[stable-v2.2 backport] topology1: Use DYNAMIC for ADL and RPL topologies #8229
Conversation
|
This PR is same as #8093(merged in main), direct cherry-pick had conflicts , so pushed the patch without cherry-pick. |
Vamshigopal
force-pushed
the
dynamic-pipelines
branch
from
d3a0d4d to
4a2bb44
Compare
There was a problem hiding this comment.
This PR is same as #8093 in main), direct cherry-pick had conflicts , so pushed the patch without cherry-pick.
We still want the line from the git cherry-pick -x command in the commit message even when there are conflicts. You must also explain how you resolved conflicts. These are especially important now because we're trying to get rid of some topology1 in the main branch and we must make sure everything has been backported.
The regular git process is:
- run
git cherry-pick -x - fix the conflicts using any editor, git tool or copy you like
- run
git cherry-pick --continue - explain the git conflicts.
This gives you a good commit message with everything required.
Speaking of backports, the easiest way to fix conflicts is often to git cherry-pick -x MORE commits = the missing ones. Have you tried that?
|
@wszypelt https://sof-ci.01.org/sof-pr-viewer/#/build/PR8229/build12816675 has failed. It cannot be because of this pure topology commit. |
yeah, we need to do backport of some other commits for cherry-pick to apply cleanly. |
Sorry I forgot quickbuild is only compatible with the main branch, ignore. It should be disabled elsewhere (PTS-45980), that would also make it much faster! |
|
BTW git cherry-pick -x $rev_expr # see git help revisionsOr, even more powerful:
Can you help? |
It shows 450 patches we need to backport for stable-v2.2 branch from main, to apply the cherry-pick cleanly. git log --oneline --format=%h origin/stable-v2.2..6f2475b -- tools/topology* |
This was just a "semi-random" example sorry. If you find that just a few commits are enough to avoid the current conflicts then maybe go with just that? |
New target device, also move the path over to core 1 when running in no_hotword mode fixed conflict in file tools/topology/topology1/sof-tgl-max98357a-rt5682.m4 by manually adding the change Signed-off-by: Curtis Malainey <[email protected]> Signed-off-by: Pin-chih Lin <[email protected]> (cherry picked from commit 048ffd8)
Add topologies for RTNR and GOOGLE AEC integration Signed-off-by: Curtis Malainey <[email protected]> Signed-off-by: Pin-chih <[email protected]> (cherry picked from commit efd3742)
Add WAVES+AEC topology fixed conflict in tools/topology/topology1/CMakeLists.txt by manually adding changes Signed-off-by: Pin-chih Lin <[email protected]> (cherry picked from commit 058dc69)
For nau8825 codec devices e.g. Kano/Volmar/Zavala, it had been observed that when Hotword module is enabled, noticeable pop noises appear in the playback audio. While disabled, the playback audio becomes clear. Because Hotword module is optional, remove it for now as WA fix for noise issue. Signed-off-by: Pin-chih Lin <[email protected]> (cherry picked from commit 864064d)
…e designs with 10ms period size we have seen power improvements. We are working to move all topologies of Chrome to 10 ms. Starting with current RPL and ADL-N topologies. With many topologies already using GOOGLE_RTC_PROCESSING flag to set 10ms, making a simpler flag to be used across to set required SPEAKER PERIOD. This helps to scale later to all designs. fixed conflict in tools/topology/topology1/CMakeLists.txt by adding changes manually Signed-off-by: Sathya Prakash M R <[email protected]> Signed-off-by: Vamshi Krishna Gopal <[email protected]> (cherry picked from commit a036799)
Adds cases {MAX09360A + RTNR + GOOGLE_RTC_AUDIO} and
{CS35L41 + WAVES}.
fixed conflict in tools/topology/topology1/CMakeLists.txt
by adding changes manually
Signed-off-by: Pin-chih Lin <[email protected]>
(cherry picked from commit f42d015)
Add topology case {WAVES + RT1019 CODEC}
fixed conflict in tools/topology/topology1/CMakeLists.txt
resolved by adding changes manually
Signed-off-by: Pin-chih Lin <[email protected]>
(cherry picked from commit b581aa5)
Adding support for Bluetooth offload.
sof-adl-rt1019-rt5682, and sof-adl-rt1019-rt5682-waves:
rt5682 headphone connects SSP0 link.
rt1019 Amp speakers connects SSP1 link.
bluetooth offload uses SSP2 link.
Signed-off-by: Ajye Huang <[email protected]>
(cherry picked from commit 5aa1545)
Adds new topology build based on sof-adl-max98360a-rt5682-2way with appending arg `-DDMIC_DAI_LINK_16k_PDM=STEREO_PDM1` for existing ADL-P projects. Signed-off-by: Pin-chih Lin <[email protected]> (cherry picked from commit feeb51f)
use dynamic for all the adl and rpl topologies except 3p(waves,DTS), excluded Dell sdw topologies which are not tested. fixed conflict in tools/topology/topology1/CMakeLists.txt resolved by adding changes manually Signed-off-by: Vamshi Krishna Gopal <[email protected]> (cherry picked from commit 6f2475b)
Vamshigopal
force-pushed
the
dynamic-pipelines
branch
from
4a2bb44 to
bef30cc
Compare
Vamshigopal
requested review from
mmaka1,
lbetlej and
dbaluta
as code owners
Backported 10 commits from main branch which are important and necessary to have them in stable-v2.2 branch. let me know your thoughts. |
|
I don't have any valuable thoughts about topologies, only about git and backports :-) From a pure git perspective this looks great! |
marc-hb
changed the title
|
Thanks @Vamshigopal , this looks good. And yes, it seems we have a bit of a problem with TGL/RPL/ADL tplg1 files as the diff between mainline and stable-v2.2 has grown so big. We should have removed the tplg1 files earlier from mainline, to ensure patches are sent to the correct branch, but alas this was not done, so now backporting is more difficult than it should be. |
@Vamshigopal @kv2019i Dell SDW topologies are all cherry-picked to v2.2 branch. No worries. |
use dynamic for all the adl and rpl topologies except 3p(waves,DTS), excluded Dell sdw topologies which are not tested.