[FEATURE] Clean up SOF DMA interface(s) #9561
Assignees
Labels
Milestone
Comments
kv2019i
added
enhancement
|
FYI @DINESHKUMARK1 -- this shouldn't interfere with work to enable Zephyr (#9549), but something to be aware of. |
|
May be easiest to do a xtos interface removal after we have all ported to Zephyr. |
|
Tagged for v2.12, some parts may be doable by this point, we can break down work if needed. |
This was referenced Dec 10, 2024
|
Work in progress, but won't be complete for 2.12. Feature cutoff for v2.12, moving this to v2.13. |
kv2019i
added a commit
to kv2019i/sof
that referenced
this issue
Dec 20, 2024
Add "sof_" namespace to dma_get()/put() calls in and modify all uses of the interface for builds with native Zephyr drivers. Keep the old name for XTOS builds. This reduces the amount of code that needs to be touched, and in XTOS code, there is no confusion with the namespaces as all dma_foo() calls are to SOF DMA layer. Link: thesofproject#9561 Signed-off-by: Kai Vehmanen <[email protected]>
kv2019i
added a commit
that referenced
this issue
Jan 2, 2025
Add "sof_" namespace to dma_get()/put() calls in and modify all uses of the interface for builds with native Zephyr drivers. Keep the old name for XTOS builds. This reduces the amount of code that needs to be touched, and in XTOS code, there is no confusion with the namespaces as all dma_foo() calls are to SOF DMA layer. Link: #9561 Signed-off-by: Kai Vehmanen <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Is your feature request related to a problem? Please describe.
DMA engines are key part of the SOF and the DMA interface are used in many place (to transfer audio data, to transfer payloads of IPC message, to implement trace, to implement debug features, and so forth).
As described in the commit message and documentation added in commit 816b143 , the transition to use native Zephyr DMA interface has been one of the complex transition tasks, and this work is not complete.
Describe the solution you'd like
This issue is used to track cleanup of the DMA usage in SOF application code with following goals:
Describe alternatives you've considered
One option is to wait for all targets to move to Zephyr (also to native drivers) and then DMA cleanup could be done by dropping all the legacy interfaces.
Additional context
Related to enhancement work done in #9015 and #5794
The text was updated successfully, but these errors were encountered: