Thank you for the report, with the carefully crafted information.
Can we resume, that *on the same hardware*, sound works as expected on Mageia 8, but not with Mageia 9?
I ask because of:
"1. Install mga8 
 2. Install task-plasma-minimal
 3. Observer missing sound cards."
which one would expect to start with 'Install mga9'.

BTAIM Asigning to Giuseppe who has been dealing with other M9 sound problems, but not quite in this direction; and currently commits PulseAudio.

CC: (none) => lewyssmith
Assignee: bugsquad => ghibomgx

Current PA (16.1-6.mga9) in mga9 includes already latest stuff from git upstream updated to 9 days ago.

Are you running a clean installation from ISO-alpha1 or a MGA8 updated to MGA9?

Beyond what Lewis posted, to exclude there aren't the same problems as of root in #31087 and #30997 please install pipewire-media-session (this will uninstall wireplumber) and after a reboot see if you still have the same problems.

As of hardware check you have installed all the firmware (i.e. alsa-firmware and alsa-sof-firmware) packages.

Q: Can we resume, that *on the same hardware*, sound works as expected on Mageia 8, but not with Mageia 9?

A: Yes, it is the same hardware, a Lenovo T16 AMD thinkpad; my apologies for not thinking through my "how to replicate".  

Q: Are you running a clean installation from ISO-alpha1 or a MGA8 updated to MGA9?  
A: I am following the MGA8 -> Cauldron path:
a.  Install MGA8
b.  Fully patch MGA8.
c.  urpmi.removemedia -a
d.  urpmi.addmedia --distrib --mirrorlist 'http://mirrors.mageia.org/api/mageia.cauldron.i586.list'
e.  urpmi --auto-update --auto

Now I see that iso-alpha1 I will definitely give that a try.  

Lastly, I rebuilt pulseaudio from mga8/updates and the onboard speakers did work.  Nevertheless, I will definitely try the alpha-iso install.

Please do.
We are seeing occasional differences (bugs on one side only) between systems Cauldron upgraded from M8 (like yours), and virgin M9 ISO installations.

Doing an install from mga9-alpha1 and patching resulted in the same thing, only the HDMI devices appeared, no speakers.  Booting from Mageia-9a1-Live-Plasma also resulted in the same thing, no speakers.  

Downloading pulseaudio-14.2-2.1.mga8.src.rpm renaming the source from pulseaudio-14.2.2.1 to pulseaudio-16.2.2.1, rebuilding and doing an RPM -Fvh to the "new" RPMs resulted in working speakers.  The fake version also prevents me from updating to new patched pulseaudio so I'll keep an eye out for newer versions.  

This isn't the best solution but it is a functioning workaround so perhaps this ticket can be reduced in severity. 

Output from pactl list cards now:
pactl list cards | grep -A10 Card
Card #0
        Name: alsa_card.pci-0000_04_00.1
        Driver: module-alsa-card.c
        Owner Module: 6
        Properties:
                alsa.card = "0"
                alsa.card_name = "HD-Audio Generic"
                alsa.long_card_name = "HD-Audio Generic at 0x808c8000 irq 75"
                alsa.driver_name = "snd_hda_intel"
                device.bus_path = "pci-0000:04:00.1"
                sysfs.path = "/devices/pci0000:00/0000:00:08.1/0000:04:00.1/sound/card0"
--
Card #1
        Name: alsa_card.pci-0000_04_00.6
        Driver: module-alsa-card.c
        Owner Module: 8
        Properties:
                alsa.card = "1"
                alsa.card_name = "HD-Audio Generic"
                alsa.long_card_name = "HD-Audio Generic at 0x808c0000 irq 76"
                alsa.driver_name = "snd_hda_intel"
                device.bus_path = "pci-0000:04:00.6"
                sysfs.path = "/devices/pci0000:00/0000:00:08.1/0000:04:00.6/sound/card1"

IMHO you should try the opposite, i.e. merge the older PA tarball from 14.2 to 16.2 spec file, so to keep the latest PA packaging structure. It could be probably something upstream. You might want to get the tarball code from PA at gitlab and see where was the breaking point.

Out of curiosity, I grabbed the latest tarball from gitlab and built it with the 16.1 SPEC file, removing all upstream patches, and sadly the behaviour when running pulseaudio -v was the same as 16.1. 

I have not tried using the 16.1 spec file with the 14.2 tarball, there are significant changes to the spec file that I didn't wish to deal with.

I did try mga9-alpha1-LIVE on my AMD 5700g B450 system and all sound hardware initialized correctly.  This issue seems to be tied to AMD 6850u Rembrandt sound hardware https://www.phoronix.com/news/AMD-Yellow-Carp-Audio-Driver

Thank you for your various trials.
(In reply to Alan Richter from comment #7)
> I did try mga9-alpha1-LIVE on my AMD 5700g B450 system and all sound
> hardware initialized correctly.  This issue seems to be tied to AMD 6850u
> Rembrandt sound hardware
> https://www.phoronix.com/news/AMD-Yellow-Carp-Audio-Driver
You can try any Live ISO on any system without touching it; so you can easily try one on the hardware showing the problem. Your refined conclusion here is worth noting (ERRATA'd).

Keywords: (none) => FOR_ERRATA9

SOLVED!

I saw that Fedora 37 uses pulseaudio 16.1 so I booted my system on a Live Fedora 37 and discovered the onboard speakers worked.  With a little probing I discovered that Fedora was using pipewire-pulse instead of pulseaudio.  I installed pipewire-pulseaudio and observed that pulseaudio was removed, rebooted and all devices showed up in pavucontrol and all sound devices were working correctly.  I did remove my pulseaudio 14.2 RPMs and reinstalled all necessary pulseaudio 16.1-2 RPMs.  

Is pipewire-pulseaudio mature enough to replace pulseaudio or is this something that can be selected during configuration?  

The only odd thing is that my cards are enumerated as 45 and 46:

pactl list cards | grep -A 10 Card
Card #45
        Name: alsa_card.pci-0000_04_00.1
        Driver: alsa
        Owner Module: n/a
        Properties:
                api.acp.auto-port = "false"
                api.acp.auto-profile = "false"
                api.alsa.card = "0"
                api.alsa.card.longname = "HD-Audio Generic at 0x808c8000 irq 75"
                api.alsa.card.name = "HD-Audio Generic"
                api.alsa.path = "hw:0"
--
Card #46
        Name: alsa_card.pci-0000_04_00.6
        Driver: alsa
        Owner Module: n/a
        Properties:
                api.acp.auto-port = "false"
                api.acp.auto-profile = "false"
                api.alsa.card = "1"
                api.alsa.card.longname = "HD-Audio Generic at 0x808c0000 irq 76"
                api.alsa.card.name = "HD-Audio Generic"
                api.alsa.path = "hw:1"

Lastly, should this ticket be closed?

Thank you for the good news.

> I discovered that Fedora was using pipewire-pulseaudio instead of pulseaudio
> I installed pipewire-pulseaudio and observed that pulseaudio was removed
> rebooted and all devices showed up in pavucontrol
> and all sound devices were working correctly
I tested doing this, but it got bogged down in seeming to want to replace loads of 'tainted' pkgs by core ones:
$ LANG=C sudo urpmi --test pipewire-pulseaudio
In order to satisfy the 'matemixer-backend-alsa|matemixer-backend-oss' dependency, one of the following packages is needed:
 1- matemixer-backend-alsa-1.26.0-2.mga9.x86_64: ALSA backend for libmatemixer (to install)
 2- matemixer-backend-oss-1.26.0-2.mga9.x86_64: OSS backend for libmatemixer (to install)
What is your choice? (1-2) 1
Some requested packages cannot be installed:
ffmpeg-5.1.2-3.mga9.x86_64 (in order to keep ffmpeg-5.1.2-3.mga9.tainted.x86_64)
...
xmms2-0.9.1-5.mga9.x86_64 (in order to keep xmms2-0.9.1-5.mga9.tainted.x86_64)

> Lastly, should this ticket be closed?
Leave this to Giuseppe.

CC: lewyssmith => (none)

This should be fixed in pulseaudio-16.1-11.mga9. Apparently it's an upstream bug in the way they reorganized the alsa-ucm (have to remember to report upstream).

I have verified this fix has resolved the problem for Rembrant Audio.  This will definitely improve the user experience when the AMD 7000 series APUs emerge.  

From my perspective, this bug can be closed.

I've opened an upstream bug report for this bug: https://gitlab.freedesktop.org/pulseaudio/pulseaudio/-/issues/3497

(In reply to Alan Richter from comment #12)

> I have verified this fix has resolved the problem for Rembrant Audio.  This
> will definitely improve the user experience when the AMD 7000 series APUs
> emerge.  
> 
> From my perspective, this bug can be closed.

Alan, can you test with this version (where I've included the newer UCM patchset):

https://copr-be.cloud.fedoraproject.org/results/ghibo/mageia9-bonus/mageia-cauldron-x86_64/05736984-pulseaudio/

if the profile is OK?

Giuseppe.

Your updates for pulse audio 16.1.15 worked fine, all devices showed up in pavucontrol and I could switch between HDMI and speaker via Configuration.  I did need to rebuild the RPM for pulseaudio-module-gsettings and pulseaudio-module-bluetooth so that RPM -Fvh would work cleanly.

(In reply to Alan Richter from comment #15)
> Giuseppe.
> 
> Your updates for pulse audio 16.1.15 worked fine, all devices showed up in
> pavucontrol and I could switch between HDMI and speaker via Configuration. 
> I did need to rebuild the RPM for pulseaudio-module-gsettings and
> pulseaudio-module-bluetooth so that RPM -Fvh would work cleanly.

Also headphones?

Headphones are working as expected, plugging in reroutes sound from built in speakers to headphones.