November 2012 - Alsa-devel - mailman.alsa-project.org

Re: [alsa-devel] [PATCH V4 1/2] ASoC: Samsung: Get I2S src_clk from clock alias id.
by Mark Brown 28 Nov '12

28 Nov '12

On Wed, Nov 28, 2012 at 04:17:48PM +0530, Padmavathi Venna wrote: > As the I2S src clks are registered with clkdev using generic > connection id, driver can get the clk name using generic id. > So the variable representing the array of rclk src clks is > deleted. Applied, thanks.

1 0

[alsa-devel] [PATCH] ASoC: omap-abe-twl6040: Use "ti, jack-detection" DT property as boolean
by Peter Ujfalusi 28 Nov '12

28 Nov '12

Handle the jack detection inforamtion as bool from devicetree. Signed-off-by: Peter Ujfalusi <peter.ujfalusi(a)ti.com> --- Documentation/devicetree/bindings/sound/omap-abe-twl6040.txt | 4 ++-- sound/soc/omap/omap-abe-twl6040.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/Documentation/devicetree/bindings/sound/omap-abe-twl6040.txt b/Documentation/devicetree/bindings/sound/omap-abe-twl6040.txt index 65dec87..fd40c85 100644 --- a/Documentation/devicetree/bindings/sound/omap-abe-twl6040.txt +++ b/Documentation/devicetree/bindings/sound/omap-abe-twl6040.txt @@ -12,7 +12,7 @@ Required properties: Optional properties: - ti,dmic: phandle for the OMAP dmic node if the machine have it connected -- ti,jack_detection: Need to be set to <1> if the board capable to detect jack +- ti,jack_detection: Need to be present if the board capable to detect jack insertion, removal. Available audio endpoints for the audio-routing table: @@ -59,7 +59,7 @@ sound { compatible = "ti,abe-twl6040"; ti,model = "SDP4430"; - ti,jack-detection = <1>; + ti,jack-detection; ti,mclk-freq = <38400000>; ti,mcpdm = <&mcpdm>; diff --git a/sound/soc/omap/omap-abe-twl6040.c b/sound/soc/omap/omap-abe-twl6040.c index a57a4e6..1d6ea86 100644 --- a/sound/soc/omap/omap-abe-twl6040.c +++ b/sound/soc/omap/omap-abe-twl6040.c @@ -331,8 +331,8 @@ static __devinit int omap_abe_probe(struct platform_device *pdev) num_links = 1; } - of_property_read_u32(node, "ti,jack-detection", - &priv->jack_detection); + priv->jack_detection = of_property_read_bool(node, + "ti,jack-detection"); of_property_read_u32(node, "ti,mclk-freq", &priv->mclk_freq); if (!priv->mclk_freq) { -- 1.8.0

2 1

[alsa-devel] [asoc:topic/adsp 139/139] sound/soc/codecs/wm_adsp.c:556:4: warning: suggest parentheses around comparison in operand of '&'
by kbuild test robot 28 Nov '12

28 Nov '12

tree: git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git topic/adsp head: 973838a030d2a91fbd524c95281991510ed0a865 commit: 973838a030d2a91fbd524c95281991510ed0a865 [139/139] ASoC: wm_adsp: Enable DVFS for ADSP2 config: make ARCH=x86_64 allmodconfig All warnings: sound/soc/codecs/wm_adsp.c: In function 'wm_adsp2_event': sound/soc/codecs/wm_adsp.c:556:4: warning: suggest parentheses around comparison in operand of '&' [-Wparentheses] vim +556 sound/soc/codecs/wm_adsp.c 2159ad93 Mark Brown 2012-10-11 540 struct wm_adsp *dsps = snd_soc_codec_get_drvdata(codec); 2159ad93 Mark Brown 2012-10-11 541 struct wm_adsp *dsp = &dsps[w->shift]; 973838a0 Mark Brown 2012-11-28 542 unsigned int val; 2159ad93 Mark Brown 2012-10-11 543 int ret; 2159ad93 Mark Brown 2012-10-11 544 2159ad93 Mark Brown 2012-10-11 545 switch (event) { 2159ad93 Mark Brown 2012-10-11 546 case SND_SOC_DAPM_POST_PMU: 973838a0 Mark Brown 2012-11-28 547 if (dsp->dvfs) { 973838a0 Mark Brown 2012-11-28 548 ret = regmap_read(dsp->regmap, 973838a0 Mark Brown 2012-11-28 549 dsp->base + ADSP2_CLOCKING, &val); 973838a0 Mark Brown 2012-11-28 550 if (ret != 0) { 973838a0 Mark Brown 2012-11-28 551 dev_err(dsp->dev, 973838a0 Mark Brown 2012-11-28 552 "Failed to read clocking: %d\n", ret); 973838a0 Mark Brown 2012-11-28 553 return ret; 973838a0 Mark Brown 2012-11-28 554 } 973838a0 Mark Brown 2012-11-28 555 973838a0 Mark Brown 2012-11-28 @556 if (val & ADSP2_CLK_SEL_MASK >= 3) { 973838a0 Mark Brown 2012-11-28 557 ret = regulator_enable(dsp->dvfs); 973838a0 Mark Brown 2012-11-28 558 if (ret != 0) { 973838a0 Mark Brown 2012-11-28 559 dev_err(dsp->dev, 973838a0 Mark Brown 2012-11-28 560 "Failed to enable supply: %d\n", 973838a0 Mark Brown 2012-11-28 561 ret); 973838a0 Mark Brown 2012-11-28 562 return ret; 973838a0 Mark Brown 2012-11-28 563 } 973838a0 Mark Brown 2012-11-28 564 --- 0-DAY kernel build testing backend Open Source Technology Center Fengguang Wu, Yuanhan Liu Intel Corporation

1 0

[alsa-devel] [PATCH 1/2] ASoC: wm_adsp: Enable DVFS for ADSP2
by Mark Brown 28 Nov '12

28 Nov '12

Some ADSP devices can make use of DVFS to optimise power consumption depending on the operating frequency of the DSP core. Implement support for this in the generic ADSP code. Signed-off-by: Mark Brown <broonie(a)opensource.wolfsonmicro.com> --- sound/soc/codecs/wm_adsp.c | 97 +++++++++++++++++++++++++++++++++++++++++++- sound/soc/codecs/wm_adsp.h | 5 +++ 2 files changed, 100 insertions(+), 2 deletions(-) diff --git a/sound/soc/codecs/wm_adsp.c b/sound/soc/codecs/wm_adsp.c index 43bc69f..bc780cc 100644 --- a/sound/soc/codecs/wm_adsp.c +++ b/sound/soc/codecs/wm_adsp.c @@ -18,6 +18,7 @@ #include <linux/pm.h> #include <linux/pm_runtime.h> #include <linux/regmap.h> +#include <linux/regulator/consumer.h> #include <linux/slab.h> #include <sound/core.h> #include <sound/pcm.h> @@ -102,8 +103,9 @@ #define ADSP1_START_SHIFT 0 /* DSP1_START */ #define ADSP1_START_WIDTH 1 /* DSP1_START */ -#define ADSP2_CONTROL 0 -#define ADSP2_STATUS1 4 +#define ADSP2_CONTROL 0 +#define ADSP2_CLOCKING 1 +#define ADSP2_STATUS1 4 /* * ADSP2 Control @@ -127,6 +129,13 @@ #define ADSP2_START_WIDTH 1 /* DSP1_START */ /* + * ADSP2 clocking + */ +#define ADSP2_CLK_SEL_MASK 0x0007 /* CLK_SEL_ENA */ +#define ADSP2_CLK_SEL_SHIFT 0 /* CLK_SEL_ENA */ +#define ADSP2_CLK_SEL_WIDTH 3 /* CLK_SEL_ENA */ + +/* * ADSP2 Status 1 */ #define ADSP2_RAM_RDY 0x0001 @@ -669,10 +678,41 @@ int wm_adsp2_event(struct snd_soc_dapm_widget *w, struct snd_soc_codec *codec = w->codec; struct wm_adsp *dsps = snd_soc_codec_get_drvdata(codec); struct wm_adsp *dsp = &dsps[w->shift]; + unsigned int val; int ret; switch (event) { case SND_SOC_DAPM_POST_PMU: + if (dsp->dvfs) { + ret = regmap_read(dsp->regmap, + dsp->base + ADSP2_CLOCKING, &val); + if (ret != 0) { + dev_err(dsp->dev, + "Failed to read clocking: %d\n", ret); + return ret; + } + + if (val & ADSP2_CLK_SEL_MASK >= 3) { + ret = regulator_enable(dsp->dvfs); + if (ret != 0) { + dev_err(dsp->dev, + "Failed to enable supply: %d\n", + ret); + return ret; + } + + ret = regulator_set_voltage(dsp->dvfs, + 1800000, + 1800000); + if (ret != 0) { + dev_err(dsp->dev, + "Failed to raise supply: %d\n", + ret); + return ret; + } + } + } + ret = wm_adsp2_ena(dsp); if (ret != 0) return ret; @@ -699,6 +739,21 @@ int wm_adsp2_event(struct snd_soc_dapm_widget *w, case SND_SOC_DAPM_PRE_PMD: regmap_update_bits(dsp->regmap, dsp->base + ADSP2_CONTROL, ADSP2_SYS_ENA | ADSP2_START, 0); + + if (dsp->dvfs) { + ret = regulator_set_voltage(dsp->dvfs, 1200000, + 1800000); + if (ret != 0) + dev_warn(dsp->dev, + "Failed to lower supply: %d\n", + ret); + + ret = regulator_disable(dsp->dvfs); + if (ret != 0) + dev_err(dsp->dev, + "Failed to enable supply: %d\n", + ret); + } break; default: @@ -712,3 +767,41 @@ err: return ret; } EXPORT_SYMBOL_GPL(wm_adsp2_event); + +int wm_adsp2_init(struct wm_adsp *adsp, bool dvfs) +{ + int ret; + + if (dvfs) { + adsp->dvfs = devm_regulator_get(adsp->dev, "DCVDD"); + if (IS_ERR(adsp->dvfs)) { + ret = PTR_ERR(adsp->dvfs); + dev_err(adsp->dev, "Failed to get DCVDD: %d\n", ret); + return ret; + } + + ret = regulator_enable(adsp->dvfs); + if (ret != 0) { + dev_err(adsp->dev, "Failed to enable DCVDD: %d\n", + ret); + return ret; + } + + ret = regulator_set_voltage(adsp->dvfs, 1200000, 1800000); + if (ret != 0) { + dev_err(adsp->dev, "Failed to initialise DVFS: %d\n", + ret); + return ret; + } + + ret = regulator_disable(adsp->dvfs); + if (ret != 0) { + dev_err(adsp->dev, "Failed to disable DCVDD: %d\n", + ret); + return ret; + } + } + + return 0; +} +EXPORT_SYMBOL_GPL(wm_adsp2_init); diff --git a/sound/soc/codecs/wm_adsp.h b/sound/soc/codecs/wm_adsp.h index b303b1f..ffd29a4 100644 --- a/sound/soc/codecs/wm_adsp.h +++ b/sound/soc/codecs/wm_adsp.h @@ -18,6 +18,8 @@ #include "wmfw.h" +struct regulator; + struct wm_adsp_region { int type; unsigned int base; @@ -34,6 +36,8 @@ struct wm_adsp { const struct wm_adsp_region *mem; int num_mems; + + struct regulator *dvfs; }; #define WM_ADSP1(wname, num) \ @@ -46,6 +50,7 @@ struct wm_adsp { .shift = num, .event = wm_adsp2_event, \ .event_flags = SND_SOC_DAPM_POST_PMU | SND_SOC_DAPM_PRE_PMD } +int wm_adsp2_init(struct wm_adsp *adsp, bool dvfs); int wm_adsp1_event(struct snd_soc_dapm_widget *w, struct snd_kcontrol *kcontrol, int event); int wm_adsp2_event(struct snd_soc_dapm_widget *w, -- 1.7.10.4

1 1

Re: [alsa-devel] [PATCH] sound/sgtl5000: fix initial setting for DAC data source
by Mark Brown 28 Nov '12

28 Nov '12

On Wed, Nov 28, 2012 at 10:14:28AM +0000, Oskar Schirmer wrote: You've somehow decided to CC half the world but still haven't managed to CC the maintainers like Liam asked (and as is covered in SubmittingPatches)... > According to documentation bit 3:2 in register SSS_CTRL are > reserved and zero, so the setting of 0x0008 does not make much sense. > Instead, bit 4 should be set as is with power up, to reflect > routing of I2S_IN to DAC correctly. > Looks much like an off by one error. What does "as is with power up" mean? This is a register defaults table, it should reflect what is in the register defaults. The meaning of the field and the desired configuration for runtime is not relevant here, the only thing that matters is what the default value of the register is.

1 0

[alsa-devel] [PATCH 1/4] mfd: arizona: Allow the CODEC DAPM context to be accessed elsewhere
by Mark Brown 28 Nov '12

28 Nov '12

Some other device functions need to integrate with signal sources in the audio portion (primarily for haptics) so allow CODEC to export the DAPM context by pointing to it from the core driver. Signed-off-by: Mark Brown <broonie(a)opensource.wolfsonmicro.com> --- include/linux/mfd/arizona/core.h | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/include/linux/mfd/arizona/core.h b/include/linux/mfd/arizona/core.h index dd231ac..a580363 100644 --- a/include/linux/mfd/arizona/core.h +++ b/include/linux/mfd/arizona/core.h @@ -78,6 +78,8 @@ enum arizona_type { #define ARIZONA_NUM_IRQ 50 +struct snd_soc_dapm_context; + struct arizona { struct regmap *regmap; struct device *dev; @@ -98,6 +100,8 @@ struct arizona { struct mutex clk_lock; int clk32k_ref; + + struct snd_soc_dapm_context *dapm; }; int arizona_clk32k_enable(struct arizona *arizona); -- 1.7.10.4

2 7

[alsa-devel] [PATCH] ASoC: arizona: Record FLL setting when disabling
by Mark Brown 28 Nov '12

28 Nov '12

Otherwise we skip reenables. Signed-off-by: Mark Brown <broonie(a)opensource.wolfsonmicro.com> --- sound/soc/codecs/arizona.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/sound/soc/codecs/arizona.c b/sound/soc/codecs/arizona.c index 1bc6f25..0c2c453 100644 --- a/sound/soc/codecs/arizona.c +++ b/sound/soc/codecs/arizona.c @@ -1004,6 +1004,9 @@ int arizona_set_fll(struct arizona_fll *fll, int source, if (ena) pm_runtime_put_autosuspend(arizona->dev); + fll->fref = Fref; + fll->fout = Fout; + return 0; } -- 1.7.10.4

1 0

[alsa-devel] [FT C400,PATCH RFC,v3 0/9] M-Audio Fast Track C400
by Eldad Zack 28 Nov '12

28 Nov '12

Hi, The following patches adds support for the M-Audio Fast Track C400. This is version 3 of the patch series. v2: http://mailman.alsa-project.org/pipermail/alsa-devel/2012-November/056750.h… This series applies against the latest mainline tree, 3.7-rc7 (HEAD 2844a48706e54ddda4a04269dba4250b42f449de ) * First of all, everything works: mixer controls, effects control, playback and capture (with the correct sampling frequency set). Note that the device doesn't really have a master control or left/right pan. It's just "emulated" in the software. * To deal with the mixer channel IDs, limited by cmask bit length, I introduced an offset value. This allows to get rid of the shorehorning of the IDs into cmask and device-referencing in the low-level functions. Is this a good approach? * Audio in/out seems to work good in this series, mainly because the playback is an implicit feedback endpoint, and it is now configured with the capture endpoint as its sync-buddy. Also needed to correct the calculation, because of the different channel numbers between the two endpoints on the C400 (6 playback - implicit feedback, 4 capture). * The clock sources are still named "Unit 129" (0x81, internal) and "Unit 130" (0x82, SPDIF). I'll fix this later. I also plan to fix the channel names (i.e., instead of "AIn3", use "SPDIF In L"). * I've touched one FTU mixer creation function. I'd appreciate if someone can test this with the FTU to make sure I didn't break something. (I haven't gotten any feedback on that one yet) Cheers, Eldad Signed-off-by: Eldad Zack <eldad(a)fogrefinery.com> Eldad Zack (9): usb-audio: replace hardcoded value with const usb-audio: use sender stride for implicit feedback usb-audio: add control index offset usb-audio: skip UAC2 EFFECT_UNIT usb-audio: parameterize FTU effect unit control usb-audio: M-Audio Fast Track C400 quirks table usb-audio: Fast Track C400 mixer ranges usb-audio: Fast Track C400 mixer controls usb-audio: FT C400 sync playback EP to capture EP sound/usb/endpoint.c | 9 ++- sound/usb/mixer.c | 53 +++++++++++- sound/usb/mixer.h | 1 + sound/usb/mixer_quirks.c | 216 ++++++++++++++++++++++++++++++++++++++++++++-- sound/usb/pcm.c | 15 +++- sound/usb/quirks-table.h | 71 +++++++++++++++ 6 files changed, 351 insertions(+), 14 deletions(-) -- 1.7.8.6

3 14

[alsa-devel] SOC that needs L/R data in separate buffers
by jonsmirl＠gmail.com 28 Nov '12

28 Nov '12

Is there a format in ALSA for hardware that wants the L/R stereo data in two different buffers and not inter-leaved? Hardware has two DMA channels into separate FIFOs for L/R. If not I guess I have to unpack it inside the device driver. Doing that will mess up mixing since ALSA won't know where the real DMA buffer is. FIFOs are only four samples deeps so that's probably too many interrupts/sec to do programmed IO when FIFO is half full. -- Jon Smirl jonsmirl(a)gmail.com

2 1

[alsa-devel] Non-DMA ASoC driver
by Craig McQueen 28 Nov '12

28 Nov '12

I'm exploring hardware options for an audio interface for a Freescale i.MX28-based embedded Linux device. One option is to use a Synchronous Serial Port interface to connect in slave mode to an external CODEC. I have several doubts about the feasibility of this option, but for now, I have just one question: If we use this SSP, then we can't use DMA, but we would have to use interrupts for audio transfer. But I can't find any example code for a non-DMA based ASoC driver. Is it feasible to make a non-DMA based ASoC driver? How would it be implemented? All the ASoC drivers I see use a machine driver that declares a struct snd_soc_card with field .platform that points to a struct snd_soc_platform object, which implements the DMA stuff. If I don't use DMA, do I still need to somehow make a non-DMA platform driver that implements a struct snd_soc_platform object and handles the interrupts? Or, should I handle the interrupts and audio buffer manipulation directly in the CPU DAI driver? (My current guess is that this is feasible, but charting new territory, so we'll probably end up using different hardware design that allows us to do a more conventional DMA-based driver.) Thanks, Craig McQueen

2 1