March 2017 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PLUGINS] Multichannel over I2S : extra cost of user space plugins
by Matt Flax 08 Mar '17

08 Mar '17

Recently I was advised to channel multichannel audio over stereo I2S and unpack in user space with plugins. I have questions about this approach. Right now, with a hacked machine driver, I get DMA unpacking to multichannel for free. It works with all sample rates in silicon and it also works with all word sizes. If I switch to user space ALSA plugin unpacking, as advised, can anyone answer the following questions please : If I want to handle all word sizes which are currently implemented S16 -> S32, do I need to use different plugins to handle different word sizes ? Regarding the actual channel unpacking, will it cost a lot more to do this in user space, rather then using the DMA silicon ? The bcm2835 is not a very powerful processor and this is a concern. Does anyone have an example plugin which will handle all word sizes through one plugin device ? thanks Matt

2 1

[alsa-devel] [PATCH] ASoC: imx-pcm-dma: simplify pcm_config
by Lucas Stach 08 Mar '17

08 Mar '17

The generic snd_dmaengine_pcm is able to retrieve all the needed information from the attached dmaengine and is in fact able to provide much more accurate flags to userspace, like the SDMA engine being only able to operate in batch mode. To avoid any future inconsistencies between the dmaengine and the pcm_config, rip out the fixed config and rely on the core to fill in the right flags derived from the dmaengine information. Signed-off-by: Lucas Stach <l.stach(a)pengutronix.de> --- sound/soc/fsl/imx-pcm-dma.c | 28 ---------------------------- 1 file changed, 28 deletions(-) diff --git a/sound/soc/fsl/imx-pcm-dma.c b/sound/soc/fsl/imx-pcm-dma.c index f3d3d1ffa84e..314814ddd2b0 100644 --- a/sound/soc/fsl/imx-pcm-dma.c +++ b/sound/soc/fsl/imx-pcm-dma.c @@ -33,48 +33,20 @@ static bool filter(struct dma_chan *chan, void *param) return true; } -static const struct snd_pcm_hardware imx_pcm_hardware = { - .info = SNDRV_PCM_INFO_INTERLEAVED | - SNDRV_PCM_INFO_BLOCK_TRANSFER | - SNDRV_PCM_INFO_MMAP | - SNDRV_PCM_INFO_MMAP_VALID | - SNDRV_PCM_INFO_PAUSE | - SNDRV_PCM_INFO_RESUME, - .buffer_bytes_max = IMX_DEFAULT_DMABUF_SIZE, - .period_bytes_min = 128, - .period_bytes_max = 65535, /* Limited by SDMA engine */ - .periods_min = 2, - .periods_max = 255, - .fifo_size = 0, -}; - static const struct snd_dmaengine_pcm_config imx_dmaengine_pcm_config = { - .pcm_hardware = &imx_pcm_hardware, .prepare_slave_config = snd_dmaengine_pcm_prepare_slave_config, .compat_filter_fn = filter, - .prealloc_buffer_size = IMX_DEFAULT_DMABUF_SIZE, }; int imx_pcm_dma_init(struct platform_device *pdev, size_t size) { struct snd_dmaengine_pcm_config *config; - struct snd_pcm_hardware *pcm_hardware; config = devm_kzalloc(&pdev->dev, sizeof(struct snd_dmaengine_pcm_config), GFP_KERNEL); if (!config) return -ENOMEM; *config = imx_dmaengine_pcm_config; - if (size) - config->prealloc_buffer_size = size; - - pcm_hardware = devm_kzalloc(&pdev->dev, - sizeof(struct snd_pcm_hardware), GFP_KERNEL); - *pcm_hardware = imx_pcm_hardware; - if (size) - pcm_hardware->buffer_bytes_max = size; - - config->pcm_hardware = pcm_hardware; return devm_snd_dmaengine_pcm_register(&pdev->dev, config, -- 2.11.0

2 1

[alsa-devel] [PATCH] ASoC: Intel: Skylake: use a helper macro to rounding-up calculation
by Takashi Sakamoto 08 Mar '17

08 Mar '17

In 'include/linux/kernel.h', there's a helper macro to round numerical value. Let's use it. Cc: Takashi Sakamoto <o-takashi(a)sakamocchi.jp> Signed-off-by: Takashi Sakamoto <takashi.sakamoto(a)miraclelinux.com> --- sound/soc/intel/skylake/skl-topology.c | 18 ++++-------------- 1 file changed, 4 insertions(+), 14 deletions(-) diff --git a/sound/soc/intel/skylake/skl-topology.c b/sound/soc/intel/skylake/skl-topology.c index 2dbfb1b..f38f8e6 100644 --- a/sound/soc/intel/skylake/skl-topology.c +++ b/sound/soc/intel/skylake/skl-topology.c @@ -299,8 +299,6 @@ static void skl_tplg_update_buffer_size(struct skl_sst *ctx, { int multiplier = 1; struct skl_module_fmt *in_fmt, *out_fmt; - int in_rate, out_rate; - /* Since fixups is applied to pin 0 only, ibs, obs needs * change for pin 0 only @@ -311,21 +309,13 @@ static void skl_tplg_update_buffer_size(struct skl_sst *ctx, if (mcfg->m_type == SKL_MODULE_TYPE_SRCINT) multiplier = 5; - if (in_fmt->s_freq % 1000) - in_rate = (in_fmt->s_freq / 1000) + 1; - else - in_rate = (in_fmt->s_freq / 1000); - - mcfg->ibs = in_rate * (mcfg->in_fmt->channels) * + mcfg->ibs = DIV_ROUND_UP(in_fmt->s_freq, 1000) * + (mcfg->in_fmt->channels) * (mcfg->in_fmt->bit_depth >> 3) * multiplier; - if (mcfg->out_fmt->s_freq % 1000) - out_rate = (mcfg->out_fmt->s_freq / 1000) + 1; - else - out_rate = (mcfg->out_fmt->s_freq / 1000); - - mcfg->obs = out_rate * (mcfg->out_fmt->channels) * + mcfg->obs = DIV_ROUND_UP(mcfg->out_fmt->s_freq, 1000) * + (mcfg->out_fmt->channels) * (mcfg->out_fmt->bit_depth >> 3) * multiplier; } -- 2.9.3

3 2

[alsa-devel] [PATCH] ASoC: fsl: Remove unneeded init of static variable
by Alin Grigorean 08 Mar '17

08 Mar '17

This was reported by checkpatch.pl Signed-off-by: Alin Grigorean <alin.grig93(a)gmail.com> --- sound/soc/fsl/imx-pcm-fiq.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/fsl/imx-pcm-fiq.c b/sound/soc/fsl/imx-pcm-fiq.c index dac6688540dc..92410f7ca1fa 100644 --- a/sound/soc/fsl/imx-pcm-fiq.c +++ b/sound/soc/fsl/imx-pcm-fiq.c @@ -282,7 +282,7 @@ static int imx_pcm_new(struct snd_soc_pcm_runtime *rtd) return 0; } -static int ssi_irq = 0; +static int ssi_irq; static int imx_pcm_fiq_new(struct snd_soc_pcm_runtime *rtd) { -- 2.11.1

3 2

[alsa-devel] [PATCH] ASoC: rt5665: fix getting wrong work handler container
by Bard Liao 08 Mar '17

08 Mar '17

We got rt5665 private data from wrong work. It will result in kernel panic. Signed-off-by: Bard Liao <bardliao(a)realtek.com> --- sound/soc/codecs/rt5665.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/rt5665.c b/sound/soc/codecs/rt5665.c index c866502..a0c1d9a 100644 --- a/sound/soc/codecs/rt5665.c +++ b/sound/soc/codecs/rt5665.c @@ -1241,7 +1241,7 @@ static irqreturn_t rt5665_irq(int irq, void *data) static void rt5665_jd_check_handler(struct work_struct *work) { struct rt5665_priv *rt5665 = container_of(work, struct rt5665_priv, - calibrate_work.work); + jd_check_work.work); if (snd_soc_read(rt5665->codec, RT5665_AJD1_CTRL) & 0x0010) { /* jack out */ -- 2.7.4

2 1

[alsa-devel] [PATCH 1/8] ASoC: rt5665: increase LDO level
by Bard Liao 08 Mar '17

08 Mar '17

Too low LDO level will cause a few functions unstable. Signed-off-by: Bard Liao <bardliao(a)realtek.com> --- sound/soc/codecs/rt5665.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/rt5665.c b/sound/soc/codecs/rt5665.c index fe2cf1e..a4c07c2 100644 --- a/sound/soc/codecs/rt5665.c +++ b/sound/soc/codecs/rt5665.c @@ -4798,7 +4798,7 @@ static int rt5665_i2c_probe(struct i2c_client *i2c, /* Enhance performance*/ regmap_update_bits(rt5665->regmap, RT5665_PWR_ANLG_1, RT5665_HP_DRIVER_MASK | RT5665_LDO1_DVO_MASK, - RT5665_HP_DRIVER_5X | RT5665_LDO1_DVO_09); + RT5665_HP_DRIVER_5X | RT5665_LDO1_DVO_12); INIT_DELAYED_WORK(&rt5665->jack_detect_work, rt5665_jack_detect_handler); -- 2.7.4

2 13

Re: [alsa-devel] [PATCH 6/8] ASoC: rt5665: move set_pll to codec level
by Mark Brown 08 Mar '17

08 Mar '17

On Wed, Mar 08, 2017 at 11:19:52AM +0000, Bard Liao wrote: > This patch doesn't use lower case of 's'ource in the function body. > Could we apply the first 5 patches if there is no comment? > Or ignore the error and I will send a patch to lower case in the function > body? I'll skip this patch, please send a new version.

1 0

Re: [alsa-devel] [RFC] Merge DSD_U32_BE devices list
by Jurgen Kramer 08 Mar '17

08 Mar '17

Hello Naoki, On Tue, 2017-03-07 at 15:53 +0900, Naoki MATSUMOTO wrote: > Dear Mr.lintweaker > CC:alsa-ml > > Hello nice to meet you. > I know that you work for xmos-native-dsd. > You has contributed linux, alsa-lib and mpd. > You have patches of xmos-native-dsd device list on github. > https://github.com/lintweaker/xmos-native-dsd/tree/master/SRPMS/patch > es/kernel > so, I want to merge upstreame kernel. because, MPD 0.20 over support > DSD_U32_BE. > > Current kernel has few support devices. > https://patchwork.kernel.org/project/alsa-devel/list/?q=dsd > > I want to create patch. Could you please tell me anything problems? > I am planning on sending in a patch(set) adding all remaining devices not in the kernel yet. Time frame: +/- 2 weeks. Jurgen

2 1

[alsa-devel] [PATCH] ASoC: bcm_2835: Simplify mono guard by returning error on mono
by Matt Flax 08 Mar '17

08 Mar '17

This patch simplifies the bcm2835_i2s.c code when the driver returns error on mono channel requests. It does this by testing for mono instead of using a switch statement to test for stereo. The bcm2835_i2s driver can only work in stereo mode. Whilst leaving the driver entirel compliant with the I2S protocol, it simplifies the code by removing a few redundant lines of code. Signed-off-by: Matt Flax <flatmax(a)flatmax.org> --- sound/soc/bcm/bcm2835-i2s.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/sound/soc/bcm/bcm2835-i2s.c b/sound/soc/bcm/bcm2835-i2s.c index 6ba2049..1e6a1da 100644 --- a/sound/soc/bcm/bcm2835-i2s.c +++ b/sound/soc/bcm/bcm2835-i2s.c @@ -310,15 +310,16 @@ static int bcm2835_i2s_hw_params(struct snd_pcm_substream *substream, ch1pos = data_delay; ch2pos = bclk_ratio / 2 + data_delay; - switch (params_channels(params)) { - case 2: - format = BCM2835_I2S_CH1(format) | BCM2835_I2S_CH2(format); - format |= BCM2835_I2S_CH1(BCM2835_I2S_CHPOS(ch1pos)); - format |= BCM2835_I2S_CH2(BCM2835_I2S_CHPOS(ch2pos)); - break; - default: + /* C + * Check we aren't setting channel 2 on a mono stream. + * We currently only support stereo + */ + if (params_channels(params) == 1) return -EINVAL; - } + + format = BCM2835_I2S_CH1(format) | BCM2835_I2S_CH2(format); + format |= BCM2835_I2S_CH1(BCM2835_I2S_CHPOS(ch1pos)); + format |= BCM2835_I2S_CH2(BCM2835_I2S_CHPOS(ch2pos)); /* * Set format for both streams. -- 2.7.4

3 5

[alsa-devel] DSD native format on SoC
by Michael Nazzareno Trimarchi 08 Mar '17

08 Mar '17

Hi DSD is a continuous stream of bits that flows over two separate channels, left and right, synchronized by a clock, while I2S is a single data wire and an additional wire that states if the incoming sample refers to the left channel or the right channel. There is no way to get DSD data other than a circuit which decode DSD streams coming directly from a conventional source, like, for examples, a SACD. The Volta dac has an internal decoding circuitry that still employs the I2S standard as a usual 44100Hz/16 bits, but sourcing the shift clock at twice the DSD standard frequency of 2.8224MHz, since it has to split two the16 bits words received on a single wire into the two DSD channels, as the standard wants. When DSD is enabled, the control signal PCD/DSD must be high, CK3..CK0 must be all low. The I2S interface works as a standard 44100Hz/16 bit and the DSD streaming must be packed into 16 bits lenght words left/right as per LRCK logic. The BCLK frequency supplied from the interface is 5644800Hz in case of DSD and 11289600 for DSD2. Can you explain how now DSD can be pass to audio card? The idea is to declare it on the audio card and then configure it as a PCM card but I don't know how data are suppose to arrive from alsa userspace. Michael

1 0