fix: remove reference to comp_buffer form sink/src based mod#9400
fix: remove reference to comp_buffer form sink/src based mod#9400kv2019i merged 1 commit intothesofproject:mainfrom
Conversation
When using sink/src API and DP processing, there are no comp_buffer based buffers on inputs and outputs There must not be any usage of comp_buffer in the module this commit changes a forgotten reference to comp_buffer Signed-off-by: Marcin Szkudlinski <marcin.szkudlinski@intel.com>
|
@lyakh any comments ? |
@lgirdwood not really. The commit message seems to suggest that it's wrong in its present form, so is it broken? But we cannot test igo-nr, can we? Maybe @andyross can take a look? |
|
@lyakh yes, it is wrong. The code tries to access to (&sourceb->stream) assuming there's struct comp_buffer there, but in sink/source API there may be no comp_buffer at all. |
andyross
left a comment
There was a problem hiding this comment.
Just noticed I was summoned: this is fine per my understanding too. @marcinszkudlinski is just saying that comp_buffer is no longer an API and that modules should use the arrays of sources/sinks at prepare() time to enumerate their connections.
In this particular case it was degenerate anyway: the only use of the buffer was to get the sample format, which is now an API on source/sink directly. Seems clean as I understand the design.
But also as was mentioned: it's sort of a wart that we have components in-tree that can't be unit tested. I gave some thought to a rig that would unit-test individual processing_modules in-image, though that still requires some per-type code be written to get them configured correctly.
When using sink/src API and DP processing, there are no comp_buffer based buffers on inputs and outputs
There must not be any usage of comp_buffer in the module
this commit changes a forgotten reference to comp_buffer