Skip to contentSkip to conversation

ucm2: hdadsp: add basic ucm config

PATCH Superseded Mateusz Gorski 3 participants 10 messages Mar 6, 2020 – Mar 25, 2020

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ucm2: hdadsp: add basic ucm config

Basic UCM configuration for HDA DSP generic enabling codec playback and capture on both HDA codec and DMIC ports.

---
 ucm2/hdadsp/hdadsp.conf | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
 create mode 100644 ucm2/hdadsp/hdadsp.conf
ucm2/hdadsp/hdadsp.conf new file +16 -0
new file mode 100644
index 0000000..4e98545
@@ -0,0 +1,16 @@
+# UCM for Intel CAVS platforms
+# For Audio in HDA and DMIC mode
+
+ValueDefaults {
+	PlaybackChannels "2"
+	PlaybackPriority "1"
+	CaptureChannels "2"
+	CapturePriority "2"
+}
+
+SectionDefaults [
+	cdev "hw:hdadsp"
+	cset "name='codec0_out mo media0_in mi Switch' 1"
+	cset "name='media0_out mo codec0_in mi Switch' 1"
+	cset "name='mch_cap_out mo mch_cap_in mi Switch' 1"
+]
Signature · 1 line
2.17.1
#1 lore raw

Dne 06. 03. 20 v 12:30 Mateusz Gorski napsal(a):

Quoted 2 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

Quoted 13 lines
---
  ucm2/hdadsp/hdadsp.conf | 16 ++++++++++++++++
  1 file changed, 16 insertions(+)

create mode 100644 ucm2/hdadsp/hdadsp.conf

ucm2/hdadsp/hdadsp.conf new file +0 -0
new file mode 100644
index 0000000..4e98545

ucm2/hdadsp/hdadsp.conf line 9

Quoted 9 lines
ucm2/hdadsp/hdadsp.conf new file +9 -0
@@ -0,0 +1,16 @@
+# UCM for Intel CAVS platforms
+# For Audio in HDA and DMIC mode
+
+ValueDefaults {
+	PlaybackChannels "2"
+	PlaybackPriority "1"
+	CaptureChannels "2"
+	CapturePriority "2"
+}

Remove PlaybackChannel and Priority defaults. The priorities should be defined in the device section.

ucm2/hdadsp/hdadsp.conf line 12

Quoted 3 lines
ucm2/hdadsp/hdadsp.conf new file +3 -0
+
+SectionDefaults [
+	cdev "hw:hdadsp"

It should be "hw:${CardId}" for ucm2.

ucm2/hdadsp/hdadsp.conf line 16

Quoted 4 lines
ucm2/hdadsp/hdadsp.conf new file +4 -0
+	cset "name='codec0_out mo media0_in mi Switch' 1"
+	cset "name='media0_out mo codec0_in mi Switch' 1"
+	cset "name='mch_cap_out mo mch_cap_in mi Switch' 1"
+]

This file does not define any devices. Does pulseaudio work with this configuration?

Jaroslav

Signature · 2 lines
Jaroslav Kysela <[email protected]>
Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
author Applied
#2 lore raw
Quoted 5 lines
Quoted 3 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

This file is for Intel Skylake SST driver. Information added in v2.

Quoted 8 lines

ucm2/hdadsp/hdadsp.conf line 9

Quoted 6 lines
ucm2/hdadsp/hdadsp.conf new file +6 -0
+ValueDefaults {
+    PlaybackChannels "2"
+    PlaybackPriority "1"
+    CaptureChannels "2"
+    CapturePriority "2"
+}

Remove PlaybackChannel and Priority defaults. The priorities should be defined in the device section.

Thanks, moved to device section in v2.

Quoted 5 lines

ucm2/hdadsp/hdadsp.conf line 12

Quoted 3 lines
ucm2/hdadsp/hdadsp.conf new file +3 -0
+
+SectionDefaults [
+    cdev "hw:hdadsp"

It should be "hw:${CardId}" for ucm2.

Thanks, applied in v2.

Quoted 6 lines
Quoted 4 lines

+    cset "name='codec0_out mo media0_in mi Switch' 1" +    cset "name='media0_out mo codec0_in mi Switch' 1" +    cset "name='mch_cap_out mo mch_cap_in mi Switch' 1" +]

This file does not define any devices. Does pulseaudio work with this configuration?

Devices added in v2. Regarding pulseaudio - are there any guidelines on how to test UCM files with it? These UCM files were writted mainly based on existing ones for other cards. Of course basic tests were made.

Thanks,
Mateusz

#3 lore raw

Dne 13. 03. 20 v 11:59 Gorski, Mateusz napsal(a):

Quoted 9 lines
Quoted 6 lines
Quoted 3 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

This file is for Intel Skylake SST driver. Information added in v2.

Ok, do we have this code in the vanilla linux kernel? Which .c file? The driver name 'hdadsp' looks suspicious. We usually have a delimiter in the driver name (like sof-hda-dsp).

Quoted 9 lines
Quoted 9 lines

ucm2/hdadsp/hdadsp.conf line 9

Quoted 6 lines
ucm2/hdadsp/hdadsp.conf new file +6 -0
+ValueDefaults {
+    PlaybackChannels "2"
+    PlaybackPriority "1"
+    CaptureChannels "2"
+    CapturePriority "2"
+}

Remove PlaybackChannel and Priority defaults. The priorities should be defined in the device section.

I will check...

Quoted 20 lines

Thanks, moved to device section in v2.

Quoted 5 lines

ucm2/hdadsp/hdadsp.conf line 12

Quoted 3 lines
ucm2/hdadsp/hdadsp.conf new file +3 -0
+
+SectionDefaults [
+    cdev "hw:hdadsp"

It should be "hw:${CardId}" for ucm2.

Thanks, applied in v2.

Quoted 5 lines
Quoted 4 lines

+    cset "name='codec0_out mo media0_in mi Switch' 1" +    cset "name='media0_out mo codec0_in mi Switch' 1" +    cset "name='mch_cap_out mo mch_cap_in mi Switch' 1" +]

Quoted 7 lines
Quoted 2 lines

This file does not define any devices. Does pulseaudio work with this configuration?

Devices added in v2. Regarding pulseaudio - are there any guidelines on how to test UCM files with it? These UCM files were writted mainly based on existing ones for other cards. Of course basic tests were made.

I wrote an ucm validator:

https://github.com/alsa-project/alsa-tests/tree/master/python/ucm-validator

Just type 'make' in the ucm-validator directory, the alsa-info.sh output should be added to the configs subtree.

The PA should be used latest from gitlab with the volume control and HDMI support.

Jaroslav

Quoted 5 lines

Thanks,
Mateusz

Signature · 2 lines
Jaroslav Kysela <[email protected]>
Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
#4 lore raw
Quoted 11 lines
Quoted 9 lines
Quoted 6 lines
Quoted 3 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

This file is for Intel Skylake SST driver. Information added in v2.

Ok, do we have this code in the vanilla linux kernel? Which .c file? The driver name 'hdadsp' looks suspicious. We usually have a delimiter in the driver name (like sof-hda-dsp).

Yes, it is a part of Skylake driver, "hdadsp" is the name of sound card created on machine when using HDA generic machine driver. This machine driver is made of 2 .c files:

    - skl_hda_dsp_common.c
    - skl_hda_dsp_generic.c

both are located in: sound/soc/intel/boards/

Example on production laptop:

test@test-Swift-SF515-51T:/proc/asound$ cat cards
 0 [hdadsp         ]: hda-dsp - hda-dsp
                      WL-SwiftSF515_51T-V1.02-Guinness_WL
Quoted 43 lines
Quoted 10 lines
Quoted 10 lines

ucm2/hdadsp/hdadsp.conf line 9

Quoted 7 lines
ucm2/hdadsp/hdadsp.conf new file +6 -0
+ValueDefaults {
+    PlaybackChannels "2"
+    PlaybackPriority "1"
+    CaptureChannels "2"
+    CapturePriority "2"
+}

Remove PlaybackChannel and Priority defaults. The priorities should be defined in the device section.

I will check...

Quoted 20 lines

Thanks, moved to device section in v2.

Quoted 5 lines

ucm2/hdadsp/hdadsp.conf line 12

Quoted 3 lines
ucm2/hdadsp/hdadsp.conf new file +3 -0
+
+SectionDefaults [
+    cdev "hw:hdadsp"

It should be "hw:${CardId}" for ucm2.

Thanks, applied in v2.

Quoted 5 lines
Quoted 4 lines

+    cset "name='codec0_out mo media0_in mi Switch' 1" +    cset "name='media0_out mo codec0_in mi Switch' 1" +    cset "name='mch_cap_out mo mch_cap_in mi Switch' 1" +]

Quoted 7 lines
Quoted 2 lines

This file does not define any devices. Does pulseaudio work with this configuration?

Devices added in v2. Regarding pulseaudio - are there any guidelines on how to test UCM files with it? These UCM files were writted mainly based on existing ones for other cards. Of course basic tests were made.

I wrote an ucm validator:

Quoted 8 lines

https://github.com/alsa-project/alsa-tests/tree/master/python/ucm-validator

Just type 'make' in the ucm-validator directory, the alsa-info.sh output should be added to the configs subtree.

The PA should be used latest from gitlab with the volume control and HDMI support.

                    Jaroslav

Thanks, used the validator on PATCH v2, it pointed out one more thing to be corrected. One thing that I've noticed - there are some problems when validating "included" files, for example:

<skylake-rt286/Hdmi1.conf>
<skylake-rt286/Hdmi2.conf>

raise "aconfig.AlsaConfigError: unable to load config" error

Mateusz

#5 lore raw

Dne 17. 03. 20 v 12:45 Gorski, Mateusz napsal(a):

Quoted 23 lines
Quoted 9 lines
Quoted 9 lines
Quoted 7 lines
Quoted 4 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and
capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

Quoted 5 lines
Quoted 1 line

This file is for Intel Skylake SST driver. Information added in v2.

Ok, do we have this code in the vanilla linux kernel? Which .c file? The driver name 'hdadsp' looks suspicious. We usually have a delimiter in the driver name (like sof-hda-dsp).

Yes, it is a part of Skylake driver, "hdadsp" is the name of sound card created on machine when using HDA generic machine driver. This machine driver is made of 2 .c files:

    - skl_hda_dsp_common.c
    - skl_hda_dsp_generic.c

Quoted 7 lines

both are located in: sound/soc/intel/boards/

Example on production laptop:

test@test-Swift-SF515-51T:/proc/asound$ cat cards
   0 [hdadsp         ]: hda-dsp - hda-dsp
                        WL-SwiftSF515_51T-V1.02-Guinness_WL

Ok, I see now. The 'hdadsp' is the user configurable card identification (alias to the card number) not the driver name. The UCM should be in 'hda-dsp' directory. If the UCM validator works for you, it should be corrected.

Could you point me to the alsa-info.sh output for this hardware?

Thank you,
Jaroslav

Signature · 2 lines
Jaroslav Kysela <[email protected]>
Linux Sound Maintainer; ALSA Project; Red Hat, Inc.
#6 lore raw
Quoted 27 lines
Quoted 27 lines
Quoted 9 lines
Quoted 9 lines
Quoted 7 lines
Quoted 4 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and
capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

Quoted 5 lines
Quoted 1 line

This file is for Intel Skylake SST driver. Information added in v2.

Ok, do we have this code in the vanilla linux kernel? Which .c file? The driver name 'hdadsp' looks suspicious. We usually have a delimiter in the driver name (like sof-hda-dsp).

Yes, it is a part of Skylake driver, "hdadsp" is the name of sound card created on machine when using HDA generic machine driver. This machine driver is made of 2 .c files:

      - skl_hda_dsp_common.c
      - skl_hda_dsp_generic.c

both are located in: sound/soc/intel/boards/

Example on production laptop:

Quoted 8 lines
Quoted 3 lines
test@test-Swift-SF515-51T:/proc/asound$ cat cards
   0 [hdadsp         ]: hda-dsp - hda-dsp
                        WL-SwiftSF515_51T-V1.02-Guinness_WL

Ok, I see now. The 'hdadsp' is the user configurable card identification (alias to the card number) not the driver name. The UCM should be in 'hda-dsp' directory. If the UCM validator works for you, it should be corrected.

Could you point me to the alsa-info.sh output for this hardware?

Thank you for the explanation, adjustments are coming in v4. Still, I was able to test ucm's on DUT using "alsaucm -c hdadsp" command and it worked..

I had some problems uploading the output automatically, so done it manually, here's the link:

http://www.alsa-project.org/db/?f=986bf4515b2af1de75d42f2df2f812664fb7ec6e

I also sent a patch adding output to configs subtree in alsa-tests repo.

V4 was tested with ucm-validator, no errors.

Quoted 3 lines

Thank you,
                Jaroslav

Thanks,
Mateusz

#7 lore raw
Quoted 49 lines
Quoted 27 lines
Quoted 27 lines
Quoted 9 lines
Quoted 9 lines
Quoted 7 lines
Quoted 4 lines

Basic UCM configuration for HDA DSP generic enabling codec playback and
capture on both HDA codec and DMIC ports.

Could you describe for what Linux driver (source code) is this configuration?

Quoted 5 lines
Quoted 1 line

This file is for Intel Skylake SST driver. Information added in v2.

Ok, do we have this code in the vanilla linux kernel? Which .c file? The driver name 'hdadsp' looks suspicious. We usually have a delimiter in the driver name (like sof-hda-dsp).

Yes, it is a part of Skylake driver, "hdadsp" is the name of sound card created on machine when using HDA generic machine driver. This machine driver is made of 2 .c files:

      - skl_hda_dsp_common.c
      - skl_hda_dsp_generic.c

both are located in: sound/soc/intel/boards/

Example on production laptop:

Quoted 8 lines
Quoted 3 lines
test@test-Swift-SF515-51T:/proc/asound$ cat cards
   0 [hdadsp         ]: hda-dsp - hda-dsp
                        WL-SwiftSF515_51T-V1.02-Guinness_WL

Ok, I see now. The 'hdadsp' is the user configurable card identification (alias to the card number) not the driver name. The UCM should be in 'hda-dsp' directory. If the UCM validator works for you, it should be corrected.

Could you point me to the alsa-info.sh output for this hardware?

Thank you for the explanation, adjustments are coming in v4. Still, I was able to test ucm's on DUT using "alsaucm -c hdadsp" command and it worked..

I had some problems uploading the output automatically, so done it manually, here's the link:

http://www.alsa-project.org/db/?f=986bf4515b2af1de75d42f2df2f812664fb7ec6e

I also sent a patch adding output to configs subtree in alsa-tests repo.

V4 was tested with ucm-validator, no errors.

Quoted 7 lines
Quoted 3 lines

Thank you,
                Jaroslav

Thanks,
Mateusz

Is there anything else I could do regarding this change? I think all of the issues/hints were adressed.

Thanks,
Mateusz

#8 lore raw
In reply to Gorski, Mateusz

Hey,

Quoted 6 lines · Mateusz

Yes, it is a part of Skylake driver, "hdadsp" is the name of sound card created on machine when using HDA generic machine driver. This machine driver is made of 2 .c files:

    - skl_hda_dsp_common.c
    - skl_hda_dsp_generic.c

btw, this machine driver is also used by SOF, but name is prefixed:

- HDA with DSP and SST FW -> "hda-dsp" as the card name
- HDA with DSP and SOF FW -> "sof-hda-dsp" as the card name

And of course without DSP we have also:
- HDA no DSP -> "HDA-Intel" as card name

While we try to share what is possible, the topologies will be different for all three, and thus different UCM files are needed.

Br, Kai

#9 lore raw
Quoted 15 lines

Hey,

Quoted 6 lines · Mateusz

Yes, it is a part of Skylake driver, "hdadsp" is the name of sound card created on machine when using HDA generic machine driver. This machine driver is made of 2 .c files:

    - skl_hda_dsp_common.c
    - skl_hda_dsp_generic.c

btw, this machine driver is also used by SOF, but name is prefixed:

- HDA with DSP and SST FW -> "hda-dsp" as the card name
- HDA with DSP and SOF FW -> "sof-hda-dsp" as the card name

Quoted 7 lines

And of course without DSP we have also:
- HDA no DSP -> "HDA-Intel" as card name

While we try to share what is possible, the topologies will be different for all three, and thus different UCM files are needed.

Br, Kai

UCMs for "sof-hda-dsp" and "HDA-Intel" are already present on the alsa-ucm-conf repo, only the "hda-dsp" one is missing right now. There is also a second e-mail thread with patch adding topology for hda-dsp which is compatible with this UCM configuration.

Thanks,
Mateusz