On Mon, Jan 20, 2025 at 03:31:19PM +0530, Vijendar Mukunda wrote:

> +static irqreturn_t acp70_irq_thread(int irq, void *context)
> +{
> +	struct sdw_dma_dev_data *sdw_dma_data;
> +	struct acp70_dev_data *adata = context;
> +	u32 stream_index;
> +
> +	sdw_dma_data = dev_get_drvdata(&adata->sdw_dma_dev->dev);
> +
> +	for (stream_index = 0; stream_index < ACP70_SDW0_DMA_MAX_STREAMS; stream_index++) {
> +		if (adata->sdw0_dma_intr_stat[stream_index]) {
> +			if (sdw_dma_data->sdw0_dma_stream[stream_index])
> +				snd_pcm_period_elapsed(sdw_dma_data->sdw0_dma_stream[stream_index]);
> +			adata->sdw0_dma_intr_stat[stream_index] = 0;
> +		}
> +	}
> +	for (stream_index = 0; stream_index < ACP70_SDW1_DMA_MAX_STREAMS; stream_index++) {
> +		if (adata->sdw1_dma_intr_stat[stream_index]) {
> +			if (sdw_dma_data->sdw1_dma_stream[stream_index])
> +				snd_pcm_period_elapsed(sdw_dma_data->sdw1_dma_stream[stream_index]);
> +			adata->sdw1_dma_intr_stat[stream_index] = 0;
> +		}
> +	}
> +	return IRQ_HANDLED;
> +}

I appreciate that this pattern is already (identically...) in the ps
driver but I'm not seeing anything here that looks like it can't run in
interrupt context - is there a special reason for deferring to thread
context?

> +static irqreturn_t acp70_irq_handler(int irq, void *dev_id)
> +{

This really does seem *very* similar to the acp63 code...
[unhandled content-type:application/pgp-signature]