Skip to content

Conversation

@singalsu
Copy link

This patch lets a capture PCM to operate with deep buffer by taking dsp_max_burst_size_in_ms from topology if it is defined. Earlier the value from topology was omitted for capture.

If not defined, the maximum burst size for capture is set similarly as before to one millisecond.

@singalsu
Copy link
Author

This needs thesofproject/sof#10393 . Still it looks like DMA transfers are 48 frames / 1 ms, so more work needed. Also seems playback DMA transfers are the same, so I'm not sure I'm looking the right thing.

This patch lets a capture PCM to operate with deep buffer
by taking dsp_max_burst_size_in_ms from topology if it is
defined. Earlier the value from topology was omitted for
capture.

If not defined, the maximum burst size for capture is set
similarly as before to one millisecond.

The dma_buffer_size is set similarly as for playback from
largest of deep_buffer_dma_ms or SOF_IPC4_MIN_DMA_BUFFER_SIZE
times OBS.

Signed-off-by: Seppo Ingalsuo <[email protected]>
@singalsu singalsu force-pushed the capture_deep_buffer_enable branch from 7526ff1 to e3a9b28 Compare November 27, 2025 11:48
@singalsu
Copy link
Author

This needs thesofproject/sof#10393 . Still it looks like DMA transfers are 48 frames / 1 ms, so more work needed. Also seems playback DMA transfers are the same, so I'm not sure I'm looking the right thing.

The update patch works, in FW the DMA reload now happens at about every 10th host copier copy. But there are in captured audio small glitches every few tens of ms, that is likely a FW issue.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant