Repository navigation
Adding support for two Lenovo Laptops (P1 & P16) #10087
Description
(Moving from thesofproject/linux#5480)
@bardliao looped us into a conversation - there are a few Lenovo products coming that require quirking in sof_sdw.c and some new topology files.
An issue was the ordering of the arl match table ( https://patch.msgid.link/20250626141841.77780-1-ckeepax@opensource.cirrus.com ), I've just sent a patch to Mark to add a few entries to the sof_sdw_ssid_quirk_table[] to activate the SOC_SDW_CODEC_MIC on these SSIDs ( https://lore.kernel.org/linux-sound/20250702112235.377479-1-simont@opensource.cirrus.com/ )
That done these products then look to load topologies that are new, here are the gathered IDs:
sof-arl-cs42l43-l0-cs35l56-l2-2ch.tplg
[ 9.719802] sof_sdw sof_sdw: create dai link SDW0-Playback-SimpleJack, id 0
[ 9.719805] sof_sdw sof_sdw: create dai link SDW0-Capture-SimpleJack, id 1
[ 9.719807] sof_sdw sof_sdw: create dai link SDW2-Playback-SmartAmp, id 2
[ 9.719808] sof_sdw sof_sdw: create dai link dmic01, id 3
[ 9.719809] sof_sdw sof_sdw: create dai link dmic16k, id 4
[ 9.719810] sof_sdw sof_sdw: create dai link iDisp1, id 5
[ 9.719811] sof_sdw sof_sdw: create dai link iDisp2, id 6
[ 9.719812] sof_sdw sof_sdw: create dai link iDisp3, id 7
sof-arl-cs42l43-l2-cs35l56-l3-2ch.tplg
[ 5.534309] sof_sdw sof_sdw: create dai link SDW2-Playback-SimpleJack, id 0
[ 5.534311] sof_sdw sof_sdw: create dai link SDW2-Capture-SimpleJack, id 1
[ 5.534313] sof_sdw sof_sdw: create dai link SDW3-Playback-SmartAmp, id 2
[ 5.534314] sof_sdw sof_sdw: create dai link dmic01, id 3
[ 5.534314] sof_sdw sof_sdw: create dai link dmic16k, id 4
[ 5.534315] sof_sdw sof_sdw: create dai link iDisp1, id 5
[ 5.534316] sof_sdw sof_sdw: create dai link iDisp2, id 6
[ 5.534317] sof_sdw sof_sdw: create dai link iDisp3, id 7
I made a patch to build them - but testing feedback is that the recording level of the dmic is low. I'm wondering if that is related to building the topology from an old branch (v2.11.1) or whether more cmake magic is required, I know @singalsu was doing some cmake magic recently so wonder if he had any suggestions?
I'll make a PR with the code in it, but the entries I added to tools/topology/topology2/production/tplg-targets-ace1.cmake are (and attached):
"cavs-sdw;sof-arl-cs42l43-l0-cs35l56-l2-2ch;PLATFORM=mtl,NUM_SDW_AMP_LINKS=1,
SDW_AMP_FEEDBACK=false,SDW_SPK_STREAM=Playback-SmartAmp,
SDW_JACK_OUT_STREAM=Playback-SimpleJack,SDW_JACK_IN_STREAM=Capture-SimpleJack,
NUM_DMICS=2,PDM1_MIC_A_ENABLE=0,PDM1_MIC_B_ENABLE=0,DMIC0_ID=3,DMIC1_ID=4,
HDMI1_ID=5,HDMI2_ID=6,HDMI3_ID=7"
"cavs-sdw;sof-arl-cs42l43-l2-cs35l56-l3-2ch;PLATFORM=mtl,NUM_SDW_AMP_LINKS=1,
SDW_AMP_FEEDBACK=false,SDW_SPK_STREAM=Playback-SmartAmp,
SDW_JACK_OUT_STREAM=Playback-SimpleJack,SDW_JACK_IN_STREAM=Capture-SimpleJack,
NUM_DMICS=2,PDM1_MIC_A_ENABLE=0,PDM1_MIC_B_ENABLE=0,DMIC0_ID=3,DMIC1_ID=4,
HDMI1_ID=5,HDMI2_ID=6,HDMI3_ID=7"
0001-topology2-Add-sof-arl-cs42l43-l2-cs35l56-l3-2ch-supp.patch
0002-topology2-Add-sof-arl-cs42l43-l0-cs35l56-l2-2ch-supp.patch
Activity
simontrimmer commentedon Jul 10, 2025
@bardliao & @singalsu , if we can't see anything wrong with the patches I've proffered shall we merge and work out why the DMICs are quiet from there?
bardliao commentedon Jul 11, 2025
@simontrimmer The patches are included in thesofproject/linux#5485. I will merge it if the CI test looks good. @singalsu is currently on vacation. I am not sure if it is the right solution, but can you add
DMIC0_ENHANCED_CAPTURE=true,EFX_DMIC0_TDFB_PARAMS=line2_generic_pm10deg,EFX_DMIC0_DRC_PARAMS=dmic_defaultto your tplg-targets-ace1.cmake and see if it helps?
Like
"cavs-sdw;sof-arl-cs42l43-l0-cs35l56-l2-2ch;PLATFORM=mtl,NUM_SDW_AMP_LINKS=1,
SDW_AMP_FEEDBACK=false,SDW_SPK_STREAM=Playback-SmartAmp,
SDW_JACK_OUT_STREAM=Playback-SimpleJack,SDW_JACK_IN_STREAM=Capture-SimpleJack,
NUM_DMICS=2,PDM1_MIC_A_ENABLE=0,PDM1_MIC_B_ENABLE=0,DMIC0_ID=3,DMIC1_ID=4,
HDMI1_ID=5,HDMI2_ID=6,HDMI3_ID=7, DMIC0_ENHANCED_CAPTURE=true,
EFX_DMIC0_TDFB_PARAMS=line2_generic_pm10deg,EFX_DMIC0_DRC_PARAMS=dmic_default"
And, not related to this issue, but we need to create nhlt bin for debug purpose.
Like
PREPROCESS_PLUGINS=nhlt,NHLT_BIN=sof-arl-cs42l43-l0-cs35l56-l2-2ch.binsimontrimmer commentedon Jul 15, 2025
I've got some unrelated technical problems I'm working through, but I just looked back at this and had a thought - the platform in question is an ARL so should these entries be in the ace1 cmake and should the PLATFORM be something other than mtl?
bardliao commentedon Jul 16, 2025
There is no difference between ARL and MTL on the topology part. So, it is correct that add the arl topologies in the ace1 cmake and set PLATFORM=mtl.
simontrimmer commentedon Jul 16, 2025
Good!
FYI - I sent a top-of-tree build of these topology files with these suggested changes for testing on target yesterday
simontrimmer commentedon Jul 17, 2025
Feedback was good - updated the PR !
bardliao commentedon Jul 18, 2025
@simontrimmer Sorry but which PR are you referring to?
simontrimmer commentedon Jul 18, 2025
@bardliao the branch at least! PR created #10123
singalsu commentedon Jul 21, 2025
Thanks @bardliao @simontrimmer ! Yep, the patch add to enable DRC and beamformer looks good. It enables the capture pre-preprocessing with a generic setup for the algorithms.
@simontrimmer It's also possible to use bespoke tuning for capture and render processing (SOF equalizers, DRC, beamformer) via UCM2 DMI ID based scripts, see https://github.com/alsa-project/alsa-ucm-conf/tree/master/ucm2/blobs/sof and https://github.com/alsa-project/alsa-ucm-conf/blob/master/ucm2/Intel/sof-hda-dsp/dsp.conf . Though the example is only for a very simple UP2 board. The UCM variables to set names vary a bit for HDA and SDW topologies with/without DMIC.