Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1565250 > unrolled thread

Re: [PATCH v2] ASoC: mxs-saif: add mclk enable/disable ops

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v2] ASoC: mxs-saif: add mclk enable/disable ops Mark Brown <broonie@kernel.org> - 2017-01-23 20:10 +0100

#1565250 — Re: [PATCH v2] ASoC: mxs-saif: add mclk enable/disable ops

[Multipart message — attachments visible in raw view] — view raw

On Thu, Dec 22, 2016 at 04:49:26PM +0000, Mans Rullgard wrote:

>  int mxs_saif_put_mclk(unsigned int saif_id)
>  {
> -	struct mxs_saif *saif = mxs_saif[saif_id];
> -	u32 stat;
> +	struct clk *clk;
>  
> -	if (!saif)
> -		return -EINVAL;
> +	clk = clk_get(NULL, "mxs_saif_mclk");
> +	if (IS_ERR(clk))
> +		return PTR_ERR(clk);

So, this *is* an in place refactoring but it's only half done in that
we don't have any followup patches that move the clk_get() to device
probe where it should be.

> +static void mxs_saif_mclk_disable(struct clk_hw *hw)
> +{
> +	struct mxs_saif *saif = to_mxs_saif(hw);
> +
> +	if (!saif->ongoing)
> +		__raw_writel(BM_SAIF_CTRL_RUN,
> +			     saif->base + SAIF_CTRL + MXS_CLR_ADDR);
> +
> +	saif->mclk_in_use = 0;
> +}

We silently ignore disables if the clock is in use?  That seems error
prone.  I'd expect us to at least warn in that case.

> +static unsigned long mxs_saif_mclk_recalc_rate(struct clk_hw *hw,
> +					       unsigned long parent_rate)
> +{
> +	return clk_divider_ops.recalc_rate(hw, parent_rate);
> +}

Can't we just assign these ops directly?  Having to write wrapper
functions like this looks like we're doing something wrong here.

[toc] | [standalone]

Back to top | Article view | linux.kernel

csiph-web