21 +++- src/include/uapi/ipc.h | 25 +++++ src/ipc/dma-copy.c | 13 ++- src/ipc/handler.c | 54 ++++++++--- src/ipc/ipc.c | 50 ++++++++-- src/lib/agent.c | 12 ++- src/lib/alloc.c | 18 +++- src/lib/dma-trace.c | 83 +++++++++++++++- src/lib/schedule.c | 20 ++-- src/lib/trace.c | 101 +++++++++++++++++++ 38 files changed, 1173 insertions(+), 569 deletions(-) -- 2.14.3

6 36

[Sound-open-firmware] [PATCH v3 0/4] Host based component testing
by Ranjani Sridharan 17 May '18

17 May '18

This set of patches introduces the testbench for audio processing components in SOF. The testbench enables for functional/quality testing of the components. The testbench reads in audio samples from a file and write out the processed samples to the output file. The src and test directories includes routines for topology parsing and setting up sof ipc, scheduler and processing pipelines. It also includes simpler implementations for SOF features such as tracing. It also includes a new file component that allows reading in or writing out samples from/to a file. The test directory also includes the testbench for setting up pipelines based on a test topology. It reads in input samples from a file , schedules pipeline copy and copies the processed sample to the output file. The output file can be used to perform audio quality tests using matlab/octave. Currently it supports only volume pipelines. Support for other components such as src/eq will be added later. v3 modifications: File component supports raw pcm input Fixed checkpatch style issues Fixed errors with 32-bit input samples debug message prints are handled via command line Ranjani Sridharan (4): tune: add testbench common headers and source files tune: new component for file IO tune: add testbench to test topology tune: Add README and a script to configure testbench parameters Makefile.am | 2 +- configure.ac | 4 +- tune/Makefile.am | 1 + tune/README | 13 + tune/scripts/run_testbench.sh | 22 + tune/src/include/test/common_test.h | 65 ++ tune/src/include/test/file.h | 77 +++ tune/src/include/test/topology.h | 202 ++++++ tune/src/include/test/trace.h | 49 ++ tune/test/Makefile.am | 31 + tune/test/alloc.c | 68 ++ tune/test/common_test.c | 206 ++++++ tune/test/file.c | 723 +++++++++++++++++++++ tune/test/ipc.c | 81 +++ tune/test/schedule.c | 99 +++ tune/test/testbench.c | 298 +++++++++ tune/test/topology.c | 955 ++++++++++++++++++++++++++++ tune/test/trace.c | 200 ++++++ 18 files changed, 3094 insertions(+), 2 deletions(-) create mode 100644 tune/Makefile.am create mode 100644 tune/README create mode 100755 tune/scripts/run_testbench.sh create mode 100644 tune/src/include/test/common_test.h create mode 100644 tune/src/include/test/file.h create mode 100644 tune/src/include/test/topology.h create mode 100644 tune/src/include/test/trace.h create mode 100644 tune/test/Makefile.am create mode 100644 tune/test/alloc.c create mode 100644 tune/test/common_test.c create mode 100644 tune/test/file.c create mode 100644 tune/test/ipc.c create mode 100644 tune/test/schedule.c create mode 100644 tune/test/testbench.c create mode 100644 tune/test/topology.c create mode 100644 tune/test/trace.c -- 2.17.0

2 5

[Sound-open-firmware] [PATCH 0/7] [v3]DMIC: Patch series for direct attach microphones
by Seppo Ingalsuo 17 May '18

17 May '18

This v3 patch set completes the missing changes from previous v2 patch set. Only stereo S32_LE 48 kHz capture is supported due to hard wiring of DMIC and DMA parameters. Configurable format will be introduced later. The used SOF kernel driver and topology were test versions. Patches will be upstreamed separately later. This patch version was tested with the UP2 board on top of SOF git master 1e231b9dd297bc88c324502a561be67007604b16. Since there was no DMIC exHAT board connected the capture only produced the DC compensation ramp shape when the HW tries to compensate the max. full scale DC signal caused by the non-toggling PDM data line. Ramp shape looked correct and there were no discontinuities in PCM samples. The CNL version has not been tested on a device. Patch was also tested on a BYT device without DMIC capability where DAI/SSP feature functioned OK. Seppo Ingalsuo (7): [v3]DMIC: Add PDM microphones (DMIC) support to DAI [v3]DMIC: Add configure command switch to disable DMIC [v3]DMIC: DMIC: Add dmic.c to drivers Makefile.am [v3]DMIC: Changes to APL platform [v3]DMIC: Changes to CNL platform [v3]DMIC: Add for format.h conversion from fractional to float [v3]DMIC: Move ceil_divide from numbers.c to numbers.h as static inline configure.ac | 6 + src/drivers/Makefile.am | 6 +- src/drivers/dmic.c | 1361 +++++++++++++++++++++ src/include/sof/audio/format.h | 3 + src/include/sof/math/numbers.h | 15 +- src/math/numbers.c | 14 - src/platform/apollolake/dai.c | 59 + src/platform/apollolake/include/platform/memory.h | 5 + src/platform/apollolake/platform.c | 12 + src/platform/cannonlake/dai.c | 57 + src/platform/cannonlake/include/platform/memory.h | 8 +- src/platform/cannonlake/platform.c | 11 + 12 files changed, 1537 insertions(+), 20 deletions(-) create mode 100644 src/drivers/dmic.c -- 2.14.1

3 14