Re: [PATCH 1/3] ALSA: pcm: add IEC958 channel status update

[Moise Gergaud <moise.gergaud@st.com>]

On 12/02/2015 08:00 PM, Russell King - ARM Linux wrote:
> On Wed, Dec 02, 2015 at 03:22:04PM +0100, Moise Gergaud wrote:
>> Add a helper to update only the IEC958 channel status sampling freq and
>> word length parameters from an ALSA snd_pcm_runtime structure, taking
>> account of the sample rate and sample size.
>>
>> Signed-off-by: Moise Gergaud <moise.gergaud@st.com>
>> Acked-by: Arnaud Pouliquen <arnaud.pouliquen@st.com>
>
> NAK.
>
>> diff --git a/include/sound/pcm_iec958.h b/include/sound/pcm_iec958.h
>> index 0eed397..0c84c69 100644
>> --- a/include/sound/pcm_iec958.h
>> +++ b/include/sound/pcm_iec958.h
>> @@ -3,7 +3,24 @@
>>
>>   #include <linux/types.h>
>>
>> +#ifdef CONFIG_SND_PCM_IEC958
>>   int snd_pcm_create_iec958_consumer(struct snd_pcm_runtime *runtime, u8 *cs,
>> -	size_t len);
>> +				   size_t len);
>> +
>> +int snd_pcm_update_iec958_consumer(struct snd_pcm_runtime *runtime, u8 *cs,
>> +				   size_t len);
>> +#else
>> +int snd_pcm_create_iec958_consumer(struct snd_pcm_runtime *runtime, u8 *cs,
>> +				   size_t len)
>> +{
>> +	return len;
>> +}
>> +
>> +int snd_pcm_update_iec958_consumer(struct snd_pcm_runtime *runtime, u8 *cs,
>> +				   size_t len)
>> +{
>> +	return len;
>> +}
>
> Firstly, this is dangerous.  If you don't have anything defined here,
> then the channel status ends up being undefined.
>
> The intention here is that if you're a user of this, then select the
> config symbol.  That makes providing stubs totally unnecessary and
> removes this problem.

agree

>
>> +#endif
>>
>>   #endif
>> diff --git a/sound/core/pcm_iec958.c b/sound/core/pcm_iec958.c
>> index 36b2d7a..abe3967 100644
>> --- a/sound/core/pcm_iec958.c
>> +++ b/sound/core/pcm_iec958.c
>> @@ -11,67 +11,72 @@
>>   #include <sound/pcm.h>
>>   #include <sound/pcm_iec958.h>
>>
>> -/**
>> - * snd_pcm_create_iec958_consumer - create consumer format IEC958 channel status
>> +/*
>> + * snd_pcm_update_iec958_consumer - update consumer format IEC958 channel status
>>    * @runtime: pcm runtime structure with ->rate filled in
>>    * @cs: channel status buffer, at least four bytes
>>    * @len: length of channel status buffer
>>    *
>> - * Create the consumer format channel status data in @cs of maximum size
>> - * @len corresponding to the parameters of the PCM runtime @runtime.
>> - *
>> - * Drivers may wish to tweak the contents of the buffer after creation.
>> + * Update sampling frequency and word length parameters of the consumer format
>> + * channel status data in @cs of maximum size @len.
>> + * Values correspond to the rate and format parameters of the PCM runtime
>> + * @runtime.
>>    *
>>    * Returns: length of buffer, or negative error code if something failed.
>>    */
>> -int snd_pcm_create_iec958_consumer(struct snd_pcm_runtime *runtime, u8 *cs,
>> -	size_t len)
>> +int snd_pcm_update_iec958_consumer(struct snd_pcm_runtime *runtime, u8 *cs,
>> +				   size_t len)
>>   {
>> -	unsigned int fs, ws;
>> -
>>   	if (len < 4)
>>   		return -EINVAL;
>>
>> +	cs[3] &= ~IEC958_AES3_CON_FS;
>> +
>>   	switch (runtime->rate) {
>> +	case 22050:
>> +		cs[3] |= IEC958_AES3_CON_FS_22050;
>> +		break;
>>   	case 32000:
>> -		fs = IEC958_AES3_CON_FS_32000;
>> +		cs[3] |= IEC958_AES3_CON_FS_32000;
>>   		break;
>>   	case 44100:
>> -		fs = IEC958_AES3_CON_FS_44100;
>> +		cs[3] |= IEC958_AES3_CON_FS_44100;
>>   		break;
>>   	case 48000:
>> -		fs = IEC958_AES3_CON_FS_48000;
>> +		cs[3] |= IEC958_AES3_CON_FS_48000;
>>   		break;
>>   	case 88200:
>> -		fs = IEC958_AES3_CON_FS_88200;
>> +		cs[3] |= IEC958_AES3_CON_FS_88200;
>>   		break;
>>   	case 96000:
>> -		fs = IEC958_AES3_CON_FS_96000;
>> +		cs[3] |= IEC958_AES3_CON_FS_96000;
>>   		break;
>>   	case 176400:
>> -		fs = IEC958_AES3_CON_FS_176400;
>> +		cs[3] |= IEC958_AES3_CON_FS_176400;
>>   		break;
>>   	case 192000:
>> -		fs = IEC958_AES3_CON_FS_192000;
>> +		cs[3] |= IEC958_AES3_CON_FS_192000;
>>   		break;
>
> Again, I'm not happy with this change.  The intention here is that we
> validate all arguments before we change anything.  This breaks that
> principle.
>
We propose this change for compressed mode (IEC61937/non linear PCM).
In case of compressed mode, iec958 channel status byte 0 bit1 = 1, then 
we cannot use snd_pcm_create_iec958_consumer.
So we propose a second function that only updates the sampling freq and 
word length. Other parameters are unchanged (set by user).

In a previous patch proposal, we managed channel status sampling freq 
directly in our driver. Mark suggested us to manage it in a generic 
function (could be useful for other drivers).