Hi Peter,

On Wednesday, 6 April 2022 21:57:34 CEST Péter Ujfalusi wrote:

Hi Janusz,

On 02/04/2022 15:01, Janusz Krzysztofik wrote:

...

The driver, OMAP1 specific, now omits clk_prepare/unprepare() steps, not supported by OMAP1 custom implementation of clock API. However, non-CCF stubs of those functions exist for use on such platforms until converted to CCF.

Update the driver to be compatible with CCF implementation of clock API.

Signed-off-by: Janusz Krzysztofik jmkrzyszt@gmail.com

sound/soc/ti/osk5912.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/sound/soc/ti/osk5912.c b/sound/soc/ti/osk5912.c index 40e29dda7e7a..22da3b335e81 100644 --- a/sound/soc/ti/osk5912.c +++ b/sound/soc/ti/osk5912.c @@ -134,6 +134,10 @@ static int __init osk_soc_init(void) goto err2; }

• err = clk_prepare(tlv320aic23_mclk);

would not make sense to change the clk_enable() to clk_prepare_enable() in osk_startup() and the clk_disable() to clk_disable_unprepare() in osk_shutdown() instead leaving the clock in prepared state as long as the driver is loaded?

OK, I can see the clk_prepare_enable() approach is more common than separate clk_prepare() across the kernel code, and I have no arguments against it in our case. I'll submit v2 soon.

Thanks, Janusz

...

• if (err)

• goto err3;
  

• /*
  • Configure 12 MHz output on MCLK.
  */

@@ -142,7 +146,7 @@ static int __init osk_soc_init(void) if (clk_set_rate(tlv320aic23_mclk, CODEC_CLOCK)) { printk(KERN_ERR "Cannot set MCLK for AIC23 CODEC\n"); err = -ECANCELED;

• 	goto err3;
  

• 	goto err4;
  

  } }

@@ -151,6 +155,8 @@ static int __init osk_soc_init(void)

return 0;

+err4:

• clk_unprepare(tlv320aic23_mclk);

err3: clk_put(tlv320aic23_mclk); err2: @@ -164,6 +170,7 @@ static int __init osk_soc_init(void)

static void __exit osk_soc_exit(void) {

• clk_unprepare(tlv320aic23_mclk); clk_put(tlv320aic23_mclk); platform_device_unregister(osk_snd_device);

}

On 07/04/2022 22:12, Janusz Krzysztofik wrote:

The driver, OMAP1 specific, now omits clk_prepare/unprepare() steps, not supported by OMAP1 custom implementation of clock API. However, non-CCF stubs of those functions exist for use on such platforms until converted to CCF.

Update the driver to be compatible with CCF implementation of clock API.

v2: use clk_prepare_enable/clk_disable_unprepare() (Peter)

Does the changelogn needs to be in the commit message?

I almost missed the V2 as it came as a reply to a thread ;)

Other than the nitpick with the changlog placement:

Looks good but I can not test it, Acked-by: Peter Ujfalusi peter.ujfalusi@gmail.com

Signed-off-by: Janusz Krzysztofik jmkrzyszt@gmail.com

sound/soc/ti/osk5912.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sound/soc/ti/osk5912.c b/sound/soc/ti/osk5912.c index 40e29dda7e7a..2790c8915f55 100644 --- a/sound/soc/ti/osk5912.c +++ b/sound/soc/ti/osk5912.c @@ -27,12 +27,12 @@ static struct clk *tlv320aic23_mclk;

static int osk_startup(struct snd_pcm_substream *substream) {

• return clk_enable(tlv320aic23_mclk);

• return clk_prepare_enable(tlv320aic23_mclk);

}

static void osk_shutdown(struct snd_pcm_substream *substream) {

• clk_disable(tlv320aic23_mclk);

• clk_disable_unprepare(tlv320aic23_mclk);

}

static int osk_hw_params(struct snd_pcm_substream *substream,