Re: [linux-sunxi][alsa-devel][PATCH 3/3] ASOC: sunxi: Add support for the spdif block
From: Code Kipper
Date: Thu Sep 24 2015 - 14:00:41 EST
On 24 September 2015 at 19:29, Mark Brown <broonie@xxxxxxxxxx> wrote:
> On Thu, Sep 24, 2015 at 04:30:05PM +0200, codekipper@xxxxxxxxx wrote:
>> From: Marcus Cooper <codekipper@xxxxxxxxx>
>>
>> The sun4i, sun6i and sun7i SoC families have an SPDIF
>> block which is capable of playback and capture.
>
> I'm not seeing patches 1 or 2 - what's the story here, are there
> dependencies? Please use subject lines matching the style for the
> subsystem and also don't fill your subject lines with noisy tags beyond
> "[PATCH n/x]", when I look at this in my mail client what I see is:
For some reason git-send-email wouldn't send the last patch last night
so I pushed it today using another machine. I was thinking last night
that the tagging was a bit extreme. I won't do it again.
>
> -> 432 C 09/24 codekipper@gmai ( 27K) [linux-sunxi][alsa-devel][PATCH 3/3] AS
>
>> sound/soc/sunxi/Kconfig | 10 +
>> sound/soc/sunxi/Makefile | 4 +
>> sound/soc/sunxi/sunxi-machine-spdif.c | 110 +++++
>> sound/soc/sunxi/sunxi-spdif.c | 801 ++++++++++++++++++++++++++++++++++
>
> The machine driver and controller driver should be submitted as separate
> patches for ease of review. Is there a strong reason for not using
> simple-card?
Yeah..I'm looking for some spiritual guidance here...I'll separate
this out and look at alternate methods.
>
>> +void sunxi_snd_txctrl(struct snd_pcm_substream *substream,
>> + struct sunxi_spdif_dev *host, int on)
>> +{
>> + u32 tmp;
>
> There's no meaningful sharing between the enable and disable paths and
> only one place either is called, it's better to just inline this into
> the callers.
>
>> + if (!cpu_dai->active) {
>> + ret = clk_prepare_enable(host->clk);
>> + if (ret)
>> + return ret;
>> + }
>
> Can you move the clock enables to runtime PM and let the core do runtime
> PM for you?
>
>> +static int sunxi_spdif_set_clkdiv(struct snd_soc_dai *cpu_dai,
>> + unsigned int rate, int div)
>> +{
>> + struct sunxi_spdif_dev *host = snd_soc_dai_get_drvdata(cpu_dai);
>> + int sample_freq, original_sample_freq;
>
> Why are you implementing a set_clkdiv() operation - is the driver not
> capable of working out its internal clocking automaticallly?
>
>> +static int sunxi_spdif_hw_params(struct snd_pcm_substream *substream,
>> + struct snd_pcm_hw_params *params,
>> + struct snd_soc_dai *cpu_dai)
>> +{
>
>> + ret = snd_soc_dai_set_fmt(cpu_dai, fmt);
>> + if (ret < 0)
>> + return ret;
>
> This looks very broken - what is this doing and why?
>
>> +static struct snd_soc_dai_driver sunxi_spdif_dai = {
>> + .playback = {
>> + .channels_min = 2,
>> + .channels_max = 2,
>> + .rates = SUNXI_RATES,
>
> There was code in the driver to handle mono signals but this says only
> stereo is supported?
>
>> + if (clk_prepare_enable(host->apb_clk)) {
>> + dev_err(&pdev->dev, "try to enable apb_spdif_clk failed\n");
>> + return -EINVAL;
>> + }
>
> Don't ignore the error code you got from the API, print it and pass it
> back.
All points noted and I'll work to clear them...thanks for you time in
reviewing this.
CK
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/