st-hal: Add support for SVA sound models merging
Added support for merging the SVA GMM sound models to be passed to
ADSP. Two or more different client Application keywords targetted to
the same SVA ADSP engine can be merged and single ADSP session can be
opened, though multiple framework sessions exist from each Application.
STHAL manages merging the incoming sound models and routing detections
to the appropriate framework session. Only GMM sound models for ADSP
are merged, the second stage sound models are not merged.
Change-Id: Ib11146e8f7e3d9e3d660a5a0521502132a87fed8
18 files changed