July 2022 - Alsa-devel - mailman.alsa-project.org

[PATCH AUTOSEL 5.18 43/54] ASoC: qdsp6: fix potential memory leak in q6apm_get_audioreach_graph()
by Sasha Levin 20 Jul '22

20 Jul '22

From: Jianglei Nie <niejianglei2021(a)163.com> [ Upstream commit df5b4aca7248dc5a5cae93f162eae0decf972e48 ] q6apm_get_audioreach_graph() allocates a memory chunk for graph->graph with audioreach_alloc_graph_pkt(). When idr_alloc() fails, graph->graph is not released, which will lead to a memory leak. We can release the graph->graph with kfree() when idr_alloc() fails to fix the memory leak. Signed-off-by: Jianglei Nie <niejianglei2021(a)163.com> Reviewed-by: Srinivas Kandagatla <srinivas.kandagatla(a)linaro.org> Link: https://lore.kernel.org/r/20220629182520.2164409-1-niejianglei2021@163.com Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/qcom/qdsp6/q6apm.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/qcom/qdsp6/q6apm.c b/sound/soc/qcom/qdsp6/q6apm.c index f424d7aa389a..794019286c70 100644 --- a/sound/soc/qcom/qdsp6/q6apm.c +++ b/sound/soc/qcom/qdsp6/q6apm.c @@ -75,6 +75,7 @@ static struct audioreach_graph *q6apm_get_audioreach_graph(struct q6apm *apm, ui id = idr_alloc(&apm->graph_idr, graph, graph_id, graph_id + 1, GFP_KERNEL); if (id < 0) { dev_err(apm->dev, "Unable to allocate graph id (%d)\n", graph_id); + kfree(graph->graph); kfree(graph); mutex_unlock(&apm->lock); return ERR_PTR(id); -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 42/54] ASoC: arizona: Update arizona_aif_cfg_changed to use RX_BCLK_RATE
by Sasha Levin 20 Jul '22

20 Jul '22

From: Charles Keepax <ckeepax(a)opensource.cirrus.com> [ Upstream commit f99e930655f411453170a5f332e12c2d2748822e ] Currently the function arizona_aif_cfg_changed uses the TX_BCLK_RATE, however this register is not used on wm8998. This was not noticed as previously snd_soc_component_read did not print an error message. However, now the log gets filled with error messages, further more the test for if the LRCLK changed will return spurious results. Update the code to use the RX_BCLK_RATE register, the LRCLK parameters are written to both registers and the RX_BCLK_RATE register is used across all Arizona devices. Signed-off-by: Charles Keepax <ckeepax(a)opensource.cirrus.com> Link: https://lore.kernel.org/r/20220628153409.3266932-4-ckeepax@opensource.cirru… Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/arizona.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sound/soc/codecs/arizona.c b/sound/soc/codecs/arizona.c index e32871b3f68a..7434aeeda292 100644 --- a/sound/soc/codecs/arizona.c +++ b/sound/soc/codecs/arizona.c @@ -1760,8 +1760,8 @@ static bool arizona_aif_cfg_changed(struct snd_soc_component *component, if (bclk != (val & ARIZONA_AIF1_BCLK_FREQ_MASK)) return true; - val = snd_soc_component_read(component, base + ARIZONA_AIF_TX_BCLK_RATE); - if (lrclk != (val & ARIZONA_AIF1TX_BCPF_MASK)) + val = snd_soc_component_read(component, base + ARIZONA_AIF_RX_BCLK_RATE); + if (lrclk != (val & ARIZONA_AIF1RX_BCPF_MASK)) return true; val = snd_soc_component_read(component, base + ARIZONA_AIF_FRAME_CTRL_1); -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 41/54] ASoC: cs47l92: Fix event generation for OUT1 demux
by Sasha Levin 20 Jul '22

20 Jul '22

From: Charles Keepax <ckeepax(a)opensource.cirrus.com> [ Upstream commit 870d72ab9228575b2f005c9a23ea08787e0f63e6 ] cs47l92_put_demux returns the value of snd_soc_dapm_mux_update_power, which returns a 1 if a path was found for the kcontrol. This is obviously different to the expected return a 1 if the control was updated value. This results in spurious notifications to user-space. Update the handling to only return a 1 when the value is changed. Signed-off-by: Charles Keepax <ckeepax(a)opensource.cirrus.com> Link: https://lore.kernel.org/r/20220628153409.3266932-3-ckeepax@opensource.cirru… Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/cs47l92.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/sound/soc/codecs/cs47l92.c b/sound/soc/codecs/cs47l92.c index a1b8dcdb9f7b..444026b7d54b 100644 --- a/sound/soc/codecs/cs47l92.c +++ b/sound/soc/codecs/cs47l92.c @@ -119,7 +119,13 @@ static int cs47l92_put_demux(struct snd_kcontrol *kcontrol, end: snd_soc_dapm_mutex_unlock(dapm); - return snd_soc_dapm_mux_update_power(dapm, kcontrol, mux, e, NULL); + ret = snd_soc_dapm_mux_update_power(dapm, kcontrol, mux, e, NULL); + if (ret < 0) { + dev_err(madera->dev, "Failed to update demux power state: %d\n", ret); + return ret; + } + + return change; } static SOC_ENUM_SINGLE_DECL(cs47l92_outdemux_enum, -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 40/54] ASoC: wm8998: Fix event generation for input mux
by Sasha Levin 20 Jul '22

20 Jul '22

From: Charles Keepax <ckeepax(a)opensource.cirrus.com> [ Upstream commit 15b2e5d10ccf32a1a1ae7c636511e2f51320fdb5 ] wm8998_inmux_put returns the value of snd_soc_dapm_mux_update_power, which returns a 1 if a path was found for the kcontrol. This is obviously different to the expected return a 1 if the control was updated value. This results in spurious notifications to user-space. Update the handling to only return a 1 when the value is changed. Signed-off-by: Charles Keepax <ckeepax(a)opensource.cirrus.com> Link: https://lore.kernel.org/r/20220628153409.3266932-2-ckeepax@opensource.cirru… Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/wm8998.c | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-) diff --git a/sound/soc/codecs/wm8998.c b/sound/soc/codecs/wm8998.c index 00b59fc9b1fe..ab5481187c71 100644 --- a/sound/soc/codecs/wm8998.c +++ b/sound/soc/codecs/wm8998.c @@ -108,6 +108,7 @@ static int wm8998_inmux_put(struct snd_kcontrol *kcontrol, struct soc_enum *e = (struct soc_enum *)kcontrol->private_value; unsigned int mode_reg, mode_index; unsigned int mux, inmode, src_val, mode_val; + int change, ret; mux = ucontrol->value.enumerated.item[0]; if (mux > 1) @@ -137,14 +138,20 @@ static int wm8998_inmux_put(struct snd_kcontrol *kcontrol, snd_soc_component_update_bits(component, mode_reg, ARIZONA_IN1_MODE_MASK, mode_val); - snd_soc_component_update_bits(component, e->reg, - ARIZONA_IN1L_SRC_MASK | - ARIZONA_IN1L_SRC_SE_MASK, - src_val); + change = snd_soc_component_update_bits(component, e->reg, + ARIZONA_IN1L_SRC_MASK | + ARIZONA_IN1L_SRC_SE_MASK, + src_val); - return snd_soc_dapm_mux_update_power(dapm, kcontrol, - ucontrol->value.enumerated.item[0], - e, NULL); + ret = snd_soc_dapm_mux_update_power(dapm, kcontrol, + ucontrol->value.enumerated.item[0], + e, NULL); + if (ret < 0) { + dev_err(arizona->dev, "Failed to update demux power state: %d\n", ret); + return ret; + } + + return change; } static const char * const wm8998_inmux_texts[] = { -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 39/54] ASoC: wm5102: Fix event generation for output compensation
by Sasha Levin 20 Jul '22

20 Jul '22

From: Charles Keepax <ckeepax(a)opensource.cirrus.com> [ Upstream commit 71b5ab96ffe6589abe7a2e302b83f7a426ebe099 ] The output compensation controls always returns zero regardless of if the control value was updated. This results in missing notifications to user-space of the control change. Update the handling to return 1 when the value is changed. Signed-off-by: Charles Keepax <ckeepax(a)opensource.cirrus.com> Link: https://lore.kernel.org/r/20220628153409.3266932-1-ckeepax@opensource.cirru… Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/wm5102.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-) diff --git a/sound/soc/codecs/wm5102.c b/sound/soc/codecs/wm5102.c index da2f8998df87..b034df47a5ef 100644 --- a/sound/soc/codecs/wm5102.c +++ b/sound/soc/codecs/wm5102.c @@ -680,12 +680,17 @@ static int wm5102_out_comp_coeff_put(struct snd_kcontrol *kcontrol, { struct snd_soc_component *component = snd_soc_kcontrol_component(kcontrol); struct arizona *arizona = dev_get_drvdata(component->dev->parent); + uint16_t dac_comp_coeff = get_unaligned_be16(ucontrol->value.bytes.data); + int ret = 0; mutex_lock(&arizona->dac_comp_lock); - arizona->dac_comp_coeff = get_unaligned_be16(ucontrol->value.bytes.data); + if (arizona->dac_comp_coeff != dac_comp_coeff) { + arizona->dac_comp_coeff = dac_comp_coeff; + ret = 1; + } mutex_unlock(&arizona->dac_comp_lock); - return 0; + return ret; } static int wm5102_out_comp_switch_get(struct snd_kcontrol *kcontrol, @@ -706,12 +711,20 @@ static int wm5102_out_comp_switch_put(struct snd_kcontrol *kcontrol, { struct snd_soc_component *component = snd_soc_kcontrol_component(kcontrol); struct arizona *arizona = dev_get_drvdata(component->dev->parent); + struct soc_mixer_control *mc = (struct soc_mixer_control *)kcontrol->private_value; + int ret = 0; + + if (ucontrol->value.integer.value[0] > mc->max) + return -EINVAL; mutex_lock(&arizona->dac_comp_lock); - arizona->dac_comp_enabled = ucontrol->value.integer.value[0]; + if (arizona->dac_comp_enabled != ucontrol->value.integer.value[0]) { + arizona->dac_comp_enabled = ucontrol->value.integer.value[0]; + ret = 1; + } mutex_unlock(&arizona->dac_comp_lock); - return 0; + return ret; } static const char * const wm5102_osr_text[] = { -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 38/54] ASoC: wcd9335: Use int array instead of bitmask for TX mixers
by Sasha Levin 20 Jul '22

20 Jul '22

From: Yassine Oudjana <y.oudjana(a)protonmail.com> [ Upstream commit a5d6d28e2ea38dff017cb562dfbe0259d093a851 ] Currently slim_tx_mixer_get reports all TX mixers as enabled when at least one is, due to it reading the entire tx_port_value bitmask without testing the specific bit corresponding to a TX port. Furthermore, using the same bitmask for all capture DAIs makes setting one mixer affect them all. To prevent this, and since the SLIM TX muxes effectively only connect to one of the mixers at a time, turn tx_port_value into an int array storing the DAI index each of the ports is connected to. Signed-off-by: Yassine Oudjana <y.oudjana(a)protonmail.com> Link: https://lore.kernel.org/r/20220622061745.35399-1-y.oudjana@protonmail.com Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/wcd9335.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/sound/soc/codecs/wcd9335.c b/sound/soc/codecs/wcd9335.c index 1e60db4056ad..ad887c9c7573 100644 --- a/sound/soc/codecs/wcd9335.c +++ b/sound/soc/codecs/wcd9335.c @@ -342,7 +342,7 @@ struct wcd9335_codec { struct regulator_bulk_data supplies[WCD9335_MAX_SUPPLY]; unsigned int rx_port_value[WCD9335_RX_MAX]; - unsigned int tx_port_value; + unsigned int tx_port_value[WCD9335_TX_MAX]; int hph_l_gain; int hph_r_gain; u32 rx_bias_count; @@ -1328,8 +1328,13 @@ static int slim_tx_mixer_get(struct snd_kcontrol *kc, struct snd_soc_dapm_context *dapm = snd_soc_dapm_kcontrol_dapm(kc); struct wcd9335_codec *wcd = dev_get_drvdata(dapm->dev); + struct snd_soc_dapm_widget *widget = snd_soc_dapm_kcontrol_widget(kc); + struct soc_mixer_control *mixer = + (struct soc_mixer_control *)kc->private_value; + int dai_id = widget->shift; + int port_id = mixer->shift; - ucontrol->value.integer.value[0] = wcd->tx_port_value; + ucontrol->value.integer.value[0] = wcd->tx_port_value[port_id] == dai_id; return 0; } @@ -1352,12 +1357,12 @@ static int slim_tx_mixer_put(struct snd_kcontrol *kc, case AIF2_CAP: case AIF3_CAP: /* only add to the list if value not set */ - if (enable && !(wcd->tx_port_value & BIT(port_id))) { - wcd->tx_port_value |= BIT(port_id); + if (enable && wcd->tx_port_value[port_id] != dai_id) { + wcd->tx_port_value[port_id] = dai_id; list_add_tail(&wcd->tx_chs[port_id].list, &wcd->dai[dai_id].slim_ch_list); - } else if (!enable && (wcd->tx_port_value & BIT(port_id))) { - wcd->tx_port_value &= ~BIT(port_id); + } else if (!enable && wcd->tx_port_value[port_id] == dai_id) { + wcd->tx_port_value[port_id] = -1; list_del_init(&wcd->tx_chs[port_id].list); } break; -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 37/54] ASoC: tlv320adcx140: Fix tx_mask check
by Sasha Levin 20 Jul '22

20 Jul '22

From: Sascha Hauer <s.hauer(a)pengutronix.de> [ Upstream commit 7d90c8e6396ba245da16bedd789df6d669375408 ] The tx_mask check doesn't reflect what the driver and the chip support. The check currently checks for exactly two slots being enabled. The tlv320adcx140 supports anything between one and eight channels, so relax the check accordingly. The tlv320adcx140 supports arbitrary tx_mask settings, but the driver currently only supports adjacent slots beginning with the first slot, so extend the check to check that the first slot is being used and that there are no holes in the tx_mask. Leave a comment to make it's the driver that limits the tx_mask settings, not the chip itself. While at it remove the set-but-unused struct adcx140p_priv::tdm_delay field. Signed-off-by: Sascha Hauer <s.hauer(a)pengutronix.de> Link: https://lore.kernel.org/r/20220624105716.2579539-1-s.hauer@pengutronix.de Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/tlv320adcx140.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/sound/soc/codecs/tlv320adcx140.c b/sound/soc/codecs/tlv320adcx140.c index 32b120d624b2..bf18fcfcaeb9 100644 --- a/sound/soc/codecs/tlv320adcx140.c +++ b/sound/soc/codecs/tlv320adcx140.c @@ -33,7 +33,6 @@ struct adcx140_priv { bool micbias_vg; unsigned int dai_fmt; - unsigned int tdm_delay; unsigned int slot_width; }; @@ -792,12 +791,13 @@ static int adcx140_set_dai_tdm_slot(struct snd_soc_dai *codec_dai, { struct snd_soc_component *component = codec_dai->component; struct adcx140_priv *adcx140 = snd_soc_component_get_drvdata(component); - unsigned int lsb; - /* TDM based on DSP mode requires slots to be adjacent */ - lsb = __ffs(tx_mask); - if ((lsb + 1) != __fls(tx_mask)) { - dev_err(component->dev, "Invalid mask, slots must be adjacent\n"); + /* + * The chip itself supports arbitrary masks, but the driver currently + * only supports adjacent slots beginning at the first slot. + */ + if (tx_mask != GENMASK(__fls(tx_mask), 0)) { + dev_err(component->dev, "Only lower adjacent slots are supported\n"); return -EINVAL; } @@ -812,7 +812,6 @@ static int adcx140_set_dai_tdm_slot(struct snd_soc_dai *codec_dai, return -EINVAL; } - adcx140->tdm_delay = lsb; adcx140->slot_width = slot_width; return 0; -- 2.35.1

1 0

[PATCH AUTOSEL 5.18 36/54] ASoC: ti: omap-mcbsp: duplicate sysfs error
by Sasha Levin 20 Jul '22

20 Jul '22

From: David Owens <daowens01(a)gmail.com> [ Upstream commit f0d96937d31c4615a6418e4bed5cee50a952040e ] Convert to managed versions of sysfs and clk allocation to simplify unbinding and error handling in probe. Managed sysfs node creation specifically addresses the following error seen the second time probe is attempted after sdma_pcm_platform_register() previously requsted probe deferral: sysfs: cannot create duplicate filename '/devices/platform/68000000.ocp/49022000.mcbsp/max_tx_thres' Signed-off-by: David Owens <dowens(a)precisionplanting.com> Link: https://lore.kernel.org/r/20220620183744.3176557-1-dowens@precisionplanting… Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/ti/omap-mcbsp-priv.h | 2 -- sound/soc/ti/omap-mcbsp-st.c | 14 ++------------ sound/soc/ti/omap-mcbsp.c | 19 ++----------------- 3 files changed, 4 insertions(+), 31 deletions(-) diff --git a/sound/soc/ti/omap-mcbsp-priv.h b/sound/soc/ti/omap-mcbsp-priv.h index 7865cda4bf0a..da519ea1f303 100644 --- a/sound/soc/ti/omap-mcbsp-priv.h +++ b/sound/soc/ti/omap-mcbsp-priv.h @@ -316,8 +316,6 @@ static inline int omap_mcbsp_read(struct omap_mcbsp *mcbsp, u16 reg, /* Sidetone specific API */ int omap_mcbsp_st_init(struct platform_device *pdev); -void omap_mcbsp_st_cleanup(struct platform_device *pdev); - int omap_mcbsp_st_start(struct omap_mcbsp *mcbsp); int omap_mcbsp_st_stop(struct omap_mcbsp *mcbsp); diff --git a/sound/soc/ti/omap-mcbsp-st.c b/sound/soc/ti/omap-mcbsp-st.c index 0bc7d26c660a..7e8179cae92e 100644 --- a/sound/soc/ti/omap-mcbsp-st.c +++ b/sound/soc/ti/omap-mcbsp-st.c @@ -347,7 +347,7 @@ int omap_mcbsp_st_init(struct platform_device *pdev) if (!st_data) return -ENOMEM; - st_data->mcbsp_iclk = clk_get(mcbsp->dev, "ick"); + st_data->mcbsp_iclk = devm_clk_get(mcbsp->dev, "ick"); if (IS_ERR(st_data->mcbsp_iclk)) { dev_warn(mcbsp->dev, "Failed to get ick, sidetone might be broken\n"); @@ -359,7 +359,7 @@ int omap_mcbsp_st_init(struct platform_device *pdev) if (!st_data->io_base_st) return -ENOMEM; - ret = sysfs_create_group(&mcbsp->dev->kobj, &sidetone_attr_group); + ret = devm_device_add_group(mcbsp->dev, &sidetone_attr_group); if (ret) return ret; @@ -368,16 +368,6 @@ int omap_mcbsp_st_init(struct platform_device *pdev) return 0; } -void omap_mcbsp_st_cleanup(struct platform_device *pdev) -{ - struct omap_mcbsp *mcbsp = platform_get_drvdata(pdev); - - if (mcbsp->st_data) { - sysfs_remove_group(&mcbsp->dev->kobj, &sidetone_attr_group); - clk_put(mcbsp->st_data->mcbsp_iclk); - } -} - static int omap_mcbsp_st_info_volsw(struct snd_kcontrol *kcontrol, struct snd_ctl_elem_info *uinfo) { diff --git a/sound/soc/ti/omap-mcbsp.c b/sound/soc/ti/omap-mcbsp.c index 4479d74f0a45..9933b33c80ca 100644 --- a/sound/soc/ti/omap-mcbsp.c +++ b/sound/soc/ti/omap-mcbsp.c @@ -702,8 +702,7 @@ static int omap_mcbsp_init(struct platform_device *pdev) mcbsp->max_tx_thres = max_thres(mcbsp) - 0x10; mcbsp->max_rx_thres = max_thres(mcbsp) - 0x10; - ret = sysfs_create_group(&mcbsp->dev->kobj, - &additional_attr_group); + ret = devm_device_add_group(mcbsp->dev, &additional_attr_group); if (ret) { dev_err(mcbsp->dev, "Unable to create additional controls\n"); @@ -711,16 +710,7 @@ static int omap_mcbsp_init(struct platform_device *pdev) } } - ret = omap_mcbsp_st_init(pdev); - if (ret) - goto err_st; - - return 0; - -err_st: - if (mcbsp->pdata->buffer_size) - sysfs_remove_group(&mcbsp->dev->kobj, &additional_attr_group); - return ret; + return omap_mcbsp_st_init(pdev); } /* @@ -1431,11 +1421,6 @@ static int asoc_mcbsp_remove(struct platform_device *pdev) if (cpu_latency_qos_request_active(&mcbsp->pm_qos_req)) cpu_latency_qos_remove_request(&mcbsp->pm_qos_req); - if (mcbsp->pdata->buffer_size) - sysfs_remove_group(&mcbsp->dev->kobj, &additional_attr_group); - - omap_mcbsp_st_cleanup(pdev); - return 0; } -- 2.35.1

1 0

Behringer UMC204HD HiFi "Default" profile crackling audio
by GitHub issues - opened 20 Jul '22

20 Jul '22

alsa-project/alsa-ucm-conf issue #192 was opened from acepukas: Since the recent update to 1.2.7.2, the Behringer UMC204HD "Default" profile has crackling audio through "Line A" and "Line B" ports. Switching to the "Direct" profile allows for clean audio but it's output over 4 speakers (front left and right, rear left and right). Not really sure how to trouble shoot this. Issue URL : https://github.com/alsa-project/alsa-ucm-conf/issues/192 Repository URL: https://github.com/alsa-project/alsa-ucm-conf

1 0

ASoC: Is this bug ? DAPM source vs sink
by Kuninori Morimoto 20 Jul '22

20 Jul '22

Hi ASoC Cc Lars-Peter I have noticed that DAPM source vs sink are strange. ASoC had used source/sink or input/output before, but it was exchanged by this patch a3423b02cf745c1418f1f199646c450d6fc9ca4e ("ASoC: dapm: Consolidate input and output path handling") I think source/sink pair are converted to SND_SOC_DAPM_DIR_IN/OUT by this patch. source <=> IN sink <=> OUT But, its for_each are inverted #define snd_soc_dapm_widget_for_each_sink_path(w, p) \ snd_soc_dapm_widget_for_each_path(w, SND_SOC_DAPM_DIR_IN, p) #define snd_soc_dapm_widget_for_each_source_path(w, p) \ snd_soc_dapm_widget_for_each_path(w, SND_SOC_DAPM_DIR_OUT, p) I have thought this for_each is for special handling, but, the original was sink is for sink, source is for source after all. #define snd_soc_dapm_widget_for_each_sink_path(w, p) \ list_for_each_entry(p, &w->sinks, list_source) #define snd_soc_dapm_widget_for_each_source_path(w, p) \ list_for_each_entry(p, &w->sources, list_sink) So, I guess current for_each is wrong, but I'm not 100% sure #define snd_soc_dapm_widget_for_each_sink_path(w, p) \ - snd_soc_dapm_widget_for_each_path(w, SND_SOC_DAPM_DIR_IN, p) + snd_soc_dapm_widget_for_each_path(w, SND_SOC_DAPM_DIR_OUT, p) #define snd_soc_dapm_widget_for_each_source_path(w, p) \ - snd_soc_dapm_widget_for_each_path(w, SND_SOC_DAPM_DIR_OUT, p) + snd_soc_dapm_widget_for_each_path(w, SND_SOC_DAPM_DIR_IN, p) I'm happy to create patch if I could get your ack. Thank you for your help !! Best regards --- Kuninori Morimoto

3 3