hda: set correct link dma channel #505
Conversation
host dma and link dma work in decouple mode for SOF + HDA codec. Now allocate host dma and link dma channel in host and set them in FW individually. Signed-off-by: Rander Wang <[email protected]>
|
for this issue:#433 |
|
host change:thesofproject/linux#210 |
|
@RanderWang can you clarify why this change is needed, I know it's to fix a bug but why must the host and link DMA channels be set by host drivers ? Should the host and link channels use the same ID ? Please explain why host must select them. |
ID is a confusing word. dma channel should be better. host channel is independent of link channel The detail is at :thesofproject/linux#210 |
| @@ -500,7 +500,8 @@ struct sof_ipc_stream_params { | |||
| enum sof_ipc_stream_direction direction; | |||
| enum sof_ipc_frame frame_fmt; | |||
| enum sof_ipc_buffer_format buffer_fmt; | |||
| uint32_t stream_tag; | |||
| uint32_t host_dma_ch; | |||
| uint32_t be_dma_ch; | |||
There was a problem hiding this comment.
same comment as for the driver, I don't see why the link DMA info be_dma_ch is passed in the ipc structure used for the host.
There was a problem hiding this comment.
thanks, please see my comment in driver side
|
I am fine with a temporary fix, but I don't want to change the ipc definition multiple times to accommodate HDAudio development. This will break all platforms and CI multiple times. If we need additional fields, let's add some padding than can be reclaimed for feature development or quirks upfront. we should do this for most structures btw. |
|
@plbossart ack, lets make a padding update for ipc prior to upstreaming |
host dma and link dma work in decouple mode for SOF + HDA codec.
Now allocate host dma and link dma channel in host and set them
in FW individually.
Signed-off-by: Rander Wang [email protected]