Hi Troy, On Tue, Nov 17, 2009 at 05:22:34, Troy Kisky wrote:
Fix underruns by using dma to copy 1st to sram in a ping/pong buffer style and then copying from the sram to the ASP. This also has the advantage of tolerating very long interrupt latency on dma completion.
Signed-off-by: Troy Kisky <troy.kisky@boundarydevices.com>
[...]
+static int request_ping_pong(struct snd_pcm_substream *substream, + struct davinci_runtime_data *prtd, + struct snd_dma_buffer *iram_dma) +{ + dma_addr_t asp_src_ping; + dma_addr_t asp_dst_ping; + int link; + struct davinci_pcm_dma_params *dma_data = prtd->params; + + /* Request ram master channel */ + link = prtd->ram_channel = edma_alloc_channel(EDMA_CHANNEL_ANY, + davinci_pcm_dma_irq, substream, + EVENTQ_1);
What is the reason for choosing EVENTQ_1 for this channel? EVENTQ_0 is already being used for ASP channel. I imagine it will be much easier to tune the queue usage in the system if all of audio data was using the same queue. I am working on a patch which lets platform specify the event queues for audio DMA. I am not sure if I really need to make a provision for two different queues to be specified - that's why I ask. Thanks, Sekhar