Audio: Copier: Add missing DMIC DAI index mapping to create_dai() by singalsu · Pull Request #6668 · thesofproject/sof

singalsu · 2022-11-25T10:53:19Z

The PR contains two commits

Audio: Copier: Add missing DMIC DAI index mapping to create_dai()
Tools: Topology2: Add DMIC16k capture to cavs-nocodec topologies

Edit - I removed the second topology2 commit. Now this is only a FW patch.

singalsu · 2022-11-25T10:58:46Z

The built topology works with a hack in firmware src/lib/dai.c dai_get() to shift the index >> 5. The index from kernel is made with macro

#define SOF_IPC4_NODE_INDEX_INTEL_DMIC(x) (((x) & 0x7) << 5)
It needs to be addressed separately.

Edit: This problem is solved in the first commit.

singalsu · 2022-11-25T11:32:17Z

The built topology works with a hack in firmware src/lib/dai.c dai_get() to shift the index >> 5. The index from kernel is made with macro
#define SOF_IPC4_NODE_INDEX_INTEL_DMIC(x) (((x) & 0x7) << 5) It needs to be addressed separately.

Cool, @juimonen found a solution that copier uses for SSP dai_index. Adding similar handling for DMIC fixes the dai_index issue. It's in the second commit of this PR now.

singalsu · 2022-11-25T11:45:50Z

The modified topology:

ranj063 · 2022-11-25T18:31:17Z

@singalsu can we please wait till #6592 is merged before you add the DMIC16K additions?

ranj063 · 2022-12-01T17:49:40Z

@singalsu my PR is merged. So this one is good for rebase

singalsu · 2022-12-02T10:08:28Z

@singalsu my PR is merged. So this one is good for rebase

Thanks! Done now. I kept the DMIC configuration as 4ch so now there's no change to copier.host.13.1, only add as 4ch copier.host.15.1 and copier.DMIC.16.1.

lgirdwood · 2022-12-02T15:58:58Z

Good find here, I guess this is also needed for mtl-02. @marcinszkudlinski fyi

This fix is necessary. Maybe @singalsu can submit a separate PR for this fix to speed up merging this patch.

libinyang · 2022-12-13T11:55:39Z

This fix is necessary. Maybe @singalsu can submit a separate PR for this fix to speed up merging this patch.

libinyang · 2022-12-13T11:57:10Z

The ppl15 is used by deep buffer in the latest code. We need to refresh the code.

libinyang · 2022-12-13T12:06:26Z

Maybe we can split DMIC0 and DMIC1 into 2 conf files, dmic0-generic.conf for dmic0 and dmic16k-generic.conf for dmic16k.
DMIC1 may be used by other purpose, such as WoV. If we split DMIC0/1 into different conf files. It will be easy to add such support. We just make some judgement in the top file, such as cavs-nocodec.conf to pick up the right configuration, like:
# If there is DMICs, DMIC0 is enabled by default IncludeByKey.NUM_DMICS { "[1-4]" "platform/intel/dmic-generic.conf" } # enable dmic16k if DMIC16 is defined IncludeByKey.DMIC16k { "1" "platform/intel/dmic16k.conf" } # enable dummy_wov if DUMMY_WOV is defined # DUMMY_WOV and DMIC16 are mutually exclusive IncludeByKey.DUMMY_WOV { "1" "platform/intel/dmic-wov.conf" }

Yep. I hope I can work in Q1 next year with tplg2. I would add topology and pipelines those enable testing of all playback and capture processing.

The dai_index needs to be mapped similarly as in kernel. The SSP type was handled correctly in the function but not DMIC. Signed-off-by: Jaska Uimonen <jaska.uimonen@linux.intel.com> Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

singalsu · 2022-12-16T08:58:24Z

@libinyang I removed everything but the essential FW patch. The 16k DMIC DAI cannot work without it. Please review this.

libinyang

LGTM

kv2019i · 2022-12-16T13:43:07Z

@wszypelt This has failures as well, can you check?

wszypelt · 2022-12-16T14:05:32Z

@kv2019i I repeated the tests, 20 minutes and you should get the correct results

This is a fix for a regression introduced by thesofproject#6668. For an unknown reason (probably a bug) Linux and Windows hosts use different range of node_id for DMIC. The fix workarounds the problem adding support for both ranges. Signed-off-by: Serhiy Katsyuba <serhiy.katsyuba@intel.com>

This reverts commit 8d166a4. This restores working DMIC functionality on Windows platform broken by thesofproject#6668. ATTENTION: This will break DMIC1 (DMIC with DAI index 1) functionality on Linux platform, DMIC0 should still work fine! Non working DMIC1 should be fixed on Linux driver side, not this firmware. Signed-off-by: Serhiy Katsyuba <serhiy.katsyuba@intel.com>

This reverts commit 8d166a4. This restores working DMIC functionality on Windows platform broken by #6668. ATTENTION: This will break DMIC1 (DMIC with DAI index 1) functionality on Linux platform, DMIC0 should still work fine! Non working DMIC1 should be fixed on Linux driver side, not this firmware. Signed-off-by: Serhiy Katsyuba <serhiy.katsyuba@intel.com>

singalsu force-pushed the add_cavs_tgl_dmic16k branch from b1ddf8c to a40b3bcCompare November 25, 2022 11:30

singalsu commented Nov 25, 2022
View reviewed changes

Comment threadtools/topology/topology2/cavs/cavs-nocodec.conf Outdated
Comment threadtools/topology/topology2/cavs/cavs-nocodec.conf Outdated

singalsu marked this pull request as ready for review November 25, 2022 11:48

singalsu requested review from aborisovich, dbaluta, lbetlej, lgirdwood, marc-hb, mmaka1 and plbossart as code owners November 25, 2022 11:48

singalsu requested a review from libinyang November 25, 2022 11:48

lgirdwood reviewed Nov 25, 2022
View reviewed changes

Comment threadtools/topology/topology2/cavs/cavs-nocodec.conf Outdated
Comment threadtools/topology/topology2/cavs/platform/intel/dmic-generic-add-dmic1.conf Outdated

marc-hb mentioned this pull request Nov 29, 2022
Convert the SRC component to use the module interface #6639
Merged

singalsu force-pushed the add_cavs_tgl_dmic16k branch from a40b3bc to 632f002Compare December 2, 2022 09:58

singalsu requested a review from lgirdwood December 2, 2022 13:27

lgirdwood approved these changes Dec 2, 2022
View reviewed changes

libinyang mentioned this pull request Dec 13, 2022
topology2: add dummy wov pipeline to nocodec #6782
Merged

libinyang reviewed Dec 13, 2022
View reviewed changes

singalsu force-pushed the add_cavs_tgl_dmic16k branch from 632f002 to d82d526Compare December 16, 2022 08:55

singalsu changed the title ~~Tools: Topology2: Add DMIC16k capture to cavs-nocodec topologies~~Audio: Copier: Add missing DMIC DAI index mapping to create_dai()Dec 16, 2022

singalsu requested a review from libinyang December 16, 2022 08:58

libinyang approved these changes Dec 16, 2022
View reviewed changes

kv2019i approved these changes Dec 16, 2022
View reviewed changes

lgirdwood approved these changes Dec 19, 2022
View reviewed changes

lgirdwood merged commit 8d166a4 into thesofproject:mainDec 19, 2022

serhiy-katsyuba-intel mentioned this pull request Dec 21, 2022
dmic: Fix to support Windows host DMIC node_id range #6872
Closed

serhiy-katsyuba-intel mentioned this pull request Dec 21, 2022
Revert "Audio: Copier: Add missing DMIC DAI index mapping to create_dai()" #6877
Merged

serhiy-katsyuba-intel mentioned this pull request Dec 22, 2022
Revert Audio: Copier: Add missing DMIC DAI index mapping to create_dai() #6881
Merged

singalsu deleted the add_cavs_tgl_dmic16k branch February 2, 2023 15:26

Uh oh!

Conversation

singalsu commented Nov 25, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

singalsu commented Nov 25, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

singalsu commented Nov 25, 2022

Uh oh!

Uh oh!

Uh oh!

singalsu commented Nov 25, 2022

Uh oh!

Uh oh!

Uh oh!

ranj063 commented Nov 25, 2022

Uh oh!

ranj063 commented Dec 1, 2022

Uh oh!

singalsu commented Dec 2, 2022

Uh oh!

lgirdwoodDec 2, 2022

Choose a reason for hiding this comment

Uh oh!

libinyangDec 13, 2022

Choose a reason for hiding this comment

Uh oh!

libinyangDec 13, 2022

Choose a reason for hiding this comment

Uh oh!

libinyangDec 13, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

libinyangDec 13, 2022• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

singalsuDec 16, 2022

Choose a reason for hiding this comment

Uh oh!

singalsu commented Dec 16, 2022

Uh oh!

libinyang left a comment

Choose a reason for hiding this comment

Uh oh!

kv2019i commented Dec 16, 2022

Uh oh!

wszypelt commented Dec 16, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

singalsu commented Nov 25, 2022•
edited
Loading

singalsu commented Nov 25, 2022•
edited
Loading

libinyangDec 13, 2022•
edited
Loading

libinyangDec 13, 2022•
edited
Loading