No HDMI audio under 5.19.0-35 & -37 (regression from -32)

[Dave Jones]

[Impact]
For amdgpu/i915, hdmi audio output device has disappeared.

[Fix]
The latest fix for the non-contiguous memalloc helper changed the
allocation method for a non-IOMMU system to use only the fallback
allocator. This should have worked, but it caused a problem sometimes
when too many non-contiguous pages are allocated that can't be treated
by HD-audio controller.

As a quirk workaround, go back to the original strategy: use
dma_alloc_noncontiguous() at first, and apply the fallback only when
it fails, but only for non-IOMMU case

[Test Case]
1. boot with kernel applied the patches.
2. check the cards in /proc/asound/cards.
   Get the hdmi cards.

 0 [PCH ]: HDA-Intel - HDA Intel PCH
                      HDA Intel PCH at 0xe2320000 irq 137
 1 [HDMI ]: HDA-Intel - HDA ATI HDMI
                      HDA ATI HDMI at 0xe2260000 irq 134

[Where problems could occur]
Low, this is just workaround and may have a better solution in the future.

[Misc]
All patches has been landed on OEM-6.1 and lunar.

~~~~~~
CLARIFICATION: Just to avoid any confusion for those coming to this bug report; the "Jammy: invalid" status above does *not* mean that this bug doesn't affect jammy -- it does, and the kernel team is aware of this. All it reflects is that the fix has to go into the kinetic kernel package which will then flow into the kernel-hwe package implicitly.

Currently known affected cards:

* HD 7700 (comment 8)
* R9 290 (comment 21)
* RX 550 (LP: #2012141, and comment 27)
* RX 570 (mine)
* RX 580 (LP: #2009276, and comment 28)
* WX 3200 (comment 29)
* RX 6600 (LP: #2009542)
* RX 6700 (LP: #2009275)

[ Original Description ]

After upgrading my Ubuntu jammy (22.04) desktop to the -35 release of the kernel, I found my HDMI audio output device had disappeared. Reverting to the -32 release caused it to appear again (hence why I'm filing the bug against the kernel rather than pulseaudio). I'm attaching the dmesg output from immediately after booting each kernel, but after a bit of trimming and diffing I *think* the following lines are the salient ones:

input: HDA ATI HDMI HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:03.1/0000:2b:00.1/sound/card0/input12
input: HDA ATI HDMI HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:03.1/0000:2b:00.1/sound/card0/input13
input: HDA ATI HDMI HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:03.1/0000:2b:00.1/sound/card0/input14
input: HDA ATI HDMI HDMI/DP,pcm=9 as /devices/pci0000:00/0000:00:03.1/0000:2b:00.1/sound/card0/input15
input: HDA ATI HDMI HDMI/DP,pcm=10 as /devices/pci0000:00/0000:00:03.1/0000:2b:00.1/sound/card0/input16
input: HDA ATI HDMI HDMI/DP,pcm=11 as /devices/pci0000:00/0000:00:03.1/0000:2b:00.1/sound/card0/input17
snd_hda_intel 0000:2b:00.1: bound 0000:2b:00.0 (ops amdgpu_dm_audio_component_bind_ops [amdgpu])

These lines appear in the dmesg of the -32 kernel, but not in the -35 kernel's log. Meanwhile, the following lines appear in the -35 kernel's log but not in the -32:

hdaudio hdaudioC0D0: no AFG or MFG node found
snd_hda_intel 0000:2b:00.1: no codecs initialized

I'll also attach the output of "pactl list" under each kernel (this shows the HDMI audio sink showing up under -32, but not -35) just in case that helps shed any light on things.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Andy Howell (andy-jjcftv6wldnzq84cs)]

HDMI video does not work for me on this kernel. 5.19.0-31 works fine.

I'm running ubuntu DISTRIB_RELEASE=22.10 on and HP Fury 16 G9 laptop with 12th Gen Intel(R) Core(TM) i9-12950HX.

It also causes the fan to come on. Logging in is very slugging. 5.19.0.31 is very responsive and does not force the fan on under casual use.

[Dave Jones (waveform)]

Looking through the growing list of duplicates, everyone has an AMD GPU so this is almost certainly an issue in that driver. The list of cards affected so far:

* RX570 (mine)
* RX6700 XT (from LP: #2009275)
* RX580 (from LP: #2009276)
* RX6600 (from LP: #2009542)

I'm happy to post any additional logs / traces that would help debug this and/or test patched kernels (worth looking at the duplicates for additional logs too).

[Matthias LEROUX (tokashi)]

Hello.

Same here, with a readeon HD7700 and Ubuntu 22.04.2

With kernel -35 there is no hdmi audio output, but after -32 restore everything is fine.

[Dave Jones (waveform)]

Hi Timo, this also affects jammy -- should I add that to the targeting, or is that to be handled separately?

[koba (kobako)]

@Matthias, Dave, Andy, Would you please help me to bisect it?
i would appreciate a lot.

here's test kernel.
https://drive.google.com/drive/folders/1Hyejgp0r_Dq2gsovXqxB2BNTE8uPEd6g?usp=share_link

* please disable the SecureBoot in bios and run these to install.
  # sudo dpkg -i *.deb

[Dave Jones (waveform)]

@kobako the 5.19.9000-34-generic kernel provided in comment 10 is "bad" (no HDMI audio present, same dmesg lines as seen in the -35 kernel)

[koba (kobako)]

@Dave, thanks
here's another test kernel for bisect
https://drive.google.com/drive/folders/1yRA9fYQg98yEAIeyRg6cQYGboroDEHgb?usp=share_link

[Dave Jones (waveform)]

@kobako the 5.19.9001-34-generic kernel provided in comment 12 is also "bad" (same symptoms, same dmesg lines as seen in the -35 kernel)

[koba (kobako)]

@Dave, appreciate,
I would like to check from first patch in amd driver first. thanks
https://drive.google.com/drive/folders/1bWeHxyBYhukZLic8CmRowDfr0zSjGs-N?usp=share_link

[Dave Jones (waveform)]

@kobako the 5.19.9002-34-generic kernel in comment 14 is also "bad"

[koba (kobako)]

@Dave, thanks your effort.
here's another kernel, reverted this patch.
Would you please also dump the lspci -nvvvv, thanks
~~~
fdcc4c22b7ab) ALSA: hda/hdmi - enable runtime pm for more AMD display audio
~~~
https://drive.google.com/drive/folders/1Ky7Xw6Yt8HVU0YLDbFRJX0ROuJ9O0iVU?usp=share_link

[Dave Jones (waveform)]

@kobako the 5.15.9003-68-generic kernel in comment 16 is "good"; HDMI audio detected and working happily. I'm attaching the requested lspci output.

[Dave Jones (waveform)]

Errr, let me try that again from root so it's actually got the caps output!

[bpl (benoit-planquelle)]

This bug affects also Ubuntu 22.04.2 LTS
# uname -a
Linux xxxx 5.19.0-35-generic #36~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb 17 15:17:25 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

[Dave Jones (waveform)]

@bpl indeed -- I asked Timo if the targeting should be adjusted in comment 9 as I wasn't sure if the kernel team wanted to track this separately over in the hwe package, but given this bug was moved from that package this one I may as well just add jammy to the affected set.

[Simon (sbtswr)]

R9 290 here, HDMI audio output disappeared after updating to 5.19.0-35, loaded 5.19.0-32 and it's working again.

[Timo Aaltonen (tjaalton)]

This does not affect the jammy kernel (5.15), so adding that is misleading. When it's fixed in kinetic, hwe-5.19 in jammy will get the fix automatically.

[Dustin Falgout (lots0logs)]

Still experiencing this with 5.19.0-37-generic. Any ETA on when the fix will make it in?

[Dave Jones (waveform)]

@Timo ah I see, sorry for the noise!

[koba (kobako)]

[phone],
1. latest vanilla kernel,
https://drive.google.com/drive/folders/1fyu43ZWyzL51YHgwRVJeP9K4AnNEPEUv?usp=share_link

2. latest vanilla reverted the suspicious
https://drive.google.com/drive/folders/1YHW7qHXqNQNHZfYHDLLFdPL1-wPezDJ4?usp=share_link
~~~
Revert "ALSA: hda/hdmi - enable runtime pm for more AMD display audio"
This reverts commit fdcc4c22b7ab20e90b97f8bc6225d876b72b8f16.
~~~

3. 5.19.0-35 reverted the suspicious
https://drive.google.com/drive/folders/1MOVbujtJmRpCyqA4nZ7hSToTScscbJ7-?usp=share_link

[Dave Jones (waveform)]

@koba I've tried each of the kernels in comment 25 and I'm afraid it rather refutes the theory that fdcc4c22b is the culprit:

The latest vanilla kernel (#1) is "good"; HDMI audio works happily. That suggests (although not conclusively) that fdcc4c22b isn't the culprit as it presumably includes that commit.

The latest vanilla kernel with fdcc4c22b reverted (#2) is also "good"; HDMI audio works, although I did note dmesg had an awful lot of preempt errors reported in it (which did not appear with the vanilla kernel above).

Finally, just to put the nail in the coffin of fdcc4c22b being the faulty commit, the 5.19.0-35 kernel with fdcc4c22b reverted (#3) is "bad"; exhibits the original failure with HDMI audio being unavailable.

[Duxeu niisan (duxeu)]

To go further wiyh the last remarks, fdcc4c22b commit (https://github.com/torvalds/linux/commit/fdcc4c22b7ab20e90b97f8bc6225d876b72b8f16 is just the addition of a new device hw caps (PCI_DEVICE(0x1002, 0xab30). I doubt it can be the culprit.

I don't understand how the bisect did identify it as the source of the problem. Except if there was some kind of error when patching the ubuntu kernel package to mimic it.

By the way I am also impacted with:
> Baffin HDMI/DP Audio [Radeon RX 550 640SP / RX 560/560X] [1002:AAE0]
(lshw -numeric -class multimedia)
so device 0x1002,0xaae0

[trent-- (sylvainfaivre)]

Hello, I tried all the kernels in this post and I experience the same behaviour as Dave :

5.19.0-31 OK
5.19.0-32 OK
5.19.0-35 bad

5.19.9000-34 (tk) bad
5.19.9001-34 (tk1) bad
5.19.9002-34 (tk2) bad
5.15.9003-68 (tk3) OK

tk_generic_202303131019 = latest vanilla OK
tk_genRevertfdcc4c22b7ab_202303131037 OK
tk_20230313113 (5.19.0-35 rev fdcc4c22b) bad

I am experiencing this problem with DisplayPort audio output on a RX580.
I am attaching the output of "sudo lspci -nvvvvl" as spci-sf.out

[Eric Fredericks (efredericks)]

This is also a problem for me on jammy with -35 / AMD Radeon pro wx 3200 series. When I rolled back to -32 it worked.

[Dave Jones (waveform)]

@duxeu To fill in a bit of background here, the AMD folks were emailed in the background and are of the same opinion.

Personally, I'm not sure the bisect *did* identify that as the commit, but unfortunately bisection isn't an entirely straight-forward process when dealing with packages that have a ton of Ubuntu-specific patches on top of them (as the kernel naturally does).

@kobako Are there any further test kernels to try?

In the meantime I'll stick the currently affected list of cards at the top so new people coming to the bug can easily check against the list (though I have a feeling, looking at the list, it could be summarised as "pretty much all AMD GPUs").

[koba (kobako)]

@dave, I will have a amdgpu and am going to reproduce on my side

[Richard Bartczak (gloster10)]

Also RX560, only to complete the list.

[Ryan Antkowiak (ryanantkowiak)]

I have an AMD/ATI R9 270 and I am also affected. Attached is my full output from "lspci -vvvv"

$ uname -a
Linux dell-xps-8700 5.19.0-35-generic #36~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb 17 15:17:25 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

[Markus (therealdarklord)]

I'm also affected with an iGPU on my amd cpu.

I attached my 'lspci -vvvv'. I'm currently using the working '-32' kernel.

[Serge Gotviansky (burrito-rockville)]

I would like to report that with -32 kernel sound works fine and -35 does not in a more complex configuration. AMD RX 560 is bypassed into a virtual machine using IOMMU. A display port connection is used.

I'm attaching 'sudo lspci -nvvv' for -32 and -35 kernels invoked from the virtual machine.

This is the difference between them:

diff --git a/lspci.32.iommu b/lspci.35.iommu
index e7f7560..058c719 100644
--- a/lspci.32.iommu
+++ b/lspci.35.iommu
@@ -34,9 +34,9 @@
        Capabilities: [48] Vendor Specific Information: Len=08 <?>
        Capabilities: [50] Power Management version 3
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
- Status: D3 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-
+ Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-
        Capabilities: [a0] MSI: Enable+ Count=1/1 Maskable- 64bit+
- Address: 00000000fee09000 Data: 0024
+ Address: 0000000000000000 Data: 0000
        Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel

[koba (kobako)]

[phone]
@Dave and affected guys,
test kernel is based on 5.19.0-37.
I could get the hdmi/audio card against wx3200
~~~
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Lexa XT [Radeon PRO WX 3200] (rev 10) (prog-if 00 [VGA controller])
 Subsystem: Dell Lexa XT [Radeon PRO WX 3200]
---
Linux u-Vostro-3888 5.19.9000-37-generic #38~22.04.1 SMP PREEMPT_DYNAMIC Tue Mar 21 18:06:56 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
 0 [PCH ]: HDA-Intel - HDA Intel PCH
                      HDA Intel PCH at 0xe2320000 irq 137
 1 [HDMI ]: HDA-Intel - HDA ATI HDMI
                      HDA ATI HDMI at 0xe2260000 irq 134
~~~
would you please help to verify this test kernel?thanks
https://drive.google.com/drive/folders/1B9PHrnKgaaM5qkfyS8Cyj40cRIIxcUnw?usp=share_link

the caused commit I found is
~~~
21f5c9a0ed2d - ALSA: memalloc: Don't fall back for SG-buffer with IOMMU
~~~

Applied the fixes patches,
~~~
9d8e536d36e7) ALSA: memalloc: Try dma_alloc_noncontiguous() at first
dd164fbfdc20) ALSA: memalloc: Drop special handling of GFP for CONTINUOUS allocation
a5ed0c547d50) ALSA: vx: Drop superfluous GFP setup
63bfc84672bb) ALSA: pdaudiocf: Drop superfluous GFP setup
97557ec97a24) ASoC: Intel: sst: Switch to standard device pages
~~~

[Richard Bartczak (gloster10)]

uname -r 5.19.9000-37-generic, RX560
refresh rate set to 30Hz, it works (default setting after first boot).
Further test has shown that the volume is significantly lower, compared with "...32".

But if I set the refresh rate to 60HZ, it seems the speech will also speed up, not understandable anymore.
lspci output attached.

[koba (kobako)]

@Richard,
Thanks for feedback.
I can't bootup with -32 on my machine due to nvme issue,
so does the audio behave normally against -32 on your side (speech
wouldn't speed up)?

Checked against 6.2-generic, the speed up issue still persists.

[Richard Bartczak (gloster10)]

@Koba,
Two issues with 37 on my system :
Set refresh rate to 60 Hz, audio will be like chirps
Set refresh rate to 30 Hz, audio will be OK, but the volume will be significantly lower (TV former volume value set to 10-12 (e.g.: youtube), now the TV must be set to 50.).
I did check the volume setting = 100% of Ubuntu.

[koba (kobako)]

@Richard,

How about against -32?
Do you also observe these two issues with -32

[Matt (mangoduck)]

So, it *hasn't* rolled? Curious how one of my 22.04.2 setups got -42 (comment #106) when neither have proposed enabled.

[Richard Bartczak (gloster10)]

After new installation of 22.04.02 and automatically update , got -41, installed after all 5.19.9000-37-generic (see above, bug fixed version)
Yesterday : sudo apt update, sudo apt upgrade, didn`t got -42, proposed disabled.

@Matt, it can be an additional bug on your system, that the proposed enabled/disabled check-box does not work properly, but I don`t know how to verify the status of proposed via command line (perhaps an environment entry ?).

[Matt (mangoduck)]

@Richard, the machine in question is very clean as it's a seldom used entertainment setup, has never had proposed turned on, and I don't remember seeing anything else experimental in the list like I do when I turn it on now. Just a curiosity anyway, not a concern.

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux - 5.19.0-42.43

---------------
linux (5.19.0-42.43) kinetic; urgency=medium

  * kinetic/linux: 5.19.0-42.43 -proposed tracker (LP: #2016503)

  * selftest: fib_tests: Always cleanup before exit (LP: #2015956)
    - selftest: fib_tests: Always cleanup before exit

  * Debian autoreconstruct Fix restoration of execute permissions (LP: #2015498)
    - [Debian] autoreconstruct - fix restoration of execute permissions

  * Kinetic update: upstream stable patchset 2023-04-10 (LP: #2015812)
    - drm/etnaviv: don't truncate physical page address
    - wifi: rtl8xxxu: gen2: Turn on the rate control
    - drm/edid: Fix minimum bpc supported with DSC1.2 for HDMI sink
    - clk: mxl: Switch from direct readl/writel based IO to regmap based IO
    - clk: mxl: Remove redundant spinlocks
    - clk: mxl: Add option to override gate clks
    - clk: mxl: Fix a clk entry by adding relevant flags
    - powerpc: dts: t208x: Mark MAC1 and MAC2 as 10G
    - clk: mxl: syscon_node_to_regmap() returns error pointers
    - random: always mix cycle counter in add_latent_entropy()
    - KVM: x86: Fail emulation during EMULTYPE_SKIP on any exception
    - KVM: SVM: Skip WRMSR fastpath on VM-Exit if next RIP isn't valid
    - can: kvaser_usb: hydra: help gcc-13 to figure out cmd_len
    - powerpc: dts: t208x: Disable 10G on MAC1 and MAC2
    - powerpc/vmlinux.lds: Ensure STRICT_ALIGN_SIZE is at least page aligned
    - powerpc/64s/radix: Fix RWX mapping with relocated kernel
    - uaccess: Add speculation barrier to copy_from_user()
    - wifi: mwifiex: Add missing compatible string for SD8787
    - audit: update the mailing list in MAINTAINERS
    - ext4: Fix function prototype mismatch for ext4_feat_ktype
    - Revert "net/sched: taprio: make qdisc_leaf() see the per-netdev-queue pfifo
      child qdiscs"
    - bpf: add missing header file include
    - wifi: ath11k: fix warning in dma_free_coherent() of memory chunks while
      recovery
    - sched/psi: Stop relying on timer_pending() for poll_work rescheduling
    - docs: perf: Fix PMU instance name of hisi-pcie-pmu
    - randstruct: disable Clang 15 support
    - ionic: refactor use of ionic_rx_fill()
    - Fix XFRM-I support for nested ESP tunnels
    - arm64: dts: rockchip: drop unused LED mode property from rk3328-roc-cc
    - ARM: dts: rockchip: add power-domains property to dp node on rk3288
    - HID: elecom: add support for TrackBall 056E:011C
    - ACPI: NFIT: fix a potential deadlock during NFIT teardown
    - btrfs: send: limit number of clones and allocated memory size
    - ASoC: rt715-sdca: fix clock stop prepare timeout issue
    - IB/hfi1: Assign npages earlier
    - neigh: make sure used and confirmed times are valid
    - HID: core: Fix deadloop in hid_apply_multiplier.
    - x86/cpu: Add Lunar Lake M
    - bpf: bpf_fib_lookup should not return neigh in NUD_FAILED state
    - net: Remove WARN_ON_ONCE(sk->sk_forward_alloc) from sk_stream_kill_queues().
    - vc_screen: don't clobber return value in vcs_read
    - scripts/tags.sh: fix incompatibility with PCRE2
    - usb: dwc3: pci: add support for the Intel Meteor Lake-M
    - USB: serial: option: add suppo...

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the linux-gcp/5.19.0-1024.26 kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-kinetic' to 'verification-done-kinetic'. If the problem still exists, change the tag 'verification-needed-kinetic' to 'verification-failed-kinetic'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Wiktor Tomanek (xses1)]

The bot is going wild again and -0.42 still didn't roll out for Jammy. Wtf? This bug has existed for almost 3 months and the fix was committed almost a month ago, yet it's still broken on the "stable" release.

[Richard Bartczak (gloster10)]

I hope a worst case scenario will be present, nobody did verify linux-gcp/5.19.0-1024.26 kernel and the working bug fix will be removed, all dependent bug-reports will be closed.

What then ? Another flood of bug-reports ? Again the same bug fix, and again user have to test a proposed version : "If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed." ?

I think, Ubuntu Kernel bot has to reflect its software verification procedure, especially the way to test a bug fix with unstable SW (e.g.: #96 "5.19.0-42 - no boot").

[Matt (mangoduck)]

@Richard: Yeah #96 was wiktor again, and he's stated he can't boot *lots* of versions everyone else can. Something's got to be up with his setup.

Would someone with authority please add verification-done-kinetic back to this a *third time* so we can get -42 and have audio? Why is the bot even talking about -1024 in this thread? And why does jammy still say invalid?

[ivanto (ivantu)]

Post #108 ??

[Wiktor Tomanek (xses1)]

@Matt, my setup is a pretty clean install of KDE Neon, so I think I am not the only one affected by this issue while not having the ability to boot different kernels than mainstream. I don't have time to fix this, so all I have left is to bother Ubuntu to push the fix.

[Richard Bartczak (gloster10)]

bug report 2009032 reflects 5.19.0-35
bug report 2009275 reflects 5.19.0-35
bug report 2009276 reflects 5.19.0-35
bug report 2009287 reflects 5.19.0-35
bug report 2009542 reflects 5.19.0-35
bug report 2009542 reflects 5.19.0-35
bug report 2009815 reflects 5.19.0-35
bug report 2011334 reflects 5.19.0-35
bug report 2012141 reflects 5.19.0-35
Ubuntu Kernel Bot reflects :
This bug is awaiting verification that the linux/5.19.0-41.42 kernel ... (5.19.0-35)
This bug is awaiting verification that the linux-nvidia-5.19/5.19.0-1010.10 kernel ...
This bug is awaiting verification that the linux-gcp/5.19.0-1024.26 kernel ...

Ubuntu Kernel Bot, how many verification threads we have to expect ?

[Richard Bartczak (gloster10)]

@Ubuntu Kernel bot :
"ISO 9001:2015 specifies requirements for a quality management system when an organization:

a) needs to demonstrate its ability to consistently provide products and services that meet customer and applicable statutory and regulatory requirements, and

b) aims to enhance customer satisfaction through the effective application of the system, including processes for improvement of the system and the assurance of conformity to customer and applicable statutory and regulatory requirements.

All the requirements of ISO 9001:2015 are generic and are intended to be applicable to any organization, regardless of its type or size, or the products and services it provides."

@Ubuntu Kernel bot, take a look.

[Wiktor Tomanek (xses1)]

Wow, you are really pissed off about that bot. I just want my audio to finally work after months of Ubuntu being stupid. I haven't used my GPU for anything for the last few months.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the linux-allwinner/5.19.0-1012.12 kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-kinetic' to 'verification-done-kinetic'. If the problem still exists, change the tag 'verification-needed-kinetic' to 'verification-failed-kinetic'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Wiktor Tomanek (xses1)]

@Koba, does 5.19.0-1012.12 even include the fix?

[koba (kobako)]

I think the fix must be applied on "allwinner" kernel, 5.29-1012

[Ryan Antkowiak (ryanantkowiak)]

I just got the kernel update this morning:

ryan@dell-xps-8700 [~]$ uname -a
Linux dell-xps-8700 5.19.0-42-generic #43~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Apr 21 16:51:08 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

I now have HDMI audio working again. A great big Thank You to all the developers and testers who worked toward getting this fix in. Wish it would have been resolved sooner, but I realize it's free software, so I take what i can get :)

[Malcolm (milkman-m20ss552)]

I've just run the software updater, downloaded the latest kernel updates etc and after a restart, HDMI is now back and fully functional again.

Happy days are here again :)

Thank you to all concerned, very much appreciated.

Big huggies,

M

[Richard Bartczak (gloster10)]

@Koba, I also confirm via update, upgrade I got -42, and it works as expected.
I must confess, I was a bit afraid, that the audio speed up issue I will get again with -42, but there is no bug. Only if I start with Kinetic installation, the audio speed up issue will be present.
Thank you.

[Wiktor Tomanek (xses1)]

I also confirm. -42 rolled out and it's working. Thx @Koba and everyone else involved.

It feels like fucking Xmas again. All demons in my audio output are gone.

[Richard Bartczak (gloster10)]

@Koba, from my side of view, https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2012830
does not depend on the kernel.

[KAWAKAMI, Kensuke (kenchanprfm)]

-42 works fine on my machine.
Thank you.

[koba (kobako)]

@Richard, that's important information, thanks.
it may be caused other packages and may be caused by pulseaudio/pipewire(im not familiar with audio pieces)

[Matt (mangoduck)]

Got the update but the new kernel wasn't showing up in grub. This fixed it: https://elementaryos.stackexchange.com/questions/28013/new-kernel-boot-option-not-showing-up-in-grub

Finally I can hear again! :)

[Richard Bartczak (gloster10)]

@Koba, please refer to bug report https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2012830, its still assigned to you.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the linux-aws/5.19.0-1027.28 kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-kinetic' to 'verification-done-kinetic'. If the problem still exists, change the tag 'verification-needed-kinetic' to 'verification-failed-kinetic'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[oshunluvr (stuartksmith)]

KDEneon provided 5.19.0-42 this morning and the audio devices reappeared. HOWEVER, system startup sound and Accessibility notifications (like Lock Key toggle sound) do not produce audio.

Rebooting back to the 5.15.0 series restored normal audio feedback.

[Richard Bartczak (gloster10)]

@oshunluvr, please check whether its the same issue as described here :
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2012830.
The described issue does not depend on the kernel. "Audio speed up, it sounds like chirps"

[zzq4343 (zzq4343)]

Updated to 5.19.0-43, still no sound. Reverting to -38 works, -40, -41, -42 and -43 all have no sound through HDMI.
Also, from my end:
- HDMI device is showing in GUI sound settings for both -38 and -43.
- "lspci" prints the same result.
- "cat /proc/asound/cards" prints the same for both -38 and -43 (no HDMI devices, but for -38 sound works)
- "pacmd list-cards" prints the same thing for both -38 and -43
- Connecting to an LG TV and using 4K 120Hz if that helps (tried to lower the resolution and refresh rate, same result, -38 has sound but -43 doesn't)

[Richard Bartczak (gloster10)]

Additional information, obviously there is something wrong. SW will be modified without test purpose, UHD + AMD GPU is not verifiable by Canonical. But why the modification of the SW without possibilities to verify the modification ???
Please confirm to ISO9001 !

[koba (kobako)]

@zzq4343, are you using jammy or kinetic?

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the linux-azure/5.19.0-1028.31 kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-kinetic' to 'verification-done-kinetic'. If the problem still exists, change the tag 'verification-needed-kinetic' to 'verification-failed-kinetic'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Iain Potter (iainp999)]

I'm coming back to this topic just to add what worked for me.

I had to disable Wayland in order to get audio up and running through HDMI again, following the instructions at

https://linuxconfig.org/how-to-enable-disable-wayland-on-ubuntu-22-04-desktop

I'm not sure why this helped, but at the moment I'm just happy that I have HDMI audio back!

Hope this helps someone else.

[Joseph Olstad (josepholstad)]

@ivanto, trying 5.15.0-83 now

[Joseph Olstad (josepholstad)]

ah, actually I was fiddling around with kernel installers and ended up with 6.2.0-33-generic
now I have to figure out how to put 22.04 back on the regular kernels.

[koba (kobako)]

@Joseph,
please try
#5.15
sudo apt install linux-image-generic
#6.2
sudo apt install linux-image-generic-hwe-22.04

[Joseph Olstad (josepholstad)]

joseph@joseph-desktop:~$ sudo apt install linux-image-generic

linux-image-generic est déjà la version la plus récente (5.15.0.84.81).
0 mis à jour, 0 nouvellement installés, 0 à enlever et 0 non mis à jour.

joseph@joseph-desktop:~$ uname -a
Linux joseph-desktop 6.2.0-33-generic #33~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Sep 7 10:33:52 UTC 2 x86_64 x86_64 x86_64 GNU/Linux

I uninstalled linux-image-generic-hwe-22.04 but I'm still stuck on 6.2.0-33-generic, I'll try again

[Joseph Olstad (josepholstad)]

Ok I managed to get back onto 5.15.x,
still no sound options available other than "Sortie fictive" (Fake output). - my OS is in French language.

5.15.0-84-generic #93-Ubuntu SMP Tue Sep 5 17:16:10 UTC 2023

DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=22.04
DISTRIB_CODENAME=jammy
DISTRIB_DESCRIPTION="Ubuntu 22.04.3 LTS"

12:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] Navi 21 HDMI Audio [Radeon RX 6800/6800 XT / 6900 XT]
 Subsystem: Advanced Micro Devices, Inc. [AMD/ATI] Navi 21 HDMI Audio [Radeon RX 6800/6800 XT / 6900 XT]
 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
 Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
 Latency: 0, Cache Line Size: 64 bytes
 Interrupt: pin B routed to IRQ 85
 IOMMU group: 12
 Region 0: Memory at fcb20000 (32-bit, non-prefetchable) [size=16K]
 Capabilities: <access denied>
 Kernel driver in use: snd_hda_intel
 Kernel modules: snd_hda_intel

30:00.1 Audio device: Advanced Micro Devices, Inc. [AMD/ATI] Renoir Radeon High Definition Audio Controller
 Subsystem: Micro-Star International Co., Ltd. [MSI] Renoir Radeon High Definition Audio Controller
 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+
 Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
 Latency: 0, Cache Line Size: 64 bytes
 Interrupt: pin B routed to IRQ 87
 IOMMU group: 16
 Region 0: Memory at fc900000 (32-bit, non-prefetchable) [size=16K]
 Capabilities: <access denied>
 Kernel driver in use: snd_hda_intel
 Kernel modules: snd_hda_intel

Meanwhile, I've tried 3rd party made in china asmedia asm1083 pcix soundcard, it crashes the system after 5 seconds of playing sound. as soon as sound started, crash. Removed the soundcard, no more crashing. This happened with 6.2 kernel, I tried many kernels.

Now I'm on 5.15.0-84-generic (assembled this new computer for my daughter) and with the Rx series video cards (I have 3 of them, 6500xt, 6700xt, 6800xt, no sound when used with the MSI b550 chipset motherboard.

On an older AM3 motherboard that my son has there is sound but I'm not sure if it's the AC97 chipset or the hdmi doing the sound on that computer which is working, haven't checked that.

This MSI B550 running AMD Ryzen 5500 has no sound.

I've ordered a sound blaster audigy card now, hopefully better luck than the knockoff card that claimed to be linux ready.

[Joseph Olstad (josepholstad)]

so I tried a 6500xt radeon, 6700xt, 6800xt, they all didn't like the asmedia asm1083 pcix soundcard. I think the soundcard drivers for the asmedia pcix card I tested out are bad or the card it'self is defective, kernel failure and reboot after 5 seconds of sound.

I purchased and installed this card because the MSI B550 chipset motherboard with any of the video cards I tried has no sound.

I noticed some people reporting better luck with some kernels above. the 5.15.0-84-generic kernel I just tested made no difference in my situation.

[Joseph Olstad (josepholstad)]

Disabling/reenabling Wayland didn't help for my setup. I have seen other reports elsewhere about issues with sound for those using motherboards containing the B550 chipset. The success reported above was encouraging. Knock on wood the soundblaster audigy pcix card I ordered will arrive soon and I'll test that with this chipset.