linux-kernel - Re: [PATCH] ASoC: pcm512x: Add support for data formats RJ and LJ

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 11 Nov 2020 08:54:55 +0100
From:   Kirill Marinushkin <kmarinushkin@...dec.com>
To:     Peter Ujfalusi <peter.ujfalusi@...com>
Cc:     Mark Brown <broonie@...nel.org>, Takashi Iwai <tiwai@...e.com>,
        Liam Girdwood <lgirdwood@...il.com>,
        Matthias Reichl <hias@...us.com>,
        Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>,
        alsa-devel@...a-project.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ASoC: pcm512x: Add support for data formats RJ and LJ

Hello Peter,

than you for your review!

> The bus format and
>
>>  	switch (pcm512x->fmt & SND_SOC_DAIFMT_MASTER_MASK) {
>
>>  	case SND_SOC_DAIFMT_CBS_CFS:
>>  		ret = regmap_update_bits(pcm512x->regmap,
>
> the clock generation role should be set in pcm512x_set_fmt(), in that
> way you can deny specific setups earlier.

I think we could move both checks for`SND_SOC_DAIFMT_FORMAT_MASK` and
`SND_SOC_DAIFMT_MASTER_MASK` into `pcm512x_set_fmt()`. But it would be a
different scope, and I didn't intend to do that level of refactoring.
Looking at other codecs in kernel, I would say, that doing those checks in
`pcm512x_hw_params()`, as they are done currently, is an equally valid approach.

As technically keeping checs where they are now doesn't break anything, and is
aligned with ASoC codecs design, I suggest to keep the checks where they are.
Would you agree?

> I would also add DSP_A and DSP_B modes at the same time, DSP_A would
> need a write of 1 to register 41 (PCM512x_I2S_2, offset = 1), other
> formats should set the offset to 0.

That's a good idea, than you for technical details! I just didn't know how to
use DSP_A and DSP_B. I will add them, and submit as patch v2

Best regards,
Kirill

On 11/10/2020 07:59 AM, Peter Ujfalusi wrote:
> 
> 
> On 09/11/2020 23.21, Kirill Marinushkin wrote:
>> Currently, pcm512x driver supports only I2S data format.
>> This commit adds RJ and LJ as well.
>>
>> I don't expect regression WRT existing sound cards, because:
>>
>> * default value in corresponding register of pcm512x codec is 0 ==  I2S
>> * existing in-tree sound cards with pcm512x codec are configured for I2S
>> * i don't see how existing off-tree sound cards with pcm512x codec could be
>>   configured differently - it would not work
>> * tested explicitly, that there is no regression with Raspberry Pi +
>>   sound card `sound/soc/bcm/hifiberry_dacplus.c`
>>
>> Signed-off-by: Kirill Marinushkin <kmarinushkin@...dec.com>
>> Cc: Mark Brown <broonie@...nel.org>
>> Cc: Takashi Iwai <tiwai@...e.com>
>> Cc: Liam Girdwood <lgirdwood@...il.com>
>> Cc: Matthias Reichl <hias@...us.com>
>> Cc: Kuninori Morimoto <kuninori.morimoto.gx@...esas.com>
>> Cc: Peter Ujfalusi <peter.ujfalusi@...com>
>> Cc: alsa-devel@...a-project.org
>> Cc: linux-kernel@...r.kernel.org
>> ---
>>  sound/soc/codecs/pcm512x.c | 24 ++++++++++++++++++++++++
>>  1 file changed, 24 insertions(+)
>>
>> diff --git a/sound/soc/codecs/pcm512x.c b/sound/soc/codecs/pcm512x.c
>> index 8153d3d01654..c6e975fb4a43 100644
>> --- a/sound/soc/codecs/pcm512x.c
>> +++ b/sound/soc/codecs/pcm512x.c
>> @@ -1167,6 +1167,7 @@ static int pcm512x_hw_params(struct snd_pcm_substream *substream,
>>  	struct snd_soc_component *component = dai->component;
>>  	struct pcm512x_priv *pcm512x = snd_soc_component_get_drvdata(component);
>>  	int alen;
>> +	int afmt;
>>  	int gpio;
>>  	int clock_output;
>>  	int master_mode;
>> @@ -1195,6 +1196,22 @@ static int pcm512x_hw_params(struct snd_pcm_substream *substream,
>>  		return -EINVAL;
>>  	}
>>  
>> +	switch (pcm512x->fmt & SND_SOC_DAIFMT_FORMAT_MASK) {
>> +	case SND_SOC_DAIFMT_I2S:
>> +		afmt = PCM512x_AFMT_I2S;
>> +		break;
>> +	case SND_SOC_DAIFMT_RIGHT_J:
>> +		afmt = PCM512x_AFMT_RTJ;
>> +		break;
>> +	case SND_SOC_DAIFMT_LEFT_J:
>> +		afmt = PCM512x_AFMT_LTJ;
>> +		break;
>> +	default:
>> +		dev_err(component->dev, "unsupported DAI format: 0x%x\n",
>> +			pcm512x->fmt);
>> +		return -EINVAL;
>> +	}
>> +
> 
> The bus format and
> 
>>  	switch (pcm512x->fmt & SND_SOC_DAIFMT_MASTER_MASK) {
> 
>>  	case SND_SOC_DAIFMT_CBS_CFS:
>>  		ret = regmap_update_bits(pcm512x->regmap,
> 
> the clock generation role should be set in pcm512x_set_fmt(), in that
> way you can deny specific setups earlier.
> 
> I would also add DSP_A and DSP_B modes at the same time, DSP_A would
> need a write of 1 to register 41 (PCM512x_I2S_2, offset = 1), other
> formats should set the offset to 0.
> 
>> @@ -1236,6 +1253,13 @@ static int pcm512x_hw_params(struct snd_pcm_substream *substream,
>>  		return ret;
>>  	}
>>  
>> +	ret = regmap_update_bits(pcm512x->regmap, PCM512x_I2S_1,
>> +				 PCM512x_AFMT, afmt);
>> +	if (ret != 0) {
>> +		dev_err(component->dev, "Failed to set data format: %d\n", ret);
>> +		return ret;
>> +	}
>> +
>>  	if (pcm512x->pll_out) {
>>  		ret = regmap_write(pcm512x->regmap, PCM512x_FLEX_A, 0x11);
>>  		if (ret != 0) {
>>
> 
> - Péter
> 
> Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
> Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
>