topology2: cavs-nocodec: include SSP1 pipelines conditionally #7239
Conversation
|
This patch will require alsatplg change in alsa-project/alsa-utils#195 |
aiChaoSONG
force-pushed
the
mtl_ssp1
branch
2 times, most recently
from
e2cf2b8 to
895db65
Compare
|
Updates:
|
the dependency on alsa-project/alsa-utils#195 prevents any merge from happening. Why are you asking for approvals @aiChaoSONG if there is no agreement on the dependencies and plan to update the docker image? |
|
@plbossart wrote:
I think the alsa-utils #195 dependency is now removed. In this https://github.com/thesofproject/sof/compare/b0b1be1251048b57ce1488085181048b1cdd3aca..e2cf2b866c925859e5db9f827a947c2f425d65c6 |
| } | ||
|
|
||
| # override defaults with platform-specific config | ||
| IncludeByKey.PLATFORM { | ||
| "tgl" "platform/intel/tgl.conf" | ||
| "adl" "platform/intel/tgl.conf" |
There was a problem hiding this comment.
I am having a bit of a heartburn with this PR.
Is the pinmux is a problem on one instance of the MTL RVP, or the MTL SOC in general? It's got to be the former, there is not general incompatibility between SSP1 and DMIC, it's just that the configuration of ONE board that has a restriction.
So rather than having a blanket removal for SSP1, we should only remove SSP1 on specific targets, and use udev rules or a tplg_name kernel option to work-around this limitation.
There was a problem hiding this comment.
@plbossart Let's extinguish the fire that burns your heart, how about generating two topology targets, see updated code.
- sof-mtl-nocodec-ssp0-ssp2.tplg, use this one for MTL RVP, only SSP0 and SSP2 are enabled
- sof-mtl-nocodec.tplg, use this for other MTL device on which all SSPs are enabled.
I think this should address your concern that there are boards have/haven't the pin-mux issue. And configurations are all in the cmake, and only needed if we want to disable SSP1, make it simple also for LNL.
There was a problem hiding this comment.
@plbossart good for you now ?
| @@ -0,0 +1,4 @@ | |||
| # TGL-specific variable definitions | |||
| Define { | |||
| SSP1_ENABLED "true" | |||
There was a problem hiding this comment.
so we are going to have to add SSP1_ENABLED for LNL as well? It'd be smarter to disable SSP1 when needed, fewer configurations to maintain in the end.
There was a problem hiding this comment.
so we are going to have to add SSP1_ENABLED for LNL as well? It'd be smarter to disable SSP1 when needed, fewer configurations to maintain in the end.
@plbossart @aiChaoSONG We're not sure if LNL will have the same pin mux conflict between SSP1 and MIC as MTL, but probably it will. So it's helpful if we can disable SSP1 when needed.
There is pinmux conflict between SSP1 and DMIC on MTL RVP, add a new nocodec topology target for MTL, on which only SSP0 and SSP2 are enabled. Signed-off-by: Chao Song <chao.song@linux.intel.com>
|
One multiple-pause fail in https://sof-ci.01.org/sofpr/PR7239/build4940/devicetest/index.html , but seems to be a case of #7324 so not related to this PR. Proceeding with merge. |
There is pinmux conflict between SSP1 and DMIC on MTL, as there are other SSPs can be used, we don't include SSP1 pipelines only for mtl to unlock DMIC test.