[PATCH v1 0/1] soundwire: qcom: disable stop clock on 1.3.0 and below

[PATCH v1 0/1] soundwire: qcom: disable stop clock on 1.3.0 and below

[Anton Bambura]

This patch fixes audio on sdm845 + wcd9340 with soundwire. The affected
devices are AYN Odin and Lenovo Yoga C630.

If this gets applied it should also be backported to stable and 6.6.x LTS.

Anton Bambura (1):
  soundwire: qcom: disable stop clock on 1.3.0 and below

 drivers/soundwire/qcom.c | 3 +++
 1 file changed, 3 insertions(+)

-- 
2.43.0

[PATCH v1 1/1] soundwire: qcom: disable stop clock on 1.3.0 and below

[Anton Bambura]

This patch returns back the behavior of disabling stop clock on soundwire
1.3.0 and below which seems to have been altered by accident which
results in broken audio on sdm845 + wcd9340. For example, on AYN Odin and
Lenovo Yoga C630 devices.

Fixes: 4830bfa2c812 ("soundwire: qcom: set clk stop need reset flag at runtime")
Signed-off-by: Anton Bambura <jenneron@postmarketos.org>
---
 drivers/soundwire/qcom.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
index a1e2d6c98186..bc03484a28e8 100644
--- a/drivers/soundwire/qcom.c
+++ b/drivers/soundwire/qcom.c
@@ -628,6 +628,9 @@ static int qcom_swrm_enumerate(struct sdw_bus *bus)
 			}
 		}
 
+		if (ctrl->version <= SWRM_VERSION_1_3_0)
+			ctrl->clock_stop_not_supported = true;
+
 		if (!found) {
 			qcom_swrm_set_slave_dev_num(bus, NULL, i);
 			sdw_slave_add(bus, &id, NULL);
-- 
2.43.0

Re: [PATCH v1 1/1] soundwire: qcom: disable stop clock on 1.3.0 and below

[Steev Klimaszewski]

Thank you!

On Sat, Apr 13, 2024 at 1:43 AM Anton Bambura <jenneron@postmarketos.org> wrote:
>
> This patch returns back the behavior of disabling stop clock on soundwire
> 1.3.0 and below which seems to have been altered by accident which
> results in broken audio on sdm845 + wcd9340. For example, on AYN Odin and
> Lenovo Yoga C630 devices.
>
> Fixes: 4830bfa2c812 ("soundwire: qcom: set clk stop need reset flag at runtime")
> Signed-off-by: Anton Bambura <jenneron@postmarketos.org>
> ---
>  drivers/soundwire/qcom.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
> index a1e2d6c98186..bc03484a28e8 100644
> --- a/drivers/soundwire/qcom.c
> +++ b/drivers/soundwire/qcom.c
> @@ -628,6 +628,9 @@ static int qcom_swrm_enumerate(struct sdw_bus *bus)
>                         }
>                 }
>
> +               if (ctrl->version <= SWRM_VERSION_1_3_0)
> +                       ctrl->clock_stop_not_supported = true;
> +
>                 if (!found) {
>                         qcom_swrm_set_slave_dev_num(bus, NULL, i);
>                         sdw_slave_add(bus, &id, NULL);
> --
> 2.43.0
>
>
Tested on the Lenovo Yoga C630
Tested-by: Steev Klimaszewski <steev@kali.org>

-- steev

Re: [PATCH v1 1/1] soundwire: qcom: disable stop clock on 1.3.0 and below

[Srinivas Kandagatla]

On 13/04/2024 07:42, Anton Bambura wrote:
> This patch returns back the behavior of disabling stop clock on soundwire
> 1.3.0 and below which seems to have been altered by accident which
> results in broken audio on sdm845 + wcd9340. For example, on AYN Odin and
> Lenovo Yoga C630 devices.
> 
> Fixes: 4830bfa2c812 ("soundwire: qcom: set clk stop need reset flag at runtime")
> Signed-off-by: Anton Bambura <jenneron@postmarketos.org>
> ---
>   drivers/soundwire/qcom.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/soundwire/qcom.c b/drivers/soundwire/qcom.c
> index a1e2d6c98186..bc03484a28e8 100644
> --- a/drivers/soundwire/qcom.c
> +++ b/drivers/soundwire/qcom.c
> @@ -628,6 +628,9 @@ static int qcom_swrm_enumerate(struct sdw_bus *bus)
>   			}
>   		}
>   
> +		if (ctrl->version <= SWRM_VERSION_1_3_0)
> +			ctrl->clock_stop_not_supported = true;
> +

This is not the right fix, this can be determined from codec 
clk_stop_mode1 flag.

can you try this patch:

----------------------------->cut<-----------------------------
From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Date: Wed, 17 Apr 2024 12:38:49 +0100
Subject: [PATCH] ASoC: codecs: wsa881x: set clk_stop_mode1 flag

WSA881x codecs do not retain the state while clock is stopped, so mark
this with clk_stop_mode1 flag.

Fixes: a0aab9e1404a ("ASoC: codecs: add wsa881x amplifier support")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
---
  sound/soc/codecs/wsa881x.c | 1 +
  1 file changed, 1 insertion(+)

diff --git a/sound/soc/codecs/wsa881x.c b/sound/soc/codecs/wsa881x.c
index 3c025dabaf7a..1253695bebd8 100644
--- a/sound/soc/codecs/wsa881x.c
+++ b/sound/soc/codecs/wsa881x.c
@@ -1155,6 +1155,7 @@ static int wsa881x_probe(struct sdw_slave *pdev,
  	pdev->prop.sink_ports = GENMASK(WSA881X_MAX_SWR_PORTS, 0);
  	pdev->prop.sink_dpn_prop = wsa_sink_dpn_prop;
  	pdev->prop.scp_int1_mask = SDW_SCP_INT1_BUS_CLASH | SDW_SCP_INT1_PARITY;
+	pdev->prop.clk_stop_mode1 = true;
  	gpiod_direction_output(wsa881x->sd_n, !wsa881x->sd_n_val);

  	wsa881x->regmap = devm_regmap_init_sdw(pdev, &wsa881x_regmap_config);
-- 
2.21.0
----------------------------->cut<-----------------------------


thanks,
Srini

>   		if (!found) {
>   			qcom_swrm_set_slave_dev_num(bus, NULL, i);
>   			sdw_slave_add(bus, &id, NULL);

Re: [PATCH v1 1/1] soundwire: qcom: disable stop clock on 1.3.0 and below

[Steev Klimaszewski]

Hi Srini,

On Wed, Apr 17, 2024 at 6:43 AM Srinivas Kandagatla
<srinivas.kandagatla@linaro.org> wrote:
>
> This is not the right fix, this can be determined from codec
> clk_stop_mode1 flag.
>
> can you try this patch:
>
> ----------------------------->cut<-----------------------------
> From: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> Date: Wed, 17 Apr 2024 12:38:49 +0100
> Subject: [PATCH] ASoC: codecs: wsa881x: set clk_stop_mode1 flag
>
> WSA881x codecs do not retain the state while clock is stopped, so mark
> this with clk_stop_mode1 flag.
>
> Fixes: a0aab9e1404a ("ASoC: codecs: add wsa881x amplifier support")
> Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
> ---
>   sound/soc/codecs/wsa881x.c | 1 +
>   1 file changed, 1 insertion(+)
>
> diff --git a/sound/soc/codecs/wsa881x.c b/sound/soc/codecs/wsa881x.c
> index 3c025dabaf7a..1253695bebd8 100644
> --- a/sound/soc/codecs/wsa881x.c
> +++ b/sound/soc/codecs/wsa881x.c
> @@ -1155,6 +1155,7 @@ static int wsa881x_probe(struct sdw_slave *pdev,
>         pdev->prop.sink_ports = GENMASK(WSA881X_MAX_SWR_PORTS, 0);
>         pdev->prop.sink_dpn_prop = wsa_sink_dpn_prop;
>         pdev->prop.scp_int1_mask = SDW_SCP_INT1_BUS_CLASH | SDW_SCP_INT1_PARITY;
> +       pdev->prop.clk_stop_mode1 = true;
>         gpiod_direction_output(wsa881x->sd_n, !wsa881x->sd_n_val);
>
>         wsa881x->regmap = devm_regmap_init_sdw(pdev, &wsa881x_regmap_config);
> --
> 2.21.0
> ----------------------------->cut<-----------------------------
>
>
> thanks,
> Srini
>
> >               if (!found) {
> >                       qcom_swrm_set_slave_dev_num(bus, NULL, i);
> >                       sdw_slave_add(bus, &id, NULL);
>

I tested it here on my c630 and can confirm that your patch does fix
the audio as well.