[PATCH v2 2/9] ASoC: q6asm: make commands specific to streams

Srinivas Kandagatla srinivas.kandagatla at linaro.org
Wed Jul 22 10:59:47 CEST 2020


Thanks Pierre for review,

On 21/07/2020 20:31, Pierre-Louis Bossart wrote:
> 
> 
> 
>> diff --git a/sound/soc/qcom/qdsp6/q6asm-dai.c 
>> b/sound/soc/qcom/qdsp6/q6asm-dai.c
>> index 941f3216399c..fb0488e7beb9 100644
>> --- a/sound/soc/qcom/qdsp6/q6asm-dai.c
>> +++ b/sound/soc/qcom/qdsp6/q6asm-dai.c
>> @@ -67,6 +67,8 @@ struct q6asm_dai_rtd {
>>       uint16_t bits_per_sample;
>>       uint16_t source; /* Encoding source bit mask */
>>       struct audio_client *audio_client;
>> +    /* Active */
> 
> nit-pick: what does this 'Active' comment try to say? the stream_id 
> seems to be used for RUN/EOS/CLOSE operations.

Active mean its the active stream id which is consuming the data at the 
point in time. As we toggle stream ids between 1 and 2. This active 
stream_id is used for every command sent to dsp.

> 
>> +    uint32_t stream_id;
>>       uint16_t session_id;
>>       enum stream_state state;
>>   };
>> @@ -184,8 +186,8 @@ static void event_handler(uint32_t opcode, 
>> uint32_t token,
>>       switch (opcode) {
>>       case ASM_CLIENT_EVENT_CMD_RUN_DONE:
>>           if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK)
>> -            q6asm_write_async(prtd->audio_client,
>> -                   prtd->pcm_count, 0, 0, NO_TIMESTAMP);
>> +            q6asm_write_async(prtd->audio_client, prtd->stream_id,
>> +                   prtd->pcm_count, 0, 0, 0);
> 
> In the V1 review we discussed this
> 

Sorry, I missed that! will address this in next version!

> "
>  > sound/soc/qcom/qdsp6/q6asm.h:#define NO_TIMESTAMP    0xFF00
>  >
>  > is the change on the previous line intentional?
> 
> May be not!
> 
> Plan is that the users of these apis will send flags directly instead of 
> boiler plating this!
> 
> This change should go as part of next patch("[PATCH 04/11] ASoC: q6asm: 
> use flags directly from asm-dai") which would make it much clear!
> "
> 
> doesn't look like there was a change here?
> 
> 


More information about the Alsa-devel mailing list