topology2: add an HDA topology with DP SRC on core 2 - #11127
Conversation
Add a topology, similar to the existing HDA SRC one, but with SRC in DP scheduling domain and with it and connected pipelines running on core 2. Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
There was a problem hiding this comment.
Pull request overview
Adds a new development topology target variant for the existing HDA+SRC configuration, allowing the SRC to be placed in the DP scheduling domain and moving the relevant pipelines (including deep-buffer) onto core 2.
Changes:
- Add
DEEP_BUFFER_COREoverride support to the Intel deep-buffer include, wiring it into pipelinecore_idand the pipeline widgetcore. - Add
SRC_CORE/SRC_DOMAINdefaults to the HDA+SRC config, wiringSRC_COREinto pipelinecore_id/pipeline widgetcoreandSRC_DOMAINinto the SRC widgetscheduler_domain. - Register a new development build target
sof-hda-src-generic-dpthat setsSRC_DOMAIN=DPand pins the involved pipelines to core 2.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.
| File | Description |
|---|---|
| tools/topology/topology2/platform/intel/deep-buffer.conf | Adds a core selection variable for deep-buffer playback and applies it to pipeline scheduling. |
| tools/topology/topology2/development/tplg-targets.cmake | Introduces a new dev topology target variant that enables DP SRC and pins pipelines to core 2. |
| tools/topology/topology2/cavs-src-mixin-mixout-hda.conf | Adds SRC scheduling-domain and core selection knobs and applies them to the HDA SRC pipelines. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
| DEEP_BUFFER_PCM_NAME 'Deepbuffer HDA Analog' | ||
| SRC_CORE 0 | ||
| SRC_DOMAIN "default" | ||
| } |
| Define { | ||
| DEEP_BUFFER_CORE 0 | ||
| } |
kv2019i
left a comment
kv2019i
left a comment
There was a problem hiding this comment.
I'd prefer the "FOO_CORE_ID" naming approach as that's used in other topologies. Not blocking though...
There was a problem hiding this comment.
Ok, this moves the whole hw:0,0 pipeline to another core. But the problem we currently have and what #11071 addresses is the DP module alone put to another core than the rest of the pipeline. To accomplish that you should put core_id property to the DP src node. The nocodec topology has possibility to move the DP src nodes around:
Maybe this topology change has some use too, and it looks correct, so I'll approve it. But if this was to address my request, then its not filling it. My hope was to get some CI test to run with a topology where the DP module is on a different core than the rest of the pipeline.
Add a topology, similar to the existing HDA SRC one, but with SRC in DP scheduling domain and with it and connected pipelines running on core 2.