topology1: Use DYNAMIC for ADL and RPL topologies #8093
Conversation
Vamshigopal
requested review from
plbossart,
ranj063,
lgirdwood,
johnylin76 and
sathyap-chrome
Vamshigopal
force-pushed
the
dynamic-pipeline
branch
from
90e11a4 to
781667a
Compare
There was a problem hiding this comment.
This looks a bit too much of a sledgehammer patch that will break more eggs than intended.
What is the problem we are trying to fix anyways? Why is this necessary and why now?
And there's also a kernel dependency, so there's a risk of breaking existing platforms if they update sof-bin but still use a kernel < 5.19
| "sof-tgl-max98357a-rt5682\;sof-tgl-max98357a-rt5682\;-DCODEC=MAX98357A\;-DFMT=s16le\;-DPLATFORM=tgl\;-DLINUX_MACHINE_DRIVER=sof_rt5682\;-DAMP_SSP=1" | ||
| "sof-tgl-max98357a-rt5682\;sof-adl-max98357a-rt5682\;-DCODEC=MAX98357A\;-DFMT=s16le\;-DPLATFORM=adl\;-DLINUX_MACHINE_DRIVER=sof_rt5682\;-DAMP_SSP=2" | ||
| "sof-tgl-max98357a-rt5682\;sof-adl-max98357a-rt5682-rtnr\;-DCODEC=MAX98357A\;-DFMT=s16le\;-DPLATFORM=adl\;-DLINUX_MACHINE_DRIVER=sof_rt5682\;-DAMP_SSP=2\;-DCHANNELS=2\;-DRTNR" | ||
| "sof-tgl-max98357a-rt5682\;sof-tgl-max98357a-rt5682\;-DCODEC=MAX98357A\;-DFMT=s16le\;-DPLATFORM=tgl\;-DLINUX_MACHINE_DRIVER=sof_rt5682\;-DAMP_SSP=1\;-DDYNAMIC=1" |
There was a problem hiding this comment.
ack, it was added by mistake. I'll correct it.
With Dynamic pipelines enabled, some of the power KPI show better numbers. We discussed with our customer and have agreed to move to dymanic pipeline in a phased manner - this is first set for ADL and RPL programs which are already verified.
We are planning to keep dynamic from cavs2.5 onwards.
our customer kernels for ADL , ADL- N and RPL programs all have the required kernel support too. I believe older kernels will not check for DYNAMIC flag and hence this feature wont be impacted ? |
Vamshigopal
force-pushed
the
dynamic-pipeline
branch
from
781667a to
7b174f9
Compare
|
@plbossart may i know your thoughts ? |
|
IIRC there are dependencies on the kernel to support the DYNAMIC pipelines. I am not sure there is any merit in changing all topologies blindly, clearly you have not tested the Dell/SoundWire topologies so there's at least a very large validation gap.... @ranj063 please chime in. |
I think you made a good point that I just overlooked. Apparently my proposed code lacks many details. Although in my view that is still worth when the new config is targeted to be set massively and can be deduced from the existing ones. I don't mind to add the config case by case in CMakeLists.txt since the deduction logic is a bit complicated. |
Can we target this on case by case, IIUC @plbossart can approve that. |
I don't mind changes to the Chrome topologies if they have been tested and are compatible with the Chrome kernel. |
Vamshigopal
force-pushed
the
dynamic-pipeline
branch
from
7b174f9 to
af1ac3b
Compare
Ack, Updated PR excluding Dell SoundWire topologies. |
use dynamic for all the adl and rpl topologies except 3p(waves,DTS), excluded Dell sdw topologies which are not tested. Signed-off-by: Vamshi Krishna Gopal <[email protected]>
Vamshigopal
force-pushed
the
dynamic-pipeline
branch
from
af1ac3b to
0bf88ce
Compare
|
Thanks @plbossart for suggestions and reviews. |
|
@Vamshigopal I'll proceed with merge. Can you make a backport to stable-v2.2 ? I will soon remove this file from mainline as we no longer support IPC3 for Intel targets in mainline, so any patch that is not in stable-v2.2 is at risk of getting lost. |
ack, #8229 for stable-v2.2 branch |
use dynamic for all the adl and rpl topologies except non 3p(waves,DTS)