July 2015 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH] ASoC: xtensa: Remove unnecessary snd_pcm_lib_preallocate_free_for_all()
by Vaishali Thakkar 23 Jul '15

23 Jul '15

The ALSA core takes care that all preallocated memory is freed when the PCM itself is freed. There is no need to do this manually in the driver. Signed-off-by: Vaishali Thakkar <vthakkar1994(a)gmail.com> --- sound/soc/xtensa/xtfpga-i2s.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/sound/soc/xtensa/xtfpga-i2s.c b/sound/soc/xtensa/xtfpga-i2s.c index 1cfb19e..039f65e 100644 --- a/sound/soc/xtensa/xtfpga-i2s.c +++ b/sound/soc/xtensa/xtfpga-i2s.c @@ -474,11 +474,6 @@ static int xtfpga_pcm_new(struct snd_soc_pcm_runtime *rtd) card->dev, size, size); } -static void xtfpga_pcm_free(struct snd_pcm *pcm) -{ - snd_pcm_lib_preallocate_free_for_all(pcm); -} - static const struct snd_pcm_ops xtfpga_pcm_ops = { .open = xtfpga_pcm_open, .close = xtfpga_pcm_close, @@ -490,7 +485,6 @@ static const struct snd_pcm_ops xtfpga_pcm_ops = { static const struct snd_soc_platform_driver xtfpga_soc_platform = { .pcm_new = xtfpga_pcm_new, - .pcm_free = xtfpga_pcm_free, .ops = &xtfpga_pcm_ops, }; -- 1.9.1

2 1

[alsa-devel] [PATCHv2 0/3] regmap: Use different lockdep class for each regmap init call
by Nicolas Boichat 23 Jul '15

23 Jul '15

Second version. Small fixup in the regmap patch itself [3/3] (the declaration of __devm_regmap_init_ac97 was missing lock_key/name parameters). Also, there are 2 drivers that are using function pointers named "regmap_init", which causes problems as it is now a macro: put these in parentheses to prevent the preprocessor from substituting them. Patches 1+2 should be pushed first to avoid breaking git bissect. Compile tested with allyesconfig on x86_64 and arm. Nicolas Boichat (3): mfd: vexpress: Add parentheses around bridge->ops->regmap_init call thermal: sti: Add parentheses around bridge->ops->regmap_init call regmap: Use different lockdep class for each regmap init call drivers/base/regmap/regmap-ac97.c | 22 +++-- drivers/base/regmap/regmap-i2c.c | 22 +++-- drivers/base/regmap/regmap-mmio.c | 27 ++++-- drivers/base/regmap/regmap-spi.c | 22 +++-- drivers/base/regmap/regmap-spmi.c | 44 +++++---- drivers/base/regmap/regmap.c | 31 +++--- drivers/bus/vexpress-config.c | 2 +- drivers/thermal/st/st_thermal.c | 2 +- include/linux/regmap.h | 192 ++++++++++++++++++++++++++++---------- 9 files changed, 252 insertions(+), 112 deletions(-) -- 2.4.3.573.g4eafbef

3 6

[alsa-devel] [PATCH] ASoC: rt5677: Remove NULL test for desc before gpiod_set_value_cansleep call
by Axel Lin 23 Jul '15

23 Jul '15

It's safe to call gpiod_set_value_cansleep() with NULL desc. Signed-off-by: Axel Lin <axel.lin(a)ingics.com> --- sound/soc/codecs/rt5677.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) diff --git a/sound/soc/codecs/rt5677.c b/sound/soc/codecs/rt5677.c index 016db7d..8d70911 100644 --- a/sound/soc/codecs/rt5677.c +++ b/sound/soc/codecs/rt5677.c @@ -4763,10 +4763,8 @@ static int rt5677_remove(struct snd_soc_codec *codec) struct rt5677_priv *rt5677 = snd_soc_codec_get_drvdata(codec); regmap_write(rt5677->regmap, RT5677_RESET, 0x10ec); - if (rt5677->pow_ldo2) - gpiod_set_value_cansleep(rt5677->pow_ldo2, 0); - if (rt5677->reset_pin) - gpiod_set_value_cansleep(rt5677->reset_pin, 0); + gpiod_set_value_cansleep(rt5677->pow_ldo2, 0); + gpiod_set_value_cansleep(rt5677->reset_pin, 0); return 0; } @@ -4780,10 +4778,8 @@ static int rt5677_suspend(struct snd_soc_codec *codec) regcache_cache_only(rt5677->regmap, true); regcache_mark_dirty(rt5677->regmap); - if (rt5677->pow_ldo2) - gpiod_set_value_cansleep(rt5677->pow_ldo2, 0); - if (rt5677->reset_pin) - gpiod_set_value_cansleep(rt5677->reset_pin, 0); + gpiod_set_value_cansleep(rt5677->pow_ldo2, 0); + gpiod_set_value_cansleep(rt5677->reset_pin, 0); } return 0; @@ -4794,10 +4790,8 @@ static int rt5677_resume(struct snd_soc_codec *codec) struct rt5677_priv *rt5677 = snd_soc_codec_get_drvdata(codec); if (!rt5677->dsp_vad_en) { - if (rt5677->pow_ldo2) - gpiod_set_value_cansleep(rt5677->pow_ldo2, 1); - if (rt5677->reset_pin) - gpiod_set_value_cansleep(rt5677->reset_pin, 1); + gpiod_set_value_cansleep(rt5677->pow_ldo2, 1); + gpiod_set_value_cansleep(rt5677->reset_pin, 1); if (rt5677->pow_ldo2 || rt5677->reset_pin) msleep(10); -- 2.1.0

2 1

[alsa-devel] [PATCH] ASoC: tas2552: Fix off-by-one for max_register setting
by Axel Lin 23 Jul '15

23 Jul '15

The latest valid register is TAS2552_VBAT_DATA. Signed-off-by: Axel Lin <axel.lin(a)ingics.com> --- sound/soc/codecs/tas2552.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/tas2552.h b/sound/soc/codecs/tas2552.h index 5746f8f..e34752b 100644 --- a/sound/soc/codecs/tas2552.h +++ b/sound/soc/codecs/tas2552.h @@ -42,7 +42,7 @@ #define TAS2552_BOOST_APT_CTRL 0x14 #define TAS2552_VER_NUM 0x16 #define TAS2552_VBAT_DATA 0x19 -#define TAS2552_MAX_REG 0x20 +#define TAS2552_MAX_REG TAS2552_VBAT_DATA /* CFG1 Register Masks */ #define TAS2552_DEV_RESET (1 << 0) -- 2.1.0

3 2

[alsa-devel] [PATCH] ASoC: tas5086: Improve the logic for de-emphasis sampling rate selection
by Axel Lin 23 Jul '15

23 Jul '15

Slightly improve the logic for de-emphasis sampling rate selection by break out the loop if the rate is matched. Signed-off-by: Axel Lin <axel.lin(a)ingics.com> --- sound/soc/codecs/tas5086.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/sound/soc/codecs/tas5086.c b/sound/soc/codecs/tas5086.c index 0a49fc8..d49d25d 100644 --- a/sound/soc/codecs/tas5086.c +++ b/sound/soc/codecs/tas5086.c @@ -266,10 +266,14 @@ static int tas5086_set_deemph(struct snd_soc_codec *codec) struct tas5086_private *priv = snd_soc_codec_get_drvdata(codec); int i, val = 0; - if (priv->deemph) - for (i = 0; i < ARRAY_SIZE(tas5086_deemph); i++) - if (tas5086_deemph[i] == priv->rate) + if (priv->deemph) { + for (i = 0; i < ARRAY_SIZE(tas5086_deemph); i++) { + if (tas5086_deemph[i] == priv->rate) { val = i; + break; + } + } + } return regmap_update_bits(priv->regmap, TAS5086_SYS_CONTROL_1, TAS5086_DEEMPH_MASK, val); -- 2.1.0

3 2

Re: [alsa-devel] Intel Baytrail Linux Audio
by Vinod Koul 23 Jul '15

23 Jul '15

On Thu, Jul 09, 2015 at 02:57:58PM -0400, Tyler Yeomans wrote: > Antonio/Vinod, > > I have made the change in the sst-acpi.c file for the atom specific boards. > I've attached the dmesg output here with dynamic debug enabled. > > To clarify- > > snd-intel-sst-core is loaded > FE-BE fix has been applied > I'm using the ssp0 firmware. > > Interrupts are still 0 > > Still no sound. > > Would leaving the device on and providing remote access (i.e though > teamviewer) assist in any way? The fw download seems to fail on your device, which can also be due to incorrect interrupt mapping... [ 20.386850] intel_sst_acpi 80860F28:00: Wait timed-out condition:0x0, msg_id:0x1 fw_state 0x3 [ 20.386862] intel_sst_acpi 80860F28:00: fw returned err -16 [ 20.386873] sst-mfld-platform sst-mfld-platform: Enter: enable=1 port_name=ssp2-port [ 20.752559] intel_sst_acpi 80860F28:00: Wait timed-out condition:0x0, msg_id:0x1 fw_state 0x3 [ 20.752570] intel_sst_acpi 80860F28:00: fw returned err -16 [ 21.163378] intel_sst_acpi 80860F28:00: Wait timed-out condition:0x0, msg_id:0x0 fw_state 0x1 [ 21.163388] intel_sst_acpi 80860F28:00: fw download failed -16 Can you tell me which device you have, what was preinstalled on it -- ~Vinod

2 4

[alsa-devel] [PATCH] ASoC: mediatek: Add suspend/resume callbacks
by Koro Chen 23 Jul '15

23 Jul '15

This adds suspend/resume callbacks, which are common for each DAI. To be able to continue the last playback/capture after resume when suspend was done during a playback/capture, in the callbacks we do backup/restore of registers which were set before prepare stage. Registers to be backup/restore are defined in a backup list array. Signed-off-by: Koro Chen <koro.chen(a)mediatek.com> --- sound/soc/mediatek/mtk-afe-common.h | 8 ---- sound/soc/mediatek/mtk-afe-pcm.c | 77 +++++++++++++++++++++++++++++++++++++ 2 files changed, 77 insertions(+), 8 deletions(-) diff --git a/sound/soc/mediatek/mtk-afe-common.h b/sound/soc/mediatek/mtk-afe-common.h index a88b175..cc4393c 100644 --- a/sound/soc/mediatek/mtk-afe-common.h +++ b/sound/soc/mediatek/mtk-afe-common.h @@ -98,12 +98,4 @@ struct mtk_afe_memif { const struct mtk_afe_irq_data *irqdata; }; -struct mtk_afe { - /* address for ioremap audio hardware register */ - void __iomem *base_addr; - struct device *dev; - struct regmap *regmap; - struct mtk_afe_memif memif[MTK_AFE_MEMIF_NUM]; - struct clk *clocks[MTK_CLK_NUM]; -}; #endif diff --git a/sound/soc/mediatek/mtk-afe-pcm.c b/sound/soc/mediatek/mtk-afe-pcm.c index de38d2e..55471ee 100644 --- a/sound/soc/mediatek/mtk-afe-pcm.c +++ b/sound/soc/mediatek/mtk-afe-pcm.c @@ -45,18 +45,21 @@ /* Memory interface */ #define AFE_DL1_BASE 0x0040 #define AFE_DL1_CUR 0x0044 +#define AFE_DL1_END 0x0048 #define AFE_DL2_BASE 0x0050 #define AFE_DL2_CUR 0x0054 #define AFE_AWB_BASE 0x0070 #define AFE_AWB_CUR 0x007c #define AFE_VUL_BASE 0x0080 #define AFE_VUL_CUR 0x008c +#define AFE_VUL_END 0x0088 #define AFE_DAI_BASE 0x0090 #define AFE_DAI_CUR 0x009c #define AFE_MOD_PCM_BASE 0x0330 #define AFE_MOD_PCM_CUR 0x033c #define AFE_HDMI_OUT_BASE 0x0374 #define AFE_HDMI_OUT_CUR 0x0378 +#define AFE_HDMI_OUT_END 0x037c #define AFE_ADDA2_TOP_CON0 0x0600 @@ -127,6 +130,34 @@ enum afe_tdm_ch_start { AFE_TDM_CH_ZERO, }; +static const unsigned int mtk_afe_backup_list[] = { + AUDIO_TOP_CON0, + AFE_CONN1, + AFE_CONN2, + AFE_CONN7, + AFE_CONN8, + AFE_DAC_CON1, + AFE_DL1_BASE, + AFE_DL1_END, + AFE_VUL_BASE, + AFE_VUL_END, + AFE_HDMI_OUT_BASE, + AFE_HDMI_OUT_END, + AFE_HDMI_CONN0, + AFE_DAC_CON0, +}; + +struct mtk_afe { + /* address for ioremap audio hardware register */ + void __iomem *base_addr; + struct device *dev; + struct regmap *regmap; + struct mtk_afe_memif memif[MTK_AFE_MEMIF_NUM]; + struct clk *clocks[MTK_CLK_NUM]; + unsigned int backup_regs[ARRAY_SIZE(mtk_afe_backup_list)]; + bool suspended; +}; + static const struct snd_pcm_hardware mtk_afe_hardware = { .info = (SNDRV_PCM_INFO_MMAP | SNDRV_PCM_INFO_INTERLEAVED | SNDRV_PCM_INFO_MMAP_VALID), @@ -722,11 +753,53 @@ static const struct snd_soc_dai_ops mtk_afe_hdmi_ops = { }; +static int mtk_afe_runtime_suspend(struct device *dev); +static int mtk_afe_runtime_resume(struct device *dev); + +static int mtk_afe_dai_suspend(struct snd_soc_dai *dai) +{ + struct mtk_afe *afe = snd_soc_dai_get_drvdata(dai); + int i; + + dev_dbg(afe->dev, "%s\n", __func__); + if (pm_runtime_status_suspended(afe->dev) || afe->suspended) + return 0; + + for (i = 0; i < ARRAY_SIZE(mtk_afe_backup_list); i++) + regmap_read(afe->regmap, mtk_afe_backup_list[i], + &afe->backup_regs[i]); + + afe->suspended = true; + mtk_afe_runtime_suspend(afe->dev); + return 0; +} + +static int mtk_afe_dai_resume(struct snd_soc_dai *dai) +{ + struct mtk_afe *afe = snd_soc_dai_get_drvdata(dai); + int i = 0; + + dev_dbg(afe->dev, "%s\n", __func__); + if (pm_runtime_status_suspended(afe->dev) || !afe->suspended) + return 0; + + mtk_afe_runtime_resume(afe->dev); + + for (i = 0; i < ARRAY_SIZE(mtk_afe_backup_list); i++) + regmap_write(afe->regmap, mtk_afe_backup_list[i], + afe->backup_regs[i]); + + afe->suspended = false; + return 0; +} + static struct snd_soc_dai_driver mtk_afe_pcm_dais[] = { /* FE DAIs: memory intefaces to CPU */ { .name = "DL1", /* downlink 1 */ .id = MTK_AFE_MEMIF_DL1, + .suspend = mtk_afe_dai_suspend, + .resume = mtk_afe_dai_resume, .playback = { .stream_name = "DL1", .channels_min = 1, @@ -738,6 +811,8 @@ static struct snd_soc_dai_driver mtk_afe_pcm_dais[] = { }, { .name = "VUL", /* voice uplink */ .id = MTK_AFE_MEMIF_VUL, + .suspend = mtk_afe_dai_suspend, + .resume = mtk_afe_dai_resume, .capture = { .stream_name = "VUL", .channels_min = 1, @@ -774,6 +849,8 @@ static struct snd_soc_dai_driver mtk_afe_hdmi_dais[] = { { .name = "HDMI", .id = MTK_AFE_MEMIF_HDMI, + .suspend = mtk_afe_dai_suspend, + .resume = mtk_afe_dai_resume, .playback = { .stream_name = "HDMI", .channels_min = 2, -- 1.8.1.1.dirty

2 1

[alsa-devel] [PATCH 1/2] ASoC: cs35l32: Fix define for CS35L32_SDOUT_3ST
by Axel Lin 23 Jul '15

23 Jul '15

According to the datasheet, the CS35L32_SDOUT_3ST is BIT(3). Signed-off-by: Axel Lin <axel.lin(a)ingics.com> --- sound/soc/codecs/cs35l32.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/cs35l32.h b/sound/soc/codecs/cs35l32.h index 31ab804..1d6c250 100644 --- a/sound/soc/codecs/cs35l32.h +++ b/sound/soc/codecs/cs35l32.h @@ -80,7 +80,7 @@ struct cs35l32_platform_data { #define CS35L32_GAIN_MGR_MASK 0x08 #define CS35L32_ADSP_SHARE_MASK 0x08 #define CS35L32_ADSP_DATACFG_MASK 0x30 -#define CS35L32_SDOUT_3ST 0x80 +#define CS35L32_SDOUT_3ST 0x08 #define CS35L32_BATT_REC_MASK 0x0E #define CS35L32_BATT_THRESH_MASK 0x30 -- 2.1.0

3 4

[alsa-devel] [PATCH] ASoC: cs42l73: Fix mask for setting CS42L73_SP_3ST bit
by Axel Lin 23 Jul '15

23 Jul '15

CS42L73_SP_3ST is BIT(7), so the mask field is wrong. Fix it. Signed-off-by: Axel Lin <axel.lin(a)ingics.com> --- sound/soc/codecs/cs42l73.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sound/soc/codecs/cs42l73.c b/sound/soc/codecs/cs42l73.c index 7cb1d70..be6477f 100644 --- a/sound/soc/codecs/cs42l73.c +++ b/sound/soc/codecs/cs42l73.c @@ -1236,8 +1236,8 @@ static int cs42l73_set_tristate(struct snd_soc_dai *dai, int tristate) struct snd_soc_codec *codec = dai->codec; int id = dai->id; - return snd_soc_update_bits(codec, CS42L73_SPC(id), - 0x7F, tristate << 7); + return snd_soc_update_bits(codec, CS42L73_SPC(id), CS42L73_SP_3ST, + tristate << 7); } static const struct snd_pcm_hw_constraint_list constraints_12_24 = { -- 2.1.0

3 2

[alsa-devel] [PATCH 0/4] i915 to call hda driver on HDMI plug/unplug
by David Henningsson 23 Jul '15

23 Jul '15

This patch set aims to resolve three problems: * The first - and most serious one - is that the audio driver is not woken up properly when in power save modes, especially not when the HDA controller is in D3. By having the i915 driver call directly into the hda driver, the HDA driver is always notified that an HDMI hotplug event has happened. * Second, there is currently no way for userspace to match an HDMI audio output with an HDMI video output. We fix this by sending connector_type and connector_type_id in the HDMI hotplug callback. * Third, writing ELD info to the hardware just so the HDA driver can read it from the hardware seems a bit inefficient. We could just pass that information in the callback, too. The patch in its current form fixes the first of these problems and provides most of the infrastructure for the second and third problem. The patch set is based on 4.2rc2 + my recent codec wakeup patch. So far, it has been tested (and working) on one Skylake machine. David Henningsson (4): drm/i915: Add audio hotplug info struct drm/i915: Call audio hotplug notify function ALSA: hda - Dispatch incoming HDMI hotplug i915 callback ALSA: hda - Wake the codec up on hotplug notify events drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_audio.c | 46 ++++++++++++++++++++++++++++++++++++ include/drm/i915_component.h | 19 +++++++++++++++ include/sound/hdaudio.h | 4 ++++ sound/hda/hdac_i915.c | 24 +++++++++++++++++++ sound/pci/hda/patch_hdmi.c | 22 ++++++++++++++++- 6 files changed, 115 insertions(+), 1 deletion(-) -- 1.7.9.5

5 22