Re: [PATCH V3] ASoC: fsl_ssi: refine ipg clock usage in this module

[Markus Pargmann <mpa@pengutronix.de>]

[-- Attachment #1: Type: text/plain, Size: 6863 bytes --]

On Mon, Sep 15, 2014 at 06:22:27PM +0800, Shengjiu Wang wrote:
> On Mon, Sep 15, 2014 at 12:05:41PM +0200, Markus Pargmann wrote:
> > On Fri, Sep 12, 2014 at 06:35:15PM +0800, Shengjiu Wang wrote:
> > > Check if ipg clock is in clock-names property, then we can move the
> > > ipg clock enable and disable operation to startup and shutdown, that
> > > is only enable ipg clock when ssi is working and keep clock is disabled
> > > when ssi is in idle.
> > > But when the checking is failed, remain the clock control as before.
> > > 
> > > Signed-off-by: Shengjiu Wang <shengjiu.wang@freescale.com>
> > > ---
> > > V3 change log:
> > > update patch according Nicolin and markus's comments
> > > 
> > > 
> > >  sound/soc/fsl/fsl_ssi.c |   53 ++++++++++++++++++++++++++++++++++++++++-------
> > >  1 file changed, 45 insertions(+), 8 deletions(-)
> > > 
> > > diff --git a/sound/soc/fsl/fsl_ssi.c b/sound/soc/fsl/fsl_ssi.c
> > > index 2fc3e66..6d1dfd5 100644
> > > --- a/sound/soc/fsl/fsl_ssi.c
> > > +++ b/sound/soc/fsl/fsl_ssi.c
> > > @@ -169,6 +169,7 @@ struct fsl_ssi_private {
> > >  	u8 i2s_mode;
> > >  	bool use_dma;
> > >  	bool use_dual_fifo;
> > > +	bool has_ipg_clk_name;
> > >  	unsigned int fifo_depth;
> > >  	struct fsl_ssi_rxtx_reg_val rxtx_reg_val;
> > >  
> > > @@ -530,6 +531,11 @@ static int fsl_ssi_startup(struct snd_pcm_substream *substream,
> > >  	struct snd_soc_pcm_runtime *rtd = substream->private_data;
> > >  	struct fsl_ssi_private *ssi_private =
> > >  		snd_soc_dai_get_drvdata(rtd->cpu_dai);
> > > +	int ret;
> > > +
> > > +	ret = clk_prepare_enable(ssi_private->clk);
> > > +	if (ret)
> > > +		return ret;
> > >  
> > >  	/* When using dual fifo mode, it is safer to ensure an even period
> > >  	 * size. If appearing to an odd number while DMA always starts its
> > > @@ -544,6 +550,21 @@ static int fsl_ssi_startup(struct snd_pcm_substream *substream,
> > >  }
> > >  
> > >  /**
> > > + * fsl_ssi_shutdown: shutdown the SSI
> > > + *
> > > + */
> > > +static void fsl_ssi_shutdown(struct snd_pcm_substream *substream,
> > > +                             struct snd_soc_dai *dai)
> > > +{
> > > +	struct snd_soc_pcm_runtime *rtd = substream->private_data;
> > > +	struct fsl_ssi_private *ssi_private =
> > > +		snd_soc_dai_get_drvdata(rtd->cpu_dai);
> > > +
> > > +	clk_disable_unprepare(ssi_private->clk);
> > > +
> > > +}
> > > +
> > > +/**
> > >   * fsl_ssi_set_bclk - configure Digital Audio Interface bit clock
> > >   *
> > >   * Note: This function can be only called when using SSI as DAI master
> > > @@ -1043,6 +1064,7 @@ static int fsl_ssi_dai_probe(struct snd_soc_dai *dai)
> > >  
> > >  static const struct snd_soc_dai_ops fsl_ssi_dai_ops = {
> > >  	.startup	= fsl_ssi_startup,
> > > +	.shutdown       = fsl_ssi_shutdown,
> > >  	.hw_params	= fsl_ssi_hw_params,
> > >  	.hw_free	= fsl_ssi_hw_free,
> > >  	.set_fmt	= fsl_ssi_set_dai_fmt,
> > > @@ -1168,17 +1190,22 @@ static int fsl_ssi_imx_probe(struct platform_device *pdev,
> > >  	u32 dmas[4];
> > >  	int ret;
> > >  
> > > -	ssi_private->clk = devm_clk_get(&pdev->dev, NULL);
> > > +	if (ssi_private->has_ipg_clk_name)
> > > +		ssi_private->clk = devm_clk_get(&pdev->dev, "ipg");
> > > +	else
> > > +		ssi_private->clk = devm_clk_get(&pdev->dev, NULL);
> > >  	if (IS_ERR(ssi_private->clk)) {
> > >  		ret = PTR_ERR(ssi_private->clk);
> > >  		dev_err(&pdev->dev, "could not get clock: %d\n", ret);
> > >  		return ret;
> > >  	}
> > >  
> > > -	ret = clk_prepare_enable(ssi_private->clk);
> > > -	if (ret) {
> > > -		dev_err(&pdev->dev, "clk_prepare_enable failed: %d\n", ret);
> > > -		return ret;
> > > +	if (!ssi_private->has_ipg_clk_name) {
> > > +		ret = clk_prepare_enable(ssi_private->clk);
> > > +		if (ret) {
> > > +			dev_err(&pdev->dev, "clk_prepare_enable failed: %d\n", ret);
> > > +			return ret;
> > > +		}
> > >  	}
> > >  
> > >  	/* For those SLAVE implementations, we ingore non-baudclk cases
> > > @@ -1236,8 +1263,9 @@ static int fsl_ssi_imx_probe(struct platform_device *pdev,
> > >  	return 0;
> > >  
> > >  error_pcm:
> > > -	clk_disable_unprepare(ssi_private->clk);
> > >  
> > > +	if (!ssi_private->has_ipg_clk_name)
> > > +		clk_disable_unprepare(ssi_private->clk);
> > >  	return ret;
> > >  }
> > >  
> > > @@ -1246,7 +1274,8 @@ static void fsl_ssi_imx_clean(struct platform_device *pdev,
> > >  {
> > >  	if (!ssi_private->use_dma)
> > >  		imx_pcm_fiq_exit(pdev);
> > > -	clk_disable_unprepare(ssi_private->clk);
> > > +	if (!ssi_private->has_ipg_clk_name)
> > > +		clk_disable_unprepare(ssi_private->clk);
> > >  }
> > >  
> > >  static int fsl_ssi_probe(struct platform_device *pdev)
> > > @@ -1321,8 +1350,16 @@ static int fsl_ssi_probe(struct platform_device *pdev)
> > >  		return -ENOMEM;
> > >  	}
> > >  
> > > -	ssi_private->regs = devm_regmap_init_mmio(&pdev->dev, iomem,
> > > +	ret = of_property_match_string(np, "clock-names", "ipg");
> > > +	if (ret < 0) {
> > > +		ssi_private->has_ipg_clk_name = false;
> > > +		ssi_private->regs = devm_regmap_init_mmio(&pdev->dev, iomem,
> > >  			&fsl_ssi_regconfig);
> > 
> > Sorry if I was unclear about that. My suggestion was to enable the clock
> > right here:
> > 		clk_prepare_enable(ssi_private->clk);
> > 
> > Then you can remove ssi_private->has_ipg_clk_name and all
> > clk_prepare_enable() and clk_disable_unprepare() from above. Also you
> > can move the devm_clk_get() into this block.
> >
> 
> ipg clock not only need to be enabled when accessing register, but also 
> need to be enabled when ssi is working. So I add clk_enable in startup.
>  
> > It seems you really want to implement this for devicetrees where the
> > "ipg" clock-name is missing, but I don't understand why? I really can't
> > see any benefit of adding all these clk_prepare_enable() calls for all
> > cornercases that may occure. For example the clocks for AC97 are still
> > missing in this version.
> 
> When ipg clock-name is missing, I just remain the code logic as before.
> So AC97 case is same as before too. 
> When have ipg clock-name, register it to regmap for accessing register,
> and enabling ipg clock in startup also is useful for AC97.
> 
> I don't understand why you think clock for AC97 is still missing?

Sorry got it all wrong, monday morning tiredness perhaps ;-).

The patch looks good for me at the second look. I will test it later
today and give you feedback.

Best regards,

Markus

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]