Re: [PATCH v8 19/22] ASoC: tegra: Enable audio mclk during tegra_asoc_utils_init

20.01.2020 07:10, Sameer Pujar пишет:

On 1/19/2020 8:44 PM, Dmitry Osipenko wrote:

quoted

External email: Use caution opening links or attachments


14.01.2020 10:24, Sowjanya Komatineni пишет:

quoted

Tegra PMC clock clk_out_1 is dedicated for audio mclk from Tegra30
through Tegra210 and currently Tegra clock driver keeps the audio
mclk enabled.

With the move of PMC clocks from clock driver into pmc driver,
audio mclk enable from clock driver is removed and this should be
taken care by the audio driver.

tegra_asoc_utils_init calls tegra_asoc_utils_set_rate and audio mclk
rate configuration is not needed during init and set_rate is actually
done during hw_params callback.

So, this patch removes tegra_asoc_utils_set_rate call and just leaves
the audio mclk enabled.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
---
  sound/soc/tegra/tegra_asoc_utils.c | 11 +++++++++--
  1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/sound/soc/tegra/tegra_asoc_utils.c

b/sound/soc/tegra/tegra_asoc_utils.c
index 1dce5ad6e665..99584970f5f4 100644

--- a/sound/soc/tegra/tegra_asoc_utils.c
+++ b/sound/soc/tegra/tegra_asoc_utils.c

@@ -216,9 +216,16 @@ int tegra_asoc_utils_init(struct

tegra_asoc_utils_data *data,
               data->clk_cdev1 = clk_out_1;
       }

-     ret = tegra_asoc_utils_set_rate(data, 44100, 256 * 44100);
-     if (ret)
+     /*
+      * FIXME: There is some unknown dependency between audio mclk
disable
+      * and suspend-resume functionality on Tegra30, although audio
mclk is
+      * only needed for audio.
+      */
+     ret = clk_prepare_enable(data->clk_cdev1);
+     if (ret) {
+             dev_err(data->dev, "Can't enable cdev1: %d\n", ret);
               return ret;
+     }

       return 0;
  }

Shouldn't the clock be disabled on driver's removal?

I am not sure if we really need to do in this series as it does not
change the behavior from what was there earlier. Also there is already a
FIXME item here and we end up adding clock disable in remove() path of
multiple drivers, which is going to be removed once we address FIXME.

Well, perhaps this is indeed good enough for the time being.

BTW, I didn't spot any suspend-resume problems using v8.

Tested-by: Dmitry Osipenko <digetx@gmail.com>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>